Respiratory Critical Care

Respiratory Critical Care

Edited by

Craig DavidsonConsultant Physician in Intensive Care and Respiratory Medicine, Guy's and St Thomas' Hospital Trust and

GKT School of Medicine, London, UK

and

David TreacherConsultant Physician in Intensive Care and Respiratory Medicine, Guy's and St Thomas' Hospital Trust and

GKT School of Medicine, London, UK

A member of the Hodder Headline GroupLONDON

First published in Great Britain in 2002 byArnold, a member of the Hodder Headline Group,338 Euston Road, London NW1 3BH

http://www.arnoldpublishers.com

Distributed in the United States of America byOxford University Press Inc.,198 Madison Avenue, New York, NY10016Oxford is a registered trademark of Oxford University Press

© 2002 Arnold

All rights reserved. No part of this publication may be reproduced ortransmitted in any form or by any means, electronically or mechanically,including photocopying, recording or any information storage or retrievalsystem, without either prior permission in writing from the publisher ora licence permitting restricted copying. In the United Kingdom such licencesare issued by the Copyright Licensing Agency: 90 Tottenham Court Road,London W1T4LP.

Whilst the advice and information in this book are believed to be true andaccurate at the date of going to press, neither the authors nor the publishercan accept any legal responsibility or liability for any errors or omissions thatmay be made. In particular (but without limiting the generality of thepreceding disclaimer) every effort has been made to check drug dosages;however it is still possible that errors have been missed. Furthermore, dosageschedules are constantly being revised and new side-effects recognized. Forthese reasons the reader is strongly urged to consult the drug companies'printed instructions before administering any of the drugs recommended inthis book.

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication DataA catalog record for this book is available from the Library of Congress

ISBN 0 340 76289 6

1 2 3 4 5 6 7 8 9 1 0

Commissioning Editor: Joanna KosterProduction Editor: James RabsonProduction Controller: Martin KeransCover Designer: Mouse Mat Design

Typeset in 10/12 Minion and Ocean Sans byIntegra Software Services Pvt. Ltd, Pondicherry, IndiaPrinted and bound in Italy by Giunti

Contents

Colour plate section appears between pages 116 and 117

Contributors vii

Foreword x

Preface xi

1 Respiratory muscles, pulmonary mechanics and ventilatory control 1

Annabel H Nickol and Michael I Polkey

2 Mechanical ventilation: the basics 21

John CGoldstone

3 Mechanical ventilation: ventilatory strategies 32

Hilmar Burchardi

4 Ventilator-patient interaction 47

John CGoldstone

5 Non-invasive mechanical ventilation in acute respiratory failure 58

Bernd Schbnhofer

6 Contemporary issues in critical care physiotherapy 70

Michael Barker, Sheric G Ellum and Sarah EJ Keilty

7 Diagnostic methods in respiratory intensive care medicine 80

Torsten T Bauer and Antoni Torres

8 Monitoring 88

Richard Beale

9 Respiratory emergencies I: Medical 105

Richard M Leach

10 Respiratory emergencies II: Chest trauma, air leaks and tracheostomy 124

Richard M Leach and David A Waller

vi Contents

11 Pathophysiology of acute lung injury 138

S John Wort and Tim W Evans

12 Management of acute lung injury 153

Keith G Hickling and Andrew Bersten

13 Weaning from mechanical ventilation 170

Stefano Nava, Michele Vitacca and Annalisa Carlucci

14 Community-acquired pneumonia 181

Wei Shen Lim and John T Macfarlane

15 Nosocomial pneumonia 192

Jean-Louis Vincent, Baudouin Byl and Daliana Peres Bota

16 Infection in the immunocompromised patient 201

David Ghez, Jean-Francois Timsit and Jean Carlet

17 Pleural disease 217

Wolfgang Frank and Robert Loddenkemper

18 Acute interstitial lung disease 235

Richard Marshall

19 Pulmonary embolism and pulmonary hypertension 249

Graham F Pineo, Russell D Hull and Gary E Raskob

20 Organizational issues in respiratory critical care 263

Adrian J Williams

21 Ethical issues in the intensive care unit 271

Sean P Keenan and William J Sibbald

22 Respiratory failure: new horizons, new challenges 278

A Craig Davidson and David F Treacher

Index 294

Contributors

Michael Barker

Physiotherapy Department, Guy's and St Thomas' Hospital Trust and GKT School of Medicine, London, UK

Torsten T Bauer

Abteilung fur Pneumologie, Allergologie und Schlafmedizin, Medizinische Klinik und Poliklinik,

Bergmannsheil Klinikum der Ruhr-Universitat, Bochum, Germany

Richard Beale

Intensive Care Medicine, Guy's and St Thomas' Hospital Trust and GKT School of Medicine, London, UK

Andrew Bersten

Department of Critical Care Medicine, Flinders Medical Centre, South Australia, Australia

Hilmar Burchardi

Zentrum Anaesthesiologie, Rettungs- und Intensivmedizin, University Hospital, Gottingen, Germany

Baudouin Byl

Infectious Diseases Clinic, Erasme University Hospital, Brussels, Belgium

Jean Carlet

Reanimation Polyvalente, St Joseph Hospital, Paris, France

Annalisa Carlucci

Respiratory Intensive Care Unit, Centro Medico di Pavia IRCCS, Fondazione S. Maugeri, Pavia, Italy

A Craig Davidson

Intensive Care and Respiratory Medicine, Guy's and St Thomas' Hospital Trust and GKT School of Medicine,

London, UK

Sheric G Ellum


Tim W Evans

Royal Brompton Hospital, London, UK

viii Contributors

Wolfgang Frank

Klinik III, Johanniter Krankenhaus im Flaming, Treuenbrietzen, Germany

David Ghez

Reanimation Polyvalente, St Joseph Hospital, Paris, France

John C Goldstone

The Centre for Anaesthesia, The Middlesex Hospital, London, UK

Keith G Hickling

Goldcoast Hospital, Southport, Queensland, Australia

Russell D Hull

Department of Medicine, Thrombosis Research Unit, Foothills Hospital, Calgary, Alberta, Canada

Sean P Keenan

Royal Columbian Hospital, New Westminster, British Columbia, Canada

Sarah EJ Keilty


Richard M Leach

Intensive Care Medicine, Guy's and St Thomas' Hospital Trust and GKT School of Medicine, London, UK

Wei Shen Lim

Respiratory Infection Research Group, Respiratory Medicine, Nottingham City Hospital, Nottingham, UK

Robert Loddenkemper

Lungenklinik Heckeshorn, Berlin, Germany

John T Macfarlane

Respiratory Medicine, Nottingham City Hospital, Nottingham, UK

Richard Marshall

Centre for Respiratory Research, University College London, Rayne Institute, London, UK

Stefano Nava

Respiratory Intensive Care Unit, Centro Medico di Pavia IRCCS, Fondazione S. Maugeri, Pavia, Italy

Annabel H Nickol

Respiratory Muscle Laboratory, Royal Brompton Hospital, London, UK

Daliana Peres Bota

Department of Intensive Care, Erasme University Hospital, Brussels, Belgium

Graham F Pineo

Department of Medicine, Thrombosis Research Unit, Foothills Hospital, Calgary, Alberta, Canada

Contributors ix

Michael I Polkey


Gary E Raskob

Departments of Biostatistics and Epidemiology, and Medicine, University of Oklahoma Health Sciences Center,

Oklahoma City, Oklahoma, USA

Bernd Schonhofer

Krankenhaus Kloster Grafschaft, Zentrum fur Pneumologie, Beatmungs- und Schlafmedizin,

Schmallenberg-Grafschaft, Germany

William J Sibbald

Department of Medicine, Sunnybrook and Women's Health Sciences Centre, Toronto, Ontario, Canada

Jean-Francois Timsit

Reanimation medicate et infectieuse, Bichat Hospital, Paris, France

Antoni Torres

Hospital Clinic i Provincial, Serve! de Pneumologia i Allergia Respiratoria, Barcelona, Spain

David F Treacher

Intensive Care and Respiratory Medicine, Guy's and St Thomas' Hospital Trust and GKT School of Medicine,

London, UK

Jean-Louis Vincent

Department of Intensive Care, Erasme University Hospital, Brussels, Belgium

Michele Vitacca

Respiratory Intensive Care Unit, Istituto Scientifico di Gussago, Fondazione S. Maugeri, Pavia, Italy

David A Waller

Glenfield Hospital, Leicester, UK

Adrian J Williams

Lane-Fox Respiratory Unit, Guy's and St Thomas' Hospital Trust and GKT School of Medicine, London, UK

S John Wort


Foreword

Intensive care is increasingly being recognized ona worldwide basis as a specialty within its ownright, and the appropriate training programmes anddiplomas have now been introduced in manycountries. It is quite clear to many of us within theintensive care community that a multi-disciplinaryapproach to the care of the critically ill patientprovides better care in every sense of the word.

Organ dysfunction is, of course, the majorproblem with most intensive care patients, andpulmonary dysfunction is particularly common.Patients range from those with severe obstructiveairway disease, where the prime objective might beto try and prevent intubation and ventilation, tothose with severe acute respiratory distresssyndrome who are profoundly hypoxic and require

urgent ventilation using the most up-to-datetechniques. Within this spectrum lies a largenumber of complex issues which require veryconsiderable expertise.

This book covers these issues with particularconciseness but without losing anything in clarity.The authors are internationally recognized as expertsin their fields and have produced an authoritativetext that can and should be used by all physicianswho are interested in the management of patientswith acute respiratory problems in the intensive andhigh dependency environments.

David BennettProfessor of Intensive Care Medicine

St George's Hospital, London, UK

Preface

Major developments in respiratory critical care haveoccurred in the past decade. These have resultedin important changes in the management of acuterespiratory problems, both within the intensive careunit and in other acute care areas. The improvedunderstanding of patient - ventilator interactions,advances in ventilator technology and respiratorymonitoring and the recognition of ventilator-inducedlung injury have all produced significant changes inclinical practice. Non-invasive ventilation has becomeestablished as the preferred option in respiratoryfailure resulting from chronic obstructive pulmonarydisease (COPD) and can be used to speed weaningfrom invasive ventilation in this condition. Its use hasbeen extended to the management of non-COPDrespiratory failure and it should now be used, at leastinitially, in immunocompromised patients presentingwith acute respiratory failure.

Other areas of change include new strategies forventilation in acute lung injury and weaning frommechanical ventilation following critical illness. Usingmore precise classification of patient populations anddefinition of ventilatory strategies, randomized trialshave resulted in important changes in practice in boththese areas. The science of aerosol delivery has alsobecome established. New therapies have come, othershave gone and others have yet to find wide applica-tion. The initial hopes for extracorporeal membraneoxygenation and nitric oxide in refractory hypox-aemia have not been realized, and prone positioning,although increasingly practised, was not shown toprovide any outcome benefit in a multicentre study.Simpler, less dramatic strategies, such as nursingintensive care unit patients in the semi-recumbentposition, have been shown to reduce the incidence ofnosocomial pneumonia. Finally, cost-effectivenessand ethical dilemmas are being more widely debatedand, in certain cases, will influence decisions concern-ing the appropriateness of admission to the high

dependency or intensive care unit as institutionsattempt to justify, and governments to quantify, thecost of caring for these patients.

Against this exciting background of clinicalprogress, the training and provision of critical caremedicine is changing. Historically, in the USA, itsdevelopment has been linked to respiratory medicineand more than 80% of trainees opt for dual accredi-tation with critical care. In Australia, on the otherhand, it has emerged as a separate specialty and isincreasingly recognized as such in Europe. In the UK,intensive care has traditionally been dominated innumber by anaesthetists, who often have, in addition,a full anaesthetic practice with relatively few formalsessions available for intensive care. There is nowrecognition that intensive care should be a separatespecialty, with sufficient dedicated sessions to ensurequality of care and to permit training of juniordoctors from other medical disciplines. In Americaand many European countries, intensive care isorganized by specialty rather than by having generalor mixed intensive care units. The re-creation ofrespiratory intensive care units, in which bothintubated and non-invasively ventilated patients aremanaged, has seen greater involvement of respiratoryphysicians in the delivery of care. These changes havebroadened the appeal of critical care and, in Europeand in the UK, respiratory physicians are becomingmore involved.

For all these reasons, there is now a need for themajor developments in respiratory critical care to bereviewed and for a text that provides a balanced clin-ical approach with state-of-the-art commentary. Thecontributors to this book are both acknowledgedexperts and practising clinicians, with a wide spreadof cultural backgrounds to reflect the internationalnature of medicine in the twenty-first century.Excessive referencing has been avoided in favour ofreadability and, inevitably, some readers will feel that

xii Preface

important topics have been either omitted or dealt their background or interests. Most particularly, thiswith superficially. Standard textbooks already book is addressed to respiratory physicians, who, weprovide excellent reviews of cardiopulmonary physi- believe, are now recognized as increasingly import -ology, the management of the difficult airway or the ant and necessary contributors to the practice andfiner details of blood-gas analysis. By omitting these development of critical care medicine.subjects, we have been able to focus on the areas inwhich new evidence is driving changes in clinical Craig Davidson and David Treacherpractice. We hope that this book will appeal to allthose working or training in critical care, whatever July 2002

1Respiratory muscles, pulmonarymechanics and ventilatory controlANNABEL H NICKOL AND MICHAEL I POLKEY

Introduction

The respiratory muscle pump

Pulmonary mechanics

1 Respiratory drive and control

1 Conclusion

7 References

14

18

18

INTRODUCTION

Passage of air into and out of the lungs is essentialfor the maintenance of O2 and CO2 homeostasis. Inspontaneously breathing humans this is achieved bycontraction of the respiratory muscles. The activity ofthese muscles is governed at both a voluntary and aninvoluntary level by specific areas of the brain, anddisease processes that interfere with either ventilatorycontrol or the respiratory muscle pump (Fig. 1.1)may cause ventilatory failure or difficulty in weaningfrom mechanical ventilation. Similarly, if the loadplaced on the respiratory muscle pump exceeds itscapacity (even if the pump is normal), ventilatoryfailure results (Fig. 1.2). Different diseases impose dif-ferent types of load on the respiratory muscle pumpand understanding these differences is important.

In this chapter, the respiratory muscle pump, thecontrol of breathing, respiratory muscle fatigue andinteractions with sleep are reviewed in both normalsubjects and patients with respiratory failure.Methods for testing the function of the respiratorymuscle pump and the influences of various condi-tions on pulmonary mechanics are considered andcritical care conditions that affect the respiratorymuscle pump are also reviewed.

THE RESPIRATORY MUSCLE PUMP

Anatomically, the muscles forming the respiratorymuscle pump may be considered as either inspira-tory (diaphragm and extradiaphragmatic inspiratorymuscles) or expiratory, of which the abdominalmuscles are the most important. Histologically, therespiratory muscles are similar to skeletal muscle,with approximately 50% Type I 'slow' fibres and 25%each of Type Ha and Type IIb, and are therefore suscep-tible to the same physiological processes, includingthe development of fatigue. Similarly, diseases thataffect skeletal muscle or its innervation may alsoinvolve the diaphragm and occasionally patients withthese diseases present with respiratory failure.

The diaphragm deserves special considerationbecause it is the most important respiratory muscle,accounting for approximately 70% of resting ventila-tion. Moreover, because the phrenic nerves that sup-ply it have a long course from their origin in the neckfrom the 3rd, 4th and 5th cervical roots, the nervesmay be damaged by a variety of diseases, trauma oriatrogenically. As well as being uniquely vulnerable,the diaphragm is the only respiratory muscle that hasa nerve supply, with surface landmarks allowing it tobe stimulated in isolation and in which tension

2 Respiratory muscles, pulmonary mechanics and ventilatory control

Figure 1.1 Schematic representation of the respiratory muscle pump. On the left are examples of disease processes that occur in critically ill

patients and can compromise the pump. (Adapted from reference 37.)

generation can be estimated (as transdiaphragmaticpressure, Pdi). When the diaphragm contracts, itmoves caudally, creating a negative intrathoracicpressure and a positive abdominal pressure. Thispressure is transmitted laterally through the zone ofapposition to the lower ribcage, causing outwarddisplacement of the lower ribs. The function of theextradiaphragmatic inspiratory muscles (for examplethe scalenes and parasternal intercostals) during res-piration in normal humans is to prevent collapse ofthe upper ribcage. Consequently, in patients withunopposed diaphragm activity, such as hightetraplegics fitted with diaphragm pacers, diaphragmcontraction results in an inward, expiratory move-ment of the upper ribcage.

The abdominal muscles have important functionsin relation to coughing, laughing and speaking. Inparticular, an effective cough depends on achieving

dynamic airway closure, which, in turn, depends onthe development of a gastric pressure greater than 50cmH2O.1 The contribution of the abdominal musclesto ventilation is more controversial; in spontaneouslyventilated subjects without airflow limitation, thesemuscles are recruited during movement from thesupine to the erect posture and when minute ventila-tion is increased for any reason, for example duringexercise. The mechanism of action is thought to bethat expiratory muscle activity drives the diaphragmto a lower lung volume and functional residualcapacity (FRC); this in turn assists the subsequentinspiration because gravity assists the inspiratorydescent of the diaphragm, thereby increasing its lengthand the force it generates.2 However, this mechanismcannot operate in patients who are supine or in thosewho are flow limited at rest, as are the majority ofpatients with chronic obstructive pulmonary disease

The respiratory muscle pump 3

Figure 1.2 Load-capacity imbalance of the respiratory muscle pump results in ventilatory failure or, if the patient is already receiving

ventilatory support, difficulty weaning. CINMA, critical illness neuropathy and myopathy.

(COPD) requiring mechanical ventilation. In thesecircumstances, abdominal muscle action may becounterproductive, both because the energy expendedis wasted and because it may contribute to patient-ventilator asynchrony.3

Pathophysiological processes affectingthe respiratory muscle pump

DISEASE PROCESSES

Respiratory muscle dysfunction due to neurolo-gical disease may precipitate respiratory failure;4

important causes are shown in Table 1.1. Manyof these conditions can be excluded by clinicalhistory or simple measurements. Myasthenia gravismerits particular review. In addition to thera-peutic approaches to modify disease activity(such as thymectomy or steroid therapy), patientswith myasthenia gravis are usually treated withanticholinesterases. This therapy affects musclegroups differentially and patients with apparentlyoptimally controlled myasthenia gravis may havesignificant respiratory muscle weakness.5 Suchpatients are predisposed to acute ventilatory failureif the dose of anticholinesterase is either too low ortoo high.

Table 1.1 Neurological causes of acute respiratory failure

Trauma to nerve or highcervical spine

Sedative drugs- prescribed

Overdose (narcotic or other)

Guillain-Barre

Organophosphate poisoning

Botulism

Envenomation/shellfishpoison

Drugs with neuromuscularblocking effects(as main or side effect)

Myasthenia gravis

Lambert-Eaton syndrome

Biochemical disturbance,e.g hypokalaemia

Periodic paralysis


MUSCLE SHORTENING

Like all skeletal muscles, the diaphragm and otherrespiratory muscles have an optimum length,defined by the length at which a given stimulus gen-erates the greatest tension. For the humandiaphragm, the optimum length (usually measuredas lung volume) has not been determined in vivo, butit must be below FRC because numerous studieshave established that the pressure-generating capa-city of the diaphragm increases between total lungcapacity (TLC) and FRC (Fig. 1.3). Importantly, inboth normal subjects and patients with COPD,6

the reduction is primarily in the capacity ofthe diaphragm to lower intrathoracic pressure. As wellas pre-existing COPD, lung volume is increased inthe intensive care unit (ICU) by acquired obstructivedefects and the application of extrinsic positiveend expiratory pressure/continuous positive airwaypressure (PEEP/CPAP).

MUSCLE FATIGUE

If skeletal muscle is subjected to increased load, areduction in tension generation occurs, whichresolves with rest; this process is termed fatigue.Failure of neural output is termed central fatigue, butthis is impossible to differentiate in vivo from lack ofmotivation. Fatigue may also result from defects aris-ing at the neuromuscular junction, e.g. myastheniagravis or neuromuscular blockade. However, theform of fatigue thought to be of greatest relevance tothe critically ill patient is low-frequency fatigue,because it is long lasting (24 hours or more) and

in-vivo respiratory motoneurons discharge at lowfrequencies. For skeletal muscle, the tension gener-ated increases with increasing stimulation frequency,reaching a plateau at around 100 Hz. In low-frequency fatigue, the tension generated at low-frequency stimulation (10-20 Hz) is reduced, butthere is little reduction at higher frequencies. Low-frequency diaphragm fatigue has been demonstratedin normal subjects after voluntary hyperventilationand exhaustive treadmill exercise. However, evidencethat low-frequency diaphragm fatigue contributes toventilatory failure in clinical practice is thus farlacking and attempts to produce it in stable patientswith COPD have failed.

Does the respiratory muscle pumpreally fail?

There are no clear-cut 'markers' of respiratory musclepump failure, although failure is suggested by a rise inPaCO2 without a fall in PaOr There have as yet beenno serial studies that have demonstrated a decrease instrength as a patient develops respiratory failure, orthat respiratory muscle strength per se distinguishespatients failing and succeeding in a weaning trial.Measurable changes in respiratory muscle physiologysuch as slowing of the maximal relaxation rate or adecrease in the ratio of high-frequency to low-fre-quency electromyogram (EMG) signal merely reflectthe fact that the muscle is loaded. Respiratory failureoccurs as the result of an unfavourable loadxapacityratio,7 an increase in respiratory load being the more

Lung volume relative to FRC (% VC)

Figure 1.3 Effect of lung volume change (VC) relative to functional residual capacity (FRC) on the pressure generated in response

to a single bilateral supramaximal stimulation of the phrenic nerves. Mean data from eight subjects are shown. Pga, gastric pressure;

Pdi, transdiaphragmatic pressure; Pes, oesophageal pressure. (Modified from reference 38.)

The respiratory muscle pump 5

usual reason for ICU admission or the need for con-tinuing ventilation. Indeed, laboratory studies invol-ving inspiratory loading have demonstrated thatrespiratory muscle fatigue develops when the meaninspiratory capacity during each breath becomes ahigh proportion (> 15-20%) of maximum inspira-tory pressure. Similarly in the ICU, long-term ventila-tor patients have been shown to fail a weaning trialwhen the oesophageal pressure required to achieveadequate ventilation is a large fraction of maximuminspiratory pressure.

Sleep

The normal physiological changes that occur in sleephave particular relevance in those with significantweakness of the diaphragm, in obstructive lungdisease, congestive cardiac failure and in the pro-foundly obese. In the context of the acutely unwell,who may have incipient respiratory failure, it may bea critically important time as respiratory failure mayworsen, with episodes of severe hypoxaemia and thepotential for cardiac arrhythmias or even cardiores-piratory arrest. Sleep will be also important duringthe weaning period, not only because of the sleepfragmentation that often occurs in the ICU,8 but alsobecause it is a risk period for the recently extubated.Inadequate alveolar ventilation may occur duringweaning, when spontaneous modes of ventilatorysupport may provide insufficient ventilatory control.

The features of sleep-disordered breathing (SDB)include hypoventilation, with resulting respiratoryacidosis, hypoxaemia and recurrent arousal.9 Arousalwill, at times, also produce, profound sympatheticand parasympathetic activation. SDB results fromboth obstructive and non-obstructive (central) sleepapnoea, a mixture of the two or a greater than nor-mal diminution in alveolar ventilation, particularlyduring rapid eye movement (REM) sleep. Alveolarhypoventilation results from either chest-wall orneuromuscular disease, when there is a reduction inthe 'power' of the respiratory pump, or when theload is increased, e.g. COPD. It is a consequence ofthe normal reduction in striated muscle activity atsleep onset, with a further reduction during REM,which leads to significant hypoventilation with theloss of the accessory muscles' contribution to venti-lation. A reduction in tone of the pharyngeal musclesmay also promote obstructive events. Additional

changes in chest or abdominal wall compliance,only partially explained by body position, may alsoaffect VQ and lead to hypoxaemia. Arousal, which istypically recurrent in OSA, results in sympatheticstimulation, the importance of which is increasinglybeing recognized in congestive cardiac failure as con-tributing to sympathetic activation. Monitoring ofrespiratory function during sleep is therefore impor-tant in several risk groups: the acutely unwell, non-ventilated patient; the at-risk neuromuscular orCOPD patient, especially during weaning; andrecently extubated patients who may be at more riskof upper airway obstruction.

Acquired damage to the respiratorymuscle pump

CRITICAL ILLNESS NEUROMUSCULARABNORMALITIES

Neurological abnormalities are common in patientsin whom weaning is difficult. Spitzer et al., forexample, concluded that, in 'difficult to wean'patients, 62% had neuromuscular disease sufficientlysevere to account for ventilator dependency.10 Theidentification of previously unsuspected neurologi-cal disease is therefore important. Full neurologicalexamination of the ICU patient is difficult, but itshould still be possible to identify muscle wasting, fas-ciculation and the presence or absence of tendonjerks. Preservation of tendon reflexes is importantas it demonstrates retained motor nerve function.11

In some cases, a demyelinating neuropathy occurs,which may be considered an acquired Guillain-Barresyndrome. An EMG may be helpful, althoughmyopathy can be difficult or impossible to distin-guish on electrophysiological grounds from anaxonal neuropathy. That myopathy, rather thanneuropathy, occurs in some patients is supported byhistological and biochemical data.12

Only a few studies have investigated the electro-physiology of the respiratory muscles in ICU andnone has systematically assessed respiratory musclestrength. The frequency of reported abnormalities ishigh and does not have a straightforward relation-ship with the frequency of abnormalities of theperipheral nervous system. Neuromuscular abnor-malities of the respiratory muscles and peripheralmuscles frequently co-exist. Moreover, patients with


critical illness axonal polyneuropathy involving non-respiratory nerves are likely to require longer periodsof ventilatory support than those without.13

The causes of critical illness neuromuscular abnor-malities (CINMA) are not well established (for a fullerdiscussion, see references 14 and 15), but multipleorgan dysfunction is a recognized risk factor. Bothneuromuscular blocking agents and corticosteroidshave been implicated in the aetiology, but CINMAcommonly occurs without exposure to these drugs.12

In renal failure, the accumulation of active drug ormetabolite such as 3-desacetyl-vecuronium can occur,leading to persistent neuromuscular failure.

IATROGENIC DAMAGE TO THE RESPIRATORYMUSCLE PUMP

Phrenic nerve injury is a recognized complication ofsurgery to the heart, liver or upper gastrointestinal(GI) tract and central venous cannulation. Chest-wall pain and upper GI surgery may also impairdiaphragm function.

Assessment of the respiratory musclepump in the intensive care unit

The function of the inspiratory muscles is to producean intrathoracic pressure below atmospheric pres-sure so that inspiration occurs. Theoretically, thepump can be assessed at any level from the cortex toflow in the respiratory airways. Although measure-ments of tidal volume and vital capacity broadlyindicate whether a patient has sufficient respiratoryfunction to avoid progressive ventilatory failure,their value in the detailed assessment of pump func-tion is limited because they are influenced by lungmechanics. However, in patients with isolated respira-tory muscle disease, changes in vital capacity areuseful in predicting the need for ventilation andin evaluating recovery.

PRESSURE MEASUREMENTS

In ambulant patients, the pressure developed at themouth or in the oesophagus during a maximalvoluntary effort is often used as a measure of inspira-tory muscle strength. This method has been adaptedfor use in the ICU by using a valve that only permitsexpiration, but the test fails to predict weaning out-come, presumably because patients in the ICU are

seldom able to make a truly maximal voluntaryeffort. Clearly, patients who can generate a high pres-sure do not have respiratory muscle weakness, butthis seldom applies to ICU patients.

To measure respiratory muscle strength independ-ent of patient effort, it is necessary to stimulatethe nerve supplying the muscle artificially, usingelectrical or magnetic stimulation, and measure theforce output. The only muscle in which this canbe performed in vivo is the diaphragm, but becauseit accounts for approximately 70% of resting ventila-tion in humans, this is useful when respiratorymuscle weakness is suspected.

Phrenic nerve stimulation allows the force outputof the diaphragm to be measured independently ofpatient effort. It is quantified by the transdiaphrag-matic or mouth/endotracheal tube pressure (Pdi orPm/PET, respectively) generated in response to a sin-gle supramaximal stimulus applied to both phrenicnerves, a 'twitch' (Tw). The measurement of Tw Pdirequires the use of oesophageal and gastric balloons,which is not always possible in intubated patients. Analternative is to measure the Tw PET by occluding theendotracheal tube at end-expiration. Although thisapproach has clear attractions, Tw PET is similar tooesophageal rather than to transdiaphragmatic pres-sure and, being 50-60% smaller, may be harder tomeasure accurately because the 'noise-to-signal' ratiowill be large. Similarly, increases in lung volume, forexample with PEEP, disproportionately influence thevalue.

Magnetic nerve stimulation is a novel techniquewith advantages over direct electrical stimulation.Either bilateral anterior magnetic stimulation (Fig.1.4) or a single circular coil anteriorly over the uppermediastinum can be employed to confirm or refute aclinical diagnosis of respiratory muscle weakness.The use of both techniques requires experience.16

ELECTROPHYSIOLOGICAL MEASUREMENTS

These have the disadvantage that they do not giveinformation regarding the force-generating capacityof the muscle. Nevertheless, investigation of theintegrity of the phrenic nerve may be indicated in thefollowing situations.

• To determine prognosis if weakness related tomedical intervention is demonstrated. For ex-ample, if hemidiaphragm paralysis follows cardiacsurgery, one would expect the prognosis to be

Pulmonary mechanics 7

must be taken to avoid brachial plexus contamina-tion. An example of an action potential is shown inFigure 1.5.

Quantification of the action potential requiressupramaximal stimulation to ensure that all axonsare recruited; this is difficult with electrical stimula-tion, and the combination of magnetic stimulationand an oesophageal electrode is preferred.17 Thistechnique does, however, require specialist expertise,as do other techniques such as needle electromyogra-phy. Neither of these techniques is routinely used inEuropean ICUs.

PULMONARY MECHANICS

To obtain optimal mechanical ventilatory supportrequires an understanding of pulmonary mechanics.Interventions within the ICU often aim to achievethis by increasing pulmonary compliance anddecreasing both airway resistance and intrinsic PEEP,thus reducing the work of breathing.

Lung volumes

Figure 1.4 Bilateral anterior magnetic stimulation of the phrenic

nerves in a patient with chronic obstructive pulmonary disease with

difficulty wean ing from mechanical ventilation.

better if an action potential is still demonstrable.This distinction can occasionally be of medico-legal importance.

• Occasionally where it is considered necessaryto distinguish axonal from demyelinatingneuropathies. In the former, the amplitude ofthe action potential is diminished, whereas in thelatter, the conduction time is prolonged.

The basic measurement of phrenic nerve electro-physiology is conduction time (PNCT). For thismeasurement, it is critical that the action potentialmeasured originates from the diaphragm. The prob-ability that this is so can be increased by selectivelystimulating the phrenic nerve with electrical stimula-tion or, alternatively, using an oesophageal electrodeto record selectively from the diaphragm. PNCT isonly mildly influenced by stimulus intensity, so thesimplest practical option is to use electrical stimulationin conjunction with surface electrodes, though care

Lung volume is often reduced in the critically ill by avariety of factors: underlying lung disease, e.g. atelec-tasis, or reduced muscle strength from critical illnessmyopathy or factors that affect diaphragm functionsuch as abdominal distension.

Figure 1.5 Example of an action potential recorded from an

oesophageal electrode using unilateral electrical stimulation. (Figure

courtesy of Dr YM Luo, Kings College Hospital, London UK.)


VITAL CAPACITY AND FORCED EXPIRATORY

VOLUME

The measurement of vital capacity is most commonlyused to monitor progress in conditions in which thereis acute respiratory muscle weakness. It is of use inpredicting the need for ventilatory support inGuillain-Barre syndrome but not in myastheniagravis,18 possibly reflecting the tendency for acutedeterioration in the latter condition. In general terms,a vital capacity of more than 10 ml kg appears to berequired to sustain adequate spontaneous ventilation.Vital capactiy has not been shown to be of use inpredicting the requirement for ventilatory supportin conditions other than respiratory muscle weakness.

Portable spirometers permit the bedside measure-ment of vital capacity and forced expiratory volume(FEVj) in the non-intubated, spontaneously breathingsubject. Accurate vital capacity measurements requirepatient co-operation and may be difficult to undertakein the ICU setting if patients are confused or sedated.

FUNCTIONAL RESIDUAL CAPACITY

Functional residual capacity is the lung volume inthe neutral position with complete respiratory mus-cle relaxation. At this point, the inward elastic recoilpressure of the lung is equal to the outward recoilpressure of the chest wall and corresponds to lungvolume at end-expiration. Measurement of FRC isnot commonly carried out in the ICU. The two tech-niques most commonly used are helium dilution andthe nitrogen wash-out method.19 Both methods onlymeasure the volume of gas in ventilated parts of thelung and will underestimate gas volume in lung unitswith long time constants.

Helium dilutionThe technique relies on measuring the differences inhelium concentration before and after the subject isswitched into a closed-circuit rebreathing systemwhen they are at FRC. By knowing the volume of thecircuit and helium concentrations, FRC can then becalculated.

Nitrogen wash-out techniqueThe patient is ventilated with air before beingswitched to a ventilator at end-expiration (FRC) andventilated with 100% O2 at the end of expiration (i.e.at FRC). Breath by breath N2 concentration and flowmeasurements are made until the concentration of

N2 is less than 1%. FRC is then given by the totalvolume of N2 washed out divided by the fall in thepercentage of N2 from the start to the end of the test.

A theoretical disadvantage of the helium dilutiontechnique is that the value of FRC is determined atthe end of the test when expiratory resistance due tothe rebreathing circuit may have artificially increasedFRC; whereas in the nitrogen wash-out technique thevalue of FRC is determined at the start of the test. Infact, comparison of the two methods has shownthem to produce similar results.

Breathing pattern

In many critically ill patients, breathing pattern isof greater interest than lung volumes per se.Abnormalities of respiratory frequency and tidal vol-ume are common and in several studies an elevatedrespiratory frequency has been shown to predictan adverse outcome. In a case-controlled studyof patients discharged from the ICU, respiratoryfrequency and haematocrit were the only continuousvariables that predicted re-admission to the ICU,with a resulting high mortality.20 Tidal volume,respiratory frequency and minute volume are easy tomeasure in intubated patients and the valuesare continuously displayed on all modern ventilators.

Compliance

Compliance of the respiratory system (Crs) isreduced (made stiffer) in conditions such aspulmonary fibrosis that affect the lungs or respirato-ry muscle weakness, where the change is more inchest-wall elasticity (the reciprocal of compliance).Of particular relevance in the ICU, is the fact thatcompliance is reduced by increases in pulmonaryvenous pressure or in alveolar oedema, acute respira-tory distress syndrome (ARDS) and pneumonia.Total respiratory system compliance is made up oflung, chest-wall and abdominal components and, insome conditions, such as intra-abdominal sepsis, thefall in abdominal compliance is very significant.

Crs is the change in lung volume (DV) producedper unit change in recoil pressure across therespiratory system (DPrs): Crs = DV/APrs. The meas-urement of compliance may be useful, particularly


when assessing the response to therapeutic interven-tions, and several methods exist.

MEASUREMENT OF STATIC COMPLIANCE

DURING MECHANICAL VENTILATION

Compliance is measured at zero airflow so thatchanges in pressure reflect changes in elastic recoil ofthe lung and chest wall and are not influenced byairway resistance, which will increase the drivingpressure required to generate a given airflow.Accurate compliance measurements require totalrelaxation of the inspiratory and expiratory muscles,making it one of the few tests that is more readilyperformed in the ICU than in the lung functionlaboratory! It should be noted that distensible venti-lator tubing may contribute significantly to themeasured compliance, particularly in those withabnormally stiff lungs.

Rapid airway occlusion techniqueAfter ensuring that the patient is relaxed and onlyoccasionally triggering the ventilator, a series of end-inspiratory airway occlusions is made at differentinflation volumes from end-expiratory lung volumes(EELV) to EELV + 1000 ml. Between each testbreath, normal ventilation is resumed. Differentinflation volumes may be achieved by changing therespiratory frequency, with the inflation volumebeing derived from integration of the flow signalusing a pneumotachograph connected to a differen-tial pressure transducer inserted between the ventila-tor circuit and the endotracheal tube. When theairway is occluded (by pressing the expiratory holdbutton on the ventilator at end inflation), the pres-sure at the airway opening (Pao) rises to a peak andthen plateaus at a pressure equal to alveolar pressureP2, as shown in Figure 1.6. This plateau pressure lessthe sum of intrinsic and extrinsic PEEP representsthe elastic recoil of the respiratory system at endinflation, Prs. Equilibration is usually completewithin 3 s, although it may take longer with airflowlimitation. This method is derived from the 'super-syringe' method, in which lung inflation is carriedout in 100-mL increments using a large, calibratedsyringe during a prolonged apnoea, and Prs isdetermined in the same way.

The inspiratory pressure-volume curve can thenbe constructed by plotting volume against thecorresponding static Pao (see the inspiratory limb of

Fig. 1.9). Compliance is given by the gradient of thelinear portion of the curve, which tends to be betweenFRC and FRC + 500 mL in normal patients. Withinthis range (expanding pressure of about —2 to —10cmH2O), the lung is remarkably distensible (verycompliant) at around 0.2 L/cmH2O. At extremes oflung volume, compliance is reduced, as reflected byflattening of the pressure-volume curve. At high lungvolume this is due to increased inwards elastic recoiland at volumes below FRC it is attributed to increasedoutwards elastic recoil of the chest wall and increasedairway closure. It is over the linear portion of thepressure-volume curve that most efficient ventilationtakes place, with the greatest change in lung volumefor a given applied pressure.

Figure 1.6 Representation of flow, volume and airway opening

pressure (Pao) for determination of compliance and resistance in a

mechanically ventilated patient using the rapid airway occlusion

technique. Following a constant inspiratory flow, the airway is

occluded at end-inspiration. Pao falls rapidly from a peak (Pmax)

to P1, and more slowly to a plateau, P2. The elastic recoil of the

respiratory system is given by (P2 - PEEP). Prs, respiratory system

pressure.


Pulse method21

As with the rapid occlusion method, flow is meas-ured using a pneumotachograph positioned eitherattached to the ETT or integral to the ventilator.Transthoracic pressure is taken as the differencebetween mouth pressure measured at the proximalpneumotachograph port and atmospheric pressure(Pao). The ventilator is adjusted to deliver a con-stant rate of airflow, V. When inflation begins, thepressure tracing shows an initial step rise related tothe flow resistance of the subject, followed by a sec-tion with a slower rise and a constant slope,(Pao)slope. Compliance is then given by Crs =V/(Pao)slope where V is measured in L/s and(Pao)slope in cmH2O/s. This method has severaladvantages, the most important of which is itsability to be used in patients in assist modes ofventilation. The patient's respiratory effort is detectedas an irregular flow tracing, and these breaths maybe discarded from the analysis. Some ventilatorshave this method incorporated into the monitoringoptions. It is a method that has been shown to correlatewell with values of Crs obtained using the rapidairway occlusion method.

MEASUREMENT OF DYNAMIC COMPLIANCE

During spontaneous breathing, measurement ofstatic compliance requires patient co-operation with

Figure 1.7 Pressure-volume curve of the lung during a respiratory

cycle, illustrating hysteresis with a greater lung volume achieved for

any given pressure during expiration than inspiration. (Reproduced

from reference 39.)

a difficult technique and the equipment is not veryportable. Measurement of dynamic compliance,CDyn, may, however, be carried out. The patientbreathes either via a pneumotachograph to give vol-ume or via a spirometer so that tidal volume (V"T)may be determined breath by breath. Continuousmeasurement of oesophageal pressure (Pes) as anestimate of pleural pressure is made using anoesophageal balloon. DPgs between points of zeroairflow is determined breath by breath from end-inspiration to end-expiration, as an assumption ismade that at these points there is complete respirato-ry muscle relaxation and that no airflow is present.Values are averaged over 10-15 breaths and dynamiccompliance determined: CD n = VT/DPoes. For anygiven measurement of pleural pressure, the associ-ated lung volume will be greater during expirationthan during inspiration due to the hysteresis of thepressure-volume curve (see Fig. 1.7). This is due tothe pressure required to overcome surface tensionforces within the alveoli during inspiration. Dynamiccompliance measurements should therefore be madeduring the expiratory phase. Dynamic compliance isunreliable in airflow obstruction, especially at higherrespiratory frequencies, as airflow within the lungsmay be present even when the airway is occluded.

TECHNIQUES USED TO DETERMINE

COMPLIANCE

Super-syringeProvides full pressure-volume curve during infla-tion and deflation, but:

(i) the patient is temporarily disconnected fromthe ventilator, changing both the previousvolume history and the end-expiratory lungvolume,

(ii) it is time consuming,(iii) it fails to start the test from a Prs equivalent

to intrinsic PEEP,(iv) continuing gas exchange during the man-

oeuvre reduces thoracic volume, causing anover-estimation of the hysteresis loop area,

(v) it usually requires a temporary increase inFiO2, which may affect the curve due to the

development of atelectasis,(vi) measurement depends on the patient mak-

ing no respiratory effort.


Rapid airway occlusionRequires no additional equipment, the patientremains connected to the ventilator, it onlyinterrupts ventilation during an inspiratory pauseof 3-5 s and has minimal impact on gas exchange,but:

(i) the expiratory limb of the pressure-volumecurve is less readily examined,

(ii) measurement depends on the patient mak-ing no respiratory effort.

Constant flow inflation or pulse methodCan be used in assist modes of ventilation anddisconnection from the ventilator; intermittentairway occlusion or end-inspiratory pause is notnecessary, but errors may arise if different timeconstants within the lung prevent a linear relation-ship between flow and transthoracic pressure.

Dynamic complianceRequires an oesophageal balloon to be positioned,cannot be used in patients with airflow limitationbecause within-lung airflow is still present even atend-inspiration and end-expiration and relies onthe assumption that respiratory muscles arecompletely relaxed during end-inspiration andend-expiration.

Intrinsic positive end-expiratorypressure

Intrinsic positive end-expiratory pressure (PEEPj orautoPEEP) is the presence of a positive alveolarpressure at the end of expiration (Fig. 1.8, (a) and(b)). Under these conditions, expiratory flow hasnot stopped before the next inspiration occurs,either during spontaneous or positive pressureventilation. It may arise in three different cir-cumstances:22

1. Insufficient Te to allow expiration to the equilib-rium (relaxed) volume due to airflow obstructionand/or high ventilatory requirements.

2. Dynamic airway collapse and resulting flow lim-itation due to emphysema.

3. Continuing expiratory muscle contraction atend-expiration (often contributing to apparentPeepj although not normally assigned as acause).

PEEPj increases inspiratory work during sponta-neous breathing and reduces the ability to triggerthe ventilator during assisted modes of ventilationas greater inspiratory effort is needed. These aspectsmay be overcome by applying external PEEPapproximately equal to that of the patient's PEEP;.Measurement of PEEP; may therefore be used toallow better synchrony of the machine with patientdemand. It can also be employed to assess theresponse to bronchodilators. Values of up to 20cmH2O are not uncommon in asthma or COPD.Recognition of PEEPj and consideration of its aeti-ology are helpful. A patient with PEEPi due toairflow limitation and hyperinflation, for example,would potentially benefit from a prolonged expira-tory time and application of extrinsic PEEP, where-as extrinsic PEEP applied to a patient with PEEPidue to expiratory muscle activation alone wouldmerely increase the patient's work of breathingfurther.

Figure 1.8 Measurement of static PEEPi. In the absence of airflowobstruction, alveolar pressure (Palv) at the end of expiration equalsthe pressure at the airway opening (Pao) (a). With airflow obstruction,however, Palv is greater than Pao at end-expiration (b). This increase inPalv, static PEEPi, may be measured from the airway pressure duringan end-expiratory pause with the expiratory port occluded so Pao thenequals Palv (c). (Reproduced from reference 40.)


MEASUREMENT OF INTRINSIC POSITIVE

END-EXPIRATORY PRESSURE

Static PEEP; can be measured by occluding the air-way at the end of expiration, with the resultingplateau pressure representing the average PEEPjpresent within the non-homogeneous lung(Fig. 1.8 (c)). This method is automatically avail-able on some commercial ventilators, but requiresthe absence of expiratory effort otherwise a falselyhigh value will be recorded. A dynamic measure-ment of PEEPj can be obtained by recordingthe airway pressure at which inspiratory flow com-mences during inspiration (Fig. 1.9). Inspiratoryflow commences only after airway pressure hasexceeded the value of PEEPj. This dynamic mea-surement reflects the lowest regional value of PEEPjand may be considerably less than static PEEPiin patients with airflow limitation and in thosewith significant pulmonary inhomogeneity.Measurements of PEEPi should be made withoutany external PEEP applied. This method can beapplied in the non-intubated patient but, com-monly, active expiration, e.g. in patients withairflow limitation, overestimates PEEPi. This maybe taken into account by measuring gastric andthus intra-abdominal pressure. The rise in gastricpressure may then be subtracted from the measuredvalue of dynamic PEEPi.

Airways resistance

Several factors apart from the elastic recoil of thelungs and chest wall must be overcome to move airin and out of the lungs, including the inertia of therespiratory system itself, factional resistance at thelung-chest-wall interface, and chest-wall and pul-monary resistance. In most cases, pulmonary resist-ance is the only factor of significance, with 80% ofthis being attributable to airway resistance and theother 20% to resistance of the lung tissue itself.Greatest airflow resistance is present in medium-sized airways, as distal to this the increased resistanceof individual airways is offset by their large number.Airway resistance falls with increasing lung volume(Fig. 1.10). This is in part due to increased tractionon small airways lacking cartilaginous supportpulling them open as lung elastic recoil increasesduring a large inspiration. Conversely, during a

Figure 1.9 Airflow and oesophageal pressure (Pes) in a patient

with obstructive lung disease. Dynamic PEEPi is given by Pes at

end-inspiration, as indicated by the point of zero airflow.

forced expiration at low lung volumes, the positivepleural pressure is transmitted to the airways, whichmay be compressed or even collapse. Thus, airwaysresistance is significantly higher during active expira-tion due to dynamic compression of the airways.

CAUSES OF INCREASED RESPIRATORYSYSTEM RESISTANCE

Airways

• Reduced lung volumes• Dynamic compression during expiration• Increased flow rates• Non-laminar airflow (e.g. in COPD)

Visco-elastic

• Pulmonary fibrosis• Presence of oedema or areas of consolidation

Determination of airway resistance is usefulwhen assessing the causes of high airway pressuresduring mechanical ventilation and the response tobronchodilators. Although increased airway resis-tance usually results from airflow limitation, it isalso increased in ARDS or cardiogenic pulmonaryoedema; this may reflect oedema in the airway walland secretions within the airway lumen. An addi-tional factor is a reduction in the number of patentairways due to the marked loss of functional lungvolume.


Figure 1.10 Increase in airways resistance with decreasing lung

volume. There is a linear relationship between the reciprocal of

airway resistance (conductance) and lung volume. (Reproduced

from reference 39.)

During laminar airflow, the relationship betweenpressure difference (P), flow (V) and resistance (R)is: P = VR, so:

described for the assessment of static compliance.Before the end of a mechanical inflation with con-stant inspiratory flow, V, the airway opening isoccluded temporarily until a plateau in airway pres-sure is achieved (by pressing the expiratory holdknob). The peak airway pressure (Pmax) falls rapidlyto a lower value, PI, and then more slowly to aplateau pressure, P2, as illustrated in Figure 1.6.Maximum respiratory resistance (Ptotmax) is thengiven by Rtotmax = (Pmax - P2)/V and minimumrespiratory resistance, R . is given by Pvtotmin =

(Pmax-p1)/V.Rtotmin represents the instantaneous resistance of

the respiratory system, mainly representing the airwayresistance. Rtotmax also reflects the component attrib-utable to time constant inequalities within the respira-tory system, and to visco-elastic pressure dissipationwithin the thoracic tissues. It should be borne in mindthat resistance of the respiratory tubing and ETT willcontribute towards the values of resistance obtained.Although pre-determined values of in-vitro ETT resis-tance may be subtracted, it has been demonstratedthat in-vivo values are higher.23

and Poiseuille's law applies: resistance being relatedto viscosity of the air (h), length (/) and radius (r) ofthe tube:

Thus, doubling the length of an endotrachealtube would double the resistance to airflow, whereashalving its radius would increase the resistance16-fold. The resistance posed by an endotracheal tubeis usually not clinically significant, but may beconsiderable at high levels of minute ventilation orwith short inspiratory times.

Increasing flow rates, diameter of the tube andincreased gas density all predispose to turbulent flow.During turbulent flow a much greater driving pres-sure is then required to generate the same flow:

DETERMINATION OF AIRWAYS RESISTANCE

In the ventilated patient, resistance may be deter-mined with the rapid airway occlusion technique as

Work of breathing

The work of breathing may be increased for a varietyof reasons, some of which may co-exist. Lung orabdominal compliance may be reduced, the chestwall may be stiff or airway resistance may beincreased. During mechanical ventilation, the workof breathing is more in the assist than control modes,because the patient makes effort in triggeringbreaths. It will also be increased when there is poorpatient-ventilator synchronization. The work ofbreathing is rarely clinically quantified, but appreci-ation of workload is useful when excessive or whenpredicting weaning from ventilation.

A number of methods have been used. For example,the O2 consumption of the respiratory muscles can bemeasured. In normal subjects, it is less than 5% of totalO2 consumption; with voluntary hyperventilation, it ispossible to increase this to 30%. In some patients, theO2 cost of breathing is a limiting factor on exercise per-formance. In spontaneously breathing subjects inwhom an oesophageal balloon has been positioned, thework of breathing is proportional to the magnitude ofnegative intrathoracic pressure and the length of timefor which it is maintained. Thus, if oesophageal pres-sure is plotted against time, the area under the curve,


termed the pressure-time product, is proportional tothe work of breathing (Fig. 1.11). Care is requiredwhen using this method to pinpoint the beginning ofeach inspiratory effort with the externally set PEEP asthe reference point (rather than zero pressure). Thismeasurement is less meaningful during assisted venti-lation because, with positive pressure ventilation, thenormal negative pressure swings are reversed duringinspiration, whereas with negative pressure support,the normal pressure swings will be augmented as lesswork is actually performed by the respiratory muscles.An alternative method is to use the diaphragmaticEMG, although this requires specialist expertise.

Examination of the airway pressure contour in aventilator-dependent patient can provide usefulinformation on the activity of the patient's respira-tory muscles.24 During mechanical ventilation, thework of breathing is given by the product of pressureacross the respiratory system (i.e. from alveolus toatmosphere) and inflating volume. Thus, if Pao isplotted against lung volume, the work of breathing isgiven by the area under the curve. Computerizedbedside monitoring is available for automated deter-mination of the work of breathing using thismethod.25 In control mode with the patient fullyrelaxed, the area under the pressure-volume curveindicates the work carried out by the ventilator toexpand the chest. The additional work of breathingcarried out by the patient in assist mode is given bythe difference between the area under the pres-sure-volume curve in assist and control modes. Ahigh work of breathing as assessed using this method

may be an indicator of poor patient-ventilator syn-chronization, and may arise when the inspiratoryflow is too low. This typically occurs with high respira-tory rates or drive, e.g. in asthma. The phenomenonis called flow deprivation. Some ventilators have thecapability of detecting flow deprivation and willautomatically increase flow rate to compensate.

RESPIRATORY DRIVE AND CONTROL

Background

In health there is remarkable constancy of arterialCO2, and ventilation is matched breath by breath tometabolic CO2 load. Many central and peripheralstimuli influence respiratory drive, which in turnlead to changes in ventilation. One way of evaluatingthe function of the respiratory control system is toobserve the change in output, e.g. minute ventilationin relation to a given input such as an increase inPaCO2. There is, however, considerable redundancyin the system. At the onset of exercise, for example,the sum of the increase in ventilation expected fromeach of a number of sensory inputs equates to morethan is observed in reality. Central integration ofthese input parameters is complex and in critically illpatients becomes even more so. Increased drive mayarise from factors such as sepsis, circulatingcytokines or pulmonary afferent stimulation, where-as decreased drive may arise from sedatives/opiatesor brain injury. In addition, the normal controlmechanisms are disrupted, with time delays betweenthe patient initiating a breath and gas flow from theventilator, potential poor matching of ventilatorysupport to the patient's demands and, in heart fail-ure, delay in the circulation time between lung andchemoreceptors. The aim of this section is to high-light a number of ways in which respiratory controlmay be affected in critically ill patients and themethods of assessment available.

Figure 1.11 The pressure-time product during inspiration, as

indicated by the hatched area, is an indicator of the work of

breathing. Points of zero flow are required to define the start and

end of inspiration. Pes, oesophageal pressure. (Figure courtesy

of Dr N Hart, Muscle Laboratory, The Royal Brompton Hospital,

London, UK.)

The respiratory centres and CNSinput

INTRINSIC RHYTHM GENERATION

Neurons located within the pons and medulla of thebrainstem form the respiratory centres. They are

Respiratory drive and control 15

responsible for the generation of periodic inspirationand expiration, the nature of which is modifieddepending upon inputs from elsewhere. Variousgroups of neurons with discrete functions have beendescribed within the respiratory centre. One view isthat, within the medulla, a dorsal or ventral respira-tory group is responsible for inspiration and expira-tion respectively. Cells of the inspiratory area havethe property of intrinsic periodic firing and areresponsible for the basic rhythm of ventilation. Burstsof action potentials result in nervous impulses to thediaphragm and other inspiratory muscles. The intrin-sic rhythm pattern of the inspiratory area starts witha latent period of several seconds, during which thereis no activity. Action potentials then begin to appear,increasing in a crescendo ramp, over the next few sec-onds. During this time, inspiratory muscle activityprogressively increases. Finally, the inspiratory actionpotentials cease and inspiratory muscle tone falls.

MODIFICATION OF THE INTRINSIC RHYTHM

This inspiratory ramp can be 'turned off' prematurelyby inhibitory impulses from the pneumotaxic centre,thereby shortening inspiratory time and increasing res-piratory frequency. The expiratory area is quiescentduring normal quiet breathing as exhalation occurs bypassive recoil of the lungs and chest wall to their equi-librium position at FRC. On more forceful breathing,for example in a patient with increased airway resis-tance or during exercise in a normal subject, activity ofthe expiratory cells results in active expiration.

An apneustic centre in the lower pons has an exci-tatory effect on the inspiratory area of the medulla,tending to prolong the ramp action potentials. Insome types of brain injury, an abnormal breathingpattern is seen in which prolonged inspiratory gaspsare interrupted by transient expiratory efforts due todamage to the apneustic centre. A pneumotaxiccentre in the upper pons appears to 'switch off' orinhibit inspiration, thereby fine-tuning inspiratoryvolume and rate.

During wakefulness, volitional influences areimportant and the cortex can override the functionof the brainstem within limits, e.g. during voluntaryhyperventilation or speech. In the ventilated patient,sedatives/opiates are commonly used to reduce corti-cal drive. Anxiety or pain alters the pattern of breath-ing via the limbic system and hypothalamus and maymake it difficult to match ventilator and patient,

particularly if assist modes rather than pressure sup-port are being employed.

Central chemoreceptors

The central chemoreceptors are situated near the ven-tral surface of the medulla in the vicinity of the exit ofcranial nerves IX and X and have been separated intorostral and caudal areas on each side. They stimulatebreathing in response to an increase in brain extracel-lular hydrogen ions. The composition of cerebrospinalfluid (CSF) is governed by the activity of theblood-brain barrier (important in making adjust-ments over days to weeks), whereas brain ECF is pri-marily affected by cerebral metabolism and cerebralblood flow, the latter leading to virtually instantaneousadjustments of ECF pH with changes in arterial PCO2

as molecular CO2 diffuses readily across the blood-brain barrier. The precise location of the sensors isnot known. Thus, the CO2 level in blood regulatesventilation chiefly by its effect on the pH of brain ECF,and the resulting hyperventilation reduces jRaCO2 in theblood and therefore in brain ECF and bulk CSF.

In the presence of CO2 retention, the expected CSFacidosis is compensated for by homeostatic adjust-ments, mediated by the blood-brain barrier, andsimultaneous renal retention of bicarbonate ions.Thus, in chronic hypercapnia, the CSF pH is normal,with a consequent low minute ventilation for a givenPaCO2. This may be advantageous in reducing thework of breathing, but it may slow weaning from aventilator or increase the risk of nocturnal hypoven-tilation in the spontaneously breathing patient.An example of this phenomenon is illustrated inFigure 1.12, which shows the hypercapnic ventilatoryresponse breath by breath of a patient with chronic,stable hypercapnic respiratory failure. Before noc-turnal non-invasive ventilation (closed circles), thereis marked blunting of the ventilatory response.Following a period of nocturnal non-invasive ventila-tion, there is an increase in the ventilatory responseto CO2. During this period, the patient's daytimearterial blood gases improved (PaCO2 and -PaO2

before ventilation were 6.3 and 9.7 kPa respectively,and after ventilation 5.74 and 10.6 kPa). Much of thisincrease in ventilatory drive is due to adjustments inbicarbonate and pH (arterial HCO2 and H+ beforeventilation = 31.2 and 36.4 mM L-1, and after venti-lation = 27.9 and 37.2 mM L-1).


Peripheral chemoreceptors

These are located in the carotid bodies at the bifurca-tion of the common carotid arteries and in the aorticbodies above and below the aortic arch. The carotidchemoreceptors are the most important in humans.They are stimulated by a decrease in arterial O2 andpH or an increase in CO2. There is a relatively smallincrease in firing rate of the carotid sinus nerve asPaO2 is decreased from hyperoxia to normoxia andthen an exponential increase with the inflection pointaround 8.0 kPa. The peripheral chemoreceptors arethe only receptors to increase ventilation in response tohypoxia and, in their absence, hypoxaemia depressesventilation due to a central inhibitory effect.

In health, PaCO2 is maintained almost constant(±0.4 kPa or 3 mmHg). In other words, the 'error sig-nal' in the control system is so small as to be negli-gible. This remarkable control is achieved by the rapidresponse of the peripheral chemoreceptors to oscilla-tions in PaCO2 generated as CO2 rises within alveoliduring expiration, which then falls during inspirationas fresh air is drawn into the lungs. The rate of rise ofthe PaCO2 oscillation is directly proportional tometabolic CO2 load, making it a suitable signal for

Figure 1.12 Graph to show a part of the hypercapnic ventilatory

response curve during C02 re-breathing breath by breath, in a

subject with chronic, stable hypercapnic respiratory failure before

(closed circles) and several days after (open circles) commencing

nocturnal non-invasive ventilation.

respiratory control.26 Indeed, modification of thePaCO2 oscillation profile using inspiratory CO2 pulsesdelivered either early in inspiration or after a smalldelay results in a greater increase in ventilation withthe early pulses, despite the same mean PaCO2 inboth instances.27 In chronic obstructive airways dis-ease, there is damping of the PaCO2 oscillation, aneffect most pronounced in emphysematous patients.Those in whom PaCO2 oscillations are most dampedhave the highest PaCO2 levels, probably reflecting theseverity of their disease and reduced ventilatory drivesecondary to damping of PaCO2 oscillations.

In normal subjects given a CO2 mixture tobreathe, only 20% of the increase in ventilation isattributable to the peripheral arterial chemorecep-tors. In simple terms, one could regard the centralchemoreceptors as providing most of the driveto breathe, whereas the peripheral chemoreceptorsprovide fine precision control of breathing.

Cheyne-Stokes respiration

An exception to the rule of tight respiratory controlarises in Cheyne-Stokes respiration, in which peri-ods of apnoea of 10-20 s alternate with periods ofhyperventilation of equal duration. In the critically ill,Cheyne-Stokes respiration is often a poor prognosticindicator. Factors predisposing to Cheyne-Stokesrespiration include low end-tidal PCO2, high chemo-reflex sensitivity, increased chemoreflex time andlow alveolar lung volume. Thus, a low cardiac outputor hypoxia predisposes to periodicity of breathing.

Carbon dioxide retention duringoxygen supplementation

Chronically hypercapnic patients may retain CO2 ondelivery of supplementary O2. Four mechanismshave been proposed to explain this phenomenon.

1. Removal of the hypoxic ventilatory drive byoxygen supplementation in patients who haveincreased dependence on this respiratory drivedue to reduced ventilatory sensitivity to CO2.Reduced sensitivity to CO2 may be attributable toa tendency to normalize CSF and arterial pHdespite persistent hypercapnia, as described in thesection on central chemoreceptors.

Respiratory drive and control 17

2. Reversal of hypoxic pulmonary vasoconstrictionin poorly ventilated areas of the lung leading toincreased physiological shunting and so V:Qmismatching.

3. Increased ratio of dead space to tidal volume(VD: VT), e.g. due to a change in breathing patternwith an increase in respiratory frequency: VT, ordilation of proximal airways.

4. Loss of PaCO2 oscillations reducing peripheralchemoreceptor ventilatory drive.

There is no agreement among studies as to which isthe predominant mechanism, possibly reflectinggenuine differences in patients. In one study of patientswith an acute exacerbation of COPD given 100% O2 tobreathe, ventilation was depressed in CO2 retainers butnot in those who did not retain CO2,

28 giving supportto theory (1). Reversal of hypoxic pulmonary vasocon-striction, as indicated by the ventilation:perfusiondispersion ratio, occurred to an equal degree in bothCO2 retainers and non-retainers, suggesting that thiswas not an important mechanism. In another study ofpatients during remission, however, supplementary O2

resulted in a substantial increase in PaCO2 with mini-mal change in ventilation, suggesting predominance ofchanges in VQ matching or VD: VT

29

Other receptors influencingrespiratory control

A number of pulmonary afferents project centrally byway of the cervical vagus nerves, such as pulmonarystretch receptors and lung epithelial irritant recep-tors, and these modulate the ventilatory response toCO2 and hypoxia. Vagal inhibition using cooling ofthe vagal nerves in anaesthetized rabbits has beenshown to stimulate ventilation during modest hyper-capnia (shifting to the left and steepening of thehypercapnic ventilatory response curve), but toinhibit ventilation during more intense hypercapnia(flattening of the hypercapnic ventilatory responsecurve). In humans, pulmonary stretch receptors arelikely to be of greater importance in newborn babiesthan in adults. Juxta-capillary receptors, lying in thealveolar walls, result in rapid, shallow breathingduring moderate stimulation and in apnoea duringintense stimulation. There is evidence that engorge-ment of pulmonary capillaries and increases in theinterstitial fluid volume stimulate these receptors.

They may therefore play a role in the rapid, shallowbreathing and dyspnoea associated with heart failureand interstitial infiltration. Joint and muscle recep-tors are likely to be of most importance in increasingventilation in the early stages of exercise. Many mus-cles contain muscle spindles that sense elongation ofthe muscle. This information is used to control thestrength of contraction. These receptors may beinvolved in the sensation of dyspnoea that occurswhen unusually large respiratory efforts are requiredto move the lung and chest wall, for example becauseof airway obstruction. An increase in arterial bloodpressure can cause reflex hypoventilation or apnoeathrough stimulation of the aortic and carotid sinusbaroreceptors. Conversely, a decrease in blood pres-sure may result in hyperventilation and a sensation ofair hunger. The pathways of these reflexes are largelyunknown. Other afferent nerves can bring aboutchanges in ventilation: pain may cause a period ofapnoea followed by hyperventilation, and heatingof the skin may result in hyperventilation.

Assessment of ventilatory drive

This is of most use when assessing the continuedneed for ventilatory support. Measurement is madeduring a period of spontaneous breathing and com-pared with the value after a period of unsupportedbreathing.

ARTERIAL BLOOD GASES

The measurement of alveolar or arterial -PCO2 givesa better indication of the normality, or otherwise,of alveolar ventilation than does formal measure-ment of minute ventilation. In the presence ofhypoxia, PaCO2 will be reduced in a predictablefashion.30

THE HYPERCAPNIC VENTILATORY RESPONSE

In normal subjects, minute ventilation gives anadequate reflection of respiratory drive, butmechanical constraints, e.g. airflow obstruction,may limit ventilation and so other indicators ofdrive need to be considered. If there is a defect atany stage in the output chain from the ventilatorycentres down, transformation into ventilation willbe impaired.31


Changes in pleural pressure: P01 and

dpOesMax/dtThe oesophageal occlusion pressure, i.e. theoesophageal pressure at 100 ms during an inspiratoryattempt against a closed airway at functional residualcapacity PQ 1, has been proposed as a more reliableindicator of drive in airflow obstruction. P0 l hasbeen used within the ICU to:

• assess respiratory centre output while changingthe fraction of inspired O2,

32

• predict weaning outcome (values of P0.. l < 6cmH2O suggest that discontinuation of mechan-ical ventilation is likely to be unsuccessful, where-as values of < 4 cmH2O predict success),33

• set the appropriate level of assisted ventilation34

and pressure support.35

Technical difficulties determining the true onset ofinspiratory muscle activity from pressure data mayreduce the reliability of this technique. The maximumrate of change of oesophageal pressure, dPoesMax/dt,has been proposed as an alternative to P0 l and hasbeen shown to have a good correlation with end-tidalpressure (PetCO2) in normal subjects during CO2

rebreathing and with walk time during an exhaustiveexercise test in patients with COPD.36

Diaphragmatic EMGHigh values of P0 l and dPoesMax/dt always indicatehigh ventilatory drive. Low values may indicate lowdrive, a defect of the phrenic nerves or muscle weak-ness. For instance, in severe COPD, hyperinflationmay render the diaphragm so mechanically dis-advantaged that there is functional weakness and it isno longer capable of generating high inspiratorypressures. PO 1 and dPoesMax/dt will therefore bothbe reduced for a given central drive. However, drivemay be examined by an oesophageal electrode tomeasure diaphragmatic EMG.

BREATHING PATTERN AND ACCESSORY

RESPIRATORY MUSCLE USE

Patients in whom the work of breathing is increasedor respiratory muscle strength decreased often com-pensate by adopting a shallow breathing pattern withincreased respiratory frequency. It should be bornein mind that measures that do not take into accountrespiratory frequency, such as P0.1 and poesMax/dtwill give a less accurate reflection of drive. Ventil-

atory drive is particularly difficult to assess inpatients with respiratory muscle weakness: outputcannot be seen as flow, intrathoracic pressure changeor muscle tension as these are all below points on theoutput chain where the problem is. The diaphrag-matic EMG may be distinctly abnormal due to thedisease process itself. In these patients, other clues,such as respiratory frequency and detection of activ-ity of respiratory muscles not normally used in quietbreathing (neck accessory muscles, shoulder girdlemuscles and abdominal muscles), all assume greaterimportance.

CONCLUSION

In conclusion, imbalance of the load and capacityof the respiratory muscle pump may result inventilatory failure. The clinician must be alert to thepossibility of respiratory muscle pump overload inpatients with a persistent requirement for mechan-ical ventilation. Strategies to assess the load on thepump and its capacity have been outlined, but fur-ther research is needed to determine the prevalence,causes and magnitude of respiratory muscle pumpdysfunction. Automated assessment of manyaspects of pulmonary mechanics has undoubtedlymoved it from a predominant research tool intothe clinical field. Intuitively, an appreciation ofpulmonary mechanics, with carefully tailoredventilatory strategies, should decrease the work ofbreathing and so improve outcome. However, fur-ther studies are required in this area. It is hopedthat future work may lead to the development of'intelligent ventilators', which assess and respond tothe work of breathing undertaken by the patientthroughout each breath and are sufficiently specificthat they are able to recognize activities such asswallowing or talking.

REFERENCES

1. Polkey, Ml, Lyall, RA, Green, M, Leigh, PN, Moxham, J.

Expiratory muscle function in amyotrophic lateral

sclerosis. Am J Respir Crit Care Med 1998; 158: 734-41.

2. Gibson, G. Diaphragmatic paresis: pathophysiology,

clinical features and investigation. Thorax 1989; 44:

960-70.

References 19

3. Parthasarathy, S, Jubran, A, Tobin, MJ. Cycling of

inspiratory and expiratory muscle groups with the

ventilator in airflow limitation. Am J Respir Crit Care

Mo/1998; 158: 1471-8.

4. Polkey, M, Lyall, R, Moxham, J, Leigh, P. Respiratory

aspects of neurological disease. J Neurol Neurosurg

Psychiatry 1999; 66: 5-15.

5. Quera-Salva, MA, Guilleminault, C, Chevret, S, et al.

Breathing disorders during sleep in myasthenia gravis.

Ann Neurol 1992; 31: 86-92.

6. Polkey, Ml, Kyroussis, D, Hamnegard, CH, etal.

Diaphragm strength in chronic obstructive pulmonary

disease. Am J Respir Crit Care Med 1996; 154(5):

1310-17.

7. Moxham, J, Goldstone, J. Assessment of respiratory

muscle strength in the intensive care unit. Eur Respir J

1994; 7: 2057-61.

8. Gabor JY, Cooper, AB, Hanly, PJ. Sleep disruption in the

intensive care unit. Curr Opin Crit Care 2001; 7: 21-7.

9. Resta, 0, Guido, P, Foschino-Barbaro, MP, et al. Sleep-

related breathing disorders in acute respiratory failure

assisted by non-invasive ventilatory treatment: utility

of portable polysomnographic system. Respir Med

2000; 94(2): 128-34.

10. Spitzer, AR, Giancarlo, T., Maher, L, Awerbuch, G.

Neuromuscular causes of prolonged ventilator

dependency. Muscle Nerve 1992; 15: 682-6.

11. Coakley, JH, Nagendran, K, Honavar, M, Hind, SJ.

Preliminary observations on the neuromuscular

abnormalities in patients with organ failure and

sepsis. Intensive Care Med 1993; 19: 323-8.

12. Deconinck, N, Van Parijs, V, Bleeckers-Bleukx, G, Van

den Bergh, P. Critical illness myopathy unrelated to

corticosteroids or neuromuscular blocking agents.

Neuromusc Dis 1998; 8: 186-92.

13. Leitjen, FSS, de Weerd, AW, Poortvliet, DCJ, et al.

Critical illness polyneuropathy in multiple organ

dysfunction syndrome and weaning from the

ventilator. Intensive Care Med 1996; 22: 856-61.

14. Dejonghe, B, Cook, DJ, Outin, H. Risk factors for

polyneuropathy of critical illness. In 7999 yearbook of

intensive care and emergency medicine, J-L Vincent, ed.

Springer: Berlin, 1999: 322-30.

15. Bolton, CF. Sepsis and the systemic inflammatory

response syndrome: neuromuscular manifestations.

Crit Care Med 1996; 24: 1408-16.

16. Polkey, Ml, Duguet, A, Luo, Y, et al. Anterior magnetic

phrenic nerve stimulation: laboratory and clinical

evaluation. Intensive Care Med 2000; 8: 1065-75.

17. Luo, YM, Lyall, RA, Harris, ML, et al. Quantification of the

esophageal diaphragm EMG with magnetic phrenic

nerve stimulation. Am J Respir Crit Care Med 1999;

160(5 P+ 1): 1629-34.

18. Rieder, P, Louis, M., jollliet, P, Chevrolet, J. The repeated

measurement of vital capacity is a poor indicator of

the need for mechanical ventilation in myasthenia

gravis. Intensive Care Med 1995; 21: 663-8.

19. Numa, AH, Newth, CJL. Assessment of lung function

in the intensive care unit. Pediatr Pulmonol 1995;

19: 118-28.

20. Durbin, C, Kopel, R. A case-control study of patients

readmitted to the intensive care unit. Crit Care Med

1993; 21: 1547-53.

21. Suratt, PM, Owens, D. A pulse method for measuring

respiratory system compliance in ventilated patients.

chest`1981;80:34-8.

22. Marini, JJ. Should PEEP be used in airflow obstruction?

Am Rev Respir Dis 1989; 140: 1-3.

23. Conti, G, Blasi, RAD, Lappa, A, et al. Evaluation of

respiratory system resistance in mechanically

ventilated patients: the role of the endotracheal tube.

Intensive Care Med 1994; 20: 421-4.

24. Marini, JJ, Capps, JS, Culver, BH. The inspiratory work

of breathing during assisted mechanical ventilation.

Chest-\985; 87(5): 612-18.

25. Banner, M, Jaeger, M, Kirby, R. Components of the

work of breathing and implications for monitoring

ventilator-dependent patients. Crit Care Med 1994;

22(3): 515-23.

26. Band, D, Wolff, C, Ward, J, Cochrance, G, Prior, J.

Respiratory oscillations in arterial in carbon-dioxide

tension as a control signal in exercise. Nature 1980;

283: 84-5.

27. Datta, AK, Nickol, AH. Dynamic chemoreceptiveness

studied in man during moderate exercise breath by

breath. Adv Exp Med Biol (US) 1995; 393: 235-8.

28. Robinson, TD, Freiberg, DB, Regnis, JA, Young, IH.

The role of hypoventilation and ventilation-perfusion

redistribution in Gyinduced hypercapnia during acute

exacerbations of chronic obstructive pulmonary

disease. Am J Respir Crit Care Med 2000; 161: 1524-9.

29. Dick, C, Liu, N, Sassoon, C, Berry, R, Mahutte, C.

02 induced changes in ventilation and ventilatory drive

in COPD. AmJ Respir Crit Care Med 1997; 155: 609-14.

30. Wolff, C. The control of ventilation in hypoxia I. The

Newsletter of the International Society for Mountain

Medicine 1997; 8(1): 3-5.

31. Whitelaw, W. Assessment of output of the respiratory

controller. Semin Respir Crit Care Med 1998; 19(4):

361-5.

32. Aubier, M, Murciano, D, Fournier, M, et al. Central

respiratory drive in acute respiratory failure patients


with chronic obstructive pulmonary disease.


33. Murciano, D, Boczkowski, J, Lecocguic, Y, et al. Tracheal

occlusion pressure: a simple index to monitor

respiratory muscle fatigue during acute respiratory

failure in patients with chronic obstructive pulmonary

disease. Ann Intern Med 1988; 108: 800-5.

34. Sassoon, C, Mahutte, C, Simmons, D, Light, R. Work

of breathing and airway occlusion pressure during assist-

mode mechanical ventilation. Chest 1988; 3 : 571-6.

35. Alberti, A, Gallo, F, Fongaro, A, Valentia, S, Rossi, A.

P0.1 is a useful parameter in setting the level of

pressure support ventilation. Intensive Care Med 1995;

21: 547-53.

36. Hamnegard, CH, Polkey, Ml, Kyroussis, D, et al.

Maximum rate of change in oesophageal pressure

assessed from unoccluded breaths: an option where

P0.1 is impractical. Eur RespJ 1998; 12: 693-7.

37. Green, M. Respiratory muscle testing. Bull Eur

Physiopathol 1984; 20: 433-6.

38. Polkey, Ml, Hamnegard, CH, Hughes, PD, et al.

Influence of acute lung volume change on contractile

properties of the human diaphragm. J Appl Physiol

1998;85:1322-8.

39. West, J. Respiratory physiology - the essentials, 5th edn,

PA Coryell, ed. Baltimore: Williams & Wilkins; 1990.

40. Hughes, J, Pride, N. Physiological principles and clinical

applications. WB Saunders; 1999.

2Mechanical ventilation: the basicsJOHN CGOLDSTONE

Introduction

Indications for mechanical ventilation

Airway pressure

Modes of ventilation

21 Partial assistance

22 Recent advances in mechanical ventilation

23 Practical aspects of mechanical ventilation

24 References

25

27

28

31

INTRODUCTION

Mechanical ventilation is used to replace or aid thework usually carried out by the respiratory muscles.As with any technology, several milestones werereached before its general introduction. Anaesthetistswere adept at airway control, and intubation of thetrachea was required to perform the complex headand neck surgery that resulted from the First WorldWar. Advances in pulmonary physiology improvedthe understanding of gas exchange, and the intro-duction of neuromuscular paralysis enabled experi-ence to be gained with positive pressure ventilationto facilitate thoracic surgery.

Surprisingly, the event that precipitated the wide-spread introduction into clinical practice of devicesto ventilate the lungs did not originate in the operat-ing theatre, but was an outbreak of paralytic polio inScandinavia in the 1950s. It is fortunate that theacute onset of respiratory muscle paralysis presentedthe simplest scenario for a ventilator to perform, totake the place of the respiratory muscles and pushgas into the chest with underlying normal lungs.1

There was an urgent need for new technology. At theheight of the outbreak, medical students were work-ing in 2-hour shifts performing ventilation by handusing simple breathing circuits for many days at atime. New cases overwhelmed the capacity of the

acute hospitals. A transfer of technology from indus-try to medicine occurred with a speed of develop-ment that would be unlikely to be possible today.What emerged were devices that were robust andsimple. Indeed, for many years these original charac-teristics applied to most ventilators.

Important events in the 1960s and 1970s changedthe potential for ventilation. First, the flow of gasdelivered to the patient was controlled directly froma valve that could be opened and shut rapidly.Second, the control of the valve was via a micro-processor. Flow volume and pressure were measuredin inspiration and expired gas. The aperture of thevalve was controlled by feedback from the flow andvolume sensors, and this could be adjusted in milli-seconds. The advent of this form of technologyenabled the ventilator to be controlled in a myriad ofdifferent ways.

Although modern ventilators are capable of gen-erating complex ventilatory modes, they are of avery simple conceptual design (Fig. 2.1). A source ofhigh-pressure air and oxygen (O2) is connected toa blender by a solenoid valve, the flow from whichis controlled by a microprocessor. The operatorpre-sets the characteristics of the inspiratory andexpiratory phase of the ventilator. By constantlychecking measured parameters such as flow, vol-ume and pressure, the ventilator delivers the breathto the patient.

22 Mechanical ventilation: the basics

Figure 2.1 A simplified diagram of

a modern ventilator. Sensors that

measure flow and pressure control

the position of the solenoid valve in

the form of a feedback loop.

The ventilator may be used to perform the normalwork of breathing when either the respiratorymuscles are weak or the motor nerve function iscompromised. Additionally, despite normal respira-tory muscles and motor output, when the workinvolved in breathing is raised, ventilation can be usedto perform some of the excessive work.

A simple task for the ventilator would be to providea constant flow rate over a given time. The valveis therefore adjusted to ensure that the flow is notaffected by changes in airway pressure. A more com-plex task is to provide a constant pressure over a giventime period. In this case, the valve may need to be fullyopen at the beginning of inspiration to produce a con-stant pressure in the airway. Later on, the flow rate mayhave to be decreased substantially as soon as the targetpressure is achieved, and flow rates may need to beincreased or decreased rapidly. The ventilator has theability to produce a variety of pressure waveformsby rapidly altering flow and sensing pressure. Forexample, an accelerating or decelerating pattern ofinspiratory flow can be used to adjust peak airwaypressure, and this feature is used when the lungs are stiff.

INDICATIONS FOR MECHANICALVENTILATION

Mechanical ventilation is indicated where establishedor impending respiratory failure exists, which isdefined as the inability of the breathing apparatus tomaintain normal gas exchange. Mechanical ventila-tion is used most commonly following major surgery

when haemodynamic variables are being normal-ized. The majority of ventilated patients are post-operative patients (>65% of all patients) who have hadmajor surgery such as cardiac, aortic or neurosurgeryand they rarely need mechanical ventilation for morethan 24 hours (Table 2.1).2 The other major groupsrequiring ventilation are patients with head or chesttrauma (<10%), poisoning (<8%) and the criticallyill with severe primary respiratory disease (<13%).

Table 2.1 Main indication for mechanical ventilationin the adult

Respiratory impairment (parenchyma!, airway or chest wall)• Pneumonia, asthma, lung contusion• Acute exacerbation in chronic bronchitis or emphysema• Adult respiratory distress syndrome, cystic fibrosis• Chest trauma with flail segment, ruptured diaphragm• Chest wall burns, kyphoscoliosis

Central nervous system or neuro-muscular impairment• Drug overdose: narcotics, anaesthetics, barbiturates• Trauma, meningoencephalitis, tumours, infarction• Brain oedema, raised intracranial pressure• Intracranial bleed, status epilepticus, tetanus, rabies• Central hypoventilation• Polyneuritis, Guillain-Barre, Lambert-Eaton• Myasthenia gravis, myopathies, paralysing poisons

Circulatory failure• Cardiac arrest, severe shock (sepsis or other causes)• Left ventricular failure (pulmonary oedema)

Airway pressure 23

AIRWAY PRESSURE

Airway pressure is dependent on two main factors:how stiff the lungs are and the resistance to airflow.3

As gas enters the lung, part of the pressure requiredovercomes flow resistance within the branching air-ways and part of the pressure is related to the elasticproperties of the system (Fig. 2.2). The stiffness of thelungs and chest wall is the compliance of therespiratory system. Pressure is proportional tovolume X 1/total compliance. Airway pressure is alsoproportional to the flow through the branching air-ways. Pressure is proportional to flow X resistance.

Airway pressure = (volume X 1/compliance)+ (inspiratoryflow X resistance)

The relationship between airway pressure flow andcompliance is known as the equation of motion and isexpressed as three simultaneous graphs, of pressure,volume and flow, running together with time (Fig. 2.3).Many ventilators have such displays available at thebedside. Most modes of ventilation control one vari-able of the equation of motion, with the remainingvariables being dependent. Compliance and resistance

Figure 2.2 When gas is pushed into the chest, the airway pressure

rises to overcome the elastic recoil of the lung and chest wall and

the resistance to airflow through the major branching airways.

can be measured at the bedside during mechanicalventilation.4 If no breathing effort is made, the ventila-tor performs the work to inflate the chest. Much infor-mation can be obtained from the pressure waveformgenerated during constant flow ventilation.

At the beginning of inspiration, very little (if any)volume has entered the chest, yet there is a rapidupstroke of the airway pressure graph. The early rise inairway pressure occurs because the ventilator has toovercome resistance to airflow (Fig. 2.4). An estimateof resistance can be calculated from the initial upstroke

Figure 2.3 Simple breath patterns

during mechanical ventilation. The

left-hand panel shows a constant

pressure during inspiration. The

right-hand panel shows a constant

flow delivered to the patient.


Figure 2.4 During constant flow ventilation, airway pressure rises

as time progresses in inspiration. (A) Early in inspiration, the rise in

pressure occurs without a large change in lung volume, and this is due

to the resistance of the branching airways. (B) As time continues, airway

pressure increases and includes the effect of the recoil of the lung and

chest wall. (C) At the end of inspiration, the pressure due to the

stiffness of the lung and chest wall (triangular-shaped element) can

be calculated from either an inspiratory breath hold or by measuring

the height of the triangular part of the diagram.

of the pressure waveform. If flow is constant, and set bythe user, it is therefore easy to obtain from the ventila-tor display. If the height of the initial upstroke in air-way pressure (in cmH2O) is divided by the flow, theresistance (in cmH2O L-1 s-1) can be easily calculated.

As time progresses, gas enters the chest and the air-way pressure then rises as an upward ramp. The totalpressure at this time is the sum of the resistance andelastic forces. From the graphs, it can be seen that theresistance to airflow is a constant increment, and this isrepresented as equal-sized bars in Figure 2.4. At the endof inspiration, the elastic recoil of the lungs and chestwall is maximal, and the pressure generated peaks.

The effective compliance of the system (lung andchest wall) can be calculated from the volume deliv-ered and the pressure during an inspiratory breathhold, at zero gas flow. When the breath is held ininspiration, pressure drops to equal the static recoilholding the chest open. This can either be measuredor estimated as the height of the triangular diagram.Compliance is then calculated as the volume per unitpressure (mL cmH2O-1).5

MODES OF VENTILATION

The way in which the ventilator delivers a breath tothe patient is described by a confusing nomenclature.As the ventilator is controlled by a microprocessor,numerous possible combinations of breaths are pos-sible, and it is estimated that over 22 different modesof ventilation have been introduced into clinical prac-

Table 2.2 A descriptive classification of ventilators by theway in which the breath is delivered to the patient

1 How is the breath controlled by the ventilator?• Pressure control• Volume control• Time control

2 How is the mandatory breath delivered to the patient?• All breaths are mandatory• Mandatory breaths are intermittent• No mandatory breaths

3 What type of inspiratory trigger is used?• Pressure triggering• Flow triggering• No triggered breaths

4 Are the spontaneous breaths assisted?• Pressure support

5 Is expiration assisted or augmented?• CPAP

tice. Although the range of potential breath typesdelivered to the patient is large, few comparative clin-ical trials have been performed relating the mode ofventilation with eventual outcome, sometimes mak-ing the rational use of such modes difficult.

Often, a mode of ventilation has many synonym-ous terms. A simple example is the description of fullmechanical ventilation. Controlled mandatory venti-lation (CMV) is the term used to describe ventilationin which all the breaths delivered to the patient arefixed by the ventilator. In the UK, this is oftenreferred to as intermittent positive pressure ventila-tion (IPPV). Assist control ventilation is used as aterm to describe a mandatory form of ventilation inwhich all the breaths delivered to the patient areassisted. However, these terms are confusing, as theyare vague and ill defined.

Ventilators can be described in terms of the way inwhich the breath is delivered to the patient. The vari-able that is controlled could be volume, pressure ortime. Each breath from the ventilator can be manda-tory or spontaneous, and the breath can be triggeredby the machine or the patient, or both (Table 2.2).6

Volume control

In this mode, volume is controlled and the otherfactors from the equation of motion will vary. The

Partial assistance 25

ventilator seeks to deliver the pre-set volume to thepatient, and this is an advantage in most circum-stances in which the patient is not able to make aninspiratory effort and the underlying lungmechanics are normal, as is the case for patientsundergoing general anaesthesia. However, pressureat the airway is not controlled and will depend onthe dynamics of the system. If the lung and chestwall are stiff, or if airway resistance is high, pres-sure will rise and, unless this is limited, excessivepressures may result.

Recently, the effect of one form of mechanicalventilation was investigated in patients with acuterespiratory distress syndrome (ARDS). In thisstudy, the volume delivered to the patients wascontrolled and the results were compared withthose from a group who received higher tidal vol-umes. Importantly, mortality when tidal volumewas restricted was less than with the standardtreatment. With this exception, few other modes ofventilation have had a demonstrable effect onpatient outcome.7'8

Pressure control

During pressure control ventilation, the airway pres-sure is pre-set and the volume delivered to thepatient is dependent on the mechanics of the system.Tidal volume may vary, and will be high when thesystem has a low resistance and a high compliance. Ifresistance and compliance alter, as they may dorapidly in the critically ill, a changing delivered vol-ume to the patient occurs. As pressure is pre-set, highairway pressures are avoided, and this may be advan-tageous in the presence of stiff lungs with the ever-present danger of barrotrauma.

In the critically ill patient, the lung is not homoge-neous in nature; frequently, there are areas of thelung that are stiffer or have a higher resistance to gasflow. Ventilation to the stiffer lung units may requireprolonged inspiratory times. Pressure control exertsa constant pressure and may succeed in expandingareas of the lung, particularly if inspiration is pro-longed. This may be contrasted to volume control,with which much of the initial volume is delivered tothe few remaining lung units whose characteristicsare more normal. This may cause over-distension ofthe normal lung units and may not inflate the stifferlung units.

PARTIAL ASSISTANCE

In the critically ill, spontaneous breathing is oftenmaintained and ventilatory support is provided as aproportion of the total effort. Another force is actingon the lungs and must be added to our simple dia-gram, that of the patient's respiratory muscles, Pmus.Pmus is acting to inflate the lungs and, if the ventila-tor is in total harmony with the patient, Pmus andPaw will work together and are additive (Fig. 2.5).

Triggering the ventilator

The mechanism by which the ventilator senses aninspiratory effort is termed the trigger, and thechange in airway pressure during inspiration is themethod that is most often used today. Pressure gen-erated by the respiratory muscles within the chest istransmitted into the upper airway and is detected inthe mouth. In normal subjects, there is little phasedifference between pleural pressure and mouth pres-sure. Pressure is easily measured using a transducerand most methods used are robust. The transducermeasures airway pressure continuously and is pro-grammed to detect the change in pressure that occursduring inspiration. The trigger can be adjusted toreact to a smaller change in airway pressure, makingit more sensitive. This enables the ventilator to betriggered closer to the beginning of the breath.

All trigger systems have a time delay from themoment inspiration begins to the point when the

Figure 2.5 When the patient is breathing spontaneously, the

ventilator assists the breathing effort. If the ventilator and the

patient efforts are synchronous, the effects can be seen as additive.


ventilator delivers a breath. The rate of response of atrigger can be divided generally into two phases.Phase I describes the time during which the pressurebuilds up to the trigger point and Phase II describesthe time from the trigger point to the opening of theinspiratory valve and breath delivery.9

If the trigger is insensitive or if the reaction time isdelayed, the patient will be breathing against a closedairway, and this ineffective ventilation can amount toa considerable amount of work for the patient.Pressure triggering may occur from other fluctu-ations of pressure within the airway and it is possibleto trigger a ventilator from cardiac oscillations thatappear as artefacts.

In lung pathology, e.g. chronic obstructive pul-monary disease (COPD), the rate of transmission ofa pressure signal from the pleural space to the airwaymay be delayed and a phase difference may occur.This slows the response of the ventilator.10

Inspiratory flow may also be detected and used asa trigger. Two methods are employed to measureflow. Flow may be measured directly at the mouth ora continuous low flow may be introduced by the ven-tilator during expiration (bias flow) and the differ-ence between inspiratory and expiratory flowcontinuously measured. When inspiration begins,some of the bias flow enters the lungs and less of thebias flow is detected at the expiratory sensor of thepatient breathing circuit. This inspiratory/expiratoryflow difference triggers the breath. Flow triggeringtends to increase the sensitivity of ventilators com-pared with pressure triggering, and the time delayfrom the start of the breath to delivery to the patientcan be reduced considerably.

Intermittent mandatory ventilation

Partial assistance can be delivered to the patient bydelivering a pre-set tidal volume at a pre-set flowrate. The number of breaths is also pre-set, and thisform of ventilation was originally termed intermit-tent mandatory ventilation (IMV).

Several problems occur in adults with IMV. Theventilator breath is not timed to spontaneous breath-ing and the machine breath may coincide with anyphase of spontaneous respiration. If the patient isbreathing out when the ventilator delivers an inspi-ratory breath, high airway pressure may be obtained.Furthermore, if the patient receives a number of

machine breaths together, there may be insufficienttime for adequate exhalation, leading to an increasein functional residual capacity (FRC) and the phe-nomenon of 'breath stacking'. In order to solve theseproblems, the ventilator needs to 'know' when thepatient is breathing and which phase of breathing isoccurring. Synchronized IMV (SIMV) senses inspir-ation and has the capacity to trigger a breath.

In order to make SIMV work effectively, the venti-lator must be inhibited from delivering a machinebreath during the initial phase of exhalation, andshould deliver a breath in synchrony with a spontan-eous breath. During SIMV, the ventilator is inhibitedat the first part of the expiratory cycle. The amountof inhibition is related to expiratory time. At the startof exhalation, the ventilator is insensitive but, as timeprogresses, it becomes more responsive. DuringSIMV, triggering is achieved by monitoring either thepressure generated in the airway or the flow thatoccurs during a spontaneous breath.'

Delivering a volume to the patient from the venti-lator does not imply that all work previously per-formed by the respiratory muscles is now providedby the machine. Inspiratory efforts often occurthroughout a mandatory machine breath, especiallywhen the inspiratory flow rate from the ventilator isslow. This is common in COPD, when inspiration isdifficult to sense by the ventilator. When machineinspiratory flow is low, the respiratory muscles maycontract during the whole of the ventilator breath, atthe same level that occurs during spontaneousbreathing. The implication is that respiratory musclerest may not occur, despite the ventilator providingmachine breaths that are designed to provide forrespiratory muscle work.

Pressure support ventilation

Pressure support ventilation (PSV) has several differ-ences compared with SIMV. During PSV, the onlyvariable that is pre-set is the target airway pressure.11

The patient determines the respiratory rate and PSVis triggered from airway pressure or from airwayflow. When inspiration is sensed, the target pressureis established by delivering maximum flow to the air-way, which is then reduced as soon as the target pres-sure is achieved. As inspiration continues, a constantpressure is maintained. The duration of inspirationis also related to patient effort. Inspiration is most

Recent advances in mechanical ventilation 27

commonly terminated when inspiratory flow isreduced to 25% of the maximum inspiratory flowrate. PSV may also cycle to expiration if the airwaypressure exceeds the pre-set level of support, or wheninspiratory time is excessive or in response to otherparameters designed to prevent excessive inspijatorytimes.

PSV is controlled by the patient, and the rate andsize of tidal volume are not pre-set by the user. Tidalvolume is not assured, and this may change beyondthat which is clinically acceptable. During sleep, or ifcentrally acting drugs are administered, PSV will notmaintain a minimum level of ventilation.

It is possible either that PSV may be deliveringmore assistance than is needed or that support isinsufficient. The correct amount of pressure supportis difficult to judge and, as with SIMV, chronicunder-ventilation or over-ventilation is possible.

Proportional assist ventilation

The equation of motion has been used so far todescribe how ventilators work when one element(volume, pressure or flow) is controlled. Pro-portional assist ventilation (PAV) is an attempt to solvethe equation of motion rather than just to controlone respiratory variable.12 The pressure required toachieve a tidal volume is, in the partially assistedpatient, related to the amount of pressure deliveredby the ventilator and the amount of pressure gener-ated by the respiratory muscles. The ventilator maydeliver most of the pressure and the respiratorymuscles would then be unloaded and the amountof work performed by the muscles would be mini-mal. The adjustment that is set by the user is thedegree of off-loading that the ventilator delivers(expressed either as a ratio or a percentage).13

Compliance and resistance can be measured bythe ventilator. As the ventilator also measures thetidal volume that is delivered, the pressure that isgenerated can be calculated. If the patient werebreathing spontaneously, the pressure calculatedwould then represent the pressure developed by therespiratory muscles. The ventilator can be instructedto provide some of this pressure, and this can be var-ied from all of the pressure (100%) through to noneof the pressure. In the simplest version of PAV, valuesfor compliance and resistance can be pre-set, or maybe taken from measurements during full mechanical

ventilation. More usefully, measurements of compli-ance and resistance, measured on-line over milli-second time frames, enable the total amount ofpressure required to be measured simultaneouslyduring a breath. If the tidal volume changes, so theamount of pressure changes, reducing for a smallbreath and increasing for a large breath. Whatremains constant is the amount the ventilator pro-vides as a fraction of maximum.

RECENT ADVANCES IN MECHANICALVENTILATION

Greater importance is now placed on maintainingspontaneous ventilation during critical illness. Thedesign of the expiratory valve now enables ventila-tors to provide cyclical changes in airway pressureand patients can superimpose their own breathingpattern 'on top' of the changing level of backgroundpressure and each spontaneous breath can be add-itionally assisted.

The advent of the modern ventilator has meantthere has been a rapid increase in the number of'modes' or patterns of mechanical ventilation.Although many studies have investigated intermedi-ate outcome of new ventilator modes, few of thesemodes have been the subject of prospective trials andthere is less evidence to suggest one form of ventila-tion is superior to another.

Positive end-expiratory pressure

In health, the lungs are held open by the outwardrecoil of the chest wall. At the end of expiration, thetendency of the lung to collapse is balanced by theoutward recoil of the chest wall, such that lung vol-ume is at steady state. The resting end-expiratoryvolume is termed functional residual capacity. FRCreduces when body posture is changed to the supineposition. During anaesthesia and after sedation, FRCfalls and the critically ill patient therefore tendstowards a FRC which is substantially reduced.14

The physiological effect of a low FRC is dependenton the state of the small dependent airways andalveoli. At some stage, airways close and at this pointhypoxaemia is common. The application of a posi-tive background pressure throughout the respiratory


cycle has the effect of increasing lung volume, splint-ing the airways open and preventing lung volumefalling to low levels at which airway closure is pos-sible. Such pressure is termed continuous positiveairway pressure (CPAP) if the patient is breathingspontaneously or positive end-expiratory pressure(PEEP) if this is applied during positive pressureventilation.

PEEP has been used extensively to treat hypox-aemia and is most effective when lung volume isrecruited. This effect is gradual, with improvementsin arterial blood gas measurements 1 hour afterchanging PEEP/CPAP. On the other hand,de recruitment tends to happen promptly.

Periodically, recruitment of lung units mayrequire much higher pressures to be applied to thelungs to open collapsed or partially inflated lungunits, and PEEP is important to maintain lung vol-ume once alveoli have been recruited, the so-calledopen lung strategy.15 Such forms of lung ventilationmay have complex interactions and need to beapplied with care.

Increasing the mean pressure in the chest also hasan effect on the mediastinum and cardiac chambers.The effect of raising intrathoracic pressure wasoriginally considered to act mainly on the venoussystem, impeding filling of the ventricles andend-diastolic volume and hence reducing stroke vol-ume and cardiac output. The relationship betweenthe heart and the lungs is now recognized to be morecomplex and may alter when ventricular function isdiminished. Nonetheless, it is important to note thatraised intrathoracic pressure may have a deleteriouseffect on cardiac output, and the net effect may be todepress O2 delivery to the tissues.

Non-invasive ventilation

Mechanical ventilation does not require endo-tracheal intubation, and respiratory support is fre-quently provided with a tightly fitting facemaskduring resuscitation and by anaesthetists in theoperating theatre. This technique has been adaptedin order to support patients on the intensive careunit (ICU), using a simple ventilator and acceptingthat some of the ventilated gas may leak away. Suchsupport is non-invasive and has many advantages.The patients do not need to be intubated, are notsedated and may spontaneously cough and expecto-

rate. Additionally, the facemask may be removedintermittently and the patient may be able to speakand communicate during therapy. This type oftreatment is now termed non-invasive ventilation(NIV). NIV is a generic term used to describe anyform of positive pressure facemask therapy; it beginswith CPAP and includes volume or pressure con-trolled mechanical ventilation. NIV was initiallyused for patients with severe chronic hypercapnicrespiratory failure as an alternative to mechanicalventilation with an artificial airway. Success in thisarea is widespread and the treatment is no longerrestricted to the ICU, ward-based care now beingusual. Frequently, patients require long-term treat-ment and domiciliary NIV is provided with specific-ally designed ventilators.

NIV is used for acutely ill patients in respiratoryfailure. It is particularly successful when used at anearly stage when the patient can co-operate withtreatment effectively. Several studies have demon-strated that NIV decreases morbidity and mortalityin acute ventilatory failure due to COPD, possiblydue to the decrease in pneumonia associated withendotracheal intubation.16

However, not all patients are suitable for this formof therapy. Patients with a decreased level of con-sciousness, those who are unable to cough andexpectorate or to swallow or protect their upper air-way are difficult to manage with NIV and may needto be intubated. NIV demands an effective team ofnurses, doctors and physiotherapists. In the acutelyill, the first few hours of treatment demand intensiveinput from the team. The case for NIV in hypoxaemicrespiratory failure is more complex, some studieshaving suggested that treatment may not preventmore invasive therapy involving endotracheal intu-bation.

PRACTICAL ASPECTS OF MECHANICALVENTILATION

Some dangers of mechanical ventilation apply to allpatients. It is not possible to establish effective, long-term ventilation without securing a sealed connec-tion with the airway via an endotracheal ortracheostomy tube. However, the insertion of thistube requires either general or local anaesthesia, withits attendant risks.

Practical aspects of mechanical ventilation 29

Anaesthesia

The risks of the anaesthesia needed for endotrachealintubation include:

• myocardial depression caused by general or localanaesthetic drugs,

• aspiration of gastric contents,• a further fall in arterial O2 tension, especially if

intubation is difficult,• an idiosyncratic reaction to anaesthetic drugs,• reflex worsening of bronchoconstriction follow-

ing tracheal intubation or suction of secretions.

These risks are not substantially reduced if topicallocal, instead of general, anaesthetic is used for theintubation of the trachea.

Sedation and paralysis

IPPV through a nasal or an orotracheal tube ispoorly tolerated without some sedation. Often,paralysing drugs are also required initially, especiallyin neurological disease (status epilepticus, for exam-ple) or trauma. In general, the ideal sedative shouldbe very short acting, given by constant intravenousinfusion and have minimal side effects, especiallyaffecting the circulation. None of the currently avail-able sedatives is devoid of important side effects. Thebenzodiazepines often cause tolerance, requiringincreasing dosage that leads to a build up of activemetabolites and prolonged depressant effects on thecentral nervous system, which last for days after stop-ping administration, especially in those patients withrenal or hepatic failure. Propofol, an intravenousanaesthetic, and isoflurane, a volatile inhalationanaesthetic, have been used for continuous sedation.They are both short-acting agents with no cumula-tive effects and their cardiovascular and respiratoryeffects at sedative doses are minimal in the fit patient.However, the cardiac depressant effects of these twodrugs may be important in the patient with poormyocardial function.

There is a wide choice of suitable neuromuscularblocking drugs: vecuronium and atracurium haveminimal side effects and are sufficiently short actingto allow rapid regulation of the state of paralysis. AllICU staff should be aware that neuromuscularblocking agents have no sedative effects and thatpatients may be awake and paralysed if sedation is

not prescribed. Another danger of paralysis is theinability of the patient to make spontaneous breath-ing efforts should there be an accidental ventilatordisconnection.

Equipment failure

The risks of equipment failure include accidentaldisconnection of the ventilator, undetected leaks ormalfunction of the endotracheal tube, all leading toalveolar hypoventilation and hypercapnia.

Hyperinflation

Airflow obstruction is defined by the failure of lungemptying and the consequent increase in restinglung volume. At the higher volume, the pressure inthe peripheral airway is increased. This pressure isdifficult to measure and is termed occult or intrinsicpositive end-expiratory pressure or autoPEEP. Thiseffect may be worsened by mechanical ventilationwhen the mandatory machine breaths may be set insuch a way as to allow insufficient time for lungdeflation. Subsequent breaths delivered from theventilator increase lung volume until a new equilib-rium is established.

Endotracheal tubes

Mechanical ventilation often requires tracheal intub-ation and this is a hazardous moment for any critic-ally ill patient. Anaesthetic drugs always depress thecirculation and neuromuscular paralysis removes thelaryngeal reflexes that may protect against aspiration.Intubation may not be technically straightforwardand patients with inadequate lung volume desaturatequickly. A checklist for items needed during intub-ation is useful.

All artificial airways offer a resistance to airflow.The flow of air through the tube is seldom ideal andnever laminar in nature. Turbulent airflow is com-mon, due to the shortness of the tube relative to theinternal diameter, the presence of secretions and thepossibility of changes in diameter due to kinking.

The work performed by gas flowing through theendotracheal tube depends largely on the diameterof the tube and on the inspiratory flow rate.


The highest work is seen when the internal diameteris 7.0 mm or less, with a minute ventilation greaterthan 15 L min-1. When inserted, endotracheal tubesin adults are seldom less than 7.00 mm and func-tional tube diameter changes with use. If secretionsare present inside the tube, it is possible for a largerendotracheal tube to function as one of a muchsmaller diameter. This phenomenon is most promi-nent after 7 days of use.

As the airway resistance due to the endotrachealtube increases, several effects occur. When thepatient initiates a breath, work is performed acrossthe tube. As a result, inspiration occurs before theventilator can sense a pressure change in the airway,and a time delay occurs such that the respiratorymuscles contract before the ventilator senses thebeginning of inspiration. Also, when the ventilatordoes trigger, some of the pressure generated is againdissipated across the endotracheal apparatus andtherefore less is available for lung inflation.

It may be possible to remove the effect of theendotracheal tube if airway pressure is sensed closerto the patient. Inspiration is sensed at the earliestpossible moment and the work to breath across thetube itself is reduced if the target pressure is alsosensed at the distal endotracheal tube. A system ofcompensation for the resistance of endotrachealtubes is a facility of modern ventilators. Detailsconcerning the type of tube may be entered into theventilator programme and the resistance of the tubeis automatically compensated for.17

The position of the endotracheal tube within theairway frequently requires attention. The plasticnature of the tube tends to soften when it reachesbody temperature and the tube adopts a configur-ation that is acutely angled. It is possible for thepatient to move the tube to the point where the cuffof the tube lies above the larynx, with only the tipof the tube within the airway. When this happens,air leaks can be heard. A temporary seal can be re-established by further inflation of the cuff. Eventually,the over-inflated cuff sits at the back of the throat,with the endotracheal tube resting above the larynx,and the airway may not in fact be intubated. In thiscircumstance, it is often necessary to re-intubate thepatient.

If humidification of the inspired gases is notadequate, endotracheal tubes may become acutelyblocked with secretions, which can completelyocclude the airway.

Endotracheal tubes may migrate beyond thecarina, entering a major bronchus. This results inpartial or complete occlusion, with collapse of theoccluded lung and hypoxaemia.

Nosocomial infection

Mechanical ventilation with sedation and theupper airway bypassed with an endotracheal tubecarries with it a high risk of infection. Despiteimmaculate care of the upper airway, secretionscollect at the back of the pharynx and can be silentlyaspirated into the lungs. Additionally, manypatients receive some form of sedation duringintubation and ventilation, with the result that thecough reflex is impaired. Ciliary clearance isreduced when the airway is intubated. Pathogensthat are present in such ventilator-acquired pneu-monias (VAPs) may often be of the hospital-acquired type and may be difficult to treateffectively. It has been noted that pneumoniasacquired in this manner may be reduced if thepharynx is continuously aspirated with the aid ofspecially designed endotracheal tubes. If the supineposture is changed to a head-up position, the inci-dence of VAPs decreases.

Cardiovascular effects of intermittentpositive pressure ventilation

Not all the effects of IPPV on the cardiovascularsystem are adverse. They result from the rise inintrathoracic pressure, especially if PEEP is used, andare mediated through direct mechanical interferencewith the heart, through indirect reflexes of the auto-nomic nervous system and through hormone releaseor changes in blood gases. The predominant directadverse effects of IPPV upon the right heart are areduction in venous return (pre-load) and anincrease in pulmonary vascular resistance (after-load). The direct effects upon the left heart are lessmarked and less well established, the widely heldview being that both pulmonary venous return (pre-load) and after-load decrease. This effect on leftventricular after-load is due to a fall in ventriculartransmural pressure because of the increase inintrathoracic pressure (this also applies to the rightventricle). This mechanism provides a form of'assis-

References 31

tance' to ventricular work that may be beneficial incardiac failure.18

The reflex responses are complex, dependingupon multiple neural and chemical feedback loops.The neural reflexes are mediated initially by thevagus nerve, predominantly affecting the heart rate,but stronger reflexes involve the whole sympatheticsystem, affecting vascular resistances and circulat-ing catecholamines. The reflexes originate fromlung and atrial stretch receptors and from the arter-ial baroreceptors and chemoreceptors (the latteronly if arterial CO2 tension falls or arterial O2 ten-sion rises in response to IPPV). The humoral reflexresponse to IPPV includes an increase in antidiuretichormone (ADH) and reninangiotensin and adecrease in atrial natriuretic peptide, which may bepartly responsible for the sodium retention seen inventilated patients. The changes in catecholaminesare partly mediated by PaCO2 changes. The patternof circulatory changes is variable. The predominanteffect is a decrease in both cardiac output (typicallyby 25%) and arterial blood pressure, an increase inheart rate and a slight increase in systemic vascularresistance; right and left atrial pressures increaserelative to atmospheric pressure (transmural pres-sures decrease). This pattern is often modified byblood-gas changes associated with mechanical ven-tilation because of the powerful stimulant effects ofCO2 upon the sympathetic system.

REFERENCES

1. Snider, GL Historical perspective on mechanical

ventilation: from simple life support system

to ethical dilemma. Am Rev Respir Dis 1989;

140: S2-7.

2. Esteban, A, Anzueto, A, Alia, I, et al. How is mechanical

ventilation employed in the Intensive Care unit? Am J

Respir Crit Care Med 2000; 161: 1450-8.

3. Mead, J, Milic-Emili, J. (1964). Theory and

methodology in respiratory mechanics with glossary

and symbols. In Handbook of physiology. Volume 1,

Respiration. Washington: American Physiological

Society, 1964; 363-6.

4. Gattinoni, L, Mascheroni, D, Basilico, E, Foti, G, Pesenti, A,

Avalli, L. Volume/pressure curve of the total respiratory

system in paralyzed patients: artefacts and correction

factors. Intensive Care Med 1987; 13:19-25.

5. Tobin, MJ, Van de Graaff, WB. (1994). Monitoring of

lung mechanics and work of breathing. In Principles

and practice of mechanical ventilation. New York:

McGrawHill, 1994; 1300.

6. Branson, RD, Chatburn RL. Technical description and

classification of modes of ventilator operation. Respir

Core 1992; 37: 1026-44.

7. The Acute Respiratory Distress Syndrome Network.

Ventilation with lower tidal volumes as compared to

traditional tidal volumes for All and ARDS. N EnglJ

Med 2000; 342: 1301-8.

8. Amato, et al. Beneficial effects of the open lung

approach with low distending pressures in acute

respiratory distress syndrome. A prospective

randomized study on mechanical ventilation. Am J


9. Sassoon, CSH. Mechanical ventilator design and

function: the trigger variable. Respir Care 1992; 37:

1056-69.

10. Murciano, D, Aubier, M, Bussi, S, Derenne, JP, Pariente,

R, Milic-Emili, J. Comparison of esophageal, tracheal

and occlusion pressure in patients with chronic

obstructive pulmonary disease during acute

respiratory failure. Am Rev Respir Dis 1982; 126:

837-41.

11. Maclntyre, NR. Respiratory function during pressure

support ventilation. Chest 1986; 89: 677-83.

12. Younes, M. Proportional assist ventilation, a new

approach to ventilatory support. Am Rev Respir Dis

1992; 145: 114-20.

13. Younes, M, Puddy, A, Roberts, D, et al. Proportional

assist ventilation; results of an initial clinical trial. Am

Rev Respir Dis 1992; 145: 121-9.

14. Sykes, K. Respiratory support. London: BMJ Publishing,

1995.

15. Lachmann, B. Open the lung and keep the lung open.


16. Vitacca, M, Rubini, F, Foglio, K, Scalvini, S, Nava, S,

Ambrosino, N. Non-invasive modalities of positive

pressure ventilation improve the outcome of acute

exacerbations in COLD patients. Intensive Care Med

1993; 19:450-5.

17. Guttman, J, Bernard, H, Mols, G, et al. Respiratory

comfort of automatic tube compensation and

inspiratory pressure support in conscious humans.


18. Lemaire, F, Teboul, JL, Cinotti, L, et al. Acute left

ventricular dysfunction during unsuccessful weaning

from mechanical ventilation. Anesthesiology 1988;

69:171-9.

3Mechanical ventilation: ventilatorystrategiesHILMARBURCHARDI

Ventilatory strategies for acute lung injury 32

Ventilatory strategies for acute bronchial asthmaand acute exacerbation of chronic obstructivepulmonary disease 36

Mechanical ventilation in brain injury

Mechanical ventilation in cardiac failure

Special considerations

References

40

44

44

45

VENTILATORY STRATEGIES FOR ACUTELUNG INJURY

Pathophysiological conditions

In acute lung injury and early acute respiratory distresssyndrome (ARDS), pulmonary membrane per-meability is critically increased as a result of an acutesystematic inflammatory reaction, either due to directtissue damage (e.g. aspiration) and/or indirectly (e.g.sepsis). This results in a severe interstitial and intra-alveolar non-cardiogenic oedema.1 Under the influenceof gravity, the fluid-filled lung tissue is compressed,particularly in the dependent parts,2 and the surfacearea for gas exchange reduced ('baby lung'), with severeimpairment of oxygenation. Surfactant production isimpaired so that alveoli tend to collapse. Gas exchangeis still maintained in the aerated non-dependent areasof the lungs, although these potentially may bedamaged by hyperinflation, and the compressed orfluid-filled alveoli can be re-aerated by lung recruitmentas long as interstitial fibrosis has not yet occurred.

The abnormalities in respiratory mechanics dependupon the underlying aetiology.3 In primary ARDS (i.e.pulmonary, such as pneumonia), lung compliance is

decreased but the chest-wall component remainsnormal. In secondary ARDS (i.e. extra-pulmonary, forexample due to surgical causes), chest-wall compli-ance is decreased (e.g. by abdominal distension),whereas lung compliance may be preserved.

There is also increasing evidence that mechanicalventilation may itself damage the lungs. The mechan-isms of this ventilator-associated lung injury (VALI) arevarious:4 regional alveolar over-distension, caused byhigh inspiratory pressure and/or volume, iscompounded by the inhomogeneity of lung injury andgenerates damaging shear forces and, in some areas,repeated opening and closing of collapsed alveoli, whichfurther damages the lung and increases microvascularpermeability ('stress failure') and oedema. Thesepotential risks and adverse effects consequentlydetermine the strategy of mechanical ventilation.

Ventilatory principles

The main principles for mechanical ventilation inacute lung injury are:

• 'open up the lungs and keep them open',• limiting airway pressure and tidal volume,


• permissive hypercapnia,• variation of the inspiration:expiration (I:E) ratio,• supplementary spontaneous breathing efforts.

'OPEN UP THE LUNGS AND KEEP THEM OPEN'

To recruit collapsed alveoli requires a higher pressurethan is necessary to keep them open. The clinical aimis to 'open up the lungs and keep them open'. Theapplication of external positive end-expiratory pres-sure (PEEP) increases lung volume and can recruitcollapsed alveoli. The distribution of additional gasvolume induced by external PEEP will depend on theregional compliance of different lung areas, whichmay be very variable in inhomogeneous lungs. Thus,it is unlikely that one level of PEEP will be optimal forthe whole lung. In ARDS, any increase in lung volumeby external PEEP not only reduces intrapulmonaryshunt, but also simultaneously hyperinflates non-compressed lung areas, thereby potentially convertingwell-ventilated alveoli into non-perfused dead space.

Consequently, a compromise is necessary thatkeeps the lesser compliant alveoli open without over-distending the more compliant areas. Some inten-sivists recommend applying external PEEP slightly(e.g. 2 cmH2O) above the lower inflection point(Pflex) of the pressure/volume (P/V) curve. However,this is cumbersome and problematic to measure andsometimes no inflection point can be identified. (Fora fuller discussion, see Chapter 12.) PEEP inducesmore recruitment in secondary or extrapulmonaryARDS than in primary or pulmonary causes ofARDS. The increase in intra-abdominal counter-pressure is important in secondary ARDS and it hasbeen proposed that measurement of intra-abdominalpressure (IAP) may better define the optimal level ofPEEP than estimating Pflex.

An alternative lung recruitment strategy is to stat-ically distend the lungs with high continuous posi-tive airway pressures (CPAP), e.g. 35-40 cmH2O for40 s followed by a return to previous PEEP levels.5

Convincing evidence for this potentially riskymanoeuvre is not available.

LIMITING AIRWAY PRESSURE AND TIDALVOLUME

Transpulmonary pressure should be kept within thenormal range that applies at maximum lung cap-acity. This corresponds to a maximum airway plateau

pressure of about 35 cmH2O.4 In ARDS 'baby lungs',the consequence is a need to reduce tidal volumeto avoid high inflation pressures and alveolar over-distension. Tidal volumes of 12-15 mL kg-1, as for-merly proposed, cause over-distension. The tidalvolume must also be adjusted to the prevailing PEEPlevel: when using higher PEEP, tidal volumes have tobe reduced to avoid some lung areas being in the flat-ter part of the pressure-volume curve, indicatingover-distension. Restriction of the ventilatory excur-sion may be even more important in inhomogen-eous lungs to reduce local tissue stress forces.Consequently, low tidal volumes and a higher level ofPEEP are now advised, with tidal volumes as low as6 mL kg-1 being recommended.10 The consequencemight be an increase in PCO2, although alveolarventilation can be increased by increasing frequency.There is no benefit in increasing respiratory rateabove 25 min-1, however.

PERMISSIVE HYPERCAPNIA

If arterial -PCO2 is allowed to increase from 40 to80 mmHg, alveolar ventilation can be reduced by50%. Hickling and co-workers6 were the first todemonstrate in ARDS that permissive hypercapniawas well tolerated. They showed that mortality wasreduced when peak airway pressures were cut bydecreasing tidal volumes and limiting peak inspiratorypressures to 40 cmH2O. This concept is nowcommonly accepted.7 Acute hypercapnia increasessympathetic activity, cardiac output and pulmonaryvascular resistance and dilates both bronchi andcerebral vessels. However, a slow and gradual elevationof PCO2 is remarkably well tolerated. Contra-indications are co-existing head injury or other risk ofcerebral oedema, a recent cerebrovascular accident orsignificant cardiovascular dysfunction.

VARIATION OF THE I:E RATIO

Once open, the alveoli can be kept open by externalapplied or intrinsic PEEP (autoPEEP). IntrinsicPEEP occurs when expiration (regional or total)remains incomplete at the end of the available expira-tory time. This is a dynamic phenomenon thatdepends on the actual conditions of ventilation.Thus, intrinsic PEEP can be caused by high tidal vol-umes, short expiratory time or high ventilatory timeconstants. Commonly, there is a spectrum of slower

34 Mechanical ventilation: ventilatory strategies

and faster alveolar compartments. Fast compart-ments may be able to expire completely, e.g. 0.5 s. Byshortening expiratory time, intrinsic PEEP can delib-erately be manipulated: this is the concept of theinverse ratio ventilation or IRV mode, whereby sloweralveolar compartments may be kept open by 'indi-vidual' intrinsic PEEP, i.e. using regional air trappingto prevent alveolar collapse.

IRV has been advocated in ARDS.8 The possibleadvantages are:

• the prolonged inspiration ensures a more homo-geneous ventilation,

• during the short expiration, slower compartmentswill not exhale completely and will remain dis-tended by an intrinsic PEEP.

It has been argued that the pressure-controlled mode(PC-IRV) may be better than the volume-controlledmode (VC-IRV). PC-IRV ensures that alveolar pres-sures never exceed the set value anywhere within theinhomogeneous lung. Of course, pressure limitationcould also be achieved in a volume-controlled mode bysetting a pressure limit. A more important argument forthe PC-IRV mode may be that inspiratory pressureremains constant, even if lung compliance improves byalveolar recruitment. This will lead to a further increasein tidal volume. In contrast, in VC-IRV, airway pressuredecreases when compliance improves, which reducesthe chance for further alveolar recruitment unless thepre-set tidal volume is manually re-adjusted.

However, the use of IRV, with the associatedincrease in mean intrathoracic pressure, will interferewith cardiac output. The benefit from IRV in oxy-genation may be lost by reducing O2 transport, unlessthese effects are counterbalanced by fluid volumeand/or vasoactive drug therapy. Clinical studies donot convincingly demonstrate the superiority of IRV.An important criticism of IRV is that most of theclaimed advantages are either unproven or can beachieved more safely with controlled mechanical ven-tilation (CMV) and an appropriate external PEEP.

SUPPLEMENTARY SPONTANEOUS BREATHING

EFFORTS

In modern strategies, there is a general tendency toincorporate spontaneous breathing efforts even ifclearly insufficient. Even a small contribution byspontaneous breathing reduces peak airway pres-sures. This makes mechanical ventilation less 'inva-

sive'. Venous return will also be less affected and O2

delivery may improve as a consequence of increasedcardiac output. Even more importantly, the ventila-tory movements of the diaphragm during spon-taneous breathing predominantly affect the mostdependent, and therefore most collapsed, areas of thelungs. Furthermore, sedation can be kept lower.9

This may be beneficial for several reasons:

• less disturbance of other organ function, e.g.gastrointestinal motility,

• less accumulation of sedatives,• easier dosing of analgesics to individual needs,• improved diagnosis of complications, e.g. cerebral

function disturbance,• enhanced coughing and clearance of bronchial

secretions.

Ventilatory strategies

LUNG-PROTECTIVE VENTILATION STRATEGY

Lung-protective strategies have been compared toconventional ventilation in several controlled studies.In one, reduced mortality was shown in the protectiveventilation group, but the mortality in the controlgroup was excessive.5 Another randomized studyfailed to prove the benefit of limiting tidal volumes.However, a large, multicentre study10 has providedevidence for this mode of ventilation: volumecontrolled ventilation (VCV), I:E ratio = 1:1-1:3,peak airway pressures < 30 cmH2O, tidal volumes of6 mL kg-1, PEEP levels linked to the required FiO2

(ranging from 5 cmH2O for FiO2 = 0.3 up to 24cmH2O for FiO2 = 1.0). The hospital mortality wasreduced from 40% to 31% and the number ofventilator-free days within the first 28 days increased.Furthermore, an index of lung inflammation showedan advantage from the limited tidal volume strategy.

AIRWAY PRESSURE RELEASE VENTILATION

This mode of ventilation allows spontaneous breathingon a pre-set CPAP level interrupted by short (0.5-1.5s)releases from that pressure level for further expiration.The principle of reducing rather than increasing lungvolume when applying tidal volume distinguishes thistechnique from other modes of ventilatory support. Itmaintains a moderately high airway pressure (about20-30 cmH2O) for most of the time, thereby keeping


the alveoli open. Further, during the short expiratoryrelease, slow compartments remain open by intrinsicPEEP, which therefore resembles the effect of theIRV mode.11 An essential advantage, however, is thepreservation of spontaneous breathing. At any timeduring the respiratory cycle, the system is open and thepatient cannot Tight against the ventilator'.

Airway pressure release ventilation (APRV; avail-able on the Drager EVITA ventilator) seems to beparticularly effective when ventilating ARDS lungs.12

It fulfils all the conditions of lung-protective strat-egies, but additionally (and most importantly)includes the beneficial effects of spontaneous breath-ing. Even a minimal amount of spontaneous breath-ing will contribute to an improvement in V/Qmismatch and an increase in systemic blood flow.

APRV would appear to offer several potentialadvantages in ARDS:

• a short expiratory time, which favours ventilationin the fast compartments,

• continuous airway pressure level to keep thealveoli open,

• spontaneous breathing, which avoids the need formuscle paralysis or deep sedation,

• minimal deviation from an individually adapted'optimal' lung volume, i.e. level of mean airwaypressure, which may reduce the risk of baro-trauma or volutrauma.

Although a consensus conference on mechanicalventilation13 concluded that there are no convincingdata to indicate that any ventilatery mode is superiorto another in ARDS, studies comparing the differentmodes in well-defined clinical conditions are stillneeded. However, it may be difficult to prove super-iority of any one mode by measuring outcome ifrespiratory failure is only one (and not the mostcommon) reason for a fatal outcome. An improve-ment in physiological parameters may still be a goodway to assess new strategies.

Case 1: Polytrauma with ARDS

KTh, a 21-year-old woman, suffered multipletrauma in a traffic accident, with blunt chestinjury and bilateral lung contusions. She wasintubated at the accident site. On arrival in hospi-tal, thoracic drains were inserted to treat bilateralhaemo-pneumothoraces. Because of progressiverespiratory deterioration, she was transferred thefollowing day to the university hospital withARDS. On arrival in the ICU, her arterial PO2 =67 mmHg and PCO2 - 30 mmHg and her FiO2=0.6 (Fig. 3.1; Table 3.1).

(cont.)

Figure 3.1 KTh, a 21-year-old female with polytrauma and ARDS:

(a) X-ray and (b) CT scan at arrival in our hospital. In the X-ray, the

bilateral lung contusions appear less impressive due to the bilateral

pneumothoraces, despite drainage.


Table 3.1 KTh, a 21-year-old female with polytrauma and ARDS

MVPhighPlowI:ERRFi02

P<->2PC02

APRV30

8

3:0.8180.6

67

30

APRV

27

8

3:0.8180.3

63

37

APRV

27

8

3:0.8180.4

137

36

APRV

25

83:1.4

190.5

77

44

APRV

22

64:1200.4

84

53

APRV

14

52.5:0.6250.341a

58a

APRV as CPAP

6

52:0.6

250.3

41a

52a

aVenous blood samples.The patient received no further ventilatory support after day 20. MV, mode of mechanical ventilation; APRV, airway pressure release ventilation; CPAP, continuouspositive airway pressure; Phigh, upper APRV pressure (cmH20); Plow, lower APRV pressure (analogous to PEEP; cmH20); I:E, inspiratory/expiratory duration (s);RR, respiratory rate (per min); Fi02, inspiratory 02 fraction; P02, PC02, 02 and C02 partial pressures (arterial on day 0-9, venous on day 16 and 20; mmHg).

(cont.)Pressure controlled mechanical ventilation was

performed using APRV to allow a contributionfrom spontaneous breathing. Re-positioning of thethoracic drains was necessary. As an extendedperiod of mechanical ventilation was anticipated,a percutaneous tracheostomy (FANTONI method)was performed 3 days after the accident. Intermittentprone/supine positioning was carried out during thenext few days. The drains were withdrawn 21 daysafter the accident, despite a small pneumothorax inthe right lung. As the patient was able to breathewith a PS of only 5 cmH2O, the tracheostomawas allowed to close and she was transferred toa rehabilitation ward at the primary hospital.

VENTILATORY STRATEGIES FOR ACUTEBRONCHIAL ASTHMA AND ACUTEEXACERBATION OF CHRONICOBSTRUCTIVE PULMONARY DISEASE

Pathophysiological conditions

Many of the basic pathophysiological conditions rele-vant for mechanical ventilation are similar in acutesevere bronchial asthma and in the acute exacerba-tion of chronic obstructive pulmonary disease(COPD) and therefore these are described together.

The following basic conditions determine thestrategy of mechanical ventilation.

• Increase in lung volume due to incomplete expir-ation. Increased airway resistance, decrease in

elastic recoil and decrease in expiration time leadto dynamic hyperinflation. In COPD, this isaggravated by chronic morphologic changes ofemphysema.

• Increase in work of breathing, combined with ahigh ventilatory demand, results in acute and/orchronic respiratory muscle fatigue.

• The consequences are a severe impairment ofpulmonary gas exchange. In status asthmaticus,acute failure of the ventilatory 'pump' leads tohypercapnia and hypoxaemia. In COPD, oxygen-ation is impaired and progressive pump failureresults in chronic hypercapnia, which maymarkedly worsen in an acute exacerbation such asacute pulmonary infection.

• The increase in intrathoracic pressure comprom-ises haemodynamics. In COPD, this leads tochronic cor pulmonale and episodes of acuteright heart failure.

The impact of these different factors varies inacute asthma and COPD. In status asthmaticus, air-way smooth muscle contraction, wall inflammationand intraluminal mucus cause a marked increase inresistance, which varies within the lungs. In COPD,the loss of elastic recoil and a chronic increase inbronchial secretions prevail, which may criticallyincrease in acute infection.

Increased resistance to flow, high ventilatorydemands and short expiratory time, all present toa variable degree, prevent the respiratory systemreaching static equilibrium volume at the end ofexpiration. Inspiration therefore begins at a highlung volume associated with intrinsic or autoPEEP(PEEP.) due to this positive recoil pressure at

Ventilatory strategies for acute bronchial asthma and acute exacerbation of chronic obstructive pulmonary disease 37

end-expiration. This phenomenon of dynamichyperinflation largely dictates the strategy ofmechanical ventilation. Dynamic hyperinflationraises intrathoracic pressure, increases the elasticwork of breathing and forces the inspiratorymuscles to operate at high lung volume, whichis a disadvantageous position for pressuregeneration.14 Increased intrathoracic pressure andhyperinflation may cause cardiovascularcompromise, barotrauma and patient-ventilatordyssynchrony. High alveolar pressure may alsoincrease alveolar dead space and hence ventilatoryrequirements. The increase in airway resistanceresults in increased resistive work of breathing andinhomogeneous ventilation distribution with lowV/Q ratio regions ('slow compartments'), leading toarterial hypoxaemia (Fig. 3.2). The increased workof breathing may lead to respiratory muscle failureand, as a consequence, to hypercapnia. Furthermore,the interaction of mechanical ventilation and elasticproperties of the respiratory system may causeserious complications.15'16

In the spontaneous breathing patient, dynamichyperinflation and the large pleural pressure swingsmay reduce stroke volume due to changes in pre-load and after-load of both ventricles.17 The situ-ation is aggravated when initiating mechanicalventilation. The positive intrathoracic pressuresthroughout the respiratory cycle further decrease

venous return and cause hypotension immediatelyafter the institution of mechanical ventilation, acommon event in these patients. If the patient isover-ventilated to correct respiratory acidosis, sud-den cardiovascular collapse may occur.

Indications for 'invasive' mechanicalventilation

Respiratory function may deteriorate despite conser-vative treatment in status asthmaticus. In this life-threatening situation, mechanical ventilation isindicated. In patients with COPD, exacerbation ismost often caused by acute pulmonary infectionsand may again necessitate mechanical ventilatorysupport. Intubation and mechanical ventilationin these circumstances are associated with a higherincidence of complications than in patients ventila-ted for other causes of respiratory failure. Therefore,the risk:benefit ratio of this decision has to be care-fully considered. On the one hand, the risks of mechan-ical ventilation such as dynamic hyperinflation,barotrauma and nosocomial infection have to bebalanced with the risk of acute respiratory or car-diac decompensation and severe arterial hypox-aemia resulting in cardiac or respiratory arrest.Careful observation is necessary in the intensive careunit (ICU) or high dependency unti (HDU), with

Figure 3.2 Causes and conse-

quences of dynamic hyperinflation

(see text for details).


readiness for immediate action if there is failure toimprove.

GOALS OF MECHANICAL VENTILATION

The goals of mechanical ventilation in asthma areto support pulmonary gas exchange and to unloadrespiratory muscles whilst allowing time for othertherapeutic interventions (such as steroids) to reduceairway inflammation and bronchial reactivity. Inasthma, ventilatory support is generally only neededfor a short time, e.g. hours to a few days. In COPD,the goals of mechanical ventilation are similar, butventilatory support may be needed for longer.

COMPLICATIONS OF MECHANICAL VENTILATION

Mechanical ventilation is associated with a numberof complications,15'16 which increase morbidity andmortality:

• mucus plugging, e.g. atelectasis, occlusion of theendotracheal tube,

• ventilation-associated pneumonia (VAP) andnosocomial infection,

• barotrauma, e.g. pneumothorax, subcutaneousemphysema,

• hypotension.

INDICATIONS FOR INVASIVE VENTILATION

IN STATUS ASTHMATICUS

The absolute indications for intubation are: coma,apnoea, cardiac arrest and severe arterial hypox-aemia despite high FiO2. Continued deterioration,despite maximal conservative treatment, often forcesthe physician to institute mechanical ventilation.Evidence of fatigue, excessive work of breathing andsomnolence (cerebral hypoxia) should be looked forcarefully: fatal apnoea can occur very suddenly andunexpectedly. Changes in alertness, speech, respira-tory rate and extent of accessory muscle use indicatedeterioration. In patients with significant bronchialsecretions, the decision to proceed to endotrachealintubation should be taken earlier.

Hypercapnia per se is not an indication for intub-ation. Patients who become more comfortable andmore able to speak should continue with medicaltherapy despite a high PaCO2. A progressive increasein PaCO2 with acidaemia (pH < 7.25) may force

matters. Evidence of CVS compromise, arrhythmiaor pneumothorax will necessitate ventilatory sup-port. Pneumothorax is best drained before intub-ation. The indication for intubation and mechanicalventilation is therefore a clinical decision based on anestimate of the risks and benefits.

IntubationEndotracheal intubation may be extremely difficultin status asthmaticus. Reflex bronchospasm and car-diac arrhythmia, or cardiac arrest, may occur inhypoxaemic patients. Intubation should thereforebe performed by the most experienced clinicianavailable. Deep sedation with benzodiazepines isnecessary. Ketamine (dosage 3-6 mg kg-1) may beuseful as it also causes sympathomimetic stimula-tion. Some recommend topical anaesthesia, but thismay irritate the airways and cause further bron-chospasm. In emergency cases, oral intubation iseasier to perform. In addition, it allows for place-ment of a large endotracheal tube (8 mm orgreater), which is important for suctioning viscousmucus. A large tube also reduces airflow resistanceand facilitates bronchoscopy, which may be neces-sary for lavage.

INDICATIONS FOR INVASIVE VENTILATIONIN ACUTE EXACERBATION OF CHRONICOBSTRUCTIVE PULMONARY DISEASE

Before initiating invasive mechanical ventilation inCOPD patients, it is essential to consider that thismaybe the starting point of a long process with com-plications such as nosocomial infection, barotrauma,weaning problems etc.

In patients with COPD, the use of non-invasiveventilation (NIV) should be considered if there isno compelling reason to intubate, such as severelife-threatening hypoxaemia, multiple organ fail-ure or an impending surgical procedure.

The decision to intubate should be based on theclinical situation. Symptoms of life-threateningdecompensation are excessive respiratory rate (>35min"1), severe respiratory inco-ordination, increas-ing agitation or coma; a pH < 7.25 will not be toler-ated for a long period. In contrast to the situation in

Ventilatory strategies for acute bronchial asthma and acute exacerbation of chronic obstructive pulmonary disease 39

status asthmaticus, early percutaneous tracheo-stomy may be useful if extubation to NIV is notplanned.

The practice of mechanical ventilation

CONTROLLED MECHANICAL VENTILATION

Controlled modes are used initially. It is important tocontrol and limit airway pressure. Pressure con-trolled ventilatory modes (PCV) prevent excessiveairway pressures that might otherwise occur if resist-ance suddenly increases. There are no controlledstudies to support their use, however.

Initiating mechanical ventilation can be difficultand requires continuous adaptation of the ventila-tory settings. Patients should initially be deeplysedated to prevent fighting against the ventilator.Muscle paralysis is needed in the acute situationbut, if it is used, it should be only for a short period.COPD patients may not adapt to the ventilator eas-ily and a high dose of opiates may be needed todepress respiratory drive. High airway pressure perse does not necessarily indicate hyperinflation andmay be caused by high airflow resistance in theendotracheal tube or the central airways (e.g. secre-tions, mucous plugs).

In acute asthma, correction of arterial hypox-aemia is the first priority. In COPD patients, a rea-sonable target is to keep arterial O2 saturationabout 90% (PaO2 60-70 mmHg). If this cannot beachieved by FjO2 < 0.4, a pulmonary shunt, e.g.atelectasis, pneumonia etc. is likely and requirestreatment.

In asthma and COPD, minimizing dynamichyperinflation is essential.

In controlled modes, there are three strategies thatcan decrease dynamic hyperinflation:

1 decrease of minute ventilation,2 increase of expiratory time,3 decrease of resistance to expiratory flows.

Decrease of minute ventilationControlled hypoventilation in mechanically venti-lated patients is associated with lower mortality and

a reduced number of complications in asthma aswell as in COPD patients.18,19 Large tidal volumeand minute volume promote dynamic hyperinfla-tion.20 Hypoventilation can be performed bydecreasing tidal volume or breathing frequency, orboth. Tidal volume and breathing frequency as lowas 5 ml kg-1 and 6 breaths min- l , respectively, havebeen successfully applied in patients with severeasthma.18 The degree and duration of hypoventila-tion depend on the severity of obstruction, whichvaries greatly between patients. To minimize pulmon-ary hyperinflation, end-inspiratory airway pressureshould be limited to <50 cmH2O (some intensivistseven recommend keeping end-inspiratory Paw <30 cmH2O).21

The risks of permissive hypercapnia are minimalas long as PaCO2 changes do not occur quickly.22

Provided oxygenation is preserved and conditionsof increased susceptibility to a high PaCO2, such asincreased intracranial pressure, are not present, val-ues of PaCO2 in excess of 80 mmHg are acceptable.Most authors agree that correction of acidaemiashould only be done when it is severe (pH < 7.2). Acarefully titrated buffer therapy (TRIS, sodiumbicarbonate) or a small increase in alveolar ventila-tion is then indicated. Attempts to decrease VCO2

by manipulation using nutritional support arecommon, but their value is uncertain. Fever reduc-tion and treatment of infection will also reduceVCO2.

Increase of expiratory timeDynamic hyperinflation can be reduced by increas-ing expiratory time to allow an almost completeexpiration despite airflow limitation. This must beachieved at the expense of shortened inspiration byincreasing inspiratory flow (>70 L min-1) and byeliminating the end-inspiratory pause (if volume-controlled modes are used). The strategy of increas-ing expiratory time, although less powerful thancontrolled hypoventilation, decreases dynamichyperinflation and improves cardiovascular functionand gas exchange.

Decrease resistance to expiratory flowsDecreasing airway resistance by the use of bron-chodilator drugs and corticosteroids is of greatimportance.16 External resistances should be mini-mized by using large endotracheal tubes and opti-mizing the connecting devices (i.e. PEEP valves,


circuit connections etc.). These patients have exces-sive work of breathing due to multiple factors and itmay be important to rest the respiratory muscles bycontrolled ventilation and sedation. Muscle paralysisshould only be used at the initiation of mechanicalventilation. After 24 hours, ventilatory support canbe switched to assist modes or mechanically sup-ported spontaneous breathing.

PATIENTS VENTILATED ON ASSISTED MODES

When the patient's status improves, the ventilator isswitched to an assist mode (pressure or volumeassist) and the process of weaning begins (seebelow). Assisted modes may be used initially, ifNIV is applied. In either case, patient-ventilatorinteraction needs close attention. By improvingpatient-ventilator synchrony, weaning may be facili-tated. To improve patient-ventilator synchrony thesetting of the ventilator must be adapted to thepatient and frequently checked.

• Maximize the trigger sensitivity. Decrease thethreshold for triggering to a level at whichautocycling does not occur; use a ventilatorwith a short response time; and decrease theresistance of the inspiratory circuit. The set-tings should be repeatedly reviewed and appro-priately adjusted.

• Minimize dynamic hyperinflation. Decreaseairway resistance (bronchodilators, cortico-steroids) and ventilatory demands (correctionof hypoxaemia, sedation and even short-termdepression of central respiratory drive byopioids).

• Low levels of external PEEP may increase triggersensitivity substantially by reducing the elasticthreshold load and the work of breathing. Thisbeneficial effect of PEEPe is most evident inpatients exhibiting flow limitation during tidalexpiration. On the other hand, if flow limitationdoes not exist, external PEEP will present a back-pressure to expiratory flow and will cause furtherhyperinflation. The increase in end-expiratoryalveolar pressure counterbalances the beneficialeffect on trigger sensitivity. As a rough rule,external PEEP greater than 8 cmH2O should beavoided. Careful re-evaluation of patient statusshould be performed after applying external PEEP.

• Initial inspiratory flow must meet the patient'sflow demand. Inspiratory flow and machine

inspiratory time must be set high enough to sat-isfy the needs of the patient, which may alter,necessitating new ventilator adjustments. Withpressure support mode (PSV), the patient has theability to influence the machine breathing patternas well as tidal volume and this mode might there-fore be preferred. In the conventional assistmodes, e.g. SIMV, this ability is compromised bythe mechanical properties of the respiratory sys-tem and the function of the ventilator.

Changes in ventilator settings may modify thepatient's respiratory response via modifications inmechanics, chemical reflexes and behavioural feed-back systems. For example, by increasing inspiratoryflows, an increase in spontaneous breathing frequencymay occur due to a reflex excitatory effect of highflows on respiratory output counterbalancing theirbeneficial effect on dynamic hyperinflation. Thus, incontrast to controlled ventilation, changes in ventila-tory settings in patients on assist modes may notalways be successful.

Restoring respiratory muscle functionby intermittent controlled ventilationA common cause of delay in weaning COPDpatients is an inability of the respiratory pump tomeet the ventilatory demand. During assisted venti-lation, the respiratory muscle load is difficult toassess and could remain excessive in assistedspontaneous modes. However, recovery can be pro-vided by intermittent controlled ventilation, particu-larly at night, and other periods of spontaneousT-piece breathing,23 or after extubation, employingNIV (see Chapter 5).

MECHANICAL VENTILATION IN BRAININJURY

Early management

Hypoxia is the second most important cause of mor-tality and morbidity following traumatic braininjury (see references 24 and 25). In the emergencysituation, clearing the airway and providing oxy-genation and adequate ventilation must be achievedwithout delay.

In patients with additional injuries that increasethe risk of hypoxia, e.g. chest trauma or massive

Mechanical ventilation in brain injury 41

EMERGENCY MANAGEMENT FOROXYGENATION AND VENTILATION

• Secure (or maintain) airway.+• Provide high-flow O2 to all patients with trau-

matic brain injury (regardless of severity).• Provide intubation and ventilation for patients

with a Glasgow Coma Score (GCS) of <8 (noteye opening, speaking or obeying commands)or a motor score <5 (withdrawing from painor worse).

• Avoid and/or treat aspiration.• Keep arterial O2 saturation >95%.• Avoid excessive hyperventilation, e.g. end-tidal

CO2 30-35 mmHg.

bleeding, intubation and mechanical ventilationshould be considered at a higher GCS.

INTUBATION

The procedure can considerably increase intracranialpressure and should only be carried out with deep seda-tion and sufficient pre-oxygenation. Oral intubation ispreferred as it is easier and safer if cervical spine injuryhas not been excluded. On the other hand, experiencedphysicians sometimes prefer 'blind' nasal intubation,especially for tube replacement in already intubatedpatients. In suspected or proven cervical spine injury,intubation may provoke further neurological damageand in-line axial stabilization must be ensured manually.A cervical collar should be removed as meticulous man-ual stabilization is regarded as being safer.

Tracheostomy should be performed at anearly stage if a prolonged period of ventilatorysupport is anticipated. Optimal timing is still asubject of controversy. In a multicentre, randomized,prospective trial, no differences in length of stay,infections, mortality and tracheopharyngeal injurywere found between early (3-5 days') and late (10-14days') tracheostomy.

Further management

Arterial hypoxaemia, hypercapnia and systemichypotension must be avoided.

These may occur during suctioning, positioning,physiotherapy, especially through lack of analgesia or

In patients with increased intracerebral pres-sure

• hypoxaemia• hypercapnia• hypotension

may cause secondary brain damage.

sedation, or during the replacement of endotrachealtubes, diagnostic or therapeutic manipulations ortransportation for external diagnostic procedures oroperations. Close monitoring with pulse oximetry,capnometry, arterial pressure and intracranial pres-sure monitoring and awareness of the problems aremandatory.

PRINCIPLES FOR MECHANICAL VENTILATION

To avoid hypoxaemia or hypercapnia, mechanicalventilation should be considered early.

Target criteria are arterial O2 saturation >95%,arterial PCO2 35-40 mmHg (if hyperventilationis not indicated, see below). A high FiO2 (0.5-0.6)is often required to give a good margin of safety.

Positive end-expiratory pressureThe easiest way to improve oxygenation is to applyexternal PEEP. Increasing intrathoracic pressurePEEP may impair venous return and cardiac outputand reduce cerebral perfusion. An optimal balancebetween the respiratory and the cardiovasculareffects must be found. The effect of PEEP on cerebralcirculation depends on the intracranial compliance,on the absolute level of the intracerebral pressureand on the haemodynamic effects related to the stiff-ness of the lungs. As long as the intracranial pressureis greater than the venous pressure, PEEP will notincrease it. Whenever patients with acute braininjury need PEEP to optimize arterial oxygenation,careful monitoring of changes in intracranialpressure and cerebral perfusion pressure (CPP) ismandatory.


Controlled ventilatory modesWhen controlled ventilatory modes are required,pressure or volume control can be used. PCV may bepreferred to avoid unexpected airway pressureincreases. Under stable conditions, and when peakairway pressure can be kept <30 cmH2O, the volumecontrolled mode (VCV) has the advantage of ensur-ing minute volume and thereby the arterial PCO2.However, sudden changes, e.g. a pneumothorax, oreven coughing, may result in an increase in airwaypressure, and a pressure limit control should be set atabout 35 cmH2O. PEEP should be set at a moderatelevel (<8 cmH2O) that has no influence on cerebralperfusion if arterial systemic pressure is maintained(systolic >120 mmHg or mean >90 mmHg). Thesemi-recumbent position (15° to maximum 30°) ofthe upper part of the body will compensate forpotential effects of PEEP. Fighting against the venti-lator must be prevented. In severe brain injury, anal-go-sedation will be kept at relatively high levels asa general treatment of the brain damage. The conti-nuous application of muscle relaxants is obsolete,although for acute situations, such as severe cough-ing attacks, during adaptation to new ventilatorysettings or in severe shivering, their short-termapplication may be useful.

Assisted ventilatory modesIn the later course, and assuming that respiratorydrive is intact, assisted modes and supportedspontaneous breathing (IMV, PSV, BIPAP) can bechosen. This author prefers the BIPAP modesbecause the ventilatory system is open at any time.The patient can breathe spontaneously whenever heor she wants to (even during the inspiration phase ofthe ventilator's cycle) and the pre-set peak airwaypressure can never be exceeded (even whenthe patient is coughing or fighting against theventilator).

WeaningThe decision as to when the patient can be weanedfrom mechanical ventilation depends on manyfactors:

• the severity of the brain damage,• the clinical situation,• age,• complications and concomitant diseases,• organizational aspects, e.g. the availability of

competent personnel.

Weaning is a step-wise process, which must be care-fully monitored. It should be started as early as pos-sible because unnecessarily prolonged mechanicalventilation will increase the risk of complications. Anunsuccessful weaning trial will not endanger thepatient as long as a critical increase in intracranialpressure and a reduction of cerebral perfusion are pre-vented. If the intracranial pressure is continuously> 30 mmHg, a weaning trial should not be undertaken.

SPECIFIC PROCEDURES

Therapeutic hyperventilationThe prophylactic use of hyperventilation (PCO2 <35 mmHg) during the first 24 hours is not recom-mended because it may compromise cerebral perfu-sion. Hyperventilation therapy (PCO2 = 30-35mmHg) may be necessary for brief periods whenthere is an acute neurologic deterioration or for alonger period of intracranial hypertension refract-ory to conventional therapy. In this situation, hyper-ventilation must be carefully titrated; an arterialPCO2 < 30 mmHg increases the risk of cerebralischaemia and must be strictly avoided.24 In theabsence of increased intracranial pressure, chronicprolonged hyperventilation (PCO2 < 25 mmHg) isnot indicated.

COMPLICATIONS

Acute lung injury is a known complication of acutebrain injury. The real frequency and cause(s) arecontroversial. If acute lung injury is not the result oftrauma to the chest, the most common causes are:

• aspiration (often during resuscitation phase),• pneumonia (the second most frequent complica-

tion, occurs in 40-50%),• atelectasis (due to impairment of the normal

cough reflex and/or to frequent disconnectionsfrom the ventilator for endotracheal suctioning),

• neurogenic pulmonary oedema (due to pul-monary vasoconstriction from increased sympa-thetic tone, arterial hypertension, or an increasein pulmonary capillary permeability).

These complications aggravate the process andmay worsen the secondary brain damage. The thera-peutic principles to treat these complications aredescribed elsewhere. However, treatment of the acutelung injury must comply with the special limitationsregarding acute brain injury (see above).

Mechanical ventilation in brain injury 43

Case 2: ARDS plus Traumatic BrainInjury

KO, a 28-year-old male, had a motorcycle accidentcausing severe traumatic brain injury. He wasdeeply comatose and was immediately intubatedby the emergency team before transportation tothe local hospital. Aspiration was considered likelybefore intubation. A few hours later, he was trans-ferred to another hospital with a neurosurgicaldepartment. Scanning revealed right parietal con-tusion with subarachnoidal bleeding into the ven-tricle, a subdural haematoma, fracture of the leftcondylus occipitalis and a pneumo-encephalon.There was also blunt chest trauma with bilaterallung contusion and probable aspiration.

A severe sepsis syndrome with multiple organfailure (severe ARDS, barotrauma, acute renal fail-ure) developed during the next days and thepatient was transferred to our ICU 9 days after theaccident. On arrival, he was deeply analgo-sedatedand under controlled ventilation (Table 3.2).Bilateral pulmonary leakage was treated by thora-cic drains. The systemic circulation was supportedby 40 u,g min-1 of nor-epinephrine. X-rays andcomputerized tomography scans revealed a severeARDS with some degrees of fibrotic consolidation,barotrauma including a pneumothorax, andemphysema in the mediastinum (Fig. 3.3).

We began mechanical ventilation with theAPRV mode. Under this ventilatory setting,intracranial pressure could be kept within a range

(cont.)

Figure 3.3 KO, a 28-year-old male with polytrauma, sepsis and

ARDS: CT scan (a) sagittal and (b) transversal at arrival in our

hospital (9 days after trauma) with intrathorack drains. The patient

had severe ARDS and pneumothorax.

Table 3.2 KO, a 28-year-old male with polytrauma, sepsis and ARDS

MV

PhighPlowI:ERRFi02

P02

PC02

ICP

IRV7

7

2:1.07

1.0

121

45

-

APRV

29

5

2:0.5

29

0.9

85

53

11

APRV

29

8

2:1.0

29

0.8

87

45

11

APRV

30

10

2:0.8

31

0.5

99

56

20

APRV

26

12

2:0.8

21

0.8

70

56

18

APRV

26

11

2:1.5

26

0.45

84

51

19

APRV

20

8

2:2.0

19

0.35

102

53

-

ASB-SP

AP10

8-

48

0.3

52a

47a

-

ASB-SP

AP10

5

-

27

0.3

50a

49a

-

aVenous blood samples.IRV, inversed ratio ventilation; ASB-SP, assisted spontaneous breathing (AP, assist pressure (cmH20); ICP, intracranial pressure (mmHg). For otherabbreviations, see Table 3.1.


Figure 3.4 KO, a 28-year-old male with polytmuma, sepsis and

ARDS: CT scan at day 26 (35 days after trauma). The ARDS is

considerably improved; there is somefibrosis and some bullae.

(cont)of 11-20 mmHg (see Table 3.2). A percutaneoustracheostomy was carried out and intermittentprone/supine positioning performed.

Inotropic support, blood transfusions and con-tinuous veno-venous haemofiltration were neces-sary. The patient's recovery was slow due to theseverely damaged cerebral function (intracerebralhygroma, ventricular drainage, infection). Later,weaning was further delayed by a critical illnesspolyneuropathy, so that it was 3 months after theaccident before the patient was discharged forfurther rehabilitation (Fig. 3.4).

MECHANICAL VENTILATION IN CARDIACFAILURE

Ventilatory support in ischaemic heartdisease

An increasing in intrathoracic pressure inhibitsvenous return but decreases left ventricular (LV)transmural pressure, which reduces LV after-load.In the failing heart, which is very sensitive to after-load, the net effect is to improve cardiac output. Incontrast, decreasing intrathoracic pressure enhancesvenous return and may increase LV after-load.17

Consequently, CPAP and mechanical ventilationwith PEEP have been used to improve cardiac out-put in LV failure and acute cardiac pulmonaryoedema.

When pulmonary oedema is severe in acuteheart failure, Type 2 respiratory failure occurs26

because the work of breathing is markedlyincreased. The excessive intrathoracic pressureswings may also impair diastolic function.Respiratory support is urgently required in thissituation and can be achieved by mask CPAP, NIV(bi-level pressure support) or by intubation andmechanical ventilation. CPAP decreases venousreturn and LV after-load. If LV contractility ismarkedly impaired, this will result in an increase incardiac output. If, however, LV contractility isnormal, the gain in cardiac output will only besmall due to the reduction in venous return.Although CPAP may be sufficient in itself tocorrect a respiratory acidosis, bi-level NIV is moreeffective and better tolerated when the PCO2 ismarkedly elevated, presumably because NIVprovides better unloading of the respiratorymuscles. In some patients, perhaps those withpredominant systolic dysfunction, intubation andmechanical ventilation will be required, especiallyif there are centra-indications to NIV (see Chapter5). Inotropes and LV assist procedures may beneeded.

In chronic congestive heart failure, the use ofCPAP is controversial. In some studies, long-termnocturnal CPAP has been reported to improve day-time cardiac function, but this has not been con-firmed by other studies. The mechanisms that couldexplain a benefit are not fully understood, e.g.decrease in venous return, decrease in LV after-load, vasodilatation by activation of parasympa-thetic vasodilator mechanisms or reduction insympathetic tone.27

SPECIAL CONSIDERATIONS

Massive obesity increases intra-abdominal pressure,impedes diaphragm function and increases the like-lihood of OSA. It may also increase the risk of respira-tory failure by impairing sputum clearance andpromoting atelectasis and veno-embolism. In acutelung injury, obesity will also increase the risk of com-plications.

References 45

REFERENCES

1. Bernard, G, Artigas, A, Brigham, K, et al. Report of the

American-European Consensus Conference on ARDS:

definitions, mechanisms, relevant outcomes and

clinical trial coordination. Intensive Care Med 1994;

20: 225-32.

2. Gattinoni, L, Pelosi, P. Pathophysiologic insights into

acute respiratory failure. Curr Opin Crit Care 1996;

2: 8-12.

3. Gattinoni, L, Pelosi, P, Suter, PM, Pedoto, A, Vercesi, P,

Lissoni, A. Acute respiratory distress syndrome caused

by pulmonary and extrapulmonary disease. Different

syndromes? Am J Respir Crit Care Med 1998; 158: 3-11.

4. American Thoracic Society (ATS), European Society

of Intensive Care Medicine (ESICM), Societe de

Reanimation de Langue Francaise (SRLF). International

Consensus Conferences in Intensive Care Medicine:

Ventilator-associated lung injury in ARDS. Am J Respir

Crit Care Med 1999; 160: 2118-24.

5. Amato, MB, Barbas, CS, Medeiros, DM, etal. Effect of a

protective-ventilation strategy on mortality in the

acute respiratory distress syndrome. N Engl J Med

1998; 338: 347-54.

6. Hickling, KG, Walsh, JSH, Jackson, R. Low mortality

rate in adult respiratory distress syndrome using

low-volume, pressure limited ventilation with

permissive hypercapnia: a prospective study. Crit

Care Med 1994; 22: 1568-78.

7. Slutsky, AS. Consensus Conference on Mechanical

Ventilation -January 28-30,1993, at Northbrook,

Illinois, USA. Part II. Intensive Care Med 1994;

20: 150-62.

8. Sydow, M., Burchardi, H. Inverse ratio ventilation and

airway pressure release ventilation. Curr Opin

Anesthesiol 1996; 9: 523-8.

9. Rathgeber, J, Schorn, B, Falk, V, Kazmaier, S, Spiegel,

T, Burchardi, H. The influence of controlled

mandatory ventilation (CMV), intermittent

mandatory ventilation (IMV) and biphasic

intermittent positive airway pressure (BIPAP) on

duration of intubation and consumption of

analgesics and sedatives. A prospective analysis in

596 patients following adult cardiac surgery. Eur J

Anaesthesiol 1997; 14: 576-82.

10. Acute Respiratory Distress Syndrome Network.

Ventilation with lower tidal volumes as compared

with traditional tidal volumes for acute lung injury

and the acute respiratory distress syndrome. N Engl J

Med 2000; 342: 1301-8.

11. Hormann, C, Baum, M, Putensen, C, Mutz, N, Benzer, H.

Biphasic positive airway pressure (BIPAP) - a new mode

of ventilatory support. Eur J Anaesthesiol 1994;

11:37-42.

12. Sydow, M, Burchardi, H, Ephraim, E, Zielmann, S,

Crozier, TA. Airway pressure release ventilation versus

volume controlled inverse ratio ventilation in patients

with acute lung injury. Am J Respir Crit Care Med 1993;

149:1550-6.

13. Slutsky, AS. Consensus Conference on Mechanical

Ventilation-January 28-30,1993, at Northbrook,

Illinois, USA. Part I. Intensive Care Med 1994;

20: 64-79.14. Rossi, A, Polese, G, Brandi, G, Conti, G. Intrinsic

positive end-expiratory pressure (PEEPi). Intensive

Care Med 1995; 21: 522-36.

15. Georgopoulos, D, Brochard, L. Ventilatory strategies in

acute exacerbations of chronic obstructive pulmonary

disease. In Mechanical ventilation from intensive care

to home care. Roussos, C, ed. Sheffield: European

Respiratory Society, 1998; 12-44.

16. Georgopoulos, D, Burchardi, H. Ventilatory strategies

in adult patients with status asthmaticus. In

Mechanical ventilation from intensive care to home

care. Roussos, C, ed. Sheffield: European Respiratory

Society, 1998; 45-83.

17. Pinsky, MR. Mechanical ventilation and the

cardio-vascular system. Curr Opin Crit Care 1996;

2: 391-5.

18. Darioli, R, Perret, C. Mechanical controlled

hypoventilation in status asthmaticus. Am Rev Respir

D/S1984; 129:385-7.

19. Tuxen, D. Permissive hypercapnic ventilation. Am J


20. Tuxen, D. Detrimental effects of positive

end-expiratory pressure during controlled mechanical

ventilation of patients with severe airflow obstruction.


21. Corbridge, T, Hall, J. The assessment and the

management of adults with status asthmaticus

(state of the art). Am J Respir Crit Care Med 1995;

151: 1296-316.

22. Feihl, F, Perret, C. Permissive hypercapnia. How

permissive should we be? Am J Respir Crit Care Med

1994; 150:1722-37.

23. Esteban, A, Frutos, F, Tobin, M, etal. A comparison of

four methods of weaning patients from mechanical

ventilation. N Engl J Med 1995; 332: 345-50.

24. Maas, A. Pathophysiology, monitoring and treatment of

severe head injury. New York: Churchill Livingstone,

1993; 565-78.


25. Bullock, R, Chesnut, RM, Clifton, G, et al.

Guidelines for the management of severe head

injury. New York: Brain Trauma Foundation,

1995; 1-166.

26. Bersten, AD, Holt, AW. Acute cardiogenic pulmonary

edema. Curr Opin Crit Care 1995; 1: 410-19.

27. Scharf, SM. Ventilatory support in cardiac failure. Curr

Opin Crit Care 1997; 3:71-7.

4

Ventilator-patient interactionJOHN CGOLDSTONE

Introduction 47 Heart-lung interactions

Triggering mechanical ventilation 47 Respiratory drive

Respiratory muscle work during mechanical ventilation 49 References

52

54

57

INTRODUCTION

During mechanical ventilation, the relationshipbetween the ventilator and the patient is not simplya passive one in which the patient's lungs are inflatedand the ventilator settings are defined by the user.Rather, many interactions occur and there exists aninterface between the patient and the ventilator. Forexample, ventilation is frequently switched on by thepatient inspiration and the ventilator must thereforebe able to sense the onset of inspiration. Many othersuch interactions occur and there may be consider-able overlap between them (Fig. 4.1).

TRIGGERING MECHANICAL VENTILATION

Mechanical ventilation is usually delivered so that thepatient is able to initiate the breath from the ventil-ator. Before inspiration occurs, there are several stepsthat need to be performed and this process is termedtriggering (Fig. 4.2).1>2 In addition, a further set ofconditions needs to occur to continue the breathdelivered to the patient. The ventilator is constantlychecking that inspiration is continuing in order toavoid high airway pressures, which would occur if thepatient tried to breathe out during inspiration. Theeffect of the poorly set ventilator may be harmful.

The trigger can be adjusted to make the ventilatormore or less sensitive. In older ventilators, a demand(mechanical) valve was employed as the device thatcould be opened by the patient at the start of inspir-ation. Demand valves are difficult to breathe throughand, although they are not used in modern mechan-ical ventilators, they serve to illustrate the problemsthat a poorly adjusted trigger system imposes on thepatient.

The inspiratory trigger

At the start of inspiration, the pressure differencegenerated in the chest is transmitted to the upperairway, and this is the physiological signal that issensed by the ventilator to begin inspiration.However, unlike spontaneous breathing, flow doesnot begin as soon as the pressure at the airway is lessthan atmospheric. In fact, there is a delay imposedby the breathing system until the pressure in theupper airway equals the opening pressure of theinspiratory valve. This is termed the 'trigger phase'.

As soon as the opening pressure of the demandvalve is reached, inspiratory flow begins. Ifinspiratory flow from the ventilator were unlimited,little further load would be placed on the patient.However, a demand valve imposes a further loadbecause the flow produced by the valve is low, often

48 Ventilator-patient interaction

Clinical effects

Figure 4.1 The patient is rarely passively ventilated. There are

many interactions that may occur between the patient and the

ventilator.

The demand valve system had a number of clinicaleffects on the patient. The ventilator trigger systemimposes a considerable absolute delay when no gascan flow to the patient. During this time, allinspiratory effort that is expended is wasted.Furthermore, inspiratory muscle activity is continuedand increased, even when the demand valve opens, asflow is insufficient. Such native inspiratory muscleactivity, once initiated, does not subside, even whenthe ventilator begins inspiration.3 The amount ofinspiratory muscle activity can equal or exceed thatduring spontaneous breathing. This implies that,if the ventilator is adjusted in this manner, nooff-loading of muscle activity occurs. The graphicterm 'fighting the ventilator' describes, in part, apoorly set inspiratory trigger.

much lower than required. This is particularly thecase for patients with high airways resistance, e.g.chronic obstructive pulmonary disease (COPD), whooften require high inspiratory flows. When the flow isinadequate for the patient demand, inspiratory effortcontinues and further negative pressure is generated.The period between the opening of the demand valveand full flow occurring has been termed the'post-trigger' phase. When insufficient inspiratoryflow occurs, the term 'flow starvation' has been used.

Modern triggering systems

A demand valve is nowadays seldom used. In its place,proportional valves are controlled by a microprocessor.A solenoid opens the valve quickly. Additionally, thevalve settings can change over small ranges quickly andwith minimal time delay, enabling flow to be adjustedprecisely and complex pressure waveforms to becreated within the airway. The response time tovalve opening may be of the order of milliseconds.

Figure 4.2 There are two phases that occur before the ventilator allows the patient to breathe freely. The pressure in the airway must reach

the set trigger pressure; and, as negative pressure continues, inspiratory flow then occurs.

Respiratory muscle work during mechanical ventilation 49

Moreover, high inspiratory flow rates can be achieved,such that the time to full flow may be less than 100 ms.The clinical problems of such systems tend now to berelated to the way in which the software functions.

Flow triggering

The signal that is used to initiate inspiration may betaken from the inspiratory flow waveform and notthe pressure tracing at the mouth. Flow at the mouthwill rise when the pressure gradient between mouthand chest is negative. Whereas the pressure .changesdetected by the ventilator are small, flow signalsincrease substantially, and may improve the sensitiv-ity of the triggering system.

Although it is possible simply to measure flowwithin the breathing circuit, this arrangement itselfimposes load on the subject. In order to remove thisload, flow in the breathing circuit is continuous andis provided at a base level. The flow through theinspiratory limb of the breathing circuit is continu-ously measured and compared with the flow throughthe expiratory limb. If no effort is made, the twoflows should be equal. During the beginning ofinspiration, a small amount of flow enters the patientand the inspiratory flow is greater than expiratoryflow. This, then, is the signal to the ventilator to assistthe breath.

Clinical problems with triggering

The sequence of events that occur to initiate inspir-ation is complex and may be described as the inspir-atory chain of command (Fig. 4.3). The processinvolves activation of the inspiratory motoneurons,mainly the phrenic nerve, neuromuscular transmis-sion and then contraction of inspiratory muscles.Movement of the thoracic ribcage then occurs andthe volume of the system changes. The pressurewithin the thorax decreases. At this stage, a pressuredifference between the chest and the mouth is estab-lished and inspiratory gas flow occurs.

A time delay between the chest and the mouth is acommon feature of obstructive lung diseases such asCOPD.4 Initially, inspiratory muscle contractionoccurs and pressure in the chest changes prior to anypressure change at the mouth. The time delaybetween the onset of pressure change in the chest

and that detected at the mouth can be substantial.During this time, the patient will receive no assist-ance and the term asynchrony has been used.Asynchronous ventilation occurs frequently inmechanical ventilated patients and may go unnoticedunless it is actively looked for.5

Other methods of triggering theventilator

Signals from the brainstem or from the phrenicnerve are difficult to detect and cannot be used totrigger the ventilator. As the muscle is activated, elec-tromyograph (EMG) activity occurs and this can bedetected by percutaneous electrodes on the chestwall or over other inspiratory accessory muscles suchas the sternocleidomastoid. Electrodes in these pos-itions tend to be helpful when the onset of inspirationneeds to be sensed and are useful when substantialcontractions of the inspiratory muscles occur.However, they are susceptible to contamination fromother adjacent, non-respiratory muscle groups, andthe baseline noise sometimes obscures weak inspir-atory contractions. Although it is technically possibleto trigger ventilation, the system is not robustenough for clinical practice.

Recently, oesophageal EMG has been used to senseinspiration. This technique has been refined toincorporate multi-sensors and is resistant to elec-trode movement as a source of artefact. Such a sys-tem could theoretically move the trigger much closerto the patient and avoid many of the problems asso-ciated with triggering.

Improvements in the speed of response can bemade if the pressure signal is moved closer to thepatient. The pressure waveform may be obtainedfrom the ventilator tubing next to the patient or atthe distal part of the endotracheal tube.

RESPIRATORY MUSCLE WORK DURINGMECHANICAL VENTILATION

Respiratory muscle activity during controlled venti-lation occurs commonly and may be detected clinic-ally or by inspecting airway pressure waveforms.During controlled ventilation, the inspiratory pres-sure waveform should be the same for each breath.


Figure 4.3 A breath is initiatedin the central nervous systemand transmitted via the motornerves to the respiratorymuscles. Lung mechanics willthen influence thetransmission of intrapleuralpressure to the airway. Theinspiratory signal may besensed at several points alongthis chain.

However, this is often not the case, as contraction ofinspiratory muscles causes distortion and variationin the pressure waveform from breath to breath.6'7

The work of breathing

Mechanical work is defined as a force moving anobject over a given distance. This form of work iseasy to observe and may be quantified if the forceand distance are known. Whereas work commonlyoccurs during skeletal muscle contraction, not allwork may be measured. Internal work occurs whenthe muscle contracts, for example when holding anobject stationary against gravity (Fig. 4.4). As nomovement has occurred, no external work is per-

formed, yet the muscle is using energy internally,dissipated as heat and light. Such internal work maybe very great indeed. Furthermore, as the muscletension may be large and is maintained constantly,blood flow is reduced. Failure of force generationfrom the muscle is rapid and, despite maximal cen-tral nervous system drive, fatigue occurs.

EXTERNAL WORK

It is easy to imagine quantifying work when a simplelever and pulley system such as a limb muscle is stud-ied. The respiratory muscles have a complex geomet-rical arrangement surrounding the ribcage. Moreover,the main inspiratory muscle has a dome-like shapeand the force generated by the diaphragm is therefore

Respiratory muscle work during mechanical ventilation 51

Figure 4.4 Work is performed during muscle contraction. Internal

work may not be easily measured, whereas external work involves

movement and can be measured more easily.

affected by the radii of curvature of the muscle itself.This makes the measurement of work difficult.

External work is performed when gas moves at themouth. When this happens, pressure within theribcage decreases and a gradient occurs between theribcage and the atmosphere. If at this stage no gasflows between the mouth and the lungs, no move-ment of gas has occurred and no external work isperformed. By contrast, internal work is high. Whengas flows at the mouth, external work is performedand the work of breathing can be measured. Thework of breathing is the integral of pressure andinspiratory volume and is expressed in terms ofJoules per minute or Joules per litre of ventilation.

MEASUREMENT

In controlled circumstances, the work of breathing isstraightforward to measure.8 The pressure differenceduring each breath is recorded. This is the pressuredifference that exists between the chest and mouthand is termed transpleural pressure. It is not possibleto measure pleural pressure directly. A surrogate isfound by measuring pressure elsewhere within theribcage. Pressure changes in the oesophagus reflectpressure changes within the pleural space.Oesophageal pressure can be measured by passing aballoon catheter transnasally into the stomach andthen withdrawing the tube until it is within the mid-point of the oesophagus. The pressure differencebetween the oesophagus and the mouth is then usedfor the calculation.

If the subject breathes through a pneumotacho-graph, inspiratory flow can be calculated and volumemeasured per breath. On an X-Y plot of volume andpressure, the area of each breath can be clearly seenand overlaid, one breath on top of another, enablinga visual display of work per breath over many itera-tions. This is typically measured digitally and soft-ware-based integration enables the work per breathto be calculated.

OTHER MEASUREMENTS OF WORK DURING

BREATHING

Work associated with breathing can be estimatedfrom other techniques. During inspiration, informa-tion can be obtained from the oesophageal pressurechanges alone and measured as the area under theoesophageal pressure curve during inspiration. Themeasurement is made from an oesophageal ballooncatheter. The pressure changes during inspiration arerecorded and the timing of inspiration is made frominspiratory flow measured at the mouth.Oesophageal pressure is integrated from the begin-ning to the end of inspiratory flow, when flowchanges direction. The area under the oesophagealpressure curve is then termed the pressure timeproduct (PTP).9

Clearly, PTP is different from the traditional meas-urement of inspiratory work calculated from volumeand pressure measurements. This need not be aproblem, providing the two methods are not com-pared directly. The greatest strength of the PTP meas-urement may be in establishing trend data. As it iseasy to measure and can be performed at the bedside,PTP has been used widely to report how inspiratorywork changes during respiration.

The electrical activity of the diaphragm can beused to assess the amount of work performed by themuscle. Diaphragmatic EMG may be detected fromelectrodes placed adjacent to the diaphragm on thechest wall, or via needle electrodes placed under theskin on the chest wall, or from electrodes mountedon a balloon catheter and swallowed into theoesophagus. The different placements of the detect-ing electrodes yield different signals and the mostrobust technique is from oesophageal probes.

During quiet respiration, EMG is phasic and it maynot be easy to distinguish it from background noise.As respiratory effort increases, the raw EMG increasesin power and is clearly visible. Oesophageal electrodes


detect the signal more easily as they are closer to themuscle. Originally, movement of the oesophageal sen-sors had a great effect on the quality of the EMG sig-nal acquired, limiting the technique to the laboratory.Multiple electrodes have now been developed, strad-dling the crura of the diaphragm. As the cathetermoves, for example during swallowing, one electrodefrom the array has the maximal signal and this is iden-tified by a computer controlling the array.

A different approach to the problem of measur-ing inspiratory work is to measure the amount ofO2 used due to respiration.9'10 In the critical caresetting, this is possible because the O2 consumedduring mechanical ventilation can be compared tothe amount of O2 used when respiration is unas-sisted. The difference between these two states iscalculated as the amount of O2 consumed by therespiratory muscles themselves. Although technic-ally possible, several problems exist at the bedsidefor critically ill patients. Obtaining steady-statemeasurements both at rest on the ventilator andthen during spontaneous breathing demands littleor no other muscle activity, and this is seldom thecase in the critically ill. Furthermore, the amount ofO2 consumed may be less than 100 ml min"1. Ifthe total amount of O2 during breathing is high(the product of minute ventilation and concentra-tion of O2), the measurement system may berequired to detect less than 1% change in the totalO2 presented to the system, a level of fidelity thatmay not be possible clinically.

CLINICAL APPLICATION

The normal work of breathing is less than 0.6 J min- l

at rest. When measured in patients with lung diseasesuch as obstructive lung defects, this may rise by up tofivefold at rest.11

In the critically ill, the work of breathing can beused to identify those patients who are breathingagainst a heavy load. The upper limit of inspiratorywork has been identified, above which level breath-ing cannot be sustained. In small groups of patientsstudied, this may be at the level of 5 J min-1.

Few large studies have been performed involvingcritically ill patients and the level of inspiratory workmust be judged against the capacity of the inspiratorymuscles to perform the work. With normal musclestrength, higher levels of work could be performedand the measurement of work should not be used inisolation to the performance of the whole system.

HEART-LUNG INTERACTIONS

Cardiovascular fluctuations during spontaneousbreathing are well recognized and it is not surprisingthat mechanical ventilation can have profoundeffects on the circulation.12 Mechanical ventilation,through its changes in both lung volume andintrathroacic pressure, has an influence on the deter-minates of stroke volume for both right and left ven-tricles. It is now appreciated that intermittentpositive pressure ventilation (IPPV) has a complexeffect on cardiac output rather than merely thedecrease in venous blood flow suggested originally.

When lung volume increases, pulmonary vascularresistance (PVR) changes.13 Pulmonary blood vesselshave two major anatomical types. Alveolar blood ves-sels are closely related to the alveoli and are affected bychanges in alveolar pressure. During inflation, alveolarblood vessel resistance increases and the capacity ofthe vessels decreases. By contrast, extra-alveolar vesselsare exposed to intrathoracic pressure changes and vol-umes. During inflation, the calibre of these vesselsincreases, resistance falls and their capacitance rises.Clearly, the net effect on PVR during changes in lungvolume is a balance, and in health PVR increases withlung inflation. In lung diseases characterized by hyper-inflation, PVR is often elevated.

With lung deflation, little change in PVR occurs inthe alveolar vessels, whereas the reduced lung volumecompresses and reduces the calibre of the extrathor-acic vessels and the net effect is to increase PVR atlow lung volumes. In the critically ill, diseases thatlead to a reduced lung volume, e.g. acute respiratorydistress syndrome (ARDS), will tend to increase PVRand this effect will be exacerbated with other changesin the pulmonary vasculature. A goal of ventilatorytherapy is to restore lung volume to functional residualcapacity (FRC) to normalize PVR in hyperinflatedand hypo-inflated lungs.

Lung volume will also impact on the cardiac sys-tem through direct mechanical compression of theheart. This may cause a restrictive effect similar totamponade when pre-existing lung expansionoccurs. As inflation pressures are transmitted to thecardiac chambers, so the measurement of cardiac fill-ing pressures becomes inaccurate. Additionally, thecardiac septum may become deviated. This shift inthe position of the septum may occur towards oraway from the left ventricle, impairing function ofeither the right or left ventricle.

Heart-lung interactions 53

As the volume of the lungs increases, so intrathor-acic pressure changes. During mechanical ventilation,intrathoracic pressure impedes blood flow to the rightatrium and decreases right ventricular diastolic filling.

Intrinsic positive end-expiratorypressure

In normal lungs, the pressure in the alveoli reachesatmospheric as the lung empties at the end of eachbreath and, in the brief moment before the nextbreath, the pressure in the ribcage is equal to the bal-ancing forces and the system stands still and no gasflows. Thus, the natural tendency of the lungs todeflate further is balanced against the tendency of theribcage to spring open. This equilibrium point is theFRC.

If too brief a time is allowed for lung deflation,some of the exhaled gas will be trapped inside thelung and this will add volume to the system whenthe next breath is taken. A new equilibrium pointwill be established, where FRC is at a higher volume,and the pressure in the alveolus is increased. Thiscan be achieved if the controls of the ventilatorare set inappropriately for normal patients who areundergoing mechanical ventilation.

In obstructive lung disease, FRC increases. The rest-ing pressure in the alveolus may be substantially raised.

The effect of intrinsic positive end-expiratorypressure (autoPEEP) was noted in the classicdescription by Pepe and Marini.14 They observedthe autoPEEP effect in a patient who was admittedto the intensive therapy unit (ITU) with a diagno-sis of heart failure and who had a high right-sidedfilling pressure, a low blood pressure and poorperipheral flow. The patient was receiving inotro-pic support. The patient was temporarily discon-nected from mechanical ventilation to facilitatetracheal suction and physiotherapy, and it wasnoted that, when disconnected, the filling pressuredropped, the blood pressure increased and theheart rate fell. In this patient, lung over-inflationraised alveolar pressure and this was responsiblefor the haemodynamic compromise. AutoPEEPacts to raise intrathoracic pressure, impedesvenous return and therefore decreases stroke vol-ume and cardiac output.

AutoPEEP was simply measured by Pepe andMarini by using the pressure gauge of the simple

ventilator and occluding the expiratory limb of theventilator tubing (Fig. 4.5). This allows the pressurein the alveolus to equilibrate along the airways and tobe transmitted to the inspiratory pressure gauge. Atocclusion, the inspiratory pressure jumps up to thelevel approximate to the level of autoPEEP.

AutoPEEP is now measured automatically on ven-tilators. The method occludes the inspiratory andexpiratory valves and allows the distal airways toequilibrate. It measures the upper airway pressureafter a set period of up to 15 s. However, in severeobstruction, this may not be sufficient time for allairways to equilibrate and care should be taken if thistechnique is employed in these patients.

For many patients, autoPEEP occurs during spon-taneous ventilation, with or without some form ofventilator assistance such as pressure support. Theocclusion method of measuring autoPEEP is notappropriate in this circumstance, as the patients can-not stop breathing to enable the measurment to takeplace. The presence of autoPEEP may be suspected inpatients who continue to breathe out throughoutexpiration up to and until the next breath deliveredfrom the ventilator.

Clinical implications

If high, autoPeep prevents venous filling, reducesend-diastolic ventricular volume and thereforereduces stroke volume. When severe, as in the casedescribed by Pepe and Marini, the effect is similar tothat seen in cardiac failure, as the pressure within thechest is easy to overlook.

Intrinsic PEEP must be overcome before gasflows into the chest. It acts as an additional load tobreathing and ventilators will only cycle to inspira-tion when autoPEEP is equalled in the upperairway.

The effect of a threshold load induced byautoPEEP can be offset to some extent by the addi-tion of external PEEP. If the pressure in the upperairways is increased by the addition of external PEEP,the effect is to balance the distal and proximal pres-sures.15. As the difference between the two sites ofpressure measurement is reduced, so the pressure tobegin flow in the upper airway decreases. When theyare perfectly matched, the inspiratory effort neces-sary from the patient to initiate inspiratory gas flowand to make the ventilator 'cycle' to inspiration is


Figure 4.5 (a) The alveolus at the

end of expiration cannot empty

because of the obstruction, and the

pressure inside is raised

(15 cmH20). In the upper airways,

in the ventilator tubing and at the

ventilator, the pressure rapidly falls

to atmospheric (zero). In order to

measure the pressure inside the

alveolus, the expiratory limb of the

ventilator tubing is occluded, as

illustrated in (b). The pressure in

the alveolus is now transmitted to

the upper airways, the ventilator

tubing and the pressure gauge

inside the ventilator itself. This is

clearly seen when the pressure

gauge 'flicks' upwards, indicating

the higher pressure inside the

alveolus (15 cmH20)

minimal. Such a reduction in 'triggering threshold'also ensures that the ventilator provides pressuresupport for each spontaneous breath made by thepatient (Fig. 4.6).

If external PEEP is added to the point in excess ofthe level of internal PEEP, the effect will be to causefurther hyperinflation and the effect of reducing thethreshold load is offset. Care should be taken toavoid a further increase in lung volume with theapplication of distending patterns of ventilation.

RESPIRATORY DRIVE

Breathing during mechanical ventilation is com-monplace and respiratory drive may be affected pro-foundly in the critically ill. At the simplest level,sedative drugs and neuromuscular agents maydecrease drive and patients may not trigger mechan-ical ventilation. The effect of sedative drugs tends tooffset the response curve of the central nervous

Respiratory drive 55

Figure 4.6 Tracings of flow, airway

pressure and oesophageal pressure

during mechanical ventilation. The

upper panel shows many inspiratory

efforts that are not sensed by the

ventilator (asynchrony). When

external PEEP is added to balance

the autoPEEP effect, triggering is

successful (bottom panel).

system to CO2 and respiratory acidosis, such thathigher stimulation levels are needed to begin respira-tion. Modest over-ventilation may then suppressventilation and apnoea occurs unless ventilation isadjusted to achieve a higher level of PaCO2 whenspontaneous respiration will occur. This may be oflittle significance in most normal patients recoveringafter brief periods of ventilation after, for example,elective surgery. For patients who may require highlevels of central nervous system drive to breathe,reduction of drive may be significant.

The response of the respiratory muscles to stimu-lation may be expressed as a frequency-force curve.The relationship between stimulation and responseof a skeletal muscle is sigmoid shaped and the curvemarkedly flattens off as stimulation frequency reaches40 Hz (Fig. 4.7). Above this frequency, little increasein force is generated.

Normal ventilation occurs at low stimulationfrequency and lies on the steep part of theforce-frequency curve and consequently largeincreases in minute ventilation will result from an


Figure 4.7 The force generated during respiratory muscle

contraction is plotted against the frequency of stimulation. Force

rapidly increases as stimulation rises above 5 Hz. There is little

increase in force after 40 Hz.

increase in stimulation frequency. This is wellillustrated in patients with a metabolic acidosis whoare able to near-normalize their acid-base status bysubstantial increases in ventilation. Many patientswho are critically ill are breathing at the upper rangeof their frequency-force curve, for example patientswith intrinsic respiratory disease such as COPD. Inthis circumstance, maintenance of respiratory driveis essential. Modest changes in output provokesubstantial changes in force generation, precipitatingrespiratory failure.

The measurement of respiratory drive

It is not possible to measure the output of a motornerve directly in humans because the amplitude ofthe electrical potentials cannot be detected percutan-eously. Phrenic neurograms may be measured whenthe nerve can be directly exposed, and this is the casewhen the phrenic nerve is stimulated to enable nativerespiration when the spinal cord is damaged at a highlevel and the patient is ventilator dependent.

An alternative to direct measurement is to assessthe result of central drive in terms of the responseof the respiratory muscles to the stimulus. For the

respiratory muscles, tension within the muscle pro-duces a change in pressure within the chest andthus the response to stimulation is a change in pres-sure. The pressure is measured at the airway by sim-ple apparatus. Central to the technique is that theairway has to be occluded (often by a rapidlyresponding solenoid valve) and, if this is donebriefly, the respiratory muscles are isometric andthe subject cannot detect this brief occlusion andtherefore the respiratory pattern is not influencedby the manoeuvre.

In order to standardize the measurement, thepressure generated in the first 100 ms of anoccluded breath is measured and this is termedairway occlusion pressure, or P0.1 P0.1 is raised whenrespiratory drive is elevated artificially duringa hypercapnic challenge and is also high in patientsin ventilatory failure. As P0.1 is measured before gasflows into the chest and prior to lung inflation,changes in lung mechanics do not affect itsmeasurement. However, as changes in the length ofthe respiratory muscles will alter the generatedpressure for a given stimulus, it is important thatlung volume does not alter.

The value of P0.1 in assessing patientsweaning from mechanical ventilation

It is easy to apply the technique to ventilatedpatients and, when respiratory drive is elevated, themeasured pressure exceeds 5.5 cmH2O. Duringweaning, when a raised P0.1 is found, patients fail towean. Interestingly, patients who are able to breatheduring weaning trials not only have a low P0.1 butare also able to increase drive and minute ventila-tion during a hypercapnic challenge. Patients whoare able to breathe spontaneously do so with alower central drive and also have some ventilatoryreserve, contrasting with the fixed capacity ofpatients who fail.

When airway resistance is high, a considerabletime delay may occur between the onset of a negativeintrathoracic pressure and pressure changes in theupper airway or mouth. In severe disease, airwayocclusion pressure no longer reflects the pressuregenerated by the muscles and intrathoracic and air-way pressures are not related.

Recently, P0.1 has been used to assess how muchhyperinflation and autoPEEP is present in patientsbreathing spontaneously. As hyperinflation increases,

References 57

respiratory drive must accommodate the increasedthreshold load mentioned previously. The increaseddrive can be detected as an increase in P0.1 and dis-tinguishes hyperinflated patients.16

REFERENCES

1. Sassoon, CSH. Mechanical ventilator design and

function: the trigger variable. Respir Care 1992; 37:

1056-69.

2. Sassoon, CSH, Gruer, SE. Characteristics of the

ventilator pressure and flow-trigger variables.


3. Flick, GR, Bellamy, PE, Simmons, DH. Diaphragmatic

contraction during assisted mechanical ventilation.

C/7«M 989; 96: 130-5.

4. Murciano, D, Aubier, M, Bussi, S, Derenne, JP, Pariente,

R, Milic-Emili, J. Comparison of esophageal, tracheal

and occlusion pressure in patients with chronic

obstructive pulmonary disease during acute respiratory

failure. Am Rev Respir Dis 1982; 126: 837-41.

5. Fabry, B, Guttman, J, Eberhard, L et al. An analysis of

desynchronization between the spontaneously

breathing patient and ventilator during inspiratory

pressure support. Chest 1995; 107: 1387-94.

6. Marini, JJ, Capps, JS, Culver, BH. The inspiratory work

of breathing during assisted mechanical ventilation.

Chesn 985; 87: 612-18.

7. Marini, JJ, Rodriguez, RM, Lamb, V. The inspiratory

workload of patient-initiated mechanical ventilation.


8. Marini, JJ, Rodriguez, RM, Lamb, V. Bedside

estimation of the inspiratory work of breathing

during mechanical ventilation. Chest 1986;

89: 56-63.9. Collett, PW, Perry, C, Engel, LA. Pressure-time

product, flow, and oxygen cost of resistive

breathing in humans. J Appl Physiol 1985; 58:

1263-72.

10. Harpin, RP, Baker, JP, Downer, JP, Whitwell, J,

Gallacher, WN. Correlation of the oxygen cost of

breathing and length of weaning from mechanical

ventilation. Crit Care Med 1987; 15: 807-9.

11. Fleury, B, Murciano, D, Talamo, Cetal. Work of

breathing in patients with chronic obstructive

pulmonary disease in acute respiratory failure.


12. Pinsky, MR. The effects of mechanical ventilation on

the cardiovascular system. Philadelphia: WB Saunders,

1990.

13. Hakim, TS, Michel RP, Chang HK. Effect of lung

inflation on pulmonary vascular resistance by

arterial and venous occlusion. J Appl Physiol 1982;

53: 1110-15.

14. Pepe, PE, Marini, JJ. Occult positive end-expiratory

pressure in mechanically ventilated patients with

airflow obstruction. Am Rev Respir Dis 1982; 126:

166-70.

15. Tobin, MJ, Lodato, RF. PEEP, Auto-PEEP, and waterfalls.

Chest *\ 989; 96: 449-51.

16. Mancebo, J, Albaladejo, P, Touchard, D, et al. Airway

occlusion pressure to titrate positive end-expiratory

pressure in patients with dynamic hyperinflation.

Anesthesiology 2000; 93: 81-90.

5Non-invasive mechanical ventilationin acute respiratory failureBERNDSCHONHOFER

Introduction and historical background 58

Non-invasive versus invasive mechanical ventilation 58Spectrum of acute respiratory failure treated

with non-invasive mechanical ventilation 59

Technical aspects: interface 59

Ventilator modes for non-invasive mechanicalventilation 59

Practical guidelines to manage acute respiratoryfailure 61

Implementation and duration of non-invasivemechanical ventilation in acute respiratory failure 62

Duration of non-invasive mechanical ventilation inacute respiratory failure 65

Non-invasive mechanical ventilation and outcome 65

References 67

INTRODUCTION AND HISTORICALBACKGROUND

Mechanical ventilation can be provided by either neg-ative or positive pressure ventilation. During thepolio epidemics of the 1950s and 1960s, negativepressure ventilation (NPV) was widely used.1

Ventilation is supported by exposing the chest to sub-atmospheric pressure to produce inspiration andallowing expiration by returning the pressure aroundthe chest wall to atmospheric pressure. Severaldevices, such as the Emerson iron lung, cuirass andthe Ponchowrap, are available. Although NPV can besuccessfully employed in acute respiratory failure,2

because of the size of the devices, the lack of access tothe patient and the danger of inducing upper airwayobstruction3 it is not widely used. In the second halfof the twentieth century, invasive mechanical ventila-tion (IMV) was provided through a cuffed endotra-cheal or tracheostomy tube and became the standardmethod of providing artificial ventilation as it guar-antees control of the airway and the ability to correct

acid-base and gas disturbances with security. It alsoallows close monitoring and control of airway pres-sures and of tidal volume. Non-invasive mechanicalventilation (NIV), using facial or nasal masks as inter-faces, was introduced as an alternative in the late1970s and early 1980s, usually for domiciliary ther-apy. According to the underlying pathophysiology ofrespiratory failure, NIV may be effective in improvingalveolar ventilation, reducing dyspnoea, resting therespiratory muscles and reducing dynamic hyperin-flation. This chapter focuses on the impact of NIV inacute respiratory failure (ARF), in which it is increas-ingly applied following a pilot study involvingpatients with hypercapnic ARF in the early 1990s.4

NON-INVASIVE VERSUS INVASIVEMECHANICAL VENTILATION

The main advantage of IMV is the secure access to andprotection of the airways. However, a range of adverseeffects or complications is associated with IMV (Table

Ventilator modes for non-invasive mechanical ventilation 59

Table 5.1 Advantages (+) and disadvantages (-) of

invasive mechanical ventilation (IMV) and non-invasive

mechanical ventilation (NIV)

Ventilator-associated penumonia — +Additional resistive work of breathing — +Early and late tracheal injuries — +Deep sedation, paralysis - +

Intermittent application of ventilator — +Cough possible - +

Eating possible - +Communication possible — +Difficult weaning — +

Protection of airways + -Access to airways + -Facial side effects + -Leakage + —

Claustrophobia + —Aerophagia + -

5.1). Ventilator-associated pneumonia (VAP) is per-haps the most important,5'6 the incidence of which isdependent on the duration of IMV as, after 3-5 days,the rate of VAP increases significantly.7 The rate ofnosocomial infection appears to be lower in patientsreceiving NIV as first-line therapy.8,9

SPECTRUM OF ACUTE RESPIRATORYFAILURE TREATED WITH NON-INVASIVEMECHANICAL VENTILATION

Controlled and uncontrolled trials of NIV have beenconducted in a broad spectrum of causes of ARF10 Alimiting factor is that exclusion criteria in manyresulted in only a minority of patients beingrandomized to NIV.11 As a consequence, the resultsapply to a restricted patient population and may notbe generally applicable. From the pathophysiologicalpoint of view, it is useful to differentiate between thehypercapnic and hypoxaemic type of ARF (Table 5.2;see also section 'Non-invasive mechanical ventilationand outcome', below).

TECHNICAL ASPECTS: INTERFACE

One of the most crucial issues is the interface. Severaltypes are commercially available: full facemask, nasal

Table 5.2 Causes of hypercapnic acute respiratory failure

(ARF) in which non-invasive ventilation has been applied

Hyercapnic ARFAcute exacerbation of COPDPost-extubationalAcute exacerbation of asthma

Cystic fibrosis

Hypoxaemic ARFCardiogenic pulmonary oedema

Organ transplantationCommunity-acquired pneumonia

PostoperativeTraumaAtelectasisOpportunistic pneumonia in HIVImmunocompromisedNear-drowningLung cancer

Pulmonary embolus

YesYesNoNo

YesYesYesYesNoNoNoYesNoNoNo

YesYesYesYes

YesYesYesYesYesYesYesYesYesYesYes

RCT, randomized, controlled trials; NRCT, non-randomized controlledtrials; COPD, chronic obstructive pulmonary disease; HIV, humanimmunodeficiency virus.

mask, nasal pillows (Fig. 5.1). In some cases, mouth-pieces and custom fabricated masks may also beuseful. In ARF, the full facemask is usually preferableto nasal masks in our experience. The advantages anddisadvantages of mask types are given in Table 5.3.

VENTILATOR MODES FOR NON-INVASIVEMECHANICAL VENTILATION

Both volume-targeted12 and pressure-targeted venti-lation11'13 can be employed. Critical care ventilators,characterized by high technical quality and elaboratemonitoring, and simpler and smaller ventilators,often used for home mechanical ventilation, can beused. In one comparison of the intensive care unit(ICU)-type ventilator and six portable 'home'devices, there were differences between the smaller'home' devices in terms of re-breathing, the speed toa stable level of pressure support and expiratory resis-tance.14 These differences may have clinical impact.Another study investigated the technical performanceof nine ventilators used for acute NIV compared to anICU ventilator.15 The authors found that most pres-sure ventilators evaluated were able to respondto high ventilatory demands and even outperformedthe ICU device! If the ventilator only supports

60 Non-invasive mechanical ventilation in acute respiratory failure

Table 5.3 Advantages (+) and disadvantages [spectrum

from (-) to -] of nasal masks andfacemasks

Mouth leak + —Mouth breathing

and quality of NIV + -Influence of dental status + —Airway pressure + —Dynamic of improved ABG + -Dead space (-) +Communication - +Eating, drinking - +Expectoration — +Risk of aspiration - +Risk of aerophagia - +Claustrophobia — +Comfort (-) +

NIV, non-invasive ventilation; ABG, arterial blood gases.

Figure 5.1 Interfaces:

facemask (a); nasal mask (b);

nasal mask with little dead

space (c); nasal pillows (d).

spontaneous breathing, each breath being triggered,the mode is termed pressure support ventilation(PSV) or assisted spontaneous breathing (ASB). PSVis the most commonly used mode in ARF as it allowsthe patient more control, which aids tolerance. Animportant advantage of PSV (over volume support)is the compensation for mild to moderate leak.Patients with ARF are often characterized by agitation,irregular breathing, intrinsic positive end-expiratorypressure (PEEPi) and sleep deprivation. PSV mayfacilitate patient-ventilator synchrony, whereas theaddition of external PEEP will reduce inspiratorymuscle work as PEEPi must be overcome beforeinspiration can begin.16 In the sleep-deprived patient,however, triggering cannot be guaranteed and atimed mode of ventilation may be better in thepatient with more severe respiratory failure.

CO2-re-breathing may occur with NIV. The risk isgreater with a single delivery circuit without a trueexhalation valve.14'17 High respiratory rates and lowexternal PEEP increase the risk of CO2-re-breathingbecause of the shorter expiratory time and lowerwash-out of the circuit. A minimal expiratory posi-tive airway pressure (EPAP) of 2-4 cmH2O is usuallynecessary to reduce CO2-re-breathing in a singletube circuit.

At present, there are no generally accepted re-commendations on how to set up NIV. Considerationof gas exchange, muscle load and breathing pattern isappropriate when setting the optimal PSV for eachindividual patient. Published studies have used avariety of end-points as targets for therapy, such aspatient comfort, the level of pressure or volume sup-port, blood gases or breathing pattern.10-13'18'19

When using bi-level NIV, we start with inspiratorypositive airways pressure (IPAP) of 8-10 cmH2O andEPAP of 2-4 cmH2O and quickly increase IPAP as thepatient settles. Subsequent adjustment will thendepend on the underlying diagnosis, the patient'stolerance and comfort and the physiological para-meters such as O2 saturation (SaO2), minute ventila-tion (aiming for an estimated tidal volume of 10-15mL kg -1), fall in respiratory rate and disappearanceof accessory muscle activity. In ARF due to chronicobstructive pulmonary disease (COPD), an inspira-tory pressure support of 15-25 cmH2O and EPAP of4-6 cmH2O would commonly be used. NIV can be

Practical guidelines to manage acute respiratory failure 61

applied with a mandatory back-up rate, e.g. 12breaths min-1 as patient effort may fall in sleep.Supplemental O2 (2-10 L min-1) should be admin-istered via the ventilator tubing to maintain the SaO2

>90%. Some higher-specification machines allowcontrol of inspired O2.

Triggering is crucial to the success of PSV, i.e. thedetection of patient inspiration. Some ventilatorshave a fixed and others a variable trigger. In thepast, ventilators were often pressure triggered. Flowtriggering is more sensitive,20 although so-calledauto-cycling may develop due to air leak.21 Thisresults in rapid switching between IPAP and EPAP,with no real benefit to the patient. Increasing theflow needed to trigger (making it less sensitive) willmake this less likely to occur. A new development isa 'moving time' analysis of the pressure contourduring a delivered breath (within milliseconds) andthese technical advances may improve triggeringand comfort. Another technical challenge is howquickly and adequately the ventilator can reachpressure despite variable patient demand.Inspiratory flow depends on the underlying patho-physiology, e.g. resistance and compliance, thegiven pressure support and inspiratory rise time.The quicker the machine can reach pressure, thelower the work of breathing.22 Depending on thetype of machine, the pressure rise time, or slope,may be adjustable or be manufacturer fixed. A slowrise time is primarily provided for domiciliary use,e.g. for comfort in the non-breathless neuromuscu-lar patient. In ARF, rise time should be short.Indeed, depending on leak, some devices are insuf-ficiently powered to support the short inspiratorytime and high pressure support requirements of thedyspnoeic COPD patient. Finally, the detection ofend-inspiration may be important for patient com-fort. It is adjustable on some ventilators and fixedon others, i.e. switching to expiration when inspirat-ory flow reaches 70% of the initial flow rate. Again,this is more relevant to domiciliary, long-term ven-tilation, but may be important when using thespontaneous mode of ventilatory support.

Allowing the patient to maintain control of thebreathing pattern may increase compliance.Proportional assist ventilation (PAV) has been pro-posed as a mode of synchronized partial ventilatorysupport that unloads both the resistive and elasticburdens and in which support provided is propor-tional to instantaneous patient effort.23-25 The

majority of studies investigating PAV have been shortterm. The clinical value of PAV as a NIV mode totreat ARF has not yet been convincingly shown.Controlled ventilation, either as volume assist con-trol (ACV) or pressure-controlled ventilation (PCV),does not require patient effort and cycles automatic-ally if there is no or insufficient ventilatory effort. Incontrast to chronic domiciliary ventilation, ACV orPCV is less frequently used when treating ARF, butmay be indicated when there is severe overload of therespiratory muscles, profound sleep deprivation oracute O2-induced narcosis resulting in minimal res-piratory effort. These modes should be considered inthe failing patient before resorting to endotrachealintubation if to do so is considered safe.

Some NIV studies have compared PSV with vol-ume assist control ventilation in ARE18'26'27 Nodifferences were found in outcome, blood-gaschanges and degree of muscle rest.26'27 PSV did,however, have a reduced incidence of side effectsand was more comfortable,18'27 with better leakcompensation. 18

PRACTICAL GUIDELINES TO MANAGEACUTE RESPIRATORY FAILURE

Based on the literature and clinical experience, apractical algorithm is depicted in Figure 5.2.Compared to conventional treatment with bron-chodilators and steroids, NIV is not indicated inpatients with a pH >7.35.28 In other words, patientsin whom ARF is not severe enough do not profitfrom the additional application of NIV. Patients withARF being considered for NIV should have pH>7.20 and <7.35.11,29 The use of pH as severityindicator is better than PCO2 as it distinguishes theacute component of respiratory failure. Using PCO2

alone would not allow for the chronic component. Itis important, when inspecting the pH and other arter-ial blood-gas results, to take note of any metaboliccomponent that may indicate failure of tissue oxy-genation and the development of a lactic acidosis.

IMV and NIV should be seen as complementarytreatments in a patient needing ventilatory support.IMV remains the preferred treatment if there are con-traindications to NIV (Table 5.4). However, in theearly stages, NIV has definite advantages. Treatmentstrategies should be based on both clinical aspects


Figure 5.2 Algorithm for the use

of non-invasive ventilation in acute

respiratory failure.

Table 5.4 Contraindications to non-invasive mechanical

ventilation

Severe acidosis at admission (pH <7.1)Coma and massive confusionMassive psychomotor agitationSignificant co-morbidityOrofacial abnormalitiesIrreversible mask intolerance and recent facial surgeryHaemodynamic instability (systolic blood pressure

<70 mmHg)Irreversible hypersecretionLife-threatening, refractory hypoxaemiaRespiratory arrestIntubation needed to protect upper airways (coma, acute

abdominal process)Glottic oedema or closure

(comfort, mental status and mouth and mask leak)and monitoring of physiological parameters (SaO2,pH, breathing frequency and tidal volume) (Fig. 5.3).Success or failure should be judged within the first1-2 hours (see Fig. 5.2). A lack of improvement inblood gases, a high initial severity of illness score andpoor tolerance of NIV are predictors of failure.30,31 Ifpatients fail to improve, it is important that thereis rapid access to intubation as delay increases mor-tality.32 In addition to using NIV to prevent intuba-

tion and IMV, it may be employed to accelerate extu-bation (see Fig. 5.2), thereby reducing the potentialfor complications, by shortening the period of IMV,and also avoiding the need for re-intubation.33

IMPLEMENTATION AND DURATION OFNON-INVASIVE MECHANICALVENTILATION IN ACUTE RESPIRATORYFAILURE

Despite the evidence for the benefits of NIV, provi-sion is still limited by the lack of resources and,especially, trained personnel. In a survey investigat-ing NIV in acute COPD in the UK, staff and equip-ment were available in fewer than half of the acutecare hospitals.34 In those hospitals in which NIVwas available, it was generally underused. Lack oftraining, problems with funding and doubts aboutits value were given as reasons for the failure to pro-vide a comprehensive 'out of hours' NIV service. Ameta-analysis of randomized trials10 may haveoverestimated its value due to the highly selectedstudy populations included in these studies. Resultsachieved in enthusiastic departments as part ofclinical trials may not be achievable in the realworld. In a French multi-centre trial, involving

Implementation and duration of non-invasive mechanical ventilaiton in acute respiratory failure 63

Figure 5.3 Variables to evaluate non-invasive ventilation. SaO-,, Q-, saturation; fB, breathing frequency; Vt, tidal volume.

more than 40 ICUs, poor patient tolerance was amajor cause of failure.35 Success may depend on thequality of NIV provision and yet a complex spec-trum of variables influences the implementationphase (Table 5.5). The location for NIV depends onlocal factors such as equipment, the skills of doc-tors, nurses and therapists and their 24-hour avail-ability. Patients with a high likelihood of failure(pH <7.25) should be initiated in the ICU or emer-gency room and, if stabilized, transferred to a highdependency unit (HDU) or specialist ward. In alarge, randomized study, the practicality of the earlyuse of NIV in a respiratory ward was examined.29

For patients with mild to moderate acidosis (pH

Table 5.5 Factors influencing the success of non-invasive

mechanical ventilation

LocationInterfaceVentilator type, mode and settingMonitoringIndicationMotivation of the staffTraining of the staffExperience of the staffSize of staffTime consumptionOrganization of team

between 7.25 and 7.35), NIV let to a more rapidimprovement in physiological parameters, a reduc-tion in the need for IMV and a significant reductionin hospital mortality. Patients were treated accord-ing to a simple protocol using a spontaneous modebi-level ventilator. Although the providers of theservice were not experts, a considerable amountof training was needed, which was required to beon-going. Some practical aspects are illustrated bythe following case reports.

SUCCESSFUL NON-INVASIVE MECHANICALVENTILATION IN HYPERCAPNIC ACUTE

RESPIRATORY FAILURE (FIG. 5.4)

A patient with end-stage COPD (male, 64 years,FEV1 as an outpatient: 0.70 L, 37% predicted) wasadmitted with acute exacerbation of respiratoryfailure. Initial blood gases: PCO2 8.5, PO2 4.5 (air)and 6.5 (kPa) with O2 at 2 L min-1, pH 7.28 andrespiratory rate 27 min-1. He was co-operative,but agitated, and haemodynamically stable.

The team decided to initiate NIV with a nasalmask. An explanation was given. The head of thebed was maintained at an angle of 45° and a lowdose of morphine (5 mg) was given to reduceagitation. Due to mouth breathing, the nasal mask

(cont.)


(cont.)needed to be switched to a full facemask and, toimprove the acceptance of NIV, the facemask wasinitially connected to an Ambu-bag. The patientwas ventilated manually in time with his breathingfrequency. Thereafter PSV was given with IPAP at10 cmH2O, progressively increasing to 20 cmH2O.However, excessive respiratory secretions limitedacceptability (the patient kept taking the mask offto cough and, after an initial fall in PCO2, it beganto increase again). Fibreoptic bronchoscopy wasperformed, employing a specially adapted mask,without interrupting NIV.36 After 4 hours ofcontinuous NIV, PCO2 was 7.4 and pH 7.33. AnNIV-free interval of 15 min allowed the patient tocommunicate, expectorate and drink. He thencontinued with NIV. During the first 24 hours onthe ICU, NIV was applied for a total of 20 hours.The patient was then transferred to the general res-piratory ward, but continued with NIV intermit-tently and overnight until blood gases normalized.

NON-INVASIVE MECHANICAL VENTILATIONFAILURE IN A PATIENT WITH HYPOXICACUTE RESPIRATORY FAILURE DUE TOPNEUMONIA (FIG. 5.5)

A patient (male, 42 years) was admitted to hospital.Chest X-ray showed pneumonia involving the right

upper and middle lobe. On the ICU, blood gaseswere: PO2 4.2 (air) and with 60% O2 8.8 kPa, PCO2

3.3 (kPa) and pH 7.45. He was agitated and tachy-pnoeic (respiratory rate 34 min-1).

NIV was attempted, but the initial period wascharacterized by several drawbacks: the nasalmask was ineffective and the full facemask leakedand produced claustrophobia. PSV mode waschosen with an inspiratory support of 16 cmH2O.However, patient-ventilator dyssynchrony wasapparent. During the first 30 min of NIV, thepatient's clinical state deteriorated and he becameconfused, the blood pressure decreased and theheart rate increased to 130 min-1. Arterial bloodgases also failed to improve. NIV failure wasobvious and the patient was intubated 2 hoursafter NIV was initiated. He was ventilated for thefollowing 6 days and required inotropes for theinitial 3 days. The pneumonia resolved withbroad-spectrum antibiotic treatment. Afterwards,he was successfully weaned without furtherproblems.

Comment This type of patient will often not besuccessfully managed by NIV and attempting itmay expose the patient to risk. In some patientswith type 1 respiratory failure, a trial of NIVis appropriate, especially when rapid recovery islikely, for example from pulmonary oedema.

Figure 5.4 Case report: non-invasive ventilation (NIV) in

hypercapnic acute respiratory failure, (fb = breathing frequency.)

Figure 5.5 Case report: non-invasive ventilation (NIV) failure

in non-hypercapnic acute respiratory failure. (IMV= invasive

mechanical ventilation; fb = breathing frequency.)

Non-invasive mechanical ventilation and outcome 65

DURATION OF NON-INVASIVEMECHANICAL VENTILATION IN ACUTERESPIRATORY FAILURE

Acute NIV is employed to overcome a life-threaten-ing crisis and the duration of treatment will dependon the precipitating cause. The shortest duration ofNIV has been reported in cardiopulmonary oedema(3 hours37 or 9.3 ± 4.9 hours38). In contrast, it takeslonger to recover from ARF in COPD. The first 24hours is the most crucial time whilst awaiting bene-fit from conventional therapy such as steroids,antibiotics and bronchodilators. We use NIV as muchas possible in the first 48 hours. According to the liter-ature, the mean compliance during the first day variesfrom 4 hours39 to 6-8 hours.11,12,29,40 The mean totalduration of NIV ranges between 4 and 25 days.11-13,41

Table 5.6 Randomized, controlled trials investigating non-

invasive mechanical ventilation in non-hypercapnic acute

respiratory failure

Kramer (13)a

Wysocki (19)a

Antonelli (46)Wood (32)Confalonieri (45)a

Antonelli (47)Martin (43)a

\\=/=/=

\\0=

aOnly a subpopulation of patients with type 1 failure. ET, management byintubation and the provision of mechanical ventilation via an endotrachealtube.

trials. Although in hypoxaemic ARF NIV didnot reduce mortality, a significant reduction inintubation rate was found in three trials.43'45'47

NON-INVASIVE MECHANICALVENTILATION AND OUTCOME

Several aspects have been investigated that reflectoutcome: intubation and mortality rate, length ofstay in ICU and hospital, duration of mechanicalventilation, complication rate (VAP, sinusitis, aspira-tion, nasal ulceration), cost-effectiveness and timeconsumption of nursing or medical staff.

Hypercapnic and hypoxaemic acuterespiratory failure

The addition of NIV to standard therapy in patientswith hypercapnic exacerbations of COPD improvessurvival and decreases the intubation rate.10 Twofurther controlled trials have confirmed thesefindings.42,43 COPD patients treated with NIV alsohave a better survival following discharge comparedto conventional treatment,44,45 although this apparentbenefit could be due to selection by survival duringthe critical illness rather than the effect of NIV versusIMV. Aspects, such as muscle wasting and generalnutritional status at discharge could also affectlong-term survival. NIV has not been unequivocallyshown to be superior to other management strategiesin hypercapnic ARF due to causes other than COPDand cardiogenic pulmonary oedema. Table 5.6 sum-marizes the results of seven randomized, controlled

Cardiogenic pulmonary oedema

The mechanisms by which continuous positive airwaypressure (CPAP) or PSV plus CPAP (bi-level PSV) areeffective in cardiogenic pulmonary oedema are multi-factorial and include lung recruitment, counterbal-ancing PEEPi, so reducing the work of breathing,decreasing shunt and reducing pre-load and after-load. When compared to conventional treatment,CPAP improves gas exchange and reduces the intuba-tion rate.37,38,48 Improvement in other outcomes, suchas ICU complications, length of stay or mortality, hasnot been demonstrated. The best modality suited toacute cardiogenic pulmonary oedema has not beeninvestigated. However, adding PSV to CPAP providesgreater assistance to inspiration. Therefore, in cardio-genic pulmonary oedema, those with compromisedrespiratory muscles - indicated by hypercapnia - maybenefit most from NIV.49 PSV may increase the risk ofacute myocardial infarction in the presence of symp-tomatic angina.50 Whether this is a true causal rela-tionship is unclear, but those at risk may be patientswith relative hypovolaemia, from diuretic therapy, orthose in atrial fibrillation.

Post-operative patients

Most published evidence for this indication is anec-dotal. Both PSV and CPAP have been used. However,


some randomized, controlled trials in cardiac, pul-monary, abdominal and transplant surgery havebeen performed.47,51,56 It was found that NIVimproves physiological parameters without apparentserious side effects. Whether NIV can also modifyclinical outcome is unclear.

Patients who are not candidates forinvasive mechanical ventilation

The concept of employing NIV when the patient is nota candidate for intubation is an important aspect.Reasons may include advanced physiological age,cachexia or end-stage disease or advance directive. NIVmay also provide time for the decision-makingprocess. It aims both to provide effective ventilatorysupport and to increase patient comfort. Availableevidence on comfort is based on retrospective oruncontrolled prospective series. In addition to efficacy,which may be up to 70%,31,57,58 NIV does appear to bewell accepted or tolerated. However, NIV should notbe used to prolong the inevitable course towards death.The ethical and economic problems of management inARF59 should be better examined in future studies.

Non-invasive mechanical ventilationin weaning, avoiding re-intubationand use at home

In more than 20% of COPD patients receiving IMV,weaning is delayed and sometimes impossible.60 NIVmay be a useful addition to existing weaning strat-egies. The use of NIV in this way has been reported inuncontrolled studies.61-63 However, in a multi-centre,randomized study, NIV was performed after short-term IMV.33 It improved weaning success, survivalrate, total time of mechanical ventilation and length ofstay in an ICU. Post-extubation failure is also relativelycommon64 and the prognosis is poor, with a hospitalmortality exceeding 30%. The high mortality ofpost-extubation failure may relate to clinical deteri-oration during the period of unsupported ventilation.

A historically controlled study demonstrated the useof NIV to treat post-extubation failure65 The need forre-intubation, the duration of ventilatory assistanceand the length of stay in the ICU were all significantlyreduced by NIV. The following is illustrative (Fig. 5.6).

Figure 5.6 Case report: non-invasive ventilation (NIV) following

invasive mechanical ventilation (IMV) in advanced chronic

obstructive pulmonary disease, (fb = breathing frequency.)

CASE HISTORY

A severely dyspnoeic patient with end-stage COPD(female, 67 years, FEVj 0.58 L, 32% predicted) wasintubated by the paramedics at home. Chest radio-graphy (CXR) did not show pneumonia. After 36hours, sedation was discontinued and, following asuccessful T-piece trial, the patient was extubatedand NIV was initiated. During the next 6 days, herrespiratory and clinical state improved, but afterdiscontinuing NIV, she again deteriorated, withprogressive hypercapnia. Nocturnal NIV was initi-ated and led to improved daytime vigilance and afall in -PCO2 and the patient was discharged to con-tinue this at home.

Comment The role of NIV domiciliary for COPD iscontentious, but the practice is increasing. Frequentadmissions or significant evidence of sleep-disorderedventilation, with consequent daytime hypersomno-lence, are probably the best guides to therapy. Designingproperly controlled trials is difficult because of mixedaetiology (OSA, obesity hypoventilation). In currentlyreported studies, OSA has been specifically excluded,and patients with advanced disease, and predominatelyadvanced emphysema, may explain why, so far, suchstudies of domiciliary NIV have been negative.

References 67

REFERENCES

1. Drinker, P, Shaw, LA. An apparatus for the prolonged

administration of artificial respiration. I. A design for

adults and children.) Clin Invest 1929; 7: 229-47.

2. Corrado, A, Gorini, M, Villella, G, De Paola, E.

Negative pressure ventilation in the treatment of

acute respiratory failure: an old noninvasive

technique reconsidered. Eur RespirJ 1996; 9:

1531-44.

3. Levy, RD, Bradley, TD, Newman, SL, et al. Negative

pressure ventilation. Effects on ventilation during

sleep in normal subjects. Chest 1989; 95: 95-9.

4. Meduri, GU, Abou-Shala, N, Fox, RC, et al. Noninvasive

face mask mechanical ventilation in patients with

acute hypercapnic respiratory failure. C/»esM991;

100: 445-54.

5. Torres, A, Aznar, R, Gatell, JM, et al. Incidence, risk, and

prognosis factors of nosocomial pneumonia in

mechanically ventilated patients. Am Rev Respir Dis

1990; 142: 523-8.

6. Cook, DJ, Walter, SD, Cook, RJ, et al. Incidence of and

risk factors for ventilator-associated pneumonia in

critically ill patients. Ann Intern Med 1998; 129:

433-40.

7. Ruiz-Santana, S, Garcia Jimenez, A, Esteban, A, et al.

ICU pneumonias: a multi-institutional study. CritCare

Med 1987; 15: 930-2.

8. Guerin, C, Girard, R, Chemorin, C, et al. Facial mask

noninvasive mechanical ventilation reduces the

incidence of nosocomial pneumonia. A prospective

epidemiological survey from a single ICU. Intensive

Care Med 1997; 23:1024-32.

9. Nourdine, K, Combes, P, Carton, MJ, et al. Does

noninvasive ventilation reduce the ICU nosocomial

infection risk? A prospective clinical survey. Intensive

Care Med 1999; 25: 567-73.

10. Keenan, SP, Kernerman, PD, Cook, Dj, et al. Effect of

noninvasive positive pressure ventilation on mortality

in patients admitted with acute respiratory failure: a

meta-analysis. Crit Care Med 1997; 25: 1685-92.

11. Brochard, L, Mancebo, J, Wysocki, M, et al.

Noninvasive ventilation for acute exacerbations of

chronic obstructive pulmonary disease. N Engl J Med

1995; 333: 817-22.

12. Bott, J, Carroll, MP, Conway, JH, et al. Randomised

controlled trial of nasal ventilation in acute

ventilatory failure due to chronic obstructive airways

disease. Lancet 1993; 341: 1555-7.

13. Kramer, N, Meyer, TJ, Meharg, J, et al. Randomized,

prospective trial of noninvasive positive pressure

ventilation in acute respiratory failure. Am J Respir Crit

Care Med 1995; 151: 1799-806.

14. Lofaso, F, Brochard, L, Hang, T, et al. Home versus

intensive care pressure support devices. Experimental

and clinical comparison. Am J Respir Crit Care Med

1996; 153: 1591-9.

15. Bunburaphong, T, Imanaka, H, Nishimura, M,et al.

Performance characteristics of bilevel pressure

ventilators: a lung model study. Chest 1997; 111:

1050-160.

16. Appendini, L, Patessio, A, Zanaboni, S, et al.

Physiologic effects of positive end-expiratory pressure

and mask pressure support during exacerbations of

chronic obstructive pulmonary disease. Am J Respir

Crit Care Med 1994; 149:1069-76.

17. Ferguson, GT, Gilmartin, M. C02 rebreathing during

BiPAP ventilatory assistance. Am J Respir Crit Care Med

1995; 151: 1126-35.

18. Girault, C, Richard, JC, Chevron, V, et al. Comparative

physiologic effects of noninvasive assist-control and

pressure support ventilation in acute hypercapnic

respiratory failure. Chest 1997; 111:1639-48.

19. Wysocki, M, Trie, L, Wolff, MA, et al. Noninvasive

pressure support ventilation in patients with acute

respiratory failure. A randomized comparison with

conventional therapy. Chest 1995; 107: 761-8.

20. Aslanian, P, El Atrous, S, Isabey, D, et al. Effects of

flow triggering on breathing effort during partial

ventilatory support. Am J Respir Crit Care Med 1998;

157:135-43.

21. Bernstein, G, Heldt, GP, Mannino, FL Synchronous

mechanical ventilation of neonates. Crit Care Med

1993; 21: 1984-5.

22. Bonmarchand, G, Chevron, V, Chopin, C, et al.

Increased initial flow rate reduces inspiratory work of

breathing during pressure support ventilation in

patients with exacerbation of chronic obstructive

pulmonary disease. Intensive Care Med 1996; 22:1147-54.

23. Grasso, S, Puntillo, F, Mascia, L, et al. Compensation for

increase in respiratory workload during mechanical

ventilation. Pressure-support versus proportional-assist

ventilation. Am J Respir Crit Care Med 2000; 161: 819-26.

24. Younes, M. Proportional assist ventilation, a new

approach to ventilatory support. Theory. Am Rev

Respir Dis 1992; 145: 114-20.

25. Younes, M, Puddy, A, Roberts, D, et al. Proportional

assist ventilation. Results of an initial clinical trial. Am

Rev Respir Dis 1992; 145:121-9.


26. Meecham Jones, DJ, Paul, EA, Grahame-Clarke, C,

Wedzicha, JA. Nasal ventilation in acute exacerbations

of chronic obstructive pulmonary disease: effect of

ventilator mode on arterial blood-gas tensions. Thorax

1994; 49: 1222-4.

27. Vitacca, M, Rubini, F, Foglio, K, et al. Non-invasive

modalities of positive pressure ventilation improve

the outcome of acute exacerbations in COLD patients.


28. Barbe, F, Togores, B, Rubi, M, et al. Noninvasive

ventilatory support does not facilitate recovery from

acute respiratory failure in chronic obstructive

pulmonary disease. Ear RespirJ 1996; 9: 1240-5.

29. Plant, PK, Owen, JL, Elliott, MW. Early use of non-

invasive ventilation for acute exacerbations of chronic

obstructive pulmonary disease on general respiratory

wards: a multicentre randomised controlled trial.

Lancet 2000; 355: 1931-5.

30. Ambrosino, N, Foglio, K, Rubini, F, et al. Non-invasive

mechanical ventilation in acute respiratory failure

due to chronic obstructive pulmonary disease:

correlates for success. Thorax 1995; 50: 755-7.

31. Soo Hoo, GW, Santiago, S, Williams, AJ. Nasal

mechanical ventilation for hypercapnic respiratory

failure in chronic obstructive pulmonary disease:

determinants of success and failure. Crit Care Med

1994; 22: 1253-61.

32. Wood, KA, Lewis, L, Von Harz, B, Kollef, MH. The use

of noninvasive positive pressure ventilation in the

emergency department: results of a randomized

clinical trial. Chest 1998; 113:1339-46.

33. Nava, S, Ambrosino, N, Clini, E, et al. Noninvasive

mechanical ventilation in the weaning of patients

with respiratory failure due to chronic obstructive

pulmonary disease. A randomized, controlled trial.

Ann Intern Med 1998; 128: 721-8.

34. Doherty, MJ, Greenstone, MA. Survey of non-invasive

ventilation (NIPPV) in patients with acute

exacerbations of chronic obstructive pulmonary disease

(COPD) in the UK. Thorax 1998; 53: 863-6.

35. Carlucci, A, Richard, JC, Wysocki, M, et al. French

multicenter survey: non-invasive versus conventional

mechanical ventilation. Am J Respir Crit Care Med

1999;159:A367.

36. Antonelli, M, Conti, G, Riccioni, L, Meduri, GU.

Noninvasive positive-pressure ventilation via face

mask during bronchoscopy with BAL in high-risk

hypoxemic patients. Chest 1996; 110: 724-8.

37. Rasanen, J, Heikkila, J, Downs, J, et al. Continuous

positive airway pressure by face mask in acute

cardiogenic pulmonary edema. Am J Cardiol 1985;

55: 296-300.

38. Bersten, AD, Holt, AW, Vedig, AE, et al. Treatment of

severe cardiogenic pulmonary edema with continuous

positive airway pressure delivered by face mask.

N EnglJ Med 1991; 325:1825-30.

39. Foglio, C, Vitacca, M, Quadri, A, et al. Acute

exacerbations in severe COLD patients. Treatment

using positive pressure ventilation by nasal mask.

Chest 1992; 101: 1533-8.

40. Hilbert, G, Gruson, D, Gbikpi-Benissan, G, Cardinaud,

JP. Sequential use of noninvasive pressure support

ventilation for acute exacerbations of COPD. Intensive

Care Med 1997; 23: 955-61.

41. Meduri, GU, Turner, RE, Abou-Shala, N, et al.

Noninvasive positive pressure ventilation via face

mask. First-line intervention in patients with acute

hypercapnic and hypoxemic respiratory failure. Chest

1996; 109: 179-93.

42. Confalonieri, M, Potena, A, Carbone, G, et al. Acute

respiratory failure in patients with severe

community-acquired pneumonia. A prospective

randomized evaluation of noninvasive ventilation. Am

J Respir Crit Care Med 1999; 160: 1585-91.

43. Martin, TJ, Hovis, JD, Costantino, JP, et al. A

randomized, prospective evaluation of noninvasive

ventilation for acute respiratory failure. Am J Respir

Crit Care Med 2000; 161: 807-13.

44. Bardi, G, Pierotello, R, Desideri, M, et al. Nasal

ventilation in COPD exacerbations: early and late

results of a prospective, controlled study. Ear RespirJ

2000;15:98-104.

45. Confalonieri, M, Parigi, P, Scartabellati, A, et al.

Noninvasive mechanical ventilation improves the

immediate and long-term outcome of COPD patients

with acute respiratory failure. Eur RespirJ 1996; 9:

422-30.

46. Antonelli, M, Conti, G, Rocco, M, et al. A comparison

of noninvasive positive-pressure ventilation and

conventional mechanical ventilation in patients

with acute respiratory failure. N Engl J Med 1998;

339: 429-35.

47. Antonelli, M, Conti, G, Bufi, M, et al. Noninvasive

ventilation for treatment of acute respiratory

failure in patients undergoing solid organ

transplantation: a randomized trial. JAMA 2000; 283:

235-41.48. Lin, M, Yang, YF, Chiang, HT, et al. Reappraisal of

continuous positive airway pressure therapy in

acute cardiogenic pulmonary edema. Short-term

References 69

results and long-term follow-up. Chest 1995; 107:1379-86.

49. Rusterholtz, T, Kempf, J, Berton, C, et al. Noninvasive

pressure support ventilation (NIPSV) with face mask in

patients with acute cardiogenic pulmonary edema.


50. Mehta, S, Jay, GD, Woolard, RH, et al. Randomized,

prospective trial of bilevel versus continuous positive

airway pressure in acute pulmonary edema. Crit Care

Med 1997; 25: 620-8.

51. Aguilo, R, Togores, B, Rons, S, et al. Noninvasive

ventilatory support after lung resectional surgery.

C/?«M997; 112: 117-21.

52. Gust, R, Gottschalk, A, Schmidt, H, et al. Effects of

continuous (CPAP) and bi-level positive airway

pressure (BiPAP) on extravascular lung water after

extubation of the trachea in patients following

coronary artery bypass grafting. Intensive Care Med

1996; 22: 1345-50.

53. Joris, JL, Sottiaux, TM, Chiche, JD, et al. Effect of

bi-level positive airway pressure (BiPAP) nasal

ventilation on the postoperative pulmonary restrictive

syndrome in obese patients undergoing gastroplasty.

CfcesM997; 111: 665-70.

54. Jousela, I, Rasanen, J, Verkkala, K, et al. Continuous

positive airway pressure by mask in patients after coro-

nary surgery. Acta Anaesthesiol Scand 1994; 38: 311-16.

55. Matte, P, Jacquet, L, Van Dyck, M, Goenen, M. Effects

of conventional physiotherapy, continuous positive

airway pressure and non-invasive ventilatory support

with bilevel positive airway pressure after coronary

artery bypass grafting. Acta Anaesthesiol Scand 2000;

44:75-81.

56. Pinilla, JC, Oleniuk, FH, Tan, L, et al. Use of a nasal

continuous positive airway pressure mask in the

treatment of postoperative atelectasis in aortocoro-

nary bypass surgery. Crit Care Med 1990; 18: 836-40.

57. Benhamou, D, Girault, C, Faure, C, et al. Nasal mask

ventilation in acute respiratory failure. Experience in

elderly patients. Chest 1992; 102: 912-17.

58. Meduri, GU, Fox, RC, Abou-Shala, N, et al. Noninvasive

mechanical ventilation via face mask in patients with

acute respiratory failure who refused endotracheal

intubation. Crit Care Med 1994; 22: 1584-90.

59. Clarke, DE, Vaughan, L, Raffin, TA. Noninvasive

positive pressure ventilation for patients with terminal

respiratory failure: the ethical and economic costs of

delaying the inevitable are too great. Am J Crit Care

1994; 3: 4-5.

60. Esteban, A, Frutos, F, Tobin, MJ, et al. A comparison of

four methods of weaning patients from mechanical

ventilation. Spanish Lung Failure Collaborative Group.

N EnglJ Med 1995; 332: 345-50.

61. Goodenberger, DM, Couser, Jl Jr, May, JJ. Successful

discontinuation of ventilation via tracheostomy by

substitution of nasal positive pressure ventilation.

D?«M992;102: 1277-9.62. Restrick, LJ, Scott, AD, Ward, EM, et al. Nasal

intermittent positive-pressure ventilation in weaning

intubated patients with chronic respiratory disease

from assisted intermittent, positive-pressure

ventilation. Respir Med 1993; 87: 199-204.

63. Udwadia, ZF, Santis, GK, Steven, MH, Simonds, AK.

Nasal ventilation to facilitate weaning in patients

with chronic respiratory insufficiency. Thorax 1992;

47:715-18.

64. Torres, A, Gatell, JM, Aznar, E, et al. Re-intubation

increases the risk of nosocomial pneumonia in

patients needing mechanical ventilation. Am J Respir

Crit Care Med 1995; 152: 137-41.

65. Hilbert, G, Gruson, D, Portel, L, et al. Noninvasive

pressure support ventilation in COPD patients with

postextubation hypercapnic respiratory insufficiency.

Eur Respir J 1998; 11: 1349-53.

6Contemporary issues in critical carephysiotherapyMICHAEL BARKER, SHERIC G ELLUM AND SARAH EJ KEILTY

Introduction

Techniques for avoiding intubation

Tracheostomy and decannulation

70

70

73

Rehabilitation

References

75

78

INTRODUCTION

Physiotherapists have a crucial role to play in thetreatment of critically ill patients. Over the last 30years, many innovations have been introduced thathave influenced the approach to patient manage-ment. At the same time, the case mix of patientsseen on the modern intensive care unit (ICU) andhigh dependency unit (HDU) has changed.Physiotherapy has responded to these changes andhas become more evidence based. Traditionalphysiotherapy practices on the ICU have beenchallenged1-4 and this has changed the therapyprovided by the critical care physiotherapy ser-vices.

This chapter describes the model of physiother-apy practice in our institution. The authors appre-ciate that structural considerations, e.g. staffing,skill mix, traditional barriers etc., in other hospi-tals may cause difficulties in the wholesale applica-tion of some of the recommendations made. Theintention is to focus on the important issues thathave lead to a successful implementation of thisphysiotherapy model.

Implicit in the following discussion is the essentialneed for holistic assessment to direct the most appro-

priate treatments. These treatments are groupedtogether as techniques for preventing intubation,those concerned with tracheostomy management,techniques for secretion clearance and those associ-ated with the rehabilitation needs of the criticallyill patient.

TECHNIQUES FOR AVOIDING INTUBATION

Intubation and mechanical ventilation are usedto provide ventilatory control and to protect theupper airway during acute severe illness. Positivepressure ventilation via an endotracheal ortracheostomy tube is standard practice. Theindications are well documented and the majority ofpatients wean from ventilation on recovery from theillness prompting ICU admission. However,complications in intubated patients in ICU arerelatively common and weaning may be difficult. Insome cases, avoidance of intubation by the use ofnon-invasive ventilatory assistance (NIV) maytherefore be appropriate. The physiotherapist will,however, aim to identify reversible factors leading toventilatory failure in this type of patient and act toprevent deterioration by improving the balance

Techniques for avoiding intubation 71

between the respiratory load and capacity of theventilatory pump. Strategies include improvingairway secretion clearance, humidification andpatient positioning and techniques of non-invasiverespiratory assistance.

Airway clearance techniques

There is a variety of interventions that can beperformed to facilitate expectoration of retainedpulmonary secretions. The active cycle ofbreathing techniques (ACBT),5 used alone or inconjunction with pressure support such asintermittent positive pressure breathing (IPPB),continuous positive airways pressure (CPAP), canbe used to aid airway clearance. These techniquesmay increase tidal volume,6 improve collateralventilation,6 decrease the work of breathing7 andimprove expectoration.

Humidification of the airway is extremelyimportant because the efficiency of mucus transportis dependent on hydrated mucous membrane andcorrectly functioning cilia. During infection, theviscosity and quantity of mucus are increased.Combined with systemic dehydration and inspiringdry medical gases, this can result in a reduction ofmucociliary escalator function, leading to sputumretention, increased airflow resistance and impairedgas exchange.

Positioning to reduce respiratory load

Traditionally, positioning was seen only as a meansto facilitate bronchial drainage. Studies show thatpositioning can reduce the work of breathing byreducing ventilatory demand.7 The specific posi-tion required and the focus of the treatment willdepend on the pathology and clinical signs. Withunilateral acute lung disease, e.g. consolidation,positioning the patient with the unaffected lungdown has been shown to optimize gas exchange byimproving ventilation perfusion matching. In thepresence of bilateral pulmonary pathology, how-ever, positioning the patient in right-side lying hasshown improvements in PaO2 compared with posi-tioning on the left.7 This may be attributed to theright lung having a greater volume than the leftlung.

Positioning to enhance respiratorymuscle function

The forward lean sitting position improves the cap-acity of the respiratory muscles8 and reduces thesensation of breathlessness9 in patients with chronicairflow limitation (CAL). It has been postulated that,in this position, the diaphragm becomes less flattenedand longer as it assumes the more usual domed shape,enhancing the length-tension relationship.9 If thepatient is too exhausted to tolerate this position, it canbe modified to high side lying, with similar results.

Positive pressure techniques

Non-invasive respiratory assistance has been used sincethe 1930s. Advances in the technology of ventilators, orflow generators, and interface design have revolution-ized the use of non-invasive positive pressure respira-tory support. It can be applied to non-intubatedpatients by a tight-fitting full facemask or nasal maskand by a mouthpiece. It can be applied continuously(CPAP), intermittently with inspiratory assistancealone (PSB) and by combining the two provides bothinspiratory positive airways pressure (IPAP) and expira-tory positive airways pressure (EPAP) or bi-level posi-tive airways pressure. Patients who do not improvewith NIV or who deteriorate may require intubation. Astrategic plan in the event of failure of NIV is import-ant at the start of therapy (see Chapter 5).

Continuous positive airways pressure

CPAP was first described in the 1930s for the treat-ment of extravascular lung fluid. It produces animprovement in oxygenation and in pulmonarycompliance and an increase in lung volume.7 Indoing so, CPAP can reduce the work of breathing,10

has been used successfully to treat acute respiratorydistress syndrome (ARDS) in adults11 and may pre-vent re-intubation in selected patients.12

Intermittent positive pressurebreathing

IPPB is patient-triggered inspiratory pressure sup-port and was first described in 1947 . Since then, formal

72 Contemporary issues in critical care physiotherapy

studies have provided conflicting results. Physiologicalstudies have demonstrated that IPPB can bringabout a rise in PaO2 and a reduction in PaCO2,

13

decrease the work of breathing and improve minuteventilation.14 IPPB, or pressure support breathing, isuseful in the treatment of the exhausted patient whohas shallow breathing resulting in hypercapnia. It isimportant that the patient is awake and co-operative.The technique can be applied non-invasively bymouthpiece or mask, but this has to be held inplace by the operator and is rarely tolerated bypatients for protracted periods of time. This meansthat IPPB is usually limited to physiotherapytreatment sessions.

Non-invasive ventilation

If patients require longer periods of pressure sup-port, NIV may be effective. Originally, this treat-ment was used for patients with chronic respiratoryfailure due to chest-wall disease. There has beenconsiderable improvement in ventilator and inter-face design and NIV is now used in the treatment ofa range of causes of acute respiratory failure (for afull review of NIV in acute respiratory failure, seeChapter 5). The broadening of the application ofNIV has moved its use from the ICU to the HDU,postoperative recovery and specialist wards and tothe accident and emergency department, creating arequirement for a 24-hour NIV service. Althoughideal, this is not always possible because to havesuitably trained staff available 24 hours a dayrequires considerable organization. Significantfinancial resources are required to fund disposableequipment (masks, circuits etc.) and provide staffand on-going training.

PRACTICAL ASPECTS OF NON-INVASIVE

VENTILATION

Many patients tolerate NIV well. Intolerance mayrelate to mask discomfort, feelings of claustrophobiaand asynchrony with the ventilator. It is importantthat the patient has confidence in the operator andshould feel in control of the situation rather thancontrolled by the situation. A variety of interfacesis available, which have their own advantages anddisadvantages. Many manufacturers producemeasuring gauges to facilitate fit, but it may be

necessary to try several before the best fit can beachieved. Overall, the nasal mask is preferred as itreduces the feeling of claustrophobia and allowscommunication with the clinician. The maindisadvantage is leak from the mouth, made worseduring sleep, and only partially helped by theuse of a chinstrap. Full facemasks are usuallynecessary for extremely dyspnoeic patients. Theyare, however, claustrophobic and present a risk ofaspiration if the patient vomits. Masks may causeskin necrosis, and a strip of foam dressing, placedover the bridge of the nose, may prevent or delaythis happening. This is important because NIV willbe poorly tolerated if painful. Nasal pillows maymaintain ventilatory support while alleviatingpressure over the bridge of the nose. They are,however, not a suitable option in severe acuterespiratory distress.

Ventilator settings and setting apatient up on non-invasive ventilation

Initially, we set a low pressure, e.g. 10-15 cmH2O,and demonstrate the flow of air on the back of thehand for reassurance. We then attach the tubing tothe mask and hold it to the patient's face for a fewbreaths, allowing a few triggered breaths.Encouragement should be given to breathe slowlyand the spontaneous mode of the ventilator willtrigger accordingly. If the respiratory rate is lessthan 8 breaths/min, or the patient is too tired orweak to trigger breaths, set the ventilator to thespontaneous/timed mode so that machine-initiatedbreaths will occur in the event of hypoventilation.If supplemental O2 is required, it may beprecautionary to set the spontaneous/timed modein case of O2-induced apnoea. Once NIV isestablished, the head strap can be attached andgently tightened. Over-tightening will exacerbateskin necrosis. Large air leaks should be avoided,especially around the eyes. Once the patient issettled, the operator should watch the chestmovement and observe the respiratory pattern. Wethen slowly increase the pressure to a comfortablelevel, e.g. 15-25 cmH2O, and reduce the respiratoryrate so that a more efficient breathing pattern isestablished. It is essential that the operator stayswith the patient for the first 30 min so thatencouragement and re-assurance can be given.

Tracheostomy and decannulation 73

Monitoring patients on non-invasiveventilation

Monitoring should include clinical assessment com-bined with physiological variables such as O2 satura-tion, blood gases, blood pressure, heart rate andrespiratory rate. Arterial blood gases should be meas-ured following 30-40 min of ventilation and ventila-tor settings should be changed with the aim ofincreasing pH to more than 7.3. Patients should useNIV for as long as possible on the first day oftherapy and especially during sleep. As gas exchangeimproves, patients can spend increasingly longerperiods off the ventilator to allow for airway clear-ance, mouth care, nutrition and mobility. To assessthe need for nocturnal ventilatory support, a trial ofovernight oximetry or capnography without ventila-tory support may be appropriate before the patient isdischarged to the general ward.

NEGATIVE PRESSURE TECHNIQUES

Negative pressure ventilatory support has become lessfrequently used since the re-emergence of positivepressure ventilation. It may be an option for patientswho fail to adapt to positive pressure ventilation. Thereare two main types of negative pressure devices avail-able: the 'tank' ventilator in which the whole body,except the head and neck, is enclosed; and the cuirassshell, which encloses the anterior thorax and abdomenonly. Sub-atmospheric pressure is intermittentlyapplied to the chest wall and abdomen by a pumpattached to the chamber and air is drawn into the tho-rax as the transpulmonary pressure is increased.

The efficiency of negative pressure ventilation isdetermined by the compliance of the chest wall, air-way resistance and the ability to create an airtightseal. It has several limitations. Access is poor with thetank respirator, and synchrony with the patient'sbreathing pattern can be a problem. The most ser-ious problem with negative pressure ventilation isthe tendency to induce upper-airway obstruction,although triggered ventilators are reportedly betterin this respect. Alternatively, upper-airway obstruc-tion can be overcome by applying positive pressureby nasal mask and we have employed this approachwhen intubation was refused or inappropriate forother reasons.

Cuirass ventilation, although of value in domicil-iary ventilatory support, is insufficient to assist ven-

tilation in acute CAL respiratory failure. Cuirass venti-lation and provision of novel therapy techniques are,however, possible with the Hyeck or RTX (Medivent,UK) machines. These devices have been shown toimprove gas exchange in patients with CAL andother causes of respiratory failure, but more clinicaltrials are needed. A physiotherapy mode using oscil-lations in pressure up to 600 cycles min-1 may facilit-ate secretion removal. This also warrants furtherinvestigation.

SUPPORTING PATIENTS POST-EXTUBATION

Not all patients manage spontaneous breathing inthe first few hours post-extubation. Patients at risk ofdeterioration or who require ventilatory assistancemay benefit from NIV. Patients who fail spontaneousbreathing trials during assessment for extubationshould also be considered for early extubation andNIV as an alternative to tracheostomy.

In the slowly weaning patient, who will usuallyhave a tracheostomy, ambulation can be possibleeither by bagging or using portable ventilatory assistdevices such as the Respironics 'Bi-PAP HarmonyST' (Medicaid, UK) or the 'Voyager' (B&D MedicalUK). These can be battery powered and offer trig-gered and non-triggered ventilatory support modes.They are light and compact devices, which can becarried in a shoulder bag by the therapist or patient.We believe that providing ambulatory respiratorysupport facilitates rehabilitation.

TRACHEOSTOMY AND DECANNULATION

Tracheostomy is indicated when prolonged ventila-tory support is required or the patient is unable toprotect his or her airway. There is a variety of tra-cheostomy tubes available, such as single and doublelumen, cuffed and uncuffed, and fenestrated andunfenestrated tubes. Cuffs are most commonly usedto provide airway protection from aspiration, whichis more likely if the cuff pressure is too low. Too higha pressure, on the other hand, will result in mucosalischaemia. Cuff pressure should therefore be moni-tored.15 A cuff pressure between 15 and 20 cmH2O isrecommended to prevent complications. Single-lumen cuffed tubes are often the initial choice of tra-cheostomy for economic reasons. The single lumenmaximizes the size of the internal diameter for a


given external diameter, so reducing resistive load.This resistive load may be of significance duringweaning from mechanical ventilation.

Double-lumen tubes have a removable inner tubeto prevent the build-up of secretions that may other-wise narrow the internal diameter. In severe situ-ations, this could lead to complete obstruction of thetube. In the extreme example of complete tubeobstruction, removal of the inner tube rapidly cor-rects the emergency, whereas the whole tracheostomytube would need to be removed in the case of asingle-lumen tube. This is perhaps of more relevancein the HDU, where the nurse/patient ratio is lower.A fenestration in the outer double-lumen tube ismatched to similar holes in the 'speaking' inner tubeto allow air up through the vocal cords. Cuff defla-tion is often required as well to provide an effectivevoice. This aspect is important during the weaningprocess. It is obviously important that the correct, i.e.unfenestrated, inner tube is inserted during periodsof controlled ventilation.

A small, cuffless catheter, or mini-tracheostomy,can be inserted when only secretion management isrequired. It has an internal diameter large enough topass a size 10-FG suction catheter. Mini-tracheostomymay be favoured as a 'step down' for patients who canbreathe spontaneously but require access to aid theclearance of bronchial secretions. A spigot isremoved to allow catheter access to clear secretions.Although designed for secretion clearance, its use toprovide ventilatory support has been described usinginspiratory pressures >50 cmH2O. Intratrachealpressure is, of course, much lower, and surgicalemphysema has not been a problem when we haveused this approach to avoid formal intubation.Cuffless tubes are rarely used in normal patient care,but have advantages for patients who require long-term tracheostomy and ventilatory support.

Cuff deflation

Once the need for ventilatory support is reduced to aminimum, the primary function of the tracheostomychanges from a means of providing ventilatory sup-port to that of clearance and management ofbronchial secretions. Patients breathing spontan-eously via a tracheostomy may be managed on theHDU or general ward. When considering removal ofthe tube, it is important to ensure the patient has an

intact swallow. We employ blue dye added to drink-ing to water. Aspiration can be clearly seen by theappearance of blue dye via the tracheostomy. Whendemonstrated, further assessment by a speech andlanguage therapist is recommended. However, thesignificance of minor aspiration is uncertain andswallow often only improves when practised with thecuff deflated.

Speech

Establishing effective communication is important.If the patient can tolerate cuff deflation, a unidirec-tional speaking valve provides a good voice yet allowscontinued inspiratory support. Alternatively, fenes-trated tracheostomy tubes may be an advantage if apatient finds the addition of a speaking valve increasesthe work of breathing. It should be noted that, if theleak around the tube within the trachea is minimal,the addition of a speaking valve may lead to distressdue to the development of hyperinflation.

Expectoration and cough

It is important when making the decision todecannulate patients that they are able to clear respira-tory secretions. Generalized muscle weakness,reduced ventilatory capacity and bypassing of thenormal mechanisms of humidification mean the tra-cheostomized patient is vulnerable to retention ofsecretions. An assessment of cough effectivenessshould be undertaken before decannulation isattempted.

A cough consists of four phases. The inspiratoryphase, of volumes varying from 200 to 3500 mLabove functional residual capacity (FRC), is the first,leading to the compressive phase, in which theexpiratory muscles contract against a closed glottis,creating an intrathoracic pressure up to 300 cmH2O.During the explosive expiratory phase, expiratoryflow rates of 6-12 Ls-1 are generated in normaladults. A transient supramaximal cough expiratoryflow, or cough spike, can be recorded during the initialpart of the expiratory phase associated with glotticopening. Although the glottis plays an importantrole in cough generation, its closure is not essential,as demonstrated by the effective expectoration ofsputum by tracheostomized or endotracheal tube-

Rehabilitation 75

intubated patients. Expiratory flow rates are lower,however, than those generated with a closed glottisand, if combined with respiratory muscle weakness,may reduce cough effectiveness. Laryngeal muscleco-ordination plays an important role because it isessential that the vocal cords are fully abducted toallow rapid expulsion of air. Finally, the relaxationphase, with expiratory muscle relaxation and elasticrecoil of the lung, allows a return to normal levels ofpressures within the thorax.

Evaluation of cough

Normal subjects reach 85-90% of their inspiratorycapacity in the inspiratory phase of coughing. Coughmay be ineffective when the vital capacity (VC) is< 1.5 L and will be so if the VC is <850 ml. A PECFof 160 L min"1 may be the minimum value associat-ed with the ability to clear the airways of sputum.16

This can be measured using a hand-held peak expira-tory flow meter or, if available, a pneumotachograph,which is more accurate. Measurement is taken using amouthpiece with the tracheostomy cuff deflated andthe trachesotomy capped off. The testing of res-piratory muscle strength on the ICU is fully consid-ered in Chapter 2. Although non-volitional tests mayat times be necessary, the invasive nature of these testsmakes their use in spontaneously breathing, alertpatients difficult. Volitional tests usually provide areasonable reflection of respiratory muscle strength.Maximal inspiratory muscle strength (MIP), sniffnasal inspiratory pressure (SNIP) and maximalexpiratory muscle strength (MEP) have all beenshown to be useful when assessing inspiratory orexpiratory muscle function. As with all volitional tests,a high value can exclude respiratory muscle weakness,although a low value is more difficult to interpret.

Cough augmentation

A number of techniques can be employed to improvean ineffective cough.

MECHANICAL IN/EXSUFFLATOR

(CoughAssist™ MI-E)

The mechanical in/exsufflator (Emerson, MA,USA) assists the clearance of bronchopulmonary

secretions. Application of positive pressure, viaa facemask, is quickly changed to negative pressuremanually or by time cycling to generate a highexpiratory flow rate. Positive inflation pressuresusing 25 cmH2O and subsequent cycling to anegative pressure of —30 cmH2O are suggestedstarting pressures.17 Although well established inthe USA, the MIE is still largely unused in the UKand Europe. The evidence to support its use hasbeen confined to patients with neuromuscularcauses, but these data suggest that MIE canincrease PECF rates during unassisted coughingfrom 1.81 Ls-1 to 7.47 Ls-1.18 Preliminaryexperience at this centre of improved radio-tracerclearance, compared with standard techniques,suggests that MIE may provide a more effectivealternative to cough augmentation in patients withrespiratory muscle weakness. It may also be avaluable asset for a broader spectrum of patientsacross the critical care environment.

MANUALLY ASSISTED COUGH

Manually assisting cough but co-ordinating abdom-inal pressure by the therapist with patient effort canimprove the clearance of respiratory secretions. It ismost commonly used for patients with long-termneuromuscular weakness, such as spinal lesions ormotoneuron disease. This technique may be benefi-cial for patients in the critical care environment withunderlying restrictive lung disease or respiratorymuscle weakness, but requires further study.

REHABILITATION

The negative sequelae associated with immobility arewell documented19 The physiotherapist will aim toimprove functional outcomes by the maintenance ofjoint range and soft-tissue length and increasingstrength and endurance by using passive movements,positioning regimens and active therapy. The sameexercise principles that apply to training healthyindividuals for strength, endurance or power areused in the recovery from critical illness. The conceptis of sport for the sick. It is useful to considerrehabilitation strategies as those concerned with'in-bed rehabilitation' and those concerned with'out-of-bed rehabilitation'.


In-bed rehabilitation strategies

POSITIONING

Positioning is important. Of all the positions used,prone positioning has recently received a lot ofattention.20 It is primarily employed in refractoryhypoxaemia in the patient with acute lung injury.Approximately 70% of patients respond with signi-ficant improvement in PaO2:FiO2 ratio, although, insome, the benefit is not maintained. Since it was firstdescribed in 1977, there have been a number ofreports and editorials that have mostly explored themechanisms behind the response. Few have citedcomplications associated with turning the criticallyill patients prone. Some of these complications arerelatively minor, such as the loss of central venousaccess or dependent oedema, and others are moreimportant, such as extubation or cardiovascularinstability. Chronic complications such as contrac-tures21 and myositis ossificans22 are reported.Guidelines have been published on the managementof proned patients,20 with a focus on the protectionof neural structures, such as the brachial plexus andcommon fibular nerves, in tension-relieving pos-tures. Typical postures to protect the brachial plexusare shown in Figure 6.1. Normally, the illustratedpositions should be altered once every 4 hours.

CONTINUOUS LATERAL ROTATION THERAPY

Continuous lateral rotation therapy (CLRT) is con-sidered as an 'in-bed' rehabilitation option, becauseit is commonly used in severely unstable patientswho do not tolerate manual positioning and han-dling. A continuous rotation of the patient along thelongitudinal axis is employed in a specialized bed.These beds can be divided into two categories: thosethat use rigid table-based rotation units to 'lock' thepatient in (mostly used in spinal injury), and thosethat have air cushion systems (contraindicated inspinal injury). The purpose is to continually changethe area of lung dependency to prevent alveolar col-lapse and to aid the mobilization of secretions.Pressure relief on the skin is also achieved. Studieshave shown that changing the area of lung depend-ency by CLRT improves respiratory function.23 Themechanisms by which this is achieved are thought tobe a combination of increased mobilization of secre-tions and redistribution of inspired gas to previouslycollapsed areas of lung.

Controversies surround the use of CLRT and we arenot aware of any cost-benefit analyses. Clinical trialsof CLRT have suffered from difficulties recruiting suf-ficient patient numbers to show statistically significantresults. A meta-analysis of six trials24 revealed thatCLRT was associated with a decreased incidence ofatelectasis and nosocomial pneumonia as well as witha decrease in the ventilation time and intensive carestay. Some patient groups appear to benefit more thanothers, such as those with sepsis or blunt trauma. Allthe studies included in the meta-analysis have beencarried out in the USA and there are argumentsagainst generalizing these findings. For example,ARDS patients were generally excluded; the practice ofrespiratory therapy is different in the USA from that inother countries; and the benefit for spinal-injurypatients without lung injury may not be transferableto general ICU patients. The optimal rotation settings(depth of rotation, pause time and number of rota-tions per 24 hours) have also not been established.

ACTIVE MOVEMENTS

If the patient cannot be mobilized out of bed, activemovements can still be effectively used when he or sheis sufficiently awake to participate. The physiothera-pist's primary concern at this stage would be therecruitment and maintenance of anti-gravity musclegroups (quadriceps, gluteal, abdominal and erectorspinae muscles), but the triceps, biceps and handmuscles should also be targeted if possible. Activemovements should be performed to mimic activefunctional movement patterns, e.g. extension of thelower limb would be performed as a composite move-ment involving extension at the hip, knee and ankle.

Out-of-bed rehabilitation strategies

The patient should be mobilized out of bed as soonas the opportunity arises. Getting a patient out ofbed may seem an easy task, but there are manyassessment issues that have to be taken intoaccount to execute the task with minimal risk andoptimal efficiency. Even when sitting the patient onthe edge of the bed - a commonly used techniqueto initiate the 'out-of-bed' rehabilitation phase -careful assessment must be conducted. Con-siderations should be given to the respiratory capacity,haemodynamic stability, muscle strength and

Rehabilitation 77

Figure 6.1 (a) The neck is rotated to the right with the left upper limb in flexion and internal rotation. The right upper limb is placed at

the patient's side with the shoulder internally rotated and elbow extended, (b) The mirror image of (a). The neck is rotated to the opposite

direction and the upper limbs positioned accordingly.

neurological co-ordination. The patient's ability toassist in rehabilitation is crucial because this influencesthe use of moving and handling devices. The methodsemployed include the following.

• Tilt table. The patient is transferred laterally ontoa rigid plinth with footrests and is secured in theanatomical position with straps. The patient isgradually tilted from the horizontal to the verti-cal/standing position.

• Sitting over the edge of the bed with support,standing from sitting and standing transfers to achair (via pivoting or stepping around).

• Ambulatory ventilation.

AMBULATORY VENTILATION

Drawing on the use of NIV during exercise inCAL,25 we have treated similarly deconditionedpatients in the ICU. This progressive approach tomobilizing ICU patients is not widely practised and

outcome parameters have not been established. Webelieve that an exercise regimen positively influencesthe speed of weaning and decreases ICU stay and cost.

Challenges to 'conventional' chestphysiotherapy techniques

PERCUSSION

The use of chest-wall percussion in critical care iswaning, due to critiques on its efficacy in acute care.1

However, percussion may have value in suppurative,chronic lung diseases such as bronchiectasis and cys-tic fibrosis.26

MANUAL HYPERINFLATION

Manual hyperinflation is still commonly used tomobilize retained pulmonary secretions and torecruit collapsed lung. It is effective in treating acute


lobar collapse,27 but its value for patients with lunginjury must be questioned. The current ventilationstrategy in lung injury is to protect the lung fromfurther trauma by using low-volume ventilation withhigh positive end-expiratory pressure (PEEP).Disconnecting the patient to manually hyperinflatethe lungs thus appears irrational.

HIGH-FREQUENCY OSCILLATION VENTILATION

High-frequency oscillation ventilation (HFOV) is astrategy employed in refractory hypoxaemia usinghigh mean airway pressures with sub-dead-space tidalvolumes. A diaphragm that oscillates at 5-6 cycles s-1is used, cycling on top of a maintained CPAP. Thephysiotherapist is faced with a patient who oughtnot to be disconnected for chest therapy becausede-recruitment and worsening hypoxaemia will ensue.Regular suctioning, even with in-line catheters, cansimilarly de-recruit the lung. The chest is also visiblyoscillating, which precludes manual physiotherapytechniques. Despite this, therapeutic managementincludes optimal patient positioning and modifyingsuctioning to only when necessary. (As auscultation isof limited value in HFOV, other parameters are usedsuch as a gradual increase in the proximal operatingpressure of the oscillator, known as the 'delta P').

REFERENCES

1. Dean, E. Oxygen transport: a physiologically-based

conceptual framework for the practice of

cardio-pulmonary physiotherapy. Physiotherapy 1994;

80(6): 347-53.

2. Jenkins, SC, Soutar, S, Loukota, JM, Johnson, LC,

Moxham, J. Physiotherapy after coronary artery

surgery: are breathing exercises necessary? Thorax

1989; 44: 634-9.

3. Eales, CJ, Barker, M, Cubberley, NJ. Evaluation of a

single chest physiotherapy treatment to

post-operative, mechanically ventilated cardiac surgery

patients. Physiother Theory Practice 1995; 11: 23-8.

4. Barker, M, Eales, CJ. The effects of manual

hyperinflation using self-inflating manual

resuscitation bags on arterial oxygen tensions and

lung compliance - a meta-analysis of the literature.

SAfrJ Physiother 2000; 56(1): 7-16.

5. Pryor, JA, Webber, BA, Hodson, ME, Batten, JC.

Evaluation of the forced expiration technique as an

adjunct to postural drainage in the treatment of cystic

fibrosis. 6M/1979; 2: 417-18.

6. Anderson, JB, Qvist, j, Kann, T. Recruiting collapse lung

through collateral channels with positive expiratory

pressure. Scand J Respir Dis 1979; 60: 260-6.

7. Dean, E. Effect of body positioning on pulmonary

function. Phys Therl985; 65: 613-18.

8. O'Neil, S, McCarthy, DS. Postural relief of dyspnoea

in severe chronic airflow limitation: relationship to

respiratory muscle strength. Thorax 1983; 38: 595-600.

9. Sharpe, JT, Drutz, WS, Moisan, T. Postural relief of

dyspnoea in severe chronic obstructive pulmonary

disease. Am Rev Respir Dis 1980; 122: 201-11.

10. Gherini, S, Peters, RM, Viirgilio, RW. Mechanical work

of breathing with positive end expiratory pressure and

continuous positive airways pressure. Chest 1979; 76:251-6.

11. Greenbaum, DM, Millen, JE, Eross, B, Snyder, JV,

Grenvic, A, Safar, P. Continuous positive airways

pressure without tracheal intubation in spontaneously

breathing adults. Chest 1976; 69: 615-20.

12. Dehaven, CB, Hurst, JM, Branson, RD. Post extubation

hypoxaemia treated with continuous positive airways

pressure mask. Crit Care Med 1985; 13: 46-8.

13. Torres, G, Lyons, HA, Emerson, P. The effects of

intermittent positive pressure breathing on

intra-pulmonary distribution of inspired air. Am J Med

1960; 29: 946-54.

14. Motley, HL, Werko, L, Cournand, A, Richards, DW.

Observations of the clinical use of intermittent

positive pressure. J Aviation Med 1947; 18: 417-35.

15. Barry, BN, Bodenham, AR. Airway management in the

ICU. BrJ Intensive Care 2000; 10: 22-9.

16. Bach, JR, Saporito, LR. Criteria for extubation and

tracheostomy tube removal for patients with ventilatory

failure. A different approach to weaning. Chest 1996:

110:1566-71.

17. Whitney, J, Harden, BA, Keilty, SEJ. Assisted cough. A

new technique? Physiotherapy 2001; 68: 201-7.

18. Bach, JR. Mechanical insufflation-exsufflation.

Comparison of peak expiratory flows with manually

assisted and unassisted coughing techniques. Chest

1993; 104: 1553-62.

19. Yasuda, K, Hayashi, K. Changes in biomechanical

properties of tendons and ligaments from joint disuse.

Osteoarthritis Cartilage 1999; 7(1): 122-9.

20. Barker, M, Beale, R. Optimal positioning for the adult

intensive care patient while prone. In Yearbook of

intensive care and emergency medicine, ed. J-L Vincent.

Berlin: Springer-Verlag, 2000; 256-62.

References 79

21. Fridich, P, Krafft, P, Hochleuthner, H, Mauritz, W.The effects of long-term prone positioning inpatients with trauma induced adult respiratorydistress syndrome. Anesth Analg 1996;83: 1206-11.

22. Willems, MCM, Voets, AJ, Welten, RJTJ. Twounusual complications of prone-dependency insevere ARDS. Intensive Care Med 1998;24:276-81.

23. DeBoisblanc, BP, Castro, M, Everret, B, Grender,J, Walker, CD, Summer, WR. Effect of air supportedcontinuous postural oscillation on the risk of early ICUpneumonia in non traumatic critical illness. Chest1993; 103: 1543-7.

24. Choi, SC, Nelson, LD. Kinetic therapy in critically illpatients: combined results based on meta-analysis.J Crit Care 1992; 7(1): 57-62.

25. Keilty, SEJ, Ponte, J, Fleming, TA, Moxham, J. Effectof inspiratory pressure support on exercise tolerance andbreathlessness in patients with severe stable chronicobstructive pulmonary disease. Thorax 1994; 49: 990-4.

26. Gallon, A. Evaluation of chest percussion in thetreatment of patients with copious sputumproduction. Respir Med 1990; 85: 45-51.

27. Stiller, K, Jenkins, S, Grant, R, Geake, T, Taylor, J, Hall,B. Acute lobar atelectasis: a comparison of fivephysiotherapy regimens. Physiother Theory Practice1996; 12: 197-209.

7Diagnostic methods in respiratoryintensive care medicineTORSTEN T BAUER AND ANTON I TORRES

Introduction

Portable chest radiography

Bronchoscopy

80 Bedside diagnostic ultrasound 86

80 Other bedside procedures in the intensive care unit 86

82 References 86

INTRODUCTION

Diagnostic methods in the intensive care unit (ICU)are partly limited by availability. Some methods maybe available within the hospital, e.g. computedtomography (CT) of the chest, but transportation ofthe patient within the hospital may be difficult oreven hazardous. However, some techniques are read-ily available and integrated into the daily routine.The anterior-posterior chest radiograph is routinelyperformed on a daily basis in most ICUs and longi-tudinal comparison may facilitate judgement, e.g.new pulmonary infiltrates. Moreover, bronchoscopycan be easily performed in the intubated patient andmay even be done while respiratory support is pro-vided using non-invasive ventilation.

The purpose of this chapter is to summarize thepotential of the most commonly applied methods forthe diagnosis of respiratory disorders and to outlinebasic practical issues.

PORTABLE CHEST RADIOGRAPHY

Patients usually have multiple medical problems, andportable chest radiography requires careful posi-

tioning of the patient to avoid displacing monitoringdevices and indwelling catheters. Anterior-posteriorchest radiographs are optimally taken with the patentin the upright position and during deep inspiration.This is usually not possible in the intubated patient,for whom supine films will normally be taken. X-rayexposure timed to the ventilator is difficult, but useof the inspiratory hold button ensures deepinspiration. Portable chest radiography remains thecornerstone of imaging techniques in respiratorycritical care medicine, despite its disadvantages. Thevalue of routine examinations is debatable.l

Monitoring of indwelling patientdevices

The correct position of endotracheal tubes should bechecked on the chest radiograph. With flexion andextension of the patient's neck, the endotracheal tubemoves approximately 3 cm upward or downward.The ideal position is therefore 5 cm above the carina,with the head in a neutral position. Although anendotracheal tube inserted too deeply rarely causesinjury, it always extends into the right mainbronchus and impairs ventilation of the left lung,despite the fact that modern tubes have side holes to

Portable chest radiography 81

ensure a degree of ventilation to the left lung. Oneuncommon but serious complication during tubeplacement is rupture of the posterior membranousportion of the distal trachea or proximal mainbronchi. Severe respiratory distress and subcuta-neous emphysema, pneumomediastinum or pneu-mothorax are the symptoms and signs of thiscomplication. Central venous lines and pressuremonitors are frequently used to guide fluid therapyand a chest radiograph should be obtained followinginsertion of a central line catheter to ensue properpositioning. Pressures within the central venous sys-tem are measured most reliably when the catheter tipis located between the right atrium and the mostproximal venous valves. The catheter tip shouldtherefore be visualized medial to the anterior portionof the first rib, at the junction of the brachiocephalicvein and the superior vena cava or within the super-ior vena cava itself. Catheters that are advanced toofar increase the risk of cardiac perforation or induc-tion of ventricular arrhythmias.

Pulmonary capillary wedge catheters are helpfulfor the differentiation of cardiac and non-cardiacpulmonary oedema and optimizing fluid or inotropetherapy. The catheter is introduced via an antecu-bital, jugular or subclavian vein and advancedthrough the right side of the heart into the pul-monary artery. Ideally, the tip should not extendbeyond the proximal interlobar pulmonary arteries.Complications of pulmonary artery cathetersinclude arrhythmias, pulmonary infarction, pul-monary artery perforation, intracardiac knotting,endocarditis and sepsis. Their routine value has beenquestioned by the Pulmonary Artery CatheterConsensus Conference.2 However, this position is notunanimously accepted and a recent meta-analysisfound that substituting a non-invasive studyfor Swan-Ganz catheter placement in the initialevaluation of acutely ill patients may slightly reduceprocedure-related events, but may also increase thenumber of procedures performed.3

Thoracostomy tubes are used to drain air and/orpleural effusions, empyemas and haemothoraces. Thetubes should be positioned in an anterior-superiorlocation for pneumothorax and in a posterior-inferiorlocation to drain liquids. Chest radiographs may beused to verify the position after the procedure, butultrasound is more useful in guiding tube placement,especially with loculated collections. Otherindwelling devices that may be visualized by the

chest radiograph are nasogastric tubes, transvenouspacing wires and intra-aortic counterpulsationdevices.

Pulmonary and pleural abnormalities

Atelectasis occurs most often in the left lower lobe,followed by the right lower and right upper lobes.Despite significant intrapulmonary shunting, radio-graphic appearances may vary from entirely normalto linear, patchy opacities to lobar collapse. Lobarcollapse with positive air bronchograms (the positiveimaging of the main bronchi) should make the clin-ician think of consolidation, whereas lobar collapsein the absence of air bronchograms is usually associ-ated with obstruction, e.g. a mucous plug. Both typesof lobar collapse may prompt bronchoscopy toobtain microbiological samples or to remove theobstruction. In general, radiograph opacities due topneumonia appear later and resolve more slowlythan do those due to aspiration or atelectasis.Therefore, new pulmonary infiltrates that persist for48 hours are more suggestive of nosocomial pneu-monia. However, the radiographic appearance maybe complicated by underlying conditions such aschronic obstructive pulmonary disease or pul-monary oedema.

Because of the effects of gravity in the supinepatient, pulmonary vascular redistribution to theupper zones occurs and pulmonary oedema mayhave an atypical distribution. In addition, the dis-tinction of non-cardiogenic from cardiogenic oedemamay be difficult because the usual radiographic signsof pulmonary venous hypertension, cardiomegaly orpleural effusion may not be present and/or are non-specific. Adult respiratory distress syndrome (ARDS)appears as non-cardiogenic oedema on the chestradiograph. Infiltrates due to ARDS develop rapidlyand bilaterally and differentiation from cardiogenicoedema may require determination of the pul-monary capillary wedge pressure with a pulmonaryartery catheter, by echocardiography or by othermethods (see Chapter 8). Initially, the chest radi-ograph may be normal. Within 24-36 hours, theremay only be little evidence of interstitial perihilaroedema. On days 2-5, the radiographic patternchanges to patchy, ill-defined opacities and then tomore confluent, diffuse, bilateral airspace opacities.Massive pulmonary embolism may mimic these

82 Diagnostic methods in respiratory intensive care medicine

appearances and the diagnosis may be difficult toestablish. Dynamic spiral contrast-enhanced CThas high sensitivity and specificity for proximalpulmonary embolism (see Chapter 19). Findings onthe chest radiograph in lesser pulmonary embolisminclude peripheral, wedge-shaped pleural-basedopacities, atelectasis, elevation of the diaphragmand/or pleural effusion.

Pneumothorax is the most frequently recognizedmanifestation of extra-alveolar air on the chest X-ray(CXR), but pneumomediastinum may also be pres-ent. Mechanical ventilation with positive end-expirat-ory pressure (PEEP) will exacerbate an air leak onceit has developed. In the supine patient, free air with-in the pleural space rises anteriorly and medially. Ifthere are pre-existing adhesions, the pneumothoraxmay remain encapsulated, even when the patient iserect. Diagnosis may require CT scanning becausedifferentiation from intrapulmonary air cysts orabscesses may be difficult and significant pneumo-thorax may not even be apparent on the plain CXRin severe ARDS. Furthermore, scanning may be use-ful in the evaluation of the later fibroproliferativestage of ARDS.

BRONCHOSCOPY

Flexible bronchoscopy has greatly influenced diag-nostic capability in the ICU. It is easy to performbecause the airways are readily accessible through theendotracheal or tracheostomy tube. Bronchoscopy maybe both diagnostic and therapeutic. For example,in a patient with pulmonary collapse and worseningarterial oxygenation, bronchoscopy may demonstratemucoid impaction causing lobular collapse, whichcan be removed at the same time. However, itmay sometimes be difficult to identify an obstructedorifice, either because of the variability of bron-chial anatomy or because of distal impactionbeyond the field of vision. In the USA, pulmono-logists with bronchoscopic training are often incharge of ICU patients, but this is not usually thecase in Europe. We would therefore encouragetraining in bronchoscopy to be given to intensivecare physicians. Acute respiratory deteriorationduring bronchoscopy may result from a reduction inalveolar ventilation consequent upon the introductionof airway obstruction. It may also relate to

Table 7.1 Indications for bronchoscopy in the intensivecare unit and high dependency unit

Emergency interventionsEndoscopic intubationMalpositioning of the endotracheal tubeHaemoptysisChest traumaChemical or thermal burns of the tracheobronchial treeMassive witnessed aspirationSuspected foreign-body aspiration

Elective interventionsPulmonary infiltrates associated with clinical signs of

infectionUnexplained lung collapse or pleural effusionPercutaneous tracheostomyPostoperative evaluation of patents after lung resection or

of lung transplant recipients

sedative rugs. If local anaesthesia is employed,only minimal additional sedation is necessary (ifany), but coughing may provoke transiently high airwaypressure in the intubated patient. In cases ofspontaneously breathing patients with respiratoryfailure who need a bronchoscopic procedure,non-invasive respiratory support can be employed.The bronchoscope is then introduced via the fullfacemask, but the procedure may be difficult whenair leaks cannot be managed.4

Table 7.1 summarizes the major indications forbronchoscopy in patients in the ICU or high depend-ency unit (HDU).

Endotracheal intubation

At intubation, the laryngoscope is usually used tovisualize the glottis. The bronchoscope is an altern-ative if abnormal upper airway anatomy makes intu-bation difficult, e.g. following trauma or in cervicalscoliosis or oropharyngeal disease. The endotrachealtube is slipped over the bronchoscope and used as aguide for intubation when the upper trachea is visu-alized. Bronchoscopy can also ensure correct endo-tracheal tube positioning because the distancebetween the tip of the tube and the carina can be eas-ily assessed. In cervical spine injury, direct laryngo-scopic intubation entails the risk of further injury,and fibreoptic nasal intubation is advised in this situ-ation.5 Although the chest radiograph is usually

Bronchoscopy 83

sufficient to exclude malposition, it is an alternativeif profound hypoxaemia calls for urgent assessment.

Haemoptysis

Accurate diagnosis is essential in the management ofhaemoptysis because massive pulmonary bleedingmay rapidly lead to respiratory failure and the needfor endotracheal intubation. If severe, elective intub-ation of the patient prior to bronchoscopy is appro-priate. However, massive haemoptysis (between 300and 1000 mL in 24 hours) is only present in about5% of cases and not all of these patients need urgentbronchoscopy.6 Distinction between gastrointestinal,pharyngeal, nasal and pulmonary sources oftenrequires endoscopy. Blood from the lungs is coughedrather than vomited, is partly frothy, alkaline andsometimes mixed with pus. Nausea and vomiting aresuggestive of haematemesis, whereas a history ofcough, weight loss, fever and night sweats is a clue forhaemoptysis. The goals of bronchoscopy are to iden-tify the source of bleeding and prevent blood spillinginto unaffected parts of the lung. The options avail-able include endobronchial suctioning, cold salinelavage, use of topical vasoconstrictors (e.g. epineph-rine in a dilution of 1:20000) and/or coagulants,endobronchial balloon tamponade or selectivebronchial intubation (see Chapter 9).7

Chest trauma, thermal or chemicalburns of the tracheobronchial tree

Although car seat belts and the airbag have greatlyreduced the incidence of blunt trauma to the thoraxin motor vehicle accidents, bronchoscopy is neces-sary to assess airway damage. Physical and radio-graphic findings are important determinants of theneed for bronchoscopy, e.g. pneumothorax, subcuta-neous or mediastinal emphysema, haemothorax, flailchest, atelectasis and haemoptysis. The identificationof discontinuation of the tracheobronchial treeranges from bronchial contusion and laceration tocomplete tracheal transection.8

Thermal inhalation injury to the lungs can be dev-astating and reliance on the absence of clinical criteria,such as facial or oropharyngeal burns, carbonaceoussputum, wheezing and hoarseness, is insufficient toexclude burn injury. Acute injury produces erythema,

mucosal sloughing and severe airway oedema.Bronchoscopy allows earlier diagnosis and appropri-ate therapy, including corticosteroids, humidified air,antibiotics and assistance in clearing airway plugs. Therole of bronchoscopy after chemical injury is to deter-mine the extent of the chemical burn and to assess orremove secondary complications such as necroticdebris, stenosis or other life-threatening sequelae .9

Massive aspiration and suspectedforeign-body aspiration

Massive aspiration of gastric contents almostinevitably leads to pneumonia, the severity of whichdepends on the amount and type of aspirate.Radiographically, aspiration pneumonia most com-monly involves the posterior segments of the upperlobes and the superior segments of the lower lobes.Rapid diagnosis and therapeutic intervention decreasemorbidity and mortality, although there is no consen-sus on the role of bronchoscopy after aspiration. Thesuction channel of the flexible bronchoscope is toosmall to recover solid material, and the introductionof saline to dilute the aspirate and remove it from thetracheobronchial tree may be harmful if it allowsspread into more distal parts of the lung. Therefore,bronchoscopy is probably limited to those patientswho develop signs of acute airway obstruction due tothe aspiration of larger food particles.

Respiratory compromise due to foreign-bodyaspiration is rarely severe enough to cause respiratoryfailure, although acute distress may result if theobstructing object is large or localized centrally.Foreign-body aspiration typically occurs in childrenor the elderly and manifests as obstructive lobar orsegmental over-inflation or atelectasis. If a definitehistory of aspiration is given or a foreign body is seenradiographically, rigid bronchoscopy is preferable toeffect removal. Flexible bronchoscopy is more readilyavailable, however, and may be more appropriate forinitial inspection, particularly of the distal parts ofthe tracheobronchial tree.

Pulmonary infiltrates associated withclinical signs of infection

New pulmonary infiltrates in mechanically venti-lated patients suggest nosocomial pneumonia,


Table 7.2 Diagnostic tools for nosocomial pneumonia in

the intensive care unit and high dependency unit

Sputum

CultureGram stain

Blood cultureSerology

Legionella pneumophilaAtypical bacteriaViruses

Endotracheal aspirateb

Urinary antigenLegionella pneumophilaStreptococcus pneumoniae

Molecular techniquePolymerase chain reaction

Percutaneous needleaspiration

Lung needle3

Pleural effusionProtected specimen brushBronchoalveolar lavage

ConventionalProtectedBlind

aNot recommended in mechanically ventilated patients.bln intubated patients.

especially when associated with other signs ofinfection. Quantitative bacterial cultures may behelpful to differentiate between colonization andinfection (see Chapter 15) and various diagnostictools are available (Table 7.2). The best approach iscurrently contentious, but bronchoscopy mayreduce mortality in ventilator-associated pneumo-nia.10 A consensus statement11 provides a clinicalalgorithm to guide management (Fig. 7.1). As thisissue is discussed in detail in Chapter 15, only thebasic aspects of diagnostic techniques are providedhere.

NON-INVASIVE METHODS

The sensitivity of blood culture in nosocomial pneu-monia is low and lacks specificity in the critically ill.Micro-organisms isolated in blood cultures shouldonly be considered to be the definitive cause whenalso isolated from respiratory samples. The quantita-tive culture of samples from the lower respiratorytract may assist in differentiation from colonization.For endotracheal aspirates, a cut-off of 105 colony-forming units mL-1 (cfu) gives an acceptable sensit-ivity and specificity in diagnosis in the presence ofclinical signs.12 Detection of bacterial antigen inurine is useful for Legionella or Streptococcus pneu-raom'ae.13'14 whereas molecular methods currentlylack specificity.

BRONCHOSCOPIC METHODS

Bronchoscopic specimens may be contaminated bymaterial from the upper airway or the endotrachealtube. To avoid this, one approach is the use of theprotected specimen brush (PSB), positioned via thebronchoscope at the orifice of a radiographicallyidentified segmental bronchus. The catheter isadvanced approximately 3 cm out of the fibreopticbronchoscope and the inner cannula is then pro-truded to eject a distal carbon wax plug. After suc-tioning material, the brush is retracted back insidethe inner cannula, which is then pulled back into theouter cannula and removed from the bronchoscope.The threshold for quantitative bacterial culturesfrom a PSB is 103 cfu mL-1.

Bronchoalveolar lavage (BAL) samples a largerportion of lung parenchyma, has good sensitivityand is especially useful in immunocompromizedpatients. The specificity of BAL is limited by con-tamination with upper-airway bacteria found in upto one-third of specimens. There are several waysof performing BAL, such as directed and non-directed non-bronchoscopic and directed bron-choscopic methods. Growth above 104 cfu mL- 1

in quantitative bacterial cultures is usually sig-nificant, no matter which BAL technique isemployed.11

Unexplained lung collapse or pleuraleffusion

Atelectasis is often encountered in the ICU/HDUsetting because of relative hypoventilation in non-intubated patients, e.g. post-laparotomy or in asso-ciation with pressure-limited ventilation strategies.In the ventilated patient, a recruitment manoeuvremay be successful in re-expanding the lung.Persistent lobar collapse may be caused by anendobronchial lesion or a mucous plug. Endoscopymay therefore be both diagnostic and therapeutic.Pleural effusions have many causes but, in the crit-ically ill, they are commonly related to congestivecardiac failure, pulmonary infection, neoplasm,excessive fluid replacement during resuscitation orhypoproteinaemia. Pleural effusion due to heartfailure is usually bilateral and evidence for this maybe provided by echocardiography or a pulmonaryartery catheter. In unilateral effusion, diagnostic

Bronchoscopy 85

Figure 7.1 Clinical algorithm for the

diagnosis of nosocomial pneumonia.

"Two or more of the following

criteria: body temperature >38 °C or

<36 °C, leucocytosis or leucopenia,

purulent tracheal secretion, decreased

Pa02. f Radiographic evidence of

alveolar infiltrates, air

bronchograms, new or worsened

infiltrates. There is no definitive

evidence to support either option A

or B. Therefore, the clinician should

choose the appropriate test based

on its sensitivity and specificity,

potential adverse effects, availability

and cost. (Adapted, with permission,

from reference 11.)

aspiration is indicated, with care being taken tominimize the risk of pneumothorax in mechani-cally ventilated patients.

wire. Although experienced personnel may not needto employ bronchoscopy, it is mandatory to have abronchoscope available in case complications develop.

Percutaneous dilatation tracheostomy

Tracheostomy prevents complications associated withprolonged translaryngeal intubation and easesweaning from mechanical ventilation.15 Percutaneousdilatation tracheostomy is a safe alternative to surgicalplacement. Bronchoscopy may be helpful during thisprocedure to verify a midline insertion of the guide

Postoperative evaluation of patientsafter lung resection or of lungtransplant recipients

Patients with lung resection or transplantation maydevelop a number of complications, such as suturegranuloma or anastomotic air leak. Bronchoscopyis of critical importance in the diagnosis of

Image Not Available


anastomotic dehiscence. In addition, pulmonaryinfiltrates may indicate infection or rejection in lungtransplant recipients.

Bronchoscopy allows the identification of pul-monary infections with BAL or pulmonary rejectionby transbronchial biopsy. Bronchiolitis obliterans syn-drome remains a major late complication and may bean immediate problem in the lung transplant patient.16

Bronchoscopy is again needed for tissue sampling.

BEDSIDE DIAGNOSTIC ULTRASOUND

Ultrasound is important in the diagnosis of variousconditions in the ICU or HDU.17 It has considerablevalue in the assessment of the pleura or pleuralspace. For the aspiration of fluid, simple marking ofthe overlying skin is adequate, without the need forreal-time imaging. The method of drainage dependson the fluid aspirated. If it is clear and odourless, anon-infectious aetiology is usual and simple aspira-tion is often adequate. Bacterial culture and otherlaboratory parameters will be employed to identifythe cause, e.g. haemoglobin, glucose, lactate dehy-drogenase and cholesterol (see Chapter 17). Whenthe ultrasound shows loculations, or the aspirate isbloody or purulent, a chest tube should be consid-ered. The application of ultrasound in the respira-tory ICU is not limited to the diagnosis of pleuralcollections or the placing of drains.18

Transoesophageal echocardiography is a Dopplertechnique that uses the oesophagus as an acousticwindow. In ventilated patients, it may be used to assessleft ventricular function, valvular disease, endocarditisand prosthetic valve dysfunction. Transoesophagealechocardiography has been increasingly used for thediagnosis of pulmonary embolism and is superior totransthoracic echocardiography in evaluating a car-diac source of embolism. In comparison with radio-logical procedures, it had limited accuracy fordetecting pulmonary embolism with acute cor pul-monale in one study.19 Only when the pulmonaryembolism was located in the main or right pulmonaryartery could transoesophageal echocardiography bereliable in confirming the diagnosis. Its performancewas better in the presence of shock due to massivepulmonary embolism. In comparison with pulmonaryscintigraphy or autopsy, the sensitivity of trans-oesophageal echocardiography for the diagnosis of

massive pulmomonary embolism was 99%, with aspecificity of 100%.20

OTHER BEDSIDE PROCEDURES IN THEINTENSIVE CARE UNIT

Portable CT of the chest is now available in somecentres and is of value for critically ill patients whocannot be moved from the ICU.21 It is especiallyuseful in the evaluation of ARDS, differentiatingintrapulmonary from pleural abscess, assessingthe mediastinum and in tracheo-oesophageal or tra-cheopulmonary fistula. CT is the modality of choicefor establishing the diagnosis of exogenous lipoidpneumonia resulting from the aspiration of hydro-carbons and mineral oil. The measurement of lungpermeability employing nuclear isotopes in acutelung injury may be useful, although other indirecttechniques such as PICCO are more likely to be use-ful in monitoring (see Chapter 8).

ACKNOWLEDGEMENT

Dr Torsten Bauer has been in part supported by theBochumer Arbeitskreis fur Pneumologie undAllergologie, Bochum, Germany (BAPA).

REFERENCES

1. Tocino, I. Chest imaging in the intensive care unit. Eur

J Radiol 1996; 23: 46-57.

2. Pulmonary Artery Catheter Consensus Conference:

consensus statement. Crit Care Med 1997; 25: 910-25.

3. Duane, PG, Colice, GL Impact of noninvasive studies

to distinguish volume overload from ARDS in acutely

ill patients with pulmonary edema: analysis of the

medical literature from 1966 to 1998. Chest 2000;

118: 1709-17.

4. Vitacca, M, Nava, S, Confalonieri, M, et al. The appropri-

ate setting of noninvasive pressure support ventilation

in stable CQPD patients. Chest 2000; 118:1286-93.

5. Fuchs, G, Schwarz, G, Baumgartner, A, Kaltenbock, F,

Voit-Augustin, H, Planinz, W. Fiberoptic intubation in

327 neurosurgical patients with lesions of the cervical

spine. 7 NeurosurgAnesthesiol 1999; 11: 11-16.

References 87

6. Dweik, RA, Stoller, JK. Role of bronchoscopy in

massive hemoptysis. Clin Chest Med 1999; 20: 89-105.

7. Jean-Baptiste, E. Clinical assessment and management

of massive hemoptysis. Crit Care Med 2000; 28:1642-7.

8. Hara, KS, Prakash, UB. Fiberoptic bronchoscopy in the

evaluation of acute chest and upper airway trauma.

Chest 1989;96: 627-30.

9. Freitag, L, Firusian, N, Stamatis, G, Greschuchna, D.

The role of bronchoscopy in pulmonary complica-

tions due to mustard gas inhalation. Chest 1991;

100: 1436-41.

10. Fagon, JY, Chastre, J, Wolff, M, et al. Invasive and

noninvasive strategies for management of suspected

ventilator-associated pneumonia. A randomized trial.

Ann Intern Med 2000; 132: 621-30.

11. Grossman, RF, Fein, AM. Evidence-based assessment of

diagnostic tests for ventilator-associated pneumonia.

Executive summary. Chest 2000; 117:177s-81s.

12. El-Ebiary, M, Torres, A, Gonzalez, J, Puig de la

Bellacasa, J, Garcia, C, Jimenez de Anta, MT.

Quantitative cultures of endotracheal aspirates for the

diagnosis of ventilator-associated pneumonia. Am J

Respir Crit Care Med 1993; 147:1552-7.

13. Dominguez, J, Gali, N, Blanco, S, et al. Detection of

Streptococcus pneumoniae antigen by a rapid

immunochromatographic assay in urine samples.

Chest 200V, 119:243-9.

14. Benson, RF, Tang, PW, Fields, BS. Evaluation of the

Binax and Biotest urinary antigen kits for detection

of Legionnaires' disease due to multiple serogroups

and species of Legionella. J Clin Microbiol 2000; 38:

2763-5.

15. Kearney, PA, Griffen, MM, Ochoa, JB, Boulanger, BR,

Tseui, BJ, Mentzer, RMJ. A single-center 8-year

experience with percutaneous dilational tracheostomy.

dm? Swrg 2000; 231: 701-9.16. Meyers, BF, Lynch, J, Trulock, EP, Guthrie, TJ, Cooper,

JD, Patterson, GA. 1999. Lung transplantation: a

decade of experience. Ann Surg 1999; 230: 362-71.

17. Beagle, GL Bedside diagnostic ultrasound and

therapeutic ultrasound-guided procedures in the

intensive care setting. Crit Care Clin 2000; 16: 59-81.

18. Lichtenstein, D, Meziere, G, Biderman, P, Gepner, A.

The comet-tail artifact: an ultrasound sign ruling out

pneumothorax. Intensive Care Med 1999; 25: 383-8.

19. Vieillard-Baron, A, Qanadli, SD, Antakly, Y, et al.

Transesophageal echocardiography for the diagnosis

of pulmonary embolism with acute cor pulmonale: a

comparison with radiological procedures. Intensive

Care Med 1998; 24: 429-33.

20. Krivec, B, Voga, G, Zuran, I, et al. Diagnosis and

treatment of shock due to massive pulmonary

embolism: approach with transesophageal

echocardiography and intrapulmonary thrombolysis.

Chest 1997; 112:1310-16.

21. White, CS, Meyer, CA, Wu, J, Mirvis, SE. Portable CT:

assessing thoracic disease in the intensive care unit.

Am J Roentgenol 1999; 173:1351-6.

8MonitoringRICHARD BEALE

Introduction

Arterial blood-gas analysis

Continuous blood-gas analysis

Oximetry

Capnography

Transcutaneous carbon dioxide tensionmeasurement

8889909094

95

Ventilatory monitoring during mechanicalventilation

Haemodynamics and the measurement oflung water

Conclusion

References

96

102

104

104

INTRODUCTION

Respiratory critical care involves much more than anarrow consideration of the respiratory system. Mostpatients who require respiratory critical care alsohave involvement of other organ systems. Lunginjury and the requirement for respiratory supportmay be either the primary problem, often leading tosecondary systemic consequences, or the secondaryconsequence of another primary disease process.Nevertheless, optimum management of the respira-tory system is crucial if the patient is to recover,because very few critically ill patients avoid respira-tory problems altogether. Effective respiratory sys-tem monitoring is therefore vital, and this chapteraims to provide an overview of the more commonlyused techniques.

Monitoring constitutes such a routine part of thecare of critically ill patients that it is often taken forgranted. In a modern intensive care unit (ICU),increasingly complex equipment is employed tomeasure and display various aspects of patient physi-ology. Indeed, as computer technology advances,measurements and variables that were once solelythe preserve of the research laboratory are now read-ily available at the bedside. These advances challenge

our ability to understand and use this extra informa-tion to benefit our patients. In some cases, there areseveral different technologies that will provide a par-ticular piece of information, and the decision aboutwhich choice is made can have far-reaching organ-izational consequences.

In order to make such choices appropriately, it isimportant to consider the purpose and process ofmonitoring as it applies to a critically ill patient.Monitoring is much more than making a singlemeasurement, because it implies the ability to take ameasurement repeatedly, even continuously, inorder to follow over time the physiological processthat the measurement represents. Usually, this iswith the aim of detecting deterioration or improve-ment, often in relation to therapeutic interventions.Frequently, the ability to track a physiologicalprocess by scientific measurement may be limited bythe technology available. Although this seems anobvious statement, its implications are more subtle.Necessarily, processes and problems can becomedefined by the methods used to measure them, evenif these methods are far from ideal. In turn, this canmislead the clinician if the limitations and surrogatenature of a measurement are forgotten or ignored.Arguably, much of the current controversy over the

Arterial blood-gas analysis 89

use of the pulmonary artery catheter arises fromprecisely this phenomenon. An understanding ofthe strengths and weaknesses of alternative moni-toring approaches is therefore essential in order toavoid these pitfalls.

The first major area of respiratory critical caremonitoring involves the adequacy of gas exchange,which is most frequently assessed using arterialblood-gas analysis, oximetry and capnography. Thesecond involves monitoring of the most frequentlyused supportive therapy, mechanical ventilation, andinvolves increasingly sophisticated bedside measure-ments of respiratory mechanics. The third is theclosely related area of haemodynamic monitoring,especially as it relates to heart-lung interaction andthe effects of fluid resuscitation upon the injuredlung. It is only possible to touch the surface of theseareas in this chapter, but specialist texts are availableto provide extra detail.

ARTERIAL BLOOD-GAS ANALYSIS

Monitoring arterial blood gases through repeatedintermittent measurement of withdrawn blood is amainstay of respiratory monitoring in any ICU.Standard blood-gas analysers provide measurementof the O2 (PO2) and CO2 (PCO2) tensions and thepH of the blood sample being analysed. Althougharterial blood is used most frequently, the sameprinciples can also be applied to venous samples,and arterio-venous differences in these variables canprovide useful information. It is crucial to rememberthat the measured PO2 and PCO2 tensions derivefrom gas dissolved in the plasma rather than the totalamount of O2 or CO2 present in the sample.

Principles of measurement

Most blood-gas analysers use an O2 electrode of theClark1 or polarographic type to measure the PO2. Inthis system, electrons are provided from a cathodeand react with dissolved O2:

Four electrons are used to reduce each molecule ofO2, so the overall current generated is proportionalto the amount of dissolved O2, i.e. the PO2. In theClark electrode, the cathode and anode are within anelectrolyte solution separated from the blood (orother fluid being measured) by a membrane perme-able to O2.

Measurements of pH and PCO2 share the samebasic principle. In a pH electrode, blood is drawninto a capillary tube of pH-sensitive glass bathed in abuffer solution that maintains the pH outside theglass at a constant value. A potential is thus generat-ed across the glass that is proportional to the pH ofthe blood sample within the capillary.

In the CO2 electrode, a pH-sensitive glass elec-trode is bathed in bicarbonate solution and coveredby a CO2-permeable membrane. When blood ispassed across the other side of the membrane, CO2

diffuses into the bicarbonate solution and reacts withthe water present:

The hydrogen ion concentration (pH) is thenmeasured in the same way as before, and the systemcan be calibrated for PCO2 because the pH change islinearly related to the logarithm of the PCO2.

Temperature correction of blood-gasmeasurement

Most blood-gas analysers maintain the electrodetemperature at 37°C, but few patients have blood atprecisely that temperature. If the patient is relativelyhypothermic, the blood sample will therefore beheated in the electrode, resulting in decreased gassolubility and decreased O2 affinity for haemoglobin,and therefore an artificially raised blood-gas tension.The reverse is true if the blood sample is cooled inthe analyser (i.e. the patient is hyperthermic). Mostanalysers include formulae to correct for this effecton PO2, PCO2 and pH in their software, if the truepatient temperature is entered into the machine.

During cardiac surgery, there are two differentapproaches to the effect of temperature upon pHthat are commonly utilized. The first is to correct forthe effect of temperature during patient cooling oncardiopulmonary bypass and aim at a normal pH forthe temperature (the pH increases approximately

90 Monitoring

0.015 pH unit per degree centigrade of cooling), andis known as the pH-stat approach. The second is tokeep the pH normal at 37°C regardless of thepatient's actual temperature, and is known as thealpha-stat approach. It remains unclear as to whichapproach is more beneficial.

CONTINUOUS BLOOD-GAS ANALYSIS

The major drawbacks of the conventional approachto blood-gas analysis are the intermittent nature ofthe measurement and the requirement to drawblood. In an attempt to overcome these disadvan-tages, continuous on-line blood-gas sensors havebeen developed.

Principles

Most critically ill patients, especially those who areunstable, have indwelling arterial lines in place tofacilitate frequent arterial blood sampling and con-tinuous arterial blood pressure measurement. Oneapproach that overcomes the problems of traditionalblood-gas sampling is to place small sensors withina cassette that can be placed in the arterial line circuitclose to the cannula in such a manner as to allowblood withdrawal, measurement and return in aclosed fashion. A more sophisticated solution is toutilize miniaturized sensors for the measurement ofPO2, PCO2 and pH that are sufficiently small to fit ona probe that can be passed through a 20-G arterialline. Technically, it is possible to miniaturize a Clarkelectrode sufficiently for this purpose,2 but attemptsto do the same for pH and PCO2 electrochemicalelectrodes have been more troublesome.

An alternative approach is to use optical sensortechnology, sometimes called optodes. Such sensorsmay utilize the principles of absorbance, fluores-cence, phosphorescence or chemiluminescence, thedetails of which are described elsewhere. The bestknown and most widely used of the available com-mercial devices is probably the Paratrend 7® devicefrom Diametrics. This device previously used acombination of an electrochemical (Clark-type) PO2

sensor and absorbance -PCO2 and pH sensors, butthe latest probes rely entirely upon optical techno-logy and are similarly sized to a standard arterial line.

Although continuous pulse oximetry andcapnography had initially appeared to make the useof continuous indwelling (and therefore invasive)blood-gas analysis largely unnecessary, develop-ments in intensive care have demonstrated some ofthe limitations of these techniques. Pulse oximetry iscritically dependent upon the quality of the perfu-sion signal in the tissue within the probe area, andperipheral perfusion is often severely limited in verysick patients. Moreover, due to the flat shape of theupper part of the haemoglobin O2 dissociationcurve, there is very little change in O2 saturation(SO2) for dramatic changes in PO2. This has becomeof greater importance with the increasing use of lungrecruitment manoeuvres, where it is often not feasi-ble to measure changes in absolute lung volume, andwhere changes in PO2 may be the best and mostimportant bedside surrogate for judging clinicalresponse (Fig. 8.1). This is greatly facilitated by theuse of continuous, online measurement.

OXIMETRY

Under normal circumstances, nearly all the O2

carried in the blood is attached to haemoglobin,yet traditional blood-gas analysis only provides infor-mation about that very small proportion dissolvedin the plasma. Oximeters are devices designed tomeasure the degree to which haemoglobin is saturated

Figure 8.1 A high-pressure recruitment manoeuvre in a patient

with acute respiratory distress syndrome performed using a

high-frequency oscillator (Sensormedics 3100B). Changes in P02 act

as a surrogate for changes in lung volume, and demonstrate

recruitment and hysteresis. Sp02 measurements would not

illustrate this effect because this patient is on the flat, upper

portion of the 02-dissociation curve.

Oximetry 91

with O2, and fall into two major categories. Both usethe principles of light absorption, but serve differentpurposes. The first group comprises the bench-topoximeters, often called co-oximeters. Thesemachines utilize light at several wavelengths to sepa-rate the quantities of oxyhaemoglobin. (HbO2),reduced haemoglobin (Hb), methaemoglobin(MetHb) and carboxyhaemoglobin (COHb) in aheparinized blood sample, and express the arterialO7 saturation (SaCO as:

Originally, co-oximeters were freestandingdevices, but now many modern blood-gas machinescan be supplied with in-built oximeters, providingboth blood-gas and oximetry profiles.

Unlike co-oximeters, pulse oximeters are truemonitoring devices, providing continuous measure-ment of in-vivo HbO2, usually referred to as SpO2.The introduction of pulse oximetry constituted amajor advance in respiratory monitoring, and todaythese devices are ubiquitous in modern respiratory,critical care and anaesthetic practice.

Principles of pulse oximetry

Pulse oximeters use the principles of spectrophoto-metry to provide a value for the arterial haemoglobinO2 saturation. Light at two wavelengths, usually 660nm (red) and 940 nm (infrared), is shone from twolight-emitting diodes (LEDs) situated in a finger orear probe. Oxyhaemoglobin (HBO2) absorbs more

infrared light than reduced haemoglobin, whilstreduced haemoglobin absorbs more red light thanoxyhaemoglobin. The two diodes are switched on andoff rapidly (approximately 600 times s-1), and eachsequence allows the detector within the probe to mea-sure the transmission of red light and infrared light,and to control for ambient light. There is a pulsatile(alternating current, AC) component to the absorp-tion signal due to the pulse of arterial blood flowingwithin the detection area, and a non-pulsatile (directcurrent, DC) component due to absorption of lightby non-pulsatile arterial blood, venous and capillaryblood and other tissues. The device relates the pul-satile component to the non-pulsatile component ateach wavelength, so calculating a ratio:

Using an algorithm that describes the relationshipbetween R and measurements of haemoglobin O2

saturation made in vivo using a bench-top oximeter,the pulse oximeter saturation (SpO2) is then calcu-lated. In terms of accuracy, most manufacturersclaim that the 95% confidence limit for measure-ments is ±4% at SaO2 levels above 70%. Accuracydeteriorates when SaO2 falls below this level, partlybecause of the difficulty in obtaining reliable humancalibration data in conditions of extreme hypo-xaemia. Different probe types can also alter deviceperformance, with ear probes tending to have a fasterresponse time than finger probes. Clearly, the averag-ing and sampling algorithms employed by differentmanufacturers are also of importance, and these areconstantly being improved.

Figure 8.2 02-dissociation curve for haemo-

globin. 2,3-DPG, 2,3-diphosphoglycerate.

92 Monitoring

Although pulse oximeters are widely used, theyhave a number of significant shortcomings that maymislead the clinician. These include:

• lack of responsiveness at high PO2 values,• dependence upon adequate tissue perfusion,• inability to discriminate between abnormal

haemoglobin species,• interference from dyes,• susceptibility to ambient light,• false alarms and susceptibility to motion artefacts.

LACK OF RESPONSIVENESS AT HIGH P02 VALUES

Because pulse oximeters measure haemoglobin O2

saturation, they provide very little information aboutchanges in arterial PO2 once the flat portion of theO2 dissociation curve has been reached (Fig. 8.2).This means that the dramatic changes in PO2 seenduring recruitment manoeuvres in responsivemechanically ventilated patients cannot readily betracked with a pulse oximeter.

DEPENDENCE UPON ADEQUATE TISSUEPERFUSION

Under circumstances in which cardiac output and tis-sue perfusion fall, the ability of the device to detect apulsatile signal becomes impaired. Many manufactur-ers also provide a plethysmograph trace, which givessome information as to the adequacy of perfusion.Some devices increase the amplification of the dis-played trace as the signal decreases, whereas othersprovide the ability to alter the amplification manually,or display a perfusion index. The latter approach ismore useful if the plethysmograph trace is being usedto assess perfusion. Eventually, perfusion may becometoo poor for an analysable signal to be detected, andan error message will be displayed. Unfortunately, insome situations, inaccurate SpO2 values may be dis-played first, especially if perfusion is sufficiently poorto become gradually impaired by the contact pressurefrom the probe itself. Under these circumstances, thevalues obtained should be checked by arterial blood-gas analysis and bench-top co-oximetry.

INABILITY TO DISCRIMINATE BETWEENABNORMAL HAEMOGLOBIN SPECIES

The two most commonly encountered abnormalhaemoglobin species are COHb and MetHb. Both

interfere with pulse oximeter measurements if pre-sent in significant amounts. COHb has a very similarabsorption spectrum to HbO2, and pulse oximetersprovide readings for SpO2 that are actually the sumof the HbO2 and COHb present. Therefore, in situ-ations of severe carbon monoxide poisoning, pulseoximetry may lead the user to believe that thepatient is well saturated when the opposite is actuallytrue.

MetHb absorbs similar amounts of light at both660 nm and 940 nm, giving an absorption ratio of 1.This equates to a SpO2 value of about 85% on theoximeter calibration curve, so the more MetHb pre-sent, the more the SpO2 reading will tend towards85%. If methylene blue is used to treat the methamo-globinaemia, although it is highly effective, it willalso result in falsely low SpO2 readings because of itsblue colour.

INTERFERENCE FROM DYES

Both methylene blue and indocyanine green may begiven intravenously to critically ill patients for a vari-ety of purposes, and both interfere with the signaldetected by the pulse oximeter.

AMBIENT LIGHT

Although the LED activation sequence within theoximeter probe is designed to correct for ambientlight levels, bright sunlight and some forms of high-intensity artificial light may still interfere with signaldetection.

FALSE ALARMS AND SUSCEPTIBILITY TO

MOTION ARTEFACTS

A persistent problem with the use of pulse oxime-ters is the number of false alarms generated. Someauthors have suggested that only one in five alarmsis genuine. Movement artefact is usually the cause,due either to interference from motion or theprobe becoming detached. Poor perfusion may alsohave this effect, when the local SpO2 falls as a con-sequence. Although staff using these devices havelargely become inured to this phenomenon, it canbe problematic when attempting to analyse trendsretrospectively. It can also lead to genuine alarmsbeing ignored. Certainly, no other routinely used

Oximetry 93

monitoring device provides quite so many falsealarms.

Oxygen content

Because most of the O2 present within a blood sam-ple is bound to haemoglobin, calculation of the O2

content requires knowledge of the PO2, SaO2 andHb:

Ca02 (per 100 ml blood) = [Hb (g dL-1) X 1.34 X

SaOj + [0.0031 X Pa02 (mmHg)]

For a patient with normal results, i.e. a Hb con-centration of 15 g dL- l

PO2 of 100 mmHg, this equates to:SaO2 of 97% and arterial

Ca02 (per 100 ml blood) - [15 (g dL-1) X 1.34 X

0.97] + [0.0031 X 100 (mmHg)]

This calculation clearly illustrates how little of theO2 present in an arterial blood sample is reflected inthe PO2 measurement.

The systemic O2 delivery (DO2I) can be calculatedby multiplying the arterial O2 content by the cardiacindex as follows:

where CI represents cardiac index, the contentis multiplied by 10 to convert it to 'per litre' and thedissolved O2 content is ignored because it is largelyinsignificant.

MIXED VENOUS OXYGENATION

The principles of oximetry can also be applied tothe venous circulation, to facilitate the assessmentof systemic O2 consumption. There are twoapproaches to this and both require the presenceof a pulmonary artery catheter. Aspiration ofblood from the distal lumen of a pulmonary arterycatheter provides a true 'mixed venous' sample,the saturation of which can then be measured inthe normal fashion with a bench-top oximeter.

Thelated:

mixed venous O2 content can then be calcu-I:

Cv02 (per 100 ml blood) = [Hb (g dL-1)] X 1.34 X

Sv02 + [0.0031 X Pv02 (mmHg)]

From this, systemic O2 consumption (VO2I) can becalculated:

where CI represents cardiac index, and the dissolvedO2 content is ignored.

With the development of fibreoptic pulmonaryartery catheters, continuous measurement of SvO2 invivo using the principles of reflectance photo-metry has become possible. Light at two or threewavelengths (usually including 660 nm and 805 nm)is shone through the pulmonary artery catheterfibreoptic bundle and the reflectance signal analysed.Proprietary empirical formulae are used that allowblood to be regarded as a particulate suspension.Correction for the haematocrit is required, and thismay be done manually or by analysis within thedevice of the reflectance data.

SvO2 changes as the relationship between O2

delivery and O2 consumption changes. If O2 deliveryfalls but O2 consumption remains constant, it fol-lows that O2 extraction must increase. This will bemanifest by a fall in SvO2.

02 extraction ratio (OER) =

Changes in SvO2 occur almost immediately inresponse to changes in O2 delivery, i.e. changesin cardiac output, Hb concentration or SaO2, andin response to changes in O2 consumption.Continuous monitoring of this variable thereforeprovides a rapid and sensitive monitor of thecondition of a critically ill patient, although inter-pretation of the precise cause for the change can bemore difficult.

Continuous monitoring of SaO2 and SvO2 canalso be used to calculate shunt fraction as follows:

94 Monitoring

This provides a convenient approximation for thestandard equation that can be readily performed atthe bedside.

CAPNOGRAPHY

Various techniques have been used to provide similarinformation for CO2. These have taken two majorforms; expired gas CO2 analysis (capnometry andcapnography) and transcutaneous CO2 analysis.

Principles of expired carbon dioxideanalysis

CO2 is produced as the product of aerobic metab-olism and diffuses passively into the circulation.About 5-10% is dissolved in plasma and a similaramount binds to terminal amine groups onhaemoglobin and other proteins within the blood,with the remainder being buffered to HCO3-. CO2

is eliminated through the lungs by alveolar ventila-tion (VA), but there is also a proportion of dead-space ventilation (VD) that does not contribute toCO2 clearance, with the sum of these being thetotal expired minute ventilation ( VE). The alveolarPaCO2 varies according to the ventilation-perfu-sion (V/Q) relationships within the lung. In areaswithout perfusion (true dead space), V/Q — andthe PaCO2 equals the inspired PCO2 and is zero.At the other extreme, where ventilation falls andso VIQ also falls, the PaCO2 becomes very close tothe venous PvCO2 (45 mmHg under normalcircumstances).

The alveolar dead-space fraction can be calculatedfrom a modification of the Bohr equation and pro-vides an indication of the V/Q relationship:

where VD/VTis the dead-space fraction, PaCO2 isthe arterial PCO2 and PaCO2 is the alveolar PCO2,which can be represented by the end-tidal CO2

measured with a capnograph.

Techniques for measuring expiredcarbon dioxide

Measurement of CO2 in expired gases is now rou-tinely used in many situations in critically ill patientsand has become an obligatory safety requirementduring general anaesthesia, especially in the intub-ated patient. In addition to providing a numericdisplay of the CO2 concentration (capnometry),virtually all modern devices provide a graphicaldisplay of the CO2 waveform (capnography), whichprovides important extra information. There areseveral techniques available for making the CO2

measurement: mass spectrometry, Raman spectroscopy,colorimetry and infrared spectroscopy - which is byfar the most widely used. The infrared absorptionpeak for CO2 is at 4.26 |xm, but this is close to that ofN2O, and between those of H2O, leading to potentialinterference. The presence of other gases (helium,O2, nitrogen and nitrous oxide) may also causebroadening of the CO2 absorption spectrum.Nevertheless, these problems have been overcome inmodern devices through the use of specific filters,reference cells and known correction factors. To bereliable in clinical use, capnographs require regularcalibration using appropriate calibration gasmixtures.

Mainstream and sidestreamcapnographs

In mainstream capnographs, a measuring cell ispositioned directly in the breathing circuit, usuallymounted on a specially designed catheter mount inclose proximity to the endotracheal tube. The cell isusually heated to around 40°C to prevent condensa-tion on the cell window interfering with the mea-surement. The cell, which may be quite heavy andcumbersome, must be supported in such a mannerthat the patient's skin is protected and the endotra-cheal tube is not kinked or pulled out. In sidestreamcapnographs, gas is aspirated from the breathing sys-tem at a constant rate through a fine-bore samplingtube and passed through the measuring chamber.Such systems are often more convenient to use, butthere is a slower response time compared with main-stream systems due to the delay while the aspirated

Transcutaneous carbon dioxide tension measurement 95

gas sample reaches the cell. The sampling rate also hasto be appropriate to the overall ventilatory pattern,and is usually of the order of 50-500 mL min- l,which may have a significant impact upon thevolume measurements recorded by the ventilator. Ifthe sampling rate is higher than the expiratory gasflow, contamination of the sample with fresh gas mayoccur. Some systems return the sampled gas to thebreathing system, but most vent it to air, which mayhave implications for scavenging. Another problemwith these systems is that of condensation from thewarm expired gases being sampled, particularly ifartificial humidification is being used. There is nearlyalways a water trap in the system to prevent thesensing chamber becoming wet, but the samplingtube may still become blocked with water or secre-tions. The multiple connectors required with thesesystems also provide potential sites for leaks from thebreathing circuit.

Clinical use of capnography

Although the end-tidal PCO2 (PETCO2) is clearlyvery closely related to the PaCO2, it is not possibleto predict one from the other, especially if thereis a degree of lung disease. For that reason, thePETCO2 is best considered as a monitored variablein its own right, rather than as a surrogate forPaC02.

The most common and probably most importantapplication is to detect oesophageal intubation or

Figure 8.3 A typical capnograph waveform. The slope of phase III

is increased (as is the a angle) in patients with reduced V/Q match-

ing, e.g. chronic obstructive pulmonary disease. The p angle

increases as rebreathing increases.

endotracheal tube displacement.3 Sometimes, theremay be some CO2 present in the stomach, initiallyproviding potentially misleading informationabout tube position. Because any CO2 is not replen-ished, the presence or otherwise of a normal capno-graph waveform still allows rapid confirmation ofcorrect tube placement (Fig. 8.3). The consequencesof undetected tube misplacement are catastrophicand it is now generally considered mandatory to usecapnography during anaesthesia in intubatedpatients for safety reasons. Capnographs are notroutinely employed in all ICU patients, althoughthey should certainly be available for acute situa-tions and have been built into some ICU ventila-tors. Capnograph use is also increasingly regardedas mandatory for safety reasons during the trans-port of critically ill patients.

TRANSCUTANEOUS CARBON DIOXIDETENSION MEASUREMENT

Transcutaneous CO2 measurement provides an alter-native to capnography for the non-invasive monitor-ing of CO2.

4 Unfortunately, the technology isrelatively cumbersome and expensive, and transcuta-neous PCO2 may not be a very good reflection ofPaCO2 in critically ill patients. Although there are anumber of technical approaches to measuring tran-scutaneous PCO2, most commercially available sys-tems use a solid-state CO2 electrode, often combinedwith an O2 electrode. The major advantage of theapproach is that it can be applied to patients who havenot been intubated and do not have an arterialcatheter in place.

Transcutaneous PCO2 measurement detects theCO2 that escapes through the skin surface. Thiscomes mainly from capillary blood in the dermisand cells in the epidermis. Due to the countercur-rent arrangement of the capillaries, the PCO2 ishighest here rather than in the arterioles orvenules. This CO2, together with the epidermalproduction, means that the skin -PCO2 is higherthan the PaCO2. If skin perfusion falls, the removalof CO2 away from the skin is impaired, and theskin PCO2 becomes progressively higher comparedwith the PaCO2. This can be overcome by heatingthe skin area where the PCO2 electrode is attached,but this has effects on local skin metabolism, on

96 Monitoring

the solubility of CO2 and on the dissociation ofCO2 from haemoglobin. For these reasons, thetranscutaneous PCO2 will still be higher than thePaCO2 but, because the effects of temperature arelargely predictable, appropriate correction factorscan be employed during the calibration process.Burning of the skin at the electrode site occasion-ally occurs.

These practical limitations mean that transcuta-neous CO2 monitoring is rarely used in adult ICUpractice. It is still occasionally seen in some paedi-atric ICUs, but it has largely been supplanted bycapnography and intermittent or continuousblood-gas analysis. One area where it is still usefulis that of sleep studies for obstructive sleepapnoea.

VENTILATORY MONITORING DURINGMECHANICAL VENTILATION

Monitoring respiratory system mechanics andpatient-ventilator interaction is a crucial aspect ofrespiratory critical care. At one time, gaining infor-mation of this kind was complex and relied uponapparatus and techniques usually only available inthe respiratory physiology laboratory. Today, withmodern computerised ventilators, an increasingamount of information is available at the bedside,and can both guide adjustment of ventilator settingsand allow assessment of patient progress.

Basic principles

In order to ventilate the lungs, it is necessary to over-come the elastic forces generated by both the lungsand the chest wall. Although estimating pleural pres-sure through the use of an oesophageal balloon canseparate these, this is not often done in routine clin-ical practice. Therefore, it is usually the mechanics ofthe total respiratory system that is being considered.There are also resistive forces that arise as a conse-quence of the flow of gas through the tubes of theventilatory circuit and the airways.

Elastance (E) is described by the equation:

where P is pressure and V is volume. More com-monly, the term compliance (C) is used, which is thereciprocal of elastance:

These measurements may be either static ordynamic, depending on whether they are made atzero gas flow. Static compliance uses the plateaupressure or end-inspiratory pressure as the upperpressure for the calculation of change in pressure.Dynamic compliance uses the peak inflation pres-sure, which includes that component of the airwaypressure generated as a result of airways resistance.

If the lung and the chest wall are considered sepa-rately, they can be regarded as being in series, so that:

where CRS is the compliance of the total respiratorysystem, C^ is lung compliance and Ccw is chest-wallcompliance. Separation of the lung and chest-wallcomponents requires knowledge of the pleural pres-sure, usually represented clinically by the oesophagealpressure measured with an oesophageal balloon orcatheter.

Resistance can be described as:

i.e. the change in pressure for the rate of change ofvolume.

The product of compliance and resistance is thetime constant of the system. In diseased lungs,there will be a number of different time constantsrelated to the degree of heterogeneity of the injury.This is particularly so in the case of severe acutelung injury (acute respiratory distress syndrome,ARDS).

Commonly measured pressures

Nearly all ICU ventilators provide informationabout a number of different 'airway' pressures,

Ventilatory monitoring during mechanical ventilation 97

although, in reality, these are usually measured atthe ventilator end of the patient-ventilator circuit.These pressures are listed below and each has aspecific importance.

PEAK AIRWAY OR INFLFATION PRESSURE

This is the highest airway pressure recorded duringthe respiratory cycle and comprises a componentrequired to overcome the elastance of the lung anda component to overcome airways resistance.Therefore, peak inflation pressure (PIP) is not thepressure to which the alveolus is exposed directly andis not the best pressure against which to judge thelikelihood of barotrauma.

PLATEAU PRESSURE

This is the pressure that is recorded once inspiratoryflow has ceased and there is a pause with zero gasflow. It is the pressure used to calculate static com-pliance. Although it is commonly provided on manyventilators, a true plateau pressure will not bereached unless the duration of the inspiratory pauseis sufficient for full equilibration to take place.

END-INSPIRATORY PRESSURE

This pressure is measured using either a manual orprogrammed end-inspiratory pause of sufficientduration to ensure that a true plateau has beenreached. This may be between 2 and 5 s. End-inspi-ratory pressure is therefore the 'real' plateau pressureand represents the alveolar pressure at the end ofinflation. It is a better marker of the likelihood ofbarotrauma than PAP.

In controlled ventilation, the difference betweenPIP and plateau pressure, divided by the precedingflow rate, is used as one approach to calculatingresistance.

MEAN AIRWAY PRESSURE

This is the inflation pressure averaged over the wholerespiratory cycle and appears to relate most closely toboth the haemodynamic disturbance associated withmechanical ventilation and changes in oxygenation,particularly when inverse ratio ventilation (IRV) isused. This is often explained on the basis that meanairway pressure (MAP) represents mean alveolarpressure and therefore represents 'how much lung is

open for how long'. Because it is usually measured bya pressure transducer in the ventilator, and alsoincludes the PIP, MAP clearly includes a resistivecomponent. Nevertheless, in clinical practice theconcept seems to be useful. One of the fundamentalprinciples of mechanical ventilation in patients withacute lung injury is to improve oxygenation by opti-mizing lung recruitment, thereby achieving bettermatching of ventilation and perfusion. MAP is there-fore a useful measurement when altering ventilatorsettings to improve oxygenation.

END-EXPIRATORY PRESSURE AND POSITIVEEND-EXPIRATORY PRESSURE

Positive end-expiratory pressure (PEEP) is routinelyemployed in mechanically ventilated patients inorder to improve oxygenation. It improves lungrecruitment, prevents lung collapse and may therebyimprove ventilation-perfusion matching. Changes inPEEP also alter MAP. PEEP is set on the ventilatorand is usually measured in the ventilator circuit.During the expiratory phase, gas is released from thecircuit until the pressure has fallen to the set level ofPEEP; this is usually the pressure displayed on theventilator. If there is a delay in lung emptying, thepressure may fall more rapidly in the ventilatorcircuit than in the lung itself, so that the intra-alveolarpressure is higher than the displayed pressure. Thishappens when the absolute duration of the expiratoryphase is too short to permit full emptying of thelung, resulting in gas trapping. It typically occurswith global obstruction to expiratory flow, as inasthma, or with more localized delay in lung emptying,as may occur in acute lung injury. This is termedintrinsic PEEP (PEEPi) or autoPEEP and may con-tribute very significantly to the total PEEP. Indeed,ventilatory strategies that employ inverse ratiosmay have some of their effect on oxygenation viathis mechanism. PEEPi is measured by institutingan end-expiratory hold of sufficient duration toallow gas equilibration to occur with the ventilatorcircuit closed. This is usually of the order of 2-5 s,but up to 20 s may sometimes be required.

PEEPi may also exist as a dynamic phenomenon inpatients being ventilated with spontaneous modeswhen the patient attempts to trigger a new breathbefore lung emptying is complete. The quantifica-tion of dynamic PEEPi requires the use of anoesophageal balloon to provide an approximation ofpleural pressure. Dynamic PEEPi is then measured as

98 Monitoring

the negative pressure change in oesophageal pressurerequired in order to initiate inspiratory flow. If theamount of dynamic PEEPi is significant, it may makeventilator triggering and patient weaning muchmore difficult. The presence of such dynamic airtrapping can also usually be detected from visualinspection of flow-time or flow-volume curves,where expiratory flow has not ceased before the nextbreath starts in a spontaneously breathing subject.

Pressure, flow and volume diagrams

Modern, computerized ventilators frequently displayinformation about pressure, flow and volume graph-ically, providing extra information compared withpressure measurements alone. The curves most fre-quently displayed are either pressure-time andflow-time curves, or the pressure-volume and theflow-volume relationships.

PRESSURE-TIME AND FLOW-TIME DIAGRAMS

Although easier to obtain than pressure andflow-volume loops, these simple displays of pressureand flow over time provide much important infor-mation about the patient-ventilator interaction that

is readily accessible from simple visual inspection.A number of common phenomena are illustrated inFigures 8.4 to 8.9. Figure 8.4 illustrates the airway andalveolar pressure changes during volume-controlled(constant flow) ventilation. Note the differencebetween PAP and plateau pressure, representingalveolar pressure that occurs as a consequenceof airway resistance. Figure 8.5 shows the samephenomena during pressure-controlled ventilation(decelerating flow). Note that alveolar pressure takestime to reach the set level of inflation pressure and,if an end-inspiratory hold manoeuvre demonstratesthat this has not occurred, this suggests that theabsolute duration of the inspiratory phase is insuf-ficient. Figures 8.6 and 8.7 illustrate the presenceof PEEPi during volume and pressure-controlledventilation. Note particularly the failure of expiratoryflow to reach zero before the next breath commences.This phenomenon is still seen in spontaneousventilation modes where dynamic PEEPi is present(so-called dynamic hyperinflation). Figures 8.8 and8.9 demonstrate how the pressure-time diagram canbe used as a surrogate for a pressure-volumecurve during volume-controlled ventilation. Be-cause flow is constant, so is the rate of change involume, and time can therefore be regarded as asurrogate for volume. Thus, the graph becomes

Figure 8.4 Pressure and flow

waveforms during volume-controlled

ventilation. PIP, peak inflation

pressure; Ppl, pleural pressure;

Palv, alveolar pressure.



waveforms during pressure-controlled

ventilation. PIP, peak inflation

pressure; PEEP, positive end-

expiratory pressure; Palv, alveolar

pressure.

an inverted version of a pressure-volume trace,demonstrating the same changes in shape associatedwith improving or worsening compliance. Theslower the flow, the closer to a static curve this willbecome.

PRESSURE-VOLUME CURVES

The pressure-volume curve has become an impor-tant measurement for adjusting the ventilator set-tings in patients with severe lung injury. Although it



ventilation with static intrinsic PEEP.

Set PEEP, level of PEEP set on the

ventilator ('external' PEEP); PIP,

peak inflation pressure; Ppl pleural

pressure; Palv, alveolar pressure.

100 Monitoring


waveforms during pressure-controlled

ventilation with static intrinsic PEEP.

For abbreviations, see Fig. 8.6.

is still predominantly a research tool, the lessonslearnt from its use have become part of standardclinical strategies, and modern ventilators allowapproximation of the curve at the bedside. The real-ization that ventilator-induced lung injury is animportant component of the syndrome of acutelung injury and that strategies designed to mini-

mize this can result in a reduction in mortality inpatients with ARDS has led to more attention beingdevoted to ensuring that the lung is ventilated atoptimum volume. If lung recruitment is inadequateor not maintained, there is cyclical opening andcollapse of alveoli, resulting in progressive lungdamage, inflammation and cytokine release, which



ventilation demonstrating worsening

compliance (e.g. over-distension). PIP,

peak inflation pressure; Ppl, pleural

pressure.


Figure 8.9 Pressure and flow waveforms

during volume-controlled ventilation

demonstrating improving compliance (e.g.

recruitment). PIP, peak inflation pressure;

Ppl, pleural pressure.

produces secondary systemic damage. If the lung isover-distended, there will be progressive 'volu-trauma' and 'barotrauma'. Modern strategies are there-fore designed to ensure that the lung is recruitedand kept open, but that tidal ventilation does notlead to over-distension. The static pressure-volumecurve provides a tool with which to do this (Fig.8.10). Typically, the curve is sigmoid in shape inmechanically ventilated patients with lung injury.The lower portion represents the area of collapse,where static compliance is poor. The linear centreportion represents that lung volume at which lungsegments are open and compliance is good. The

Figure 8.10 Pressure-volume curve of the lung demonstrating

potential for lung damage with inadequate recruitment or

over-distension.

flattened upper portion represents the zone of over-distension, where compliance once again deterio-rates. Lower and upper inflection points separatethese different portions of the curve. Broadlyspeaking, modern ventilator strategy is based upondetermining where the lower inflection point is andsetting the PEEP level slightly above this to ensurethat tidal ventilation occurs within the region ofmaximum compliance. Similarly, the tidal volumeis set to ensure that the area of over-distension isavoided. Typically, this means a tidal volume of theorder of 6 mL kg~ l.

Unfortunately, measurement of the pressure-volume curve is not always straightforward. Thereare a number of approaches, of which the bestknown is probably the 'super-syringe' technique.5

Originally described using a 2-L syringe, with theinflation volume being measured by an electricalsignal proportional to the administered volume,this approach utilizes stepwise inflation of thelung. Volumes of between 50 mL and 200 mL areused and static pressure measurements are taken ateach step. The procedure should start from therelaxation volume of the lung (functional residualcapacity, FRC), especially if the lower inflectionpoint is to be established accurately. This requiresthe removal of PEEP and disconnection from theventilator, and the procedure may take severalminutes, all of which may result in considerable

102 Monitoring

Figure 8.11 Flow-volume loops demonstrating a normal pattern

(left) and the typical pattern seen with airflow limitation and dynamic

hyperinflation (right). COPD, chronic obstructive pulmonary disease.

The display seen on mechanical ventilators shows inspiration as the

upward deflection and the shape of this part of the flow-volume

loop will be influenced by the ventilatory mode employed.

instability and desaturation in patients with severelung injuries. This has resulted in other approachesalso being used. These include using different known-volume ventilator breaths and recording the staticend-inspiratory pressure after each, allowing a curveto be constructed from the results of a numberof breaths together. Another approach is to use theventilator to inflate the lung with a known volumeat a constant but very low flow. This minimizesthe resistive (dynamic) component of the inspira-tory phase and provides a good approximation of astatic inspiratory curve. The expiratory curve obtainedafter such a manoeuvre does not reflect staticconditions, although a variety of occlusion tech-niques applied during expiration have beendeveloped to provide a complete loop. This thenprovides information about the degree of hysteresis(if any) that is present.

FLOW-VOLUME CURVES

In mechanically ventilated patients, the ventilatorgenerates inspiratory flow and expiratory flow islargely passive. Consequently, it is virtually impos-sible to obtain reproducible forced expiratory flowmeasurements, but useful information can still beobtained from the shape of the flow-volume curve,especially about the presence of airflow limitation(Fig. 8.11). For patients with significant expiratoryflow limitation, the expiratory phase of the curvehas a characteristic concave shape. When there aresecretions in the airways (or water in the ventilatortubing), the curve has a typical spiky nature to it, asopposed to flow being smooth.

HAEMODYNAMICS AND THEMEASUREMENT OF LUNG WATER

Optimum haemodynamic management is a crucialcomponent of the treatment of critically ill patientswith respiratory impairment, especially those whorequire mechanical ventilation. Although it isbeyond the scope of this chapter to provide a detailedappraisal of haemodynamic monitoring techniques,some aspects deserve comment. A major challengein all critically ill patients is to achieve the correctbalance between giving enough fluid resuscitationto allow adequate cardiac performance and avoidingfluid overload. This applies especially to patientswith acute lung injury, because this condition is char-acterized by the development of non-cardiogenicpulmonary oedema.

Use of cardiac filling pressures

Traditionally, central venous pressure (CVP) and pul-monary artery occlusion pressure (PAOP) have beenused to assist clinicians with the assessment of vascu-lar filling. Unfortunately, there is often a significantdegree of confusion about the true value of thesemeasurements. The first is whether they are beingemployed to assess overall volume status (as surro-gates for total blood volume) or to assess cardiac fill-ing. More usually, it is the latter that is required, butboth CVP and PAOP are prone to considerable mea-surement error (zeroing and levelling the transducer,correct waveform analysis etc.), as well as being sig-nificantly altered by mechanical ventilation and the

Figure 8.12 During a high-pressure recruitment manoeuvre,

cardiac output and intrath'oracic blood volume (ITBV) fall, while

pulmonary artery occlusion pressure (PAOP) rises, reflecting the

rise in intrathoracic pressure. MAP, mean airway pressure.

Haemodynamics and the measurement of lung water 103

Figure 8.13 Principles of extravascular

lung water (EVLW) measurement by

indicator dilution. CV = central venous,

TD = thermodilution, RAEDV = right

atrial end-diastolic volume,

RVEDV = right ventricular end-diastolic

volume, PBV - pulmonary blood volume,

LAEDV = left atrial end-diastolic volume,

LVEDV = left ventricular end-diastolic

volume, ITBV = intrathoracic blood

volume, ITTV = intrathoracic thermal

volume

presence of significant cardiac or pulmonary disease.In a study of patients with ARDS, Lichtwarck-Aschoffand colleagues6 demonstrated the lack of agreementbetween changes in conventional filling pressures andchanges in cardiac index, demonstrating the unrelia-bility of this approach. Figure 8.12 shows the effect ofa high-pressure recruitment manoeuvre on PAOP,cardiac output and intrathoracic blood volume (seebelow), demonstrating how misleading filling pres-sures can be.

Nevertheless, as measurements in their own right,CVP and pulmonary artery pressures can providevery valuable information about the performance ofthe right heart under the extra stresses imposed bysevere lung injury (where acute pulmonary hyper-tension is common), and mechanical ventilationresults in high intrathoracic pressures.

Measurement of lung water

Because severe acute lung injury is characterized byincreased oedema formation, some authorities haveproposed that measurement of lung water should beincorporated into the definitions of acute lung injuryand ARDS in order to improve their robustness. Inaddition, quantification of the abnormal increase inlung water (i.e. the degree of oedema) would haveconsiderable potential advantages for clinicianswhen judging the balance of good versus harm dur-ing the volume resuscitation of critically ill patients.Unfortunately, chest radiography, which is still themost common approach to judging the degree ofpulmonary oedema, is neither specific nor quantita-tive. Developments in indicator dilution techniqueshave now made it possible to measure lung water at

the bedside, and this has been demonstrated to be ofconsiderable potential value in the monitoring ofpatients with acute lung injury.

Initially, the technique relied upon double indicatordilution, using chilled indocyanine green solution,which provided an intravascular dye indicator anda thermal indicator that also distributed into theextravascular space. The indicator was injected intothe right atrium via a central venous line and wasdetected via a femoral artery catheter. Cardiac outputwas calculated in the standard fashion from the ther-modilution wash-out curve, and then used in combi-nation with the transit times of the two indicators tocalculate their relative distribution volumes. The dis-tribution volume of the intravascular dye representsthe blood volume between the injection anddetection points (known as the intrathoracic bloodvolume, ITBV). The distribution of the thermalindicator (known as the intrathoracic thermalvolume, ITTV) contains the intrathoracic bloodvolume, but also includes other structures. Althoughthere is negligible distribution into vessel walls andcardiac muscle, the only major extravascular distrib-ution volume is in the lungs, because this is the onlymajor capillary bed between the injection and detec-tion points. The difference between the ITTV andthe ITBV is the extravascular thermal volume, whichis a good approximation to lung water and is there-fore known as the extravascular lung water (EVLW;Fig. 8.13). A further derivation of this technique hasmade it possible to obtain bedside measurements oflung water from thermodilution alone.7 Althoughnot quite as reliable as when the double indicatorapproach is used, the thermodilution method is suf-ficiently reliable in most clinical situations and isconsiderably more convenient and less expensive.

104 Monitoring

A further benefit of this monitoring approach hasbeen the recognition that ITBV, being a central bloodvolume closely related to cardiac filling and thereforepre-load, allows a volumetric approach to fluidresuscitation that is more robust than that using fill-ing pressures. Together, knowledge of cardiac output,ITBV and EVLW greatly facilitates the managementof haemodynamic instability in the context of severeacute lung injury, although only one study currentlyexists that demonstrates that management based onthe lung water approach translates into a beneficialoutcome for patients.8

CONCLUSION

This chapter provides only a brief overview of theprinciples behind a number of the standard monitor-ing techniques applied during respiratory criticalcare. Nevertheless, as monitoring devices becomemore advanced and more ubiquitous, it is crucial tohave some understanding of how they work in orderto take full advantage of the benefits they can provide,as well as to avoid errors of misinterpretation. Allmonitoring devices and techniques have limitationsthat can mislead the unwary user. Another importantconsideration is that there are often several differentways of obtaining the same information and somedevices are more robust and have lower running coststhan others. Considerable training is required ifmodern monitoring devices are to be used to theirfull potential. Decisions about which devices to pur-chase must therefore be broadly based and shouldbalance absolute capability and the ability to harnessthat capability within the complex infrastructure that

is a modern ICU. Respiratory critical care is depen-dent to a large degree on technological support of thecritically ill patient, and keeping abreast of thattechnology is a major challenge for the clinician.

REFERENCES

1 Clark, LC. Monitoring and control of blood and tissue

oxygen. Trans Am SocArtif Inten Organs 1956; 2: 41-8.

2. Rithalia, SV, Bennett, PJ, Tinker, J. The performance

characteristics of an intra-arterial oxygen electrode.


3. Sayah, AJ, Peacock, WF, Overton, DT. End-tidal C02

measurement in the detection of esophageal intuba-

tion during cardiac arrest. Ann Emerg Med 1990; 19:

857-60.4. Rithalia, SVS. Developments in transcutaneous blood

gas monitoring: a review. 7 Med Technol 1991; 15:143-53.

5. Matamis, D, Lemaire, F, Harf, A, et al. Total respiratory

pressure-volume curves in the adult respiratory

distress syndrome. Chest 1984; 86: 58-66.

6. Lightwarck-Aschoff, M, Zeravik, J, Pfeiffer, UJ. Intrathoracic

blood volume accurately reflects circulatory volume status

in critically ill patients with mechanical ventilation.

Intensive Care Med 1992; 18:142-7.

7. Pfeiffer, UJ, Lichtwarck-Aschoff, M, Beale, RJ. Singlethermodilution monitoring of global end-diastolic

volume, intrathoracic blood volume and extravascular

lung water. Clin Intensive Care 1994; 3 (Suppl.): 38-9.

8. Mitchell, JP, Schuller, D, Calandrino, FS, Schuster, DP.

Improved outcome based on fluid management in

critically ill patients requiring pulmonary artery

catheterization. Am Rev Respir Dis 1992; 145: 990-8.

9Respiratory emergencies I: medicalRICHARD M LEACH

IntroductionMassive haemoptysisAspiration syndromes

105 Large airways obstruction 116105 Neuromuscular, infective and endocrine respiratory-110 emergencies 119

References 122

INTRODUCTION

Respiratory emergencies are a common cause of car-diorespiratory collapse in the intensive care unit(ICU) patient. The management of haemodynamicinstability associated with sudden hypoxaemia is oftena greater challenge than that associated with primarycirculatory disturbances. Prompt identification of thecause and appropriate therapy are essential. Thischapter discusses those respiratory emergencies notdiscussed elsewhere in this book, specifically massivehaemoptysis, aspiration syndromes, acute largeairways obstructions and infective emergencies.

MASSIVE HAEMOPTYSIS

48% as moderate (<500 mL or 1-2 cups daily) andonly 14% as massive (>500 mL or more than 2 cups ofblood daily).5 Most cases of haemoptysis are due to in-fective causes (approximately 80%), including tuber-culosis, pneumonia, lung abscess and bronchiectasis,and only a minority are due to malignancy (approxi-mately 20%).1-6 The bronchial circulation is usuallythe source of bleeding. Death results from asphyxia,caused by flooding of the alveoli, and only rarelyfrom circulatory collapse.6 Mortality is directlyrelated to the rate and volume of blood loss and theunderlying pathology. In patients expectorating>600 mL of blood within a 4-hour period, themortality is reported to be 71%, compared to 45%in patients expectorating the same quantity over4-16 hours and 5% during 16-48 hours.7

Massive haemoptysis is a dramatic, life-threateningclinical emergency. Delayed or inappropriate treat-ment is common and, as a consequence, patientswith potentially treatable conditions die.1'2 It is usu-ally defined as the expectoration of 500-1000 mL ofblood in 24 hours or any life-threatening haemopty-sis in a patient with co-existing respiratory compro-mise.3'4 Massive haemoptysis accounts for less than20% of all episodes of haemoptysis. In a recentreview from a tertiary referral centre, 38% of caseswere classified as trivial (flecks of blood in sputum),

Clinical evaluation

A good history is essential and may provide valuableinformation regarding the cause of haemoptysis. Thecharacteristic clinical picture of diseases such astuberculosis, bronchiectasis and bronchogenic carcin-oma may direct subsequent investigation and man-agement. However, few patients are able to localizethe site of bleeding in the thorax. Chest examinationmay reveal localized crepitations or consolidation,but widespread soiling of the tracheobronchial tree,

106 Respiratory emergencies I: medical

due to coughing, often results in diffuse clinicalsigns. The examination of expectorated blood mayprovide clues. Food particles suggest the possibilityof haematemesis, but blood in the nasogastric aspir-ate does not differentiate between haematemesis andhaemoptysis as coughed-up blood is often swal-lowed. Purulent material in the sputum may indicatebronchiectasis or a lung abscess (Fig. 9.1) and micro-biology may isolate tubercle bacilli. Associatedhaematuria raises the possibility of an alveolarhaemorrhage syndrome.

Chest radiography (CXR) should be obtained inall patients (Fig. 9.2). It may provide important diag-nostic information, including evidence of a mass,cavity or abscess. However, the CXR may be unhelp-ful6 and potentially misleading as lesions seen on

CXR do not always correlate with the site of bleed-ing. Soiling (with diffuse alveolar shadowing) due tothe widespread distribution of the tracheobronchialblood during coughing and aspiration may furtherobscure the site and cause of bleeding.

Management of massive haemoptysis

Successful management requires a team approach,involving the intensivist, chest physician, radiologist,anaesthetist and surgeon. Management decisions areoften difficult and there is little evidence to directappropriate therapy, even in specific diseases such asbronchiectasis. The key aspects of management areas follows.

Figure 9.1 A 41-year-old man with lung abscess presenting with

haemoptysis, (a) Lung abscess on chest radiography, (b) Subsequent

development of empyema necessitans over left chest wall (see also

Plate 1). (c) Subsequent discharge of pus from abscess (see also Plate 2).

Massive haemoptysis 107

Figure 9.2 Sequence of four chest radiographs following massive haemoptysis due to tracheobronchial aspergillosis. (a) Onset of minor

bleeding in the right upper lobe with some minor 'soiling' of the right lower lobe, (b) Subsequent massive haemoptysis with collapse/

consolidation of right upper lobe at 12 hours, (c) 'White out' right lung at 24 hours, with the patient lying right side down to prevent soiling

of the 'good' (left) lung, (d) Resolution at 14 days.

1. Maintain a patent airway and ensure adequate 2. Promote drainage and prevent further alveolaroxygenation by providing supplemental oxygen, 'soiling', particularly of the unaffected lung, byif necessary by endotracheal intubation and positioning the patient slightly head down in themechanical ventilation: asphyxia is the greatest lateral decubitus position with the 'presumed'immediate risk to the patient. bleeding side down.


3. Determine the cause, site and severity of thebleeding: haematemesis and upper airways bleed-ing from the nose, pharynx or larynx may be con-fused with haemoptysis.

4. Excessive chest manipulation, including physio-therapy and spirometry, should be avoided as thismay increase or restart bleeding. Cough suppres-sion with codeine 30-60 mg every 6 hours may behelpful.

5. Institute appropriate therapeutic measures,including antibiotics and bronchodilators,depending on the underlying pathology and clin-ical circumstances.

Bronchoscopy is the most useful initial investiga-tion for identifying the source of bleeding and alsoallows bronchial toilet to be performed (Fig. 9.3). Itis, however, easy to be misled when there is wide-spread soiling of the bronchial tree, and over-zealoussuctioning of blood clots may encourage furtherbleeding. CXR, computer tomography (CT) andarteriography also have important roles in establish-ing the diagnosis and assessing progress.

The unaffected lung may have to be ventilatedindependently until bleeding can be controlled. Toachieve this, the endotracheal tube should be pos-itioned in the right or left main bronchus. In theemergency situation, the endotracheal tube isadvanced into the right main bronchus (which is inthe same axis as the trachea) until breath sounds canno longer be heard in the left side of the chest. If thepatient is not bleeding from the right side, the endo-

Figure 9.3 Blood cast of the left bronchial tree removed at

fibreoptic bronchoscopy in a 16-year-old girl with haemoptysis

due to meningococcal meningitis and disseminated intravascular

coagulation.

tracheal cuff can be inflated and the right lung select-ively ventilated. If bleeding is originating from theright lung, a Foley or Fogarty catheter should bepositioned in the right main bronchus and inflated.The endotracheal tube is then withdrawn until theleft lung is ventilated. Urgent bronchoscopy shouldbe arranged to establish the site and cause of bleed-ing and reposition the endobronchial tube asrequired. Double-lumen tubes (Carlens orRobertshaw) will also isolate the affected lung, butare difficult to maintain in position, require experi-enced personnel and have the serious limitation thatthe small lumens hinder suctioning and prevent flex-ible bronchoscopy. In the acute situation, the use ofdouble-lumen tubes provides little additional bene-fit. The use of positive end-expiratory pressure(PEEP) to increase intrathoracic pressure and tam-ponade the site of haemorrhage during mechanicalventilation is rarely helpful.

Determining the site and cause ofhaemoptysis

Once the patient has been adequately stabilized, thesite and cause of bleeding must be established.

• Early bronchoscopy is essential and is generallyaccepted to be superior to other diagnostic tech-niques, including bronchial arteriograms andCT.6'8 Rigid and flexible fibreoptic bronchoscopyhas been used with similar diagnostic yields andthe choice of scope depends on the experience ofthe admitting physician.2'9 If bleeding is consider-able and ongoing, rigid bronchoscopy in theoperating theatre provides better access for suc-tioning and the patient can be ventilated duringthe procedure. However, inspection is limited tothe large, lower airways. Flexible bronchoscopyhas the advantages of being readily available, it canbe passed through an endotracheal tube (8.0 mmdiameter or larger) and allows examination ofsubsegmental and upper lobe bronchi, whichaccount for 80% of bleeding sites. The disadvan-tage is limited suctioning capability compared torigid bronchoscopy. The flexible scope may bepassed through the rigid bronchoscope to providethe benefit of both methods.

• If bronchoscopy is unsuccessful, spiral CT withcontrast may detect the site of bleeding, tumours,

Massive haemoptysis 109

vascular malformations and other structuralabnormalities. A combination of bronchoscopyand CT scanning has the highest diagnosticyeild.1'2'6 Occasionally, radionucleotide scans maydefine the site of bleeding and the underlyingpathology. Transfer to the radiology or nuclearmedicine department requires that the patient isrelatively stable, and this may limit their use.

• If bleeding continues, bronchial arteriographyand occasionally pulmonary angiography areindicated (Fig. 9.4).

Control of haemoptysis

Establishing the best practice for the treatment ofmassive haemoptysis is limited by the paucity ofcomparative trials. In general, control of bleeding isusually required during on-going investigation andincludes immediate temporizing measures orbronchial embolization. This is followed by defini-tive surgery, if feasible, when the patient's conditionhas been stabilized.

IMMEDIATE

Initial control of haemoptysis is achieved by direct-ing boluses of iced saline with adrenaline (lOmL;1:10000 dilution) at the bleeding site through thebronchoscope.10 This simple technique is successful in95% of cases for the temporary control of bleeding.4

The application of topical thrombin to the bleeding

Figure 9.4 A 40-year-old man with haemoptysis from an old tuberculous cavity, (a) Bleeding into the cavity on radiography, (b) Angiogram

demonstrating multiple sites of bleeding within the cavity, (c) Angiogram after embolization and control of bleeding.


lesion is also effective in > 70% of patients.11

Alternatively, tamponade of the affected segmental orsubsegmental bronchus with a Fogarty or Foleycatheter is successful in > 80% of cases.12 Occasionally,intravenous vasoconstrictors (vasopressin, terlipressin)may be useful to reduce heavy bleeding.

BRONCHIAL ANGIOGRAPHY AND

EMBOLIZATION

Angiography has been proposed as the primary diag-nostic procedure in place of bronchoscopy (Fig. 9.4).However, controlled comparisons of the two tech-niques demonstrated that bronchoscopy was super-ior, identifying 68% compared to 55% of bleedingsites. On this basis, bronchoscopy remains the pri-mary diagnostic modality.13 However, bronchialartery embolization is the established therapeutictechnique for the control of haemoptysis.14 It isinitially successful in 70-100% of cases.14,15 Thebest results are described in patients with dilatedbronchial arteries (e.g. bronchiectasis). The frequencywith which re-bleeding occurs is disputed. Somereports describe 80% recurrence within 10 days,others 22% recurrence over 1-48 months.3'14

Recently published data described long-term controlof haemoptysis (>3 months) in 45% of cases stud-ied.15 Embolization is associated with serious com-plications. Infarction of the anterior spinal arterywith paraplegia has been reported in up to 5% ofcases.14 Other rare complications include ischaemicnecrosis of the bronchus and arterial dissection. Ingeneral, embolization should be regarded as atemporary or palliative manoeuvre, when surgery iscontraindicated or to stabilize a patient prior todefinitive surgical therapy.

MEDICAL VERSUS SURGICAL TREATMENT

There is still dispute as to whether patients with mas-sive haemoptysis should be treated conservatively(medically) or surgically. There are few comparativestudies, most are retrospective and few take intoaccount the aetiology of the bleeding. Most studiesagree that surgical therapy is associated with the bestlong-term outcomes for isolated lesions.9'16 In onesuch study, 5-year survival was 84% in surgicallytreated patients compared to 41% in medically treat-ed patients.16 Other studies report successful out-comes with surgery in 82-99% of patients.4'9'16

However, some respiratory physicians continue toadvocate a conservative approach and suggest thatsurgical intervention is only justified if haemoptysisremains uncontrolled following arterial emboliza-tion or in patients with recurrent bleeding.17 In tworecent conservative studies, where surgery wouldhave been feasible, the long-term success rates were46% and 68%, respectively4'7 Primary medical man-agement may be mandatory because bleeding cannotbe localized or is not amenable to the surgical resec-tion of a pulmonary segment (Fig. 9.5). In otherpatients, surgery will be contraindicated because ofend-stage lung disease (FEV1 < 40% predicted),poor cardiac reserve, unresectable cancer or severebleeding diathesis.

ASPIRATION SYNDROMES

To diagnose aspiration, a high index of suspicion isrequired in those at risk because it is often not wit-nessed.18 Volume and type of fluid aspirated are crit-ical and are usually related to the clinical scenario. Inthe peri-anaesthetic situation, the aspiration of largevolumes of gastric contents rapidly progresses toaspiration pneumonia and acute respiratory distresssyndrome (ARDS). In contrast, repeated micro-aspirations in the stroke patient with bulbar palsycause nosocomial pneumonia. Examination of pha-ryngeal and tracheal aspirates may be helpful andarterial blood gases are essential to monitor theseverity of aspiration syndromes.

High-risk groups for aspiration include thosewith:

• depressed levels of consciousness (head injury,drug overdose, epileptics and hypothermia),

• laryngeal incompetence (bulbar syndromes, cere-brovascular accident, myasthenia gravis,Guillain-Barre syndrome and multiple sclerosis),

• peri-operative, ICU and emergency roompatients.

Risk factors for aspiration in ICU patients include:

• supine posture,• nasogastric feeding,• gastrointestinal haemorrhage,• non-invasive ventilation: swallowed gas during

continuous positive airways pressure (CPAP) or

Aspiration syndromes 111

Figure 9.5 Tracheobronchial aspergillosis affecting the right upper lobe (a, b), right lower lobe (c, d) and left lower lobe (e, f) in a

25-year-old, immunosuppressed patient with systemic lupus erythematosus and massive haemoptysis (see also Plate 3).


non-invasive positive-pressure ventilation maycause vomiting and subsequent aspiration as air-way clearance is impeded by the tight-fittingmask,

• mechanically ventilated patients with endotra-cheal or tracheostomy tubes,

• gastric lavage.

Tracheo-oesophageal fistulae may present withrecurrent aspiration and most commonly are due totrauma or previously undiagnosed malignancy.

Solid particulate matter

A large variety of solid particulate matter may beaspirated, including confectionary, coins, teeth orrubber from balloons that burst while being inflated.However, the aspiration of partially masticated foodduring swallowing is the most common cause, givingrise to the 'cafe coronary'. Following the aspiration ofa large particle that completely occludes the larynx ortrachea, the subject is unable to speak or breathe andrapidly becomes cyanosed. If a sharp blow tothe back of the chest fails to dislodge the particle,the Heimlich manoeuvre should be attempted.The attendant stands behind the patient with his orher arms around the upper abdomen, just adjacentto the costal margin, and the hands clenched belowthe xiphoid process. The hands are pulled sharplybackwards, compressing the upper abdomen andlower costal margin. The sudden increase in thoracicpressure may dislodge the obstructing particle,which is exhaled by the patient. Continued obstruc-tion rapidly leads to coma and death and, as a lastresort, an emergency cricothyroidotomy should beattempted. This will only be successful if the obstruc-tion is at the level of the larynx. If available, a large-bore needle should be inserted through thecricothyroid membrane, which is palpable just belowthe thyroid cartilage. Alternatively, a knife (or othersharp implement), with the blade in the horizontalaxis, may be inserted at the same site and turned verti-cally to achieve an opening in the trachea. A pointed,hollow tube (straw, biro) may be equally effective inan emergency situation. Oxygen should be fed downthe needle or tube, if available. Urgent rigid bron-choscopy and/or thoracic surgery are required toremove the obstruction.

Partial tracheal or bronchial obstruction by foodor any other material causes symptoms common to

all foreign-body aspiration, including stridor, cough,wheeze and tachypnoea. A history of recurrent pneu-monia following an episode of aspiration is occa-sionally elicited and suggests that further explorationto recover an obstructing foreign body (e.g. apeanut) is necessary. Pneumonia, atelectasis andexpiratory emphysema are common radiologicalfindings in these circumstances. Small particles maybe aspirated with gastric contents and initially onlyproduce the characteristic inflammatory response ofliquid aspiration. In the inadequately sedated ICUpatient, endotracheal tubes may be 'bitten through'and the distal end of the tube aspirated.

Fluid aspiration

Gastric contents are the fluid most frequently aspi-rated.18'19 Water (during drowning) and blood (dur-ing pharyngeal operations or haematemesis) mayalso be inhaled. Aspiration after accidental or delib-erate self-poisonings with hydrocarbons, bleach andother toxic fluids usually occurs during vomiting orgastric lavage.

ASPIRATION OF GASTRIC CONTENTS

Aspiration of gastric contents or any other acidicfluid inhalation (pH < 2.5) is associated with severelung damage.18,19 The onset and severity of symp-toms depend on the volume of gastric contents aspi-rated. Large-volume to moderate-volume aspiration(e.g. inadequate gastric emptying before intubation)can result in the rapid development of respiratoryfailure, with tachypnoea, wheeze, cyanosis, hypoten-sion and hypoxaemia (Fig. 9.6). On occasions, frankpulmonary oedema or ARDS may develop. Abouttwo-thirds of aspirations are witnessed and the diag-nosis is confirmed by the detection of gastric con-tents at tracheal suctioning and the demonstration ofacidity by testing with litmus paper. In the remainingthird, successful diagnosis depends on a high indexof suspicion in high-risk patients. About 90% ofpatients who aspirate gastric contents will developpulmonary infiltrates on CXR. However, radio-graphic changes are often delayed for 12-24 hoursand initial films may be normal. The right mainbronchus is the most direct path for aspirated mater-ial and the right lower lobe is the most commonlyaffected area when only one lobe is involved (inapproximately 60% of cases). With large-volume


Figure 9.6 Rapid development of aspiration pneumonia in a

79-year-old woman after a straightforward sigmoid volvulus repair.

The patient vomited and aspirated gastric fluid immediately

following extubation. The chest radiographs are at (a) 1 hour,

(b) 3 hours and (c) 6 hours after the aspiration.

aspiration, diffuse bilateral infiltrates may beobserved. The development of ARDS secondary toaspiration pneumonia is associated with a high mor-tality (>80%).19

As treatment is ineffective, prevention is essential.When severe regurgitation occurs in the obtunded pat-ient, immediate management involves the following.

1. Clearance of the upper airways of vomit, fluid andobstructing particles: suction should always beavailable when dealing with high-risk patients.

2. Positioning of the patient head down in therecovery position to reduce further lung 'soiling'.

3. The immediate administration of O2.4. Tracheal intubation may be necessary to secure

the airway.

5. Neither tracheal suction nor bronchoscopy willprevent acid-induced lung injury, which is usuallyimmediate, but will clear excess free fluid andremove particulate matter from the airways.18'19

Respiratory support will depend on the degree oflung damage.

• Aerosolized bronchodilators reduce aspiration-induced bronchospasm.

• CPAP may improve oxygenation and avoid theneed for intubation in the alert, co-operativepatient with even relatively severe diffuse acute lunginjury. CPAP to 10 cmH2O maybe used, althoughabove this pressure the lower oesophageal sphinctertone may be exceeded, with the risk of furthervomiting and aspiration.


• Intubation and positive pressure ventilation willbe required if the patient is obtunded or non-inva-sive support fails to maintain adequate oxygena-tion. Alveolar recruitment with high PEEP orprolonged inspiratory to expiratory ratios is oftennecessary.

• Antibiotic therapy is often instituted followinggastric aspiration, although there is little evi-dence to support this practice. Traditionally,therapy should cover the Gram-positive andGram-negative organisms colonizing the upperairways and the upper gastrointestinal tractanaerobes.

• Early steroid therapy is not beneficial.20

Smaller-volume, recurrent aspiration of gastriccontents is a serious cause of morbidity and mortal-ity in the ICU and one of the main factors in thedevelopment of ventilator-associated pneumo-nia.21'22 This small-volume aspiration often occurspast inflated endotracheal or tracheostomy cuffs.The onset is slower and the symptoms more subtle,but the end result may be severe lung damage and ahigh mortality. Micro-aspiration in tube-fed patientsmay be detected by adding dye to the feed andobserving for its appearance in tracheobronchialsecretions or by using glucose oxidase reagent stripsto test for glucose in tracheobronchial secretions,although, if negative, these tests do not necessarilyexclude the diagnosis.

The management strategies used to reduce theincidence of small-volume aspiration in the ICUinclude:21

• nursing in the semi-recumbent position (30-45%head-up),22

• preventing gastric microbial overgrowth causedby stress ulcer prophylaxis, by early enteral feed-ing or sucrulfate, although further evidence-baseddata are required.

• Subglottic drainage did not reduce the incidenceof Pseudomonas or Enterobacteriaceae respiratorytract colonization or infection in clinical studies.

NEAR-DROWNING AND WATER ASPIRATION

Aspiration of water and cases of near-drowning maypresent to any ICU. They are one of the commonestcauses of accidental death in children and adults andare often associated with alcohol consumption.Although more frequent in areas with close proxim-

ity to natural water sources, deaths may occur in anyhome and in relatively small volumes of water (e.g. ashallow bath). Near-drowning is often secondary to aprimary medical event, such as a myocardial infarc-tion, occurring whilst the subject is in water. In thissituation, the patient becomes suddenly motionlessin the water, with little or no struggling.

The amount and type of water aspirated deter-mine the severity and pathophysiology of the pul-monary lesion.23,24 About 10% of near-drowningpatients do not aspirate due to laryngospasm andbreath holding.24

• Freshwater aspiration affects pulmonary surfac-tant and subsequent atelectasis results in pul-monary shunt, venous admixture andhypoxaemia.24 Freshwater in alveoli is rapidlyabsorbed into the pulmonary circulation, and inlarge volumes may cause initial hypervolaemia.However, the patient may be hypovolaemic by thetime he or she reaches hospital due to fluid redis-tribution and the development of pulmonaryoedema. If sufficient freshwater is absorbed, theplasma becomes hypotonic, causing intravascularhaemolysis. The resulting increase in plasmapotassium and free haemoglobin has significanteffects on the heart and kidneys.

• Seawater aspiration results in fluid-filled, perfusedalveoli with shunting and venous admixture. Thehypertonic sea water pulls fluid from the plasmainto the lungs and may cause rapid hypovolaemia.

Mortality in freshwater and sea-water drowning issimilar.23,24 Those who survive the initial, short-livedpulmonary insult associated with water aspirationusually have no long-term respiratory impairment.In animal experiments, aspiration of as little as2.5 mL kg-1 of water may increase the intrapul-monary shunt from the normal value of 10% to 75%(PaO2 < 8 kPa) within 3 min. Typically, the respira-tory symptoms (tachypnoea, wheeze and cyanosis)and the pulmonary oedema depend on the amountof water aspirated, the level of contamination by pol-lutants and the risk of secondary infection fromsewage and other sources of infective organisms.Pulmonary oedema may be delayed by up to 12hours after near-drowning incidents. For this reason,it is generally recommended that all conscious andasymptomatic cases of near-drowning should beadmitted to a high dependency unit (HDU) for12-24 hours' observation before discharge.


By reducing cerebral metabolism, profoundhypothermia (<30 °C) can prevent irreversibleneurological damage, particularly in children,enabling them to survive protracted periods ofasphyxia. However, cold-water immersion andhypothermia are major factors in the pathophysiologyof near-drowning. Respiratory drive is inverselyrelated to water temperature and, below 10 °C,uncontrollable hyperventilation increases minuteventilation tenfold and breath hold times arereduced to less than 10 s. This results in anincreased risk of aspiration during escape from asubmerged vehicle or when swimming in turbulentwaters. Body temperatures <28 °C impair neuro-muscular performance, making swimming difficult,and the conduction advantage of Purkinje tissueover normal ventricular tissue is lost. Rough han-dling of the hypothermic patient easily precipitatesventricular fibrillation.

The primary aims during the initial treatment ofnear-drowning25'26 are to:

• reverse hypoxaemia,• restore cardiovascular stability,• prevent further heat loss,• correct acidosis and electrolyte imbalance,• prevent hypoxaemic brain injury.

At the scene of the accident, the victim must beretrieved from the water and cardiopulmonaryresuscitation initiated as soon as possible. TheAmerican Heart Association does not recommendroutine abdominal-thrust (Heimlich) manoeuvresto aid the drainage of fluid from the lungs becausecontrolled studies demonstrated no benefit.25

Gravitational drainage is equally effective and littlewater can be aspirated by direct suction within5 min. In addition, arrhythmias or asystole may beinduced, which are extremely difficult to reverse inthe hypothermic patient. Gastric dilatation is oftenassociated with near-drowning. Vomiting and aspi-ration may follow sudden increases in upper abdom-inal pressure.25,26

Victims who appear normal on arrival at hospitalcan deteriorate rapidly. Full resuscitative effortsshould be attempted in all near-drowned patientsand successful outcomes after several hours of man-ual resuscitation have been reported.25,26 Hypo-thermic cases must be re-warmed before any decisionis made to terminate resuscitation, because recovery

after prolonged, cold submersion has been reported.Hypothermia may be associated with resistant arrhyth-mias. If initial attempts at defibrillation are unsuccess-ful, they should be discontinued until the core bodytemperature is >29 °C. Re-warming techniques shouldaim to avoid shivering, which increases O2 demand.Cardiovascularly stable patients should be rewarmedat 1 °C h- l . This can be achieved using warm, humid-ified, inspired gas, warm intravenous fluids and warm-ing blankets. Rapid re-warming should be consideredwhen the core temperature is < 28 °C because of therisk of ventricular arrhythmias. Extracorporealrewarming with haemofiltration or cardiopul-monary bypass rapidly restores normothermia (upto 10.7 °C h"1) and allows fluid removal in the pres-ence of pulmonary oedema. When bypass is used,perfusion is re-instituted, regardless of the cardiacrhythm. Other techniques of rapid re-warminginclude peritoneal dialysis, bladder irrigation, gastricand pleural lavage.

Fluid resuscitation is essential and electrolyteimbalance, although infrequent, must be corrected.O2 therapy should be given until hypoxaemiaresolves. A normal CXR is present in up to 20% ofcases at admission to hospital, although, in themajority, changes ranging from lobar infiltrates tobilateral pulmonary oedema are observed.Controlled studies have verified the effectiveness ofCPAP, with either spontaneous or mechanical venti-lation for hypoxaemia due to sea-water aspiration. Infresh-water aspiration, CPAP is more effective whencombined with mechanical ventilation, which mayreflect alterations in pulmonary surfactant afterfresh-water aspiration.23'24 Antibiotics should bewithheld unless there is specific evidence of infectionor the aspirated water was grossly contaminated.Although chest infections are most common, bacter-aemia and brain abscesses have been reportedfollowing aspiration.

The degree of hypoxic brain injury often deter-mines the outcome, but neither intracranial pres-sure monitoring nor corticosteroid therapy isbeneficial.23 Techniques including deliberatehypothermia and barbituate-induced coma do notimprove survival and are not recommended.Prolonged submersion, delayed resuscitation, severeacidosis (pH <7.1), fixed, dilated pupils and a lowGlasgow Coma Score (<5) are usually associatedwith death and brain injury, although none of thesepredictors is infallible.


HYDROCARBON ASPIRATION

The ingestion of petrol, furniture polish and otherhydrocarbons accounts for about 15% of acciden-tal poisonings in children. Pulmonary toxicity onlyoccurs if there is inhalation during ingestion orfollowing subsequent vomiting. The characteristicodour of hydrocarbon is usually readily detectedand initially the patient experiences a burning sen-sation in the mouth and oropharynx. Centralnervous system irritability with lethargy, dizziness,twitching and, less commonly, convulsions mayoccur. Non-infective fever is common. Labouredrespiration and cyanosis follow symptoms ofcough and choking. Progressive respiratory failurewith hypoxaemia develops over the next 24 hours.Large aspirations may be associated with pul-monary oedema and haemoptysis. Patients withsmall aspirations usually recover over 2-5 days.Gastric lavage or induced emesis following theingestion of hydrocarbons is not recommendedbecause pulmonary aspiration, rather than gas-trointestinal absorption of hydrocarbon, is the life-threatening event.

BLOOD ASPIRATION

Inhalation of blood may occur during haemateme-sis, intrapulmonary haemorrhage and surgery onthe upper airways. Aspiration of blood mimics theacute phase of acid gastric content inhalation.Respiratory distress is associated with cyanosis ifsufficient blood is inhaled to cause intrapulmonaryshunting. However, unlike acid gastric contentinhalation, these symptoms usually settle rapidly,with few long-term sequelae. Blood in the upperairways may also precipitate severe laryngospasmand this is a common and serious complication ofupper airways surgery (see 'Large airways obstruc-tion', below).

TRACHEO-OESOPHAGEAL FISTULA

Recurrent occult aspiration may be due to trans-oesophageal fistula (TOF) caused by tumour,trauma or mediastinal sepsis. The diagnosis can bedifficult and may require bronchoscopy,oesophagoscopy, gastrograffin swallow and CTscan. The management will be determined by theunderlying cause.

LARGE AIRWAYS OBSTRUCTION

In the critically ill patient, large airways obstructionis a common and life-threatening emergency.27 Itusually presents as sudden hypoxaemia and, if notrapidly corrected, cardiovascular instability andeventually cardiac arrest ensue. The causes of largeairways obstruction include sputum plugs, bloodclots, aspirated particulate matter, inhaled toxicgases, burns, trauma, anaphylactic attacks, laryngealangioedema, severe laryngospasm and tracheal orother large airways stenoses. Rarely, childhoodinfections, including epiglottitis or diphtheria, maycause acute laryngeal obstruction (see below).Obstruction of the airways by the tongue in theunconscious patient should always be excluded andprevented with a pharyngeal airway. In the ICU,acute upper airways obstruction is often associatedwith the process of extubation or the complicationsassociated with tracheostomy. Extubations shouldbe undertaken with considerable care. Facilities forO2 therapy, suction, nebulizers, re-intubation andnon-invasive respiratory support should be im-mediately available.

Sputum plugs and blood clots

The commonest cause of large airways obstructionin the ICU is sputum plugs, which are often dis-lodged during physiotherapy. CXR reveals a col-lapsed lobe or segment. Physiotherapy with'bagging' and gentle lavage and suctioning is usual-ly sufficient to dislodge the obstruction, althoughbronchoscopy may be required to remove a viscidplug. Blood clots due to tracheal traumaduring suctioning or following the insertion of atracheostomy may result in obstructions. Viscidsputum or blood clots may also act as ball valves atthe lower end of the tracheostomy or endotrachealtube, causing respiratory impairment and lunghyperinflation (Fig. 9.7).

Laryngospasm

Laryngospasm is an important but uncommon com-plication in the ICU. Irritation of the upper airways byblood, pharyngeal secretions, food or aspirated gastric

Plate 1 A 41-year-old man with lung abscess presenting with

haemoptysis. Subsequent development of empyema necessitans over

left chest wall.

Plate 2 A 41-year-old man with lung abscess presenting with

haemoptysis. Subsequent discharge of pus from abscess.

Plate 3 Tmcheobronchial aspergillosis affecting the right upper lobe (a, b), right lower lobe (c, d) and left lower lobe (e, f) in a 25-year-old,

immunosup pressed patient with systemic lupus erythematosus and massive haemoptysis.

Plate 4 Sudden, severe hypoxaemia and airways obstruction

developed in a 26-year-old man 3 weeks following an episode of

blunt chest trauma. A large blood clot was found occluding the

tracheostomy at flexible bronchoscopy.

Plate 5 A 27-year-old man sustained a severe blunt chest injury

following a road traffic accident. His CXR and CT scan showed the

heart to be displaced to the right. At thoracotomy, he was found to

have a ruptured pericardial sac through which the heart had

hernia ted.

Plate 6 Magnetic resonance imaging scan: bronchoscopic

appearance looking down from the vocal cords. The stenosis

developed 2 months after the insertion of a surgical tracheostomy

in a 16-year-old boy ventilated for Guillain-Barre syndrome. Major

tracheal surgery was required to repair the defect.

Plate 7 Endoscopic view of thoracoscopic talc poudrage. Note the

homogeneous distribution of talc powder across the entire left

pariet al pleura, adhering lung upper lobe on the left upper margin,

free lower lobe at the left lower margin.

Large airways obstruction 117

Figure 9.7 Sudden, severe hypoxaemia and airways obstruction

developed in a 26-year-old man 3 weeks following an episode of

blunt chest trauma. A large blood clot was found occluding the

tracheostomy at flexible bronchoscopy (see also Plate 4).

contents can cause severe laryngeal spasm, presentingwith stridor and respiratory distress. In the ICU,laryngospasm usually occurs immediately after extu-bation due to the oedema and stimulation of theupper airways associated with the removal of theendotracheal tube. In many cases, there may be bleed-ing in the pharynx due to trauma associated with theinsertion of nasogastric tubes or following recentsurgery. Severe respiratory distress requires immediatere-intubation. If the clinical situation is less critical,nebulized adrenaline (2-3 mL 1:1000 adrenaline),salbutamol and intravenous hydrocortisone (200 mgintravenously) may be administered to reduce laryn-geal oedema, spasm and obstruction sufficiently toavoid re-intubation. Helium and O2 mixes (Heliox)are often recommended as a temporary measure toimprove gas flow through upper airways obstructions.

Anaphylaxis and angioedema

The life-threatening anaphylactic response of asensitized human appears within minutes of theadministration of a specific antigen and usuallypresents as respiratory distress and cardiovascularcollapse. Cutaneous manifestations include pruri-tis, urticaria and angioedema. A variety of materialsmay precipitate an attack, including foods, pollen,hormones and enzymes. In the ICU, drugs are thecommonest cause, particularly antibiotics and con-trast agents. Urticaria and angioedema may be asso-ciated with a variety of autoimmune or vasculiticdiseases. Rarely, angioedema is a hereditary condi-tion due to an autosomal dominant deficiency ofCl inhibitor, and it may also be acquired in certainlymphoproliferative disorders. The manifestationsof anaphylaxis are attributed to the release of hista-mine, and levels of IgE are raised. Laryngeal oed-ema presents with hoarseness and stridor andbronchial obstruction with chest tightness andwheeze. The urticarial eruptions are intensely pru-ritic, may coalesce to form giant hives, but seldompersist for longer than 48 hours. A deeper, oedema-tous, cutaneous process - angioedema - may alsobe present. Angioedema of the larynx, epiglottis orupper trachea may cause severe obstruction anddeath.

The immediate treament of anaphylaxis includesintravenous adrenaline (0.5-1 mg), hydrocortisone(200 mg) and diphenhydramine hydrochloride(20 mg). Fluid resuscitation for hypotensive shockmay be required. Bronchospasm is treated with O2

therapy, nebulized salbutamol and, occasionally,intravenous aminophylline. Upper airways obstruc-tion may require immediate intubation, but, if thesituation is less critical, nebulized adrenaline mayrelieve largyngeal oedema and avoid the need forintubation. If intubation is not possible due to laryn-geal deformity or oedema, an emergency cricothy-roidotomy should be performed.

Trauma, toxic inhalational injury andburns

Inhaled toxins (sulphur dioxide, SO2) and aspiratedchemicals (bleach, acid, hydrocarbons) can causeinjury, oedema and eventually obstruction to theupper airways.28 Similarly, burns' victims who are


exposed to very high temperatures may developserious injuries caused by the hot inspired gases.29

Severe laryngeal or upper tracheal trauma mayobstruct the upper airways. Immediate tracheostomymay be necessary in all these situations.

Many inhaled chemicals cause injury to the airwayand alveolar epithelium (Table 9.1).28 In general,chemicals are either direct respiratory toxins (NO2>SO2) or agents with systemic effects (CO, HCN).Toxic inhalational injuries depend on the propertiesof the inhaled gas, its concentration and the rate anddepth of ventilation.

• Penetration of toxic particles is dependent on par-ticle size, airways anatomy and breathing pattern.Particles > 10 |xm are efficiently filtered by thenose, those between 2 and 10 |o,m are deposited inthe tracheobronchial tree, and those <2 jxm aredeposited in the alveoli.

• Exercise and physical work increase the penetra-tion of toxic gases into the lung and the 'totaldose' of toxin.

• Highly soluble gases (such as SO2) cause upperairways damage and oedema with localized

obstructive symptoms, whereas less-soluble gases(such as NO2) penetrate deeply and causeparenchymal damage. Poorly soluble gases (e.g.CO) diffuse across the alveolar capillary mem-brane into the pulmonary circulation and maysubsequently damage distant tissues.30

The clinical manifestations and timing of symp-toms are dependent on the toxin, but include upperairways inflammation and oedema, laryngospasm,bronchospasm and pulmonary oedema. Whereasrapid death may occur with CO, delayed pulmonaryoedema may occur at 24-48 hours with NO2 orbronchiolitis obliterans at 2-7 days with smokeinhalation.

NITROGEN DIOXIDE

NO2 causes peroxidation of lipid cellular mem-branes and subsequent tissue damage. Exposureoccurs during welding, mining, fire fighting and infarmers. Gas stoves produce NO2 and, when usedin unventilated areas, may give rise to respiratoryillness. Delayed pulmonary oedema and ARDS are

Table 9.1 Inhalational toxins and sites of injury

Direct respiratory toxins

Agents with systemic effect

Acet aldehydeAcroleinAmmoniaBromineChlorineHydrogen chlorideIsocyanatesNitrogen oxidesOzonePhosgeneSulphur dioxideCadmiumManganeseMercuryNickel

Carbon monoxideHydrogen cyanideHydrogen sulphideMethane

Upper airway irritationUpper airway irritationUpper airway irritationARDSUpper and lower airway damageUpper and lower airway damageIncreased airways responsivenessAirway responsiveness, ARDS, bronchiolitis obliteransAirway responsiveness plus inflammation, ARDSLower airway injury, ARDSAirway responsiveness, ARDS, bronchiolitis obliteransEmphysema, ARDSMet al fume feverARDSAsthma

Tissue hypoxiaTissue hypoxia, lactic acidosisTissue hypoxiaAsphyxia

ARDS, acute respiratory distress syndrome.

Neuromuscular, infective and endocrine respiratory emergencies 119

the commonest clinical consequences of exposure.The concentration of gas is more important thanthe duration of exposure. In silo-filler's lung, NO2

produced by grain fermentation can reach veryhigh concentrations (200-2000 p.p.m.; safe range<8 p.p.m.) and, because NO2 is heavier than air, itcollects at the bottom of the silo. A single breathmay be fatal and reported mortality varies between20% and 65%.

CARBON MONOXIDE

CO poisoning is the commonest cause of acute fatalpoisoning (3800 deaths per year in the USA). It is acolourless, odourless and tasteless gas, which isnon-toxic to the lungs but displaces O2 fromhaemoglobin and causes damage by reducing O2

delivery to the tissues. Its affinity for haemoglobinis 220 times greater than that of O2. By the Haldaneequation, 0.1% CO in air would result in 50% carb-oxyhaemoglobin (COHb). Tissues with a high O2

demand, including the brain and myocardium, aresusceptible to CO poisoning. In the normal individ-ual, breakdown of haemoglobin gives rise to aCOHb level of less than 1%, although acutehaemolysis can increase this level to 4-6%.Cigarette smoke contains 400 p.p.m. CO and cigar-ette smokers have COHb levels between 3% and9%. CO poisoning results from poorly ventilatedheating devices and the use of internal combustionengines in enclosed spaces. The clinical symptomsdepend on the level of COHb:

• 10-20%: dizziness, headache,• 20-30%: chest pain, reduced vision,• 30-40%: nausea, vomiting, severe headache,• 40-50%: confusion, ataxia, tachycardia,• 50-60%: stupor, convulsions, coma,• >60%: coma, death.

In general, patients who do not lose conscious-ness recover without permanent sequelae.Coma may be associated with basal ganglia orcerebellar damage, altered personality, memory loss,neuropsychiatric disorders and epilepsy. Treatmentwith high O2 concentrations is essential. Thehalf-life of COHb in room air is 240 min, comparedto 30 min with 100% O2 therapy. O2 therapy shouldbe continued until the COHb level falls below 7%.Hyperbaric O2 therapy (100% O2 at 3 atmospheres)reduces the half-life to <20 min. Although it has not

been found to have any major benefit over 100% O2

therapy, there is some evidence that it may reducethe long-term neurological sequelae. Limitedavailability and the time required for transport limitits practical use.

SMOKE INHALATION

Respiratory failure is the leading cause of death fol-lowing burn injuries. The chemical composition ofsmoke depends on the material being burnt, butincludes aldehydes, acrolein, ammonia, CO, HCN,NO2, SO2 and phosgene. Animal studies demonstratethat lung dysfunction is proportional to the mass ofsmoke inhaled and is independent of the tidal vol-ume. Smoke directly injures the lung epithelial lining,causing pulmonary oedema, which is exacerbated byfluid therapy for the cutaneous burns. Early inactiva-tion of surfactant leads to atelectasis and V/Q mis-match. Smoke inhalation is often associated with COpoisoning, burns to the upper airways and severelaryngospasm. The clinical features of upper airwaysobstruction following smoke inhalation can bedelayed for several hours. Late-onset bronchiolitisobliterans may develop. Cyanide poisoning followingsmoke inhalation has been implicated as a cause ofdeath and the use of sodium nitrite and thiosulphatetherapy has been recommended.

NEUROMUSCULAR, INFECTIVE ANDENDOCRINE RESPIRATORY EMERGENCIES

There is an extensive list of neuromuscular, infectiveand endocrine diseases that may predispose torespiratory compromise and result in rapid respira-tory impairment. Several factors may be present inthe same patient.

• Central respiratory depression: coma, cerebrovas-cular accident, sedation during intubation orpostoperatively.

• Self-poisoning: sedative drugs (benzodiazepines,neuroleptics) or neuromuscular poisons (strych-nine) often result in progressive loss of respiratorydrive and respiratory failure.

• Neurological disease: myasthenia gravis, ascend-ing polyneuritis (Guillain-Barre syndrome) andrestrictive chest-wall defects (kyphoscoliosis) arefrequently referred to the ICU for monitoring of


respiratory function because of the potential riskof sudden respiratory failure. Progressive respira-tory muscle weakness with inadequate cough maycomplicate myopathies or motorneuron disease.Atelectasis due to loss of mobility or laryngealincompetence with aspiraton may occur.

• Infections: poliomyelitis, botulism, meningitis,miliary tuberculosis and tetenus may present asrespiratory emergencies. ARDS complicating gen-eralized sepsis is a frequent cause of respiratoryfailure on the general ward.

• Endocrine disorders: thyroid, adrenal and pituitarygland disorders may be complicated by respiratoryfailure due to central respiratory depression, mus-cle weakness or electrolyte imbalance.

Discussion of all the potential causes of respiratoryimpairment is not appropriate here, but thoseinfective and neurological disorders that may resultin rapid and potentially life-threatening respiratoryfailure and presentation to the ICU are reviewed inthis section.

BOTULISM

Botulism is a disorder of the myoneural junctioncaused by a Gram-positive anaerobic bacterium,Clostridium botulinum. Potent neurotoxins bindirreversibly to the presynaptic junction, preventingthe release of acetylcholine. Symptoms develop with-in 1-24 hours of exposure to the neurotoxin, usuallyfollowing ingestion of affected food or releasefrom infected tissue. Decreased vital capacity,increased residual volume and hypoxaemia parallelsevere respiratory muscle involvement and theneed for mechanical ventilation. About 30% ofpatients develop respiratory failure severe enough torequire mechanical ventilation. Aspiration is common.Therapy requires elimination of the neurotoxinfrom the gastrointestinal tract by gastric lavage andenemas, administration of trivalent antitoxin toneutralize circulating serum neurotoxin and of high-dose penicillin (3 MU 4 hourly intravenously) to erad-icate the C. botulinium organisms and preventfurther release of neurotoxin. Surgical debridementof infected wounds may be needed in some cases.

Infective respiratory emergencies

Although infective respiratory emergencies are rarein developed countries with advanced immunizationand public health programmes, they remain a seri-ous and relatively common cause for ICU admissionin Third World countries.31 The most seriousinclude the following.

POLIOMYELITIS

Flaccid paralysis and respiratory failure may developwithin hours of polio infection. Muscle weakness isusually asymmetric, widely distributed and initiallyinvolves the lower limbs and trunk. Upper-cord andbrainstem lesions are associated with loss ofdiaphragmatic function and respiratory failure.Bulbar lesions cause speech and swallowing difficul-ties, with the risk of aspiration and secondary pneu-monia. In a small proportion of cases, acutelaryngeal obstruction or rapid respiratory failuremay require urgent intubation. Therapy is support-ive and, except in areas of complete paralysis, motorfunction recovers in the majority of patients.Persisting muscle weakness or paralysis may necessi-tate long-term ventilatory support.

TETANUS

Tetanus is a toxin-mediated disease caused byClostridium tetani. Usually, it presents as the gen-eralized form, with diffuse muscle rigidity, general-ized muscle spasms, respiratory failure andcardiovascular instability. Localized presentations,with rigidity around the site of injury, or thecephalic variant with trismus, dysphagia and paraly-sis of cranial nerves may occur. Both progress tothe generalized form in 65% of cases. Appropriatemanagement of the airways is the first priority inthe treatment of tetanus. Patients with diffuserigidity, generalized spasms or who are at risk ofaspiration due to dysphagia should be intubated,even in the absence of respiratory compromise. Anon-depolarizing neuromuscular blocking agentshould be used to facilitate paralysis for intubationas depolarizing agents such as succinylcholine maycause hyperkalaemia and cardiac arrest. Muscularrigidity should be managed with high doses ofintravenous benzodiazepine or morphine therapy.Paralysis with agents such as pancuronium,atracurium and vecuronium may be necessary toprevent spasms. Survival is improved by aggressivesurgical debridement of necrotic wounds.Intramuscular human tetanus immune globulin

Neuromuscular, infective and endocrine respiratory emergencies 121

(3000-6000 IU) should be given as soon as possibleand before any surgical debridement, which maybe associated with the increased release of toxin. Ifthe patient has not progressed to generalizedspasms, intrathecal antitoxin (250 IU) may bemore effective and is reported to reduce mortality.Antibiotics (penicillin, teracycline, chlorampheni-col) are given to eradicate C. tetani and preventfurther toxin formation. Cardiovascular instabilityis best controlled with deep sedation, but high-dose magnesium therapy may be required, withcareful monitoring of the serum calcium levels andappropriate supplementation. Autonomic dysfunc-tion is usually well controlled with bolus morphinetherapy, although epidural bupivacaine may alsobe helpful. A successful outcome depends onmeticulous intensive nursing care, which may benecessary for 6 weeks or more. If recovery occurs,immunity to tetanus is not guaranteed andprimary immunization must start before thepatient leaves hospital.

node biopsies or lumbar puncture may establishthe diagnosis. Transbronchial biopsies in themechanically ventilated patient have a relativelyhigh risk of pneumothorax, and tuberculin skintesting is not often helpful in ICU patients. If AFBare detected, the species of Mycobacterium shouldbe confirmed, the HIV status considered and, inthe patient unable to give consent, the ethicalguidelines regarding HIV testing should beaddressed.

In acute respiratory failure due to tuberculosis,steroid therapy for 1-6 weeks is recommended, inconjunction with standard antimycobacterial drugs,unless there is a strong suspicion of co-existing bac-terial infection. Steroids potentiate the resolution oftuberculous pneumonia, decrease exudative reac-tions and reduce systemic toxicity. In addition,steroids address the adrenocortical deficiency andhepatic enzyme induction associated with antituber-culous medication, which result in low levels ofendogenous steroid.

TUBERCULOSIS

In the majority of cases of both pulmonary andmiliary tuberculosis, the presentation is slowlyprogressive. However, acute respiratory failure canoccur when a large quantity of infected pus rup-tures into the bronchial tree, or the vasculature,causing acute miliary tuberculosis. ARDS and dis-seminated intravascular coagulation can rapidlydevelop in these patients and signal a poor prog-nosis. Massive haemoptysis from active or healeddisease may also result in aspiration and acute res-piratory failure. The diagnosis is often missed inthe ICU because of failure to consider advancedfibrocaseous or miliary tuberculosis in a patientpresenting with acute respiratory failure.Assessment of potential risk factors and a highindex of suspicion are essential. Cavitation on CXRis often suggestive, but both pulmonary and mil-iary tuberculosis may present with radiographicchanges that may be confused with pneumonicconsolidation or ARDS. A relatively normal whitecell count should favour the diagnosis of tubercu-losis. The detection of acid-fast bacilli (AFB) isdiagnostic in pulmonary tuberculosis, but repeatedsputum smears and bronchioalveolar lavage sam-ples may be negative in miliary tuberculosis. In thissituation, bone marrow, pleural, liver and lymph

DIPHTHERIA

Diphtheria causes acute respiratory failure due tolaryngeal obstruction or as a consequence of respira-tory muscle failure due to polyneuritis. Severalother pharyngeal and laryngeal infections, includ-ing epiglottitis, and peri-tonsillar abscesses (quinsy)may cause acute upper airways obstruction. Dip-htheria is characterized by a local inflammatory lesionin the upper respiratory tract and remote effectsresulting from the release of a toxin, which affects theheart and nerves. The characteristic pharyngeal mem-brane of diphtheria may spread over the posteriorpharyngeal wall into the larynx, trachea and, lesscommonly, the bronchial tree. Bronchopulmonarydiphtheria has a high mortality, not only becauseof the risk of obstruction, but also because of thelarge surface area from which the toxin can beabsorbed.

Neuromuscular disorders

Many long-standing neurological diseases, includingthe myopathies, multiple sclerosis, motorneuron dis-ease and post-polio syndrome, are associated withintermittent respiratory emergencies (muscle weak-ness, atelectasis, sputum retention). Acute ascending


polyneuritis and myasthenia gravis are commonneurological diseases that can cause a previouslyhealthy person to develop paralysis and respiratoryfailure over the course of a few days.

GUILLAIN-BARRE SYNDROME (ACUTE

ASCENDING POLYNEURITIS)

The most immediate threat to the patient is thedevelopment of respiratory failure from intercostalor diaphragmatic paralysis. This can occur veryrapidly, with little previous distress or deteriorationin blood gases. Arterial blood-gas monitoring isessential, but it must be emphasized that hypoxaemiaand hypercapnia are late findings and indicate thatrespiratory arrest is imminent. Respiratory reserve isbest monitored by serial determinations of forcedvital capacity. Elective intubation should be per-formed when the vital capacity approaches 15 mL kg"1

(1 L) or sooner if pharyngeal paralysis impairs thehandling of secretions. About 25% of patientsultimately require ventilation for periods rangingfrom 1 week to over 1 year. With good intensive caresupport, the mortality has been reduced to 5%,although about 15% of survivors have residualneurological deficits. Respiratory function willrecover within 2 weeks in many patients, but a tra-cheostomy may be necessary if there is no sign ofrecovery after this time.

MYASTHENIA GRAVIS

Myasthenia gravis patients are often admitted to theICU after thymectomy. Occasionally, confirmationof the diagnosis with the tensilon test may precipitatea 'cholinergic crisis', requiring immediate intubation.Rarely, myasthenia gravis may develop rapidly withsevere, generalized muscle weakness and respiratoryfailure requiring urgent ventilation. Patients whocomplain of dyspnoea or difficulty swallowingshould be admitted to hospital immediately forobservation and monitoring of respiratory function.Use of the spirometer may be hampered by weakfacial muscles, in which case monitoring negativeinspiratory pressure is a useful alternative. A clearairway should be maintained by regular, gentle suc-tioning and postural drainage. Segmental atelectasisand hypostatic pneumonia may be prevented byintermittent positive pressure ventilation using avariety of techniques. Incentive spirometry mayresult in respiratory muscle fatigue and should be

avoided. The patient should be transferred to theICU if the vital capacity falls below 1.5 L, and seriousconsideration should be given to elective intubationif the vital capacity falls below 1 L. Early intubationmay prevent a respiratory arrest due to mucous plug-ging or sudden fatigue. Mechanical ventilation isrequired in about 10% of patients with myastheniagravis and may be precipitated by surgical stress, theadministration of aminoglycosides, neuromuscularblocking agents or cholinergic agents. Infection mayalso precipitate respiratory failure, and these patientsare at increased risk of infection due to immunosup-pression (steroid/azothioprine therapy) or becauserespiratory muscle weakness impairs the ability toclear secretions. If ventilation is required,withdrawal of anticholinesterases (neostigmine,pyridostigmine) for 72 hours may improve theireffectiveness when they are restarted.

REFERENCES

1. Thompson, AB, Teschler, H, Rennard, SI. Pathogenesis,

evaluation, and therapy for massive haemoptysis. Clin

Chest Med ̂ 92; 13: 69-82.

2. Cahill, BC, Ingbar, DH. Massive haemoptysis assessment

and management. Clin Chest Med 1994; 15:147-68.

3. Garzon, AA, Cerruti, MM, Golding, ME. Exsanguinating

hemoptysis. J Thome Cardiovasc Surg 1982; 84: 829-33.

4. Conlan, AA, Hurwitz, SS, Krige, L, Nicolaou, N, Pool, R.

Massive hemoptysis. Review of 123 cases. 7 Thome

Card iovasc Surg 1983; 85:120-4.

5. Hirshberg, B, Biran, I, Glazer, M, Kramer, MR.

Hemoptysis; etiology, evaluation, and outcome in a

tertiary referral hospital. Chest 1997; 112: 440-4.

6. Patel, U, Pattison, CW, Raphael, M. Management of

massive haemoptysis. BrJ Hasp Med 1994; 74: 76-8.

7. Crocco, JA, Rooney, JJ, Fankushen, DS, DiBenedetto, RJ,

Lyons, HA. Massive haemoptysis. Arch Intern Med

1968; 121: 495-8.

8. Gong, H, Salvatierra, C. Clinical efficacy of early and

delayed fibreoptic bronchoscopy in patients with

hemoptysis. Am Rev Respir Dis 1981; 124: 221-5.

9. Bobrowitz, ID, Ramakrishna, S, Shim, YS. Comparison

of medical vs surgical treatment of massive

haemoptysis. Arch Intern Med 1983; 143: 1343-6.

10. Conlan, AA, Hurwitz, SS. Management of massive

haemoptysis with the rigid bronchoscope and iced

saline lavage. Thorax 1980; 35: 901-7.

References 123

11. Sense, L. Intrabronchial selective coagulativetreatment of hemoptysis. Chest 1990; 97: 990-6.

12. Saw, EC, Gottlieb, LS, Yokayama, T, Lee, BE. Flexiblefibre optic bronchoscopy and endobronchialtamponade in management of severe haemoptysis.C/i«M976; 70: 589-91.

13. Saumench, J, Escarrabil, J, Padro, L, et al. Value offibreoptic bronchoscopy and angiography fordiagnosis of the bleeding site in haemoptysis.Ann ThoracSurg 1989; 48: 272-4.

14. Tan, RT, McGahanJP, Link, DP, Lantz, BMT. Bronchialartery embolisation in management of haemoptysis.J Intervent Radiol 1991; 6: 67-76.

15. Mai, H, Rullon, I, Mellot, f,et al. Immediate andlong-term results of bronchial artery embolization forlife-threatening hemoptysis. Chest 1999; 115:996-1001.

16. Jewkes, J, Kay, PH, Paneth, M, Citron, KM. Pulmonaryaspergillosis: analysis of prognosis in relation tohaemoptysis and survey of treatment. Thorax 1983;38: 572-8.

17. Jones, KD, Davies, RJ. Massive haemoptysis. BMJ 1990;300: 889-900.

18. Lomotan, JR, George, SS, Brandstetter, RD.Aspiration pneumonia. Strategies for earlyrecognition and prevention. Postgrad Med 1997; 102:229-31.

19. Bynum, K, Pierce, AK. Pulmonary aspiration of gastriccontents. Am Rev Respir Dis 1976; 114:1129-34.

20. Downs, JB, Chapman, RL, Modell, JH, et al. Anevaluation of steroid therapy in aspiration

pneumonitis. Anaesthesiology 1974;40:129-34.

21. Drakulovic, MB, Torres, A, Bower, TT, Nicolas, JM, Nogue,S, Ferrer, M. Supine body position as a risk factor fornosocomial pneumonia in mechanically ventilatedpatients; a randomised trial. Lancet 1999; 354:1851-8.

22. Stoutenbeck, CP, van Saene, HK. Nonantibioticmeasures in the prevention of ventilator-associatedpneumonia. Sew/A? Respir Infect 1997; 12: 294-9.

23. Modell, HH. Drowning. N EnglJ Med 1993; 328: 253-6.24. Golden, F, Tipton, MJ, Scott, RC. Immersion, near

drowning and drowning. BrJ Anaesth 1997; 79: 214-25.25. Standards and guidelines for cardiopulmonary

resuscitation (CPR) and emergency cardiac care (EEC).JAMA 1986; 255: 2905-84.

26. Ornato, JP. The resuscitation of near-drowning victims.JAMA 1986; 256: 75-7.

27. Abdulmajid, OA, Ebeid, AM, Motaweh, MM, Kleibo, S.Aspirated foreign bodies in the tracheobronchial tree:report of 250 cases. Thorax 1976; 31: 635-8.

28. Lentz, CW, Peterson, HD. Smoke inhalation is amultilevel insult to the pulmonary circulation. CurrOpin Pulm Med 1997; 3: 221-6.

29. Frampton, MF, Utell, MJ. Inhalational injuries due toaccidental and environmental exposures. Curr OpinCrit Care 1995; 1:246-52.

30. Jackson, DL, Menges, H. Accidental carbon monoxidepoisonmg.JAMA 1980; 243: 172-4.

31. Christie, AB. Infectious diseases: epidemiology andclinical practice 4th edition. Edinburgh: ChurchillLivingstone, 1987.

10Respiratory emergencies II: chest trauma,air leaks and tracheostomyRICHARD M LEACH AND DAVID A WALLER

Chest trauma 124The management of air leaks in intensive care 130

Tracheostomies, tracheal stenosis andtracheomalacia

References134136

CHEST TRAUMA

Chest trauma may be penetrating or blunt.Penetrating chest wounds are a surgical emergencyand many patients will have had definitive treatmentbefore they are admitted to the intensive care unit(ICU). However, while on the ICU they, must bemonitored for secondary complications and prob-lems that were not detected at admission. Chest traumamanagement requires a multidisciplinary approachinvolving the ICU team, cardiothoracic surgeons,radiologists and physicians. Frequently, life-threaten-ing problems are neglected and the primary functionof the ICU physician is to co-ordinate and overseethe management of these critically ill patients.1-5

General factors in chest traumamanagement

Chest injury is often a single component in the mul-tiple-trauma victim.1-5 Hypoxaemia, CO2 retentionand hypotension resulting from the chest damage canhave profound detrimental effects on other organs.This is of particular importance when chest and headtraumas occur together. Pain control is essential:inadequate analgesia results not only in an agitated

patient, who will fight the ventilator and disconnectmonitoring and intravenous lines, but also in poorventilation, CO2 retention and atelectasis.

Radiographic imaging, including chest radiog-raphy (CXR), chest computed tomography (CT) scanand screening techniques, is of particular importancein the management of thoracic trauma. An uprightCXR should be obtained if possible because filmstaken in the supine position may not give an accurateestimation of the degree of pneumothorax or pleuralfluid. The film should be assessed for the presence ofsoft-tissue and bony abnormalities, pulmonary infil-trates, pneumothorax, mediastinal shift or wideningand thoracic fluid. The position of monitoringcatheters, endotracheal and chest tubes should also beevaluated. Chest CT scans with contrast are invalu-able in the assessment of intrathoracic damage andlocalization of abnormal air or fluid collections.

Penetrating chest trauma

Penetrating chest injuries are usually caused by sharpimplements or gunshot wounds, which may be eitherlow or high velocity.1'4 Most low-velocity gunshotwounds and stab wounds to the pulmonaryparenchyma will cause haemopneumothorax andcan be managed with chest drainage alone. Chest

Chest trauma 125

drainage will re-expand the lung, evacuate blood andstop bleeding by tamponade of the low-pressure pul-monary vessels (Fig. 10.1). Bleeding from an inter-costal vessel or a large pulmonary vessel may stopspontaneously, but often requires surgical control.Although an antero-apical drain to remove air and apostero-lateral drain to remove blood are frequentlyadvocated, a single, correctly placed, large-bore pos-tero-lateral chest drain is usually adequate. Multipledrains are only required for loculated collections.

Figure 10.1 An inebriated 64-year-old man walked into the

casualty department complaining of 'stabbing' chest pain. Removal

of his coat revealed a carving knife in his upper chest. His wife had

stabbed him during a domestic dispute. When access to immediate

surgical facilities was available, the knife was removed without

complications and the patient was managed with a single left-sided

chest drain. He made a good recovery. As his wife had stabbed him

on a previous occasion, urgent marital counselling was advised!

The indications for thoracotomy in penetratingchest trauma are:

• the initial amount of blood removed from thechest is greater than 1250 mL,

• chest drainage persists at more than 250 mL h"1

for 3 consecutive hours, suggesting damage to alarge pulmonary or systemic vessel,

• an air leak sufficiently large to compromiseventilation,

• cardiac tamponade,• transmediastinal stab wound: facilities for imme-

diate surgery should be available before theimplement is removed.

In high-velocity gunshot wounds, the enormouskinetic energy released causes cavitation, resultingin extensive lung and surrounding tissue damage.Early surgical intervention is usually required toevacuate blood clots, resect severely damaged lungand prevent air leaks, empyema and haemorrhage.Combat injuries often have to be managed initiallywith chest drainage alone, but many will developcomplications that will require surgical interven-tion. High-velocity injuries are also more prone toair embolism than blunt or low-velocity injuries.This complication is more common than previ-ously realized, potentially occurring in any patientwith parenchymal injury. It may only becomeapparent when the patient is mechanically venti-lated and the increased intrathoracic pressureforces air from a ruptured bronchus into pul-monary veins. Severe dysrhythmias may occur ifair enters the coronary arteries, and neurologicalcomplications, particularly fitting, result if it entersthe cerebral arteries. A large cerebral air embolusmay mimic the appearance of a cerebrovascularaccident. Multiple small air emboli result in confu-sion and a variety of focal symptoms and signs.Any patient who develops unexplained neurologi-cal or cardiac symptoms with penetrating chesttrauma (including lacerations due to fracturedribs) should be assessed for air embolism. Chestauscultation may detect localized chest signs andair may be seen in the retinal arteries, which isdiagnostic. Echocardiography may be particularlyhelpful in detecting air bubbles in the heart. If airembolism occurs, positive pressure ventilationshould be stopped and the patient taken to theatre.Massive air embolism may require immediatethoracotomy in the ICU, with clamping of theaffected hilum to prevent further air entering the

126 Respiratory emergencies II: chest trauma, air leaks and tracheostomy

bloodstream, after which mechanical ventilationmay be restarted and definitive surgery arranged.

Blunt chest trauma

Blunt chest trauma is a common occurrence andrelatively minor trauma may have serious conse-quences.1"5 On chest examination, there maybe fewexternal signs, apart from bruising. Recognition ofthe resulting, and sometimes severe, intrathoracicdamage is often delayed or missed.4 Injury mayaffect any of the intrathoracic structures and manydistant structures, such as the brain.

MECHANISMS OF INTRATHORACIC INJURY

The types of injury resulting from blunt chest traumaare direct (e.g. rib and sternal fractures, cardiaccontusion and other soft-tissue injuries), shear andpressure related. All intrathoracic structures aretethered to adjacent tissues. Shear forces producedby differential organ motion may cause visceral orvascular tears. The most serious injuries of this typeare deceleration-induced shear tears of the aortaand tracheobronchial tree. Other common injuriesinclude pulmonary contusions and haematomas.

Sudden elevations of intrathoracic and intracavi-tary pressures may rupture air-filled or fluid-filledstructures. Thus, oesophageal rupture may result inmediastinitis or empyema, and alveolar rupture maycause pneumothorax or pulmonary haemorrhage.The diaphragm may also tear, with herniation ofabdominal contents into the thoracic cavity andrestriction of ventilation.

Blunt trauma has also been classified according tothe nature of the impact:6

• high-velocity impact (deceleration): fracturedsternum/anterior flail segment, ruptured aorta,major airway injury,

• low-velocity impact (direct blow): unilateral frac-tured ribs, pulmonary and cardiac contusion,

• crush: fractured ribs/flail segments, rupturedbronchus/oesophagus.

quently broken, usually along the posterior axillaryline, which is the site of maximum stress (Fig. 10.2).Ribs one and two are protected by the shoulder gir-dle, and fracture implies a very forceful blow andshould raise the suspicion of potential damage to theairways and great vessels.7 Ribs 10 to 12 are lessprone to injury, but fracture indicates potential dam-age to the spleen, liver and kidneys. Rib fractures cancause direct lung damage with pneumothoraces,contusions, lacerations and haemothorax. The asso-ciated pain impedes breathing, causing hypoventila-tion and atelectasis.

A flail chest occurs in the self-ventilating patientwhen multiple rib fractures, usually in two sites,result in a free segment of chest wall or sternum with'paradoxical motion' on inspiration. Dislocation ofthe costochondral junctions, which are not detectedon chest films, may combine with posterior fracturesto create a flail segment that may be missed. In thissituation, careful palpation, in addition to radiogra-phy, is essential if the two sites of abnormal motionare to be detected. Even so, the flail segment may notbe recognized for several hours due to 'splinting' byinvoluntary chest-wall muscle spasm. If the flail seg-ment involves three ribs or more, the patient is athigh risk of developing significant impairment ofventilation. Falls in vital capacity to 15 mL kg-1

CHEST-WALL INJURIES

Rib fractures occur more frequently in olderpatients. Chest-wall flexibility in younger patientsoften allows energy transfer to intrathoracic organswithout rib fracture. Ribs six to nine are most fre-

Figure 10.2 Multiple right-sided (posterior axillary line) rib

fractures and pulmonary contusion following severe, right-sided

blunt chest trauma sustained during a road traffic accident by a

23-year-old motorcyclist.

Chest trauma 127

(normal 60 mL kg- L) or blood-gas deterioration areindependent indicators that mechanical ventilationmay be required.

In patients with a 'mild to moderate' flail chest,the need for ventilation may be avoided with ade-quate pain control, drainage of air leaks, careful pul-monary toilet, nasotracheal suctioning and chestphysiotherapy. Adequate analgesia will usuallyrequire a combination of oral, intravenous orpatient-controlled opiate analgesia with intercostalnerve blocks. In general, intravenous analgesia willbe required if more than two ribs are fractured.Nerve blocks with 0.25% bupivacaine and adrenaline1:200 000 are extremely useful, but, if feasible, a thor-acic epidural is more effective and provides easierlong-term control.8 These techniques have theadvantages of causing less sedation than intravenousopiates, assisting physiotherapy and permittingmobility. The paradoxical motion of a 'severe' flailsegment in combination with pulmonary contusionsand other chest injuries will inevitably result inregional hypoventilation, retained secretions, shunt-ing and severe hypoxaemia in some patients. Severalstudies have demonstrated that morbidity and mor-tality are improved with positive pressure ventila-tion, which overcomes the paradoxical motion of thechest wall and ensures adequate ventilation.Mechanical ventilation may be required for 7-14days and patients with severe flail segments are usu-ally more comfortable and easier to manage with anearly tracheostomy. The important principle in themanagement of flail chest is that the underlying pul-monary contusion is more detrimental than anyparadoxical movement and therefore therapy shouldbe directed towards the underlying lung rather thanthe chest wall. In general, it is now accepted thatexternal chest-wall stabilization (taping, sandbag-ging, fixation) is of little benefit in these patients.

TRACHEOBRONCHIAL INJURIES

Tracheobronchial injuries caused by blunt traumaare rare and often missed. A high index of suspicionis required in any patient with severe thoracicinjury.9,10 Fractures of the first and second ribs areassociated with tracheobronchial disruption in 15%of cases.7 The diagnosis should also be suspected inany patient with haemoptysis (10% of cases) or per-sistent air leaks (90% of cases), including bilateralpneumothoraces, pneumomediastinum and massivesubcutaneous emphysema despite adequate chest

drainage. Characteristic sites of injury are a longitu-dinal tear in the posterior membranous portion ofthe tracheal wall (15%) or a spiral tear in the mainbronchi (80%), usually within 2.5 cm of the trachealcarina (Fig. 10.3). The injury occurs like a chickenwishbone breaking, with the force of the rotatingheart forcing the bronchi apart. Tracheobronchialtears may not be recognized at the time of the injuryand the process of repair is complicated by bronchialstenosis, secondary infection and the eventual devel-opment of irreversible bronchiectasis. Early rigidbronchoscopy will establish the diagnosis and allowthe clearance of debris from the airway. If there isonly a short longitudinal tear in the posterior mem-branous trachea or if the spiral tear is less than athird of the circumference and the lung is expanded,conservative management may be considered. Amini-tracheostomy may help by reducing intratra-cheal pressure and allowing air to escape from themediastinum. Surgery, if necessary, should follow assoon as the patient has been stabilized. Completeseparation of the lung from the trachea ('drop lung')due to transection of a main bronchus usually occurswithin 2 cm of the carina. In contrast to bronchialtears, complete disruption is recognized early on theCXR. There are often no long-term complicationsfollowing surgical repair.

LUNG INJURY AND FAT EMBOLISM SYNDROME

Pulmonary contusions are common after blunt chesttrauma11 and may occur in the absence of rib frac-tures. At the tissue level, localized bleeding and oedemaoccur, leading to ventilation-perfusion V7Q mis-matching and hypoxaemia. Clinically, bruising at thesite of chest trauma, haemoptysis and hypoxaemiaare common. The CXR reveals ill-defined infiltratesor opacities in the path of the trauma, or contra-coup, which usually develop within 6 hours (see Fig.10.2). Opacification may increase for 24-48 hoursand then gradually subsides over the next 7 days.However, progressive hypoxaemia in the absence ofchanges on the radiograph may be the first sign ofpulmonary contusion. Pulmonary haematomas (1-6cm) may develop, but are usually absorbed withminimal morbidity. If the trauma is severe, general-ized pulmonary injury may lead to, or co-exist with,acute respiratory distress syndrome (ARDS).Ventilatory support may be required in the event ofsevere hypoxaemia, but the management of thepulmonary contusions is essentially supportive.


Figure 10.3 The location of tmcheobmnchial tears following

blunt chest trauma.

Pulmonary torsion is a rare but life-threateningcomplication of blunt chest trauma. Early surgicalcorrection is essential.

In cases of severe trauma, fat embolism syndrome(FES) may occur from 1 hour up to 3 days later. It isusually associated with multiple long-bone or pelvicfractures. Lipases hydrolyse neutral triglycerides toliberate fatty acids, which are toxic to the lung. Inaddition, fat emboli may pass through the pul-monary to the systemic circulation, causing infarc-tions in the retina, skin and brain. The characteristicclinical triad of FES is confusion, pulmonary dys-function (with cough, dyspnoea and pleurisy) and apetechial rash over the upper torso. Initially, hypox-aemia may be associated with a normal CXR andreduced lung compliance and gas transfer.Eventually, ARDS may develop. Many patients alsodevelop a coagulopathy, with disseminated intravas-cular coagulation and retinal infarcts. The diagnosisis established by detecting increased serum andurine lipase and fat globules in urine, sputum orserum.

CARDIOVASCULAR INJURY

Cardiac contusionCardiac contusion is a common, but frequentlyunrecognized, complication of non-penetratingblunt chest trauma.12,13 Severe blows to the anteriorchest wall involving the steering wheel or seat beltduring road traffic accidents are the commonestcause. However, relatively mild chest trauma maycause severe myocardial injury (Fig. 10.4). At presen-tation, the patient may have little bruising or discom-fort and recognition is again dependent on a highindex of suspicion. Clinically, most cardiac contu-sions are mild, unrecognized and settle with supportive

therapy alone. Nevertheless, severe myocardial contu-sion is analogous to myocardial infarction, and car-diac arrhythmias are common but often delayed.Thus, cardiac monitoring is indicated for 24 hoursafter severe blunt chest injury. The anterior, thin-walled right ventricle is most commonly damaged,but the small amount of muscle involved at this sitemay not be sufficient to raise cardiac enzymes, despitesignificant dysfunction. Nevertheless, cardiacenzymes, troponin levels and serial electrocardio-grams (ECGs) are usually monitored to assessprogress. Technetium scans are non-specific but maybe helpful (Fig. 10.4). Echocardiography is essentialto assess the integrity of the valves, particularly whencardiac murmurs are detected. Coronary angiogra-phy is usually normal as the pathophysiologicalmechanism in cardiac contusion is thought to involvemicrovascular disruption secondary to tissue oedema.Treatment of arrhythmias and heart failure is asfollowing myocardial infarction. Residual functionaldisability may occur (Fig. 10.4), but is rare: most casesare unrecognized and usually recovery is complete.

Aortic and great vessel injuryAbrupt deceleration injuries due to road traffic acci-dents account for most aortic injuries.13,14 The aortacan withstand transmural pressures greater than2000 mmHg, but it is less able to tolerate shearstresses. The majority of aortic injuries are fatal at thescene of the accident and therefore, in practice, this isan injury that is suspected much more often then itactually occurs. Characteristically, 80% of aortic rup-tures occur just distal to the ligamentum arteriosum(the aortic isthmus) and 5% occur at the aortic rootjust above the aortic valve and may damage the valveand coronary arteries. Often there is no evidence ofexternal trauma, and rib and sternal fractures are notalways present. The diagnosis should be suspectedwhen chest pain occurs in the intrascapular region,with signs of aortic valve regurgitation, pulse differ-entials or acute neurological changes after bluntchest trauma. The CXR may reveal a wide medi-astinum, indistinct aortic arch, pleural fluid, a pleur-al cap or depressed left main bronchus. Angiographyis the most sensitive technique to confirm the diag-nosis. Evaluation of the aorta by helical CT scanswith contrast is less sensitive because cross-sectionalimaging may not demonstrate a transverse lesion. Ithas the advantages of speed and concurrent assess-ment of other intrathoracic structures in the

Chest trauma 129

Figure 10.4 Following arrest for suspected burglary, a 25-year-old

man fell onto the back step of the police van, sustaining a blunt,

central chest injury. The man presented to the casualty department

12 hours later with heart block (Mobitz 2) and ECG changes of an

inferior myocardial infarction. His CXR at admission showed mild

right-sided pulmonary contusion (a), but by 48 hours he had

developed marked heart failure (b). A technetiurn scan (c)

confirmed severe inferior and anterior myocardial damage, but

coronary angiograms were normal. The man was discharged on

diuretics with marked cardiomegaly.

critically ill trauma patient, but surgery is unlikely tobe undertaken without angiography.15

Traumatic injuries to the heart andpericardiumTraumatic damage to the heart can involve any of thechambers, valves, papillary muscles or the pericardialsac (Fig. 10.5).13 In most cases (>80%), transmuralrupture is rapidly fatal. Valve rupture is more commonin older males. The aortic (60%) and mitral (30%)valves are most frequently affected. Diagnosis withtransoesophageal echocardiogram and early repair isessential. Traumatic tamponade may be due to aorticroot disruption, coronary artery laceration or ruptureof the free ventricular wall. The clinical features are asfor tamponade from any cause. Early surgical decom-pression reduces mortality, but pericardiocentesis maybe attempted whilst preparing for surgery.

OESOPHAGEAL RUPTURE

Oesophageal injury in blunt chest trauma is uncom-mon.1-5,16 However, in major thoracic trauma, thesigns of a perforated oesophagus are often over-looked or attributed to other serious injuries. A per-forated oesophagus implies violent deceleration andis usually associated with other life-threatening dam-age. The poor prognosis (approximately 60% mor-tality within 10 days of admission) is usuallyattributable to these accompanying injuries. Thediagnosis should be suspected if major trauma isassociated with rapidly developing pleural effusions,pneumothorax or mediastinal or subcutaneousemphysema. If the diagnosis has been missed, it maynot be apparent until an empyema with associatedpain, fever or hypotension develops. The CXR showsmediastinal widening with mediastinal or pleural


gas. The diagnosis should be established with awater-soluble contrast swallow, which revealsextravasation into the pleural cavity or mediastinum.Endoscopy and CT scans are unhelpful in diagnosis.Early diagnosis is essential because, untreated, themortality of oesophageal rupture is about 2% h'1

and is fatal in 50% of cases within the first day due tomediastinitis. Surgical repair is usually undertakenwithin 24 hours of rupture, after which time surgeryis reserved for mediastinal debridement and lavage.Rarely, oesophagectomy with primary or delayedreconstruction may be appropriate.

DIAPHRAGMATIC INJURY

Diaphragmatic injuries complicate 7% of severethoracic trauma cases and are life threatening unlessdiagnosed and treated early (Fig. 10.6).17 Mortality inpatients with diaphragmatic injury due to blunttrauma is about 40% because it is commonly associatedwith rupture of the spleen or liver. About 80% occur onthe left because the right hemidiaphragm is protectedand supported by the liver. The diagnosis is oftenmissed as positive pressure ventilation masks therespiratory distress and may relocate herniated bowel.The diagnosis may be confirmed by thoracoscopy, buttreatment is best performed by a thoracotomy.

THE MANAGEMENT OF AIR LEAKS ININTENSIVE CARE

Figure 10.5 A 27-year-old man sustained a severe blunt chest

injury following a road traffic accident. His CXR (a) and CT scan (

showed the heart to be displaced to the right. At thoracotomy (c),

was found to have a ruptured pericardial sac through which the

heart had herniated (see also Plate 5).

he

Air leaks are a common and frequently serious prob-lem in the ICU. The aetiology is extensive, but in theintensive care setting three situations are particularlyimportant:

• ventilator-associated lung injury,• cardiothoracic surgery (iatrogenic chest-wall

penetration/parenchymal damage),• blunt or penetrating chest trauma.

Many conditions predispose to air leaks, includingnecrotizing lung pathology, non-homogeneousparenchymal disease, prolonged ventilation and youth.As a general rule, all pneumothoraces in patients in theICU require immediate tube drainage, whereas themanagement of pneumatocoeles and loculatedpneumothoraces is less well defined. However, in ourexperience, drainage of most large air collectionsunder radiological guidance may be life saving

The management of air leaks in intensive care 131

Figure 10.6 Left-sided diaphragmatic rupture in a 33-year-old

man following a road traffic accident. Characteristically, the

diaphragm is not clearly visualized on the CXR. The left-sided chest

drain had, in fact, been inserted into the abdominal cavity. This

emphasizes the importance of performing a CT scan if this

diagnosis is suspected.

(Fig. 10.7) Drainage of other air leaks, including pneu-momediastinum and subcutaneous emphysema, is notnecessary, except in exceptional circumstances.

Pneumothorax

Pneumothorax18'19 maybe categorized into:

• primary spontaneous pneumothorax, usuallyoccurring in a young adult without underlyinglung disease,

• secondary pneumothorax, usually occurring in anelderly patient with underlying lung disease, e.g.chronic obstructive pulmonary disease (COPD),

• traumatic or iatrogenic pneumothorax caused byalveolar puncture due to a foreign body.

A tension pneumothorax may complicate allthree types if air continues to accumulate in thepleural cavity faster than it can be removed orabsorbed. The resulting increase in intrathoracictension compresses the remaining functioning lung,inhibits venous return and reduces cardiac output.This condition is fatal if the tension is not rapidlyrelieved by drainage.

Figure 10.7 Two weeks after admission, a 22-year-old woman

with severe ARDS due to blunt chest injury had developed multiple

loculated pneumothoraces and pneumatocoeles on CXR (a) and CT

scan (b). The CT scan demonstrated that less than 30% of the lung

tissue was ventilated. The woman's condition progressively

deteriorated until the air collections were drained under

radiological screening. At one stage, nine functioning chest drains

were in place. She subsequently made a good recovery, with

minimal long-term respiratory impairment.


A small 'secondary' pneumothorax in a patientwith emphysema may have more seriousimplications than a large 'primary' pneumothorax inan otherwise healthy young man because therespiratory reserve to compensate for the associatedloss of ventilatory capacity is reduced in subjectswith underlying lung disease.19 The incidence ofsecondary pneumothorax increases with age andwith the severity of the underlying lung disease. It ismost commonly associated with COPD, but mayaffect a wide variety of other lung diseases thatdamage lung architecture or the pleura. The patientwith a secondary pneumothorax is more likely to beadmitted to the ICU, not only for management of thepneumothorax but also for management of theunderlying lung disease. Ventilated patients withlung disease are also at greater risk of secondarypneumothorax as a result of 'barotrauma' (damagedue to high airway pressures) and Volutrauma'(alveolar over-distension) associated with mechanicalventilation.20,21 There is now evidence that'protective' ventilation strategies using low-pressure,low-volume ventilation may reduce this risk.22 Bluntor penetrating chest trauma is associated withpneumothoraces due to direct penetration of thechest wall, increased intrathoracic pressure and shearstress causing tracheobronchial disruption oroesophageal rupture (see 'Chest trauma', above).

is performed. If a chest drain is not readily available,the largest available needle should be insertedthrough the chest wall in the third intercostal spacein the mid-clavicular line into the pneumothorax,relieving tension and preventing the pneumothoraxfrom increasing in size. Characteristically, a 'hiss ofgas' escaping through the needle occurs as tension isrelieved. A chest drain should be positioned as soonas practically possible. Bilateral pneumothoraces,often occurring after thoracic trauma, representanother situation in which immediate chestdrainage is essential (Fig. 10.8).

The management of pneumothoraces in ventilatedpatients is a challenging problem. In experimentalmodels, there is extensive evidence that high ventila-tory pressures cause 'barotrauma', including pneu-mothoraces and other air leaks.23,24 More recently, ithas been postulated that high ventilatory volumesdelivered to areas of lung unaffected by the diseaseprocess lead to damage by alveolar over-distension orVolutrauma'.24,25 This over-distension of relativelynormal lung also results in air leaks, increased alveo-lar permeability and capillary damage. In animalstudies, acute lung injury associated with air leaks hasan increased mortality. In contrast, some studies ofARDS patients have reported that, although the inci-dence of air leaks varied from 0% to 92%, the correl-ation with airway pressure, ventilatory tidal volume

MANAGEMENT OF PNEUMOTHORAX ININTENSIVE CARE

Most patients admitted with or developing a pneu-mothorax on the ICU will have either severe under-lying lung disease or traumatic lung damage.19 Themajority will be ventilated and almost all willrequire intercostal tube drainage. Recognition andearly drainage of significant air leaks can be life sav-ing in such patients. Any pneumothorax that devel-ops in a patient on positive pressure ventilationrequires drainage before it enlarges and causes car-diorespiratory compromise. Prompt recognition ofa tension pneumothorax, presenting with rapidlydeveloping respiratory distress, hypoxaemia andcardiovascular instability, is essential. Clinical fea-tures include hyper-resonance on chest percussion,a shift in the trachea and apex beat and reducedbreath sounds on auscultation over the affectedlung. The detection of a tension pneumothorax is aclinical diagnosis and should be made before CXR

Figure 10.8 Bilateral pneumothoraces (arrows) following a fall

from a roof in a 33-year-old man.

The management of airleaks in intensive care 133

and mortality was poor.21 However, these studies didnot always differentiate between the early inflamma-tory and late fibroproliferative phases of ARDS and itis our experience that air leaks often develop in thelate phase of the disease. Nevertheless, every effortmust be made to reduce the risk of air leaks inmechanically ventilated patients, by avoiding highairways pressures and volumes, preventing secondaryinfection and by careful bronchial toilet. Recent stud-ies have demonstrated that the early institution oflow-pressure, low-tidal-volume ventilation improvesboth morbidity and mortality in ARDS.22

Once an air leak or pneumothorax has developed,high airways pressures associated with mechanicalventilation and alveolar recruitment strategies - pos-itive end-expiratory pressure (PEEP) and reversedinspiratory:expiratory time (I:E) ratios - willencourage persistence of the leak. The appropriatestrategy is to use the lowest peak, mean and end-expiratory airway pressures and I:E ratios compatiblewith adequate gas exchange and to ensure adequatedrainage of any gas collection. Even a smallpneumothorax may compromise gas exchange in theventilated patient with severely damaged lungtissue, and conservative management or aspiration is

inappropriate. Multiple intercostal drains insertedunder screening may be required to ensure adequatelung re-expansion in patients with loculated pneu-mothoraces. However, this strategy requires earlydetection and precise localization of air spaces,which is not always possible with routine CXR. Inthis respect, serial CT scans and screening are anessential but insufficiently used diagnostic tool in themanagement of the critically ill patient with severelung injury (Fig. 10.9).

In the mechanically ventilated patient with severelung disease, air leaks may be large and difficult tomanage.26 If a large, persistent air leak develops,increased suction of up to 10 kPa may be applied tothe drain with the aim of obliterating the pleuralspace. The correct pressure is that which opposes thelung and chest-wall pleural surfaces, allowing spon-taneous pleurodesis. A high-flow, low-pressure sys-tem is essential. Wall suction is ideal and mostportable pump systems are incapable of coping witha large flow rate and are potentially harmful becausethey may actually obstruct the air leak and produce atension pneumothorax. Early thoracic surgicaladvice should be obtained if the pleural surfaces can-not be opposed in the presence of a large air leak.

Figure 10.9 Serial CT scans, at weekly intervals, in a 55-year-old man who developed severe postoperative ARDS with recurrent loculated

pneumothoraces in different sites that compromised respiratory gas exchange. The loculated pneumothoraces could only be accurately

located and drained with CT scanning.


Initial bronchoscopy (preferably using the rigidscope) for bronchial toilet is valuable because thelung cannot re-inflate if the proximal bronchi areobstructed by blood or mucus. The decision then liesbetween surgical closure of the air leak or an attemptat chemical pleurodesis. There is no indication forintroducing sclerosants if the pleural surfaces cannotbe opposed and surgery is required. Now the treat-ment of choice is video-assisted thoracoscopy, whichis as effective as thoracotomy but causes less respira-tory dysfunction.27,28 If the air leak persists but thelung is expanded, closed pleurodesis via the chesttube using a slurry of talc should be considered. Theterm bronchopleural fistula is frequently misused. Itshould be reserved for postoperative air leaks result-ing from disruption of a bronchial suture line and isgenerally an indication for re-operation because thesize of the air leak is so large.

Chest drains are often required to drain air orfluid from the traumatized chest to maintain bothrespiratory and cardiovascular function. They mustbe well positioned, secured and monitored to ensureproper functioning. All drains inserted withoutimaged guidance should be placed in the so-called'triangle of safety' between the anterior border oflatissimus dorsi and the posterior border of pec-toralis major and above the level of the nipple. Wheninserting a drain in a ventilated patient, remember todisconnect the ventilator on insertion. High-flowwall suction of between 2 and 10 kPa negative pres-sure may be required to ensure complete lung expan-sion. Drain patency is enhanced by preventingkinking of the tube or pooling of secretions and sub-sequent clotting due to excessive length of tubing.Clamping of drains for transport or patient reposi-tioning is not recommended because of the risk oftension pneumothorax.

Timing of chest drain removal depends on theclinical situation. In general, a chest drain can beremoved when there is clinical and radiologicalevidence of lung re-expansion and there has beenno air leakage (bubbling) through the drain for12-24 hours. It is not necessary to clamp drainsbefore removal, and this practice is discouraged.Before removal of the drain, the patient should begiven adequate analgesic and anti-emetic cover.The drain should be pulled out when the patientis in inspiration or breath-holding, and thepurse-string suture tied securely. The patientmust be carefully observed for recurrence of thepneumothorax.

Other air leaks

Pneumomediastinum describes air dissecting alongthe mediastinal-pleural reflection, outlining the heartand great vessels with air. Air may also dissect alongperivascular sheaths into the neck, causing subcuta-neous emphysema, or around the heart, resulting in apneumopericardium, which may cause tamponade.As with pneumothorax, the source of the air leak maybe spontaneous, iatrogenic or traumatic. In the ICU,it can indicate on-going ventilator-induced 'barotrau-ma' or damage to the airways from intubation or per-cutaneous tracheostomy. More serious causes includeinjury to the trachea, bronchus and oesophagus,which require surgical repair.

Physical examination reveals subcutaneousemphysema at the root of the neck; a nasal qualityto the voice and auscultation over the precordiummay reveal a 'crunch' with each heart beat (Roman'ssign). Paramediastinal air cysts are occasionallyseen in young patients who experience severe chesttrauma. Massive subcutaneous emphysema mayoccur over the entire body, leading to a grotesque,bloated appearance with closure of the eyelids and acharacteristic crackling feeling on palpation.

In general, subcutaneous emphysema andpneumomediastinum are not associated with severerespiratory complications and both conditionsusually resolve spontaneously. General managementin the ICU includes low-pressure, low-tidal-volume,protective ventilation strategies and ensuringgood drainage of existing air leaks. The failure ofextensive subcutaneous emphysema to resolve, withprogressive dyspnoea, usually indicates inadequatedrainage of a large air leak and should promptcareful investigation, including bronchoscopy, forpreviously undetected leaks or problems that woulddecrease chest drain efficiency. Occasionally patientswith severe surgical emphysema may require asuperficial skin incision in the suprasternal notch(cervical mediastinotomy) to allow subcutaneous airto escape.

TRACHEOSTOMIES, TRACHEAL STENOSISAND TRACHEOMALACIA

Prolonged endotracheal intubation causes progres-sive damage to the upper trachea as a result of cuffpressure and movement.29 Prevention requires the

Tracheostomies, tracheal stenosis and tracheomalacia 135

use of appropriate cuffed tubes, low cuff pressuresand the prevention of movement. The longer theendotracheal tube is to be left in place, the moreimportant is each of the components. Large-volume,low-pressure or foam cuffs are available that willmaintain a good seal at low pressure and preventaspiration. Cuff pressures should be measured andmaintained at the lowest levels that ensure no leaks oraspiration. Patient movement can be minimized bysecurely fixing the endotracheal tube and ensuringadequate sedation and analgesia. Appropriate endo-tracheal tube position, with the tip 3-5 cm above thecarina, is confirmed at CXR. It is important to notethat the tip will move between 2 and 4 cm from fullextension to full flexion of the neck and chin. Failureto prevent damage may lead to the development oftracheomalacia or tracheal stenosis. In addition tolocal tracheal considerations, it is becoming increas-ingly apparent that there are a number of other fac-tors that may be important in determining the timingof a tracheostomy. Early tracheostomy may be associ-ated with reduced requirements for sedation,improved intestinal motility and enteral nutrition(less sedation), early weaning and earlier dischargefrom the ICU.30

The timing, method used and care of tra-cheostomies vary amongst units. To a large extent, thetime at which a tracheostomy is fashioned dependsupon the individual case and varies from 10 to 35days. At 7-10 days, we assess whether extubation islikely within the next week. If extubation is consid-ered possible, we continue translaryngeal intubation,but, if not, a tracheostomy is formed. The recentlyintroduced technique of percutaneous tracheostomy,using a Seldinger technique with progressive dilata-tion, has reduced the cost, need for surgical personneland theatre facilities for fashioning tracheostomies,thus expediting the process by an average of nearly 3days.31 In general, this new technique is safe and eas-ily performed by appropriately trained clinicians whocan manage the life-threatening complications.32'33

Deaths have been reported during the procedure, dueeither to the loss of adequate upper airway control orto massive bleeding. The prompt availability of anaes-thetic and surgical back-up is essential. The additionof fibreoptic bronchoscopy facilitates the midlineplacement of the guide wire and dilators and mayreduce the chance of fashioning a false track in thepre-tracheal space or the oesophagus and of tears tothe posterior wall of the trachea.32,33 The primarycomplication is bleeding, which affects 6% of first-

time percutaneous tracheostomies and up to 25-30%of subsequent tracheostomies placed through a previ-ous tracheostomy scar. Bleeding may be torrential ifany of the major vessels close to the trachea are dam-aged, and deaths due to rupture of the brachiocepha-lic vein have been reported. The larger thyroid veinsand arteries are often involved and, if the tamponadeassociated with insertion of the tracheostomy fails tostem the bleeding, surgical exploration may berequired. Infection is rare with percutaneous tra-cheostomy.

Surgical tracheostomy has become much lesscommon with the advent of the technique of percu-taneous tracheostomy. Surgical tracheostomy may beadvisable for patients with abnormal neck anatomyor recent upper airways or neck surgery and in thosepatients at significant risk of bleeding. Infectionrather than bleeding is the primary complication,affecting about 10% of surgical tracheostomies. Thedevelopment of significant (>75%) late subglottic(tracheal) stenosis appears to be very low (<l-2%)with both procedures, although minor stenosis at thesite of the stoma occurs more frequently and is insome part related to the degree of laryngeotrachealinjury prior to tracheostomy (Fig. 10.10).

Once inserted, tracheostomy care involves main-taining a clear, clean and humidified airway. Clots ofblood formed at the time of procedure may causesevere obstructions shortly after the insertion of thetracheostomy. Similarly, tracheal drying and the for-mation of thick viscid secretions due to the loss ofnasal and pharyngeal humidification during ventila-tion may lead to episodes of obstruction and desatu-ration. If the patient cannot be weaned offventilatory support within 10 days, the initial tra-cheostomy tube should be replaced with a long-termtracheostomy tube with a removable inner tube thatcan be cleaned regularly. These tubes should bechanged every 4 weeks because secretions accumu-late on the outer surface of the tube and may formball valves that subsequently obstruct the lumen.Humidification of the inhaled gas will prevent theformation of sputum plugs and, during the weaningprocess, a 'Swedish nose' will reduce water loss by asmuch as 70%. As respiratory independence returns,the tracheostomy cuff may be deflated for periodsand the tracheostomy can eventually be replacedwith a non-cuffed tube. The tracheostomy lumencan be reduced with each change of tracheostomytube. The tube is eventually capped to ensure inde-pendent breathing prior to removal.


Figure 10.10 Magnetic resonance imaging scan: sagittal section

(a); coronal section (b); and bronchoscopic appearance looking

down from the vocal cords (c); see also Plate 6 of a subglottic

tracheal stenosis, indicated by the arrow in (b). The stenosis

developed 2 months after the insertion of a surgical tracheostomy

in a 16-year-old boy ventilated for Guillain-Barre syndrome.

Major tracheal surgery was required to repair the defect.

REFERENCES

1. Sabbe, MB. Recent advances in the diagnosis and

treatment of thoracic injury. Curr Opin Crit Care 1995;

6: 503-8.

2. Kshettry, V, Bolman, R. Chest trauma: assessment,

diagnosis, and management. Clin Chest Med 1994;

15: 137-46.

3. Anderson, DR. The diagnosis and management of

non-penetrating cardiothoracic trauma. Brj Clin Pract

1993; 47: 97-103.

4. Wall, MJ, Soltero, E. Damage control for thoracic

injuries. Surg Clin North Am 1997; 77: 863-78.

5. Ferguson, M, Luchette, FA. Management of blunt

chest injury. Respir Care Clin North Am 1996;

2: 449-66.

References 137

6. Westaby, S, Brayley, N. ABC of major trauma. Thoracictrauma I. BMJ 1990; 300: 1639-43.

7. Gupta, A, Jamshidi, M, Rubin, JR. Traumatic first ribfracture: is angiography necessary? A review of 730cases. Cardiovasc Surg 1997; 5: 48-53.

8. Luchette, F, Radafshar, S, Kaiser, R, et al. Prospectiveevaluation of epidural versus intrapleural catheters foranalgesia in chest wall trauma. 7 Trauma 1994; 36:865-70.

9. Guest, JL, Anderson, JN. Major airway injury in closedchest trauma. Chest 1977; 72: 63-6.

10. Huh, J, Milliken, JC, Chen, JC. Management oftracheobronchial injuries following blunt andpenetrating trauma. Am Surg 1997; 63: 896-9.

11. Hoff, S, Shotts, S, Eddy, V, Morris, J. Outcome ofisolated pulmonary contusion in blunt traumapatients. Am Surg 1994; 60: 138-42.

12. Kumar, S. Myocardial contusion following non-fatalblunt chest trauma. J Trauma 1983; 23: 4-9.

13. Pretre, R, Chilcott, M. Blunt trauma to the heart andgreat vessels. N Engl J Med 1997; 336: 626-32.

14. Ahrar, K, Smith, DC. Trauma to the aorta and aorticarch branches. CurrOpin Cardiol 1998; 13: 355-68.

15. Unsworth-White, MJ, Buckenham, T, Treasure, T.Traumatic rupture of the thoracic aorta: computedtomography may be a dangerous waste of time. Ann RColl Surg Engl 1994; 76: 381-3.

16. Stothert, JC, Buttorff, J, Kaminski, DL Thoracicoesophageal and tracheal injury following blunttrauma. J Trauma 1980; 20: 992-5.

17.Steinau, G, Bosnian, D, Dreuw, B, Schumpelick, V.Diaphragmatic injuries; classification, diagnosis andtherapy. Chirurg 1997; 68: 509-12.

18. Millar, AC, Harvey, JE, on behalf of Standards of CareCommittee, British Thoracic Society. Guidelines for themanagement of spontaneous pneumothoKax. BMJ1993; 307: 114-16.

19. Marini, JJ, Wheeler, AP. Chest trauma, pneumothoraxand barotrauma in critical care medicine. Baltimore:Williams & Wilkins, 1989; 257-70.

20. Gammon, RB, Shin, MS, Groves, RH, et al. Clinical riskfactors for pulmonary barotrauma: a multivariateanalysis. Am J Respir Crit Care Med 1995; 152:1235-40.

21. Weg, JG, Anzueto, A, Balk, RA, et al. The relation ofpneumothorax and other air leaks to mortality in acute

respiratory distress syndrome. N EnglJ Med 1998; 338:341-6.

22. The Acute Respiratory Distress Syndrome Network.Ventilation with lower tidal volumes as comparedwith traditional tidal volumes for acute lung injuryand the acute respiratory distress syndrome. N Engl JMed 2000; 342: 1301-8.

23. Kolobow, T, Morretti, MP, Fumagalli, R, et al. Severeimpairment in lung function induced by high peakairways pressure during mechanical ventilation: anexperimental study. Am Rev Respir Dis 1987; 135:312-15.

24. Dreyfuss, D, Soler, P, Basset, G, Saumon, G. Highinflation pressure pulmonary edema: respectiveeffects of high airway pressure, high tidal volume, andpositive end-expiratory pressure. Am Rev Respir Dis1988; 137: 1159-64.

25. Dreyfuss, D, Saumon, G. Role of tidal volume, FRC,and end-inspiratory volume in the development ofpulmonary edema following mechanical ventilation.Am Rev Respir Dis 1993; 148:1194-203.

26. Powner, DJ, Grenvik, A. Ventilatory management oflife-threatening bronchopleural fistulae-a summary.

Crit Care Med 198V, 9: 54-8.27. Waller, DA, Forty, J, Morritt, GN. Video-assisted

thoracoscopic surgery versus thoracotomy forspontaneous pneumothorax. Ann ThoracSurg 1994;58: 372-7.

28. Waller, DA. Video-assisted thoracoscopic surgery forspontaneous pneumothorax - a 7 year learningexperience. Ann R Coll Surg Engl 1999; 81: 387-92.

29. Gaynor, ER, Greenberg, SB. Untoward sequelae ofprolonged intubation. Laryngoscope 1985; 95: 1461-7.

30. Grower, ER, Bihari, DJ. The role of tracheostomy in theadult intensive care unit. Postgrad Med J 1992; 68:313-17.

31. Friedman, Y, Fildes, J, Mizock, B, et al. Comparison ofpercutaneous and surgical tracheostomies. Chest1996; 110: 480-5.

32. Griggs, WM, Myburgh, JA, Worthly, LIG. A prospectivecomparison of a percutaneous tracheostomytechnique with standard surgical tracheostomy.Intensive Care Med 1991; 17: 261-3.

33. Ciaglia, P, Graniero, KD. Percutaneous dilationaltracheostomy: results and long-term follow-up.Chest 1992; 101:464-7.

11Pathophysiology of acute lung injuryS JOHN WORT AND TIM W EVANS

Definitions 138

Epidemiology of acute lung injury/acute respiratorydistress syndrome 139

Histological changes within the lung 142

Pathogenesis 143

Pulmonary and systemic markers of lung injury 148

The way forward: investigating at-risk populations 151

Conclusions 151

References 151

DEFINITIONS

Acute respiratory distress syndromeand acute lung injury

Acute respiratory distress syndrome (ARDS) wasfirst described over 30 years ago in 12 patients whodeveloped acute respiratory failure in associationwith a variety of serious surgical and medicalpathologies, not all of which involved the lungdirectly. Seven of the 12 patients died.1 In 1988,Murray proposed an expanded definition of ARDSthat attempted to quantify the extent of lung injuryas well as including information about possible causesand the presence of non-pulmonary organ dysfunc-tion. Unfortunately, this lung injury score (LIS)proved unable to predict outcome and therefore hadlimited clinical usefulness (Table 11.1). It was notuntil 1994 that a Consensus Definition of the syn-drome finally emerged, by which time it was recog-nized that ARDS probably represents only theextreme end of a spectrum of lung injury2

Consequently, a lesser degree of acute lung injury(ALI) was also formally defined by the Consensusgroup (Table 11.2). However, the significance of ALI

as a separate entity, or as a predictor of progressionto 'full-blown' ARDS, remains far from clear.Although a declared goal of the ConsensusConference was to bring clarity and uniformity tothe definition of ARDS, important problems remain.

• When applied to large populations of hypoxaemicpatients, the only difference between ALI andARDS appears to be the severity of oxygenationimpairment at the time of assessment, although itmay be that more patients develop ARDS in cer-tain predisposing groups and more develop ALIin other 'at-risk' groups.

• The present definitions of ALI and ARDS stillhave serious shortcomings because ARDS attrib-utable to a direct pulmonary insult, such as pneu-monia, differs both in evolution and physiologicalcharacteristics from lung injury attributable toindirect insults, such as intra-abdominal sepsis.

• Neither definition takes into account the natureof the precipitating cause, or the presence of non-pulmonary organ dysfunction, both of which areknown to influence outcome.

• The variation in interpretation of chest radio-graphy (CXR) of patients with lung injury is con-siderable, even amongst experts.

Epidemiology of acute lung injury/acute respiratory distress syndrome 139

Table 11.1 Components of the Murray Lung Injury Score

Value

Chest radiograph scoreNo alveolar consolidation 0Alveolar consolidation in one quadrant 1Alveolar consolidation in two quadrants 2Alveolar consolidation in three quadrants 3Alveolar consolidation in all four quadrants 4

Hypoxaemia scorePa02/Fi02 >300 0Pa02/Fi02 225-299 1Pa02/Fi02175-224 2Pa02/Fi02 100-174 3Pa02/Fi02 <100 4

Respiratory system compliance>80 060-79 140-59 220-39 3<19 4

Positive end-expiratory pressure (PEEP) scorewhen ventilated (cmH20)

<5 06-8 112-14 3>15 4

The final value is obtained by dividing theaggregate sum by the number ofcomponents used

ScoreNo injury 0Mild to moderate injury 0.1-2.5Severe injury (ARDS) >2.5

Adapted from Murray et al. (1988).25

EPIDEMIOLOGY OF ACUTE LUNGINJURY/ACUTE RESPIRATORY DISTRESSSYNDROME

Variation by definition

At least ten epidemiological studies have shedlight on the incidence of ALI and ARDS. Figuresderived from these studies suggest an incidence from1.5 to 13.5 cases per 100000 inhabitants per yearfor ARDS. This wide range may be attributed toa lack of unified and agreed definitions, the varietyof ther-apies applied at the time the definingcriteria are sought, and the failure to define thepopulation within which patients with ARDS aredetected. However, two recent investigationsapplying the ARDS/ALI definitions of the 1994American-European Consensus Conference suggestthat 16% and 18% respectively of all mechanicallyventilated patients had ARDS.3,4 Surprisingly, only4-5% of these subjects had ALI, a markedly lowerfigure than that previously published.

Influence of precipitating conditions

The most important precipitating conditions forlung injury are sepsis, gastric aspiration, major traumaand multiple blood transfusions. The likelihood of apatient with any individual condition developinglung injury is also variable. The incidence of ARDSassociated with cardiopulmonary bypass surgery is1-2%, but it is 30-40% in patients with gastric aspi-ration. Although the extent to which the individualresponse to an inflammatory insult influences

Table 11.2 Criteria for the diagnosis of acute lung injury (ALI) and acute respiratory distress syndrome (ARDS)

ALI criteria

ARDS criteria

Acute onset

Acute onset

Pa02/Fi02

<300 mmHg

Pa02/Fi02

200 mmHg

Bilateral infiltrates seenon frontal chest X-ray

Bilateral infiltrates seenon frontal chest X-ray

<18 mmHg when measured or noclinical evidence of left atrialhypertension<18 mmHg when measured orno clinical evidence of left atrialhypertension

Adapted from Bernard et al. (1994).26

140 Pathophysiology of acute lung injury

progression to ARDS is unknown, high levels ofinterleukin-8 (IL-8) in broncho-alveolar lavage(BAL) fluid taken from trauma victims predict thedevelopment of ARDS with reasonable accuracy.5 Awide variety of endogenous defence mechanisms,including redox balance, may also influence individ-ual susceptibility.6'7 The extent to which the endogen-ous responsiveness to inflammatory insults correlates with the precipitating insult remains unclear,but the manifestation of lung injury probably differsamongst patients suffering indirect, as opposed todirect, pulmonary insults.8

Cause of death

The precipitating condition also influences clinicaloutcome. Trauma victims have a better prognosisthan those who develop ARDS secondary to gastricaspiration. Although the severity of respiratory fail-ure has been shown to predict mortality, patientswith ARDS rarely succumb to respiratory failure, butrather to multiple-organ dysfunction. Multivariateanalysis in a recent study indicated mortality to beassociated with septic shock, high severity of illnessscores and immunosuppression.3

Trends in mortality

Single European and American centres have report-ed mortality amongst patients with ARDS varyingfrom 25-70%. Recently, two long-term studiescontrolled for case-mix severity and performed inmajor referral centres on either side of the Atlanticshowed significant falls in mortality rates.9

However, a recent meta-analysis evaluating 101peer review articles published between 1967 and199410 did not detect a trend towards decreasedmortality rates and, world wide, any improvementis undoubtedly patchy: the reported mortality inSpain is 43%, whereas it remains around 60% inFrance.3 Such outcome variation is difficult toexplain, but it is possible that reduced mortalitymay be achieved by specialized centres, which havedefined admission criteria. In addition, manyof the published studies cited above found simi-lar mortality rates for ALI and ARDS, suggestingthat the severity of lung injury itself may not berelevant.

Impact on the conduct and outcomeof clinical trials

The failure of the majority of the large-scale clinicaltrials of therapeutic interventions in ARDS/ALI todemonstrate any mortality benefit may beexplained by

• variation in the definitions used and case-mixwith regard to the precipitating cause of theALI/ARDS,

• the intervention is ineffective under the circum-stances of the investigation or it has been appliedinappropriately, e.g. used too late, or given topatients in whom the intervention cannot workbecause the underlying condition is irreversible,

• the influence of lung injury and respiratory fail-ure on patient outcome in patients with ARDSmay be small relative to other factors,

• critically ill patients are, by definition, an inhomo-geneous group, and are therefore often treatedwith a number of co-interventions that may influ-ence the end-point under evaluation, e.g. the useof muscle relaxants and sedative agents in a trialusing ventilator days as an outcome measure.

The relationship between ALI/ARDS,sepsis and multiple organ dysfunctionsyndrome

Sepsis and its associated syndromes (Table 11.3)affect more than 1% of hospital patients and in up to40% of cases cause circulatory failure. An identifiablemicrobiological source of infection is found in lessthan 50% of patients, the remainder displaying thesystemic inflammatory response syndrome (SIRS).Many authorities consider that sepsis, SIRS and sep-tic shock represent a continuum in the severity of thehost response to non-infective and infective insultsand the extent of this response influences prognosis.Thus, the mortality for patients with SIRS is approxi-mately 7%, but rises to 50-90% for those with septicshock and multiple organ dysfunction syndrome(MODS).11,12

ARDS is now widely regarded as the pulmonarymanifestation of MODS. Pulmonary hypertensionwith increased pulmonary vascular resistance is com-mon, even in the setting of the lowered systemic vas-cular resistance that characterizes SIRS and sepsis.From the late 1980s, ARDS was known to be

Epidemiology of acute lung injury/acute respiratory distress syndrome 141

Table 11.3 Definitions of the systemic inflammatory response syndrome, sepsis, septic shock and multiple organdysfunction/failure

Systemic inflammatory response syndromeTwo or more of the following clinical signs of systemic response to endothelial inflammation:• a temperature of >38 °C or <36 °C• an elevated heart rate >90 beats min~1

• Tachypnoea, manifested by a respiratory rate of >20 breaths min-1 or hypoventilation (PaC02 <4.25 kPa)• an altered white blood cell count (>12 x 109 L-1, or <4 x 109 L-1, in the presence of more than 10% immature

neutrophils)

In the setting (or strong suspicion) of a known cause of endothelial inflammation, such as:• infection (Gram - or Gram +ve bacteria, viruses, fungi, parasites, yeasts or other organisms)• pancreatitis• ischaemia• multiple trauma and/or tissue injury• haemorrhagic shock• immune-mediated organ injury

SepsisThe systemic response to infection, manifest by two or more of the following as a result of infection:• a temperature of >38 °C or <36 °C• an elevated heart rate >90 beats min-1

• tachypnoea, manifested by a respiratory rate of >20 breaths min-1 or hypoventilation (PaC02 <4.25 kPa)• an altered white blood cell count (>12 x 109 L-1, or <4 x 109 L-1, in the presence of more than 10% immature

neutrophils)

Septic shockSepsis-induced hypotension (systolic blood pressure <90 mmHg or a reduction of >40 mmHgfrom baseline) despite

adequate fluid resuscitation

Multiple organ dysfunction syndromePresence of altered organ function in an acutely ill patient such that homeostasis cannot be maintained without intervention

Adapted from reference 27.

associated with endothelial dysfunction and disrup-tion and this has recently been characterized in vivousing non-invasive radioisotopic techniques.13 Therecognition that the refractory hypoxaemia of ARDSwas attributable to a loss of hypoxic pulmonary vaso-constriction in these patients highlighted the import-ance of vascular control mechanisms in determiningthe clinical characteristics of the syndrome and possi-bly also the development of multiple organ failure.Indeed, changes in vascular control have been docu-mented in both experimental models and patientswith sepsis uncomplicated by ARDS, and are charac-terized by systemic hypotension unresponsive to pres-sor agents and inotropes, possibly mediated throughchanges in the production of endothelially derivedvasomotor agents. The hypothesis that such sub-stances (Table 11.4) play a significant role in modulat-ing both systemic and pulmonary vascular tone underphysiological conditions was proven by the early1990s. This emphasized the importance of the barrier

and endocrine functions of the endothelium in deter-mining the clinical manifestations of SIRS/sepsis andparticularly in the development of MODS.

The way forward: epidemiology

There are a number of initiatives that may improveour understanding of the relevance of epidemiologyin determining outcome in ALI/ARDS and the wayin which the definition might be modified to benefitthe conduct of clinical trials.

• The issues that influence outcome need to bemore precisely identified, so that the influence ofpopulation homogeneity in determining trialinclusion criteria can be minimized.

• The extent to which defining criteria for ARDSmight be modified by the provision of an optimalregimen of mechanical ventilation and general


Table 11.4 Substances produced by the endothelium

ThrombomodulatoryThrombomodulinTissue plasminogen activatorHeparan sulphatesVon Willebrand factorEcto AD PasesTissue factor

VasoactiveProstacyclin, thromboxane and other prostanoidsNitric oxideEndothelins

Adhesion moleculesE-selectinICAM-1 and ICAM-2VCAM

Inflammatory moleculesPlatelet-activating factorCytokines: IL-6, IL-8, MCP-1Class II MHC molecules

ICAM, intercellular adhesion molecule; VCAM, vascular cell adhesionmolecule; IL, interleukin; MCP, monocyte chemoattractant protein, MHC,major histocompatibility molecules.

supportive measures (e.g. fluid balance, prone posi-tioning, use of nitric oxide (NO) etc.) needs to beidentified. Recruiting patients to clinical trials whomay meet ALI/ARDS criteria merely because theyare receiving suboptimal care at the time of assess-ment is clearly inappropriate. Standardization ofcare and the use of agreed management protocolsmight obviate this problem.The influence of inter-observer variability in apply-ing physiological and radiographic definitions ofALI/ARDS needs to be recognized and addressed.The therapeutic and prognostic significance ofdividing patients into ALI and ARDS categoriesneeds to be established before ALI can reasonablybe considered to be a relevant and separate entity:ALI will therefore not be discussed in the remain-der of this chapter.Criteria that define appropriate end-points forclinical trials must be agreed.

HISTOLOGICAL CHANGES WITHIN THE LUNG

the initiating event or the pathogenetic mechanisms.Characteristically, there is disruption of the alveolar-capillary unit, described as 'diffuse alveolar damage',comprising three overlapping stages: exudative, pro-liferative and fibrotic.

The exudative stage (Fig. 11.1) occurs within thefirst 48 hours and is characterized by the develop-ment of non-specific intra-alveolar oedema associ-ated with epithelial and endothelial cell damage. After24 hours, hyaline membranes are detected lining thealveoli and alveolar ducts, together with fibrinthrombi. Hyaline membranes are distinctive featuresof early ARDS and are composed of condensed plasmaproteins mixed with cell debris. Although there isevidence of early damage to the endothelium, there isoften extensive necrosis of type I pneumocytes.

The second, or proliferative, stage (>7 days) ismore variable and is characterized either by reso-lution of the changes already occurring or by organiza-tion of the intra-alveolar and interstitial exudates.The latter course involves formation of intra-alveolargranulation tissue from fibroblasts and myofibro-blasts, which eventually develops into dense fibrosis.Vascular remodelling begins at this stage and maycontribute to pulmonary hypertension.

The last, or fibrotic, stage is characterized byincreased fibrosis and collagen deposition, althoughcollagen and its precursor proteins can be identifiedmuch earlier. The lung consists of microcystic areasand irregular scarring together with larger cysticareas characteristic of chronic ARDS. There is exten-sive vascular remodelling with muscularization ofpre-acinar and intra-acinar vessels, leading (rarely)to irreversible pulmonary hypertension.

The response of the lung to injury is stereotypical,and histological examination rarely identifies either

Figure 11.1 Histopathology of the 'exudative' early phase of

ARDS, demonstrating distorted alveolar spaces filled with

inflammatory cells and lined with hyaline membranes.

Pathogenesis 143

PATHOGENESIS

The precipitating causes of ARDS can be divided intothose causing a direct insult to the lung and those inwhich the lung injury complicates a more remotedisease process. This highlights two importantpoints. First, although the response of the lung toinjury may be uniform in a histopathological sense,there is unlikely to be a single pathogenetic pathway.Second, circulating inflammatory mediators are like-ly to be important in modulating these processes, atleast as far as distant insults are concerned.Inflammatory mediators, such as cytokines, are

known to induce an acute inflammatory response inthe microvasculature of the lung and other organs.Cells activated by such a process in turn producemore inflammatory mediators, leading to endothe-lial and epithelial cell damage and ultrastructuralchanges that cause increased permeability of thealveolar capillary membrane. Vasoactive mediators,including NO, cyclo-oxygenase (COX) products andendothelin-1 (ET-1), regulate vascular tone at a locallevel under physiological conditions, but alterationsin their generation and site of release have beendescribed in animal models of sepsis/ARDS, leadingto disordered vascular control. In addition, activa-tion of platelets and the complement and coagulation

Figure 11.2 Overview of the

pathogenesis of ALI/ARDS. VSM,

vascular smooth muscle; NO, nitric

oxide; ET, endothelin; V/Q,

perfusion/ventilation; HPV,

hypoxic pulmonary vasoconstriction.


cascades leads to the formation of microthrombi.The clinical consequences of these events are impairedhypoxic pulmonary vasoconstriction and ventilation- perfusion relationships leading to refractoryhypoxaemia, alveolar oedema formation and increasedpulmonary vascular resistance (Fig. 11.2).

Neutrophil-dependent lung injury

Histological specimens from the lungs of patientsdying with ARDS demonstrate sequestration andsubsequent migration of neutrophils into lung tis-sue. In addition, BAL fluid taken from patientsearly in the course of the syndrome demonstratesneutrophilia,14 the extent of which correlates withlevels of granulocyte colony stimulating factors(G-CSF) and dictates disease progression andseverity. Increased levels of the neutrophil secret-ory products elastase and collagenase have alsobeen described. In vitro, neutrophil-mediatedcell injury can only occur when the cell is in closeproximity to the capillary endothelium. In thepulmonary capillary, this is likely to occur by twomechanisms (Fig. 11.3):

• neutrophils become rigid, inhibiting flow throughsmaller capillaries,

• increased adhesion between the neutrophil andthe endothelium develops, secondary to up-regu-lation of adhesion molecules on both cells (see thesection on endothelial activation).

Once activated and in the presence of high appliedoxygen concentrations (FiO2), neutrophils andmacrophages release reactive O2 species, leading tothe production of the toxic hydroxyl radical, which,in the presence of reactive iron species, causes oxida-tive damage to lipids and proteins.

Despite such evidence suggesting a central role forthe neutrophil in the evolution of ARDS, neu-tropenic patients can develop ARDS, and increasingcirculating neutrophil numbers with G-CSF does notlead to more severe lung injury.

Endothelial activation

The endothelial cell exerts active control over vascu-lar tone, thrombosis and permeability, through thesynthesis and release of a wide variety of substances(Table 11.4, Fig. 11.4). Under inflammatory condi-tions, the endothelial cell is activated, leading to aloss of vascular integrity, increased expression of leu-cocyte adhesion molecules, HLA molecules andcytokines, and becomes pro-thrombotic. Two stages

Figure 11.3 Early events in the

pathogenesis of ARDS, with

particular reference to neutrophil-

mediated lung injury. IL-1, IL-6,

IL-8, IL-10, interleukins 1, 6, 8, 10;

TNF-a, tumour necrosis factor-a;

ROS, reactive oxygen species; MIF,

macrophage inhibitory factor.

Pathogenesis 145

Figure 11.4 The maintenance of

vascular homeostasis by the tonic

release of mediators such as nitric

oxide (NO), prostacyclin (PGI2) and

endothelin-1 (ET-1). AA, arachidonic

acid; PLA2, phospholipase A2;

COX1, cyclo-oxygenase 1; ppET,

endothelin; ECE, eicoanoids; eNOS,

of activation occur. Type I activation does notrequire either de-novo protein synthesis or genotypicup-regulation. Endothelial cells retract from eachother, express P-selectin leading to increased neu-trophil adhesion, and release von Willebrand factor,which regulates platelet adherence to the subendo-thelium. By contrast, Type II activation requires up-regulation of mRNA expression and de-novo proteinsynthesis, particularly of cytokines and adhesionmolecules.

There appear to be common intracellular controlmechanisms involved in both processes, such as thosemediated through the intracellular messenger nuclearfactor-KB (NF-KB). The endothelium produces vascu-lar cell adhesion (VCAM-1) and intercellular adhesion(ICAM-1 and ICAM-2) molecules and E-selectin,facilitating the binding of leucocytes. Simultaneously,activated neutrophils express a complementarysequence of surface adhesion molecules termed inte-grins, the most significant of which is a CD 11/CD 18complex that determines the migration of neutrophilsinto the interstitium. This adhesion cascade isreviewed in detail elsewhere,15 but is associated withincreased expression of endotoxin/cytokine-induciblegenes that are significant in determining vasomotorcontrol, particularly those encoding for the produc-tion of NO, ET-1 and COX products (Fig. 11.5). These

mediators and their role in modulating the vascularcontrol under inflammatory conditions have recentlybeen reviewed.16 Recent work has demonstrated thatthe underlying smooth muscle can be an importantsource of vasoactive mediators under inflammatoryconditions,17 and human systemic and pulmonaryartery smooth muscle can both produce ET-1 andCOX products when stimulated with cytokinesin vitro.l7-19 These cells may be particularly impor-tant if the endothelial layer is dysfunctional or miss-ing. Additionally, autocrine mechanisms may causevasoconstriction, cellular proliferation and vascularremodelling.

Epithelial injury

The acute phase of ARDS is characterized by thepresence of protein-rich fluid in the alveolar space.Consequently, there must be increased permeabilityof the alveolar-capillary membrane due to breachesof both endothelial and epithelial cells. The degreeof epithelial cell damage is increasingly recognizedas clinically significant and is an important predic-tor of outcome.14 The normal epithelium is com-posed of both type I and type II pneumocytes. TypeI cells are more prevalent under normal conditions


Figure 11.5 Disruption of

vascular control mechanisms in

sepsis. Under these conditions the

smooth muscle itself is an impor-

tant producer of these mediators.

The relative proportions of the

vasodilators nitric oxide (NO)

and prostacyclin (PGI2) and the

vasoconstrictor and mitogen,

endothelin-1 (ET -1), determine the

vascular response in the affected

area. LPS, lipopolysaccharide;

CD14, cluster of differentiation 74;

ppET-1, prepro-endothelin 7; TNF,

tumour necrosis factor; COX2,

cyclo-oxygenase 2; IL-1, interleukin

7; /A/OS, inducible nitric oxide

synthase.

(90:10), but type II cells are more resistant to injury,as well as having important functions such as sur-factant production, ion transport and differentia-tion to type I cells after injury. Damage to type IIcells may:

• disrupt fluid transport, leading to impaired fluidremoval,

• impair surfactant production,• influence the repair process, increasing the likeli-

hood of fibrosis developing.

Resolution

Enhancing natural resolution processes may be asimportant as attenuating early inflammation.Alveolar oedema is cleared by an active processinvolving sodium transport from the distal airwaysinto the interstitium, followed by movement of wateracross the epithelium. It now appears that thisprocess involves specialized water channels termedaquaporins, located on type I cells. However, little isknown about the regulation of these channels. Theability to clear fluid appears to be associated with abetter prognosis. Soluble protein is removed by pas-sive diffusion, whereas insoluble protein is removedby endocytosis across epithelial cells and phagocyt-osis by macrophages. Epithelial integrity is restored

by type II cells, which re-epithelialize the denudedalveolar epithelium by differentiating into type Icells. This process presumably requires adhesion,spreading, migration and proliferation of the type IIcells. Although this is an area of active research, suchprocesses require cell—cell and cell-matrix interac-tions, which are probably mediated by epithelialintegrins. In addition, such processes need modula-tors such as cytokines and growth factors. Tumourgrowth factor-b (TGF-(3) is of particular interest as ithas the ability to regulate the expression of epithelialintegrins, as well as being a potent inhibitor ofepithelial cell proliferation.

Resolution of accumulated cellular infiltrate andfibrosis must also occur. Apoptosis, or 'pro-grammed cell death', may be involved in clearingneutrophils. Certainly, markers of apoptosis arehigh in the pulmonary oedema fluid of patients andsuch fluid induces epithelial cell apoptosis in vitro.This is likely to represent an important area forfuture research.

Fibrosis

In those patients in whom resolution does not occur,progression to the late stage of ARDS, characterizedby intra-alveolar and interstitial fibrosis, is inevitable.Interestingly, fibrosis appears to occur rapidly and is

Pathogenesis 147

present as early as the first week. The development offibrosis is associated with hypoxaemia, reduced lungcompliance, ventilator dependence and a pooroutcome. N-terminal procollagen III (N-PCP-III), aprecursor of collagen III, is actively synthesized in thenormal lung. N-PCP-III levels are raised in BAL fluidand in the serum of ARDS patients as early as 24—48hours after ventilation.20,21 In addition, BAL fluidtaken from ARDS patients within 48 hours is alsointensely mitogenic for fibroblasts. Such data suggestthat the mechanisms capable of producing fibrosisare in place early in the course of the disease.Although epithelial, endothelial and smooth musclecells can produce collagen, the principal source is thefibroblast. Cytokines released from these cell typesmay also stimulate collagen synthesis and fibroblastproliferation. In addition, macrophages are animportant source of TGF-b, insulin-like growthfactor (IGF-1) and ET-1, which are all important pro-fibrotic cytokines.

Cytokines

Cytokines are released by cells under stress and maybeinvolved in the initiation and progression of ARDS.Broadly, there are pro-inflammatory cytokines, suchas tumour necrosis factor-a (TNF-a), interleukin-1b(IL-1b, interferon-'Y (IFN-"y) and IL-8, and anti-inflammatory cytokines, such as IL-10, IL-11 andmacrophage inhibitory factor (MIF), as well as relatedmolecules such as IL-1 receptor antagonist (IL-1 Ra)and soluble TNF receptor and antibodies againstIL-8.14 In inflammatory conditions, there seems to bea local or systemic imbalance of these cytokines.The balance may change with the progression of thedisease. Clearly, in studying the role of cytokinesin ARDS, biological activity is of paramountimportance, rather than simply the levels.

Other cells implicated in acuterespiratory distress syndrome

EOSINOPHILS

Eosinophil activation has been related to lung dam-age in ARDS. Levels of eosinophil cationic protein(ECP) are higher in the BAL fluid taken frompatients with ARDS than in that from controls andthe levels relate to severity of the disease.

MAST CELLS

Mast cells have been detected in 50% of cases of pro-gressive ARDS studied by immunohistochemicalmethods. Although appearing after the onset offibrosis, tryptase, stored in mast cell granules, is amitogen for fibroblasts in vitro and may contributeto the propagation of the process.

PLATELETS

Progressive thrombocytopenia occurs in 50% ofpatients with non-traumatic ALI, its severity parallel-ing the development of worsening hypoxaemia. Atleast some of these platelets appear to be trapped inthe lung microvasculature, as autopsy specimensdemonstrate thrombi containing enmeshed platelets,neutrophils and fibrin. Platelets, once activated,release the contents of their granules, which have thepotential to participate in the pathogenesis of lunginjury. Serotonin and thromboxane can cause vaso-constriction, whereas TGF-a, TGF-b and platelet-derived growth factor (PDGF) are mitogens and maytherefore take part in vascular remodelling and fibro-sis. However, despite many animal and clinical studies,it is still not known whether the changes in plateletnumbers and function are directly related to thepathogenesis of ARDS or merely an epiphenomenon.

MACROPHAGES

Macrophages produce pro-inflammatory cytokinessuch as IL-1b and TNF-ct as well as growth factors suchas PDGF, TGF-b, IGF and ET-1. They may thereforeplay a key role in both the initiation and propagationof ARDS. However, whereas high neutrophil counts inBAL fluid from patients with ARDS are associated withmore severe disease and a worse outcome, an increasein BAL fluid macrophage count is associated with theresolution of lung injury and improved survival. Thus,the macrophage appears to be capable of both pro-inflammatory and anti-inflammatory actions.

Ventilator-induced lung injury

Substantial interest has recently been directed at therole of mechanical ventilation in the developmentand propagation of lung injury. Studies have demon-strated that ventilation at low tidal volumes, com-pared with traditional tidal volumes, can improveoutcome for patients with ARDS. Lower plasma levels


of IL-6 were also found in those patients ventilatedat lower tidal volumes.22 Several animal experimentshave indicated that ventilation at high tidal volumesleads to damage to the epithelium and endothelium,with consequent inflammation, oedema formation,atelectasis, hypoxaemia and the release of cytokines.Alveolar over-distension, together with repeated col-lapse and re-opening of atelectatic alveoli, alsoresults in an increase in systemic cytokine levels andthis may contribute to the development and propa-gation of MODS. A recent report demonstrated thatthe use of a 'lung-protective' ventilatory strategyreduced both the pulmonary and systemic cytokineresponses 23

PULMONARY AND SYSTEMIC MARKERS OFLUNG INJURY

Three reasons exist for measuring systemic markersin ARDS:

• they may identify patients likely to develop thesyndrome in an 'at-risk' population,

• they may provide information concerning thepathogenesis of ARDS,

• they might represent a method for monitoringprogress and predicting outcome.

Samples for analysis may be collected from blood(plasma/serum) or from the distal airways (pul-monary oedema or BAL fluid). The pros and cons ofsampling oedema fluid or BAL fluid will not be dis-cussed further, except to say that BAL is more widelyperformed but suffers from the complication of anunknown dilution factor of the alveolar fluid.

CD14 and lipopolysaccharide

CD 14 is a pattern-recognition molecule present onthe cell surface of macrophages and neutrophils. It is areceptor for lipopolysaccharide (LPS) complexes andLPS binding protein (LBP). In turn, CD 14 employs'toll-like' receptors for signal transduction. A solubleform of CD 14 (sCD14) is released from cell mem-branes and complexes with LPS and mediates LPS-dependent response in CD14-negative cells, such asendothelial and epithelial cells. Soluble CD 14 and LBPlevels are significantly increased in BAL fluid from

patients with ARDS. In addition, the levels correlatewith BAL total protein and neutrophil number, bothmajor indices of lung inflammation. However, neithersCD14 nor LBP levels alone predict mortality. LPSitself has been measured in the plasma of patients atrisk of and with ALI/ARDS. Unfortunately, there wereconflicting results with regard to the predictive valueof raised levels, although there were problems withpatient selection and measurement methodology.

Early-response cytokines

TNF-a and IL-1 b are early response cytokines pro-duced by macrophages and neutrophils. They caninduce epithelial and endothelial cells, fibroblastsand smooth muscle cells to produce furthercytokines, thereby initiating the inflammatory cas-cade (see section on pathogenesis). TNF-a isproduced initially in a membrane-associated 26-kDform, which is cleaved by TNF-a converting enzyme(TACE) into the soluble 17.5-kD cytokine.

TUMOUR NECROSIS FACTOR-a

Soluble TNF-a levels are increased in the plasma andBAL fluid of patients with ARDS. Studies have triedto establish a link between levels of plasma-solubleTNF-a and the risk of the development of ALI orARDS. Disappointingly, the results have been mixed.It appears that circulating levels of TNF-a are moreindicative of the extent of lung injury than of theactual diagnosis. In addition, the administration ofanti-TNF-a receptor antibody does not reduce mor-tality in patients with sepsis. More recently, increasedfunctionally active membrane-associated TNF-a onalveolar macrophages has been identified in patientswith ARDS, suggesting that surface expression offunctionally active TNF-a may be an importantmechanism for TNF-a-mediated lung injury. Thismay partly explain the disappointing results fromstudies investigating the use of soluble TNF-a.

INTERLEUKIN-13

IL-1b levels are increased and are biologically activein the BAL fluid from patients at risk from and in theearly stages of ARDS. These levels are sustainedthroughout the course of the syndrome and persis-tently high concentrations are predictive of a pooroutcome. Increased levels in BAL fluid from these

Pulmonary and systemic markers of lung injury 149

patients are in part due to increased production byalveolar macrophagess. Plasma levels of IL-1 b do notpredict those patients at risk of developing ARDS.14

IL-1b receptor antagonist (IL-1 Ra) is the naturallyoccurring antagonist of IL-lb and is elevated inpatients with ARDS, but IL-1b is tenfold in excess ofthis, so that the equilibrium is shifted favouring apro-inflammatory state. Low concentrations of IL-1Ra and IL-10 (another anti-inflammatory cytokine)in the BAL fluid of patients in the early phase ofARDS are associated with a poor prognosis. IL-1b isable to initiate a cascade of inflammatory events,including the release of prostaglandins, increasedsynthesis of collagenases and the migration of neu-trophils through the endothelium and the produc-tion of other inflammatory cytokines such as IL-8and macrophage inflammatory protein-1 (MIP-1).

Chemokines

Chemokines are molecules that direct leucocytemigration. Two major classes exist: the a or CCchemokines, such as MIP-1, and the b or CXCchemokines, such as IL-8 and ENA-78, which arepotent neutrophil chemotactic agents. Of these, IL-8is by far the most studied in ARDS. IL-8 is detectablein the BAL fluid of patients at risk of developingARDS and may be predictive of those who subse-quently develop the established syndrome.5 Alveolarmacrophages are an abundant source of thischemokine when exposed to LPS, although othercells are able to produce it when exposed to cytokinessuch as TNF-a and IL-1 (3. IL-8 measured in BALfluid correlates with its neutrophil concentrations,but not with the severity of the lung injury. Althoughearly studies found that IL-8 levels were highest inthe BAL fluid of non-survivors, subsequent studiessuggest that levels do not predict an adverse clinicalcourse. IL-8 cannot be easily measured in bloodbecause of binding to red blood cells. Moreover,inhibitory factors such as endogenous immuno-glubulin G (IgG) auto-antibodies and a2-macro-globulin bind and neutralize the molecule.

lnterleukin-6

IL-6 was originally described as a B-cell growthfactor although, in the context of inflammatory

conditions, it is believed to be an 'orchestrator', stimu-lating acute-phase responses in the liver. It is pro-duced by activated alveolar macrophages in responseto other cytokines such as IL-1b and TNF-a. IL-6could also be directly involved in the pathogenesis ofARDS by stimulating neutrophils to release elastase.IL-6 levels are raised in the BAL fluid of patients atrisk for ARDS, and remain elevated throughout thecourse of the disease. However, concentrations donot predict the onset or outcome of ARDS. In add-ition, a soluble receptor, IL-6R, is raised in the BALfluid of patients at risk and during the course ofARDS. Interestingly, this receptor is an agonist,unlike IL-lRa. Recently, IL-6 levels in the plasma ofpatients with ARDS have been shown to be lower inthose ventilated with protective, lower tidal volumes,suggesting a possible role for this cytokine in ventila-tor-induced lung injury.22

lnterleukin-10

IL-10, as already mentioned, is an anti-inflammatorycytokine that is believed, like IL-4 and IL-13, tocounteract or balance the pro-inflammatorycytokines. IL-10 is present in the BAL fluid ofpatients with ARDS, although in low quantities.Indeed, patients who die with ARDS appear to havevery low levels of IL-10 in their BAL fluid.

Macrophage inhibitory factor

Macrophage inhibitory factor (MIF) has a complexrole, antagonizing the effect of cortisol on cytokineproduction by alveolar macrophages. As such, MIFmay acfto sustain the inflammatory response. It ispresent in the BAL fluid of patients studied on thefirst day of ARDS, and immunoreactive MIF isdetectable in isolated macrophages. MIF is also pre-sent in the BAL fluid of patients at risk of ARDS, andthe concentration rises as the disease progresses.

Growth factors and collagen

Transforming growth factor-a (TGF-a) increasesfibroblast collagen production and may be impor-tant in the development of fibrosis in ARDS. Indeed,TGF-a is present in the BAL fluid of patients with


sustained ARDS, and high levels in BAL fluid on day7 are associated with a poorer prognosis. TGF-a isalso a mitogen for epithelial cells in vitro. Procollagenpeptide III (PGP III) is a marker of collagen synthe-sis and is detectable in BAL fluid at the onset ofARDS, with high levels being associated with a poorprognosis. In addition, PDGFs are present that stimu-late fibroblast proliferation in vitro and also the pro-duction of an angiogenic factor that is present in theBAL fluid in 70% of patients with ARDS. Clearly,more work is needed in this area as these and othergrowth factors are probably key mediators in theremodelling processes that occur in the lung duringthe course of the disease.

Coagulation cascade

Coagulation cascade activation is a feature of ARDS.However, conventional laboratory measurements ofcoagulation parameters do not distinguish betweenpatients with and without ALL Despite this, there is awealth of evidence for coagulation abnormalities inthe pathogenesis of ARDS. Fibrin deposition is foundin the lungs as well as micro-emboli in the vascula-ture of patients with ARDS. The BAL fluid of patientswith ARDS is pro-coagulant, containing tissue fac-tor-factor VII/VIIa complexes, and it is thereforecapable of activating the extrinsic coagulation cas-cade via the generation of factor X. This pro-coagu-lant property increases for 3 days following diagnosisand then decreases over a 2-week period. Althoughfibrin is present in the lungs of patients with ARDS, ithas not been possible to detect thrombin in the BALfluid of patients with ARDS. However, thrombin levelsare increased in the lungs of patients with pulmonaryfibrosis associated with systemic sclerosis. Apart fromits pro-coagulant properties, including platelet aggre-gation, thrombin is also capable of increasing vascu-lar permeability, inducing pro-inflammatory cytokinesand promoting fibrosis, by inducing fibroblastproliferation and the production of growth factors.There is also evidence for reduced fibrinolysis, encour-aging extravascular fibrin deposition. Plasminogenlevels are higher in patients with ARDS compared tocontrols and those at risk of ARDS, although mostplasminogen appears to be bound to local inhibitors.In addition, there appear to be reduced amountsof naturally occurring anticoagulants such as anti-thrombin III, protein C and protein S. There is also

evidence of increased antifibrinolytic activity in the BALfluid of ARDS patients, with increased levels ofurokinase inhibitors such as plasminogen activatorinhibitor (PAI-1). However, another study found thatplasma levels of PAI-1 (and factor VIII) were not pre-dictive of the development of ARDS in patients at risk.

Reactive oxygen species

Once activated, and in the presence of the high O2

concentrations frequently necessary to supportpatients with ARDS, neutrophils and macrophagesrelease ROS, leading to the production of thehydroxyl radical, which, in the presence of reactiveiron species, causes oxidative damage to lipids andproteins. Xanthine oxidase is a key enzyme in theproduction of ROS, and its enhanced production aswell as the increased production of its substratesxanthine and hypoxanthine are seen in non-survivors of ARDS, suggesting that they experiencemore oxidative stress than survivors.7 Increased levelsof the free radical species hydrogen peroxide (H2O2),formed by the action of O2 radicals and water, havealso been demonstrated in urine in the first 48 hoursin patients with ARDS compared to controls. Asthese pro-oxidant molecules are generally unstable,and therefore difficult to measure, it is sometimeseasier to measure anti-oxidant molecules in patientswith ARDS. Plasma levels of catalase, an enzyme thatscavenges H2O2, are higher in septic patients withlung injury than in those without. In addition,increased serum levels of catalase and manganesesuperoxide dismutase (SOD) have been shown topredict the development of lung injury in an at-riskpopulation. Furthermore, plasma iron-binding anti-oxidant activity is lower in patients with ARDS thanin controls and this correlates with the percentageiron saturation of transferrin.6 Indeed, in patientswith ARDS and abnormal liver function, low-molec-ular-mass iron has been detected in plasma, andtransferrin was shown to be fully saturated with iron,resulting in greatly decreased anti-oxidant protec-tion. Caeruloplasmin is another primary plasmaanti-oxidant and'acute phase protein, capable ofbinding to iron and therefore regulating iron-depend-ent oxidative reactions. Both proteins are raised inBAL fluid and in the serum of patients at risk of andwith established lung injury. However, despite highlevels of caeruloplasmin in the BAL fluid of patients

References 151

with ARDS, the ferroxidase activity appears dimin-ished, perhaps due to proteolytic activity.

trauma, pancreatitis and perforation of a viscus,identifying a group of patients at very high risk ofARDS.

Selectins

Neutrophils migrate from the vascular compartmentby a multistage process (discussed above in the sec-tion on pathogenesis). Selectins form a family of mol-ecules that are important in the early attachment ofneutrophils. E-selectin and P-selectin are expressedon the surface of the endothelium, whereas L-selectinis expressed only on leucocytes. Soluble L-selectin(sL-selectin) inhibits neutrophil binding to endothe-lial cells in vitro in a concentration-dependent manner.Plasma levels of sL-selectin are lower in patients atrisk of or who develop ARDS. In addition, low valuesare associated with an increased mortality. Reducedlevels of sL-selectin may therefore represent a markerof panendothelial activation in ARDS.

THE WAY FORWARD: INVESTIGATING'AT-RISK' POPULATIONS

Predicting which patients are at risk of developingALI or ARDS is potentially important for futurestudies of both the pathogenesis of ARDS and poten-tial therapeutic interventions. This may be particu-larly relevant to the identification of geneticpolymorphisms that increase the likelihood of apatient developing lung injury following a specificinsult. However, there are several problems withdefining the 'risk' or 'latent' period:

• only 5% of patients with a risk factor for ARDSdevelop the full syndrome,24

• this period may be very variable and can rangefrom hours to days if the insult is distant from thelung,

• studies examining this risk period may select theirdata from different times, often very late in thisperiod.

Studies into the subclinical lung damage will never-theless provide opportunities for defining the earlydisease mechanisms. In support of this last point,reports have been published describing altered pul-monary levels of IL-85 and blood levels of sL-selectinand neutrophil elastase within 90 min of multiple

CONCLUSIONS

Despite 30 years of interest in ARDS, its precisepathogenesis remains elusive. Many different cellsand mediators are involved. Consequently, manipu-lation of a single mediator is unlikely to producemajor therapeutic benefit. Determining whichpatients are at risk of developing ARDS may prove tobe of particular importance. This will involve notonly measuring markers in BAL fluid and plasma asearly as possible in the course of the disease, but alsodetermining which patients are genetically predis-posed to develop the syndrome.

REFERENCES

1. Ashbaugh, DG, Bigelow, DB, Petty, TL, et al. Acute

respiratory distress in adults. Lancet 1967; 2: 319-23.

2. Bernard, GR, Artigas, A, Brigham, KL, et al. The

American-European Consensus Conference on ARDS.

Definitions, mechanisms, relevant outcomes, and

clinical trial coordination. Am J Respir Crit Care Med

1994; 149: 818-24.

3. Roupie, E, Lepage, E, Wysocki, M, et al. Prevalence,

etiologies and outcome of the acute respiratory

distress syndrome among hypoxemic ventilated

patients. SRLF Collaborative Group on Mechanical

Ventilation.Soc/ete de Reanimation de Langue

Francaise. Intensive Care Med 1999; 25: 20.

4. Luhr, OR, Antonsen, K, Karlsson, M, et al. Incidence

and mortality after acute respiratory failure and acute

respiratory distress syndrome in Sweden, Denmark,

and Iceland. The ARF Study Group. Am J Respir Crit

Care Med 1999; 159: 1849-61.

5. Donnelly, SC, Stricter, RM, Kunkel, SL, et al.

lnterleukin-8 and development of adult respiratory

distress syndrome in at-risk patient groups. Lancet

1993; 341: 643-7.

6. Gutteridge, JM, Quinlan, GJ, Mumby, S, et al. Primary

plasma antioxidants in adult respiratory distress

syndrome patients: changes in iron-oxidizing,

iron-binding, and free radical-scavenging proteins.

J Lab Clin Med 1994; 124: 263-73.


7. Quinlan, GJ, Lamb, NJ, Til ley, R, et at. Plasma

hypoxanthine levels in ARDS: implications for

oxidative stress, morbidity, and mortality. Am J Respir

Crit Care Med 1997; 155: 479-84.

8. Gattinoni, L, Pelosi, P, Suter, PM, et at. Acute respiratory

distress syndrome caused by pulmonary and

extrapulmonary disease. Different syndromes? Am J


9. Abel, SJ, Finney, SJ, Brett, SJ, et al. Reduced mortality

in association with the acute respiratory distress

syndrome (ARDS). Thorax 1998; 53: 292-4.

10. Krafft, P, Fridrich, P, Pernerstorfer, T, et al. The acute

respiratory distress syndrome: definitions, severity

and clinical outcome. An analysis of 101 clinical

investigations. Intensive Care Med 1996; 22: 519-29.

11. Rangel-Frausto, MS, Pittet, D, Costigan, M, et al. The

natural history of the systemic inflammatory response

syndrome (SIRS). A prospective study. J Am Med Assoc

1995;273:117-23.

12. Marshall, JC, Cook, DJ, Christou, NV, et al. Multiple

organ dysfunction score: a reliable descriptor of a

complex clinical outcome. Crit Care Med 1995; 23:

1638-52.

13. Sinclair, DG, Braude, S, Haslam, PL, et al. Pulmonary

endothelial permeability in patients with severe lung

injury. Clinical correlates and natural history. Chest

1994; 106: 535-9.

14. Pittet, JF, Mackersie, RC, Martin, TR, et al. Biological

markers of acute lung injury: prognostic and

pathogenetic significance. Am J Respir Crit Care Med

1997;155:1187-205.

15. Albelda, SM, Smith, CW, Ward, PA. Adhesion molecules

and inflammatory injury. FasebJ 1994; 8: 504-12.

16. Wort, SJ, Evans, TW. The role of the endothelium in

modulating vascular control in sepsis and related

conditions. Br Med Bull 1999; 55: 30-48.

17. Woods, M, Mitchell, JA, Wood, EG, et al. Endothelin-1

is induced by cytokines in human vascular smooth

muscle cells: evidence for intracellular

endothelin-converting enzyme. Mol Pharmacol 1999;

55: 902-9.

18. Jourdan, KB, Evans, TW, Lamb, NJ, et al. Autocrine

function of inducible nitric oxide synthase and

cyclooxygenase-2 in proliferation of human and rat

pulmonary artery smooth-muscle cells: species

variation. AmJ Respir Cell Mol Biol 1999; 21:105-10.

19. Wort, SJ, Mitchell, JA, Woods, M, et al. The

prostacydin-mimetic cicaprost inhibits endogenous

endothelin-1 release from human pulmonary artery

smooth muscle cells. 7 Cardiovasc Pharmacol 2000;

36: S410-13.

20. Marshall, RP, Bellingan, G, Webb, S, et al.

Fibroproliferation occurs early in the acute respiratory

distress syndrome and impacts on outcome. Am J


21. Chesnutt, AN, Matthay, MA, Tibayan, FA, et al. Early

detection of type III procollagen peptide in acute lung

injury. Pathogenetic and prognostic significance. AmJ






Med 2000; 342: 1301-8.

23. Ranieri, VM, Suter, PM, Tortorella, C, et al Effect of

mechanical ventilation on inflammatory mediators in

patients with acute respiratory distress syndrome: a

randomized controlled trial. .Mm Med Assoc 1999;

282: 54-61.

24. Baumann, WR, Jung, RC, Koss, M, et al. Incidence and

mortality of adult respiratory distress syndrome: a

prospective analysis from a large metropolitan

hospital. Crit Care Med 1986; 14: 1-4.

25. Murray, JF, Matthay, MA, Luce, JM, Flick, MR. An

expanded definition of the adult respiratory distress

syndrome. Am Rev Respir Dis 1988; 138: 720.

26. Bernard, GR, Artigas, A, Brigham, KL, et al. The

American European Consensus Conference on ARDS.

Definitions, mechanisms, relevant outcomes, and clin-

ical trial coordination. Am J Respir Crit Care Med 1994;

149: 818-24.

27. American College of Chest Physicians/Society of

Critical Care Medicine Consensus Conference.

Definitions for sepsis and organ failure and guidelines

for the use of innovative therapies in sepsis. Crit Care

Med 1992; 20: 864-74.

12Management of acute lung injuryKEITH G HICKLING AND ANDREW BERSTEN

Introduction 153Ventilator-induced lung injury 153The lung in acute respiratory distress syndrome 155Ventilation strategies 156Posture 163Surfactant replacement therapy 164

Manipulation of the pulmonary vasculature 165Anti-inflammatory drugs 166Fluid therapy 167Summary 167References 167

INTRODUCTION

Until recently the mortality (50-60% in most studies)amongst patients with the acute respiratory distresssyndrome (ARDS) has remained unchanged since thecondition was first described, despite increasedunderstanding of its pathophysiology and the use ofspecific treatments directed at these pathologicalprocesses. The recognition that mechanical ventila-tion can itself cause acute lung injury (ALI), and mayhave actually contributed to the high mortality rate,has led to the development of 'lung-protective' venti-lation strategies that have resulted in improved out-come. A number of uncontrolled studies reportmortality rates of 12-30% using lung-protective ven-tilation (Table 12.1), a mortality lower than historicalcontrols. Two randomized trials have shown markedreductions in mortality in patients managed withlung-protective ventilation, from 71% to 38%,p < 0.001,n and from 39.8% to 31%, p = 0.007.12

However, three other randomized trials have not beenable to demonstrate a reduction in mortality,although the explanation may be that the controlgroups were not exposed to sufficient levels of lungdistension to be injurious. As the potential for caus-ing ventilator-induced lung injury (VILI), and its

associated morbidity, has been recognized, the goal ofmechanical ventilation in ARDS has shifted fromachieving optimum oxygenation to the avoidance ofVILI. The extent to which these goals differ is, how-ever, unclear.

VENTILATOR-INDUCED LUNG INJURY

Animal studies have shown that mechanical ventila-tion produces an acute parenchymal lung injury thatis histologically very similar to that seen in ARDS.This is associated with an inflammatory response inthe lung, characterized by granulocyte infiltration,the release of pro-inflammatory and anti-inflam-matory cytokines and chemokines into the alveolarspaces,13 and increased micro vascular permeability inthe lung and systemic tissues. Studies have also shownthat VILI results in the release of cytokines into thesystemic circulation in both animals and patientswith ARDS.13 In animals, following the instillation ofbacteria into the trachea, injurious ventilation resultedin a high frequency of positive blood cultures (termed'bacterial translocation' from the lung), whereasblood cultures were negative in most animals venti-lated with non-injurious ventilation.

154 Management of acute lung injury

Table 12.1 Uncontrolled studies of outcome in ARDS managed with lung-protective ventilation and permissive hypercapnia

50

949

53

20

53

40

23

54

25

ARDS, LIS > 2.5

ARDS

Severe ARDS referredfor ECC02RARDS, LIS > 2.5

ARDS, ECMOcrit LIS 3.6 ± 0.2ARDS, LIS > 2.5

Children with ARDS

Children with ARDSLIS > 2.5Children, burns

Trauma/sepsis

PH, PIP < 30-40 cmH20always <40 cmH20PHPH < ECC02R (24),NO (7), FPCPH, PIP<30-40cmH20

PH, NO, TGI, prone positioning

Mild PH, PH > 7.3, PPL39 ± 15, PIP 44 ± 16PH, PCV, PIP < 35-40PaC02 71 ± 23 mmHgPH, PIP < 35 cmH20,PaC02 39-94 mmHgPIP<40cmH20,PaC0239-111 mmHgConsensus criteria LIS > 2.5

16

2224

26

30

40

30

17.4

0

12

Hicklingef al.(1990)1

Tothet al. (1992)2

Lewandowski et al.(1992)3

Hicklingeffl/.(1994)4

Levy et al.(1995)5

Thomsen et al.(1994)6

Nakagawa and Bohn(1995)7

Botero et al.(1995)8

Sheridan et al.(1995)9

Stacker et al. (1997)10

PH, permissive hypercapnia; US, lung injury score; NO, nitric oxide inhalation; TGI, tracheal gas insufflation; FPC, frequent body position changes; PCV, pressure-control ventilation; PIP, peak inspiratory pressure; 1979 ECMO study criteria, criteria used for selecting patients for entry to the 1979 randomized trial ofextracorporeal membrane oxygenation in severe ARDS.

Ventilation-induced lung injury may lead to a sys-temic inflammatory response, organ dysfunctionand even sepsis.

These findings have led to speculation that VILImay result in the systemic inflammatory responsesyndrome (SIRS). This could cause, or contribute to,the development of the multiple organ dysfunctionsyndrome (MODS) and thus to increased mortalityfrom MODS and associated sepsis. A post-hoc analy-sis of data from one trial14 showed that patients withARDS randomized to lung-protective ventilationdeveloped significantly less remote organ dysfunc-tion (especially renal dysfunction) than controls. Thenumber of organs failing during the study periodcorrelated with the change in plasma interleukin-6(IL-6), tumour necrosis factor-a (TNF-a) and IL-1bconcentrations, lending further support to thishypothesis. These findings are important becausemost deaths of patients with ARDS are due toMODS, often associated with sepsis, rather than torespiratory failure. If VILI can contribute to thedevelopment of MODS, then modified ventilationstrategies designed to avoid VILI could result ina substantial reduction in mortality rate. These find-ings provide a plausible mechanism to explain the

substantial reduction in mortality rate seen withlung-protective ventilation strategies.

The aspects of mechanical ventilation resulting inVILI in animal models are end-inspiratory over-distension or end-expiratory collapse and tidalreinflation.

It is important to understand that it is regionallung volume, not high airway pressure, that is import-ant in causing injury. The term volutrauma hastherefore been suggested, rather than barotrauma, todescribe such injury. When chest and abdominalstrapping is applied to prevent thoracic hyperinfla-tion, the ventilation of animals with very high airwaypressures does not result in VILI. In contrast, whenthe animals were ventilated with negative-pressureventilation, high peak lung volume resulted in VILI,even with low (atmospheric) airway pressure. This isnot surprising and simply illustrates that it is thetranspulmonary pressure, not the airway pressurealone, that determines lung distension and associatedinjury. This is important in considering the effects ofspontaneous respiratory muscle activity and abnor-mal chest wall compliance during mechanical venti-lation (see below).

The lung in acute respiratory distress syndrome 155

The detailed mechanisms resulting in VILI havenot yet been determined, but it is thought that exces-sive stretch of alveolar walls, and perhaps of small air-ways, may provide signals for the release of cytokines.It has also been shown that lung capillary stress fail-ure, resulting from high lung microvascular pressure,occurs to a greater extent when lung volume is high.Capillary wall stress is a composite determined byboth transmicrovascular pressure and alveolar wallstretch. This process may also be important in caus-ing VILI and may account for the 'bacterial transloca-tion' from the lung. Accordingly, therefore, highmicrovascular pressure may contribute to VILI dur-ing injurious ventilation. High alveolar wall stressoccurs with excessive peak lung volume (regional orglobal) at end-inspiration. High tissue stresses duringinflation may occur at the interfaces betweenatelectatic and aerated lung, even in the absence ofhigh airway pressures, if end-expiratory collapse andtidal re-inflation occur with each breath.

In animal models, VILI associated with suchend-expiratory collapse can be prevented by usingsufficient PEEP to prevent collapse.

In most studies, for any degree of peak lung dis-tension, higher positive end-expiratory pressure(PEEP) - and higher end-expiratory volume andtherefore lower tidal volume - has been associatedwith less VILI. This has suggested that the goals inmechanical ventilation should be to use sufficientPEEP to prevent end-expiratory collapse and to limittidal volume to avoid end-inspiratory over-disten-sion. However, the best method of selecting anappropriate tidal volume and PEEP level has not yetbeen agreed. The physiological basis for making thisselection is discussed below.

THE LUNG IN ACUTE RESPIRATORYDISTRESS SYNDROME

The compliance of aerated lung in ARDS appearsto be normal.

Using computerized tomography (CT) scans,Gattinoni and colleagues have demonstrated that thereduced lung compliance usually seen in early ARDS

is not a result of a uniform increase in elastic recoilof lung, but a reduced amount of aerated lung. Theremaining lung is non-aerated and does not con-tribute to ventilation. The amount of aerated lungmay be only half to one-third of normal, but its spe-cific compliance (defined as compliance per unitweight of aerated lung tissue) is probably relativelynormal. The use of a tidal volume of 10-15 mL kg-1,as recommended in the past, will therefore result inover-distension of aerated lung, indicated byincreased peak inspiratory pressure (PIP) and end-inspiratory plateau pressure (-PpL), and may result inVILI. If the specific compliance is indeed normal,then the highest level of PpL that can safely be appliedin ARDS without over-distending the aerated lungshould be the same as in a normal lung. It has beenrecommended that PpL should be limited to35 cmH2O in ARDS.15 It is perhaps surprising thatthe specific compliance of aerated lung is normal,because it is known that surfactant is dysfunctionalin ARDS. If the specific compliance was overestimated,and is actually reduced, it is possible that a higher PpL

could be acceptable. However, preliminary data(unpublished) suggest that a PpL > 32-35 cmH2O isassociated with worse outcome, suggesting that ahigher PpL may be injurious and that specific com-pliance is relatively normal. Other evidence suggestssignificant surfactant dysfunction, both in ARDS andanimal models of ALL Perhaps inhomogeneity ofsurfactant distribution and function in ARDS mayaccount for this conflicting evidence. Further studiesare clearly needed.

Some of the non-aerated lung can be re-inflatedwhen exposed to a sufficient transpulmonary pres-sure. This occurs predominantly in the dependentregions (i.e. posteriorly in supine patients) andredistributes rapidly when patients are placed inthe prone position. The collapse of the dependentlung regions appears to be a result of compressionby the weight of the overlying lung; the amount ofPEEP required to prevent end-expiratory collapseat any lung level is similar to the superimposedpressure from the weight of the lung above thislevel.16 The lung thus behaves like a wet sponge:the distribution of water is relatively uniformthroughout, but the lower regions are compressedby those above.

The amount of PEEP required to prevent end-expiratory collapse should be equal to the hydrostaticpressure resulting from the overlying lung. This isusually not more than 15-20 cmH2O, although some


additional PEEP may be required to overcome sur-face tension forces favouring collapse. The preven-tion of end-expiratory collapse is the main mechanismby which PEEP improves oxygenation in ARDS.An important implication of these findings is that, ifsufficient PEEP is applied to prevent collaps, the non-dependent regions will be moderately distended evenat end-expiration. A low tidal volume is thus requiredto prevent over-distension injury. In the later stagesof ARDS, the pathophysiology changes and inflamm-ation and fibrosis are more prominent and PEEP lesseffective in achieving lung aeration and improvingoxygenation. PEEP is also less effective in ARDS dueto direct pulmonary insults (such as pneumonia)rather than to extrapulmonary causes of ARDS.Unilateral lung involvement, with pneumonia orother processes, also frequently results in a poor oreven paradoxical response to PEEP. PEEP may causelittle recruitment of the affected lung and over-distensionof the relatively normal lung may increase pulmonaryvascular resistance and divert pulmonary blood flowto the consolidated lung, thus increasing intrapulmonaryshunt. Lateral positioning, with the consolidatedlung downwards, may also increase blood flow to theconsolidated lung and increase shunt.

VENTILATION STRATEGIES

VILI develops rapidly, within minutes to hours, so itis important that lung-protective ventilation isapplied from the start. Although supporting sponta-neous breathing is preferred, controlled mechanicalventilation, and even muscle paralysis, may berequired. Pressure support should be titrated toreduce excessive work. However, there is no agree-ment about what represents excessive respiratorywork, but tachypnoea (>35-45 min-1) associatedwith excessive intercostal in-drawing and acces-sory muscle use should probably be avoided.Tachycardia, hypertension and sweating mayalso indicate excessive respiratory work, but maysimply result from hypercapnia without necessarilya high workload. Dyspnoea, in conscious patients,may require the use of sedation or increased ventila-tory support. If increased support in these circum-stances results in an excessive tidal volume, furtherrespiratory depression or paralysis should be con-sidered (see below).

Lung recruitment and the open-lungapproach

As discussed above, the role of PEEP is to preventend-expiratory collapse. However, as expected fromLaPlace's Law, alveoli or small airways that have col-lapsed require a higher pressure to re-inflate them(the opening pressure) than to keep them open. Re-inflation may also occur only gradually over severalminutes. Even in normal lungs, a sustained inflationof 30-40 cmH2O is required to achieve re-inflationafter collapse has occurred during thoracotomy. InALI, the opening pressure in some lung regions maybe as high as 45-60 cmH2O. Therefore, once depen-dent collapse has occurred, the application of PEEPalone may not result in complete re-inflation.

It may be necessary to apply various 'recruitmentmanoeuvres' to achieve additional re-inflationand thus to improve oxygenation.

Different recruitment manoeuvres can beemployed, such as:

• sustained inflation at 35-60 cmH2O in the paral-ysed patient,

• continuous positive airway pressure (CPAP) at30-50 cmH2O, with pressure-support ventilation(PSV) of only 5-10 cmH2O, during spontaneousbreathing,

• a period of ventilation with a higher tidal volumeand an end-inspiratory pause with a resulting PPL

of 35-60 cmH2O.

These manoeuvres are usually applied for 30s to2 min whilst monitoring the patient for hypotensionand subsequently for a pneumothorax. Alternatively,the use of 'sighs' (the intermittent delivery of largertidal volumes by the ventilator on a continuousbasis) may be effective. Such manoeuvres are reason-ably safe, but clinical experience is limited, particu-larly with airway pressures of 50-60 cmH2O. In somepatients, the result is a marked improvement in oxy-genation. The use of recruitment and sufficient PEEPto prevent end-expiratory collapse has been calledthe 'open-lung approach' and the PEEP used tomaintain full recruitment the 'open-lung PEEP'.Following a recruitment manoeuvre, a higher PEEPlevel may be required to maintain the recruitmentand improved oxygenation than previously used.

Ventilation strategies 157

Whether such 'open-lung' strategies result in lessVILI, or in improved outcome, compared withapproaches limiting PpL but using PEEP in a moretraditional manner (to provide adequate oxy-genation with an inspired O2 concentration of60% or less) is unknown.

At present, recruitment manoeuvres should possi-bly be restricted to patients with moderate or severehypoxaemia.

Similarly, 'de-recruitment' may occur when theairway pressure is allowed to fall during ventilatordisconnection or suctioning. Such drops in trachealpressure should be prevented as far as possible.Chest-wall compression during physiotherapy orloss of pressure when suctioning may also cause de-recruitment and may account for the hypoxaemiafrequently observed following physiotherapy.

example by pre-existing chest-wall abnormalities,chest-wall oedema or abdominal distension withhigh diaphragms), a higher PpL will be required andend-inspiratory pleural pressure will be increased bythe same amount.

Chest-wall compliance is reduced in manypatients with ARDS, probably as a result of chest-wall oedema, abdominal distension or pleuraleffusions.

Thus, a higher PpL may be acceptable in suchpatients. It is the transpulmonary pressure thatdetermines lung distension and, in patients withabnormal chest-wall compliance, estimation oftranspulmonary pressure using an oesophageal bal-loon may be helpful. The goal is to limit regionalpeak lung distension to a safe level.

Expiratory muscle activity in patients withtachypnoea may oppose the effect of PEEP anddecrease end-expiratory volume. Paralysis maythen improve oxygenation.

Selection of tidal volume

Having used sufficient PEEP to prevent end-expira-tory collapse, the tidal volume should be limited toachieve a safe PpL,15 which is probably between 32and 35 cmH2O. This usually requires a tidal volumeof <8 ml kg"1, and often <6 ml kg-1. However,in a trial of low tidal volume ventilation,12 therewas no significant interaction between tidal volumeand static compliance, suggesting that a tidalvolume of 6 ml kg-1 was associated with betteroutcome than in patients in whom the PpL

remained <35 cmH2O but with a higher tidal vol-ume. This requires confirmation.

Interpretation of end-inspiratoryplateau pressure with abnormal chest-wall compliance

The PpL required to produce a particular degree ofend-inspiratory lung distension (with relaxed respira-tory muscles) depends on both lung and chest-wallcompliance. If chest-wall compliance is reduced (for

Estimation of lung distension duringspontaneous breathing

In the presence of spontaneous breathing, the inter-pretation of airway pressure traces is difficult andlung over-distension can occur with airway pressuresin an apparently acceptable range. During volume-controlled ventilation (VCV), inspiratory muscleactivity results in lower airway and pleural pressuresthan would otherwise occur, but transpulmonarypressure and lung volume are unaffected. Thus, anexcessive tidal volume resulting in lung over-disten-sion may be associated with a relatively low PIP orPPL. During pressure-control ventilation (PCV) andPSV, spontaneous breathing will result in a lowerpleural pressure and an increase in transpulmonarypressure and peak lung volume while airway pressureremains unchanged; again, lung over-distension mayoccur. If the ventilator settings are initially made dur-ing paralysis, subsequent breathing during VCV willnot result in increased tidal volume (but, rather, thePIP and PpL will decrease), whereas, with PCV,increased tidal volume and over-distension willoccur. Even with low levels of PSV or PCV, a tidal vol-ume >6 mL kg"1 commonly occurs. In the trial ofAmato et al.,11 tidal volume was limited to 6 mL kg-1

during spontaneous breaths; this may be important,although it can be difficult to accomplish in patientswith high respiratory drive and may require a reduction


in the inspiratory airway pressure to much less than30 cmH2O and the use of sedatives.

In our experience, the use of paralysis for 24-48hours often allows adaptation to hypercapnia, andparalysis can then be discontinued and spontaneousbreathing resumed with much lower respiratorydrive, tidal volume and respiratory work. Amato etal. have used partial paralysis, with controlled infu-sions of muscle relaxants, with careful monitoring ofexpired tidal volume.11 The interpretation of airwaypressures during spontaneous breathing can be facil-itated using an oesophageal balloon or by temporar-ily abolishing respiratory muscle activity usingneuromuscular blockade or opiates. The expiredtidal volume should be monitored and limited tothat producing an acceptable PpL during paralysis,and preferably to <6 mL kg-1. Basic principles sug-gest that this is just as important as the limitation ofPpL during paralysis and therefore should beattempted at all times during ventilation, althoughthis is not commonly the practice.

Selection of positive end-expiratorypressure and the pressure-volumecurve

The best method of selecting open-lung PEEP is notclear. Collapse can be visualized on CT scans, but it isnot feasible to perform frequent CT scans for thispurpose. The respiratory inflation pressure-volumecurve in ARDS usually has a sigmoid shape (as shownin Fig. 12.1), with lower and upper inflexion points(Pflex). It was previously believed that the lower Pflexindicates the pressure and volume range over which

re-inflation (recruitment) of collapsed lung unitsoccurs and that no further recruitment occurs on the'linear'portion above the lower Pflex because furtherrecruitment would result in a continued increase inslope (upwards concavity). It has also been assumedthat the upper Pflex indicates the beginning of lungover-distension. It has therefore been recommendedthat tidal ventilation should occur over a pressurerange between the lower and upper Pflex.

However, a mathematical model of the ARDSlung17 suggests that recruitment of previouslycollapsed lung units can continue over the wholelength of the PV curve, that an upper inflectionpoint can occur as recruitment stops or dimin-ishes (without necessarily implying alveolar over-distension) and that the lower Pflex may not beclosely related to open-lung PEEP.

The mathematical model incorporates gravita-tional superimposed pressure increasing from zeroin the non-dependent regions to 15 cmH2O in thedependent regions, in keeping with Gattinoni's con-cept. These findings have been supported by clinicalstudies and suggest that previous interpretations ofthe pressure-volume curve are incorrect.18

Figure 12.1 shows an example of a pressure-volume curve produced by a modified version of themathematical model. As the pressure increases duringinflation, collapsed alveoli in the non-dependent lungregions become inflated (recruited) when theiropening pressure is exceeded, resulting in a greaternumber of aerated alveoli and therefore increasingcompliance (i.e. pressure-volume slope) to produce

Figure 12.1 Inflation and deflation static pressure-volume (PV)

curve generated using a mathematical model of ARDS lungs. The

lower and upper inflection points (Pflex) on the inflation curve are

indicated. Recruitment continues during inflation over the whole

length of the plot and the upper Pflex occurs as recruitment

diminishes. Inflation (bold lines) and deflation tidal PV plots are also

shown with tidal volume of 400 mL at incremental PEEP levels of

0, 5, 10, 15 and 20 cmH20. Opening pressures were normally

distributed and varied from 5 to 35 cmH20. (See reference 17 for

details of the model.)


the lower Pflex. As each alveolus is recruited, it 'snaps'open and suddenly increases its volume from zero tothat appropriate to its new transalveolar pressure.With each increment of pressure, the increase involume in newly recruited alveoli is much greater thanthat of alveoli that were already aerated prior to thepressure increment. On the steep portion of thepressure-volume curve above the lower Pflex, most ofthe volume increment with each pressure increment(i.e. slope) is a result of these large volume increases asnewly recruited alveoli snap open (i.e. to recruitment).The slope of this portion of the curve is much greaterthan the total compliance (sum of the individualcompliances, or 'total alveolar compliance') of all ofthe inflated alveoli. When the rate of recruitmentdiminishes and finally stops, the slope decreases tothat of the total alveolar compliance, causing an upperPflex. The total compliance of all alveoli that wereinflated at end-inspiration is shown by the slope of thedeflation pressure-volume curve over the upperpressure range. When the pressure during deflationfalls below open-lung PEEP, the dependent lungregions start to collapse, resulting in suddendecrements in volume as each alveolus collapses; thisagain results in a steepening of the curve. However, itis not possible to identify open-lung PEEP on thedeflation curve, either with the model or in patients. Ifthe assumptions of this model are correct, the lowerPflex indicates the commencement of recruitmentand is unlikely to be closely related to open-lung PEEP,and the upper Pflex may also be greatly influenced byrecruitment. Inflation and deflation tidalpressure-volume plots are shown in Figure 12.1within the envelope of the inflation and deflationpressure-curves.

A method that has been widely used to determine'optimum PEEP' is to increase PEEP progressivelyuntil the 'effective compliance' (tidal volume/[PpL —PEEP] ) reaches a maximum. It was suggested that,when PEEP is increased above the lower Pflex of thepressure-volume curve, ventilation would occur onthe steep portion, increasing effective compliance.Some data suggest that this is true in those who showno recruitment (i.e. upwards displacement on thevolume axis) with PEEP; PEEP then simply movesthe end-expiratory point further up on the samepressure-volume curve. However, such patients alsoshow little improvement in oxygenation with PEEP.In patients showing recruitment with PEEP andimproved oxygenation, there may be little change or

even a decrease in effective compliance as PEEP isincreased (Fig. 12.2). As PEEP increases, there is littlechange in the tidal pressure-volume slope (effectivecompliance) because the end-expiratory volumeincreases (due to prevention of end-expiratory col-lapse by PEEP) by a similar amount to the end-inspira-tory volume. As PEEP approaches open-lung PEEP,eliminating end-expiratory collapse, the effectivecompliance may decrease slightly. With a low tidalvolume, the effective compliance may furtherincrease with PEEP well above open-lung PEEP,because the PpL continues to increase as PEEPincreases, causing more end-inspiratory recruitment.This increases the number of aerated alveoli andtherefore compliance.

An incremental PEEP trial with a low tidal volumeis really testing the recruiting ability of the end-inspiratory pressure rather than the ability of PEEPto prevent end-expiratory collapse. Therefore, evenoxygenation during an incremental PEEP trial, the-oretically, would not be expected to be a good indi-cator of open-lung PEEP, and limited clinical andexperimental data support this.

Figure 12.2 Inflation and deflation pressure-volume (PV) curve

generated by mathematical model of ARDS lungs. Inflation tidal PV

plots are shown with tidal volume of 150 ml and incremental and

decremental positive end-expiratory pressure (PEEP) from 0 to 25

cmH20 in 5-cmH20 steps. Opening pressures were normally

distributed and varied from 5 to 35 cmH20. With decremental PEEP,

the tidal PV plot remains superimposed on the deflation PV curve

until PEEP falls below open-lung PEEP. The PV slope is much higher

with decremental PEEP, and the maximum tidal PV slope occurs at a

PEEP level several cmH20 below open-lung PEEP.


In contrast, if a recruitment manoeuvre is per-formed and then a decremental PEEP trial is per-formed, both oxygenation and effective complianceare likely to better indicate open-lung PEEP. In themathematical model, effective compliance during adecremental PEEP trial always underestimatesopen-lung PEEP, but is related to it more pre-dictably than during an incremental PEEP trial. AsPEEP is reduced, the effective compliance initiallyincreases because each alveolus moves onto a morecompliant portion of its pressure-volume curve.The tidal PV plot moves downwards superimposedon the deflation PV curve (see Fig. 12.2). WhenPEEP is reduced below open-lung PEEP, some end-expiratory collapse occurs. Some alveoli remain col-lapsed at end-inspiration, because their openingpressures are not exceeded, so the end-inspiratorypoint is no longer superimposed on the deflationpressure-volume curve, but is below it. The reducednumber of aerated alveoli tends to reduce effectivecompliance, but this is opposed by the increasingcompliance of each alveolus as PEEP is reduced andcompliance continues to increase. Some tidalrecruitment (alveoli that are inflated at end-inspi-ration but collapse at end-expiration) also tends toincrease compliance, but this occurs less than dur-ing incremental PEEP. As PEEP is reduced further,more alveoli collapse and eventually this has agreater effect on effective compliance than the con-tinuing increase in compliance of each alveolus;effective compliance therefore falls. The maximumvalue of effective compliance in the model with alow tidal volume (in most circumstances) occurs ata PEEP level 3-6 cmH2O below open-lung PEEP.Thus, during incremental PEEP, the PEEP givingthe highest effective compliance is highly depen-dent on tidal volume, opening pressures and 'lungmechanics' and can be well above or below open-lung PEEP. These factors have much less effect dur-ing a decremental PEEP trial, which is thereforemore reliable in indicating open-lung PEEP. Thesame principles suggest that the best oxygenationwould also indicate open-lung PEEP more reliablyduring a decremental rather than during an incre-mental PEEP trial. Clinical data concerning the useof decremental PEEP trials in ARDS are very limit-ed and the use of effective compliance duringdecremental PEEP to determine open-lung PEEPcannot currently be recommended. However, thelimitations of incremental PEEP trials (especially

with a low tidal volume) are being recognized.Limited data concerning oxygenation during adecremental PEEP trial suggest that this may be auseful method to determine open-lung PEEP.Further studies will be required, however.

Figure 12.3 shows the effect of different tidal vol-umes with the same PEEP level (15 cmH2O) gener-ated by the model. As tidal volume is increased, theend-expiratory volume increases, because most ofthe additional lung units recruited at the higher end-inspiratory pressure remain inflated at end-expira-tion. The figure shows a simulated recruitmentmanoeuvre followed by a return to a low tidal vol-ume, demonstrating a large increase in lung volumeand effective compliance during subsequent tidalventilation and a shift towards the deflation limb ofthe pressure-volume curve. A higher level of PEEPmay be subsequently required to maintain thisincreased lung volume and the associated improve-ment in oxygenation.

Figure 12.3 Inflation and deflation pressure-volume (PV) curve

and inflation (bold lines) and deflation tidal PV plots with two tidal

volumes (150 and 400 ml) and a simulated recruitment manoeuvre

with a volume of WOO mL The positive end-expiratory pressure

(PEEP) level is 15 cmH20 in each case, but the end-expiratory

volume is greater with the higher end-inspiratory pressures, because

more lung units have been recruited at end-inspiration, and most

remain inflated at end-expiration because PEEP is higher than their

closing pressure. The tidal PV slope is also much greater with a high

tidal volume (Vt) because of more end-inspiratory recruitment. The

plot shown as black triangles indicates the inflation tidal PV plot

with Vt of 150 mL and PEEP 15 cmH20 following the recruitment

manoeuvre. Effective compliance is much higher and the plot has

moved towards the deflation PV curve, as occurs during a

decremental PEEP trial. Opening pressures were normally distributed

and varied from 5 to 35 cmH20


Pressure-control versus volume-control ventilation

The main differences between these two modes relateto the effect of spontaneous breathing, as discussedabove. Either mode can be used to generate anydesired peak lung volume and flow rate. There is littleevidence that the de-accelerating flow profile that occursduring PCV offers any real benefit; indeed, the higherearly-inspiratory flow rate may cause greater VILI.The high peak pressure during coughing with VCV isnot associated with high transpulmonary pressure,so it does not result in volutrauma. Using the sametidal volume, the slightly higher peak pressure thatoccurs during VCV (with constant flow) rather thanPCV is a result of continuing flow at end-inspirationand the associated pressure drop across the airways;the alveolar pressure (and the potential for VILI)does not differ. Studies comparing PCV and VCV inARDS at equivalent levels of total (applied plusintrinsic) PEEP have shown no benefit in terms ofoxygenation.

High-frequency ventilation

Theoretically, high-frequency ventilation (HFV)offers the advantage of maintaining lung recruitmentwith a lower tidal volume yet maintaining adequateCO2 elimination. The results in neonatal respiratorydistress syndrome (RDS) appear encouraging, butoutcomes also appear to be improving with modifiedconventional ventilation. No trials have adequatelycompared HFV and optimum modified conventionalventilation, either in neonates or adults.

Inverse ratio ventilation

Extended ratio ventilation occurs when the inspirat-ory to expiratory (I:E) ratio is increased above thecommonly used 1:2 and inverse ratio ventilation(IRV) when the I:E ratio is >1:1. IRV increases themean airway pressure and this may be associated withimproved oxygenation. It has been suggested that IRVmay facilitate lung recruitment, but no studies havecompared it with current strategies using the open-lung approach. When the expiratory time is decreasedsufficiently, expiration remains incomplete anddynamic hyperinflation and intrinsic PEEP (PEEPi)

occur. This has a similar effect to applied PEEP andmay improve oxygenation. If excessive, it may resultin hypotension and barotraumas. During VCV, PEEPiresults in increased end-inspiratory lung volume andPpL (unless tidal volume is reduced), whereas duringPCV the end-inspiratory pressure is unchanged andtidal volume decreases. It has therefore been suggest-ed that PCV is safer than VCV during IRV, but eitherapproach is acceptable providing that the tidal vol-ume and PpL are carefully monitored. Preliminarydata suggest that there may be greater regional over-distension during IRV at equivalent tidal volume andtotal PEEP levels, suggesting more heterogeneous distri-bution of ventilation related to the shorter expiratorytime (unpublished data, Bersten A et al). The role ofIRV therefore remains uncertain; it has no provenbenefit and this possible disadvantage.

Newer modes of ventilation

A number of new ventilation modes have been deve-loped, a full discussion of which is beyond the scopeof this chapter. However, none has been shown tohave a real advantage over optimum conventionalventilation in ARDS. It is now possible to achieve theoptimum peak and end-expiratory lung volume andflow rates using several different modes and with dif-ferent contributions to respiratory work from theventilator and the respiratory muscles. There are nodata showing an advantage of any particular mode.

Permissive hypercapnia

The protective ventilation strategies described abovefrequently result in hypercapnia. This is usually welltolerated, especially if it occurs gradually, allowingcompensation of the resulting acidosis. The physio-logical effects of acute hypercapnia are complex andonly a few aspects will be discussed here. Most of theeffects are thought to be caused by the associatedintracellular acidosis and this is corrected quite rapid-ly (within a few hours, as opposed to 1-2 days for therenal compensation of the extracellular acidosis) bycell membrane ion transporters that protect intracel-lular pH. Thus, the direct depression of myocardialcontractility caused by acute hypercapnia in denervat-ed heart preparations is largely corrected within a fewhours. During sustained hypercapnia, therefore, the


extracellular (blood) pH may not provide a good indi-cation of the physiological effect of hypercapnia. Inintact animals and humans, myocardial contractilityand cardiac output increase because of increased sym-pathetic activity. Pulmonary artery pressure frequentlyincreases in ARDS during hypercapnia and, whensevere, this may be associated with a fall in cardiacoutput. The pulmonary hypertension can be reducedby buffering the acidosis or by inhaled nitric oxide.Hypercapnia may rarely precipitate arrhythmias.

Even in critically ill patients requiring inotropicdrug infusions, cardiac output increases duringhypercapnia, although blood pressure may fallfrom systemic vasodilatation.

The effects of hypercapnia on gas exchange arecomplex but, in ARDS, there is usually little change ora slight increase in -PaO2 and an increase in mixedvenous O2 (PvO2) as a result of an increase in cardiacoutput and a right-shift of the haemoglobin-O2 dis-sociation curve. The overall effect, therefore, is likelyto be beneficial to tissue O2 uptake. However, higherlevels of PEEP may be required to maintain recruit-ment when tidal volume is reduced. Hypercapnia hasa direct vasodilator effect on pulmonary arterioles.This is offset by the potentiation of hypoxic vaso-constriction caused by acidosis. In an animalmodel, buffering of the acidosis resulted in a markedincrease in intrapulmonary shunt and a decrease inarterial O2 saturation by removing the acidosis-inducedpotentiation of hypoxic vasoconstriction andleaving the vasodilator effect of CO2 unopposed.19

The effect of buffering during hypercapnia on gasexchange in ARDS has not been adequately studied.

Therapeutic hypercapnia

Part of the improved outcome resulting fromlung-protective ventilation strategies may be dueto the hypercapnia per se rather than solely to theventilation strategy.

Acute hypercapnia has a number of poorly under-stood effects, including the suppression of cytokinerelease and the oxidative burst in vitro. Intracellularacidosis has been shown to prevent cell death during

anoxia in cultured cells and, in an isolated perfusedheart model, hypercapnic acidosis resulted in bettercardiac function following cardioplegic ischaemia. Aseries of studies in isolated perfused lungs and, morerecently, in intact rabbits,20 showed that hypercapnicacidosis is protective against ischaemia-reperfusioninjury. The protection is probably mediated by theinhibition of xanthine oxidase and it was thereforespeculated that part of the improved outcome result-ing from lung-protective ventilation strategies maybe from the hypercapnia itself, rather than solelyfrom the ventilation strategy - hence the term 'thera-peutic hypercapnia'. These observations are intrigu-ing and merit further study. At present, hypercapniashould be regarded as an undesirable side effect oflung-protective ventilation, and severe hypercapniashould be avoided, when possible, whilst meeting thetargets for PpL and PEEP. The use of alkali to correctthe acidosis has not, as yet, been adequately studied.

Hypercapnia should be avoided in patients with, orat risk of, intracranial hypertension and should beallowed cautiously in ischaemic heart disease. In suchpatients, measures should be taken to reduce thePaCO2, which may include an increase in the ventila-tor rate, tracheal gas insufflation, and high-fat enteralfeeding. If possible, the tidal volume and PEEP levelsshould be maintained at 'protective' levels.

Extracorporeal support

Extracorporeal support has been used to improveoxygenation in patients with refractory hypoxaemia.However, a trial of extracorporeal membrane oxy-genation (ECMO) in ARDS showed no benefit. Lungtissue ischaemia may have resulted from the veno-arterial bypass, which reduces pulmonary bloodflow, and the continuation of mechanical ventilationwith a high tidal volume may have resulted in VILI.In 1980, Gattinoni and colleagues used extracorporealCO2 removal (ECCO2R). They used veno-venousextracorporeal perfusion (perfusing the lung with anormal flow of well-oxygenated blood) and empha-sized lung rest, using a low tidal volume. Initial experi-ence suggested an improved outcome, but a furthertrial in the USA21 showed a similar mortality rate inthe ECCO2R and control groups. This study has beencriticised because the investigators had limited experi-ence of the technique. The role of ECCO2R thereforeremains controversial and unproven. Outcomes with

Posture 163

conventional ventilation appear to be improving andfewer patients are being considered for ECCO2R.

Liquid ventilation

Ventilation with perfluorocarbons is usually admin-istered as partial liquid ventilation, in which tidal gasvolumes are delivered to the perfluorocarbon-treatedlung. It is a promising novel therapy. A volume ofperfluorocarbon equivalent to the functional re-sidual capacity is administered via the endotrachealtube. Its high O2 and CO2 solubility, reduction insurface tension (a surfactant like effect) and ability torecruit dependent lung offer improved oxygenationcoupled with reduced volutrauma. Because the per-fluorocarbons are denser than water, they may beviewed as a form of 'liquid PEEP'. They increase thetransalveolar pressure preferentially in the depen-dent regions, reducing the gravitational gradient oftranspulmonary pressure caused by the weight of thelung, and may also redirect blood flow away fromdependent areas, thereby improving oxygenation.When external PEEP is applied, the dependent lungcould become over-distended.

Clinical studies22 have confirmed laboratory datashowing improved gas exchange and lung mechan-ics after the administration of perfluorocarbons.

Although this has not been accompanied byimproved survival or ventilator-free days, the stud-ies have been inadequately powered. However, asignificant incidence of pneumothorax was report-ed and, while this may just reflect the severity ofthe underlying disease, appropriate ventilatorystrategies will be crucial to the design of furtherclinical trials.

POSTURE

Although it has been traditional to manage criticallyill patients supine, other approaches to body posi-tioning should be considered. The two importantissues in patients with ARDS are the prone and semi-recumbent positions.

Prone positioning

Turning a patient with ARDS prone from the supineposition increases oxygenation in approximately80% of patients and this may be sustained for somehours when returned to the supine position.

Although there are no published phase II or phaseIII studies using the prone position, the improve-ment in oxygenation may be dramatic and thismanoeuvre should be considered when there issevere hypoxaemia despite adequate PEEP andrecruitment manoeuvres.

Complications include accidental removal of tubesand lines, facial oedema, skin abrasion and apicalatelectasis. The prone position is relatively con-traindicated following recent sternotomy or duringhaemodynamic instability, and the turning procedurerequires time, personnel and preparation (SeeChapter 6).

The mechanisms behind the effect are debated. Onturning prone, there is a significant change intranspulmonary pressure that allows dorsal inflationwithout an equivalent degree of ventral collapse.Ventilation becomes more uniform and aeration oflung in the posterior diaphragmatic recess isimproved. Because pulmonary blood flow is relativelyunchanged, there is an immediate improvement inpulmonary shunt. A number of factors have beenproposed to contribute to this differential effect,including the shape of the chest wall and diaphragmand compression by the heart. Prone positioningmay also reduce VILI through the recruitment of col-lapsed air spaces.

Semi-recumbent position

Although not specific to ARDS patients, the semi-recumbent position reduces nosocomial pneumonia(but not mortality) in critically ill patients, presum-ably through reduced aspiration. Factors such ashaemodynamic instability, multiple trauma or therequirement for postural drainage limit the number ofpatients that can be managed entirely semi-recum-bent, but it is otherwise the preferred position whensupine.


SURFACTANT REPLACEMENT THERAPY

Pulmonary surfactant is essential for normal lungfunction. It is a complex mix of phospholipids, neu-tral lipids and proteins that lines the gas/liquid inter-face, reduces surface tension and allows it to varydirectly with alveolar radius. Consequently, the workof breathing is reduced and alveoli of different sizesare able to co-exist. In addition, surfactant plays animportant role in fluid balance and host defencewithin the lung.

Surfactant is synthesized primarily by alveolartype II cells and is stored in lamellar bodies. Inresponse to a number of stimuli, in particular physi-cal distortion of the type II cells, lamellar bodies areexocytosed into the hypophase, where they unravelto form tubular myelin, which in turn supplies thesurface active monolayer. The monolayer constantlybecomes inactive and, coupled with the re-uptake ofinactive surfactant, release of fresh surfactant isessential for maintaining a viable lung.

Although pulmonary surfactant is not deficient inARDS, it is dysfunctional and this correlates with thedegree of lung damage. The surface tension hystere-sis is decreased and the minimum surface tension isincreased up to twofold in patients at risk of devel-oping ARDS and fourfold in ARDS patients. Samplesobtained from patients with ARDS by bronchoalveo-lar lavage have abnormal surfactant composition(Table 12.2), which, together with surfactant inacti-vation by plasma proteins, reactive O2 species andphospholipases, results in surfactant dysfunction.

These changes probably make important contribu-tions to both alveolar instability and collapse, withconsequent shunt and hypoxaemia, and to decreases inspecific lung compliance. In addition, becausedysfunctional surfactant exaggerates the heterogeneityof air-space ventilation, it will contribute to regionalover- inflation and to the opening and closing of alveoliduring tidal ventilation, leading to further lung damage.

Despite the important role that surfactant abnor-mality plays in the pathogenesis and pathophysiologyof ARDS, promising results in animal models of lunginjury, and the proven efficacy of exogenous surfac-tant replacement therapy (ESRT) in infants withestablished RDS, it has not yet been shown to be ofbenefit in ARDS. A number of small studies havereported improved oxygenation and compliance withESRT, but neither outcome nor physiological benefitwas found in a large, multi-centre, prospective, ran-domized trial of ESRT in patients with sepsis-inducedARDS.23 There are a number of reasons why this mayhave occurred. The surfactant preparation used maynot have reached the air spaces or may have modifiedthe composition and surface tension of the alveolarepithelial lining fluid. The dose of surfactant admin-istered was probably an order of magnitude less thanoptimal. It was administered as an aerosol, whichresults in preferential distribution to well-ventilatedareas; bronchoscopically administered surfactant canreach poorly ventilated regions, where it will be mostuseful. The surfactant preparation used is sensitive toprotein inhibition and does not contain surfactantproteins that improve its surface-tension-reducing

Table 12.2 Changes in surfactant composition in ARDS

Phosphatidylcholine ,Disaturated phospholipidsPhosphatidlyglycerolPhosphatidylinositolSphingomyelinLysophosphatidylcholine

Surfactant protein-A

Surfactant protein-B ,Tubular myelin-rich aggregatesTubular myelin-poor aggregates

The major phospholipid classThe main surface active phospholipids

Probably a cell membrane componentThe first catabolic product of PCDirectly interferes with the surface active properties of surfactantImportant for reducing surface tension and host

defenceEssential for reducing surface tensionFunctionally active componentFunctionally inactive component

Manipulation of the pulmonary vasculature 165

properties. Ventilatory strategies were not controlledand recent data suggest that a low tidal volume main-tains a greater proportion of tubular myelin-richaggregates following ESRT. Finally, these patientswere enrolled with established ARDS, whereas ESRTwill be most beneficial if delivered early.

Although ESRT has not proven to be of clinicalbenefit in ARDS, there are sound reasons forbelieving that, with improved preparations,administration techniques and study design, ESRTmay find a role in management. Techniques suchas intermittent lung stretch or biological variabil-ity in tidal volume may also prove useful throughenhanced release of endogenous surfactant.

MANIPULATION OF THE PULMONARYVASCULATURE

Hypoxic pulmonary vasoconstriction causes someredistribution of pulmonary blood flow away fromthe dependent atelectatic areas to better-ventilatedlung. Inhaled nitric oxide (iNO), prostacyclin(PGI2) and intravenous almitrine may be used tomanipulate the pulmonary circulation to furtherreduce shunt. Both iNO and PGI2 are potentvasodilators. When delivered by inhalation, theyare distributed to aerated lung regions and vasodi-late the local pulmonary circulation, furtherincreasing pulmonary blood flow to well-ventilat-ed lung and improving oxygenation. Almitrine is aselective pulmonary vasoconstrictor that reinforceshypoxic pulmonary vasoconstriction. It mayimprove oxygenation alone, but has usually beentrialled with iNO. The effects are additive, allowinga lower dose of almitrine to be used, thereby reduc-ing the risk of pulmonary hypertension andpolyneuropathy that has been reported with long-term almitrine use.

Inhaled NO and PGI2 also reduce right ventricularafter-load, which is often increased in ARDS. Thismay result in improved cardiac output and O2 trans-port, especially during permissive hypercapnia. Inpatients with severe primary or secondary pul-monary hypertension, these agents may also be effec-tive. Intravenously administered PGI2 improvescardiac output in ARDS, but widespread pulmonaryvasodilatation results in increased intrapulmonaryshunt and worse oxygenation.

Inhaled nitric oxide

Most clinical experience has been with iNO. NO is anendothelium-derived smooth-muscle relaxant, buthas other crucial physiological roles, including neu-rotransmission, host defence, platelet aggregation,leucocyte adhesion and bronchodilatation. Doses aslow as 60 parts per billion iNO may improve oxy-genation, although concentrations of 1-40 parts permillion (p.p.m.) are often used in ARDS. A higherdose may be required to reduce pulmonary arterypressure. Delivery is usually as medical-gradeNO/N2, and this should be adequately mixed toavoid variable NO concentrations. It is recommend-ed that inspiratory NO and NO2 concentrations aremeasured. The electrochemical method is accurate to1 p.p.m. (adequate for clinical use) and is less expen-sive than the more accurate chemoluminescencemethod. Environmental levels of NO and NO2 dur-ing iNO therapy are usually low and predominantlyinfluenced by atmospheric concentrations. However,it is still common practice to scavenge expired gas.Binding to haemoglobin in the pulmonary circula-tion rapidly inactivates NO, and systemic effects areonly reported following high concentrations.Systemic methaemoglobin levels may be monitoredand are generally less than 5% during clinical use.NO may cause lung toxicity through combinationwith O2 free radicals and through the metabolism ofNO to NO2, but these do not appear to cause majorclinical problems.

Only 40-70% of patients with ARDS sustainimproved oxygenation with iNO (responders), prob-ably because active hypoxic pulmonary vasoconstric-tion has already minimized intrapulmonary shunt inthe remainder. Two large trials of iNO24,25 haveshown no improvement in the mortality or reversalof ALI, although the requirement for ECMO wasreduced in infants with persistent pulmonary hyper-tension. However, iNO was safe and improved oxy-genation initially. This was not sustained beyond12-24 hours, however, and some patients receivingplacebo had an increase in -PaO2 >20% at 4 hours.Consequently, the role of iNO in patients with ARDSremains uncertain.

In severe hypoxaemia, perhaps in combinationwith almitrine, iNO may provide temporaryrescue.


ANTI-INFLAMMATORY DRUGS

A complex inflammatory response is central to the devel-opment of diffuse alveolar damage (see Chapter 11).Consequently, a number of anti-inflammatory phar-macological interventions have been examined, but,in general, promising laboratory data have not heldup in clinical trials.

Glucocorticoids

Glucocorticoids inhibit several aspects of alveolarinflammation through inhibition of the transcriptionof many of the involved cytokines (including IL-1, IL-2,IL-6, TNF-cx, and interferon-gamma [IFN--y]), inhibi-tion of complement-mediated neutrophil aggregation,the synthesis of the arachidonic acid metabolites,platelet-activating factor and NO. Although short-termglucocorticoid therapy was ineffective in early ARDSand in septic at-risk patients, there is renewed interest inits use later in ARDS when it may modify persistentalveolar inflammation and fibroproliferative oblitera-tion of the blood-gas membrane.

Encouraging small case series have been followedby a small, prospective, randomized trial in whichICU and hospital mortality rate and remote organdysfunction were reduced.26

Because pneumonia frequently occurrs in theabsence of fever, surveillance using bronchoalveolarlavage is recommended. Further trials are requiredbefore the routine use of glucocorticoids can be con-fidently recommended in this situation.

Ketoconazole

This antifungal has anti-inflammatory properties,including inhibition of both thromboxane synthaseand 5-lipoxygenase - enzymes that are necessary inthe production of thromboxane A2 and leukotrieneB4. Because both are involved in alveolar inflamma-tion, ketoconazole has been trialled in ARDS. Smallstudies of ketoconazole in at-risk patients haveshown a reduction in the development of ARDS and

possibly mortality. However, in a large trial in earlyALI and ARDS, ketoconazole did not affect the mor-tality rate or lung function.27 Consequently, it cannotbe recommended at present.

Prostaglandin E1 and prostacyclin

These prostaglandins are potent vasodilators andinhibit platelet and neutrophil function. They causereduced expression of the CD lib/CD 18 neutrophiladhesion complex, and decreased neutrophil activa-tion due to increased c-AMP, with attenuated releaseof O2 radicals and leukotriene B4. Although theirintravenous administration may result in hypoten-sion, and worsen hypoxaemia by causing pulmonaryvasodilatation, they have been trialled in ARDS.Prostaglandin EI is metabolized in the lung and socauses less systemic vasodilatation than prostacyclin.It has therefore been used in most studies, often as aliposomal mixture to enhance intracellular penetra-tion. Despite encouraging animal data, in a phase IIItrial involving 350 ARDS patients, liposomalprostaglandin E1 did not reduce mortality, althoughthe time to a PaO2/FiO2 > 300 was decreased.28

Hypotension and hypoxia were common reasonsfor altering drug infusion in the liposomal pro-staglandin Ej arm (59% versus 15% in controlpatients).

Cytokine antagonism

Elevated levels of both pro-inflammatory and anti-inflammatory cytokines are present in the lung andin plasma in ARDS, but their role is uncertain.In patients with indirect risk factors for ARDS, suchas non-pulmonary sepsis, elevated circulatingcytokines fail to predict the development of ARDSand, in patients with a direct risk factor, such aspneumonia or aspiration of gastric contents, the ini-tial increase in circulating cytokines may simply beattributable to leakage from the alveolus because ofdamage to the alveolar-capillary membrane.

Most studies of cytokine antagonism have targetedsepsis, rather than ARDS. Overall, no outcome advan-tage has been shown, although, in some studies, animproved or worse outcome occurred in post-hoc sub-group analysis.'Various factors should be consideredwhen interpreting these data, including the heteroge-

References 167

neous populations, the timing of therapy, possibleeffects of the drug and its efficacy. For example,although TNF-a plays a role in increased capillary per-meability in ALI, it increases alveolar epithelial fluidclearance in models of pneumonia via a non-cate-cholamine-dependent mechanism and also promotesthe clearance of bacteria in pneumonia. Consequently,the effect of an anti-TNF-a antibody maybe influencedby the precise risk factor for ARDS and a number ofother factors. Without further clinical data, antagonismof cytokines cannot be recommended in ARDS.

Other anti-inflammatory drugs

A number of other agents have been examined asanti-inflammatory agents in ARDS, but none haslived up to the promising results in animal models.Pentoxyfilline inhibits (i) TNF-a release followingendotoxin, (ii) polymorph deformability, (iii) therelease of O2-derived free radicals, and (iv) respon-siveness to platelet-activating factor. However, liso-phylline, a related drug, does not appear to influenceoutcome in ARDS. Non-steroidal anti-inflammato-ries reduce thromboxane A2 and prostacyclinthrough inhibition of cyclo-oxygenase, but fail toreduce the incidence of shock or ARDS in septic

Although the results of anti-inflammatory thera-py in animal models have been encouraging, clin-ical trials have so far been disappointing.

patients.29 Whereas anti-oxidants such as N-acetyl-cysteine and its analogues have not proven useful, thecentral role of O2-derived free radicals in generatinglung injury suggests that further clinical research iswarranted. A platelet-activating factor antagonistalso failed to alter mortality in ARDS.

FLUID THERAPY

There is general agreement that intravenous fluidsshould be limited to the minimum necessary tomaintain adequate cardiac output and tissue perfu-sion. Elevated pulmonary microvascular pressure, aswell as increasing pulmonary oedema, may interactwith alveolar stretch to augment VILI, providing afurther reason to maintain lung microvascular pres-

sure as low as possible. However, there is still no easymethod of determining the adequacy of intravascu-lar volume (see Chapter 8). Regional tissue hypoxiamay be present with normal indices of global perfu-sion and tissue oxygenation. Therefore, fluid therapyremains controversial. It is important to recognizethe effects of cardiovascular changes on gasexchange. A low cardiac output results in a lowmixed venous O2 saturation and, in the presence ofhigh levels of intrapulmonary shunt, this may resultin a fall in arterial O2 saturation. Fortunately, areduction in cardiac output often causes a reductionin intrapulmonary shunt, opposing this effect.Nevertheless, it is important to ensure that any hypo-volaemia is corrected. CVP may be elevated in thepresence of pulmonary hypertension and right ven-tricular dysfunction and does not correlate well withPA wedge pressure in ARDS. It is also important toconsider the effect of variations in intrathoracicpressure, including intrinsic and applied PEEP, onthe measured vascular pressures. The preference forcolloid or crystalloid therapy remains controversial,although several meta-analyses have failed to showany benefit of colloids.

SUMMARY

Rapid and adequate treatment of the underlyingconditions causing ALI, including definitive surgicaltreatment of sepsis or excision of necrotic tissuewhen required, and early recognition and treatmentof secondary infection remain of fundamentalimportance. In association with current ventilationstrategies, mortality rates appear to be falling. Agreater understanding of lung mechanics is helpingto develop improved ventilation and recruitmentstrategies, but there are still many unresolved issues.Anti-inflammatory agents, cytokine modulation andsurfactant replacement have, so far, been disappoint-ing, but may eventually yield additional treatmentsand lead to even greater improvements in outcome.

REFERENCES

1. Hickling, KG, Henderson, SJ, Jackson, R. Low mortality

associated with low volume pressure limited

ventilation with permissive hypercapnia in severe


adult respiratory distress syndrome. Intensive Care

AtoH990; 16:372-7.

2. Toth, JL, Capellier, G, Walker, P, Winton, T, Marshall, J,

Demajo, W. Lung emphysematous changes in ARDS.

Am Rev Respir Dis 1992; 145: A184.

3. Lewandowski, K, Falke, KJ, Rossaint, R, Slama, K,

Pappert, D, Kuhlen, R. Low mortality associated with

advanced treatment including V-V ECMO for severe

ARDS. Intensive Care Med 1992; 19: S42.

4. Hickling, KG, Walsh, J, Henderson, S, Jackson, R. Low

mortality rate in adult respiratory distress syndrome

using low-volume, pressure-limited ventilation with

permissive hypercapnia: a prospective study. Crit Care

Med 1994; 22: 1568-78.

5. Levy, B, Bollaert, PE, Bauer, P, et al. Therapeutic

optimisation including inhaled nitric oxide in adult

respiratory distress syndrome in a polyvalent intensive

care unit.) Trauma 1995; 38: 370-4.

6. Thomsen, GE, Morris, AH, Pope, D, et al. Mechanical

ventilation of patients with adult respiratory distress

syndrome using reduced tidal volumes. Crit Care Med

1994;22:A205.

7. Nakagawa, S, Bohn, D. Pressure controlled ventilation

with limited peak inspiratory pressure below 35 to 40

cm H20 may improve survival of pediatric acute

respiratory failure. Am J Respir Crit Care Med 1995; 151:

A77.

8. Botero, C, Reda, Z, Mendoza, P, Davis, A, Harrison, R.

Pressure limited ventilation with permissive hypercapnia

(PH) in children with ARDS. Crit Care Med 1995; 23: A188.

9. Sheridan, RL, Kacmarek, RM, McEttrick, MM, et al.

Permissive hypercapnia as a ventilatory strategy in

burned children: effect on barotrauma, pneumonia

and mortality. 7 Trauma 1995; 39: 854-9.

10. Stocker, R, Neff, T, Stein, S, Ecknauer, E, Trentz, 0,

Russi, E. Prone positioning and low-volume pressure-

limited ventilation improve survival in patients with

severe ARDS. Chest 1997; 111:1008-17.

11. Amato, MB, Barbas, CS, Medeiros, DM, et al. Effect of a

protective-ventilation strategy on mortality in the

acute respiratory distress syndrome. N Engl J Med

1998; 338(6): 347-54.





Med 2000; 342(18): 1301-8.

13. Ranieri, VM, Suter, PM, Tortorella, C, et al. Effect of

mechanical ventilation on inflammatory mediators in

patients with acute respiratory distress syndrome.

JAMA 1999; 282: 54-61.

14. Ranieri, VM, Giunta, F, Suter, PM, Slutsky, AS.

Mechanical ventilation as a mediator of multisystem

organ failure in acute respiratory distress syndrome.

JAMA 2000; 284: 43-4.

15. Slutsky, A. Mechanical ventilation: report of American

College of Chest Physicians Consensus Conference.

C/M?sM993; 104: 1833-59.

16. Gattinoni, L, D'Andrea, L, Pelosi, P Vitale, G, Pesenti, A,

Fumagalli, R. Regional effects and mechanism of

positive end-expiratory pressure in early adult

respiratory distress syndrome.JAMA 1993; 269:

2122-7.

17. Hickling, K. Recruitment greatly alters the pressure

volume curve: a mathematical model of ARDS lungs.

Am J Respir Crit Care Med 1998; 158: 194-202.

18. Jonson, B, Richard, JC, Straus, C, et al. 1999.

Pressure-volume curves and compliance in acute lung

injury. Evidence of recruitment above the lower

inflection point. Am J Respir Crit Care Med 1999; 159:

1172-8.

19. Brimioulle, S, Vachiery, JL, Lejeune, P, et al. Acid-base

status affects gas exchange in canine oleic-acid

pulmonary edema. Am J Physiol (Heart Circ Physiol 29)

1991; 260: H1086.

20. Laffey, JG, Tanaka, M, Engelberts, D, et al. Therapeutic

hypercapnia reduces pulmonary and systemic injury

following in vivo lung reperfusion. Am J Respir Crit Care

Mo/2000; 162:2287-94.

21. Morris, AH, Wallace, CJ, Menlove, RL, et al. Randomized

clinical trial of pressure-controlled inverse ratio

ventilation and extra corporeal C02 removal for adult

respiratory distress syndrome. Am J Respir Crit Care

/Wed1994; 149:295-305.

22. Hirschl, RB, Pranikoff, T, Wise, C, et al. Initial

experience with partial liquid ventilation in adult

patients with the acute respiratory distress syndrome.

JAMA 1996; 275: 383-9.

23. Anzueto, A, Baughman, RP, Guntupalli, KK, et al.

Aerosolized surfactant in adults with sepsis-induced

acute respiratory distress syndrome. Exosurf Acute

Respiratory Distress Syndrome Sepsis Study Group. N

Engl} Med 1996; 334: 1417-21.

24. Dellinger, RP, Zimmerman, JL, Taylor, RW, et al. Effects

of inhaled nitric oxide in patients with acute

respiratory distress syndrome: results of a randomized

phase II trial. Crit Care Med 1998; 26:15-23.

25. Lundin, S, Mang, H, Smithies, M, Stenqvist, 0, Frostell,

C, for the European Study Group of Inhaled Nitric

Oxide. Inhalation of nitric oxide in acute lung injury:

results of a European multicentre study. Intensive Care

Med 1999; 25: 911-19.

References 169

26. Meduri, GU, Headley, AS, Golden, E, et al. Effect of 28. Abraham, E, Baughman, R, Fletcher, E, et al. Liposomal

prolonged methylprednisolone therapy in unresolving prostaglandin E1 (YTLC C-53) in acute respiratory distress

acute respiratory distress syndrome: a randomized syndrome: a controlled, randomized, double-blind,

controlled trial. JAMA 1998; 280: 159-65. multicenter trial. Crit Care Med 1999; 27:1478-85.

27. NIH ARDS Network. Ketoconazole for early treatment 29. Bernard, GR, Wheeler, AP, Russell, JA, et al. The effects

of acute lung injury and acute respiratory distress of ibuprofen on the physiology and survival of

syndrome. JAMA 2000; 283: 1995-2002. patients with sepsis. N Engl J Med 1997; 336: 912-18.

13Weaning from mechanical ventilationSTEFANO NAVA, MICHELE VITACCA AND ANNALISA CARLUCCI

The size of the problem

Factors delaying weaning

The weaning process

What ventilator mode speedsweaning?

Extubation criteria

Weaning protocols

170

170

172

173

173

173

Self-extubation 175

Post-extubation failure 176

Weaning failure 176

Nutritional, psychological and rehabilitative aspects 177

Volume-reduction surgery and transplantation 177

Conclusion 178

References 178

THE SIZE OF THE PROBLEM

There are a number of problems when consideringthe concept of 'weaning'. First, the clinician needs todistinguish between liberation from mechanical ven-tilation, when support is no longer required, andextubation, when there is no longer a need for theendotracheal tube. Accordingly, after a patient hassuccessfully undergone a trial of unassisted breath-ing, one must decide whether access to the lowerrespiratory tract is still required. The term 'weaningsuccess' therefore applies only after both conditionshave been met.l Another problem is the definition ofweaned. What is the minimum time that the patientmust remain disconnected from a ventilator to beconsidered to have 'weaned'? This is rarely addressedin published reports. In most, a time interval of 48hours after extubation or disconnection from sup-port is specified. The assumption is that, after this,failure is likely to be for non-respiratory reasons.However, this may not be the case. How large is theproblem of weaning delay or failure? About 80% ofpatients admitted for mechanical ventilation in theintensive care unit (ICU) resume spontaneousbreathing in a few days. In the remaining 20%, oftena combination of unresolved primary illness and

pre-existing cardiorespiratory or neuromuscular dis-ease renders discontinuation from mechanical venti-lation difficult. The length of weaning is, therefore,dependent upon the aetiology of respiratory failure.When the need for ventilation persists for more than15 days, as occurs in 2-5% of ICU admissions(depending on case-mix), it carries a poor prognosis(>50% mortality). Moreover, failure of extubationincreases the risk for death, prolongs ICU stayand may lead to the need for transfer to long-termcare or rehabilitation. In a national survey in Spain,2

it was reported that 41% of the overall ICU timewas devoted to weaning, with large differencesbetween different aetiologies necessitating mechanicalventilation. In patients with chronic obstructivepulmonary disease (COPD), cardiac failure or neuro-logical problems, the process of weaning accountedfor >50% of ICU stays.2

FACTORS DELAYING WEANING

The causes of weaning delay include:

• unresolved primary illness,• nosocomial infection,• pre-morbid COPD or heart failure,

Factors delaying weaning 171

• upper airway problems, such as glottic oedema orbulbar disease,

• corticosteroids,• impaired consciousness,• electrolyte disturbances,• haemodynamic instability,• critical illness neuropathy/myopathy.

Endotracheal intubation may cause complica-tions that increase morbidity and mortality.3

Furthermore, the need for sedation or paralysis toenable effective mechanical ventilation, particularlyin the initial days of critical illness, may lead to a gen-eralized myopathy. The evidence for this is con-tentious. The association with, for instance, paralyticagents may be explained by confounding factors suchas disease severity. The use of prolonged mechanicalventilation may lead to diaphragmatic atrophy. Inone report, this developed after 48 hours in a studyperformed on rats.4 However, other animal experi-ments suggest that diaphragm atrophy does notoccur if paralysis is partial or intermittent (only a fewcontractions per day are required to prevent wast-ing). Impairment in skelet al muscle strength in theICU may also be a consequence of electrolyte disturb-ances5 or a direct effect of hypercapnia, hypoxia,malnutrition, treatment with corticosteroids and alow cardiac output. Nevertheless, critically illpatients with sepsis and multiple organ failure(MOF) are at risk of developing critical illness neur-opathy, which is, in most cases, a combination ofmyopathy and neuropathy.6 The weakness of the res-piratory muscles that follows is probably one of themajor determinants of weaning failure in patientsrecovering from critical illness. In COPD, the alteredshape of the diaphragm, with flattening from hyper-inflation, will result in a mechanical disadvantagethat will impair function. The force generation of thediaphragm is not the only factor involved in weaningdelay. The respiratory pump output, in terms ofminute ventilation, will be the result of the balancebetween the load on the pump and its capacity (Fig.13.1). In many causes of acute respiratory failure,such as COPD, acute respiratory distress syndrome(ARDS), pneumonia and heart failure, the elastic andresistive loads are elevated as much as three to fourtimes.7 In COPD, in particular, respiratory muscletraining has been suggested as a strategy to increasethe strength and endurance of the diaphragm.Similowski et al.8 demonstrated that the diaphragmin COPD is able to generate a normal pressure, inresponse to bilateral phrenic nerve stimulation,

when corrected for hyperinflation. Their findingsargue against the value of inspiratory muscle train-ing (or weaning by intermittent periods of increasedload, e.g. T-tube breathing). Similowski concludesthat 'the absence of central inhibition or evidence ofchronic fatigue cast doubt on the need to treatpatients with interventions intended to improve thecontractility of the diaphragm'. More recently, Levineand co-workers9 provided support for this viewwhen they showed, in biopsy specimens taken fromthe diaphragm in severe COPD, that there is anincrease in the slow-twitch fibres, presumably as anadaptive mechanism that will lead to an increase inresistance to fatigue (see Chapter 2).

Infection is an important cause for delay inweaning. The presence of an endotracheal tube formore than 3 days significantly increases the risk ofnosocomial pneumonia (>20% at 10 days of invasivemechanical ventilation). Nosocomial pneumoniaresults in a longer hospital stay as well as an increasein mortality. The endotracheal tube predisposes topneumonia:

1. by impairing cough and mucociliary clearance,2. by allowing aspiration of contaminated secretions

that accumulate above the cuff, or

Figure 13.1 Balance between the loads imposed on the

respiratory system and its capacity in normal subjects and in

respiratory diseases. COPD, chronic obstructive pulmonary disease;

ARDS, acute respiratory distress syndrome.

172 Weaning from mechanical ventilation

3. because bacterial binding to the surface ofbronchial epithelium is increased.

The use of non-invasive ventilation reduces the riskof ventilator-associated pneumonia. For this reason, itwas recently suggested10 that the term ventilator-associated pneumonia be replaced by intubation-associated pneumonia. Unfortunately, the use ofnon-invasive ventilation is not always successful,especially in patients without alveolar hypoventilation,and non-invasive ventilation is currently employedin selected patients with hypercapnic respiratoryfailure (see Chapter 5). Aspiration of regurgitatedstomach contents is a potent cause of ventilator-associated pneumonia, and the risk of this withnasogastric feeding is high. In one study, it occurred inapproximately 50% of patients.11 A semi-recumbent,rather than fully recumbent, body position may reduceregurgitation and thus the frequency of nosocomialpneumonia in patients receiving enteral nutrition.12

Cardiac failure is another important cause ofdelayed weaning. In patients undergoing a trial ofspontaneous breathing, a progressive decrease inSvO2, caused both by a decrease in 02 transport and anincrease in 02 extraction, increased the failure rate.13 Aprotracted ICU stay also leads to complications due toconfinement to bed and general de-conditioning, withchanges in skelet al muscle composition, alteredcardiovascular response to stress, bone demineraliza-tion, protein wastage and a decrease in total bodywater. The central nervous system, endocrine functionand blood composition may also alter.14

THE WEANING PROCESS

The first step in the weaning process is the identifica-tion of the patient potentially able to sustainspontaneous breathing. The following requirementsare considered necessary.

• Clinical stability, defined as the absence of sepsis,significant bronchospasm, electrolyte imbalanceor over-sedation, evidence of profound malnutri-tion, excessive secretions and/or weak cough orhypotension (systolic <90 mmHg).

• Normal or only mildly disturbed central nervoussystem function.

• Adequate oxygenation (PaO2 > 8.0 kPa or 60mmHg) with FiO2 < 40% and end-expiratorypressure < 5.0 cmH2O in patients able to triggerthe ventilator.

The second step is the choice of the modality forliberation from mechanical ventilation.The mostcommon techniques are:

• Pressure support ventilation (PSV): a progres-sive decrease in pressure support is continueduntil minimal support is provided, e.g.8 cmH2O,

• T-piece: a period of spontaneous breathingthrough the endotracheal tube connected to aT-piece,

• SIMV: the patient can breathe spontaneouslybetween ventilator-delivered breaths.

Pressure support ventilation

PSV is a widely used method of partial mechanicalventilatory support that efficiently reduces the work-load of the inspiratory muscles. It can be used duringweaning by progressively decreasing the level ofassistance. At a pressure support level of 8 cmH2O,the work imposed by the endotracheal tube anddemand valve of the ventilator is equivalent to spon-taneous breathing, so that extubation will notincrease the work of breathing.15

T-piece

This is a method of weaning based on the philosophythat the respiratory muscles need retraining, withperiods of disconnection from the ventilator forlengthening periods. During these periods, patientsreceive humidified O2-enriched gas through aT-piece connected to the endotracheal tube.

Synchronized intermittent mechanicalventilation

In this mode of ventilation, the patient is able tobreathe spontaneously between ventilator-deliveredbreaths. Delay in the timing of machine breaths isreferred to as synchronization; a variable 'lock-out'period after machine provided breaths limits breath'stacking' or hyperventilation. Spontaneous breathsmay also be supported - so-called SIMV plus pres-sure support. The weaning procedure consists ofgradually decreasing the number of machine-deter-mined breaths.

Weaning protocols 173

Other techniques have also been proposed, e.g.mandatory minute ventilation, volume-assured pres-sure support and BiPAP. There is no evidence for theroutine use of these methods in the clinical practice.

WHAT VENTILATOR MODE SPEEDSWEANING?

Two important multi-centre trials were performed inthe mid 1990s by Brochard and co-workers16 and byEsteban and colleagues.17 Both compared the follow-ing methods: intermittent T-piece breathing, PSVand SIMV. The former study, performed on 456medical and surgical patients, concluded that theoutcome of weaning was influenced by ventilatorystrategy and found that the use of PSV (attemptedreductions of pressure support by 2-4 cmH2O twicea day until pressure support was 8 cmH2O) resultedin significantly faster weaning than the other twotechniques. On the other hand, the latter study foundthat a once-daily trial of spontaneous breathing witha T-piece of gradually increasing duration led toextubation three times more quickly than SIMV andtwice as quickly as PSV. In this study, the minimumtarget pressure support level before extubation was5 cmH2O. The explanation for these contrastingresults is that the method employed is probably lessimportant than the patient pathology in determiningthe duration of mechanical ventilation. Confidenceand familiarity with the technique adopted are likelyto be more important than the chosen method.Certainly, the rate of failure following extubation wassimilar after T-piece weaning and PSV, which sug-gests that either approach is acceptable in the libera-tion from mechanical ventilation. In both studies,weaning trials were performed only in those patientswho had first failed a 2-hour T-piece trial. Thesepatients represented about 25% of all patients whohad reached the criteria for weaning. Accordingly,the majority of patients judged 'weanable' (75%),according to the above criteria, could be safely extub-ated after a single brief trial of spontaneous breath-ing. More recently, it has been shown that successfulextubation can be achieved using a shorter trial period(30 min) of spontaneous breathing.18 Before pro-ceeding to extubation, additional aspects that requireconsideration are the volume and character of secre-tions and the ability of the patient to cough effect-ively. In addition to an adequate cough reflex, patient

co-operation and good bulbar function are neces-sary. The criteria to judge the outcome of a period ofT-piece breathing are:

• subjective comfort,• physiological stability (no significant increment

in heart rate and respiratory rate),• absence of an acute respiratory acidosis or

hypoxemia, defined as PaO2 < 8.0 kPa on 40%inspired O2.

EXTUBATION CRITERIA

Many parameters have been proposed to predict suc-cessful extubation: the Index of Rapid ShallowBreathing (RSB), or Tobin's Index.,19 has a predictiveutility superior to other proposed indices such asvital capacity, maximal inspiratory pressure (Pimax)or tidal volume, and is simpler and less invasive thancomplex measurements such as neuromuscular driveto breathe (-P01), the Work of Breathing Index,CROP (acronym for compliance, rate, oxygenationand pressure) or other parameters of respiratorymechanics. The RSB Index relates respiratory fre-quency and tidal volume (RR/Vt) with a threshold of>105 breaths mL-1 predicting weaning failure.Despite good sensitivity in predicting success whenRSB <105, the index lacks specificity. Thus, somepatients with an RR/Vt > 105 may, in fact, succeed,whereas a few with RR/Vt < 105 may fail.1

WEANING PROTOCOLS

The concept of using a standardized protocol towean patients from mechanical ventilation is popu-lar in the USA. Therapist-driven protocols combineextubation criteria with daily care plans, changes intherapy being directed by changes in measurablepatient variables. The daily screening of respiratoryfunction by nurses or respiratory therapists, followedby trials of spontaneous breathing and notificationto the patients' physicians when the trials are suc-cessful, can reduce the duration of ventilation, thecost of intensive care and the rate of complications.20

Whereas these protocols may be applied in the USA,where respiratory therapists are directly involved inthe weaning procedure, they may not be in Europe,where only 22% of respiratory therapists working in


ICUs are directly involved in ventilator manage-ment.21 Therapist-driven protocols are a consensusof medical knowledge and opinion that are summar-ized into a care plan or algorithm, with changes intherapy directed by changes in objectively measur-able patient variables. It is important to stress thespecific roles of the respiratory therapists in this pro-cedure. With the institution of therapist-driven pro-tocols, the interaction between the therapist andnurse regarding the indications for arterial blood-gasand maximal inspiratory pressure (MIP) measure-ments, bronchial secretions management or numberof hours of T-tube breathing results in a significantchange in the behaviour of nursing and medical staff.The whole therapist-driven protocol team consists ofthe physician, patient, family, nurse and a respiratorytherapist. It addresses the prevention of the deleteri-ous effects of bed rest, communication, emotionalsupport, psychological well-being and function. Theinitial evaluation includes assessment of the patientand ventilator status and patient-ventilator syn-chrony. This evaluation is performed routinely every2 hours and with each ventilator setting change. Theuse of respiratory therapists is also important duringthe application of non-invasive ventilation, one ofthe keys to the success of which is the continuous

Figure 13.2 NIV to aid early extubtion in COPD (see reference 22).

Distribution of the patients according to treatment and outcomes.

MV, mechanical ventilation; NIV, non-invasive ventilation.

monitoring, preparation and nursing of patients. Forthis reason, in the first phases of treatment, the pres-ence of the respiratory therapist and/or nurse is nec-essary to ensure correct positioning of the mask, tocoach the patient, to aspirate bronchial secretionsand to ensure compliance and tolerance. Nursinginput is clearly important and may affect success. Inone study, a significant inverse correlation was foundbetween the duration of mechanical ventilation andavailability of nurses as assessed by a nursing index.

A major limitation to 'accelerated' extubationremains the lack of criteria that guarantee success.However, removing the endotracheal tube does notnecessarily mean that respiratory support cannot beprovided. Instead of relying upon spontaneous venti-lation in the immediate post-extubation period, non-invasive ventilation, with face or nasal mask, can beemployed as a 'bridge'. This technique has been vali-dated in a randomized, multi-centre Italian study inpatients with COPD.22 The trial involved patientsintubated for acute hypercapnic respiratory failure,either after initial failure of or contraindications tonon-invasive ventilation. Patients were initially sedated,and often paralysed, and frequent bronchial toilet wasperformed in the first 6-12 hours. In the subse-quent 24-36 hours, sedation was reduced andmechanical ventilation provided with pressure sup-port. At 48 hours after intubation, a T-piece trial wasperformed if the patients were haemodynamically sta-ble, had a normal temperature, were alert and had noevidence of pneumonia. Strict criteria, similar to theBrochard eta/.16 extubation requirements, were usedto judge failure of the T-piece trial. Patients who failedthis trial (50 of the initial 68 patients) were than ran-domized to either re-institution of full mechanicalventilation and conventional weaning or temporaryreconnection to the ventilator, until previous arterialblood-gas levels were reached, and then extubationonto non-invasive ventilation. In both groups, wean-ing proceeded by daily reductions in the level of pres-sure support and intermittent spontaneous breathingtrials twice a day (Fig. 13.2). The mean durations ofventilatory support and ICU stay and the 60-day mor-tality were significantly reduced in the non-invasivelyventilated group. Importantly, none of the patientsweaned non-invasively developed nosocomial pneu-monia, whereas 7 (28%) of those who continued withendobronchial ventilation did. Another study,employing non-invasive ventilation a few days afterintubation in patients with hypercapnic respiratory

Self-extubation 175

Figure 13.3 Algorithm for weaning from

mechanical ventilation. CNS, central nervous

system; PEEP, positive end-expiratory pressure;

RSB, Index of Rapid Shallow Breathing; PSV,

pressure support ventilation; NIV, non-invasive

ventilation; COPD, chronic obstructive

pulmonary disease.

failure due to COPD or restrictive thoracic disease,showed similar results, although the reduction ininfectious complications did not achieve statisticalsignificance.23 Finally, a study using non-invasiveventilation for weaning patients without pre-existingCOPD (most following lung transplant procedures)also reported that nosocomial pneumonia can beavoided by using non-invasive PSV and that theduration of mechanical ventilation, the duration ofICU stay and the need for re-itubation can bereduced.24 These studies therefore suggest that non-invasive ventilation may allow patients to be extubatedearlier, more successfully and with fewer complicationsthan conventional weaning procedures.

However, weaning with non-invasive ventilationshould be performed with caution. Published studiespertain mostly to patients with hypercapnic respira-tory failure (pump failure) due to COPD who hadbeen carefully selected to be haemodynamicallystable with PaO2:FiO2 ratio > 1.5, no evidence ofinfection or depressed consciousness and an effective

cough. Until large-scale studies have been under-taken, it is difficult to quantify how many intubatedCOPD patients may be successfully managed thisway. Further studies are also needed to assess thefeasibility of the technique in other forms of respira-tory failure. Keeping in mind these limitations, apractical algorithm for the weaning process is pro-posed in Figure 13.3.

SELF-EXTUBATION

Unplanned extubation occurs in 8-14% ofpatients.25'26 It occurs more frequently with oralintubation, with insufficient sedation, in patientswith COPD and when the fixation of the tube ispoor. The re-intubation rate varies in different series,but is considerably lower after unplanned extub-ations that complicate a weaning trial (15-30%) thanafter those occurring when the patients are receiving


full ventilatory support (75-80%). These figures sug-gest that a considerable proportion of patients couldbe liberated from mechanical ventilation at an earliertime! An unplanned extubation is not unimportant,because it appears to increase the duration ofmechanical ventilation and the length of ICU andhospital stay when compared with patients re-intub-ated following extubation failure. The mortality rate,interestingly, is not increased. On the other hand,when successful, unplanned extubation clearlyreduces the length of mechanical ventilation, but hasno other measurable beneficial impact.26

POST-EXTUBATION FAILURE

Failure following extubation is a common problem.The incidence ranges from 3.3% to 23.5%.27 Theprognosis in these patients is poor, with a hospitalmortality of 30-40%, depending on whether thecause is respiratory or non-respiratory. The time tore-intubation is an independent predictor of out-come. As re-intubation per se is an insufficient explan-ation for the high mortality rate, it has been claimedthat clinical deterioration during the period ofunsupported ventilation may allow the developmentof secondary organ failure that then leads to a poorprognosis.28 The prolonged period of unsupportedventilation may arise because the clinician wishes toavoid re-intubation due to the severity of underlyingdisease or because of concerns about the well-recog-nized complications of continued intubation andmechanical ventilation.

Hilbert and co-workers have described the so-called 'sequential use' of non-invasive ventilation,consisting of intermittent periods of non-invasiveventilation for 30 min every 3 hours.29 During peri-ods of spontaneous unsupported ventilation, patientscan be monitored and returned to non-invasive ven-tilation if Sa02 falls to <85% or the respiratory rateincreases to >30 breaths min"1.28 This sequential usehas been successfully employed in the management ofCOPD patients and may be applied to all patientsidentified as at risk of post-extubation failure. Thelimitation of this study is the use of historical con-trols. However, the use of sequential non-invasiveventilation significantly reduced the need for endo-tracheal re-intubation, the mean duration of ventila-tory assistance and the length of ICU stay, and the

mortality was (not statistically) three times higher inthe group treated conventionally. This study alsodemonstrated a lower incidence of pneumonia in thegroup treated non-invasively (7% versus 20%).

Randomized, controlled studies are needed to con-firm the utility of non-invasive ventilation in thesecircumstances (post-extubation hypercapnia), butalso in those with hypoxaemic respiratory failure.

WEANING FAILURE

Weaning failure, defined as mechanical ventilationpersisting after recovery from the initial critical ill-ness, is commonly due to a pre-existing disease29

and/or neuromuscular disease. In a subset of COPDpatients, weaning may be particularly difficult or evenimpossible. This may relate to the advanced nature oftheir disease, with a severe mechanical load beingplaced upon the respiratory muscles. Advanced age,respiratory muscle weakness (possibly due to treat-ment with corticosteroids) or co-existent cardiac dis-ease may contribute to difficulties in the weaningprocess. Once recovered from the acute phase illness,patients may continue to require intensive nursingand/or physiotherapy care for weeks before they canbe finally weaned or even be discharged ventilatordependent. In one study, these severely compromisedpatients, although only representing 3% of the totalnumber of patients admitted to the ICU, consumedalmost 40% of the total days of care.30

Patients are often judged as having failed weaningif they are still ventilator dependent 14-21 days fol-lowing recovery from the admission illness. Thisarbitrary threshold needs clarification, as does theincidence of such patients in the ICU. There may wellbe considerable differences between hospitals andbetween countries, reflecting admission criteria anddifferences in the withdrawal of mechanical ventila-tion (see Chapter 21). In many European countries,high dependency, step-down or long-term ventila-tion units may offer the opportunity for continuedweaning of such patients. Despite recovery from theacute illness, they may require intensive nursing andphysiotherapy care, nutritional and psychologicalsupport and a more gradual weaning process beforebeing judged totally ventilator dependent. One studyshowed that 60% of patients considered 'unweanable'at the time of ICU discharge regained respiratory

Nutritional, psychological and rehabilitative aspects 177

autonomy after a relatively short (mean 17 days) stayin a specialized respiratory apecial care unit.31 Forthose patients discharged to home mechanicalventilation by tracheostomy, the mortality rate at1 year is considerably higher, at 62-87%, comparedto the group successfully weaned (23-54%).Moreover, the former group remained severelydisabled and house bound.32'33

NUTRITIONAL, PSYCHOLOGICAL ANDREHABILITATIVE ASPECTS

Malnutrition and psychological aspects are import-ant reasons for weaning difficulties.34'35 A decreasein body weight is a recognized feature of advancedCOPD and is especially severe in recurrently hospi-talized patients. Nutritional state, measured as per-centage of ideal body weight, influences mortalityindependently of the degree of airflow obstruction.It has been suggested that the main contributor toweight loss in COPD is an inadequate dietary intakefor energy expenditure. The risk of being hospital-ized for an episode of acute respiratory failure is sig-nificantly increased in patients with a low body massindex (<20 kg m-2).36. Interestingly, in one report,body mass index and serum albumin were indepen-dently related to survival in acute respiratory failure.Malnutrition may contribute to respiratory andskelet al muscle weakness and have other metabolicconsequences. On the other hand, excessive nutri-tional supplements will increase CO2 productionand thus the amount of ventilation necessary tomaintain normocapnia. Further investigations maybe necessary to determine whether nutritional inter-vention may improve clinical outcome in weaning.Metabolic disorders, such as diabetes, often associat-ed with malnutrition, may lead to a decrease inimmunological defences, so that patients are moreprone to infections that prolong mechanical ventila-tion.37 The correlation between body mass indexand impaired respiratory muscle force may explainwhy malnourished patients are more likely to takelonger to wean.

Psychological problems have also been frequentlyreported in patients in whom weaning is difficult,although the literature is limited. Overall, psycholo-gical problems are found in more than 50% of ICUpatients, compared to an incidence of <1% in non-

ICU hospitalized patients.38 Sleep disturbance in theICU is very common and may be implicated in thepsychological disturbance of patients. Beingmechanically ventilated involves a loss of independ-ence and often ineffective communication, and pro-motes passivity. Depressive reactions often develop,making the process of weaning more difficult. Somestudies have investigated the use of respiratory feed-back to reduce anxiety. By using visual and auditoryfeedback of tidal volume, coupled with a display ofthe frontalis muscle electromyograph, relaxationwas induced. The protocol significantly reducedventilation time.39 Although these preliminaryresults are encouraging, the requirement for trainedpersonnel and sophisticated equipment limits itsapplication. Indeed, there is considerable controver-sy about whether biofeedback is a useful, or neces-sary, addition to relaxation techniques. Studies havebeen often poorly controlled and the results cannotbe generalized. Biofeedback may have a place whenused in a comprehensive multi-modal treatmentplan.

Dyspnoea (and resulting anxiety) is a frequentproblem for patients with COPD. In particular, diffi-culty in gaining independence from the ventilatormay relate to paradoxical breathing patterns, exces-sive respiratory drive and anxiety. A major goal in therehabilitation process is early mobilization. In add-ition to avoiding the adverse effects of prolongedinactivity, there are several advantages to getting outof the bed, e.g. improved mechanics of ventilation,the mobilization of secretions and the promotion ofself-confidence. It has been shown that COPDpatients recovering from acute respiratory failurebenefit from early mobilization, compared topatients who received standard medical therapy.40

Patients treated with such a protocol showedimproved effort tolerance, maximal inspiratoryeffort and dyspnoea score.

VOLUME-REDUCTION SURGERY ANDTRANSPLANTATION

No studies have systematically investigated whetherlung-volume-reduction surgery (LVRS) or lungtransplantion is appropriate in the patient who hasfailed weaning. Nevertheless, it has been reported thatLVRS in selected ventilator-dependent COPD


patients can result in improved gas exchange and res-piratory mechanics, which enable successful wean-ing.41 Inability to walk at least 200 m in 6 min, beforeor after pulmonary rehabilitation, and the presence ofsignificant hypercapnia are pre-operative predictorsof a longer hospital and ICU stay, duration ofmechanical ventilation and of the need for chest tubedrainage in COPD patients undergoing elective bilat-eral LVRS.42 The implication of this in economicterms is that cost was directly related to length of stayin the ICU: the range being 11%-30 % of the totalcosts of the LVRS programme.43,44 A thoracoscopicapproach, compared to open surgery, for LVRS mayoffer a shorter ICU stay, fewer days with an endotrach-eal tube, fewer respiratory complications and a lowermortality rate.45 For patients with cystic fibrosis, themean duration on ventilatory support for survivorsof bilateral sequential lung transplantation is 3.1 days,with a range of 1-12 days; the mean ICU and hospi-tal stays were 4.7 days (range 1-13 days ) and 28 days(range 12-79 days), respectively46

CONCLUSION

Although we now have a 'science' of weaning, withevidence on which to base practice, there remains an'art' to the process. The move to less sedation, moreinteractive modes of ventilation and greater empha-sis on general rehabilitation in the weaning patientmust be leading to less morbidity and, it is hoped,mortality.

REFERENCES

1. Manthous, C, Schmidt, GA, Hall, J. Liberation from

mechanical ventilation. A decade of progress. Chest

1998; 114:886-901.

2. Esteban, A, Alia, I, Ibanez, J, Bonito, S, Tobin, MJ.

Modes of mechanical ventilation and weaning: a

national survey of Spanish hospitals. Chest 1994; 106:

1188-93.

3. Stauffer, JL. Complications of translaryngeal

intubation. In Principles and practice of mechanical

ventilation, ed. M. Tobin. New York, McGraw Hill,

1994; 711-47.

4. Le Bourdelles, G, Vires, N, Bockzowki, J, Seta, N,

Pavlovic, D, Aubier, M. Effects of mechanical

ventilation on diaphragmatic contractile properties in

rats. Am J Respir Crit Care Med 1994; 149: 1539-44.

5. Agusti, AG, Torres, A, Estopa, R, Agustividal, A.

Hypophosphatemia as a cause of failed weaning: the

importance of metabolic factors. Crit Care Med 1984;

12:142-3.

6. Hund, EF, Fogel, W, Kreger, D, et al. Critical illness

polyneuropathy: clinical findings and outcomes of a

frequent cause of neuromuscular weaning failure. Crit

Care Med 1996; 24: 1328-33.

7. Broseghini, C, Brandolese, R, Poggi, R, et al.

Respiratory mechanics during the first day of

mechanical ventilation in patients with pulmonary

edema and chronic airway obstruction. Am Rev

Respir Dis 1988; 138(2): 355-61.

8. Similowski, T, Van, S, Gauthier, AP, Macklem, PT,

Bellemare, F. Contractile properties of the human

diaphragm during chronic hyperinflation. N Engl J

Med 1991; 325: 917-23.

9. Levine, S, Kaiser, L, Leferovich, J, Tikunov, B. Cellular

adaptations in the diaphragm in chronic obstructive

pulmonary disease. N Engl J Med 1997; 337:

1799-806.

10. Kramer, B. Ventilator-associated pneumonia in critically

ill patients. Ann Intern Med 1999; 130(12): 1027-8.

11. Elpern, EH, Scott, MG, Petro, L, Ries, MH. Pulmonary

aspiration in mechanically ventilated patients with

tracheostomies. Chest 1994; 105: 563-6.

12. Drakulovic, MB, Torres, A, Bauer, TT, Nicolas, JM, Nogue,

S, Ferrer, M. Supine body position as a risk factor for

nosocomial pneumonia in mechanically ventilated

patients: a randomised trial. Lancet 1999; 354:1851-8.

13. Jubran, A, Mathru, M, Dries, D, Tobin, Mj. Continuous

recordings of mixed venous oxygen saturation during

weaning from mechanical ventilation and the

ramifications thereof. Am J Respir Crit Care Med

1998; 158: 1763-9.

14. Bortz, WM. Disuse and aging. JAMA 1982; 248: 1203-8.

15. Brochard, L, Rue, F, Lorino, H, Lemaire, F, Harf, A.

Inspiratory pressure support compensates for the

add-itional work of breathing caused by the

endotracheal tube. Anesthesiology 1991; 75: 739-45.

16. Brochard, L, Rauss, A, Benito, S, et al. Comparison,of

three methods of gradual withdrawal from ventilatory

support during weaning from mechanical ventilation.


17. Esteban, A, Frutos, F, Tobin, MJ, et al. A comparison of

four methods of weaning from mechanical ventila-

tion. Spanish Lung Failure Collaborative Group. N Engl

J Med 1995; 332: 345-50.

References 179

18. Esteban, A, Alia, I, Tobin, MJ, et al. Effect of

spontaneous breathing trial duration on outcome of

attempts to discontinue mechanical ventilation.

Spanish Lung Failure Collaborative Group. Am J Respir

Crit Care Med 1999; 159: 512-18.

19. Yang, KL, Tobin, MJ. A prospective study of indexes

predicting the outcome of trials of weaning from

mechanical ventilation. N Engl J Med 1991; 324:

1445-50.

20. Ely, EW, Baker, A, Dunagan, D, et al. Effect of the

duration of mechanical ventilation on identifying

patients capable of breathing spontaneously. N Engl

j Med 1996; 335: 1864-9.

21. Norrenberg, M, Vincent, JL, A profile of European

intensive care unit physiotherapists. Intensive Care

Med 2000; 26: 988-94.

22. Nava, S, Ambrosino, N, Clini, E, et al. Noninvasive



pulmonary disease. A randomized, controlled trial.

Ann Intern Med 1998; 128: 721-8.

23. Girault, C, Daudenthum, I, Chevron, V et al.

Non-invasive ventilation - a systematic extubation

and weaning technique in acute-on-chronic

respiratory failure: a prospective, randomized

controlled study. Am J Respir Crit Care Med 1999;

160: 86-92.

24. Kilger, E, Briegel, J, Haller, M, et al. Effects of

noninvasive positive pressure ventilatory support in

non-COPD patients with acute respiratory

insufficiency after early intubation. Intensive Care Med

1999; 25: 1374-9.

25. Chevron, V, Menard, JF, Richard, JC, Girault, C, Leroy, J,

Bonmarchand, G. Unplanned extubation: risk factors

of development and predictive criteria for

reintubation. Crit Care Med 1998; 26: 1049-53.

26. Epstein, SK, Nevins, ML, Chung, J. Effect of unplanned

extubation on outcome of mechanical ventilation. Am

J Respir Crit Care Med 2000; 161: 1912-16.

27. Epstein, SK, Ciubataru, RL, Wong, JB. Effect of failed

extubation on the outcome of mechanical ventilation.

Chest-\997; 112: 186-92.

28. Espstein, SK, Ciubotaru, RL. Independent effects of

etiology of failure and time to reintubation on

outcome for patients failing extubation. Am J Respir

Crit Care Med 1998; 158: 489-93.

29. Hilbert, G, Gruson, D, Gbikpi-Benissan, G, Cardinaud,

JP. Sequential use of noninvasive pressure support

ventilation for acute exacerbations of COPD. Intensive

Care Med 1997; 23: 955-61.

30. Daly, BJ, Rudy, ED, Thompson, KS, et al. Development

of a special care unit for chronically ill patients. Heart

Lung 1991; 20: 45-52.

31. Nava, S, Confalonieri, M, Rampulla, C. Intermediate

respiratory intensive care units in Europe: a European

perspective. Thorax 1998; 53: 798-802.

32. Menzies, R, Gibbons, W, Goldberg, P. Determinants of

weaning and survival among patients with COPD who

require mechanical ventilation for acute respiratory

failure. Chest 1989; 95: 398-405.

33. Nava, S, Rubini, F, Zanotti, E, et al. Survival and

prediction of successful ventilator weaning in COPD

patients requiring mechanical ventilation for more

than 21 days. Eur Respir J 1994; 7: 1645-52.

34. Al-Saady, NM, Blackmore, CM, Bennett, ED. Might fat,

low carbohydrate enteral feeding lowers PaC02 and

reduces the period of ventilation in artificially

ventilated patients. Intensive Care Med 1989; 15:

290-5.

35. Holliday, JE, Hyers, TM. The reduction of weaning time

from mechanical ventilation using tidal volume and

relaxation biofeedback. Am Rev Respir Dis 1990; 141:

1214-20.

36. Kessler, R, Faller, M, Fourgaut, G, Mennecier, B,

Weitzenblum, E. Predictive factors of hospitalization

for acute exacerbation in a series of 64 patients with

chronic obstructive pulmonary disease. Am J Respir

Crit Care Med 1999; 159: 158-64.

37. Almirall, J, Bolibar, I, Balanzo, X, et al. Risk factors for

community-acquired pneumonia in adults: a

population-based case-control study. Eur Respir J1999;

13: 349-55.

38. Criner, GJ, Isaac, L. Psychological problems in the

ventilator-dependent patient. In Principles and

practice of mechanical ventilation, ed. M. Tobin. New

York, McGraw-Hill, 1994; 1163-75.

39. Acosta, F. Biofeedback and progressive relaxation in

weaning the anxious patient from the ventilator: a

brief report. Heart Lung 1988; 17(3): 299-301.

40. Nava, S. Rehabilitation of patients admitted to a

Respiratory Intensive Care Unit. Arch Phys Med Rehabil

1998; 79: 849-54.

41. Criner, GJ, O'Brien, G, Furukawa, S, et al. Lung volume

reduction surgery in ventilator dependent COPD

patients. Chest 1996; 110: 877-84.

42. Szekely, LA, Oelberg, DA, Wright, C, et al.

Preoperative predictors of operative morbidity

and mortality in COPD patients undergoing

bilateral lung volume reduction surgery. Chest 1997;

111:550-8.


43. Albert, RK, Lewis, S, Wood, D, Benditt, J. Economicaspects of lung volume reduction surgery. Chest 1996;110: 1068-71.

44. Elpern, EH, Behner, KG, Klontz, B, Warren, WH,Szidon, P, Kesten, S. Lung volume reduction surgery. Ananalysis of hopsital costs. Chest 1998; 113: 896-99.

45. Roberts, JR, Bavaria, JE, Wahl, P, Wurster, A,Friedberg, JS, Kaiser, LR. Comparison of open and

thoracoscopic bilateral volume reduction surgery:complications analysis. Ann Thome Surg 1998; 66:1759-65.

46. Wiebe, K, Wahlers, T, Harringer, W, Hardt, H,Fabel, H, Haverich, A. Lung transplantation forcystic fibrosis: a single center experienceover 8 years. Eurj Cardiothorac Surg 1998;14: 191-6.

14Community-acquired pneumoniaWEI SHEN LIM AND JOHN T MACFARLANE

Background epidemiology

Patient characteristics

Getting admitted to the intensive care unit

In the intensive care unit

Microbiological investigation

181

181

182

182

184

Management strategies

Patient not improving

Summary

References

186

188

189

190

BACKGROUND EPIDEMIOLOGYrates of mechanical ventilation of patients in ICU,which vary from 50% to 87%.

Community-acquired pneumonia (CAP) is com-mon. In the UK, there are approximately 261 000general practice consultations for CAP, which annu-ally result in 83 000 hospital admissions. The esti-mated incidence of CAP is 5-11 per 1000 adultpopulation per year. The elderly carry the burden ofthe disease, those ^65 years of age comprising 65%of patients with CAP in the UK.

Most respiratory infection is managed in the com-munity. CAP accounts for about 5-12% of adultlower respiratory tract infection managed by generalpractitioners (GPs) in the community. It has beenestimated that for every 100 cases of CAP seen in pri-mary care, 20 are referred to hospital and, of these,only one to two (10%) will require admission to anintensive care unit (ICU). The proportion of patientsadmitted to ICU varies in different countries (NewZealand, 1-3%; UK, 5%; USA, 12-18%; Germany,35%) and is influenced by ICU designation andadmission criteria.

Mortality rates range from 6% to 15% in hospital-ized patients, increasing to 22-54% in ICU series.Again, this broad range reflects admission criteriaand is consequently proportionately related to the

PATIENT CHARACTERISTICS

Based on cohort studies conducted over the last 20years, the mean age of patients with CAP admitted toICUs is around 50 years, ranging from 18 to 89 years.Associated chronic illness is common (70%), with>30% of cases having underlying lung disease. InSpain, high alcohol intake has been noted in manypatients.

The time from the onset of symptoms to ICUadmission is generally about 4 days. Patients laterfound to have pneumococcal infection tend to beadmitted more quickly compared to patients withLegionella or Mycoplasma infections. Importantly,half to three-quarters of the admissions to ICUsoccur within the first 24 hours of hospital admission.Whether earlier ICU referral would result in a betteroutcome is unclear. In a cross-sectional study com-paring patients admitted to the ICU at one hospital10 years apart, no change in mortality was noted,despite a reduction in the numbers of unplannedemergency ICU admissions resulting from car-diopulmonary arrest which fell from 25% to 17%.

182 Community-acquired pneumonia

GETTING ADMITTED TO THE INTENSIVECARE UNIT

ICU admission is usually a result of acute respiratoryfailure requiring assisted ventilation. An unstableclinical condition precipitated by septic shock oracute renal failure accounts for other cases admitted.The American Thoracic Society recommended thatthe presence of either one of two major criteria, ortwo of three minor criteria (Table 14.1) shouldprompt ICU care.2 These admission criteria havebeen tested, with admission to a specialized tertiaryreferral ICU as the outcome measure. A high sensitiv-ity (98%) was demonstrated, but there was low speci-ficity (32%).3 Modifying the criteria to include bothmajor and minor features has been suggested toincrease the sensitivity and specificity. However, theserecommendations require validation.

The British Thoracic Society has defined severeCAP as the presence of two or more of the followingfeatures measured on hospital admission:4

• respiratory rate >30 min"1,• diastolic blood pressure <60 mmHg,• urea >7 mmol L-1 .• mental confusion

This severity prediction rule has been widely validatedand has an overall sensitivity of 83% and a specificityof 70% for predicting mortality.5-7 Its value inpredicting the need for ICU admission has not beentested. However, no admission criteria can be ideal,especially as there is no universally accepted definitionfor severe CAP. These recommendations help informthe decision to refer or admit a patient for ICU care,but cannot replace what is, ultimately, a clinical deci-sion. Access to a high dependency unit (HDU) ormonitored respiratory care beds will be an alternativeoption in some hospitals with such facilities, where the

Table 14.1 Criteria for severe community-acquiredpneumonia according to the 2001 American Thoracic Societyrecommendations

Major criteriaNeed for mechanical ventilationSeptic shock

Minor criteriaSystolic blood pressure > 90 mm HgPa02/Fi02 ratio < 250Multilobar disease

need for assisted ventilation is not deemed immediate.No guidelines for admission to a HDU have beenpublished, but they are likely to be similar to those forICU admission.

We recommend the following as indications forconsidering transfer to an ICU:

• severe CAP as judged by the British ThoracicSociety severity prediction rule,

• inability to maintain PaO2 > 8 kPa despite maxi-mal O2 therapy,

• severe acidosis with pH < 7.26,• the presence of, or worsening, CO2 retention,• a depressed level of consciousness or patient

exhaustion,• shock — defined as a > 1-hour decrease in systolic

blood pressure of >40 mmHg from baseline or asystolic blood pressure of <90 mmHg afteradequate volume replacement.

The influence of age and co-morbidity

Advancing age blunts the immune response, bothlocally and systemically. Older people are therefore athigher risk of developing pneumonia and are morelikely to present with atypical symptoms, such as theabsence of pyrexia. It has been suggested that, withincreasing age, the mortality of CAP also increases.Studies conducted exclusively in older patients haveyielded conflicting results and suggest that, at least inhospitalized patients, an age >65 years does notincrease mortality.8-10 Studies of prognostic factorshave yielded similar results. In two studies conduct-ed in Europe, patients aged >65 years admitted tothe ICU with CAP did not have a higher ICUmortality.11,12 However, this could be due to morestringent ICU admission criteria in the elderly.Furthermore, the risk of mortality followingdischarge from the ICU is not known. Nonetheless,on balance, the evidence does not suggest that agealone should deny ICU admission to the elderly.

IN THE INTENSIVE CARE UNIT

Microbial pathogens

Depending on the range of diagnostic testsemployed, a pathogen is identified in 40-82% of

In the intensive care unit 183

cases. The frequency of identification of pathogensin 14 CAP ICU studies is shown in Figure 14.1.

Streptococcus pneumoniae is the most commonlydetected pathogen in these studies. In areas wherethere is a high prevalence of penicillin-resistantStrep, pneumoniae, there is as yet no strong evi-dence to indicate that patients with penicillin-resistant pneumococcal pneumonia (PRPP) have amore severe illness. Currently, most cases displayonly low to intermediate (0.1-1 mg L - 1 ) levels ofresistance and remain susceptible to high doses ofpenicillin. However, as the proportion of high-levelresistant (> 2 mg L - J ) cases increases, the clinicalimportance and severity of PRPP may alter.Many of these PRPP isolates are also resistant toerythromycin (25% in the USA). In the UK, theprevalence of penicillin-resistant Strep, pneu-moniae is still relatively low (2-4% in isolates frominvasive samples).

The importance of Legionella infection in patientsadmitted to the ICU with CAP was underlined bystudies conducted in the 1980s and emerging mainlyfrom the UK and Spain. In the UK, it accounted for12-30% of cases. More recent studies have not alwaysfound Legionella infection to be as prominent. In onecentre in Spain, the frequency dropped from 14% in1984 to 2% in 1996. This fall in frequency may, inpart, be due to year-to-year variation and toincreased use of macrolides in the early treatment ofCAP in the community.13 A knowledge of localepidemiological patterns is therefore important.

Figure 14.1 Pathogens identified in patients with CAP admitted to

an ICU from 14 studies. GNEB, Gram-negative enteric bacilli.

(Figures are under-represented as they do not include studies in

which specific pathogens are mentioned, e.g. Escherichia coll.)

Legionella pneumonia is commonly perceived ashaving a high mortality, probably reflecting both thehigh proportion of patients with Legionella infectionwho are admitted to the ICU compared to those withother infections and the unusually high publicity thisinfection has attained.14 However, Legionella is notan independent risk factor for mortality in CAP anddoes not increase the mortality in those patients whorequire mechanical ventilation.15

Clinical features considered typical for Legionellapneumonia include a dry cough, neurologicalsymptoms, such as mental confusion and headache,diarrhoea and a low sodium level. In the UK, overhalf of all cases are associated with travel abroad inthe 10 days prior to onset of illness. Unfortunately,none of these features is specific for Legionellainfection and none can be relied upon to confidentlydiscriminate infection caused by Legionella sp. fromother pathogens. Rhabdomyolysis is an uncommonbut well-recognized complication. Creatininekinase levels >50 000 U L-1 have been reported,and acute renal failure may be precipitated.Interestingly, the organism has been demonstratedby immunofluorescence in renal biopsy samples insuch cases.16

In all studies, the other atypical pathogens(C. pneumoniae, C. psittici, M. pneumoniae andC. burnetti) are found in only 2-3% of cases. Mor-tality in these cases tends to be lower, consistentwith the generally less severe nature of theseinfections.

Staphylococcus aureus is well recognized as animportant pathogen in severe CAP and carries ahigh mortality (60-100%). In about 50% of cases, itis associated with recent influenza virus infection. Ittherefore needs to be considered in the wintermonths during high influenza activity. In France, itis reported to be a more frequent cause of severeCAP (15%),n although the reason for this is notclear. In a typical case, there is a history consistentwith influenza followed up by a period of relativerecovery, before further deterioration within 2weeks. Infection may be complicated by lungabscess formation, cavitation and empyema.Pneumatoceles have been described, more com-monly in paediatric cases. Methicillin-resistantStaph. aureus (MRSA) is an increasing problem,particularly in nosocomial infection in the ICU.Patients at risk for MRSA pneumonia who comefrom outside of the hospital include the elderly, the


immunocompromised and those with recentprolonged hospitalization during which MRSAcolonization may have developed. MRSA is awell-recognized pathogen in nosocomial pneumo-nia, but patients presenting with CAP as a result ofMRSA infection are now being reported. Anotherrisk factor may be admission from a nursing homefacility. A rise in the importance of community-acquired MRSA infection will have significantimplications for the empiric antibiotic therapy ofCAP, and the evolution of this situation requiresclose surveillance.

Some centres, mostly in Spain, have reporteda relatively high rate (4-5%) of infection withPseudomonas aeruginosa in CAP associated witha high mortality - up to 100%. Significant co-morbidity, in particular underlying chronic lung dis-ease, may be the explanation, and an independentassociation of P. aeruginosa infection with mortalityhas not been firmly established. Similarly, infectionby Gram-negative enteric bacilli (GNEB), especiallythe Enterobacteriaceae, has been inconsistently asso-ciated with a higher mortality. Patients admittedfrom a nursing home facility are thought to be atgreater risk of acquiring GNEB-related CAP. Thisassociation has been mainly described in NorthAmerica and may not apply in other countrieswhere different healthcare systems exist and wherethe health of nursing home residents may differsubstantially.

The importance of geographical variation in thepattern of pathogens seen is exemplified by the highfrequency of Burkholderia pseudomallei (7-21%)compared to Strep, pneumoniae infection (4-5%) inCAP ICU patients reported in retrospective studiesfrom Singapore.17 This reflects the endemicity ofB. pseudomallei in the region that includes NorthernAustralia and Thailand. In South Africa, Klebsiellapneumoniae was the second commonest organismisolated after Strep, pneumoniae in a 10-year retro-spective survey (19% versus 29%).18

A microbial agent cannot be identified in approx-imately 45% of patients, despite intensive micro-biological investigations. In the past, attemptshave been made to use clinical, laboratory andradiological features to define the responsiblepathogen. There is now good evidence that

this approach is unreliable.19,20 As a result,empiric antibiotic therapy must therefore continueto be directed at the most probable range ofpathogens determined by local epidemiologicaldata.

MICROBIOLOGICAL INVESTIGATIONS

The value of diagnostic tests to identify microbialagents in CAP has been questioned, especially as achange in antibiotic therapy as a result does notappear to reduce mortality. The limitations to micro-biological investigations include the following.

• A lack of sensitivity and specificity of diagnostictests available. Prior antibiotic use, in particular,reduces the diagnostic rate. In hospital studies ofCAP, about 35% of patients report receivingantibiotics prior to hospital admission. Apartfrom patients transferred immediately to the ICUupon presentation to hospital, almost all patientswill have received at least one dose of antibiotic.

• Time delay in obtaining results. Results will notgenerally be quickly available (except Gram stain-ing of sputum), whereas antibiotic administrationneeds to be prompt. Serological tests that rely ona rise in antibody titres yield results only 2-3weeks later.

• The presence of mixed infections. The impor-tance of co-pathogens in regard to diagnostic test-ing and antibiotic choice is a difficult issue. Mixedinfections have been reported in up to 30% ofpatients hospitalized with CAP. The identificationof a single pathogen, therefore, does not rule outthe presence of a second. Most co-infectionsinvolve an atypical pathogen or virus with a bac-terial pathogen, usually Strep, pneumoniae.Whether specific treatment for the atypicalpathogen in such circumstances is necessary isunclear. Certainly, where C pneumoniae has beenfound together with Strep, pneumoniae, recoveryis reported despite antibiotic therapy directedonly against Strep, pneumoniae.21,23

On the other hand, the early identification ofa pathogen can be useful through confirmationof the diagnosis and choice of antibiotic regimen.A more focused antibiotic regimen favourablyaffects the development of adverse drug reactions,

Microbiological investigations 185

complications and antibiotic resistance. The intro-duction of rapid detection methods, such as anti-gen detection and polymerase chain reactions, istherefore welcomed. These diagnostic tools are lessaffected by prior antibiotic use (though notimmune to it) and, in theory, can offer a resultwithin hours of the appropriate sample beingtaken. Commercial kits for the detection in urineof Legionella and pneumococcal antigen are nowavailable and may prove to be valuable as part of adiagnostic package. It is our practice to performthe diagnostic tests detailed in Table 14.2 in allcases of CAP admitted to the ICU.

Percutaneous needle biopsy for the bacteriologicdiagnosis of pneumonia has regained interest recent-ly. A positive microbiological culture result is possi-ble in 33-80% of cases. The specificity is felt to begood as there is less contamination by oropharyngealflora, although this has not been evaluated carefully.The most common complication is pneumothorax(up to 30% of procedures), although most are smalland do not require drainage. Haemoptysis occurs in1-5%. Fatal air embolism is a rare but recognizedcomplication. Contraindications to the procedureinclude:

• poor pulmonary reserve,• bleeding diathesis,• lack of patient co-operation,• mechanical ventilation.

Table 14.2 Suggested microbiological investigations inpatients with severe community-acquired pneumonia

Gram stain and culture of sputum and respiratory samplesPneumococcal antigen tests on sputum, respiratory

samples and/or urineUrine for Legionella antigenSputum or other respiratory sample for Legionella culture

and direct immunofluorescence to Legionella sp.,Chlamydia sp., influenza A and B, parainfluenza 1-3,adenovirus, respiratory syncytial virus and Pneumocystiscarinii (if at risk)

Initial and follow up viral and atypical pathogen (includingLegionella ) serology with the initial sample being testedwithout waiting for the follow-up sample

Collection of lower respiratory tract samples by moreinvasive techniques, such as bronchoscopy, and deeptracheal sampling by catheter should be considered.Percutaneous fine-needle aspiration may be useful inpatients who are not on positive pressure ventilation37

The risk of pneumothorax is important in patientswho are receiving assisted positive pressure ventila-tion, and consequently the technique has limited usefor patients with severe CAP, except when performedby a very skilled and experienced operator. There-fore, we feel that, currently, percutaneous lung biopsyis not recommended as a routine diagnostic test inCAP. Its role in selected patients, such as the severelyill, remains to be evaluated.

Prognosis

Studies conducted to identify prognostic factors inpatients admitted to the ICU differ from other hos-pital studies of severity assessment in CAP, especiallyin the use of mortality as the outcome measure. Themost consistently independently associated prognos-tic factors are the presence of either septic shock orprogression of chest radiography changes during theICU stay. (Septic shock has been defined in variousways. A common definition is a >l-hour decrease insystolic blood pressure of >40 mmHg from baselineor a resultant systolic blood pressure of <90 mmHgafter adequate volume replacement.) Other risk fac-tors are detailed in Table 14.3.

A large, French, multi-centre study over a 10-yearperiod, and involving 472 patients, identified nineimportant variables from which a prediction modelwas constructed.4 Patients were initially stratifiedinto three groups, based on six risk factors assessableon ICU admission. Mortality risk estimates weresubsequently adjusted according to the developmentof three other prognostic factors (Table 14.4). Thismodel appears promising, but, in its current form, iscomplicated and remains to be tested in different set-tings. The initial stratification of patients into threerisk groups, if subsequently validated, may prove tobe a useful prognostic tool.

Table 14.3 Factors independently associated with death inthe ICU

Anticipated death within 4-5 yearsSimplified Acute Physiology Score (SAPS) >12Acute renal failureBacteraemiaInfection with Strep, pneumoniaeInfection with Gram-negative enteric bacilliNon-pneumonia-related complications


Table 14.4 A prognostic model for predicting ICU mortality

Age >40 years 1

Anticipated death <5 years(based on co-morbid index) 1

Non-aspiration pneumonia 1

Multilobar (>1) lobe involvement 1Acute respiratory failure

requiring MV <12 hours 1Septic shock >1 hour (decrease in

SBP of >40 mmHg from baselineor resultant SBP <90 mmHgafter volume replacement) 3

Sepsis-type complication (ARDS,shock, multi-organ failure) 4

Hospital-acquired lowerrespiratory tract infection 1

Non-specific complications (deep venous thrombosis,pulmonary embolism, gastrointestinal bleed) 2

0-23-56-8

I 4II 25III 60

012

49

1-21

3949

>2508693

Adapted from reference 4. ICU, intensive care unit; MV, mechanical ventilation; SBP, systolic blood pressure; ARDS, acute respiratory distress syndrome.

MANAGEMENT STRATEGIES

Antibiotics

The early institution of appropriate antibiotics hasbeen shown to improve outcome.24 The administra-tion of antibiotics should not be postponed. This isespecially pertinent in the early admission period,when multiple tasks may need to be performedsimultaneously. The choice of antibiotic is cruciallydetermined by the local epidemiological pattern oflikely pathogens. In the UK, where pneumococcaland Legionella infection predominate, a combinationof a b-lactamase-stable antibiotic and a macrolide± rifampicin is recommended as empirical therapy(Table 14.5). A rise in liver enzymes can be expectedwith rifampicin given at a dose of 600 mg intra-venously twice daily. This is usually harmless andresolves once the drug is stopped. The authors' prac-tice is to reduce the intravenous dose after 3 days to600 mg once daily and to stop the drug after a totalof 7 days if the patient is responding adequately.Whether the addition of rifampicin to a macrolides

is more effective in proven Legionella infection isuntested. Similarly, although the quinolones havebeen found to be more active than macrolides in ani-mal studies, there are no controlled trials demon-strating their superiority. As there is extensive clinicalexperience with the macrolides in the treatment ofLegionella infection, and no evidence of emergingresistance, it seems reasonable to continue usingthem as first-line therapy, with the quinolones beingemployed as second-line agents in the event of drugintolerance, allergy or failure.

Flucloxacillin should be added when Staph. aureusinfection is suspected. It is worth noting that thethird-generation cephalosporins are less active thanthe first-generation and second-generation againstStaph. aureus. For MRSA infection, vancomycinremains the antibiotic of choice. Extra coverage forGram-negative enteric bacilli (e.g. with an amino-glycoside) and pseudomonal infection (e.g. with cef-tazidime) is not routinely required in the UK in viewof the low frequency (2-3%) of these pathogens.Where these infections are more common, differentrecommendations may apply. The EuropeanRespiratory Society and Italian guidelines for the

Management strategies 187

Table 14.5 Recommended empirical antibiotics for severecommunity-acquired pneumonia of unknown cause

PreferredAmoxicillin/davulanate 1.2 g t.d.s. i.v.orCefuroxime 1.5 g t.d.s. i.v.orCefotaxime 1 g t.d.s. i.v.orCeftriaxone 2 go.d. i.v.plusErythromycin 500 mgq.d.s. i.v.orClarithromycin 500 mg b.d. i.v.with or withoutRifampicin 600 mg o.d. or b.d. i.v.

AlternativeFor those intolerant of preferred regimen, or where there are

local concerns over Clostridium difficile-associated diar-rhoea related to p-lactam use Levofloxacin 500 mg b.d. i.v.

plusBenzylpenicillin 1-2 g 6-hourly i.v.

management of CAP do not recommend the use ofanti-pseudomonal agents, whereas in theNetherlands, they are recommended for patientswith severe chronic obstructive airways disease orstructural lung disease.25-27 Anaerobic infection isuncommon. The need for adding anaerobic covershould therefore be assessed individually and takeinto account known risk factors such as poor denti-tion and the likelihood of aspiration. The InfectiousDiseases Society of America has published recom-mendations for the empirical antibiotic treatmentof severe CAP.28

Non-antibiotic therapy

NON-INVASIVE VENTILATION

Failure to maintain adequate oxygenation (PaO2 >8kPa or SaO2 >90%), despite high-flow O2 supple-mentation, may indicate the need for assisted venti-lation. Continuous positive airways pressure(CPAP), by recruiting under-ventilated alveoli andallowing an increase in fractional FiO2, may correcthypoxaemia and avoid the need for intubation (seeChapter 5). It has been shown to be effective in

hypoxia resulting from CAP as well as in viral andPneumocystis carinii pneumonia, although there areno randomised, controlled trials. Non-invasive venti-lation (see Chapter 5) may be employed when CO2

retention develops. Confalonieri et al. have reporteda randomized, controlled trial of 56 patients withCAP in which the intubation rate was significantlyreduced by non-invasive ventilation.29 A rapid andsustained reduction in respiratory rate was evident inpatients treated with non-invasive ventilation andthe need for mechanical ventilation was reduced byhalf (21% in patients treated with non-invasive ven-tilation compared to 50% in controls). The intensityof nursing care workload, duration of hospitalizationand hospital mortality remained unchanged.Patients treated with non-invasive ventilation needto be closely monitored, particularly for evidence ofretention of respiratory secretions. Review of clinicalresponse is important and criteria for endotrachealintubation should be unambiguous. In all instances,the availability of non-invasive ventilation shouldnot result in inappropriate delay in the initiation ofmechanical ventilation when necessary. Centresexperienced in non-invasive ventilation are likely toreport the best results.

SPECIAL VENTILATORY APPROACHES

In pneumonia, hypoxia is mediated primarily byperfusion of units of low or no ventilation. Hypoxicpulmonary vasoconstriction, which attempts toimprove VQ matching through the releaseof prostaglandins, may be abnormal in pneumoniaand be partly responsible.

PositioningPositioning the patient with unilateral consolidationwith the involved side up may increase perfusion tothe dependent, uninvolved lung by gravity.Ventilation perfusion matching is thereby improved,with improvement in hypoxaemia.30

Differential lung ventilationVentilatory requirements may differ substantiallybetween the two lungs in patients with unilateralinvolvement. By using a double-lumen tube, ventila-tion to each lung can, theoretically, be individuallyoptimized. For instance, higher levels of positiveend-expiratory pressure (PEEP) may be applied to


the less compliant pneumonic lung. In practice,double ventilation is very difficult and rarely effective.

Extra-corporeal membrane oxygenationExtra-corporeal membrane oxygenation (ECMO)has been used in adults with respiratory failure dueto CAP who have failed to maintain adequate levelsof oxygenation despite aggressive ventilation tech-niques. However, as yet, there are only isolated casereports and better evidence is required before suchexperimental approaches can be justified.

Pharmacological manipulation

Anti-inflammatory Drugs Theoretically, cyclo-oxy-genase inhibitors, such as aspirin and indomethacin,should reduce prostaglandin-induced loss of hypoxicvasoconstriction. However, in exploratory studies,although improvement in intrapulmonary shuntshas been demonstrated, commensurate improve-ments in oxygenation did not occur.31

Prostacydin and Nitric Oxide Local selective vasodi-latation in better ventilated lung areas by aerosolizedprostacyclin or inhaled nitric oxide decreases intra-pulmonary shunt. An improvement in oxygenation istherefore possible. Although studied most in adultrespiratory distress syndrome (ARDS), these drugsmay be helpful in patients with pneumonia.32

ADJUVANT AGENTSSteroidsIn the past, steroids have been used in septic shockcomplicating infection, but without success. A pre-liminary study of intravenous hydrocortisone inCAP suggests that inflammatory markers, includingtumour necrosis factor-a (TNF-a), interleukin (IL)-1b, IL-6 and CRP, are reduced.33 Further work todelineate clinical benefit is awaited.

Granulocyte colony-stimulating factorGranulocyte colony-stimulating factor (G-CSF)increases the production and function of neu-trophils. It is widely used in treating neutropenia inpatients receiving chemotherapy. There is evidencethat in non-neutropenic patients with CAP, itreduces complications (empyema, ARDS, dissemi-nated intravascular coagulation) and/or length ofhospitalization. A phase I trial has shown that, givensubcutaneously for 10 days (at doses of 75-650 |xg

day !) in combination with appropriate antimicro-bial therapy, it is safe. There has, as yet, been noevidence of increased tissue damage as a resultof enhanced neutrophil activation.34

PATIENT NOT IMPROVING

Patients who do not respond to initial treatmenthave a higher mortality. Early re-assessment andappropriate intervention are therefore required.Possibilities to consider include:

• incorrect or additional unrecognized diagnosis, e.g.

1. pulmonary embolism2. foreign-body inhalation3. proximal obstructing endobronchial tumour4. bronchiolitis obliterans organizing pneumonia5. hospital-acquired infection

• a pathogen resistant to initial antimicrobialtherapy1. natural resistance, e.g.

(a) fungus: aspergillus, coccidiomycosis(b) Mycobacteria: M. tuberculosis(c) virus: chicken pox(d) parasite: Pneumocystis carinii(e) bacteria: P. aeruginosa

2. acquired resistance, e.g.(a) penicillin-resistant Strep, pneumoniae(b) methicillin-resistant Staph. aureus

• drug non-compliance, hypersensitivity or drug-related complications, e.g.

1. inadequate dose2. Clostridium difficile-associated diarrhoea3. phlebitis at intravenous cannula site

• defective host immune response, e.g.

1. undiagnosed immunocompromised state, e.g.human immunodeficiency virus (HIV) infec-tion, uncontrolled diabetes mellitus

2. underlying cancer3. cystic fibrosis

• complication of infection, e.g.

1. lung abscess2. empyema3. metastatic infection4. ARDS

Summary 189

Para-pneumonic effusions are a common com-plication, occurring in 20-60% of patients withCAP. The majority are small and resolve in stepwith the main site of pulmonary infection withantibiotics. However, 5-10% of para-pneumoniceffusions progress to become empyemas unlessadequately drained. In deciding to drain a para-pneumonic effusion, clinical factors, radiologicalfindings and pleural fluid characteristics need to betaken into consideration. Persistent fever, featuressuggestive of anaerobic infection (aspiration, alco-holism), infection with Staph. aureus, Staph. pyo-genes or K. pneumoniae and larger effusions are themost important features that indicate the need forfurther evaluation. Ultrasound and computerizedtomography (CT) scanning are both useful investi-gations (see Chapter 7). Whenever possible, theeffusion should be tapped and pleural fluid sent fordetermination of pH, Gram stain and culture.(Pleural fluid pH is tested anaerobically using aheparinized syringe through a standard blood-gasmachine.) The presence of frank pus or the growthof micro-organisms is diagnostic of an empyema,and a pH of <7.2 is highly suggestive.35 In thesecircumstances, and when loculations are demon-strated radiologically by ultrasound or CT, earlydrainage is indicated. A small to moderate-sizedcatheter (8-16 French gauge) is adequate in mostinstances, but requires irrigation with saline one tofour times per day to maintain patency. The instil-lation of intrapleural thrombolytics is useful inpromoting drainage and preventing the develop-ment of loculations within the pleural cavity.36

Streptokinase and urokinase have both been tested.Different doses and methods of instillation havebeen reported. One 'protocol' is detailed in Table14.6. Where frank pus is present, the insertion of alarge-bore chest drain (size 22-34 French gauge) isthe traditional approach and early consultationwith a thoracic surgical colleague is recommended.Pleural decortication, or open drainage with ribresection, is sometimes still required in difficultcases.

Complete evaluation of the patient who does notappear to be responding to treatment begins with areappraisal of the history of illness, with collaborat-ing information obtained from relatives, friends orwitnesses. It may also involve repeating laboratoryinvestigations, including blood cultures, assessingHIV status and performing CT scans, ultrasound

Table 14.6 Protocol for instillation of intrapleural Strep-tokinase

1. Position a size 12-14 French catheter in the mostdependent portion of the pleural collection

2. Connect to an underwater seal and keep on continuoussuction ( - 20 cmH20)

3. Flush catheter with 20 ml saline every 6 hours4. Instill 250 000 IU Streptokinase in 20 ml saline in

place of one of the saline flushes 12 hourly5. Close off the catheter for 2 hours after instilling

Streptokinase; return to suction thereafterDoses of Streptokinase may be given for 3-5 days,

depending on the clinical situation and response

assessment of pleural and abdominal cavities andechocardiography. Bronchoscopy is indicated toexclude post-obstructive pneumonia or lung abscess.It can be performed on patients supported by non-invasive ventilation see Chapter 5). Lung biopsy isonly necessary in a tiny proportion of patients inwhom the diagnosis remains unclear despite lessinvasive investigations. The differential diagnosisincludes:

• bronchiolitis obliterans organizing pneumonia,• malignancy, e.g. primary lymphoma, broncho-

alveolar cell carcinoma,• hypersensitivity pneumonitis,• eosinophilic pneumonitis,• drug-induced pneumonitis,• pulmonary vasculitis, e.g. Wegener's granulo-

matosis,• granulomatous disorders, e.g. sarcoidosis, beryl-

liosis,• pulmonary alveolar proteinosis.

SUMMARY

CAP is common and is severe in some patients.Although there are no randomised, controlled trials,for some patients, prompt and appropriate manage-ment in an ICU probably reduces mortality.Admission criteria and the likely pathogens encoun-tered are strongly influenced by geography, health-care resources and population characteristics. Thereview offered here may therefore need tailoring tothe different situations encountered in day-to-dayclinical practice.


REFERENCES

1. Hirani, NA, Macfarlane, JT. Impact of management

guidelines on the outcome of severe community

acquired pneumonia. Thorax 1997; 52: 17-21.

2. Niederman, MS, Mandell, LA, Anzueto, A, et al.

Guidelines for the management of adults with CAP.

ABRCCM 2001; 163: 1730-54.

3. Ewig, S, Ruiz, M, Mensa, J, et al. Severe community-

acquired pneumonia. Assessment of severity criteria.

Am J Respir Crit Care Med 1998; 158:1102-8.

4. The British Thoracic Society, BTS guidelines for the

management of CAP. Thorax 2001; 56: Supplement IV.

5. Anonymous. Community-acquired pneumonia in

adults in British hospitals in 1982-1983: a survey

of aetiology, mortality, prognostic factors and

outcome. The British Thoracic Society and the

Public Health Laboratory Service. Q J Med 1987;

62: 195-220.

6. Lim, WS, Macfarlane, Jl, Boswell, TC, et al. SCAPA:

Study of CAP aetiology in adults admitted to hospital:

implications for management guidelines. Thorax

2001; 56: 296-301.

7. Neill, AM, Martin, IR, Weir, R, et al. Community acquired

pneumonia: aetiology and usefulness of severity

criteria on admission. Thorax'\996; 51:1010-16.

8. Riquelme, R, Torres, A, El-Ebiary, M, et al. Community-

acquired pneumonia in the elderly: a multivariate

analysis of risk and prognostic factors. Am J Respir Crit

Care M«/1996;"l 54: 1450-5.

9. Janssens, JP, Gauthey, L, Herrmann, F, Tkatch, L,

Michel, JP. Community-acquired pneumonia in older

patients. J Am Geriatr Soc 1996; 44: 539-44.

10. Venkatesan, P, Gladman, J, Macfarlane, JT, et al. A

hospital study of community acquired pneumonia

in the elderly. Thorax 1990; 45: 254-8.

11. Leroy, 0, Bosquet, C, Vandenbussche, C, et al.

Community-acquired pneumonia in the intensive care

unit: epidemiological and prognosis data in older

people. J Am Geriatr Soc 1999; 47: 539-46.

12. Rello, J, Rodriguez, R, Jubert, P, Alvarez, B. Severe

community-acquired pneumonia in the elderly:

epidemiology and prognosis. Clin Infect Dis 1996;

23: 723-8.

13. Ruiz, M, Ewig, S, Torres, A, et al. Severe community-

acquired pneumonia. Risk factors and follow-up

epidemiology. Am J Respir Crit Care Med 1999; 160:

923-9.

14. Woodhead, MA, Macfarlane, JT, Rodgers, FG, Laverick,

A, Pilkington, R, Macrae, AD. Aetiology and outcome

of severe community-acquired pneumonia. J Infect

1985; 10:204-10.

15. Woodhead, MA, Macfarlane, JT. Legionnaires'disease:

a review of 79 community acquired cases in

Nottingham. Thorax 1986; 41: 635-40.

16. Shah, A, Check, F, Baskin, S, Reyman, T, Menard, R.

Legionnaires' disease and acute renal failure: case

report and review. Clin Infect Dis 1992; 14: 204-7.

17. Tan, YK, Khoo, KL, Chin, SP, Ong, YY. Aetiology and

outcome of severe community-acquired pneumonia

in Singapore. Eur Respir J 1998; 12: 113-15.

18. Feldman, C, Ross, S, Mahomed, AG, Omar, J, Smith, C.

The aetiology of severe community-acquired

pneumonia and its impact on initial, empiric,

antimicrobial chemotherapy. Respir Med 1995; 89:

187-92.

19. Farr, BM, Kaiser, DL, Harrison, BD, Connolly, CK.

Prediction of microbial aetiology at admission to

hospital for pneumonia from the presenting clinical

features. British Thoracic Society Pneumonia

Research Subcommittee. Thorax 1989; 44: 1031-5.

20. Fang, GD, Fine, M, Orloff, J, et al. New and emerging

etiologies for community-acquired pneumonia with

implications for therapy. A prospective multicenter

study of 359 cases. Medicine 1990; 69: 307-16.

21. Torres, A, El Ebiary, M. Relevance of Chlamydia

pneumoniae in community-acquired respiratory

infections [editorial]. Eur Respir J 1993; 6: 7-8.

22. Kauppinen, MT, Saikku, P, Kujala, P, Herva, E, Syrjala,

H. Clinical picture of community-acquired Chlamydia

pneumoniae pneumonia requiring hospital treatment:

a comparison between chlamydial and pneumococcal

pneumonia [see comments]. Thorax 1996; 51: 185-9.

23. Lieberman, D, Schlaeffer, F, Boldur, I, et al. Multiple

pathogens in adult patients admitted with community-

acquired pneumonia: a one year prospective study of

346 consecutive patients. Thorax 1996; 51: 179-84.

24. Meehan, TP, Fine, MJ, Krumholz, HM, et al. Quality of

care, process, and outcomes in elderly patients with

pneumonia. JAMA 1997; 278: 2080-4.

25. Vegelin, AL, Bissumbhar, P, Joore, JC, Lammers, JW,

Hoepelman, IM. Guidelines for severe community-

acquired pneumonia in the western world. Neth J Med

1999; 55: 110-17.

26. Gialdroni, GG, Bianchi, L. Guidelines for the

management of community-acquired pneumonia

in adults. Italian Society of Pneumology. Italian

References 191

Society of Respiratory Medicine. Italian Society ofChemotherapy. Monaldi Arch Chest Dis 1995; 50: 21-7.

27. ERS Task Force Report. Guidelines for management ofadult community-acquired lower respiratory tractinfections. European Respiratory Society. Ear Respir J1998; 11: 986-91.

28. Bartlett, JG, Breiman, RF, Mandell, LA, File, TM Jr.Community-acquired pneumonia in adults: guidelinesfor management. The Infectious Diseases Society ofAmerica. Clin Infect Dis 1998; 26: 811-38.

29. Confalonieri, M, Potena, A, Carbone, G, et al. Acuterespiratory failure in patients with severe community-acquired pneumonia. A prospective randomizedevaluation of noninvasive ventilation. Am J RespirCrit Care Med 1999; 160: 1585-91.

30. Dreyfuss, D, Djedaini, K, Lanore, JJ, Mier, L,Froidevaux, R, Coste, F. A comparative study of theeffects of almitrine bismesylate and lateral positionduring unilateral bacterial pneumonia with severehypoxemia. Am Rev Respir Dis 1992; 146: 295-9.

31. Ferrer, M, Torres, A, Baer, R, Hernandez, C, Roca, J,Rodriguez-Roisin, R. Effect of acetylsalicylic acid onpulmonary gas exchange in patients with severepneumonia: a pilot study. Chest 1997; 111: 1094-100.

32. Walmrath, D, Schneider, T, Pilch, j, Schermuly, R,Grimminger, F, Seeger, W. Effects of aerosolized

prostacyclin in severe pneumonia: impact of fibrosis.AmJ Respir Crit Care Med 1995; 151: 724-30.

33. Monton, C, Ewig, S, Torres, A, el al. Role ofglucocorticoids on inflammatory response innonimmunosuppressed patients with pneumonia:a pilot study. Eur Respir J 1999; 14: 218-20.

34. DeBoisblanc, BP, Mason, CM, Andresen, J, el al.Phase 1 safety trial of Filgrastim (r-metHuG-CSF) innon-neutropenic patients with severe community-acquired pneumonia. Respir Med 1997; 91:387-94.

35. Heffner, JE, Brown, LK, Barbieri, C, DeLeo, JM. Pleuralfluid chemical analysis in parapneumonic effusions. Ameta-analysis. [Published erratum appears in Am JRespir Crit Care Med 1995; 152(2): 823.] Am J RespirCrit Care Med 1995; 151:1700-8.

36. Davies, RJ, Traill, ZC, Gleeson, FV. Randomisedcontrolled trial of intrapleural streptokinase incommunity acquired pleural infection. Thorax 1997',52:416-21.

37. Ruiz-Gonzalez, A, Falguera, M, Nogues, A,RubioCaballero, M. Is Streptococaispneumoniaelheleading cause of pneumonia of unknown etiology?A microbiologic study of lung aspirates in consecutivepatients with community-acquired pneumonia.AmJ Med 1999; 106: 385-90.

15Nosocomial pneumonia

JEAN-LOUIS VINCENT, BAUDOUIN BYLAND DALIANA PERES BOTA

Introduction

Definitions and epidemiology

Pathogenesis

Diagnosis

Risk factors

192192193193194

Prevention

Treatment

Conclusion

References

194196198198

INTRODUCTION

Derived from the Latin word for hospital (noso-comium), a nosocomial infection is an infectionacquired in hospital as opposed to in the community.The critically ill patient is particularly at risk of devel-oping nosocomial infection. In a large, single-day pointprevalence study of nosocomial infection in intensivecare units (ICUs) across Europe (the EPIC study), 45%of the 10038 patients occupying an ICU bed had anosocomial infection.1 The most common type ofinfection was nosocomial pneumonia, reported in 47%of infected patients (Fig. 15.1). Nosocomial pneumo-nia is therefore common on the ICU and is not onlyassociated with increased morbidity and mortality, butalso with prolonged ICU stay and costs.2 Awareness ofthe associated risk factors facilitates early diagnosis andtreatment, and specific preventive strategies may helplimit the development of this complication. Thischapter discusses basic definitions and epidemiology,diagnosis, potential risk factors, prevention and therapy.

DEFINITIONS AND EPIDEMIOLOGY

ICU nosocomial pneumonia may be broadly definedas a pneumonia that develops while on the ICU but

that was not present on ICU admission. However,within this broad definition, varying time intervals andother criteria have been specified by different authors.Most agree that ICU-acquired pneumonia refers to anyinfection occurring 48 hours or more after ICU admis-sion. Many investigators also separately classify ventila-tor-acquired pneumonia (VAP) as an infection thatoccurs more than 24 or 48 hours after the patient hasbeen intubated and mechanically ventilated. The quot-ed incidence of nosocomial pneumonia on the ICUranges from 9% to 47%, with the variation relating tothe ICU case-mix and differences in the definitions anddiagnostic techniques used.3-6 Specific groups of ICUpatients such as trauma patients have a higher inci-dence of nosocomial infection.3,7

Mortality associated with the development of noso-comial pneumonia varies considerably, with somestudies reporting a crude mortality ranging from 40%to 70%,4;6>8>9, while other studies have attempted toidentify the increased mortality specifically attribut-able to its development (27-45%).10,11 Nosocomialpneumonia may itself cause increased mortality, butit could also be that it is the sicker patients withinherently higher mortality who develop nosocomialpneumonia and the pneumonia itself is only a markerof the higher mortality seen in such patients.

The most commonly identified organisms arePseudomonas aeruginosa and Staphylococcus aureus,

Diagnosis 193

Figure 15.1 Incidence of ICU-acquired infection in the EPIC

study1 by site of origin, showing pneumonia as the most common

nosocomial infection.

with other causative organisms including Klebsiella,Enterobacter, Escherichia coli and Acinetobacterspecies.5 The pattern of causative organisms variesamong units according to patient case-mix,antibiotic protocols, infection control practices,etc.12 Mortality seems to be particularly high inpatients infected with Pseudomonas or Acinetobacterspecies 10

PATHOGENESIS

For pneumonia to develop, one or more of thefollowing conditions must exist: impaired hostdefences; the presence of a sufficient number oforganisms in the patient's lower respiratory tract toovercome host defence; or the presence of a partic-ularly virulent organism. Host defences are fre-quently already reduced in the critically illpopulation, due to concurrent disease processes.Immunosuppression in the critically ill is primarilydue to the release of interleukin-10 (IL-10) andother anti-inflammatory mediators, such as IL-1receptor antagonist (IL-IRa) and tumour necrosisfactor (TNF) receptors, and produces a state of'immunoparalysis'. Patients who have been treatedwith immunosuppressive drugs represent a particu-larly high-risk group.13

The inoculation of infectious organisms into thelower airways can occur in various ways, but(micro)aspiration of colonized oropharyngealsecretions is probably the most important. Whereasmicro-aspiration of oropharyngeal secretions is acommon event in all individuals, the critically ill,

particularly those receiving mechanical ventilation,have a higher incidence of upper airway coloniza-tion with potentially pathogenic organisms andfigures as high as 80% have been reported.14'15

Aspiration of gastric contents is more likely tooccur in patients with a reduced level of conscious-ness, during insertion of nasogastric or endotra-cheal tubes and if there is co-existing oesophagealdisease. Alternative routes of entry for infectiousorganisms to the lower airways include aerosolspread, perhaps more commonly associated withviruses, Legionella species and Mycobacteriumtuberculosis, and haematogenous spread, morecommon in postoperative patients and in thosewith long-term intravenous or urinary catheters.Tracheal intubation impairs certain host defencemechanisms, such as cough and mucociliary clear-ance, putting such patients at greater risk of devel-oping pneumonia. Secretions may collect aroundthe endotracheal cuff and these are not easilyremoved by suctioning and are prone to coloniza-tion. Changes in airway size during swallowing orbreathing can allow leakage of these secretions intothe lower airways. In addition, respiratory equip-ment can harbour pathogenic organisms, permit-ting direct inoculation of the lower airways.

DIAGNOSIS

The diagnosis of nosocomial pneumonia in the ICUmay be difficult because the typical features of infec-tion, such as fever, tachycardia, tachypnoea andraised white blood cell counts, are non-specific indi-cators of an inflammatory response. Patchy alveolarinfiltrates on chest X-ray may represent pneumonia,but could also be due to pulmonary oedema, infarc-tion or atelectasis. Non-pathogenic colonization ofthe upper airways is common in the ICU patient,rendering positive cultures from tracheal aspiratesrelatively insensitive. Other techniques, includingculture of samples obtained by bronchoalveolarlavage (BAL) or by protected specimen brush (PSB),have been suggested to improve diagnostic sensitiv-ity. Although the value of these invasive methods isstill debated, a recent randomized trial suggestedthat, in patients managed according to results fromspecimens obtained by BAL or PSB, survivalimproved, antibiotic use was reduced and organdysfunction resolved earlier than in patients who

194 Nosocomial pneumonia

were managed using non-invasive clinical assess-ment and microscopic evaluation of endotrachealaspirates.16

Bronchoalveolar lavage

BAL specimens are generally obtained by wedgingthe tip of a bronchoscope in the medium-sizedbronchus relevant to the area of alveolar infiltrateidentified on the chest X-ray. The lung segment islavaged with 20-30 mL of sterile isotonic saline and,after 5-10 s, a sample of 5-10 mL is obtained usinggentle suction. A culture of >104 colony-formingunits (cfu) mL-1 is generally considered to beindicative of pneumonia.12,17 Non-bronchoscopicBAL (nBAL) specimens may also be obtained using aflexible catheter inserted blindly into the airways, butthe source of the secretions obtained will beunknown and diluted by bronchial secretions. Thismay lead to a false-positive result in a heavily colon-ized patient or a patient with bronchitis and also toproblems related to the threshold value used to diag-nose significant infection.

Protected specimen brush

The PSB is a double-lumen brush system that avoidsupper airway contamination of the sample. Thebrush is introduced, either blindly or via fibreopticbronchoscopy, into the bronchus of a lung segment,with infiltrates seen on chest X-ray. The inner can-nula is then advanced further and a specimenobtained. Generally, a culture of >103 cfu mL - l isconsidered diagnostic of pneumonia. Problems withPSB include the fact that only a small lung segmentis sampled, potentially leading to false-negativeresults, and, despite the double-lumen system, con-tamination can still occur. In addition, in patientsalready receiving antibiotics, PSB sampling has a verylow sensitivity.17

Direct comparison of BAL with PSB has producedconflicting results, with no general agreement onwhich has the greater sensitivity or specificity. Arecent meta-analysis concluded that both wereequally accurate in diagnosing pneumonia,although, in patients already on antibiotics at thetime of sampling, BAL is more sensitive.17

BAL and PSB techniques are dependent, to adegree, on operator skills and training. Blind non-bronchoscopic sampling may help reduce costs andhas been evaluated for both techniques, yieldingresults similar to those of bronchoscopic sam-pling.18,19 Early treatment is important in theoutcome from nosocomial pneumonia and concernshave been raised about delays in initiating treatmentwhile culture results from these techniques are awaited.Empiric therapy, based on likely pathogens and localresistance patterns, should be started as soon as thediagnosis is suspected, and adapted as results fromsuch tests become available.

RISK FACTORS

Many risk factors have been reported as being asso-ciated with the development of nosocomialneumonia in the critically ill population (Table 15.1).Certain patient groups are at greater risk due to thenature of current or pre-existing disease processes,such as those with trauma, chronic respiratory orcardiac disease, and those who are in a coma. Patientswith acute respiratory distress syndrome (ARDS)carry a twofold increase in risk compared to othercritically ill patients.20 Recent studies have shownthat there is a reduced risk of nosocomial pneumo-nia in patients nursed in the semi-recumbent ratherthan the supine position.21 The use of antibiotics ismore controversial, with some studies suggesting anincreased incidence of pneumonia in patients previ-ously treated with antibiotics and others showing areduced risk. However, prolonged and inappropri-ate antibiotic therapy may be expected to favour col-onization, and hence infection, with resistantorganisms.22

PREVENTION

With the high associated morbidity and mortality,prevention of nosocomial pneumonia must be seenas an important part of routine ICU patient care.Based on knowledge of the patient groups at risk andother specific aetiological factors, various preventivestrategies have been proposed (Table 15.2).

Prevention 195

Table 15.1 Risk factors for the development of nosocomialpneumonia

Patient-related factorsTrauma, burnsPre-existing disease: chronic obstructive airways disease,

diabetes mellitus, central nervous system disease, liverfailure, alcoholism

MalnutritionSeverity of illnessOlder ageSmoking habits

Intervention-related factorsDuration of mechanical ventilationPresence of indwelling catheters: urinary catheters,

central venous lines, arterial linesUse of nasogastric tubesProlonged surgeryExcessive use of sedative agentsInappropriate and prolonged use of antibioticsUse of antacids/H2 antagonistsParenteral nutritionNasal intubationSupine positioning

Environment-related factorsInadequate infection-control proceduresProlonged ICU stayContaminated respiratory equipment

Adequate infection control andequipment management

Thorough hand washing is the most effective meansof limiting the spread of infection, but it is frequentlyforgotten or inadequate, involving only a cursoryholding of the hands under the tap. Some studieshave shown that less than 50% of hospital personnelcomply with hand-washing protocols,23,24 andpatients' relatives are often better at washing theirhands than ICU staff. The use of hand disinfectionrather than washing more efficiently reduces thehand carriage of potentially pathogenic organismsand achieves better compliance, probably because itis both quicker and more convenient.24 The use ofdisposable gloves and gowns may also help limit thetransmission of bacteria between staff and patients,but the evidence for this is less compelling.25,26

Transport of patients outside the ICU, for whateverreason, is associated with an increased risk of develop-

Table 15.2 Suggested strategies to prevent nosocomialpneumonia

Adequate infection control policies: hand washing,equipment sterilization, etc.

Semi-recumbent positioningSubglottic drainageSelective digestive decontaminationMaintenance of low gastric pH by avoidance of antacids

and H2 antagonistsOral intubation versus nasal intubationJejunal versus nasogastric tubeEarly enteral feeding (particularly with immune

supplemented feeds)Avoidance of excessive sedation

ing nosocomial pneumonia and should therefore berestricted as far as possible.27

Increasing the frequency of ventilator circuitchanges does not appear to be beneficial, althoughthe use of heat and moisture exchangers (HMEs)may decrease the incidence of nosocomial pneu-monia, perhaps by minimizing the amount of con-densate in the circuit. Further research is necessary inthis area because existing information is scanty andconflicting and some HMEs may be associated withdifficulty in weaning.28,29 Oral intubation is prefer-able to nasal intubation and is associated with lowerrates of nosocomial sinusitis. Non-invasive ventila-tion has been associated with reduced rates of infec-tion and should be considered in appropriatepatients.30

Patient positioning

A recent randomised, controlled trial evaluated theassociation of nosocomial pneumonia with patientpositioning in mechanically ventilated patients.21

The study was interrupted after a preliminary analy-sis showed a significant reduction in nosocomialpneumonia in those patients nursed in the semi-recumbent compared to the supine position; this dif-ference was particularly significant in patientsreceiving enteral nutrition and in those with areduced Glasgow Coma Score. Other studies haveshown similar results, the benefit presumably beingdue to a reduced incidence of aspiration in this pos-ition. Kinetic beds may also be associated with a


reduced incidence of nosocomial respiratory infec-tion,31 but this therapy is expensive.

Subglottal drainage of secretions

The aspiration of secretions that pool above theendotracheal cuff in mechanically ventilated patientsmaybe another factor implicated in the developmentof nosocomial pneumonia. Special endotrachealtubes are now available with a separate lumen thatallows continuous subglottal suctioning of thesesecretions, but it is not yet clear whether the benefitjustifies the additional cost.

Selective digestive decontamination(SPP)

Selective decontamination of the digestive tract(SDD) is designed to prevent infection by eradicat-ing and preventing the carriage of potentially patho-genic aerobic micro-organisms from theoropharynx,stomach and gut. SDD consists of non-absorbableantibiotics (usually polymyxin, tobramycin andamphotericin B) applied topically to the oropharynxand through a nasogastric tube, plus the use of asystemic antibiotic, most commonly cefotaxime.Several studies, although not all, have shown thistechnique to reduce the incidence of nosocomialpneumonia, particularly in trauma patients, andmeta-analyses have confirmed that SDD using acombination of topical and systemic antibiotics canreduce respiratory infection32 and may have a bene-ficial effect on mortality.33 However, SDD is not rou-tinely used in most ICUs because concerns remainregarding the cost and the risk of increasing bacteri-al resistance and drug toxicity with this approach.

alternative agent that may be less likely to increasethe risk of pneumonia, but it appears to be less effec-tive than H2-blockers in preventing gastrointestinalhaemorrhage.36 On the available evidence, H2-antagonists have superior anti-ulcer activity andminimal, if any, excess risk for pneumonia.

Nutritional support

Malnutrition is a risk factor for nosocomial pneu-monia, and achieving adequate nutritional supportis an important preventative strategy. However, whenadministered via a nasogastric tube, enteral nutritionhas itself been associated with increased infectionrates, possibly because it raises gastric pH, therebyencouraging bacterial colonization. The use of jeju-nal feeding tubes has been shown both to reduce theincidence of pneumonia and to improve nutritionalstatus. The use of various immune-enhanced feedsincluding conditionally essential amino acids such asglutamine and arginine, nucleotides and omega-3fatty acids has been associated with fewer acquiredinfections than standard feeding solutions, but fur-ther studies are needed to define the most beneficialcombination of nutrients, to identify the mostappropriate target populations and to justify theincreased costs.

Avoiding excessive sedation

A reduced conscious level is associated with anincreased risk of respiratory complications, andsedative agents should therefore be titrated to theminimal level required to keep each individualpatient comfortable and co-operative. Sedationscores may be useful in assessing and adjustingsedation.37

Maintenance of low gastric pH

Antacids and H2-blockers are often used in the ICUpopulation to prevent the development of stressulcers and gastrointestinal bleeding. However, theseagents can increase gastric pH and encourage col-onization, and several publications have suggestedan increased incidence of pneumonia with thesetreatments.34,35 Sucralfate has been proposed as an

TREATMENT

Antibiotic therapy

The treatment of nosocomial pneumonia representsa considerable challenge due to the range of organ-isms encountered, the frequent polymicrobialnature of these infections and the high incidence of

Treatment 197

resistant organisms. Maintaining a high index ofsuspicion and the prompt initiation of therapy areimportant aspects of management. Appropriateantibiotic selection is essential (Fig. 15.2) becauseinadequate antibiotic therapy is associated withincreased mortality rates.38 Guidelines for empirictherapy have been drawn up by several groups, anda consensus statement by the American ThoracicSociety suggests that treatment should be based onthe subdivision of patients into groups according tothe severity of their disease and the presence of asso-ciated risk factors.12

• Group 1 patients are those with mild to moderateinfection and without significant risk factors, orwith severe but early-onset infection. Mono-therapy is considered adequate.

• Group 2 patients are those with both mild tomoderate infection and risk factors. Thesepatients are more likely to have infection withmore virulent pathogens, including Pseudomonas

aeruginosa, and need additional coverage.Monotherapy may not be adequate.

• Group 3 patients are those with severe infectionand risk factors. These patients are at greatest riskof infection with highly virulent pathogens andmany, especially those with late-onset pneu-monia, will require combination antibiotic therapy.

Figure 15.2 shows suggested antibiotic regimensfor patients with nosocomial pneumonia, but theprecise choice of antibiotic will depend on variousfactors (Table 15.3).

Importantly, once an organism has been isolatedand antibiotic sensitivity determined, the antibioticprescription should be modified accordingly to pre-vent unnecessary treatment, which increases boththe costs and the risk of inducing bacterial resistance.The duration of therapy needs to be assessed on anindividual patient basis according to clinicalresponse, but antibiotics are often given for at least 7days. In cases of Gram-negative pneumonia associated

Figure 15.2 Suggested flowchart to

direct antibiotic therapy in patients

with nosocomial pneumonia.


Table 15.3 Factors influencing the choice of antibiotic inthe therapy of nosocomial pneumonia

Bacteriological informationEpidemiolgic data (literature plus local)Length of hospitalizationPrevious antibiotic therapySeverity of infection - presence of septic shockDegree of immunosuppressionSide effects of antibioticsCosts

sedation avoided. Other strategies have given lessconsistent results, and the cost-benefit assessment ofthese approaches needs to be determined before theycan be routinely recommended. Early diagnosis andeffective antibiotic treatment can improve prognosis,and immunomodulatory therapies (unproven as yet)may become an important therapeutic adjunct.

REFERENCES

with tissue necrosis and cavitation, treatment mayneed to be prolonged (14-21 days).

Immunomodulating therapies

The suppression of host defence mechanisms is a keyfactor in the development of nosocomial pneu-monia, and therapies aimed at stimulating hostdefence may therefore be beneficial. Clinical trialshave focused on two main agents: interferon-gamma(IFN-g) and granulocyte-colony stimulating factor(G-CSF).

IFN-g has been shown to restore monocyte func-tion in vivo, but clinical trials using prophylacticIFN-g in patients with severe trauma or burns havebeen inconclusive, with no sustained improvementin infection rates.39

G-CSF has shown promising results in animalexperiments, but, in clinical trials, the results have beenless convincing, with no apparent effects of prophylac-tic use on the incidence of nosocomial pneumonia.40

Immunomodulation is an interesting area of on-going research, but remains experimental at present.

CONCLUSION

The high incidence and associated increase in morbid-ity and costs associated with nosocomial pneumoniademand that the critical care clinician remains alert tothe risk factors and early features of this condition.Prevention should play a key role in limiting thedevelopment of nosocomial pneumonia. Simpletechniques such as hand washing and disinfection, andplacing the patient in the semi-recumbent rather thanthe prone position must be promoted, and excessive

1. Vincent, JL, Bihari, D, Suter, PM, et ai, Members of the

EPIC International Advisory Group. The prevalence of

nosocomial infection in intensive care units in Europe-

the results of the EPIC study. JAMA 1995; 274: 639-44.

2. Kappstein, I, Schulgen, G, Beyer, U, Geiger, K,

Schumacher, M, Daschner, FD. Prolongation of

hospital stay and extra costs due to ventilator-

associated pneumonia in an intensive care unit. EurJ

Clin Microbiol Infect Dis 1992; 11: 504-8.

3. Chevret, S, Hemmer, M, Carlet, J, Langer, M, European

Cooperative Group on Nosocomial Pneumonia.

Incidence and risk factors of pneumonia acquired in

intensive care units. Intensive Care Med 1993; 19:256-64.

4. Fagon, JY, Chastre, J, Vuagnat, A, Trouillet, JL, Novara, A,

Gibert, C. Nosocomial pneumonia and mortality among

patients in intensive care units. JAMA 1996; 275: 866-9.

5. Richards, MJ, Edwards, JR, Culver, DH, Gaynes, RP.

Nosocomial infections in combined medical-surgical

intensive care units in the United States. Infect Control

Hosp Epidemiol 2000; 21: 510-15.

6. Tejada Artigas, A, Bello Dronda, S, Chacon Valles, E,

et al. Risk factors for nosocomial pneumonia in critically

ill trauma patients. Crit Care Med 2001; 29: 304-9.

7. Wallace, WC, Gnat, M, Gornick, WB, Lekawa, ME,

Wilson, SE. Nosocomial infections in the surgical

intensive care unit: a difference between trauma and

surgical patients. Am Surg 1999; 65: 987-90.

8. Kollef, MH. Ventilator-associated pneumonia. A

multivariate ana lysis. JAMA 1993; 270: 1965-70.

9. Torres, A, Aznar, R, Gatell, JM, et al. Incidence, risk,

and prognosis factors of nosocomial pneumonia in

mechanically ventilated patients. Am Rev Respir Dis

1990; 142: 523-8.

10. Fagon, JY, Chastre, J, Hance, AJ, Montravers, P, Novara,

A, Gibert, C. Nosocomial pneumonia in ventilated

patients: a cohort study evaluating attributable

mortality and hospital stay. Am J Med 1993; 94: 281-8.

References 199

11. Heyland, DK, Cook, DJ, Griffith, L, Keenan, SP,

Brun-Buisson, C. The attributable morbidity and

mortality of ventilator-associated pneumonia in the

critically ill patient. The Canadian Critical Trials Group.

AmJ Respir Crit Care Med 1999; 159: 1249-56.

12. American Thoracic Society. Hospital-acquired

pneumonia in adults: Diagnosis, assessment of severity,

initial antimicrobial therapy and preventative strategies.


13. Fayon, MJ, Tucci, M, Lacroix, J, et al. Nosocomial

pneumonia and tracheitis in a pediatric intensive care

unit: a prospective study. Am J Respir Crit Care Med

1997; 155: 162-9.

14. Ewig, S, Torres, A, El-Ebiary, M, et al. Bacterial

colonization patterns in mechanically-ventilated

patients with traumatic and medical head injury.

Incidence, risk factors, and association with ventilator-

associated pneumonia. Am J Respir Crit Care Med

1999; 159: 188-98.

15. Cardenosa Cendrero, JA, Sole-Violan, J, Bordes Benitez,

A, et al. Role of different routes of tracheal colonization

in the development of pneumonia in patients receiving

mechanical ventilation. Chest 1999; 116: 462-70.

16. Fagon, JY, Chastre, J, Wolff, M, et al., for the VAP Trial

Group. Invasive and noninvasive strategies for

management of suspected ventilator-associated

pneumonia. Ann Intern Med 2000; 132: 621-30.

17. de Jaeger, A, Litalien, C, Lacroix, J, Guertin, MC,

Infante-Rivard, C. Protected brush specimen or

bronchoalveolar lavage to diagnose bacterial

nosocomial pneumonia in ventilated adults: a

metaanalysis. Crit Care Med 1999; 27: 2548-60.

18. Bello, S, Tajada, A, Chacon, E, et al. 'Blind' protected

specimen brushing versus bronchoscopic techniques

in the aetiological diagnosis of ventilator-associated

pneumonia. Eur Respir J 1996; 9: 1494-9.

19. Wearden, PD, Chendrasekhar, A, Timberlake, GA.

Comparison of nonbronchoscopic techniques with

bronchoscopic brushing in the diagnosis of ventilator-

associated pneumoniaj Trauma 1996; 41: 703-7.

20. Chastre,J, Trouillet,JL, Vuagnat, A, et al. Nosocomial

pneumonia in patients with acute respiratory distress

synd rome. Am J Respir Crit Care Med 1998; 157:1165-72.

21. Drakulovic, MB, Torres, A, Bauer, TT, Nicolas, JM, Nogue,

S, Ferrer, M. Supine body position as a risk factor for

nosocomial pneumonia in mechanically ventilated

patients: a randomised trial. Lancet 1999; 354:1851-8.

22. Fagon, JY, Chastre, J, Domart, Y, et al. Nosocomial

pneumonia in patients receiving continuous

mechanical ventilation. Prospective analysis of 52

episodes with use of a protected specimen brush and

quantitative culture techniques. Am Rev Respir Dis

1989; 139:877-84.

23. Doebbeling, BN, Stanley, GL, Sheetz, CT, et al.

Comparative efficacy of alternative hand-washing

agents in reducing nosocomial infections in intensive

care units. N Engl J Med 1992; 327: 88-93.

24. Bischoff, WE, Reynolds, TM, Sessler, CN, Edmond, MB,

Wenzel, RP. Handwashing compliance by health care

workers: the impact of introducing an accessible,

alcohol-based hand antiseptic. Arch Intern Med 2000;

160: 1017-21.

25. Crowe, M, Towner, KJ, Humphreys, H. Clinical and

epidemiological features of an outbreak of

Acinetobacter infection in an intensive therapy unit.

J Med Microbiol 1995; 43: 55-62.

26. Eveillard, M, Eb, F, Tramier, B, et al. Evaluation of the

contribution of isolation precautions in prevention

and control of multi-resistant bacteria in a teaching

hospital. J Hosp Infect 2001; 47: 116-24.

27. Kollef, MH, Von Harz, B, Prentice, D, et al. Patient

transport from intensive care increases the risk of

developing ventilator-associated pneumonia. Chest

1997; 112:765-73.

28. Manthous, CA, Schmidt, GA. Resistive pressure of a

condenser humidifier in mechanically ventilated

patients. Crit Care Med 1994; 22: 1792-5.

29. Le Bourdelles, G, Mier, L, Fiquet, B, et al. Comparison

of the effects of heat and moisture exchangers and

heated humidifiers on ventilation and gas exchange

during weaning trials from mechanical ventilation.

Chest-\996; 110: 1294-8.

30. Nourdine, K, Combes, P, Carton, MJ, Beuret, P,

Cannamela, A, Ducreux, JC. Does noninvasive ventilation

reduce the ICU nosocomial infection risk? A prospective

clinical survey. Intensive Care Med 1999; 25: 553-5.

31. Nelson, LD, Choi, SC. Kinetic therapy in critically ill

trauma patients. Clin Intensive Care 1992; 3:

248-52.

32. Vandenbroucke-Grauls, CM, Vandenbroucke, JP. Effect

of selective decontamination of the digestive tract on

respiratory tract infections and mortality in the

intensive care unit. Lancet 1991; 338: 859-62.

33. Liberati, A, D'Amico, R, Pifferi, S, et al. Antibiotics for

preventing respiratory tract infections in adults

receiving intensive care. Cochrane Database Syst Rev

2000; 2: CD000022.

34. Cook, DJ, Reeve, BK, Guyatt, GH, et al. Stress ulcer

prophylaxis in critically ill patients. Resolving

discordant meta-analyses./AM/l 1996; 275: 308-14.


35. Messori, A, Trippoli, S, Vaiani, M, Gorini, M, Corrado,

A. Bleeding and pneumonia in intensive care patients

given ranitidine and sucralfate for prevention of stress

ulcer: meta-analysis of randomised controlled trials.

Br Med 72000; 321:1103-6.

36. Cook, D, Guyatt, G, Marshall, J, et al. A comparison of

sucralfate and ranitidine for the prevention of upper

gastrointestinal bleeding in patients requiring

mechanical ventilation. N EnglJ Med 1998; 338: 791-7.

37. Detriche, 0, Berre, J, Massaut, J, Vincent, JL The

Brussels Sedation Score: the use of a simple clinical

sedation scale can avoid excessive sedation in

mechanically ventilated patients in the intensive care

unit (ICU). BrJAnaesth 1999; 83 698-701.

38. Kollef, MH, Ward, S. The influence of mini-BAL

cultures on patient outcomes: implications for

the antibiotic management of ventilator-associated

pneumonia. Chest 1998; 113: 412-20.

39. Wasserman, D, loannovich, JD, Hinzmann, RD, Deichsel,

G, Steinmann, GG. Interferon-gamma in the prevention

of severe burn-related infections: a European phase III

multicenter trial. Crit Care Med 1998; 26: 434-9.

40. Heard, SO, Fink, MP, Gamelli, RL, et al. Effect of

prophylactic administration of recombinant human

granulocyte colony-stimulating factor (filgrastim) on the

frequency of nosocomial infections in patients with

acute traumatic brain injury or cerebral hemorrhage.

Crit Care Med 1998; 26: 748-54.

16Infection in the immunocompromisedpatientDAVID GHEZ, JEAN-FRANCOIS TIMSIT AND JEAN CARLET

Introduction

Pathophysiology: types of immune deficiencyand their relationship to infections

Epidemiology of severe pneumonia inimmunocompromised hosts

201 Non-infectious causes

Diagnostic strategy

201 Treatment

Conclusion

204 References

209

210

212

215

215

INTRODUCTION

Although at one time limited to rare constitution-al diseases or haematological malignancies,immune deficiency has become progressively moreprevalent in recent years. Acquired immunodefi-ciency syndrome (AIDS) accounts for a large pro-portion of this increase in cases, but the morepowerful and effective immunosuppressive drugregimens now used have also increased both theincidence and the variety of clinical patterns ofacquired immune deficiency encountered.Consequently, many more immunocompromisedpatients require intensive care unit (ICU) admis-sion with serious infections caused by both com-mon and opportunistic pathogens.

The occurrence of a pneumonia in an immuno-compromised host is a daunting challenge because itmay rapidly become life threatening, diagnosis isoften difficult, and patient survival depends greatlyon the speed with which the diagnosis is made andthe appropriate treatment instituted. Clinical presen-tation is often atypical and a wide range of micro-organisms may be responsible. In addition, other

non-infectious causes of pulmonary infiltrates, suchas intra-alveolar haemorrhage, may present aremarkably similar clinical picture and, although lessfrequent, should also be considered.

PATHOPHYSIOLOGY: TYPES OF IMMUNEDEFICIENCY AND THEIR RELATIONSHIPTO INFECTIONS

The nature and magnitude of the immune defi-ciency and its impact on specific host defencemechanisms are relevant to the clinician, becausethey largely determine which pathogens will beinvolved (Table 16.1). The aetiology of the immunedeficiency may be complex. It can be related to theunderlying disease itself, to the immunosup-pressive' therapy and, frequently, to both. Thenature of the immunosuppressive therapy - type,duration, cumulative dose - is particularlyimportant. The temporal sequence of theadministration of the different elements ofthe immunosuppressive regimen is also relevant,particularly after transplantation.

202 Infection in the immunocompromised patient

Table .16.1 Major causes of immunodeficiency and associated pulmonary pathogens (HIV infection excluded)

Neutropenia

Cell-mediated immunity

Humoral immunity

Chemotherapy inducedAgranulocytosisLeukaemiasConnective tissue diseases:

Systemic lupus erythematosusRheumatoid arthritis(Felty'syndrome)

Common variable immunodeficiency

Corticosteroid treatment

Haematological malignancies:Hodgkin's diseaseLymphomasAngioimmunoblastic lymphadenopathy

Connective tissue diseases:Systemic lupus erythematosusWegener's granulomatosis

Sarcoidosis

BMTOrgan transplantation

Constitutional immunodeficiencies:Hyper-IgM syndrome

Multiple myelomaSplenectomyBMTConstitutional immunodeficiencies:

Common variable immunodeficiencyAgammaglobulinaemia

Enterobacteriaceae,Pseudomonas aeruginosaGram-positive bacteriaAspergillus spp.

Viruses:CMV, HSV, VZV, RSV,adenovirus, rhinovirus, influenza,parainfluenza virus, enteroviruses,HSV6 (BMT)

Protozoans:Toxoplasma gondiiFungi:

Aspergillus spp

Pneumocystis cariniiCryptococcus neoformans

Coccidioides immitis3

Bacteria:Legionella pneumophilaMycobacterium tuberculosisAtypical mycobacteria

Nocardia asteroidesStrongyloides stercoralis5

Encapsulated bacteria:Streptococcuspneumoniae

Haemophilus influenzae

Salmonella sp.

aln areas of endemicity.HIV, human immunodeficiency virus; CMV, cytomegalovirus; HSV, herpes simplex virus; VZV, varicella-zoster virus; RSV, respiratory syncytial virus; BMT, bonemarrow transplantation.

Other factors include:

• alterations of natural barriers,• uraemia,• malnutrition,• concomitant infection with immunomodulating

viruses such as Epstein-Barr virus (EBV),cytomegalovirus (CMV) and hepatitis viruses.

The spectrum of potential pathogens in theimmunocompromised host is particularly broad.Although there is no rule, the nature of the organ-isms most likely to cause an infection depends mostly

on the type and duration of the immune defect(Tables 16.1 and 16.2). Infection with humanimmunodeficiency virus (HIV) produces quantita-tive and functional defects in cell-mediated immuni-ty. When antiretroviral therapy fails or is not taken,HIV-infected people experience a progressive declinein the number of circulating CD4 lymphocytes andare increasingly at risk of opportunistic infections.Functional defects in macrophage function associatedwith an impairment of some cytokine productiondiminish the host response to intracellularpathogens. Infection with pyogenic bacteria is

Pathophysiology 203

Table 16.2 Probable causative organisms according to time course and pattern of immunosuppression

Neutropenia

TransplantationBone marrow

Early infections

Late infections

Solid organEarly infections

Late infections

HIV-infected patientsCD4

>800-1000106/litre

Early declineCD4 250-800 106/litre

CD4 <200106/litre

<2 weeks: Gram-negative bacilli, Gram-positive cocci>2 weeks: increased risk of invasive pulmonary aspergillosis (less frequentlyCandida spp)

Before engraftment:Identical to that of neutropenic patients

After engraftment:Nosocomial bacterial pneumonia (also consider Legionella)Aspergillus (bimodal: day 20 and 80), other fungi (Blastomycosis,Coccidioi'domycosis) in areas of endemicityCMV (recipient + or donor + , especially between day 30 and 60)

HSV, HSV6Toxoplasmosis (Pneumocystis carinii)

Community-acquired bacteria (encapsulated bacteria: Strep, pneumoniae,Haemophilus influenzae)Respiratory viruses, VZVAspergillosis (less frequent after day 100)Pneumocystis carinii, Cryptococcus neoformansMycobacteria, Nocardia asteroides

Nosocomial bacterial pneumonia: Gram-negative bacilli, Staphylococcusaureus,Legionella sp (Pseudomonas if cystic fibrosis)CMV (donor + /recipient - )Aspergillosis (especially in lung transplant recipients)Mycobacterium tuberculosis (increased risk in renal transplant recipients)Pneumocystis cariniiToxoplasmosis (donor + /recipient - )

Community-acquired bacteria (encapsulated bacteria: Strep, pneumoniae,H. influenzae) Respiratory viruses, CMVPneumocystis carinii, Cryptococcus neoformansMycobacteria, Nocardia asteroides

Bacterial pneumonia: Strep, pneumoniae, Staph. aureus, H.influenzae, LegionellaAspiration pneumonia

As for HIV-infected patients plus: Mycobacterium tuberculosisIf non-infectious infiltrates: non-specific interstitial pneumonitis, lymphoidinterstitial pneumonitis, Kaposi's sarcoma, non-Hodgkin lymphoma) consideropportunistic infection

Pneumocystis cariniiCryptococcosis (in USA and Africa), histoplasmosis, aspergillosis, toxoplasmosis(in France)CMV infection, Mycobacterium avium intracellularePseudomonal pneumonia

CMV, cytomegalovirus; HSV, herpes simplex virus.


increased by other immunological defects, whichinclude the spontaneous activation of B-lympho-cytes accompanied by decreased responsiveness toneoantigens, a reduced level of type-specific anti-bodies and impaired chemotaxis, phagocytosis andbacterial killing of neutrophils.

EPIDEMIOLOGY OF SEVERE PNEUMONIAIN IMMUNOCOMPROMISED HOSTS

Outcome predictors and prognosis

Approximately 4-5% of hospitalized HIV-infectedpatients are admitted to the ICU, half of them withrespiratory failure.1 Several studies in the mid-1980sshowed an extremely high mortality (86-100%) forAIDS patients requiring mechanical ventilation forrespiratory failure. The prognosis of Pneumocystiscarinii pneumonia (PGP) in AIDS patients graduallyimproved and, in the 1990s, reports of improved sur-vival were noted in AIDS patients requiring mechan-ical ventilation, with a mortality of 40-70%. Thisimproved survival is probably related to a combinationof factors, including patient selection, earlier diagnosis,antiretroviral therapy, anti-Pneumocystis prophylaxisand corticosteroid therapy.2 Similar reasons might beadvanced to explain the improved survival of cancerpatients, whose overall mortality in the ICU is now40-50 %.3 General severity scores, although stronglyassociated with prognosis, nonetheless underestimatethe risk of death in this population. The outcome forcancer patients admitted to the ICU is adversely affect-ed by respiratory insufficiency, the need for mechanicalventilation and the development of septic shock, butneutropenia per se is not an independent predictor ofdeath in recent studies.4 Allogenic bone marrow trans-plantation (BMT) has the highest mortality: 18% ICUsurvival and 5 % 6-month survival in patients requir-ing mechanical ventilation.5 Autologous BMT has abetter outcome from ICU admission.

In general, the ICU outcome of immunocompromi-sed patients is related to the number of organs that failand particularly to the use of mechanical ventilation.The increasing use of non-invasive ventilation in thissetting has improved the prognosis of BMT patients.6

Non-invasive ventilation was also shown to improvemorbidity and hospital mortality (13/26 versus20/26, p = 0.02) of other immunocompromised

patients with acute respiratory failure in a recent ran-domized study.7

Causes of pulmonary infiltratesaccording to the immune status:epidemiology and preventive strategies

Infections are responsible for more than 60% of pul-monary infiltrates in immunocompromisedpatients.8 A thorough analysis of the underlying con-ditions and their treatments, and their influence onthe immune system, is required to define empiricallythe spectrum of possible pathogens (Table 16.2).Some settings are particularly frequent and shouldtherefore be detailed.

NEUTROPENIA

Neutropenia (neutrophil count < 1500 mm-3) ismore frequently encountered with the increasing useof cytotoxic drugs to treat malignancies and inflam-matory diseases. The vulnerability to infections isnot identical in every neutropenic patients as itdepends on several cofactors, including the degree ofthe neutropenia, its duration and the alterations ofthe natural barriers. An absolute neutrophil count

< 500 X 109 L-1 and a rapidly falling g r a n u l o c y t e c o u n t are major risk factors for infections, whereas

they are less frequent above 500 X 109 L-1 . Defectsin neutrophil function caused by chemotherapyincrease the susceptibility to infections, and changesin the gastrointestinal tract related to the chemother-apy facilitate invasion of the bacteria flora both fromthe mouth and via the bowel.

Bacterial pneumoniaFever has an infectious origin in more than 60% ofneutropenic patients and is often due to a primarybacteraemia. However, lung infections represent theprincipal cause of admission to the ICU of neu-tropenic patients. The majority of bacteria areacquired by the patient after hospital admission.Enterobacteriaceae, Pseudomonas spp. and nosoco-mial multi-resistant Gram-negative bacteria are theclassical pathogens, but Gram-positive organismsare being isolated with an increasing frequencyrelated to the presence of indwelling catheters(Staphylococcus species), prophylactic antibiotictherapy with fluoroquinolones and lesions of the

Epidemiology of severe pneumonia in immunocompromised hosts 205

oral mucosa (Streptococcus viridans). They nowaccount for more than 60% of the isolated organ-isms in blood cultures.9 Nosocomial outbreaks ofLegionella species pneumonia have also beenreported.

AspergillosisNeutropenia lasting for more than 2 weeks is associ-ated with a high risk of pulmonary aspergillosis.10

Because the post-therapeutic aplasia may lastbetween 30 and 40 days, patients receiving inductionchemotherapy for acute leukaemia are therefore par-ticularly at risk. Early antifungal treatment and bonemarrow recovery may allow the control of the infec-tion. In patients who remain deeply neutropenic,survival beyond 3 weeks is uncommon because dis-semination to other organs, especially the brain,often occurs.

Other pneumoniasPulmonary infections caused by common respira-tory viruses have been described mostly in patientssuffering from haematological malignancies. Fungi,including Candida, Fusarium and Trichosporonspecies, may cause lung infiltrates, usually in the set-ting of a disseminated haematogenous infection.

ALLOGENEIC BONE MARROWTRANSPLANTATION

Allogeneic BMT combines neutropenia and majoralterations of monocyte/macrophage functions inthe early phase post-transplantation with a profoundand long-lasting impairment of cell-mediated andhumoral immunity. Pulmonary diseases represent40-60% of complications and the estimated mortal-ity exceeds 30%. Diffuse alveolar haemorrhage is afrequent complication in patients with pulmonaryinfections who require mechanical ventilation.5 Thedifferent pathogens affect the recipient at differenttimes in the post-transplantation course, althoughthere are some overlaps (Table 16.2). One-half of theinfections occur in the first 4-6 weeks followingbone marrow grafting.11 Additional risk factors forinfectious complications include engraftment failure(resulting in prolonged neutropenia), HLA mis-match, volunteer-unrelated donor, use of T-depletedbone marrow, acute graft-versus-host disease(GVHD; with subsequent use of high-dose steroidsand anti-lymphocyte globulins) and chronic GVHD.

The risk periods for certain infections are changingin relation to:

• the reduction in the length of the neutropenicperiod with the use of granulocyte colony stimulat-ing factor (G-CSF) and peripheral blood stem cells,

• the increasing shift of viral (especially CMV) andfungal infections to the late period (after day 100)as effective early preventive strategies have beendeveloped, such as ganciclovir prophylaxis for theprevention of CMV reactivation in seropositiverecipients, and fluconazole prophylaxis for can-didiasis.

Early infections(< 3-months-post-transplantation)Bacterial Pneumonia Bacterial pneumoniasoccurring before engraftment, i.e. during the neu-tropenic phase, are clinically and bacteriologicallysimilar to chemotherapy-induced neutropenia. Itaffects up to 20% of patients.11 In addition to com-mon respiratory tract pathogens, the high toxicityof pre-transplant therapy (especially total-bodyirradiation) on the digestive tract and the occur-rence of digestive acute GVHD facilitates thetranslocation of enteric organisms. Legionellosis israre because of the systematic decontamination ofthe water supply.

Fungal Pneumonia Invasive pulmonary aspergillo-sis, whose overall incidence varies from 4% to 10%,10

remains one of the most feared complicationsfollowing allogeneic BMT. The risk factors are:

• prolonged neutropenia before BMT,• time to engraftment,• previous history of Aspergillus colonization or

infection,• use of immunosuppressive drugs to treat an acute

GVHD.

It has a bimodal distribution, with peaks at day 20and day 80. Its mortality remains as high as 95%,despite early diagnosis and therapy.

Systemic candidiasis frequently occurs after BMT,but the diagnosis of primary pulmonary candidiasisis controversial, with many authorities believing thatit arises secondary to hematogeneous dissemination.

The systematic use of prophylaxis has consider-ably decreased the incidence of PCP in BMT recip-ients, which is less than 10%, usually in patients


who cannot take pentamidine or cotrimoxazoleprophylaxis.11

Viral Pneumonia CMV infections remain one ofthe most important causes of morbidity and mor-tality after allogeneic BMT. CMV active infectionoccurs preferentially between day 30 and 60 afterBMT, by reactivation of a latent endogenous virus inCMV-seropositive BMT recipients. CMV pneumo-nia reflects a severe disseminated infection, which,once declared, has a 30-70% mortality, despite theuse of antiviral agents and immunoglobulins.11

Preventive approaches, such as the systematicadministration of ganciclovir to high-risk patients,have proven effective, but drug-induced neutrope-nia occurs in up to 30% of patients treated. Earlypre-emptive antiviral therapy using repeated moni-toring of pp65 antigenaemia and surveillance bron-choalveolar lavage (BAL) is preferable.

Herpes simplex virus (HSV) has become uncom-mon with the prophylactic use of aciclovir duringthe first month after transplantation. HSV-6 maycause an interstitial pneumonia.

Parasitic Pneumonia In the first 2 months post-BMT, the reactivation of a latent Toxoplasma gondiiinfection may cause pulmonary infiltrates, in the set-ting of a disseminated disease. The incidence of T.gondii is only 0.9%, but it has a mortality rate over95%. Serologic tests have no diagnostic value, butmay serve to identify the group of patients at risk.The diagnosis is difficult and often only made atautopsy; it relies on the visualization of tachyzoites inthe BAL and, more recently, polymerase chain reac-tion (PCR).

Late infections(>3 months post-transplantation)Bacterial Pneumonia Because of the long-lastingdefect in humoral immunity and the functionalasplenia induced by total body irradiation,Streptococcus pneumoniae and Haemophilus influen-zae are the main pathogens responsible for late bac-terial pneumonia. Legionellosis remains a risk inthese patients. Pulmonary infections caused byNocardia species or Mycobacterium spp are rare.

Viral Pneumonia Common viruses such asadenovirus, respiratory syncytial virus (RSV),rhinoviruses, enteroviruses, influenza and parainflu-enza viruses can cause devastating disease in severelyimmunocompromised hosts, especially after BMT, and

are frequently associated with bacterial superinfection.The incidence has increased in recent years, probablydue to improved methods of detection. The prognosisis poor in patients with pneumonia and the mortalityrate in patients with respiratory failure requiringmechanical ventilation has been reported to be over90%. Adenovirus infection can present as a dis-seminated disease responsible for severe hepatitis,gastroenteritis and encephalitis.

CMV pneumonia is less common more than 3months after transplantation, but may occur in asso-ciation with intense immunosuppression, particu-larly associated with the use of T-cell-depleted graftsand chronic GVHD. In one study, the incidence ofCMV infection after day 100 was 4.8%, with a mor-tality rate of 70%.n. Interestingly, late-onset CMVinfection has been reported in patients given pro-phylactic or early ganciclovir therapy, probably dueto an inability completely to restore their immuneresponse to CMV.

Most cases of varicella-zoster virus (VZV) infec-tions represent reactivation of a pre-existing infec-tion and occur in up to 50% of patients. Visceraldissemination, which may involve the lungs, hasbeen described in 5% of patients.

Fungal Pneumonia It is likely that aspergillosis alsoplays a significant role in fatal infections during thelate period, but its frequency has not been definedprecisely.

AUTOLOGOUS BONE MARROWTRANSPLANTATION

Intensification of the therapeutic regimen followedby autologous BMT is now routinely used to treathaematological malignancies and is increasinglyperformed in solid tumours. Although the proce-dure yields a considerably lower mortality thanallogeneic BMT (around 5-10%), the neutropenicperiod is often complicated by infections, theincidence of bacterial pneumonia being around3%.12 Because the neutropenia usually lasts lessthan 2 weeks, invasive aspergillosis has an inci-dence of 2%.10 CMV and other opportunisticinfections are rare. The use of G-CSF seems todecrease the risk of severe infectious complicationsand mortality. The use of peripheral blood stemcells also shortens the neutropenic phase, but thebenefits are debated.

Epidemiology of severe pneumonia in immunocompromised hosts 207

ORGAN TRANSPLANTATION

Although organ transplantation has achieved a 1-yearsurvival rate exceeding 80%, it can be followed by life-threatening complications, amongst which pulmonarydiseases are most prominent. Infections are the causeof approximately two-thirds of pulmonary infil-trates.13 In recent years, the incidence of infectiouscomplications has decreased with the use ofcyclosporin and tacrolimus. Pulmonary infectionshave different presentations and causes, according tothe timing post-transplantation. Early infections,which occur when the immunosuppression is maxi-mal, are more frequent and have a higher mortality.Because these patients receive a maintenance immuno-suppression to avoid graft rejection, they remain par-ticularly susceptible to opportunistic pathogens.

Bacterial pneumoniaApproximately 5% of solid-organ transplant recipi-ents develop bacterial pneumonia, more frequently inthe first 3 months. The incidence of bacterial pneu-monia depends on the tranplantation site: it is high-est in recipients of heart and/or lung (around 20%)and less common in liver or renal transplant patients.Identified risk factors are the occurrence of primaryCMV infection, splenectomy, graft rejection and anti-rejection therapy. The global mortality of bacterialpneumonia in solid-organ transplantation is above40% and that of early pneumonia exceeds 60%.

Gram-negative bacilli, Staphylococcus aureus andLegionella species predominate in the first 3 months.Lung allograft recipients are at greater risk for pneu-monia than other organ transplant recipients,because:

• the lung is directly exposed to the environment,• the transplanted lung is denervated, with

impaired mucociliary function and cough reflex,• patients typically require a higher level of

immunosuppression to prevent rejection.

Patients transplanted for end-stage disease sec-ondary to cystic fibrosis are often colonized byPseudomonas spp and Burkholderia cepacia andrequire particularly close surveillance. Empiricaltherapy must include agents against thePseudomonas strains that are prevalent in the institu-tion. Empyema and lung abscess require aggressivetherapy, with immediate drainage of potentiallyinfected collections. Late-onset bacterial pneumonia

is less frequent, primarily caused by community-acquired bacteria, e.g. Strep, pneumoniae andH. influenzae, and has a much better prognosis.

Mycobacterium tuberculosis infections, eitherprimary or reactivation, occur in approximately 1 %of solid-organ transplant recipients. M. tuberculosispneumonia only rarely develops early after trans-plantation. It occurs mostly in recipients of kidneytransplants, because the prevalence of tuberculosis ishigh in patients undergoing chronic haemodialysis.Tuberculosis is often disseminated and may be par-ticularly severe.

The incidence of Nocardia asteroides infectionsvaries between transplant centres (0-20%, mean 3%).Nocardia infection presents with pneumonia in 90%of cases, but also with brain abscess (30%), and occa-sionally involves the skin, eye or joints. The clinicalsigns are often non-specific and subacute. Chest radi-ograph findings include nodular, segmental or multi-lobar infiltrates. Cavitation and pleural effusion arefrequent. Even with appropriate antimicrobial thera-py, the mortality appraches 50%. The infectedpatients should receive cotrimoxazole indefinitely asrelapse can occur even after prolonged treatment.

Viral pneunoniaCMV infections are frequent in organ transplantrecipients and present at a similar time to thatdescribed for BMT recipents. However, unlike BMT,the donor's seropositivity represents the highest riskfactor for the development of symptomatic CMV dis-ease. Reactivation of latent virus is often less severethan primary infection following transplantation.Interestingly, CMV manifestations depend on theorgan transplanted. Recipients of lung or heart-lungtransplants are at higher risk of CMV pneumonitisand bronchiolitis obliterans. Prophylactic ganciclovirtherapy appears to reduce the incidence and severityof disease following renal transplantation and delaysthe disease in lung transplant recipients.

HSV pneumonitis has become a rare event withthe implementation of oral aciclovir prophylaxis,with an incidence of 1-10%, the highest-risk groupbeing heart and lung transplant recipients. As inBMT, EBV, VZV and common respiratory virusescan cause a severe, diffuse, interstitial pneumonia.

Parasitic pneumoniaPulmonary toxoplasmosis occurs primarily inseronegative patients who receive an allograft from a


seropositive donor. Because of the predilection of theparasite for muscle tissue, it is more frequent afterheart transplantation. In one study of T. gondii-seronegative recipients of allografts from T. gondii-seropositive donors, 57% of heart, 20% of liver and< 1% of kidney transplant recipients acquired pri-mary T. gondii infection. It is usually a fulminant dis-seminated disease with a mortality that exceeds 90%.14>15

Strongyloides stercoralis has been called the 'Trojanhorse' because it possesses an autoinfection cycle thatcan cause a chronic asymptomatic infestation fordecades after an exposure in endemic areas.Depression of the cell-mediated immunity mayresult in the so-called hyperinfestation syndromeassociated with haemorrhagic pulmonary consolida-tion, diffuse bilateral alveolar opacities leading toacute respiratory distress syndrome (ARDS) and gas-trointestinal symptoms. An associated bacteraemiaand/or meningitis caused by enteric bacteria occur in50% of cases. There is often no hypereosinophilia.The parasite is recovered on direct examination ofsputum, BAL or stool. When the diagnosis is sus-pected, immediate treatment must be instituted.Prophylactic treatment before transplantation isneeded in endemic areas. The prognosis is poor.

Fungal pneumoniaAs in BMT recipients, the implementation of pro-phylaxis against PCP has dramatically decreased therisk of this infection.

Pulmonary aspergillosis pneumonia occurs withinthe first 2 months following tranplantation, but, withthe exception of lung or heart-lung transplant recipi-ents, it has a low incidence (<5%) compared toBMT. Disseminated disease occurs in 50% of thesepatients and is almost invariably fatal.

Other causes of fungal pneumonia includeCryptococcus neoformans, Coccidioides immitis andHistoplasma capsulatum.

HUMAN IMMUNODEFICIENCY VIRUSINFECTION

The number of HIV-infected patients admitted tothe ICU has decreased following highly active anti-retroviral therapy. However, the absolute rate ofopportunistic infections remains high, e.g. PCP hadan incidence of 46 per 1000 patient-years in 1997,16

and some patients are still admitted with undiag-nosed HIV disease. The CD4 count remains the bestsurrogate marker of host immune response in HIV-

infected patients (see Table 16.2). Patients with aCD4 count >500 mm~3 are mainly at risk of com-mon bacterial pneumonia. As CD4 falls lower, infec-tions caused by relatively virulent pathogens such asM. tuberculosis may appear. Opportunistic infec-tions, especially PCP, occur when the CD4 count fallsbelow 200 mm-3. Under 100 mmT-3, cryptococcosis,M. avium complex and toxoplasmosis may cause dis-seminated infections and pneumonias.

Bacterial pneumoniaBacterial pneumonia is a frequent cause of ICUadmission for HIV-infected patients, caused by Strep,pneumoniae (penicillin-resistant strains are fre-quent), Pseudomonas aeruginosa and Staph. aureus.17

Pneumococcal pneumonia is frequently associatedwith positive blood cultures. Pseudomonal pneumo-nia is becoming a common complication, especiallyin patients with low leucocyte and CD4 counts.Chest radiographs are often atypical, mimicking PCPin half the cases. Intrapulmonary cavitations,abscesses and empyema are frequent.

Despite the widespread introduction of effectiveprimary and secondary prophylaxis, P. cariniiremains the cause in 50-90% of HIV-infectedpatients admitted to the ICU with pneumonia.Primary and secondary prophylaxis by cotrimoxa-zole, dapsone or aerosolized pentamidine is recom-mended when the CD4 count falls below 200mm-3.18 The clinical presentation is typically insidi-ous, with mild symptoms that progress over weeks ormonths. There may be surprisingly little clinical dis-tress despite severe hypoxaemia (PaO2 often <50mmHg on room air). The usual appearance is bilat-eral perihilar interstitial infiltrates, which, in severeforms, can progress to diffuse confluent alveolarshadowing. Hilar or mediastinal adenopathies arerare and suggest an alternative or concomitantprocess. Cotrimoxazole for 3 weeks in associationwith steroids remains the treatment of choice. TheICU mortality is under 20%.19 The adverse prognos-tic factors of PCP in ICU are: the need for mechan-ical ventilation after 3 days of ICU stay and/or formore than 5 days, the occurrence of nosocomialinfections and pneumothorax.

TuberculosisTuberculosis may occur at any stage in HIV diseaseand is less frequently responsible for admission tothe ICU. Positive tuberculin tests or contact with a

Non-infectious causes 209

person with active tuberculosis should lead to pre-ventive treatment.18 In the early stage of HIV disease,its clinical presentation resembles adult post-primarydisease. Later, the presentation is often atypical,cavitation is rare but hilar and mediastinallymph-adenopathy, diffuse and miliary shadowingor pleural effusion are frequent. A high proportion ofpatients have an extrapulmonary disease involvingthe bone marrow, liver, pericardium and meninges.17

Disseminated tuberculosis is responsible for septicshock and multi-system organ failure. The suspicionof tuberculosis warrants immediate conventionalanti-tuberculous therapy. Steroids are frequentlyused for severe miliary tuberculosis and may be lifesaving.

Pulmonary infection with C. neoformans usuallyoccurs as part of disseminated infection with menin-goencephalitis. This disease is common in France,the USA and Africa and rarer in other Europeancountries such as the UK. The presenting symptomsand chest radiographic appearance are non-specific.Serum, BAL and cerebrospinal fluid cryptococcalantigen are always positive. The diagnosis is estab-lished by the culture of BAL fluid or a transbronchialbiopsy specimen or, in disseminated disease, by theculture of blood or cerebrospinal fluid.

the main causes of infectious pneumonia, but awide range of opportunistic pathogens shouldalways be considered, amongst which mycobacteriaand P. carinii are the most common. Aspergillus is arare cause of pulmonary infiltrates and occurs inpatients receiving prolonged corticosteroid therapy.The aggressive treatment with high-dose steroidsand cyclophosphamide used in Wegener's granulo-matosis increases the risk of opportunistic infec-tion, particularly PCP, which was a significantcause of mortality before the systematic use ofprophylaxis.

Haematological malignanciesLymphoproliferative diseases may cause variousdefects in cellular and humoral immunity thatincrease the susceptibility to opportunisticpathogens (particularly P. carinii). Mycobacterialinfections are increased in hairy-cell leukaemia.Infectious complications occurring in acuteleukaemias are usually related to the neutropenia,but other pathogens have occasionally been reportedto cause pulmonary infections. Some treatments(especially steroids and 2-chloro-deoxyadenosine)deeply depress cellular immunity, which favours thedevelopment of opportunistic infections.

OTHER IMMUNODEFICIENT STATES

Humoral defiency and hyposplenismSplenectomy or functional hyposplenism (e.g tosickle-cell disease, SLE) renders patients particularlysensitive to encapsulated bacteria such as Strep,pneumoniae, H. influenzae or non-typhi salmonella,which can take a fulminant course. Patients withhypogammaglobulinaemia (common variableimmunodeficiency, lymphoproliferative disorders,especially multiple myeloma) that entails anopsonization defect, have a similar risk.

Collagen tissue diseasesVarious alterations in cell-mediated immunityhave been described in patients suffering from col-lagen tissue diseases, particularly SLE, rheumatoidarthritis and Wegener's granulomatosis. Whereasthese defects alone appear rarely to be responsiblefor severe opportunistic infections, the immuno-deficiency is frequently exacerbated by treatment(steroids, cyclophosphamide, methotrexate,cyclosporin). Community-acquired bacteria are

Trauma patients, severe sepsis and other ICUpatients with severe prolonged organdysfunctionCritical illness is associated with profound cellularand humoral immune dysfunction, probably relat-ed to excess production of pro-inflammatory andanti-inflammatory mediators. Hormonal, nutri-tional and genetic factors, as well as the increasinguse of corticosteroi'ds in septic patients, contributeto the impairment of immune function. Consequently,the severely ill ICU patient may develop infectionswith low-grade pathogens such as Staphylococcusepidermidis, Aspergillus fumigatus or cytomegalo virus.

NON-INFECTIOUS CAUSES

Non-infectious causes of pulmonary infiltrates arefound in more than 30% of immunocompromisedhosts and should be systematically eliminated (Table16.3). A detailed description of their characteristics isbeyond the scope of this chapter. Because they can


Table 16.3 Non-infectious causes of pulmonary infiltrates

Pulmonary oedemaCardiogenicNon-cardiogenic

Conditioning regimenVeno-occlusive diseaseGraft-versus-host disease

Alveolar haemorrhagePulmonary embolismBronchiolitisobliteransToxic: irradiationImmunoallergic

MethotrexateBleomycin

Idiopathic interstitial pneumoniaLeukaemic relapseLeukostasis (acute leukaemias)Retinoic acid syndromeCapillary leak syndromeAlveolar proteinosisPulmonary lymphomaEBV post-transplant lymphoproliferative diseaseKaposi's sarcoma

EBV, Epstein-Barr virus.

present with a wide variety of pulmonary complica-tions whose specific treatment relies on an increaseof the immunosuppressive therapy, collagen vasculardisease may cause difficult diagnostic and therapeu-tic problems. In patients with haematological malig-nancies, the aetiologies are dominated by pulmonaryhaemorrhage, pulmonary oedema, leukaemic relapseand transfusion reactions.

DIAGNOSTIC STRATEGY

The diagnostic approach to the evaluation of pneu-monia in the immunocompromised is based on themedical history.20 The understanding of the expect-ed 'timetable of infections' depending on the par-ticular programme of immunosuppression plays amajor role in the differential diagnosis (see Table16.2). Knowledge of the modes of onset and pro-gression of the pneumonia process is also useful. Anacute or fulminant onset suggests conventionalpyogenic bacterial infections (and, for non-infectious causes, pulmonary embolism, pulmonaryoedema, leukoagglutinin reaction or pulmonaryhaemorrhage). A subacute onset suggests viral,

Legionella, Mycoplasma, Pneumocystis, Aspergillusor Nocardia infection. A more protracted course,over several weeks, suggests fungal or tuberculousinfections.

Prior prophylaxis must also be reviewed. Forexample, compliance with cotrimoxazole prophy-laxis practically rules out the diagnosis of PCR

Clinical features

Even if the main symptom at admission is pneumo-nia, a careful physical examination is needed. The clin-ical presentation of these pneumonias may overlapconsiderably. Moreover, some processes are occasion-ally caused by mixed infection or can reflect infectionssuperimposed on another non-infectious process.Clinical features specific to particular aetiologies havealready been discussed. In neutropaenic patients, theclinical presentation is either pneumonia or ARDSdue to sepsis caused by bacterial or fungal infections.

The initial examination should include a chestradiograph, standard laboratory tests, two peripheralblood cultures and blood gases.

Radiological findings

Although no particular chest X-ray pattern is specificfor a given pathologic process, particularly in theimmunosuppressed patient, some patterns are morecharacteristic. The distribution and location of chest-imaging abnormalities might help in the differentialdiagnosis (Table 16.4). Neutropenia may greatly modifyor delay the appearance of pulmonary lesions.Atelectasis maybe the only radiological clue to the pres-ence of a clinically important bacterial pneumonia.

Computerized tomography (CT) scanning pro-vides more information, which may aid diagnosis,and it is also the best method for predicting whetherbronchoscopy will be helpful. It also enables a CT-guided biopsy to be performed, if necessary to makethe diagnosis, and allows for transparietal needlebiopsy of pulmonary nodules.

Pulmonary samples

Immunocompromised patients with pulmonarydysfunction should have a sputum Gram stain. If

Diagnostic strategy 211

Table 16.4 Diagnosis according to chest radioeranh aooearances in oneumonia of acute and subacute onset

Consolidation

Peri-bronchovascular/interstitial

Nodular infiltrate

Cavitation

Hilarand mediastinal lymph nodes

Pleural effusion

Pyogenic bacteria3

Legionnaires'diseaseIntra-alveolar haemorrhagePulmonary oedema(cardiogenic or not)

Pulmonary oedemaLeukoagglutinin reaction

(Right-sided endocarditis)

Gram-negative bacilliStaphylococcus aureus

Cardiogenic pulmonary oedema(bilateral)Bacteria

FungalNocardialTuberculosis

ViralPneumocystis (diffuse alveolar in severe

forms)Radiation (adjoin the edges of theradiation field)Drug induced

TumoursKaposi's sarcomaAspergillosisToxoplasmosisNocardia (macronodules)Tuberculosis (miliary)

Necrotic tumoursTuberculosisNocardiaAspergillosis(Pneumocystis: atypical cysts)

Tumours ++Mycobacteria +++Fungus

MycobacteriaFungus (except

P. carinii)Tumours

aBacterial pneumonia frequently leads to discrete infiltrates in neutropenic patients.

pleural fluid is present, it should be sampled andexamined for infection and malignancy. Sputumstain and culture for mycobacteria should also beroutine. Blood cultures should be taken and serumcryptococcal antigen requested. Induced sputum canbe useful in diagnosing pneumocystosis in HIV-infected patients, but these tests have little value inother immunocompromised patients.

In neutropenia, bronchoscopy with BAL is of lim-ited value in the diagnosis of pulmonary infection,particularly fungal infection, because of a high rateof false-positive bacterial isolates caused by contam-ination from the upper airway and a poor yield in thediagnosis of fungal infection.

Usually, in non-neutropenic patients, fibreopticbronchoscopy is needed to make the diagnosis.11'20

However, if patients are still breathing spontaneous-ly, bronchoscopy can precipitate the need formechanical ventilation. In this case, empiricalantimicrobial treatment should be given.Bronchoscopy should be delayed to confirm thediagnosis after the initial improvement of the patientor performed if the patient deteriorates and requiresmechanical ventilation. Nevertheless, an aggressiveapproach to diagnosis limits drug toxicity and inter-actions and the risk of potentially lethal superinfec-tion without exposing the patient to potentiallyinadequate therapy.


BAL has become the procedure of choice in theevaluation of diffuse pulmonary infiltrates because ithas a high diagnostic yield, especially with viruses,P. carinii and M. tuberculosis. In order to maximizethe sensitivity, bacterial, fungal and viral culture ofBAL fluid should be performed in addition to shellvial centrifugation culture and staining for viral anti-gens. Cytological evaluation should be performed byan experienced pathologist to look for viral inclusionbodies and fungal elements using specific stains. Itcan also confirm a non-infectious process such asalveolar haemorrhage. Non-diagnostic findings fromBAL are common, especially in the evaluation ofnodular lesions. In such a situation, a diagnosisshould be pursued with transbronchial biopsy, whichhas a higher sensitivity, but there is an increased riskof pneumothorax and haemorrhage and it cannot beperformed in patients with thrombocytopaenia orsevere ARDS. Transthoracic needle aspiration forperipheral lesions has a low yield, with significantrisks of pneumothorax (5-10%) and haemoptysis(3-5%). Open-lung biopsy will provide a diagnosis inthe majority of cases and should be considered bothwhen the empirical treatment is toxic or interactswith another medication and, particularly, if thepatient has a reasonable prognosis from the underly-ing disease.

Because obtaining lung tissue is potentially haz-ardous, other possible sites from which diagnostictissue could be obtained should be sought. Thisapplies especially to moulds that invade the blood-stream because the organism can often be identifiedin more easily accessible sites such as the skin, sinus-es and bone marrow. Diagnostic tests according tospecific pathogens are summarized in Table 16.5.

Empirical antimicrobial treatment

The sudden onset of symptoms and an often fulmi-nant course of the disease are characteristic of infec-tions in immunocompromised patients, particularlythe neutropenic patient. Therefore, the antimicrobialtherapy has to be instituted as early as possible andusually before microbiological results are available.

NEUTROPENIA

The majority of pathogens associated with newepisodes of acute pneumonia are bacteria. Because ofthe absence of neutrophils, clinical and radiologicaldata are frequently absent or not specific. Con-sequently, the first-choice antibacterial treatmentmust be broad spectrum and adapted to the bacterialspecies and their antibiotic susceptibility prevalentin the institution. A combination of an aminoglycosideand an extended-spectrum b-lactamin is recommend-ed for initial therapy.9 Vancomycin is appropriate ininstitutions where methicillin-resistant staphylococ-ci are frequent or when the direct stain of the sputumexamination shows Gram-positive cocci. Duringprolonged neutropenia, amphotericin B should beadded when Aspergillus or Candida spp infection issuspected.

The use of G-CSF has been shown to shorten theduration of neutropenia after low-risk chemothe-rapy. However, the results of several randomized tri-als have shown that G-CSF was ineffective inreducing either the incidence of infectious complica-tions or the overall mortality. However, G-CSF orGM-CSF might be useful in cases of severe sepsis orseptic shock or pneumonia.21 G-CSF may be respon-sible for the development of ARDS.22.

TREATMENT

Treatment of respiratory insufficiency

Mechanical ventilation is associated with an increasedrisk of death in immunocompromised patients,5

because of the very high risk of nosocomial infec-tions. CPAP and non-invasive ventilation6'7 should betried, before intubation and mechanical ventilation,when O2 therapy via facemask has failed. These tech-niques are also useful during and after bronchoscopywith BAL in patients with severe hypoxaemia.

INFECTIONS IN NON-NEUTROPENIC PATIENTS

In non-neutropenic patients with haematologicalmalignancy, the causes of pulmonary infection aresimilar to those in immunocompetent patients. Inthe case of acute onset, common pyogenic bacteriasand Legionella spp should be the targets of the initialtreatment.

When the onset of symptoms is delayed in patientsreceiving immunosuppressive therapy, P. cariniiinfection and tuberculosis should be considered,especially in cases of bilateral diffuse infiltrates. Onthe basis of AIDS experience, prednisolone should beadded when PCP is suspected.

Treatment 213

Table 16.5 Value of diagnostic tests for different pathogens

Pyogenic bacteria

Legionnaires' disease

Nocardia a stem ides

Tuberculosis

FungiP. carinii

Aspergillus spp.

Cryptococcus neoformans

ParasitesT. gondii

Strongyloides stercoralis

VirusesRSV

CMV

SputumBronchoscopy with trachealaspirate and BALBlood culturesDirect imunofluorescence (sputum)Urinary antigen + +Culture (BCYE)Sputum or bronchoscopy:Gram-positive beaded filaments(Gram stain) and culture (Ziehl-Nielsenor Kinyoun stain)

Sputum tracheal aspirate or BALTissue biopsyZiehl-Nielsen staining orimmunofluorescenceCulture (Lowenstein)

BAL: cyst or trophozoites(immunofluorescence or)toluene blueDefinite diagnosis: septatednon-pigmented hyphae in tissuebiopsy (Gomory methamine silveror acid Schiff stains) andpositive cultureBAL: smear, culture andantigen detection

BAL: direct examination and culturePositive serum cryptococcal antigen

BAL: tachyzoites (eosine/methyleneblue or Giemsa staining)

PCR methods, Culture on MRC5 cellsDirect stool (concentration)Sputum BAL

Nasopharynx, tracheal aspirate orBAL: culture (gold standard) or rapidantigen detectionBAL: shell vial assay, antigen,cytopathogenic effectBiopsies

Oropharyngeal contaminationIf possible according to severity ofrespiratory symptoms

Poor sensitivity

Culture: 5-21 days to grow the organism

Direct examination is often negative inimmunocompromised hosts

90% sensitivity, especially in AIDS patients;PCR is a very specific method

Isolation of Aspergillus spp. from therespiratory specimen: positive predictivevalue of \r\vas\veAspergillus infection of72% in patient with haematological malignancy, and 82% in BMT recipientsAspergillus antigen (EIA, ELISA orimmunoblot) serially, probablyvaluable in BMTPCR assays

The absence of serum T. gondii IgGmakes the diagnosis unlikely,especially in AIDS

Repeat tests + + +

RSV, respiratory syncytial virus; CMV, cytomegalovirus; BAL, bronchoalveolar lavage; BCYE, buffered charcoal yeast extract; PCR, polymerase chainreaction; MRC5 cells, AIDS, autoimmune deficiency syndrome; BMT, bone marrow transplantation; EIA, enzyme immune assay; ELISA, enzyme linkedimmunoadsorbent assay.


Table 16.6 Suggested initial treatment of severe pneumonia according to suspected pathogens

Pyogenic bacteria

Legionnaires' disease

Nocardia asteroides

Mycobacterium tuberculosis

Pneumocystis carinii

Aspergillus fumigatus

Cryptococcus neoformans

Histoplasma capsulatum

Respiratory syncytial virus

CMV

VZV

Toxoplasma gondii

Erythromicin 4 g day 1 i.v. andrifampicin 20 mg kg-1 day-1 for 21 daysCotrimoxazole + amikacin

Isoniazid (5 mg kg-1 day-1) +rifampidn(10 mg kg-1 day-1) +pyrazinamide (15 mg kg-1 day-1)

Cotrimoxazole (sulphamethoxazole100 mg kg-1 day-1, trimethoprim20 mg kg-1 day-1) for 21 days(second choice: pentamidine i.v.4 mg kg-1 day-1) and prednisone40 mg b.d. for 5 days, 20 mgb.d. for 5 days, 20 mg o.d. for 11 days

Amphotericin B 1.5 mg kg-1 day-1

(and 5-flucytosine 100-150 mg kg-1

day-1, dosage + + )If creatinine clearance < 30%:amphotericin B lipid complex 5 mgkg-1 day-1 or amphotericin B colloidaldispersion 5 mg kg-1 day-1 or liposomalamphotericin 3-5 mg kg-1 day-1

Amphotericin 0.7-1 mg kg-1 day-1

plus flucytosine 100 mg kg-1 day-1

for 2 weeks, then fluconazole400 mg day-1 or fluconazole400-800 mg day-1

Amphotericin B 0.7 mg kg-1 day-1

(or one of the lipid preparations at3 mg kg-1 day-1 for patients withrenal impairement)Then itraconazole 200 mg day-1 orfluconazole 800 mg day-1

Ribavirin (aerosolized) 6 g in 300 ml ofwater continuous nebulization 18-24 hoursdaily for 3-7 days and i.v. immunoglobulin

Ganciclovir 5 mg kg-1 day-1 b.d. pluspolyvalent immunoglobulinAciclovir 10 mg kg-1 8 h-1 for 7 days

Pyrismethamine 50 mg day-1 +(sulphadiazine 4 g day-1 orclindamycin 2.4 g day-1)

Broad-spectrum antimicrobialsif neutropenia

Imipenem + amikacin or ceftriaxone +imipenem are alternative treatments

Consider ethambutol (15 mg kg-1 day-1)if risk of drug resistancePrednisone 1-2 mg kg-1 day-1

in miliary tuberculosis

Fever and impairment of respiratoryfunction may occur when steroids aredecreased

G-CSF therapy has to be discussed inneutropenic patientsAmphotericin continued until 3 g or for 2weeks after resolution of all clinical signsand symptoms. Consolidation therapy forthe entire period of immunosuppression:itraconazole 400-800 mg day-1

(dosage + + + )Consider surgical resection of aspergilloma

In HIV-infected patients, consolidationtreatment must be continued for at least10 weeks

Consider prednisone 60 mg day 1 f o r 2weeks

Use closed system to prevent exposure tohospital personnel (teratogenicity)

References 215

Strongyloides stercoralis Thiabendazole25 mg kg 1 for 7-10 daysor ivermectine 200 (Jig kg-1 day1,2, 15, 16 + + +

Associated bacteraemia or bacterialmeningitis is frequentThiabendazole 25 mg kg-1 for 3 daysmust be given before immunosuppressivetherapy in areas of endemicity

CMV, cytomegalovirus; VZV, varicella-zoster virus; HIV, human immunodeficiency virus; G-CSF, granulocyte colony stimulating factor; i.v., intravenous; b.d.,twice daily; o.d., once daily.

In AIDS patients with a CD4 count > 200 mm 3,bacterial pneumonia is the most common cause ofrespiratory infection leading to admission to the ICU.In the case of diffuse alveolar infiltrates with miliarylesions, hilar or mediastinal lymph nodes or cavita-tion, an anti-tuberculous treatment should be empir-ically started. For diffuse lesions and subacute onsetof symptoms, PCP should be treated. Macrolidesshould be added if atypical pneumonia is suspected.

Other specific treatments of pathogens should bechosen according to previous anti-infective prophy-laxis, clinical history and results of specific tests andare summarized in Table 16.6. These are often toxicand should be carefully chosen on the results of diag-nostic tests.

CONCLUSION

Pulmonary infections in the immunocompromisedhost are an increasing problem for intensive careclinicians. In the neutropenic patient, the prognosisis related to the promptness of a broad-spectrumantibacterial treatment. In the non-neutropenichost, treatment should be given that covers the com-mon respiratory pathogens. Diagnostic tests, particu-larly bronchoscopy and BAL, remain the keyinvestigations in making the diagnosis and avoidingpotentially toxic empirical therapy. If possible, respira-tory support should be provided non-invasively andtracheal intubation avoided.

REFERENCES

1. Casalino, E, Mendoza-Sassi, G, Wolff, M, et al. Predictors

of short- and long-term survival in HIV-infected

patients admitted to the ICU. Chest 1998; 113: 421-9.

2. Sprung, C, Eidelman, LA. Triage decisions for intensive

care in terminally ill patients. Intensive Care Med

1997; 23:1011-14.

3. Kress, JP, Christenson, J, Pohlman, AS, Linkin, DR,

Hall, JB. Outcomes of critically ill cancer patients in

a university hospital setting. Am J Respir Crit Care

Med 1999; 160:1957-61.

4. Staudinger, T, Stoiser, B, Mullner, M, et al. Outcome

and prognostic factors in critically ill cancer patients

admitted to the intensive care unit. Crit Care Med

2000; 28: 1322-8.

5. Huaringa, AJ, Leyva, Fj, Giralt, SA, et al. Outcome of

bone marrow transplantation patients requiring

mechanical ventilation. Crit Care Med 2000; 28:1014-17.

6. Antonelli, M, Conti, G, Bufi, M, et al. Noninvasive

ventilation for treatment of acute respiratory failure

in patients undergoing solid organ transplantation:

a randomized trial. JAMA 2000; 283: 235-^1.

7. Hilbert, G, Gruson, D, Vargas, F, et al. Non invasive

ventilation in immunosuppressed patients with

pulmonary infiltrates, fever, and acute respiratory

failure. N Engl J Med 2001; 244: 481-7.

8. Rubin, HR, Greene, R. Clinical approach of the

compromised host with fever and pulmonary infiltrates.

In Clinical approach to infection in the compromised

host, ed. RH Rubin, LS Young. New York: Plenum

Medical Book Company, 1994; 121-62.

9. Pizzo, PA. Management of fever in patients with

cancer and treatment induced neutropenia. N Engl

J Med 1993; 328: 1323-32.

10. Stevens, DA, Kan, VL, Judson, MA, et al. Practice

guidelines for diseases caused by Aspergillus. Clin

Infect Dis 2000; 30: 696-709.

11. Soubani, AO, Miller, KB, Hassoun, PM. Pulmonary

complications of bone marrow transplantation. Chest

1996; 109: 1066-77.

12. Nosanchuk, JD, Sepkowitz, KA, Pearse, RN,

White, MH, Nimer, SD, Armstrong, D. Infectious

complications of autologous bone marrow and

peripheral stem cell transplantation for refractory

leukemia and lymphoma. Bone Marrow Transplant

1996; 18: 355-9.

13. Fishman, JA, Rubin, RH. Infection in organ transplant

recipients. N Engl J Med 1998; 338:1741-51.


14. Speirs, GE, Hakim, M,. Wreghitt, TG. Relative risk of

donor-transmitted Toxoplasma gondii infection in

heart, liver and kidney transplant recipients. Clin.

Transplant W88; 2: 257.

15. Gallino, A, Maggiorini, M, Kiowski, W, et al.

Toxoplasmosis in heart transplant recipients. EurJ Clin

Micmbiol Infect Dis 1996; 15: 389.

16. Kovacs, JA, Masur, H. Prophylaxis against opportunistic

infections in patients with human immunodeficiency

virus infection. N Engl J Med 2000; 342: 1416-29.

17. Miller, R. HIV-associated respiratory diseases. Lancet

1996; 348: 307-12.

18. 1999 USPHS/IDSA guidelines for the prevention of

opportunistic infections in persons infected with human

immunodeficiency virus. Clin Infect Dis 2000; 30:

S29-65.

19. Bedos, JP, Dumoulin, JL, Gachot, B, et al. P. carinii

pneumonia requiring intensive care management:

survival and prognostic study in 110 patients with

human immunodeficiency virus. Crit Care Med

1999; 6:1109-15.

20. Mayaud, C, Cadranel, J. A persistent challenge: the

diagnosis of respiratory disease in the non-AIDS

immunocompromised host. Thorax2000; 55: 511-17.

21. Ozer, H, Armitage, JO, Bennett, CL, et al. 2000 update

of recommendations for the use of hematopoietic

colony-stimulating factors: evidence-based, clinical

practice guidelines. J Clin Oncol 2000; 18: 3558-85.

22. Schilero, GJ, Oropello, J, Benjamin, E. Impairment in

gas exchange after granulocyte colony stimulating

factor (G-CSF) in a patient with the adult respiratory

distress syndrome. Chest 1995; 107(1): 276-8.

17Pleural diseaseWOLFGANG FRANK AND ROBERT LODDENKEMPER

Introduction

Pleural effusion

Empyema

217

217

226

Pneumothorax and bronchopleural fistula

References

229

233

INTRODUCTION PLEURAL EFFUSION

The pleura is a delicate, double-layer structure thatis physiologically characterized by a regulated fluidturnover that maintains a low-protein pool of lessthan 30 ml at a production rate of OJmLkg"1

day"1. The mechanisms for fluid transport (ultrafil-tration and active clearance) also provide the energysupply for the creation of the ventilation-modulatedand gravity-modulated negative pleural pressure.The gliding function of the fluid film results inmechanical coupling with the chest wall to ensurelung expansion in any physiological condition.Whereas the pleural cleft does not appear to be ofvital importance for respiratory function, it isresponsible for its vulnerability to a variety of localand distant disease processes that may interfere withlung expansion. Fluid accumulation due to inflam-matory or non-inflammatory causes and lung col-lapse from traumatic or spontaneous pneumothoraxare the most important. Empyema and bron-chopleural fistula may complicate these conditions.Significant pleural problems occur in 30% of allgeneral intensive care patients, mainly in the formof effusion (20%). The incidence of empyema andpara-pneumonic effusion is around 5% andpneumothorax and bronchopleural fistula accountfor roughly 10%, with some overlap between thesefigures.1

Effusion is the most common manifestation ofpleural disease, and the biochemical profile of pleuralfluid provides the basis for the classical distinctionbetween a transudate and an exudate. The clinicalimportance of this distinction relates to the aetio-logical causes and consequent therapeutic implications.Low-protein-containing transudates, representing aplasma ultrafiltrate, usually result from extrapleuraldisease interfering with hydrostatic-oncotic fluidbalance. Protein-rich exudates indicate directinflammatory or neoplastic pleural involvement.Increased vascular permeability and vascular injuryare the basic mechanisms, but impaired lymphdrainage may be a contributory factor in both tran-sudates and exudates. In the non-critically ill patient,approximately 40% of effusions are transudates, sothat the odds for a patho-anatomic pleural cause inan undefined effusion are 0.6 or less. Taking intoaccount the high prevalence of cardiac dysfunctionin intensive care, the ratio is probably less in this set-ting. Indeed, the majority of exudates in surgicalcases (e.g. post-thoracotomy) may still have cardiacdysfunction and/or excessive fluid administration asthe principal cause. When assessing the clinical rele-vance of pleural effusion in the medical patient,however, it is important to realize that even smallfluid collections indicate severe impairment of

218 Pleural disease

regulatory mechanisms, because the pleural drainagesystem may accommodate a 20-fold increase (= 700mL) of the low normal fluid influx.

Causes of pleural effusion

TRANSUDATIVE EFFUSION

Extrapleural or systemic diseases associated withtransudative effusions are listed in Table 17.1.Congestive cardiac failure (CCF) is the most impor-tant cause. Approximately half of all patients withCCF will develop transudative effusion, which willbe bilateral in 88% of cases. Unilateral effusion ismore common on the right side. Left heart (mainlydiastolic) dysfunction is the essential pathophysio-logical mechanism, as indicated by a correlationwith pulmonary venous pressure and reflected bypulmonary capillary wedge pressure. In CCF, eleva-tion of systemic venous pressure may be contribut-ory by increasing systemic capillary filtration.Another mechanism appears to be elevated 'backpressure' at the lymphatic-venous junction (tho-racic duct-jugular vein), with a consequent decreasein lymphatic clearance. Both these mechanisms arebelieved to explain a transudate in constrictive peri-carditis due to diastolic dysfunction of the right ven-tricle. Isolated right ventricular systolic dysfunction,for example due to increased after-load in pul-

monary hypertension, does not cause pleural effu-sion, even when severe (PAP 5=50 mmHg). Theoccurrence of pleural effusion in right heart failuretherefore suggests either associated left heart dys-function or pericardial constriction. In acute leftventricular failure with pulmonary oedema, over90% of patients will have radiologically obvious, orconcealed, effusions.

In hepatic dysfunction complicating cirrhosis,about 8% of patients will have an effusion, with astriking 90% right-sided predominance. Hypo-albuminaemia is an important pathogenic mechanism,but a fluid shift from peritoneum to pleura contri-butes - a condition referred to as 'porous diaphragmsyndrome'. The right-sided preference is so char-acteristic as to require exclusion of a peritoneal causein recurrent or therapeutically unresponsive right-sided transudates. The same mechanism applies inchronic pleural effusions seen in continuous ambulantperitoneal dialysis (CAPD).

Other causes of transudates, usually bilateral,include nephrotic syndrome and renal failure. Bothhypoproteinaemia and fluid overload result in grossascites, pleural effusions and generalized peripheraloedema. Transudative effusion may occur in pleuralinflammation such as pneumonia, thromboembolicdisease or malignancy, but is unusual. Interestingly,long-standing transudates may undergo partialresorption and appear as exudates (so-called pseudo-exudates), especially with diuretic therapy.

Table 17.1 Causes of transudative effusions

Congestive heart failureHepatic cirrhosis

Nephrotic syndromeOther (total)Continuous ambulant peritoneal dialysis

Hypoalbuminaemia

UrinothoraxAtelectasis3

Constrictive pericarditisTrapped lung3

Superior vena cava obstruction

Sarcoidosis3

Pulmonary embolism3

80848

Elevated hydrostatic pressure, hypoproteinaemiaand pleuro-peritoneal communication

Hypoproteinaemia, systemic fluid overload

Pleuro-peritoneal communication

Hypoproteinaemia,Obstructive uropathyEffusion 'e vacuo'Increased central venous pressureAdjacent inflammationVenous hypertension and lymphatic obstructionInflammationElevated hydrostatic pressure ± inflammation

aAlso exudative effusion possible.

Pleural effusion 219

EXUDATIVE EFFUSION

Exudative effusion results from inflammation of thepleura, with leakage of blood constituents (in partic-ular protein) from damaged pleural capillariescaused by systemic (mostly inflammatory or neoplas-tic) disorders, adjacent organ disease (pulmonary,mediastinal, abdominal) or primary pleural pathology.Differential diagnosis is therefore difficult, asindicated in Table 17.2. With respect to critical care,the important causes are more restricted:

• infection accounts for about 50% of cases,• malignant disease for 25.5%,• thromboembolic disease for 19%,• gastrointestinal disease for 4%,• autoimmune disease, tuberculosis and others

for 1.5%.J

Malignant effusion ('pleuritis carcinomatosa') iscommon, with carcinoma of the lung (22%), breast(20%) and gastrointestinal tract (17%) predominat-ing and accounting together for about 60% of malig-nant causes.7 Effusions with no detectable tumourseeding are termed para-neoplastic, occasionally, theymay have transudative characteristics or result fromsecondary problems such as atelectasis, pneumonia,impaired lymph drainage or systemic factors.Tuberculous pleurisy should be suspected when

Table 17.2 Causes of exudative effusion

there is a predominance of lymphocytes, multilocu-lation on imaging and a positive PPD test, althoughthis may be falsely negative in 30% of cases.

Pulmonary embolism is an important, and oftenoverlooked, cause of pleural (mainly exudative)effusion, with a reported incidence of up to 18% ofall non-cardiac pleural effusions.1 The possibility ofpulmonary embolism should be considered in anynon-attributable effusion, a situation not infre-quently encountered in the respiratory critical carepatient.

Chylothorax is a rare, but serious, mostly right-sided, specific cause of exudative effusion. Its clinicalsignificance is related to the specific problem of con-trolling the spilling of chyle into the pleural cavity.The most common cause is injury to the thoracicduct by malignancy (in particular lymphoma), andtrauma is the second most common cause, withinflammation (e.g. tuberculosis, sarcoidosis) andeven benign lymphatic disease or central venousthrombosis as rare aetiologies. Chylothorax is easilydiagnosed by its characteristic milky appearance andthe diagnosis is confirmed by chemical analysis withtriglyceride levels >110mgdL~1 or lipid electro-phoresis in ambiguous cases. In pseudo-chylothorax,also referred to as 'cholesterol pleurisy' - pathophysio-logically an entirely different condition - the turbidity

InfectiousBacterial pneumoniaAtypical pneumoniasParasitesNocardiosisFungal diseaseTuberculosisAIDSPost-transplant

Other inflammatory conditionsPulmonary embolismRadiation therapyAsbestos pleuritisSarcoidosisARDSPost-thoracotomy syndromeUraemic pleurisyPost-transplant

Gastrointestinal diseasePancreatitisPancreatic pseudocystSubphrenic abscessHepatic, splenic abscessHepatitisChylous ascitesOesophageal rupture

MalignancyCarcinomaSarcomaLymphoma/leukaemiaMesotheliomaChylothoraxAIDS

Connective tissue diseaseLupus pleuritisRheumatoid pleurisyMixed connective tissue diseaseSjogren syndromeChurg-Strauss syndromeWegener's granulomatosisFamilial Mediterranean fever

latrogenic and traumaticMisplaced CVP catheterPerforated oesophagusHaemothorax

MiscellaneousMeigs syndromeLeiomyomatosisYellow nail syndromeCholesterol effusionChylothoraxHypothyroidism

AIDS, autoimmune deficiency syndrome; ARDS, acute respiratory distress syndrome

220 Pleura I disease

is due to a high and diagnostic cholesterol content(>200 mg dL- 1). The condition may develop in long-standing pleural collections, usually non-bacterial,such as post-traumatic effusion, rheumatic ortuberculous pleurisy. Haemothorax is considered tobe present when the fluid approximates to blood.Because sanguinous effusion with haematocrit valuesas low as 5% may appear to be blood, the diagnosisof haemothorax requires laboratory proof by adefining haematocrit > 50% of circulating blood.Blunt or penetrating chest trauma is the main cause,but malignancy, pulmonary infarction, (iatrogenic)bleeding disorders and spontaneous pneumothoraxmay be causative.

Clinical findings

Clinical signs and symptoms are closely related to theamount of intrathoracic volume displacement andwill cause varying degrees of dyspnoea. With largeeffusions, respiratory failure type I is common, oftenassociated with hyperventilation. Respiratory failuretype II, unless in compressive effusion, is unusualand would suggest co-existing obstructive airwaydisease. Major distress (orthopnoea) will requireseveral litres of fluid, resulting in lung compressionor mediastinal shift and possibly even causing cen-tral venous congestion and a low cardiac outputsecondary to constriction. The sensitivity of physicalexamination in small effusions is limited (300-400mL) and loculated collections may escape detection.

Imaging

Conventional chest radiography (CXR) and ultra-sonography are the basic imaging techniques.Blunting of the costo-phrenic angle is the firstevidence, requiring >175mL for detection. In thesupine position of debilitated patients, the classicallaterally ascending contour sign is lost and at least 500mL is necessary to produce a diffuse, ground-glassopacity on supine CXR. Ultrasound is more sensitivethan CXR and will detect effusions as small as 50-100mL. It is particularly useful in evaluating multilocularand encapsulated effusions. Echogeneity to high-frequency (5-10 MHz) ultrasound has been shownto correlate with pleural fluid density and may havesome predictive value in the distinction between

transudates and exudates.2 In addition, extendedtechnical capabilities using convex or sector scannersand additional probing windows (suprasternal,parasternal, infraclavicular, subcostal) have expandedthe investigational range to a depth of 26 cm, there-fore reaching more central areas of the diaphragmand major portions of the mediastinum. Thus, itallows evaluation of diaphragm motility in the spon-taneously breathing patient, although, in controlledmechanical ventilation, paralysis may be missed.Ultrasonography is also valuable for the distinctionbetween a subdiaphragmatic and a pleural cause ofsuspected effusion. However, the posterior medi-astinum, hilum, paravertebral area and subscapularregion still remain blind areas.

Therefore, the advantages of ultrasonography are:

• precise bedside visualization and localization ofeffusion,

• guidance for diagnostic and therapeutic interven-tions,

• detection of loculations, adhesions and septae,• it may give clues as to the effusion profile,• it allows evaluation of diaphragm motility and

exclusion of subdiaphragmatic processes,• it is safe and repeatable,• it is easy to perform and inexpensive.

Computed tomography (CT) clearly offersgreater imaging information and is:

• often necessary to assess the relative contributionsof pleural or lung pathology in complex situ-ations, such as multiple air-fluid levels that mayresult from pleural or intrapulmonary infection,

• essential for investigating mediastinal involvement.

It also provides an image that is more accessible tothe non-radiologist and interpretation is less oper-ator dependent than ultrasonography. However, itsvalue is limited by the fact that it is only usuallyavailable outside of the intensive care unit/highdependency unit (ICU/HDU), with attendant riskswhen transporting the critically ill.

Portable CT facilities are now available in largerICU settings. CT is unreliable in assessing the pres-ence of septal loculations, for which ultrasonographyis superior. Technical developments, such as highresolution (HRCT) or helical scanning, allow multi-planar or three-dimensional reconstruction, and intra-venous contrast enhancement will increase the informa-tion yield. The ability to scan rapidly, e.g. during a


breath hold, has revolutionized image quality. Theoverall accuracy of HRCT in diffuse pleural thickeningwas 97% in a recent study.3 The sensitivity of CTfor malignant infiltration of the chest wall remainslimited. Three-dimensional techniques have beenused successfully to study pleural surface infiltration,identifying visceral pleural involvement in 92%,compared to 17% of patients with two-dimensionalCT imaging.3

Magnetic resonance imaging (MRI) is now thegold standard for imaging infiltrative chest-wallprocesses. Its better soft-tissue assessment comparedto CT, improved spatial resolution and imagingversatility make it the best option for detectingdisruption, invasion or thickening of the pleuralmembranes. In one study, Tl-weighted, contrast-enhanced MRI sequences correctly identified 15/18cases of malignant and 16/18 cases of benign chest-wall disease.4 The problems of access, however, makethis rarely a realistic possibility in the critically ill.

Diagnostic procedures

THORACOCENTESIS

Once clinical or imaging findings have establishedthe presence of fluid, thoracentesis is the next diag-nostic step. If there is reasonable evidence for a con-dition associated with transudates, thoracentesis isonly required should it fail to respond to medicaltherapy. With large effusions, thoracentesis can bepercussion guided; with small effusions, ultrasoundguidance provides increased safety and efficacy. Fordiagnostic purposes, 10-20 mL should be aspiratedinto an optionally anticoagulant-conditioned (e.g.sodium citrate) syringe and appearance, smell andspecific gravity should be noted, as well as requestingbasic biochemical, cytological and microbiologicstudies. PF may be serous (clear or turbid), suppura-tive, sanguinous, chylous or frankly haemorrhagic.Light's criteria, based upon the protein and lactatecontents and their ratio compared to serum, havebeen traditionally employed to differentiate transu-dates from exudates and have 98% sensitivity, 77%specificity and an overall accuracy of 95%.5 Morerecently, consideration of cholesterol has shownadditional value, increasing specificity (91%) butreducing sensitivity (81%).5 Therefore, their com-bined use is advised.

The following are the cut-off values for discrimin-ating between exudates and transudates (with thepleuralrplasma ratio in parentheses):

• protein </>3 g dL-1 (0.5)• LDH </>200 IU (</>0.6)• cholesterol </>60 mg dL-1 (</>0.3).

Glucose and pH values are also helpful para-meters. Low values of both are characteristic ofempyema, tuberculosis, rheumatic effusion andmalignancy, and reflect enhanced local anaerobic metabolism. The analysis of cellular components helps inthe differentiation of inflammatory and neoplasticaetiologies. An exclusive neutrophil leucocytosischaracterizes empyema and para-pneumonic effusion,a relative lymphocytosis or eosinophilia suggestvarious inflammatory and non-inflammatoryconditions, as listed up in Table 17.3.6

Cytological analysis requires considerable experi-ence to reliably discriminate inflammation frommalignancy. The diagnostic yield of conventionalcytology is generally reported to be 40-70%, but87% sensitivity has been reported.1'6'7 The yieldwith advanced cytological analysis (immunocytol-ogy, immunocytometry, flow cytometry) is70-80%, with up to 91% sensitivity and 100%specificity being reported. A panel of commerciallyavailable markers - mostly monoclonal antibodies(MAb) - is now often used to detect and identifytumour cells.

Immunological parameters such as rheumatoidfactor, anti-nuclear antibodies, anti-neutrophil cyto-plasmatic antibodies (c-ANCA) and components ofthe complement cascade (C3, C4) may confirm asystemic clinical condition as causative, but theirsensitivity is limited. Measurement of interleukin andother cytokines, e.g. interferon-gamma (IFN-*y andtumour necrosis factor-alpha (TNF-a) may provehelpful in the diagnosis of tuberculous or rheumaticeffusions. Adenosine deaminase (ADA) has beensuggested as useful in suspected tuberculosis, withreported 100% sensitivity and 87.5% specificity;however, its accuracy appears largely modifiedby epidemiological factors (age and regionaltuberculosis prevalence).8 Initial enthusiasm con-cerning the potency of polymerase chain reaction(PCR) and nucleic acid amplification techniques(NAAT) in general for the diagnosis of tuberculouspleurisy has been dampened by a reported limitedoverall sensitivity of 47-87%, in both pleural fluid

222 Pleural disease

Table 17.3 Conditions associated with lymphocytosis (>80%) and eosinophilia (>10%) in the pleura! fluid

Pleural lymphocytosisTuberculosisLymphomaChylothoraxRheumatic effusionSarcoidosisMalignancyYellow nail syndrome

Pleural eosinophiliaPneumothoraxHaemothoraxPrevious thoracentesisPulmonary embolismBenign asbestos pleuritis

Parasitic diseaseFungal diseaseAllergic and immunological conditionsLymphomaCarcinoma

Most frequent cause (90-95% lymphocytes)—100 %, in particular non-Hodgkin

Associated with trapped lungVery rare (>90% lymphocytes)In about 50%, but <70% lymphocytesVery rare

Most common cause, up to 50% eosinophiliaDelayed occurrencePneumothorax and bleeding relatedMay be haemorrhagicUp to 50% eosinophilia, presumably often underlying

'idiopatic' pleural effusionVarious parasitesHistoplasmosis, coccidioidomycosisDrugs, Wegener's granulomatosisM. HodgkinUncommon, even with blood-stained effusion

According to Sahn.1

samples and paraffin-embedded tissue specimens,particularly in cultural negative (pauci-bacillary)pleurisy.9 Carcino-embryonic antigen (CEA) is theonly tumour marker of value in malignant effusion,its presence distinguishing adenocarcinoma frommesothelioma.

PLEURAL BIOPSY AND ENDOSCOPY(THORACOSCOPY)

When there is clinical suspicion of pleural pathol-ogy, but imaging techniques and thoracentesisprovide inconclusive or conflicting results, blindpleural biopsy or thoracoscopy will be required fordiagnosis. A recommended scheme is depictedin Figure 17.1.7,10 Bacterial pleurisy may progressfrom benign para-pneumonic effusion toempyema.

Blind needle biopsy is useful in suspected tubercu-lous or malignant effusion. The Tru-cut needle maybe preferable to the older Abraham's needle. It shouldbe diagnostic in 74% of malignant causes and, com-bined with microbiology, in 60% of cases of tubercu-losis.1'6'7'10 However, the ability to biopsy suspiciouspleural areas at thoracoscopy, when technically feasi-ble, provides a significantly higher diagnostic yield.

Video-assisted thoracoscopy is a technical expansionof the original direct vision, single-entry technique,which has also led to the development of video-assisted thoracic surgery (VATS). Unlike traditionalthoracoscopy, which can be performed with localanaesthesia in the endoscopy unit, VATS requiresdouble-lumen intubation and a general anaesthetic.Whereas medical thoracoscopy may be basicallyincorporated in an ICU setting, this is probably onlyexceptionally a realistic option for VATS. Relativecontraindications include coagulation disorders,severe cardiac dysfunction and respiratory failure(unless the patient is intubated and ventilated). Apartfrom the advantage of direct visualization, adhesionsand loculations can be broken down, placement ofdrains optimized and air leaks evaluated inpneumothorax.

The value of medical thoracoscopy may thus besummarized as follows:10

• >90% diagnostic ability in exudative effusion,• staging of mesothelioma or bronchial carcinoma,• provision of optimum pleurodesis,• breaking down of loculations and debridement in

tuberculosis and empyema,• assessment of pleural leaks in pneumothorax.


Management of large pleuraleffusions

In high dependency or intensive care medicine, themanagement of large effusions is aimed at the urgentrestoration of lung expansion to improve gas exchangeand allow restoration of venous return in the presenceof cardiac embarrassment. Interventions can be dividedinto systemic and local approaches and also into acutepalliative and elective control strategies.

ACUTE PALLIATION

Transudates will respond to therapy aimed at treat-ing the primary cause, for example improvingmyocardial contractility or correcting fluid overload.Diuretics, and albumin replacement in the case ofsevere hypoproteinaemic states, are effective,although thoracocentesis may be required initially inrespiratory distress or failure. Correction of markedhypoproteinaemia with intravenous albumin shouldbe implemented in a prolonged fashion to avoiddetrimental over-expansion of the intravascular fluidcompartment. Frusemide infusions are more effect-ive than bolus therapy, but potassium and

Figu re 17.1 General algorithm for

the clinical diagnosis of pleural

effusion 7 TB, tuberculosis.

magnesium depletion should be avoided. Spiron-olactone is indicated in hypoproteinaemic states. Inview of the evidence favouring its long-term use inCCF, it should now be used in most conditions withgeneralized oedema unless there is severe impair-ment of renal function. In the case of a large exudate,tube drainage is the first management strategy. Theinsertion of a chest drain is conventionally per-formed in the lateral decubitus position, especially inthe ventilated patient. In severe respiratory distressin the spontaneously breathing individual (who maybe orthopnoeic), it should be carried out with thepatient sitting upright and supported. As vago-vagalsyncope may occur, pre-medication with atropine isa wise precaution if this approach is being adopted.Local anaesthesia should be generously infiltratedinto the relevant rib interspace (avoiding potentialdamage to the subcostal neurovascular bundle), withblunt dissection and separation of muscle planes downto the pleura. Alternatively, thoracoscopy-derived tro-car systems using a 9-mm external diameter sleeveand a sharp obturator are in common use.

It is important to avoid damage to underlyinglung, and drains should not be inserted using force.If there is any doubt as to the presence of loculations


and adhesions, ultrasonography should be used toguide placement, although the infiltrating needle willoften give warning if the effusion is unexpectedlyshallow. Particular attention needs to be paid to fix-ing the drain after insertion, although purse-stringsutures are no longer recommended. Kinking at theskin is a common cause of subsequent failure, andblockage of the tube will increase the risk of surgicalemphysema. We favour the use of large-bore, trans-parent silicone or PVC tubes (>24 F), which provideexcellent suppleness with optimum patient toleranceand resistance to kinking and also allow visual paten-cy control. The angle of the drain should be acute,with the skin incision being over the rib immediate-ly below the relevant rib space. A transparent dress-ing is preferable to allow inspection of the drain site.Immediate relief will usually result from relativelysmall (500 ml) drainage. Pulmonary re-expansionoedema is unlike to occur, even after the evacuationof large effusions, except in the case of long-standingcompressive collections, when the lung has devel-oped complete atelectasis with significant surfactantdepletion. However, in order also to avoid hypoten-sive circulatory effects, drainage should not exceedI Lhour"1 and the suction level should be low(10cmH2O). Cough provoked by re-expansion iscommon and may be distressing. Pre-medicationwith an opiate may therefore be useful if there are nocontraindications. However, these aspects are rarelyan issue in the ventilated patient.

Long-term control

The majority of effusions that require definitiveaction for long-term control are caused by malig-nancy. After initial tube drainage and confirmationof the malignant cause, instillation of pleural irri-tants to produce a pleurodesis will usually be neces-sary. The value of large-bore drains (3=24 F) inpleurodesis is to optimize instillation and reduce therisk of tube obstruction by viscous, fibrinous exu-dates or blood. Should this occur, as indicated by theabsence of fluctuations that reflect pleural pressureswings in the water column of the drainage bottle,mechanical manipulation to mobilize fibrin depositsor clots can be performed ('milking': Fig. 17.2).Disposable commercial systems that combine fluidcollection and suction control systems (including ahigh-pressure safety valve) are available (Fig. 17.3).

Mechanical wall suction, using a central pressurizedair supply, is still in common use and may be com-bined with a variety of collection systems. Oneadvantage of the venturi-operated system is an(almost) infinite suction reserve with a range in pres-sure up to 100 cmH2O. For safety reasons, a low-pressure 'thoracic' device is preferred and the routineuse of suction levels of 10-20 cmH2O suffices.

Thoracoscopy, as described above, providesoptimum placement of chest tubes and helps in pre-dicting likely lung re-expansion. The recognition ofthe degree of lung encasement by thickened visceralpleura or complicating adhesions is helpful as lungre-expansion is critical for successful pleurodesis.Pleurodesis may be achieved by the instillation oftalc powder ('poudrage') with a hand-bulb operateddevice (intra-thoracoscopic pleurodesis), ensuringwidespread powder distribution (Fig. 17.4). Dosesup to 8 g have been evaluated as safe. Talc poudragehas been evaluated in controlled, prospective trialsand has been shown to provide >90% long-termcontrol in malignant pleurisy.10,11 Alternatively, agentssuch as tetracycline or doxycycline can be instilledvia the chest tube, with reported success ratesbetween 54% and 96%. A figure around 70% is real-istic.11 The instillation of talc as a slurry, in equiva-lent doses to poudrage, may be both more effectiveand better tolerated than tetracycline. Topical analge-sia (200-250 mg lidocaine intrapleurally) is usefulwith additional opiate systemic analgesia when per-forming pleurodesis. The reported complication rateis around 5%.7 The clinical course is characterized byan exponential fall in drainage over 3-5 days. Whendrainage is below 100 mL day-1, the chest tube canbe removed. Pleurodesis may also be used in chronictransudative effusion of hepatic origin. However, thepresence of ascites carries a lower success rate. Theresponse rate is around 85% in the absence of ascites,but falls to 40% in patients with ascites.7

In otherwise refractory chronic transudative andexudative effusions, a pleuro-peritoneal Denvershunt can be useful. This may be inserted using localanaesthesia without important complications.Indications are failure of pleurodesis, particularlyin the presence of so-called 'trapped lung' andchylothorax. The device consists of a double-valvedpump with an afferent (pleural) and efferent (peri-toneal) tube. The pump is implanted subcutaneous-ly and operated by the patient or an assistant two tothree times a day. Although peritoneal seeding by


Figure 17.2 Schematic representation

of optimum drainage position and

venturi-ejector-operated, two-bottle suction

and fluid collection system.

Figure 17.3 Schematic

representation of an integrated

three-bottle suction and

fluid-collection system. The arrows

and bubbles indicate airflow

direction, both from the ambient air

(suction control system) and from the

pleural cavity. If no suction is

applied, the system operates as

a two-bottle, gravity-dependent,

combined water seal and

fluid-collection system. The

integrated manometer in the

fluid-collection system (C) allows

monitoring of the actual suction

level applied on the patient

( = primary suction control gradient

delivered by system A (cmH20) minus

height of the water seal column); it

also acts as a safety valve to positive

pressure.


Figure 17.4 Endoscopic view of thoracoscopic talc poudrage. Note

the homogeneous distribution of talc powder across the entire left

pariet al pleura, adhering lung upper lobe on the left upper margin,

free lower lobe at the left lower margin. (See also Plate 7.)

tumour cells occurs, the benefit of controlling thepleural collection offsets this risk in the short-termpalliation of symptoms.

EMPYEMA

Definition, pathogenesis and clinicalfeatures

Empyema is defined as a suppurative effusion due tobacterial infection. The most frequent cause isunderlying pneumonia. The recovery of obvious pusfrom the pleural cavity establishes the diagnosis.Serous or turbid effusions that are sterile are termedpara-pneumonic effusions. They may, however, pre-date the later development of empyema. Infectionaccounts for 20% of all pleural effusions, with bacter-ial pleuritis complicating pneumonia in 20-57%. In23% of cases, empyema is the result of surgery, andtrauma accounts for about 6%. Other causes includeoesophageal perforation (5%) pneumothorax (2%)and other miscellaneous causes. Empyema has a peakincidence in the middle-aged and elderly population.Predisposing morbidity has been reported in up to82% of patients, with alcohol abuse as the leading

risk factor in between 29% and 40% of patients,depending on case-mix. It is clinically important todistinguish:

• the early exudative stage, corresponding to apauci-cellular sterile effusion of 1 to several daysduration,

• the fibrino-purulent stage, representing classicalempyema with abundant leucocytes and bacteria,which progresses to the formation of fibrinousmembranes and loculations between 3 days and3 weeks,

• chronic empyema, characterized by organizingadherent peels encapsulating the lung and even-tually leading to rupture transcutaneously(empyema necessitans).

Empyema can take a highly variable course, froma well-preserved general status (silent empyema) tosevere septic shock, depending on pre-morbidity,antibiotic treatment, immune status, age andaetiology.

Empyema should be suspected if there is:

• persisting or unexplained fever after adequatelytreated pneumonia,

• persisting elevation of inflammatory markers(CRP,WBC,ESR),

• relevant pre-conditions, such as thoracic oroesophageal surgery or aspiration,

• suggestive faeculent sputum production (broncho-pleural fistula),

• imaging suggesting multi-loculations.

Diagnosis

IMAGING TECHNIQUES

The features distinguishing a simple effusion fromempyema are a multi-locular collection, membraneformation indicating thickening of the visceral andpariet al pleura and fibrin strands, but empyema mayalso be monolocular in about 16% of cases.Multi-loculation is not a specific indicator because itmay also occur with rheumatic and tuberculouspleurisy. Conventional X-ray may show a convexrather than the typical concave crescent of a freeeffusion and air within the pleural space, whichsuggests a bronchopleural fistula. Both ultrasonogra-phy and CT have limited sensitivities but highspecificity (around 96%) for the detection of pariet al

Empyema 227

and visceral pleural thickening suggestive of bacterialpleuritis. Contrast-enhanced CT is also importantin helping to distinguish empyema from lungabscess (which will not require thoracostomy) by thefollowing criteria:

Empyema:

• signs of lung compression,• smooth margins of membranes,• dissection of the thickened visceral and pariet al

pleura ('split-pleura' sign),• blunt angle with the chest wall,

Lung abscess:

• spherical shape with irregular thick-wallstructures,

• absence of lung compression,• sharp angle with the chest wall,• visible airway connection,• demonstration of vasculature around abscess

(definite proof).

Bronchoscopy can provide important contribu-tions to the diagnosis of empyema by:

• adding to the microbiological yield in patientswith underlying pneumonia,

• allowing definite proof and localization of a bron-chopleural communication,

• detecting potential causative conditions such asforeign-body aspirate, tumour or an oesophago-tracheobronchial communication.

Likewise, bronchoscopy may be indicated thera-peutically for clearing the airways of secretions oraspirate.

Pleural fluid analysisAlthough clinical features may suggest empyema,thoracentesis establishes the diagnosis. Thoracentesismay need to be performed at different locationsbecause, in loculated empyema and para-pneumoniceffusion, sterile fluid and pus may be found indifferent compartments. Pus is instantly recog-nizable by its appearance and/or the characteristicallyoffensive smell suggesting anaerobic infection. Con-fusion with chyle or pseudo-chyle may occur whenaspiration reveals odourless, whitish-turbid fluid.The parameters in Table 17.4 (referred to as Light'scriteria) unequivocally establish the diagnosis anddefine the stages of bacterial pleuritis from

• uncomplicated para-pneumonic effusion,• complicated empyema,• frank empyema.1,12

The determination of amylase (±isoenzymes)may be useful in oesophageal rupture or pancreatitis.Empyema and a complicated para-pneumonic effu-

Table 17.4 Pleural fluid analysis in para-pneumonic effusion and empyema: indications for tube drainage

Absolute indication for drainageFrank empyemaPositive bacterial culture (stain)

±Loculations

Biochemical characteristics:(Light's criteria)glucose <40 mg dL-1

LDH >1000IUL-1

pH < 7.00WBC> 15nL-1

Bronchopleural fistula

Drainage

Relative indication for drainageTurbid effusion ± bacterial culture (stain)Large fluid amounts (>2000 ml)

Loculations

Biochemical characteristics:glucose 40-60 mgdL-1

LDH < 1000 IU L-1

pH 7.00-7.20WBC10-15 nL-1

Serial determination + clinical follow-up

No indication for chest drainageSerous-clear effusionNegative bacterial culture (stain)Fluid amounts (< 2000 ml)No loculations

Biochemical characteristics:glucose > 60 mg dL-1

LDH < 1000 IU L-1

pH < 7.3 0WBC< 10 nL-1

Serial determination

Resolution


sion are characterized by low pH (<7.0) and glucose(<40 mg dL-1), a raised LDH (>1000 IU L- 1 ) andneutrophil count (>15 nL-1), whereas bacterial cul-ture will usually be positive. The clinical significanceof subdividing bacterial pleurisy (Table 17.4) is inmanagement: in uncomplicated para-pneumoniceffusion, a conservative approach is justified, where-as in empyema and complicated para-pneumoniceffusion, drainage is required. Serial determinationof Light's criteria is an adjunct to clinical observationin ambiguous cases (indeterminate effusion).

Microbiologic investigation should extend to cul-tures of blood, sputum and bronchoalveolar lavage(BAL). A wide range of bacterial isolates has beenreported (24-94%), obviously due to different ratesand intensity of antibiotic pre-treatment and to dif-ferent methods of sample collection and isolate cul-tivation (which is particularly true for anaerobes).Importantly, the causative aetiology is changing,with an increasing contribution of Gram-negativemicro-organisms, anaerobic isolates and multipleinfectious agents. In one series, comprising morethan 400 isolates from 336 patients, 56% were mono-infections and 44% multiple (up to four pathogens):46% Gram positive, 23% Gram negative and 21%anaerobes.13

Treatment optionsThe management of empyema involves both antibi-otics and drainage; surgical intervention may berequired. Parenteral antimicrobial therapy will needto take account of clinical features such as faeculent

fluid or pre-treatment. An empiric scheme is shownin Table 17.5. Therapy should be adjusted by micro-biologic isolation as soon as possible.

Tube drainage is indicated:

• with severe sepsis,• with large quantities of effusion (>2 L),• in the presence of air in the pleural space (indi-

cating bronchopleural fistula),• following Light's criteria-based fluid assessment

(see Table 17.4).

We favour the use of double-lumen catheters(Fig 17.5), with a diameter of at least 20 F, as thisallows closed-circuit, large-volume irrigation withnormal saline (± aseptic additives) and the option ofthe instillation of fibrinolytics if required. Irrigationis continued until clear sterile fluid is recovered andthe net fluid production falls below 50-100 mLday-1. The instillation of fibrinolytic agents (strep -tokinase, urokinase) is indicated if drainage fails toclear thick pus and/or membranes and loculations arepresent. The value of fibrinolysis has been shown incontrolled, prospective studies. 200 000-250 000 IUstreptokinase, or of an equipotent dose of urokinase(50 000-100 000 IU), is instilled once or twice daily.14

Our protocol is given in Table 17.6. Instillation of fib-rinolytic agents is usually required for 5-6 days.Failure, defined as persisting clinical features orultrasonography-demonstrated loculations after 2weeks, occurs in about 15% of cases.12,14 Similarresults have been claimed with the use of small-borecatheters (8-14 F).15 Contraindications to throm-

Table 17.5 Options for empiric antimicrobial therapy in empyema and para-pneumonic effusion

Community-acquired pneumonia

Nosocomial pneumonia

PneumococciStreptococcus spp.Staphylococcus aureusHaemophilus influenzaeLegionella spp.Anaerobes

Enterobacteriaceae spp.Pseudomonas spp.Staphlyococcus aureusAcinetobacterPeptostrepococcusBacteroidesFusobacterium

2nd or 3rd generation cephalosporinor Augmentin + clindamycin ormetronidazole + macrolide ifLegionella possible

3rd or 4th generation cephalosporin(e.g. ceftazidime) + aztreonam3

or carbapenem oracylamino-penicillin/tazobactamcombination

aAztreonam is recommended instead of aminoglycosides to avoid intrapleural inactivation.

Pneumothoraxand bronchopleural fistula 229

Figure 17.5 Schematic representation of double-lumen chest

tubes for irrigation and instillation therapy.

bolytics include bronchopleural fistula, significantcoagulation disorders or allergy (streptokinase). Prioruse of streptokinase (e.g. in acute myocardial infarc-tion) will not invalidate its use. The use of medicalthoracoscopy to break down lung parenchymal adhe-sions and loculations may be helpful. Instillation ofantibiotics has been suggested, with the rationale thatlow penetration of systemic antibiotic therapy mightprovide suboptimal therapy. However, with inflamedpleural membranes, concentrations well above theminimal inhibitory concentrations (MIC) have beendemonstrated in empyema fluid (except aminoglyco-sides) in many studies.

Surgical therapy (using VATS rather than formalthoracotomy) is indicated:

• when medical treatment fails - although early(e.g. 4-7 days) intervention may be indicated inseverely ill patients,

• when the empyema is encapsulated (empyemanecessitans), which maybe resected extrapleurallywithin the encapsulated empyema sack (empye-mectomy),

• when empyema is traumatic or postoperative,• for long-term open management (rib resection)

of a chronically infected pleural cavity.

Even with adequate antibiotic therapy and judi-cious medical or surgical management, empyemaremains a serious condition, with a mortality vary-ing from 6% up to 21%, depending on case-mix.The most important late sequelae is fibrothoraxwith associated impairment of pulmonary func-tion. This should occur in less than 10% of patients,in whom decortication would then be indicated.

Table 17.6 Large-volume irrigation (LVI) and fibrinolytic

therapy in the local treatment of empyema

Drainage

Irrigation

Fibrinolysis

Duration

Side effects

Contraindicationorcautions

Thoracoscopic/ultrasound-guideddouble-lumen trocar-catheter inser-ion, diameter 20-28 F, length 40 cm

1000 ml normal salinesolution + 20 ml2% polyvidone-iodine 1-2 times aday until clear irrigation fluidrecovered

200 000 IU streptokinase, tubeinitially clamped 4-8 hours

< 14 days

Fever (> 38 °C) in 42%, pain 10%

Bronchopleural fistula, allergy,previous thrombolysis for myocardialinfarction

PNEUMOTHORAX ANDBRONCHOPLEURAL FISTULA

Aetiology

Pneumothorax is a fairly common event in respira-tory intensive care. Its importance is related to thefacts that many cardiorespiratory conditions thatrequire ICU treatment can precipitate pneumo-thorax and that mechanical ventilation and otherinterventions entail both an increased incidence andcomplication rate. Pneumothorax is commonlydivided into traumatic and spontaneous pneumo-thorax, the latter occurring either without apparentpre-existing lung disease (primary pneumothorax)

230 Pleural disease

1M1 Subclavian region soft-tissue injury (soft-tissue emphysema)

^ Trauma to the trachea (mediastinal emphysema, soft-tissue emphysema)

£ Trauma to the bronchus (mediastinal emphysema, interstitial emphysema)

^ Alveolar rupture (interstitial emphysema)

Q Visceral pleura rupture (pneumothorax)

(§) Rupture of preformed bullae or blebs (spontaneous pneumothorax)

jgfc Trauma to the external chest wall and pariet al pleura (pneumothorax,:"* soft-tissue emphysema)

I Oesophagus rupture (mediastinal emphysema, soft-tissue emphysema)

Jfjt Entrance of abdominal air (mediastinal emphysema, pneumothorax)

Figure 17.6 Causes (mechanisms)

of air penetration to the pleura

and soft tissue and their

immediate sequelae.

or secondary to a condition involving structuraldamage to the lung, such as chronic obstructive lungdiasease (COPD) or interstitial lung disease.Important focal pleuro-pulmonary causes includepneumonia, lung abscess, neoplasms, tuberculosisand empyema. In respiratory critical care, secondaryspontaneous pneumothorax may be iatrogenic(traumatic), related to diagnostic and therapeutic

interventions, or may result from barotrauma inpositive pressure ventilation. Accidental pleuralinjuries during subclavian vein catheterization, thora-centesis, transthoracic needle aspiration and trans-bronchial biopsy are other important causes. Theadult respiratory distress syndrome (ARDS) andopportunistic Pneumocystis carinii infection are par-ticular risk factors, with a pneumothorax incidence

Pneumothorax and bronchopleural fistula 231

up to 60%; bilateral occurrence is not uncommon.Aspiration pneumonia (37%) and COPD (10%) areother important causes.1

Respirator-associated pneumothorax combinesthe features of secondary and iatrogenic (traumatic)pneumothorax. The risk of pneumothorax withpositive pressure ventilation has an overall incidenceof 4%. Positive end-expiratory pressure (PEEP)increases the risk to approximately 17%, with atwofold to fourfold further increase in risk withPEEP > 15 cmHp.1 Interestingly, in a large ARDSstudy, no relationship was found between peak air-way pressure and pneumothorax risk.16 However,only the initial period of ICU stay was reported andpneumothorax risk is greater in the later fibro-proliferative stage.

The most important aspect of ventilatory pres-sures is the difference between plateau pressure andthe level of PEEP as this determines tidal volume.The basic mechanism in respirator-associated pneu-mothorax is alveolar rupture due to increased shearforces resulting from inhomogeneity of lung path-ology, with resulting regional over-distension andrupture of alveoli. Therefore, respirator-associatedpneumothorax is more appropriately described asvolutrauma rather than barotrauma.

Alveolar rupture may also lead to mediastinalemphysema, bronchopleural fistula and soft-tissueemphysema. The complications and mechanismsinvolved in pneumothorax are schematically depict-ed in Figure 17.6. Efforts to minimize the risk aim tolimit pressure differences between inflation and expi-ration (small tidal volume) and avoid hyperinflation.The strategies for limiting airway pressure includethe use of spontaneous modes of ventilation, pres-sure rather than volume control (de-acceleratingflow) and tidal volume reduction, i.e. 'permissive'hypoventilation.16

Signs and diagnosis

CLINICAL SIGNS

It is important immediately to consider the possibil-ity of pneumothorax in any patient with a knownrisk and compatible symptoms and signs. In thenon-intubated patient, a sudden increase in dysp-noea and chest pain are the principal symptoms,with hypoxaemia and hypercapnia occurring in16-17% of cases, cyanosis in 9% and shock in

yo/0 1,17,18 in respirator-associated pneumothorax,the severity of cardiorespiratory compromise ismuch higher and may be disproportionate to the sizeof pneumothorax. The differential diagnosisincludes myocardial infarction, pulmonaryembolism and ruptured aneurysm. Classical signsare unilateral hyper-resonance and attenuated breathsounds (silent chest). An immobile hemi-thoraxwith central venous congestion and evidence of areduced cardiac output suggest tension pneumotho-rax. Tension pneumothorax may also induce electro-cardiogram changes due to air interposition ormediastinal displacement (e.g. dextrocardia).Ventilatory asynchrony in patients on mechanicalventilation may be recognized by a sharp rise of thepeak and plateau airway pressures, a fall in tidal vol-ume and concomitant deterioration of gas exchange.

RADIOLOGICAL SIGNS

The diagnosis is easily made in large and moderate-sized pneumothorax by standard X-ray. Problems ofrecognition may arise:

• with small pneumothorax and consolidated lung,• with localized or mediastinal pneumothorax,• in the presence of pre-existing bullous or general-

ized emphysema,• with air-fluid levels (confusion with intrapul-

monary cavities),• with superimposed chest-wall artefacts.

Postural, e.g. decubitus, views and follow-up filmsmay be helpful. Air-fluid levels, in the absence ofprior chest aspiration, indicate a pulmonary leak andmay be seen associated with hydrothorax, serotho-rax, pyothorax chylothorax and haemothorax.Standard X-ray films may also reveal additional signsof air in the mediastinum, pericardium or chest wall(surgical emphysema). The distinction betweenintrapulmonary and pleural air-fluid levels mayrequire CT scanning and is important for treatmentdecisions. .

ENDOSCOPIC DIAGNOSIS

As with pleural effusion, medical thoracoscopy cancontribute significantly to diagnosis and manage-ment and is only marginally more time consumingthan the standard therapeutic intervention of insert-ing a chest drain. The staging or classification systemproposed by Swierenga and Vanderschueren19 pro-

232 Pleural disease

Table 17.7 Endoscopic staging in spontaneous pneumo-

thorax (according to Vanderschueren)

Stage I PTX with endoscopically normal-appearinglung (40%)

Stage II PTX with pleuropulmonary adhesions thatmay be accompanied by haemothorax(12%)

Stage III PTX with small bullae and blebs (< 2 cm in

diameter) (31%)Stage IV PTX with numerous large bullae (> 2 cm in

diameter) (17%)

vides a precise description of the endoscopic findings(Table 17.7). We recommend the incorporation ofthoracoscopy into the routine management of pneu-mothorax, even in the mechanically ventilatedpatient. The advantages may be summarized as:

• better pleural and lung assessment than with CT,• visualization of complicating bronchopleural fis-

tula (with interventional option),• optimum drain placement,• ability to induce pleurodesis if required,• assessment of need for surgical intervention.19

Therapeutic options

In terms of therapeutic aims, the management ofpneumothorax is straightforward: re-expansion ofthe lung and prevention of recurrence. However, theimplementation of these aims is controversial.

RE-EXPANSION

Based on a calculated pleural gas absorption rate of1.2-1.8% day" l, the indication for draining or aspiratinga pneumothorax is somewhat arbitrarily set at 5= 15% ofthe hemi-thorax volume in uncomplicated primaryspontaneous pneumothorax. In secondary spontaneouspneumothorax and respirator-associated pneumotho-rax, however, treatment will depend more on the imme-diate physiological consequences or the danger of thesubsequent development of a large pneumothorax in theventilated patient. Simple aspiration may be sufficient inspontaneous pneumothorax, but the success rate(defined as complete lung expansion maintained for atleast 1 month) is only 48-85% in primary and 31-80%in secondary spontaneous pneumothorax.20 In the criti-

cal care patient, tube drainage (> 24 F) is required, par-ticularly in the presence of a concomitant effusion(serothorax, pyothorax, chylothorax, haemothorax),with reported success rates of 96% in primary and 92%in secondary spontaneous pneumothorax.1,18

Low suction (10-20 cmH2O) will usually suffice toachieve full lung expansion. Drainage times usuallyvary from 3 to 7 days. Importantly, the chest tubeshould not be removed prematurely. We recommenda probationary period of clamping (12 hours) beforeremoval of the drain, but this is controversial and thepatient should be closely observed. Failure of expan-sion to low suction pressures or early recurrence(< 7 days) indicates one of the following:

1. consolidated or atelectatic lung,2. reduced lung compliance (stiff) or trapped lung,3. loculated pneumothorax with adhesions,4. pulmonary air leak (bronchopleural fistula),5. technical problems such as kinked or plugged

tube, water in the filter or excessively long con-necting tubes reducing or preventing the creationof a negative pressure within the pleural cavitythroughout the respiratory cycle.

Combinations of these complications may alsooccur and require specific management strategies.With consolidated or collapsed lung, pneumoniamay be specifically treated and bronchoscopy mayallow occluding mucous plugs to be removed. Withstiff or trapped lung, re-expansion may eventuallyoccur with prolonged suction at higher pressure levelsand, as previously outlined, thoracoscopy may behelpful in removing fibrin or membranes when theselimit re-expansion. Suction levels >20 cmH2O inmal-expansion due to trapped lung must be appliedin a cautious, incremental fashion and with carefulobservation of the patient, because negative-pressuretransmission to the mediastinum may cause painand an ipsilateral mediastinal shift with centralvenous obstruction.

In loculated pneumothorax, some compartmentsmay be inadequately drained and multiple drainsmay be required. Bronchopleural fistula will beapparent by flow of air through the drainage tube,which may be continuous (large fistula) or dis-continuous (small fistula). Air leakage may varysignificantly, from < 1 to 16 L min"1. In the ventilatedpatient, quantitative assessment is easily possible bysubtracting expiratory minute volume from inspira-

References 233

tory volume. Tube size may become critical, becausean internal diameter of at least 4.72 mm (= 20 Fdrainage) is required to accommodate a flow of10 L min-1 at a standard suction level of 10 mmHg,and of 5.87 mm (= 24 F drainage) to accommodate15 L min-1 air flow.20 Attempts to re-expand thelung using higher suction levels may merely increasethe leak and it may be difficult to achieve adequateoxygenation, although the respiratory 'steal' willeliminate CO2.

If the lung can be re-expanded, the air leak usu-ally ceases on re-institution of pleural contact andthis can usually be achieved with suction if the leakis not too large. In large pleural leaks, it may be bet-ter to avoid suction, with the drainage system actingmerely as a safety valve to prevent tension pneu-mothorax. Lung expansion may then be attemptedafter closure of the defect. The management ofbronchopleural fistula may require a variety ofmeasures (summarized in Table 17.8)21, such asreduced ventilatory pressure and attempts to closethe leak bronchoscopically using fibrin sealant(which has, in our experience, a 50% success rate)or via thoracoscopy employing cautery or talc pleu-rodesis. Unilateral or differential bilateral ventila-tion can be tried, but is rarely successful and can betechnically difficult. High-frequency jet ventilationand high-frequency oscillation therapy are ofteneffective in the temporary stabilization of the diffi-cult patient, but surgical closure may still berequired when/if the patient survives the acute peri-od of respiratory failure.

DEFINITIVE THERAPY: PREVENTION OFRECURRENCE

The recurrence rate in secondary spontaneous pneu-mothorax is reported to be as high as 54%.1'20

Recurrence is most likely to occur in the immediatepost re-expansion period, which is particularly rele-vant to the intensive care patient. Pleurodesis shouldbe performed only after a recurrent pneumothoraxin primary spontaneous cases, but should be serious-ly considered after the first episode in secondaryspontaneous and respirator-associated pneumo-thorax. This is particularly true in the presence ofthoracoscopy-demonstrated stage II or III pleuro-pulmonary changes and in the elderly patient (>50years). The technique of pleurodesis is the same asfor chronic pleural effusion, with the exception that

Table 17.8 Ventilator-associated pneumothorax

Barotrauma preventative strategiesInterstitial lung disease

BiPAP or CPAP/PCV rather than VCV ventilationInverse ratio ventilation with Pawmax < 35 or PEEP

max

<15cmH20Permissive hypercapniawith V<6 r n L k g - 1

Obstructive airway diseaseLow minute volumeAvoiding further hyperinflation bylimiting PEEPe < PEEPi

Recognition of PTXSudden increase in Pawmax and Pawplateau

and deterioration in gas exchangeTreatment of PTX with bronchopleural fistula

Large-bo re drain(s)Postural manoeuvres: diseased side downMonitor size of fistula leak: = inspiratory minusexpiratory minute volumeLow level suction with adjustment to minimize leakReduce tidal volume, PEEP and inspiratory timeAttempt to close fistula by occlusion of air leak withfibrin or thoracoscopic cautery or talcHigh-frequency oscillation ventilationDouble-lumen intubation for unilateral or differentialbilateral ventilation

continued pleural drainage is not relevant. The long-term success rates (up to 95%) with talc poudrage inpneumothorax closely approach those of surgicalprocedures such as VATS or formal thoracotomy andpleurectomy (98%).1,2° Talc slurry may be a less-favourable agent due to unequal distribution, buttetracycline has been reported to produce long-termcontrol in 84% of cases and thus remains a usefulalternative.1

REFERENCES

1. Light, RW. Pleural diseases, 3rd edition, Williams and

Wilkins, 1995.

2. Yang, PHC, Luh, KT, Chang, DB, et al. Value of

sonography in determining the nature of pleural

effusion: analysis of 320 cases. Am J Radiol 1992; 159:

29-33.

3. Verschakelen, JA. Spiral CT of the chest: diaphragm,

chest wall and pleura. In Spiral CTof the chest, ed.

234 Pleural disease

M Remy-Jardin, J Remy. New York: Springer, 1996;

305-19.

4. Bittner, RC, Felix, R. Magnetic resonance (MR) imaging

of the chest: state of the art. Eur RespirJ 1998 ; 11:

1392-404.

5. Burgess, LJ, Maritz, FJ, Taljaard, FFJ. Comparative

analysis of the biochemical parameters used to

distinguish between pleural transudatesand

exudates. Chest 1995; 107: 1604-9.

6. Sahn, SA. The diagnostic value of pleural fluid analysis.

Semin Respir Crit Care Med 1995; 16(4): 269-78.

7. Loddenkemper, R, Frank, W. Pleural effusion,

nemo-thorax, chylothorax. In Pulmonary diseases,

ed. C Grassi. McGraw-Hill International, 1999; 41:

391^04.

8. Valdes, L, Alvarez, D, Sanjose, E, et al. Value of

adenosine deaminase in the diagnosis of tuberculous

pleural effusions in young patients in a region of high

prevalence of tuberculosis. Thorax 1995; 50: 600-3.

9. Ferrer, J. Pleural tuberculosis. Eur RespirJ 1997;

10: 942-7.

10. Loddenkemper, R. Thoracoscopy: state of the art. Eur

RespirJ 1998; 11: 213-21.

11. Walker-Renard, PB, Vaughan, LM, Sahn, SA. Chemical

pleurodesis for malignant pleural effusion. Ann Intern

Med 1994; 120: 56-64.

12. Hamm, H, Light, RW. Parapneumonic effusion and

empyema. Eur RespirJ 1997; 10: 1150-8.

13. Frey, DJM, Klapa, J, Kaiser, D. Irrigation drainage

and fibrinolysis in the treatment of

parapneumonic pleural empyema. Pneumologie

1999; 53: 583-642.

14. Bouros, D, Schiza, S, Patsurakis, G, et al. Intrapleural

streptokinase versus urokinase in the treatment of

complicated parapneumonic effusions: a prospective,

double-blind study. Am J Respir Crit Care Med 1997;

155:291-5.

15. Sahn, SA. Management of complicated

parapneumonic effusions. Am Rev Respir Dis

1993; 148: 813-17.




and the acute respiratory distress syndrome. N Engl

J Med 2000; 342: 1301-8.

17. Shields, TW, Oilschlager, GA. Spontaneous

pneumothorax in patients 40 years of age and

older. Ann ThoracSurg 1966; 2: 377-83.

18. Seremetis, MG. The management of spontaneous

pneumothorax. Chest 1970; 8: 57-65.

19. Vanderschueren, RG. The role of thoracoscopy in the

evaluation and management of pneumothorax. Lung

1990; Suppl., 1122-5.

20. Baumann, MH. Treatment of spontaneous

pneumothorax, a more aggressive approach?

Chest-\997; 112:789-804.

21. Baumann, MH, Sahn, SA. Medical management

and therapy of bronchopleural fistulas in the

mechanically ventilated patient. Chest 1990;

97: 721-28.

18Acute interstitial lung diseaseRICHARD MARSHALL

Introduction 235

Terminology 235

Making the diagnosis 236

Therapy 238

Acute interstitial pneumonia 239

Cryptogenic organizing pneumonia 240

Eosinophilic pneumonia

Drug-induced interstitial lung disease

Acute vasculitic lung disease

Diffuse alveolar haemorrhage

Summary

References

242243246246247247

INTRODUCTION

The term interstitial lung disease (ILD) encompassesa group of disorders in which there are varyingdegrees of inflammation and fibrosis in the inter-stitial space and distal airway. Many of these disor-ders present with a progressive decline in respiratoryfunction, permitting management in the out-patientsetting. Less commonly, they present acutely withmore severe respiratory failure and may reach theattention of critical care physicians. Their incidencein the intensive care unit (ICU) is largely unknown,but it is likely that a number of cases are misclassifiedas either diffuse infection or acute respiratory dis-tress syndrome (ARDS), with which they share manyfeatures.

ILD often generates understandable nihilismamongst clinicians, the result of complex classifica-tion systems and a perceived paucity of therapeuticoptions, which can act as a deterrent to furtherunderstanding and investigation. In this chapter, apragmatic approach is adopted to both the classifica-tion and management of acute ILD. In particular,although often essential for accurate diagnosis, suit-able biopsy material is seldom available in the ICUsetting, and an emphasis on pathological description

is avoided as far as possible. Similarly, a detailed dis-cussion of potential pathological mechanisms maybe found elsewhere.1

Clinical suspicion is the key to diagnosis and mustremain high if these conditions are not to be missedand delays in administering appropriate therapy areto be avoided.

TERMINOLOGY

The study and management of ILD suffer from anover-complex and changeable classification system.ILD presenting acutely may broadly be considered asinterstitial pneumonias, organizing pneumonias ordiseases primarily of the small airways. By this defin-ition, ARDS itself is a form of ILD, but is consideredelsewhere (see Chapters 11 and 12).

Interstitial and organizing pneumonias areparenchhymal lung diseases characterized by amononuclear cell and proteinaceous infiltrate distalto the terminal bronchiole. All have the capacity toprogress to established fibrosis to a varying degree.1

A number of subtypes have been classified on thebasis of histological appearance. At present, it is use-ful to think of the following clinical disease entities.2

236 Acute interstitial lung disease

• Usual interstitial pneumonia (UIP): a subtype ofcryptogenic fib rosing alveolitis (CFA), also knownas idiopathic pulmonary fibrosis (IFF). This is aprogressive, fibrotic lung disease that may rapidlyaccelerate in a minority of patients. Generally,there is a poor response to current therapy.

• Desquamative interstitial pneumonia (DIP) andnon-specific interstitial pneumonia (NSIP): furtherCFA subtypes that appear to have a better prog-nosis than UIP.

• Diffuse alveolar damage (DAD): this is the charac-teristic histological pattern seen in ARDS (seeChapter 11).

• Acute interstitial pneumonia (AIP): pathologicallyvery similar to ARDS, but idiopathic. It was previ-ously known as the Hamman-Rich syndrome.

• Organizing pneumonia (OP): a pathologicaldescription in which small airway and alveolargranulation tissue predominates. It has a numberof known causes, particularly drug toxicity andcollagen vascular disease, and is sometimesreferred to as bronchiolitis obliterans organizingpneumonia (BOOP) in this context. A cryptogenicform (COP) is described. OP usually demon-strates an excellent response to corticosteroids.

MAKING THE DIAGNOSIS

The presentation of acute ILD is typified by thedevelopment of respiratory failure in the presence ofwidespread radiological opacification on chest X-ray(CXR). The diagnosis must often be made in theabsence of histological analysis or specific diagnostictests, and is more often based upon a composite ofclinical, radiological and laboratory features. In suchcases, the diagnosis is one of exclusion and, in par-ticular, differentiating acute ILD from pulmonaryinfection and ARDS. It is because of this overlap thatthe presence of non-ARDS ILD may be overlooked.A classification system for ILD based upon aetiologyis perhaps more useful in this context (Table 18.1).

Clinical features

Obtaining an adequate history remains vital. A moreprolonged symptomatic period prior to hospitaladmission is an important clue and includes fever,

malaise and anorexia, which may precede pulmonarysymptoms by weeks or even months. The presenceof non-pulmonary symptoms, clinical evidence ofsystemic disease (such as with vasculitis) and unusualradiological features should also alert suspicion(Table 18.1). Pulmonary signs, including the ubiqui-tous coarse crepitations, are non-specific. In practice,it may be the failure of antimicrobial therapy thatprompts the search for ILD. An approach to diagno-sis is outlined in Figure 18.1.

Radiology

Plain films should be reviewed regularly with a radi-ologist. The pattern of opacification, its distributionand changes over time are important. A computer-ized tomography (CT) scan should also be per-formed where possible to help to resolve the patternof disease (i.e. reticular, alveolar or bronchocentric)and to uncover unexpected pathology such as malig-nancy, lymphadenopathy, pre-existing ILD, pleuraleffusions or pneumothoraces. Occasionally, prog-nostic information may also be obtained.

Radionucleotide imaging is not usually possible inmechanically ventilated patients, but may be con-sidered in the less severe cases. 99m TcDTPA clear-ance, where available, is superior to 67Ga scanning inthe detection of an interstitial inflammatory processat early stages, being a more sensitive index of epithe-lial integrity. The results are usually non-specific, butmay have prognostic value, at least in chronic formsof ILD.2

Bronchoalveolar lavage

Although sampling secretions from the lowerrespiratory tract is routine in ventilated patients,more formal assessment of alveolar lining fluid bybronchoalveolar lavage (BAL) is variable in its use.This may partly depend upon local experience of thetechnique, but perhaps more so on its perceivedusefulness and safety. Competency on the part ofthe operator is vital to ensure safety and improvediagnostic yield. Although it is not always possible toextrapolate results published by large centres togeneral ICU practice, the safety of BAL has beendemonstrated repeatedly in the literature,3 and itsuse should be considered early in the diagnostic

Making the diagnosis 237

Table 18.1 Classification of interstitial lung disease by cause

Trauma/sepsis/haemorrhage/surgery/burns etc.

Infection

Pulmonary oedema

Allergy

Drug induced

Inhalation injury

Pulmonary haemorrhage

Vasculitis

Idiopathic

ARDS

MiliaryTB, RSV, CMV

HIVInvasive aspergillosisCardiac failureRapid lung re-expansionFluid overloadEosinophilic pneumonia

Amiodarone lungCytotoxic agentsSmokeChlorineWegener's granulomatosisGoodpasture's syndrome, Behcet's

syndromeAcute SLEAcute SLERheumatoid pneumonitis

Eosinophilic pneumoniaAcute interstitial

pneumoniaCOP

Presence of immunosuppressivefactors

Serology

Presence of left ventricularimpairment (echocardiography,previous history etc.)

History, Eosinophilia on BALPeripheral, migratory opacification

onCXR

HistoryHistoryBronchoscopic appearanceHistoryPresence of systemic diseaseSerologyRadiological appearancePresence of systemic diseaseSerologyProdromal symptomsProdromal symptomsAbsence of causative agents

ARDS, acute respiratory distress syndrome; TB, tuberculosis; RSV, respiratory syncytial virus; CMV, cytomegalovirus; HIV, human immunodeficiency virus; BAL,bronchoalveolar lavage; CXR, chest X-ray; SLE, systemic lupus erythematosus; COP, cryptogenic organizing pneumonia.

pathway. BAL will aid the exclusion of a typicalinfection, but may also yield specific diagnosticfeatures, e.g. eosinophilia.

non-specific and blind immunosuppressive therapyis contemplated, a biopsy should be stronglyconsidered.

Biopsy Additional tests

Trans-bronchial biopsy is to be avoided on the ICU.Its diagnostic yield in ILD is usually poor and it car-ries a significant risk of morbidity. Open lung biop-sy is clearly not available to most intensive carephysicians, yet, as with BAL, published series attest toits usefulness and safety.4 Many issues surround thedecision to perform an open lung biopsy, not least itstolerance by the patient, the potential to precipitatethe need for intubation, and the impact of the infor-mation obtained on management decisions. As ageneral rule, if clinical and radiological features are

An assessment of immunological markers may be ofgreat help, including: microbial antigens, anti-nuclear antibodies (ANAs) and anti-cytoplasmicantibodies (ANCAs), helping to exclude infectionand identify autoimmunity. Evidence of systemicdisease should also be sought, including an examina-tion of the urine for proteinuria, and casts, an assess-ment of the peripheral blood film and furtherradiological imaging of affected tissues. A markedlyelevated or normal erthrocyte sedimentation rate orC-reactive protein may steer the clinician towards


Figure 18.1 Diagnostic algorithm

for acute interstitial lung disease.

ARF, acute respiratory failure; BAL,

bronchoalveolar lavage; CT,

computed tomography.

systemic disease or infection, but both are non-spe-cific. They may be of use in monitoring the course ofthe disease and the response to intervention.

THERAPY

Once infection is excluded, as far as is possible, treat-ment falls into two main categories: supportive andimmunosuppressive (there are currently no therap-

ies that effectively target fibroproliferative pathways).Acute ILD is uncommon and the evidence support-ing specific therapy is poor. There have been no ran-domised, controlled trials.

Setting the ventilator

Given the clinical and pathological features thatmany causes of acute ILD share with ARDS, it would

Acute interstitial pneumonia 239

seem reasonable to adopt a comparable approach toventilator management. Ventilator-induced lunginjury is likely to be an issue for all patients withsevere respiratory failure, and the avoidance of hightidal volumes (6-10 mL kg-1) coupled with a levelof positive end-expiratory pressure (PEEP) titratedto FiO2 would seem prudent. Permissive hypoxia(8-10 kPa) and hypercapnoea (6-10 kPa) may alsobe preferable, although the evidence for either isweak at present. Interestingly, hyperoxia rapidlyinduces lung injury in vivo5 and may interactsynergistically with NO.6 Although there is littledocumentation of its effects in humans, it at leastsuggests that limiting FiO2 may be theoreticallybeneficial.

The role of other ventilatory strategies, such asrotational therapy and nitric oxide (NO), is unclear.Neither has a sufficient evidence base to be recom-mended unequivocally, but their use in patientswith severe hypoxaemia seems as pertinent to otherforms of acute respiratory failure as it is to ARDS.Their use will also depend on the availability ofsuitable resources and expertise in individualcentres.

Immunosuppression

The use of immunosuppressive agents remains themainstay of treatment for all ILD. In the absence ofinfection, these should only be given blindly whena biopsy is not possible or has been non-diagnostic,and only when all other serological tests have beenrequested.

Again, the evidence from randomised, controlledtrials for any therapy is weak at best andlimited to patients with less severe disease thanwill be encountered on the ICU. High-dose systemiccorticosteroids form the mainstay of initial therapy.Methyl-prednisolone has less mineralocorticoidactivity, dexamethasone requires a smaller volumefor administration, but otherwise there is littleevidence to support the use of one agent over theother. Other immunosuppressants, e.g.cyclophosphamide and azathioprine, may be addedas steroid-sparing agents in prolonged therapy ormay be specifically indicated, i.e. in Wegener'sgranulomatosis.7

ACUTE INTERSTITIAL PNEUMONIA

Clinical features

Hamman and Rich, at the Johns Hopkins Hospital,described three cases of 'acute diffuse fibrosis of thelung' over a period of 3 years. They had to wait 10years for a fourth case before publishing their reportin 1944. These initial cases were acute in onset andproduced a rapidly fatal course. Although theHamman-Rich syndrome has been synonymous withIFF in the past, it is no longer a useful term and isnow referred to as AIP, which it resembles clinicallyand radiologically.8

Confusion occurs as to its relationship withARDS. AIP in reported cases occurs at a mean ageof 49 years (range 7-83). It presents with a rapiddeterioration, usually over a few days; however, incontrast to ARDS, a flu-like prodrome is characteris-tically present and no predisposing cause is found.AIP is thus, by definition, idiopathic. In addition,although almost all patients require mechanical ven-tilation, one does not observe the profound systemicinflammatory response and multi-organ failuremore typical of ARDS.

Pathology

Histologically, AIP is characterized by a patternknown as DAD, which is also seen in ARDS (seeChapter 11). An initial profound exudation of fluidand protein into the lung accompanied by neutro-philic infiltration is followed by intense fibroprolifer-ation and, subsequently, established fibrosis. Hyalinemembranes consisting of organized fibrin clot aretypical. These changes are spatially uniform, whichmay distinguish AIP from the more intra-alveolarand patchy distribution seen in OP.9

Investigations

Radiologically, widespread infiltration of the lung isseen, with ground-glass opacification and consolida-tion in the absence of the cystic changes and grossparenchymal distortion seen in more chronic formsof ILD (Fig. 18.2).


Figure 18.2 Acute interstitial pneumonia. This 34-year-old man

presented with a 3-day history of flu-like symptoms. 5evere

respiratory failure ensued. A thorough screen for pulmonary

infection (including bronchoalveolar lavage) was negative.

He later died despite immunosuppressive therapy.

this poor prognosis, the early use of corticosteroidsis probably justified, provided infection and ARDScan be excluded. Few data exist, but intravenousprednisolone (250 mg day-1), cyclophosphamide(1.5g day-1) and vincristine (2 mg) have beenreported to halt the otherwise rapid progression ofthe disease.11

Lung-protection strategies, including low tidalvolume ventilation and high PEEP, are also likely tobe beneficial in AIP given the pathological similaritieswith ARDS. One case of lung transplantationhas been reported in which there was an improve-ment in the native lung. This is probably the result ofa more intense immunosuppressive regimen and hasalso been reported in patients with IPF.

In summary, AIP could be considered as an idio-pathic form that mimics ARDS, but which may bemore susceptible to immunosuppression early in itscourse and which carries a worse prognosis. Thepresence of a prodromal syndrome and a clinicalpicture of ARDS for which no predisposingcause can be identified are the most useful dia-gnostic features.

CT scan appearances mirror those found on theCXR. It is most useful in the early stages of illness todifferentiate AIP from an acute acceleration ofunderlying IPF, which may mimic the condition.Honeycombing or traction bronchiectasis in theearly phase strongly suggests the latter, althoughtheir presence has been described in late AIP, butonly after many weeks of illness. The differentialdiagnosis also includes diffuse infection, lymphopro-liferative infiltration, OP, pulmonary vasculitis andtoxic pulmonary reactions to drugs or inhaledagents. Although OP cannot be easily be distin-guished radiographically from AIP, it is generallycharacterized as a patchy, migratory infiltrate withmore diffuse involvement of all lobes.

Additional tests such as serology, erythrocyte sedi-mentation rate (ESR), C-reactive protein (CRP) andBAL, although likely to be abnormal, are not specificto AIP, but are helpful to exclude infection and othercauses of ILD.

Treatment and prognosis

The reported mortality is 78%, although, as withARDS, this may have improved recently.10 Given

CRYPTOGENIC ORGANIZING PNEUMONIA

OP is a histological feature of pulmonary inflamma-tion, comprising buds of granulation tissue fillingthe alveolar space and terminal bronchiole. Thereare varying degrees of accompanying interstitialinflammation and fibroproliferation. OP may have adefinable cause (e.g. infectious pneumonia,bronchiectasis, IPF, drug reactions) or occur in thecontext of another disease (e.g. collagen vasculardisease, ulcerative colitis, leukaemia).12 Here, we areconcerned with the cryptogenic form. The termBO OP is probably best avoided because it is easilyconfused with bronchiolitis obliterans, which pre-dominantly affects small airways and presents withairways obstruction.

Clinical features

A subacute presentation is usual, which will immedi-ately distinguish COP from ARDS and AIP. The meanage at presentation is 55-60 years, but it has beendescribed in all adult age groups. A flu-like illnesscoupled with non-productive cough, chest pain and

Cryptogenic organizing pneumonia 241

arthralgia is the typical picture. In severe cases, thedifferential diagnosis includes ARDS, AIP, acuteeosinophilic syndromes and vasculitic lung disease.Importantly, there is no evidence of systemic disease,and immunological testing will be non-specific in COP.Thus, these investigations are of value only in exclud-ing underlying autoimmune and collagen vasculardisease.

Radiology

Migrating peripheral air-space opacification is char-acteristic of COP, and an examination of sequentialCXRs is crucial (Fig. 18.3). Other patterns includediffuse bilateral infiltration, a single opacity (whichmay cavitate and which is usually associated witha more prolonged illness) and, less commonly,crescentic opacities, subpleural bands or even apneumatocoele. A CT scan can be invaluable in estab-lishing the diagnosis.

Lung function and histology

The prime abnormality on lung-function testing isa restrictive pattern. Severe hypoxia results fromright-to-left shunting and the alveolar-arterial O2

difference will be increased on 100% O2. BAL changesare usually non-specific.

If clinical and radiological features are unhelpful, abiopsy is indicated. This is largely because of theexcellent response of this condition to corticos-teroids, which can be started earlier in the presence ofa histological diagnosis. A biopsy will also help ensureagainst the potential hazards of immunosuppressionin infectious disease. Granulation tissue fills the alve-olar space, often 'budding' into adjacent alveoli viathe pores of Kohn. Of note, bronchiolitis obliterans isclearly bronchocentric by comparison. A diagnosis ofCOP also demands that the granulation tissue is thepredominant lesion and not merely in associationwith other features of vasculitis or granuloma. Insevere cases, large areas of lung tissue will be involved,but care should be taken to ensure imaging is per-formed immediately prior to biopsy, due to themigratory and transient nature of the lesions.

Figure 18.3 Cryptogenic organizing pneumonia. This 56-year-old

man complained of progressive dyspnoea over a 2-week period. His

condition deteriorated despite broad-spectrum antibiotics. There

was no previous history of drug exposure. Organizing pneumonia

was diagnosed by open lung biopsy and responded to cortico-

steroids. No aetiological agent was identified.

Therapy and clinical course

Occasional spontaneous improvement occurs, butthis is unlikely in severe cases. Generally, patientsrespond rapidly to corticosteroid therapy (e.g. 0.75mg kg-1 day-1 prednisolone). Some show improve-ment in hypoxaemia within 48 hours and almost allwill have improved within 7 days. Although COP thatis refractory to corticosteroids has been reported, it ismore likely that these cases represent alternative diag-noses, including ARDS, AIP or a rapid acceleration ofIFF. In severe, biopsy-proven COP, one to three intra-venous boluses of cyclophosphamide may also beconsidered.

Treatment is usually required for between 6 and 12months. Corticosteroids should be tapered after 1-3months, depending on the speed of resolution.Relapses occur in some individuals and the dose mayhave to be temporarily elevated or prolonged in suchinstances. The prognosis of COP with patchy alveo-lar opacities is excellent providing the diagnosis ismade. On the ICU, diagnostic delay and secondarypathology - notably infection - are potentialhazards.


EOSINOPHILIC PNEUMONIA

The term 'pulmonary eosinophilia' is commonly usedto describe any condition in which pulmonary opa-cities are associated with a peripheral eosinophilia.This is misleading because it includes conditions inwhich there is a blood eosinophilia and an increasedsusceptibility to non-eosinophilic pneumonia. Theterm eosinophilic pneumonia is preferable becauseit implies the presence of pulmonary rather thanperipheral eosinophilia. The eosinophilic pneu-monias are classified on the basis of cause and thelength of the clinical presentation into simple, acuteand chronic forms.

The term simple eosinophilic pneumonia shouldnow be used to describe the disorders originallyencompassed by Loeffler's syndrome, which wascharacterized by peripheral eosinophilia and pul-monary opacities and was either idiopathic orcaused by a variety of agents including drugs andparasites. Such cases are usually mild and transient.

Acute eosinophilic pneumonia (AEP) is arelatively new entity, first described in 1989 andincreasingly reported in the literature.13 Charac-teristically, it affects younger individuals (mean age30 years). A number of reports from Japan describea strong association with tobacco-smoke inhalation,but only a proportion of patients are smokers.The severity is variable and, from the list of otheragents (Table 18.2) implicated in its aetiology, thereis clearly potential overlap with simple eosinophilicpneumonia. However, the short clinical course andabsence of recurrence in AEP are distinguishingfeatures. Radiologically, more peripheral, migratoryair-space opacification may help differentiate AEPfrom other ILD, but the distribution may be variable(Fig. 18.4).

Table 18.2 Factors associated with acute eosinophilic

pneumonia

Tobacco smoking

Heroin smoking

Ranitidine

Pentamidine

Aspirin

Carbamazepine

Clomipramine

Venlafaxine

Trazadone

Minocyline

Figure 18.4 Acute eosinophilic pneumonia (AEP). This chest radi-

ograph demonstrates peripheral, patchy, air-space shadowing typi-

cal of AEP. This 64-year-old woman presented with a 2-day history

of fever and dyspnoea and deteriorated rapidly after this chest X--

ray was obtained. Subsequently, bronchoalveolar lavage revealed

38% eosinophilia. Corticosteroid therapy led to a full recovery.

Clinical course and treatment of acuteeosinophilic pneumonia

The presentation is typically less than 1 month induration, with fever and malaise being the promin-ent symptoms. Peripheral blood eosinophilia is vari-able, but BAL eosinophilia is generally diagnostic,with the demonstration of increased cellularity and20-50% eosinophils. The histological features arethose of DAD with profound eosinophilic infiltra-tion. Histologically diagnosed AEP without BALeosinophilia has been reported.

The prognosis is excellent, with a good clinicalresponse to steroid therapy. Treatment should initiallybe with methylprednisolone at 1 mg kg-1 every6 hours for 2-3 days, followed by prednisolone40-60 mg day- l tapered over 4-6 weeks. Recurrencehas not been reported to date.

Chronic eosinophilic pneumonia presents withgeneral malaise, weight loss and fever. Investigationsreveal a marked peripheral blood eosinophilia witha polymorph leucocytosis, anaemia and a raised ESR.The cause is unknown, but there is a good response

Drug-induced interstitial lung disease 243

to corticosteroid therapy. Admission to an ICU israrely necessary.

DRUG-INDUCED INTERSTITIAL LUNGDISEASE

The list of agents associated with pulmonary ILDcontinues to expand. The more common associ-ations are considered here by way of example, particu-larly those that may present with more severerespiratory failure.

General considerations

Clinical signs, radiological features and lung-func-tion tests are non-specific and thus clinical suspicionis the main diagnostic prompt. A detailed history andthorough search of the medical notes should bemade to identify agents associated with pulmonarysyndromes. Symptoms may develop weeks, monthsor even years after drug administration and after thedrug responsible has been stopped. Changes in dif-fusing capacity may precede the onset of symptomsand radiographic changes by days or weeks, at whichtime the agents are discontinued to halt progression.BAL is useful primarily to exclude infection, but mayalso reveal eosinophilia. Once suspected, a thoroughsearch for infection should be made and all poten-tially responsible drugs should be stopped and, ifclinically indicated, an alternative agent substituted.

Most patients will recover from drug-inducedpulmonary disease, although, in a minority of cases,respiratory symptoms will persist or even progressafter drug withdrawal. Treatment for these individualsis largely supportive.

Pathology

Drug reactions present with a variety of pulmonarypathologies, including interstitial pneumonia, OP,eosinophilic pneumonia and diffuse reticular fibrosis(Table 18.3). These different patterns may occur inresponse to the same drug.

The mechanisms of drug-induced lung injury arepoorly understood. Sequestration or metabolism ofthe drug in the pulmonary circulation/parenchymamakes toxicity more likely. Although severe reactions

Table 18.3 Patterns of acute pulmonary drug reactions

Pattern

Interstitial pneumonitis

Organizing pneumonia

Interstitial fibrosis

Eosinophilic pneumoniaPulmonary oedemaDiffuse alveolar

haemorrhage

Typical agents implicated

Methotrexate, nitrofurantoin,(3-blockers

Amiodarone, bleomycin,cytotoxics

Bleomycin, amiodarone,cytotoxics

Antibiotics, NSAIDS, cytotoxicsAspirin, opiates, drug overdoseQuinidine, thrombolytics,

anticoagulants

NSAIDS, non-steroidal anti-inflammatory drugs.

are not common, subclinical evidence of toxicity canbe found in a much larger proportion of individuals.This suggests modifying factors that influence theprogression from subtle injury to clinically manifestdisease. Hypotheses concerning the mechanisms ofinjury include direct cellular toxicity, immunologicalreaction, redox imbalance, phospholipidosis (amio-darone), apoptosis and DNA scission.

Chemotherapeutic agents

The increasing use of cytotoxic therapy, particularlybleomycin, cyclophosphamide and methotrexate, hasled to an increased incidence of pulmonary drug tox-icity. Cumulative dose and concomitant therapy arefactors that have been associated with an increasedsusceptibility to pulmonary drug reactions and,although genetic factors are proposed, none has beendescribed to date. Interestingly, cyclophosphamidehas been used to treat not only non-drug inducedILD but also methotrexate and amiodarone pneu-monitis, highlighting the complex interactionbetween external and internal environments.

Fever, a dry cough and progressive dyspnoea overdays or weeks may precede more severe respiratorycompromise. In the immunocompromised patient,the major differential diagnosis is infection. Patientspresenting with respiratory failure in the absence ofneutropenia should alert particular suspicion.Diffuse malignant infiltration by the underlying dis-ease is also a possibility, but an open lung biopsy isnecessary to confirm the diagnosis. The diagnosticyield is approximately 50% for malignancy-associatedpulmonary infiltrates, but this is highly variable and


will almost certainly depend on the adequacy ofsamples obtained, the size and choice of site of thebiopsy, the skill of operator, the stage of illness andthe strength of clinical suspicion. The decision toundertake open lung biopsy in such cases should beconsidered on an individual basis, but its early use,prior to broad-spectrum antimicrobial therapy, is tobe recommended.

In addition to the withdrawal of the drug, treat-ment is supportive. Evidence for the use of cortico-steroids is variable.

Clinical features include dyspnoea, fever, cough, newparenchymal infiltrates and pleural thickening. Twopatterns have been defined. Early-onset pneumonitisoccurs within the first 6 months of therapy andgenerally responds to withdrawal of the drug. Incontrast, late-onset pneumonitis commences aftermany months or years of cyclophosphamide therapyand manifests with progressive pulmonary fibrosis andbilateral pleural thickening. The late-onset varietyhas a minimal response to withdrawal of the drug orthe use of corticosteroid therapy.

Bleomycin Methotrexate

This is the chemotherapeutic agent that most com-monly causes pulmonary ILD. There is a significantincrease of bleomycin-related pulmonary disease inpatients over 70 years of age and in those who havereceived a cumulative dose >550 U. Prior or con-comitant thoracic radiotherapy increases the inci-dence of severe pulmonary toxicity and there is asynergistic effect between prior bleomycin exposureand subsequent exposure to high O2 concen-trations.14

Intense screening with pulmonary function tests,CT scanning and 67Ga scanning suggests as many as40% of bleomycin-treated patients develop pul-monary disease, but this is clinically relevant inapproximately 5-20%. Frequent monitoring of theCO-diffusing capacity may predict toxicity. CT mayalso be useful in establishing an early diagnosis. Upto 3% of those treated with bleomycin may die fromsevere drug-induced fibrosis.15

Diffuse pneumonitis, OP and severe progressivefibrosis with honeycombing are the pulmonarypathologies most frequently seen with bleomycin.Treatment is supportive and the changes may bereversible if detected at an early stage. However, ifsignificant fibrosis has already developed, theprocess may progress despite corticosteroidtherapy.16

Cyclophosphamide

The incidence of cyclophosphamide-induced pneu-monitis is certainly underestimated. There is noclear-cut relationship between toxicity and dose.

Pneumonitis is one of the most serious complica-tions of methotrexate therapy. This reaction is unre-lated to dose. Dyspnoea, non-productive cough andfever usually develop a few days to several weeks afterstarting treatment and, in rare cases, even severalmonths or years later. Diffuse pulmonary infiltrateswith or without hilar lymphadenopathy and pleuraleffusions may be seen on CXR. Peripheral and/orpulmonary eosinophilia is seen in over half of thecases. The process is almost always reversible withor without the addition of corticosteroids and onlya few deaths have been reported.17

Nitrofurantoin

Nitrofurantoin is one of the commonest causes ofdrug-induced pulmonary ILD, although the clinicaluse of this antibiotic is waning. The incidence is esti-mated at between 1 in 500 and 1 in 5000 individualstreated. The onset of symptoms is usually withinhours to several days after the initiation of therapyand is not dose related. Fever and dyspnoea arealmost always present and pleuritic chest pain isreported in about one-third of patients. Other com-mon findings include a peripheral leucocytosis,eosinophilia and a high ESR.

The CXR shows an alveolar and/or interstitialprocess and there may be a pleural effusion, which isusually unilateral. There are no specific laboratorytests to confirm the presence of acute nitrofurantoinpneumonitis. It is not known whether cortico-steroids accelerate the resolution. One per cent ofnitrofurantoin pulmonary reactions are fatal.18

Drug-induced interstitial lung disease 245

Amiodarone

This anti-arrhythmic agent has become an importantcause of drug-induced ILD in recent years.Pneumonitis occurs in up to 6% of patients. The timefrom starting treatment to onset is inverselyproportional to the dose and it is uncommon inpatients taking less than 200 mg day-1. The majorityof the patients have been receiving the drug for at leasta month, usually at a dose of at least 400 mg day-1.The average time to onset is 1-2 years. Recently, caseshave been described of ARDS developing within a fewdays of cardiac surgery in patients previously receivingamiodarone (Fig. 18.5)19 and this may be anunsuspected cause of ARDS.20

Most patients complain of dyspnoea, non-produc-tive cough and, occasionally, a low-grade fever.Pleuritic chest pain occurs in 10% of patients. Thepathogenesis is uncertain, but there is evidence ofincreased oxidant stress, and amiodarone-inducedapoptosis in alveolar epithelial cells has been demon-strated in vitro. The pulmonary toxicity appears to berelated to its sequestration in the lung.

As withdrawal of the drug may precipitate life-threatening arrhythmia, the diagnosis should be con-firmed by thorough investigation, but may remainone of exclusion. Laboratory studies show a normalto mildly elevated leucocyte count, generally withouteosinophilia, and an elevated ESR. Pulmonary func-tion studies reveal hypoxaemia, with a decreasedtotal lung capacity and CO-diffusing capacity. 67Galung scanning may be useful in differentiating amio-darone pneumonitis from congestive heart failure,but echocardiography and an assessment of pul-monary capillary wedge pressure may also berequired. Cardiac failure is also distinguishable by itsclinical response to conventional therapy.

Although helpful in excluding infection, BAL isnon-diagnostic in amiodarone pneumonitis as cellu-lar patterns range from normal to lymphocytic, neu-trophilic and mixed pictures. The absence of foamycells (phospholipid-filled macrophages) eliminatesthe diagnosis, whereas their presence only confirmsexposure to the drug and does not necessarily indi-cate toxicity. Recently, KL-6, a mucin-like glycopro-tein secreted by type II alveolar cells, was found to beelevated in amiodarone pneumonitis21 and intersti-tial pneumonitis. This is likely to be a non-specificbut potentially useful marker of type II cell injury.

Amiodarone pneumonitis is primarily an intersti-tial or alveolar process seen on CXR (Fig. 18.5). CT

Figure 18.5 Amiodarone pneumonitis. This 62-year-old woman

developed an acute respiratory distress syndrome-like respiratory

failure following coronary bypass surgery, but had also been on

amiodarone for over 2 years. Bronchoalveolar lavage revealed a

23% eosinophilia. No evidence of infection was found and her con-

dition dramatically improved on withdrawal of the amiodarone.

scanning may provide further definition becauseamiodarone is iodinated and consequently radio-opaque, making the lesions of amiodarone pneu-monitis denser than the surrounding soft tissue inthe chest wall.

A biopsy is not usually necessary to establish thediagnosis. Histological features include foam cells asevidence of abnormal phospholipid turnover, OPand diffuse interstitial fibrosis, which develops in10% of affected patients.

Eighty per cent of patients respond to stopping thedrug and treatment with steroids, which are usuallyrequired for at least 2-6 months. Approximately 25%of patients will demonstrate long-term pulmonarysequelae in the form of persistent radiological orlung-function abnormalities. A few individuals showrecurrence of pulmonary opacities, usually as thecorticosteroids are withdrawn. There are many casereports of patients who have continued on amio-darone because it was the only drug that controlledtheir ventricular dysrhythmia, and who were concur-rently given corticosteroids to suppress amiodaronepneumonitis. The overall mortality is about 20% andincludes death from respiratory and cardiac causes.


ACUTE VASCULITIC LUNG DISEASE

Pulmonary involvement has been described in associ-ation with all the systemic vasculitides, including sys-temic lupus erythematosus (SLE), rheumatoidarthritis, dermatomyositis/polymyositis, Sjogren'ssyndrome, polyarteritis, giant cell arteritis, micro-scopic polyarteritis, Wegener's granulomatosis andHenoch-Schonlein purpura. Pathology varies bothwithin and between diseases. Most cases present sub-acutely, but pulmonary involvement, characterized byOP, pneumonitis, DAD or diffuse alveolar haemor-rhage (DAH), can present with acute respiratoryfailure.

General investigations

There may be a preceding history of joint/otherorgan involvement, but a vasculitis may present pri-marily with pulmonary pathology. Clinical evidencefor systemic vasculitis should be sought and serolog-ical tests, including ANA and ANCA requested. Thepresence of ANA suggests collagen vascular disease.Immune complex deposition, hypergammaglobuli-naemia and complement consumption are also fea-tures of these disorders.

The ANCA-positive vasculitides are Wegener'sgranulomatosis, microscopic polyarteritis andChurg-Strauss syndrome. ANCA is positive in 90%of patients with acute Wegener's, is usually positivein microscopic polyarteritis and Churg-Strausssyndrome, but will be negative in collagen vasculardiseases.

monia. Both respond to increased immunosuppres-sion and generally have a better prognosis than whenno cause is found.

ANCA-associated vasculitis

Wegener's granulomatosis may present solely in thelung. Its manifestation depends on whether the pre-dominant pathological lesion is granulomatous orvasculitic. Radiological appearances are often dis-tinct from other causes of acute ILD. Most patientspresent with either large opacities (70%) or multi-ple small opacities that change over time. The clin-ical picture will be one of systemic upset andprogressive respiratory failure, usually over weeksor months. Mechanical ventilation may be requiredin severe cases, particularly those with DAH (seebelow).

In suspected cases, a histological diagnosis shouldbe made from the most accessible tissue affected(nose, skin, kidney or lung being the most commonsites for biopsy). BAL adds little to the specific diag-nosis of ANCA-associated vasculitis, but may behelpful in excluding other diagnoses. Untreated, thedisease is rapidly fatal. Treatment is with cortico-steroids and cyclophosphamide. If severe, parenteraltherapy may be required, usually with three doses ofcyclophosphamide.

Churg-Strauss syndrome occurs at a mean age of35 years, initially presenting as asthma. Untreated, itprogresses to eosinophilic pneumonia with highperipheral blood eosinophilia. Rarely, it presents latein the disease course as severe pulmonary disease dueto overwhelming vasculitis. The response to steroidsis excellent.

Collagen vascular disease

SLE may present acutely as lupus pneumonia, diffusepulmonary haemorrhage or OP, but all are uncom-mon. BAL will exclude infection and determine thepresence or absence of DAH in such individuals.Both lupus pneumonia and OP will generallyrespond to steroids, although some patients willrequire the addition of cyclophosphamide. DAH, bycontrast, is associated with a high mortality. Itsresponse to immunosupression is variable.

Other collagen vascular diseases may presentacutely, rarely as OP or accelerating interstitial pneu-

DIFFUSE ALVEOLAR HAEMORRHAGE

DAH is a clinical diagnosis and may occur in associ-ation with a number of disorders, includingWegener's granulomatosis, Goodpasture's syndrome,microscopic polyarteritis, SLE, rheumatoid arthritis,polymyositis, lymphangioleiomyomatosis, pul-monary vascular occlusive disease and lung allograftrejection. Of importance is the absence of haemo-ptysis as a presenting symptom in approximatelyone-third of patients. Progressive pulmonary infil-trates, a falling haemoglobin level and a haemorrhagic

References 247

BAL with the presence of haemosiderin are usuallydiagnostic. In subacute cases, there may be an increasein diffusion capacity on pulmonary function testingand this may also herald recurrence in previouslydiagnosed cases. The radiological appearance isone of diffuse air-space shadowing (Fig. 18.6).

Screening for ANCA, ANA, rheumatoid factor,anti-phospholipid antibodies, complement, cryoglo-bin and coagulation abnormalities may help identifyan underlying associated syndrome such asGoodpasture's syndrome or collagen vascular dis-ease. In most cases, the diagnosis of the collagen vas-cular disease pre-dates the occurrence of DAH. Renalinvolvement may be present secondary to vasculitisand collagen vascular disease.

Histology may show small-vessel vasculitis orbland pulmonary haemorrhage. Characteristically,DAH is characterized by an oedematous interstitiumwith fibrinoid necrosis, infiltration of neutrophilsand leakage of red blood cells into the alveolar space.

The outcome of DAH depends on the underlyingdisease, with an early mortality approaching 50% in

SLE and a 5-year survival of only 20%. In contrast,isolated vasculitis enjoys a better prognosis, with25% early mortality and a 5-year survival of over50%.

DAH is treated by controlling the underlying vas-culitis with cyclophosphamide and prednisolone. Insevere cases, other therapies, including plasmaphere-sis and pooled intravenous immunoglobulins, havebeen used, but such therapy remains empirical.

SUMMARY

The broader differential diagnosis of acute ILDshould be considered in all patients presenting withacute respiratory failure and radiological opacities.As individual entities, these disorders are rare, butcollectively they form an important diagnostic dif-ferential together with ARDS and severe pulmonaryinfection. In particular, the absence of a predisposingfactor for ARDS coupled with a prodromal phaseshould alert suspicion. A vigorous and promptsearch for infection should be made to avoid delay incommencing immunosuppressive therapy, to whichmany of these disorders respond well. It is probablethat patients with acute ILD will benefit from similarventilatory strategies and adjunctive therapies tothose currently used in ARDS. Because these are rarediseases, improvements in diagnosis and therapy willbe slow to materialize, but developments in the treat-ment of ARDS and the chronic forms of ILD mayprovide valuable insights into the stereotypicalresponses of the lung to diffuse injury.

REFERENCES

Figure 18.6 Diffuse alveolar haemorrhage. A 72-year-old man pre-

sented with a 2-month history of progressive dyspnoea, anaemia

and leucocytosis. Bronchoalveolar lavage was positive for Heme

and also revealed a lymphocytosis (26%). Antinuclear cytoplasmic

antibody (ANCA) testing was positive and a renal biopsy confirmed

Wegener's granulomatosis. The patient continued to deteriorate

despite methylprednisolone and cyclophosphamide.

1. Phan, S, Thrall, R, eds. Pulmonary fibrosis. New York:

Marcel Dekker, 1995.

2. Wells, AU, Hansell, DM, Harrison, NK, Lawrence, R,

Black, CM, du Bois, RM. Clearance of inhaled 99mTc-

DTPA predicts the clinical course of fibrosing alveoli-

tis. EurRespirJ 1993; 6: 797-802.

3. Steinberg, KP, Mitchell, DR, Maunder, RJ, Milberg, JA,

Whitcomb, ME, Hudson, LD. Safety of bronchoalveolar

lavage in patients with adult respiratory distress

syndrome. Am Rev Respir Dis 1993; 148: 556-61.

4. Papazian, L, Thomas, P, Bregeon, F, et al. Open-lung

biopsy in patients with acute respiratory distress

syndrome. Anesthesiology 1998; 88: 935-44.


5. de los Santos, R, Seidenfeld, JJ, Anzueto, A, et al. One

hundred percent oxygen lung injury in adult baboons.

Am Rev Respir Dis 1987; 136: 657-61.

6. Capellier, G, Maupoil, V, Boillot, A, et al. L-NAME

aggravates pulmonary oxygen toxicity in rats. Eur

Kesp/>yi996;9:2531-6.

7. Fauci, AS, Wolff, SM, Johnson, JS. Effect of

cyclophosphamide upon the immune response in

Wegener's granulomatosis. N Engl J Med 1971;

285(27): 1493-6.

8. Bouros, D, Nicholson, AC, Polychronopoulos, V, du

Bois, RM. Acute interstitial pneumonia. Eur Respir J

2000; 15:412-18.

9. Primack, SL, Hartman, TE, Ikezoe, J, Akira, M,

Sakatani, M, Muller, ML Acute interstitial pneumonia:

radiographic and CT findings in nine patients.

Radiology 1993; 188: 817-20.

10. Suntharalingam, G, Regan, K, Keogh, BF, Morgan, CJ,

Evans, TW. Influence of direct and indirect aetiology

on acute outcome and 6-month functional recovery in

acute respiratory distress syndrome. Crit Care Med

2001; 29: 562-6.

11. Vourlekis, JS, Brown, KK, Cool, CD, et al. Acute inter-

stitial pneumonitis. Case series and review of the

literature. Medicine (Baltimore) 2000; 79: 369-78.

12. Cordier, JF. Organising pneumonia. Thorax 2000; 55:

318-28.

13. Allen, JN, Pacht, ER, Gadek, JE, Davis, WB. Acute

eosinophilic pneumonia as a reversible cause of

noninfectious respiratory failure. N Engl J Med 1989;

321: 569-74.

14. Tryka, AF, Godleski, JJ, Brain, JD. Differences in effects

of immediate and delayed hyperoxia exposure on

bleomycin-induced pulmonary injury. Cancer Treat

Rep 1984; 68: 759-64.

15. White, DA, Stover, DE. Severe bleomycin-induced

pneumonitis. Clinical features and response to

corticosteroids. Chest 1984; 86: 723-8.

16. Maher, J, Daly, PA. Severe bleomycin lung toxicity:

reversal with high dose corticosteroids. Thorax 1993;

48: 92-4.

17. Emery, P, Breedveld, FC, Lemmel, EM, et al. A com-

parison of the efficacy and safety of leflunomide

and methotrexate for the treatment of rheumatoid

arthritis. Rheumatology (Oxford) 2000; 39: 655-65.

18. Jick, SS, Jick, H, Walker, AM, Hunter, JR.

Hospitalizations for pulmonary reactions following

nitrofurantoin use. Chest 1989; 96: 512-15.

19. Van Mieghem, W, Coolen, L, Malysse, I, Lacquet, LM,

Deneffe, GJ, Demedts, MG. Amiodarone and the

development of ARDS after lung surgery. Chest

1994; 105: 1642-5.

20. Ashrafian, H, Davey, P. Is amiodarone an underrecog-

nized cause of acute respiratory failure in the ICU?

Chest 2001; 120: 275-82.

21. Endoh, Y, Hanai, R, Uto, K, et al. KL-6 as a potential

new marker for amiodarone-induced pulmonary

toxicity. AmJCardiol 2000; 86: 229-31.

19Pulmonary embolism and pulmonaryhypertensionGRAHAM F PINEO, RUSSELL D HULL AND GARY E RASKOB

Introduction 249

Aetiology and pathogenesis 249

Clinical factors predisposing to the development ofvenous thromboembolism 250

Clinical features 250

Laboratory features 251

Diagnosis 251

Treatment of venous thromboembolism 254

Massive pulmonary embolism 257

Pulmonary embolectomy 258

Percutaneous clot extraction or disruption 259

Inferior vena caval interruption 259

Chronic thromboembolic pulmonary hypertension 260

Conclusion 260

References 260

INTRODUCTION

Venous thromboembolism is a common disorderwith an estimated annual incidence of symptomaticvenous thromboembolism of 117 per 100 000 popu-lation.1 The incidence increases with each decadeover the age of 60. Accordingly, with an ageingpopulation, it is an increasingly important problemfor the health services in many countries.

Venous thrombosis commonly develops in thedeep veins of the leg, but may occur in the arm or inthe superficial veins of the limbs. Superficial venousthrombosis is a relatively benign disorder unlessextension into the deep venous system develops.Thrombosis involving the deep veins of the leg isdivided into two prognostic categories:

1. thrombi that remain confined to the deep calfveins (calf-vein thrombosis),

2. proximal-vein thrombosis involving the poplitealfemoral or iliac veins.

Pulmonary embolism (PE) originates from thrombiin the deep veins of the leg in 90% or more ofpatients. Other less common sources include the deeppelvic veins, renal veins, inferior vena cava, right sideof the heart or the axillary veins. Although most clin-ically important emboli arise from proximal deep-vein thrombosis (DVT) of the leg, upper-extremitythrombosis may be the source.2 DVT and/or PE arereferred to collectively as venous thromboembolism.

AETIOLOGY AND PATHOGENESIS

Venous thrombi are composed mainly of fibrin andred blood cells, with a variable platelet and leucocytecomponent. The formation, growth and breakdownof venous thromboemboli represent a balancebetween thrombogenic stimuli and protective mech-anisms. The thrombogenic stimuli are venous stasis,activation of blood coagulation and vein damage.The protective mechanisms are:

250 Pulmonary embolism and pulmonary hypertension

• the inactivatioh of activated coagulation factorsby circulating inhibitors (e.g. antithrombin,protein C),

• clearance of activated coagulation factors andsoluble fibrin polymer complexes by the reticulo-endothelial system and by the liver,

• lysis of fibrin by fibrinolytic enzymes derivedfrom plasma, endothelial cells and leucocytes.

CLINICAL FACTORS PREDISPOSING TOTHE DEVELOPMENT OF VENOUSTHROMBOEMBOLISM

Epidemiologic studies, particularly in hospitalizedpatients, have identified a number of clinical factorsthat predispose to DVT. Common risk factors areshown in Table 19.1. When designing clinical trialsfor the prevention of DVT, these risk factors are usu-ally taken into account, but, in some studies, high-risk patients are excluded, making the results lessgeneralizable. The identification of naturalinhibitors that predispose to DVT is important incancer, pregnancy and the use of the oral contracep-tive pill. Patients with cancer have been known,since the days of Trousseau, to be at risk for thedevelopment of DVT, but it has been difficult tounderstand why patients with minimal diseasedevelop thrombosis whereas others with advancedmalignancy do not. Similarly, although DVT is rela-tively uncommon in users of oral contraceptive pillsor in pregnancy, it is now clear that the presence ofan inhibitor deficiency markedly increases the likeli-hood of DVT.

The known inhibitor deficiencies predisposing tothrombosis include antithrombin, protein C defi-ciency, protein S deficiency, activated protein C resist-ance, the prothrombin mutant and decreasedfibrinolytic activity. These disorders have beenreviewed and an attempt has been made to identifythe incidence, the ethnic distribution and the degreeof risk in pregnancy, surgery and cancer, and recom-mendations have been made for prophylaxis andtreatment.3,4 Homocysteinaemia (sometimes refer-red to as hyperhomocysteinaemia) predisposes to bothvenous and arterial thrombosis through mechanismsthat are poorly understood. The same applies topatients with the antiphospholipid syndrome andheparin-induced thrombocytopenia.

Activated protein C resistance is the most commonabnormality predisposing to DVT. The defect is due tosubstitution of glutamine for arginine at residue 506in the factor V molecule, making factor V resistant toproteolysis by activated protein C. The gene mutationis known as factor V Leiden and follows autosomaldominant inheritance. Factor V Leiden is present inapproximately 5% of the normal population, in 16%of patients with a first episode of DVT and in up to35% of patients with idiopathic DVT.3 ProthrombinG20210 A is a recently identified gene mutation pre-disposing to DVT. It is present in approximately 2% ofapparently healthy individuals and in 7% of thosewith DVT. In approximately 60% of patients withidiopathic DVT, an inherited abnormality cannot bedetected, suggesting that other gene mutations arepresent and may have an aetiological role.

CLINICAL FEATURES

PE occurs in 50% of patients with objectively docu-mented proximal-vein thrombosis. Many of theseemboli are asymptomatic. The clinical importance ofPE depends on the size of the embolus and on thepatient's cardiorespiratory reserve. Usually, only partof the thrombus embolizes and 70% of patients withPE demonstrated by angiography have detectableDVT at presentation. DVT and PE are not separatedisorders, but a continuous syndrome in which theinitial clinical presentation may be with symptoms ofeither DVT or PE. Strategies for detection includetests for PE - lung scanning, pulmonary angiographyor spiral computerized tomography (CT) - and testsfor DVT - ultrasound, impedance plethysmographyor venography.

The clinical features of venous thrombosis includepain, tenderness and swelling, a palpable cord (i.e. athrombosed vessel that is palpable as a cord), discol-oration, venous distension and prominence of thesuperficial veins and cyanosis. The clinical diagnosisof DVT is non-specific because symptoms or signsmay be caused by non-thrombotic disorders. Therare exception is the patient with phlegmasia ceruleadolens (swollen leg with cyanosis of the skin due tomarked obstruction of venous outflow), in whomthe diagnosis of massive ileofemoral thrombosis isusually obvious. This syndrome occurs in <1% ofpatients with symptomatic DVT. In most, the symp-toms and signs are non-specific and yet, in 50-85%,

Diagnosis 251

Table 19.1 Clinical risk factors predisposing to the development of venous thromboembolism

Surgical and non-surgical traumaPrevious venous thromboembolismImmobilizationMalignant diseaseHeart diseaseLeg paralysisAge > 40ObesityOestrogensParturitionVaricose veins

Activated protein C resistanceProthrombin polymorphismProtein C deficiencyProtein S deficiencyAntithrombin deficiencyAnticardiolipin syndromeHeparin-induced thrombocytopenia

thrombosis will not be confirmed by objective tests.Patients with extensive DVT may have minor symp-toms and signs. Conversely, patients with florid legpain and swelling may have no objective evidence ofDVT. Patients can be assigned pre-test probabilitiesof DVT based on their clinical features and history.However, these pre-test probabilities are neithersufficiently high to give anticoagulant treatmentnor sufficiently low to withhold treatment withoutperforming objective testing.

The clinical scenarios of acute PE include thefollowing syndromes, which may overlap:

• transient dyspnoea and tachypnoea,• the syndrome of pulmonary infarction including

pleuritic chest pain, haemoptysis and pleural effu-sion and infiltrates on chest radiograph

• right-sided heart failure with severe dyspnoea andtachypnoea,

• cardiovascular collapse with hypotension, syncopeand coma (massive PE),

• less common and non-specific presentations,including unexplained arrhythmia, resistantcardiac failure, wheezing, cough, pyrexia, anxi-ety/apprehension and confusion.

All of these clinical scenarios are non-specific andmay be caused by a variety of cardiorespiratory dis-orders. Objective testing is therefore mandatory toconfirm or exclude the presence of PE. Classifyingpatients into categories of pre-test probability (low,intermediate or high) is useful in a minority of patientswhen combined with lung-scan findings. Pre-testclinical probabilities for PE have not been as repro-ducible as they have been for the diagnosis of DVT.

LABORATORY FEATURES

Venous thromboembolism is associated with non-specific laboratory changes of the acute-phaseresponse. This response includes elevated levels offibrinogen and factor VIII, increases in the leucocyteand platelet counts, systemic activation of bloodcoagulation, fibrin formation and breakdown, andincreases in the plasma concentrations of prothrom-bin fragment 1.2, fibrinopeptide A, complexes ofthrombin-antithrombin and D-dimer. All of theseare non-specific and may occur as the resultof surgery, trauma, infection, inflammation orinfarction. None can reliably be used to predict thedevelopment of DVT.

The fibrin breakdown fragment D-dimer can bemeasured by an enzyme-linked immunosorbentassay (ELISA) or by a latex agglutination assay. Someof these assays have a rapid turnaround time andsome are quantitative. The D-dimer may be useful asa test for exclusion of patients with suspected DVTor PE (see below). A positive result is non-specific.

DIAGNOSIS

The differential diagnosis in DVT includes musclestrain, lymphangitis, popliteal cyst rupture and cel-lulitis. Without objective testing, it is impossible toconfirm or exclude DVT. Diagnosis can often bedetermined once DVT has been excluded, althoughin 25% the cause remains uncertain, even after care-ful follow-up.5


Objective tests for deep-veinthrombosis

The tests of value in clinically suspected DVT areultrasound imaging, impedance plethysmographyand venography.6 Each has been validated by clinicaltrials, including prospective studies with long-termfollow-up. These have established the safety of with-holding anticoagulant treatment in patients withnegative test results.6 Ultrasound imaging or imped-ance plethysmography (IP) are both effective indetecting proximal-vein thrombosis. They have limitedsensitivity for calf-vein thrombosis and require seri-al testing to detect extension into the popliteal veinor more proximally. When performed serially, thesetests can safely replace venography in symptom-atic patients. Venography continues to have animportant role in selected patients, such as those forwhom serial testing is impractical and those withabnormal non-invasive test results who have condi-tions known to produce false-positive results. Thewide availability of ultrasound has meant that it hassupplanted IP as the principal non-invasive test forDVT in most centres. Venography remains useful inpatients with suspected acute recurrent DVT. A diag-nostic approach for a patient with a first episode ofDVT is shown in Figure 19.1. Ultrasound imaginghas two practical advantages. It is more sensitive thanIP for small, non-occlusive thrombi and this enablesserial testing to be limited to a single repeat test done5-7 days after presentation. Also, it is not influencedby congestive cardiac failure or by disorders thatimpair deep-venous filling (e.g. peripheral arterialdisease), which may produce false-positive IP results.In the absence of these conditions, IP has a similarhigh positive predictive value (>90%) to ultrasoundimaging, but clinical examination of the patient isrequired to exclude potential causes of false-positiveresults.

IP is a valuable test for patients with suspectedrecurrent DVT because it returns to normal earlierthan ultrasonography in proximal-vein thrombosis.7

Compression ultrasound may remain abnormal for2 years or more due to persistent non-compressibili-ty of the vein caused by fibrous organization of theoriginal thrombus. A normal IP result obtained atthe time of completing anticoagulant treatment pro-vides a useful baseline for future comparison. An IPresult that has changed from normal to abnormal is

Figure 19.1 Diagnosis and therapy of patients with suspected

first-episode deep-vein thrombosis (DVT). (For details see text and

reference 3.)

highly predictive of acute recurrent proximal-veinthrombosis. The finding of a new non-compressiblevenous segment by ultrasound imaging is probablypredictive of acute recurrent thrombosis, but thiscriterion is of limited value because many patientshave persistently non-compressible venous segmentsdue to the initial episode. To date, no ultrasoundimaging criteria have been validated by prospectivefollow-up studies to establish the safety of withhold-ing anticoagulant treatment in patients with recur-rent DVT.

Studies have demonstrated a high negative predict-ive value of D-dimer for acute DVT.8 The measure-ment of plasma D-dimer may be useful in patientswith initially negative ultrasound imaging to excludethe presence of DVT and to avoid the need forrepeated testing. However, this is less important inpatients with suspected first-episode DVT, becausethe need for repeat testing has been reduced to a sin-gle test at 5-7 days.9 Most require a follow-up clinicvisit anyway, so the return visit is not a major incon-venience. The value of D-dimer is potentially greaterin patients with suspected acute recurrent DVT.However, the safety of withholding anticoagulanttreatment in such patients with negative D-dimerresults 'has not been established by adequatelydesigned clinical trials.

Diagnosis 253

Differential diagnosis of pulmonaryembolism

The differential diagnosis is wide, depending uponthe clinical scenario. It includes pneumonia, pneumo-thorax, pulmonary oedema, pericarditis, rib fracture,myocardial infarction,and septicemia. The key testsinclude echocardiography, spiral CT, pulmonaryangiography, ventilation perfusion (VQ) lung scan-ning, magnetic resonance imaging (MRI) and objec-tive tests for proximal DVT.6 A diagnostic approachis summarized in Figure 19.2.

Objective testing for DVT is useful in patients withsuspected PE, particularly those with non-diagnosticlung-scan results (indeterminate, intermediate or lowprobability categories).6 The detection of proximalDVT provides an indication for anticoagulant treat-ment, regardless of the presence or absence of PE. Anegative result does not exclude PE. If the patient hasadequate cardiorespiratory reserve, serial ultrasoundimaging may be used as an alternative to pulmonaryangiography. The rationale is that the clinical object-ive in such patients is to prevent recurrent PE, whichis unlikely in the absence of proximal-vein thrombo-sis. For patients with inadequate cardiorespiratoryreserve, the clinical objectives are to prevent deathand morbidity from an existing embolus and to allowfurther testing for the presence or absence of PE.

OBJECTIVE TESTS

The value of echocardiography has recently beenfirmly established and has four main advantages:

• it is non-invasive and easily available,• it can exclude other causes of cardiogenic shock,

e.g. extensive left ventricular infarction, pericar-dial tamponade or dissecting aortic aneurysm,

• it allows an estimate of pulmonary artery pressureand so provides information on the severity ofpulmonary artery obstruction,

• it can be used serially to assess response to treat-ment.

The echocardiographic findings are not specificand reflect the response of the right heart to acutepulmonary artery hypertension. They consist of dis-tension of the pulmonary artery trunk, right ventricu-lar (RV) dilatation and hypokinesis, reduced leftventricular (LV) size and an increased RV/LV diam-eter, diastolic and systolic flattening of the intraven-tricular septum and paradoxical systolic wallmotion. This pattern is mimicked, in particular, byRV infarction associated with LV dysfunction.

Scoring systems have been developed that correl-ate well with the angiographic severity index, and thevalue of serial echocardiography has been demon-strated in patients treated with thrombolytic agents.Rarely, RV thrombus may be visualized, a clinical sit-uation associated with a high mortality rate, 30% ofsuch patients succumbing as a result of massive pul-monary thromboembolism.

There are several limitations to the applications ofechocardiography. At least 40% of the pulmonaryvascular bed needs to be obstructed to producedetectable features. Co-existent cardiorespiratorydisease also limits its value because of the non-specific nature of the abnormalities and because

Figure 19.2 Diagnosis and therapy

of patients with suspected pulmonary

embolism (PE). DVT, deep-vein

thrombosis. (For details see text arid

reference 3.)


imaging via the transthoracic route may be difficult.Transoesophageal echocardiography may be valuablein this situation, particularly in making a full haemo-dynamic assessment in the shocked, intubatedpatient.

Both spiral CT imaging and MRI are now ofproven value in PE. Spiral CT imaging is highly sen-sitive for large emboli (segmental or greater arteries),but is less sensitive for emboli in subsegmental pul-monary arteries.10 These smaller emboli may still beclinically important in patients with inadequate car-diorespiratory reserve. MRI is highly sensitive for PE.However, one study documented significant inter-observer variation in the sensitivity, ranging from70% to 100%.n The safety of withholding anticoagu-lant treatment in patients with negative results byspiral CT imaging or MRI has not been establishedby prospective clinical trials.

The assay for plasma D-dimer is potentially usefulto exclude PE, based on a high negative predictivevalue reported from centres with research expertise.8

However, further studies are required to establish theplace of D-dimer testing and, in particular, to evalu-ate the safety of withholding anticoagulant treatmentin patients with a negative result.

TREATMENT OF VENOUSTHROMBOEMBOLISM

The objectives of treatment are:

• to prevent death,• to prevent recurrent DVT and PE,• to prevent the post-phlebitic syndrome.

Anticoagulant drugs constitute the mainstay oftreatment, and graduated compression stockings (for24 months) significantly decrease the incidence ofthe post-thrombotic syndrome.

Heparin therapy

The anticoagulant activity of unfractionated heparindepends upon a unique pentasaccharide that bindsto antithrombin and potentiates the inhibition ofthrombin and activated factor X (Xa) by antithrom-bin. About one-third of all heparin moleculescontain the unique pentasaccharide sequence,regardless of whether they are low or high molecular

weight (HMWH) fractions.12 It is the pentasacchar-ide sequence that confers the molecular high affin-ity for antithrombin. In addition, heparin catalysesthe inactivation of thrombin by another plasmacofactor (co-factor II), which acts independently ofantithrombin.

Heparin has a number of other effects. These includethe release of tissue factor pathway inhibitor, bindingto proteins located on platelets, endothelial cells andleucocytes, suppression of platelet function and anincrease in vascular permeability. The anticoagulantresponse to a standard dose of heparin varies widelyamongst patients. This makes it necessary to monitorthe anticoagulant response, using either the activatedpartial thromboplastin time (aPTT) or heparin levels,and to titrate the dose to the individual patient.

Conventional therapy comprises a combination ofcontinuous intravenous heparin and oral warfarin.6

The length of the initial intravenous heparin therapycan be reduced to 5 days, thus shortening the hospi-tal stay. The simultaneous use of initial heparin andwarfarin has become standard practice in medicallystable patients. Exceptions include patients who mayrequire immediate interventions such as in throm-bolysis or insertion of a vena cava filter or patients atvery high risk of bleeding. Heparin is continued untilthe INR has been within the therapeutic range (2-3)for 2 consecutive days.

The efficacy of heparin therapy depends uponachieving a critical therapeutic level of heparin with-in the first 24 hours of treatment. Data from threeconsecutive, double-blind clinical trials indicate thatfailure to achieve the therapeutic aPTT threshold by24 hours is associated with a 23.3% subsequentrecurrent venous thromboembolism rate, comparedwith 4-6% for the patient groups who reached thera-peutic anticoagulation at 24 hours.13 The recur-rences occurred throughout the 3-month follow-upperiod and could not be attributed to inadequateoral anticoagulant therapy. The critical therapeuticlevel of heparin, as measured by the aPTT, is 1.5times the mean of the control value or the upperlimit of the normal aPTT range. This corresponds toa heparin blood level of 0.2-0.4 U mL-1 by the pro-tamine sulphate titration assay and 0.35-0.70 U mL-1

by the anti-factor Xa assay.There is a wide variability in the aPTT and in

heparin blood levels with different reagents and evenwith different batches of the same reagent. It is there-fore vital for each laboratory to establish the minimal

Treatment of venous thromboembolism 255

therapeutic level of heparin, as measured by theaPTT, that will provide a heparin blood level of atleast 0.35 U mL- ] by the anti-factor Xa assay for eachbatch of thromboplastin reagent being used, particu-larly if the batches of reagent are provided by differ-ent manufacturers.

Although there is a strong correlation between sub-therapeutic aPTT values and recurrent thrombo-embolism, the relationship between supra-therapeuticaPTT and bleeding (aPTT ratio 2.5 or more) is lessdefinite. Indeed, bleeding during heparin therapy ismore closely related to underlying clinical risk factors.A lower body weight and age > 65 are independentrisk factors for bleeding on heparin.14,15

Numerous audits indicate that administration ofintravenous heparin is difficult and that the clinicalpractice of using an ad hoc approach to heparin dosetitration frequently results in inadequate therapy. Forexample, an audit of physician practices at threeuniversity-affiliated hospitals documented that 60% ofpatients failed to achieve an adequate aPTT response(ratio 1.5) during the initial 24 hours of therapy andthat 30-40% of patients remained 'sub-therapeutic'over the next 3-4 days.16 Several practices were iden-tified that led to inadequate therapy. The commontheme is an exaggerated fear of bleeding complica-tions on the part of clinicians. Consequently, it hasbeen common practice for many clinicians to starttreatment with a low heparin dose and to increasethis dose cautiously over several days to achieve the

therapeutic range. The use of a protocol has beenevaluated in two prospective studies involvingpatients with venous thromboembolism.17,18 In one,for the treatment of proximal DVT, patients weregiven either intravenous heparin alone followed bywarfarin or intravenous heparin and simultaneouswarfarin.17 The heparin nomogram is summarized inTables 19.2 and 19.3. Only 1% and 2% of the patientswere under-treated for more than 24 hours in theheparin group and in the heparin and warfaringroup, respectively. Recurrent DVT (objectively docu-mented) occurred infrequently in both groups (7%).In the other trial, a weight-based heparin dosagenomogram was compared with a standard-carenomogram (Table 19.4).18 Patients on the weight-adjusted heparin nomogram received a starting doseof SOU kg-1 as a bolus and ISUkg-1 h-11 as an infu-sion. The heparin dose was adjusted to maintain anaPTT of 1.5-2.3 times control. In the weight-adjustedgroup, 89% of patients achieved the therapeuticrange within 24 hours, compared with 75% in thestandard-care group. The risk of recurrent throm-boembolism was more frequent in the standard-caregroup. The weight-based nomogram has gainedwidespread acceptance.

COMPLICATIONS OF HEPARIN THERAPY

The main adverse effects of heparin therapy includebleeding, thrombocytopenia and osteoporosis.

Table 19.2 Heparin protocol2

Adapted from Hull, RD, et al. Arch Intern Med 1992; 152: 1589-95, with permission.17

Image Not Available


Table 19.3 Intravenous heparin dose titration nomogram according to the activated partial thromboplastin time (aPTT)

a

Adapted from Hull, RD, et al. Arch Intern Med 1992; 152:1589-95, with permission.17

Table 19.4 Weight-based nomogram for initial intra-

venous heparin therapy

Adapted from Raschke, RA, et al. Ann Intern Med 1993; 119; 874-81, withpermission.19

Patients at risk of bleeding include those who havehad recent surgery or trauma or who have other clini-cal factors that predispose to bleeding, such as activepeptic ulcer, liver disease, haemostatic defects or age>65 years. The management of bleeding will dependon its severity, the risk of recurrent venous throm-boembolism and the aPTT. Heparin should be dis-continued temporarily or permanently and insertionof an inferior vena cava filter considered. If urgentreversal of heparin effect is required, protamine sul-phate can be administered.

Heparin-induced thrombocytopenia is a well-recognized complication, usually occurring 5-10 days

after heparin treatment has started. Approximately1-2% of patients receiving unfractionated heparinwill experience a fall in platelet count to less than thenormal range. In the majority, this mild to moderatethrombocytopenia appears to be a direct effect ofheparin on platelets and is of no consequence.However, approximately 0.1-0.2% of patients developan immune thrombocytopenia mediated by IgG anti-body directed against a complex of PF4 and heparin(HIT). This may be accompanied by arterial or venousthrombosis, which may lead to death or limb amputa-tion.19 The diagnosis of HIT, with or without throm-bosis, must be made on clinical grounds, because theassays with the highest sensitivity and specificity arenot readily available and have a slow turnaround time.

When the diagnosis of heparin-induced thrombo-cytopenia is made, heparin in all forms must bestopped immediately. This is a particularly difficultproblem in the critically ill, in whom thrombo-cytopenia is usually the result of a variety of morecommon causes, e.g. sepsis, bleeding or otherintravascular or extravascular consumption.Typically, HIT is sudden in onset, with a precipitousfall in platelet count. Therefore, a slowly falling countmay not necessitate the complete cessation of allheparin (e.g. arterial flushes). In those patientsrequiring on-going anticoagulation, several alterna-tives exist; the agents most extensively used are theheparinoid Danaparoid and Hirudin.19

Osteoporosis has been reported in patients receiv-ing unfractionated heparin in dosages of 20 000 U

Image Not Available

Image Not Available

"Massive pulmonary embolism 257

day l (or more) for more than 6 months.Demineralization can progress to the fracture of ver-tebral bodies or long bones and the defect may notbe entirely reversible.

LOW MOLECULAR WEIGHT HEPARIN

Heparin currently in clinical use is polydispersed,unmodified heparin, with a mean molecular weightranging from 10 to 16 kD. In recent years, low molecu-lar weight derivatives of commercial heparin havebeen prepared that have a mean molecular weight of4-5 kD.

The low molecular weight heparins (LMWHs)that are commercially available are made by differentprocesses (such as nitrous acid, alkaline or enzymat-ic depolymerization) and they differ chemically andpharmacokinetically. The clinical significance ofthese differences is unclear and there have been veryfew studies comparing different LMWHs withrespect to clinical outcomes. The doses of the differ-ent LMWHs have been established empirically andare not necessarily interchangeable.

The LMWHs differ from unfractionated heparinin numerous ways. Of particular importance are thefollowing:

• increased bio-availability (>90% after subcuta-neous injection),

• prolonged half-life and predictable clearance,enabling once or twice daily injection,

• predictable antithrombotic response based onbody weight, permitting treatment without labora-tory monitoring.

Other possible advantages are their ability to inac-tivate platelet-bound factor Xa, resistance to inhibi-tion by platelet factor IV and their decreased effecton platelet function and vascular permeability (pos-sibly accounting for fewer haemorrhagic effects atcomparable antithrombotic doses).

There has been a hope that LMWHs will have fewerserious complications than unfractionated heparin.Evidence is accumulating that these complications areindeed less serious and less frequent. LMWHs cross-react with unfractionated heparin and cannot there-fore be used as alternative therapy in patients whodevelop HIT. The heparinoid Danaparoid possesses a10-20% cross-reactivity and can be used instead.

In a number of clinical trials, LMWH by subcuta-neous or intravenous injection was compared with

continuous intravenous unfractionated heparin,with repeat venography being the primary end-point. These studies demonstrated that LMWH wasat least as effective as unfractionated heparin.Subcutaneous unmonitored LMWH has also beencompared with continuous heparin in a number ofclinical trials.20,21 These studies indicate a signifi-cant advantage for LMWH in the reduction ofmajor bleeding and mortality. Three recent studiesindicated that LMWH used predominantly out ofhospital was as effective and safe as intravenousunfractionated heparin given in hospital,21 andLMWH is as effective as intravenous heparin in thetreatment of patients presenting with PE. Economicanalysis has also shown that LMWH is cost-effective.

Oral anticoagulant therapy

Oral anticoagulant treatment is continued for 3-6months in patients with a first-episode DVT or PE.22

Stopping oral anticoagulant treatment at 4-6 weeksresults in a high incidence (12-20%) of recurrenceduring the following 12-24 months.23 Warfarinshould be continued for 1 year to indefinitely inpatients with a second episode of objectively docu-mented DVT or PE. Stopping treatment at 3 monthsin these patients results in a 20% incidence of recur-rent DVT and a 5% incidence of PE.24

Patients at high risk of recurrent thromboembolismare those with idiopathic thrombosis, homozygousfactor V Leiden gene mutation or cancer.22

MASSIVE PULMONARY EMBOLISM

Patients with acute massive PE usually have adramatic presentation with sudden onset of severeshortness of breath, hypoxaemia and right ventricu-lar failure. Symptoms include central chest pain,often identical to angina, severe dyspnoea and, fre-quently, syncope, confusion or coma. Examinationreveals a patient in severe distress with tachypnoea,cyanosis and hypotension. The marked increase inpulmonary vascular resistance leads to acute rightventricular failure, with the presence of largeA-waves in the jugular veins and a right ventriculardiastolic gallop. With pulmonary hypertension, thereis marked right ventricular dilatation with a shiftof the intraventricular septum, decreasing cardiac


output and further decreasing coronary perfusion,which may result in cardiorespiratory arrest. Ifpatients with a massive PE survive, they are at greatrisk from further thromboembolism.

The emergency management of massive PEincludes the use of intravenous heparin, O2, mechan-ical ventilation, volume resuscitation and the use ofinotropic agents or even vasodilators. In addition tothese supportive measures, specific treatmentoptions for massive PE include:

• thrombolysis,• pulmonary thrombectomy, with or without car-

diopulmonary bypass support,• transvenous catheter embolectomy or clot dissol-

ution,• insertion of an inferior vena cava filter.

Thrombolytic therapy

Several studies have demonstrated that the mortalityfrom PE can be decreased by heparin. Treatmentwith intravenous heparin and oral anticoagulantsreduces the mortality rate to less than 5% and thismay be further reduced with the use of LMWH. Inthe prospective investigation of PE diagnosis(PIOPED) trial,25 only 10 of 399 (2.5%) patientswho had angiographically confirmed PE died.However, patients who present with acute massivePE and hypotension have a mortality rate of approxi-mately 20%. For such patients, the appropriate use ofthrombolytic agents has a role.

Several trials have compared thrombolytic drugswith heparin.26 Outcome measures for acceleratedthrombolysis included quantitative measures onrepeat pulmonary angiograms, quantitative scoreson repeat pulmonary perfusion scans and measuresof pulmonary vascular resistance. Although all stud-ies demonstrated superiority of thrombolysis (and inparticular with tPA) in radiographic and haemody-namic abnormalities within the first 24 hours, thisadvantage was short lived.26 Repeat perfusion scansat 5-7 days revealed no significant differencebetween the patients treated with thrombolyticagents or with heparin. However, at 1 year, thosereceiving thrombolytic therapy had higher CO diffu-sion capacity and lung blood capillary volume com-pared to patients receiving heparin. Follow-up of23 patients at 7 years showed that patients who had

been treated with thrombolytic therapy had lowerpulmonary artery pressure and pulmonary vascularresistance. The clinical relevance of these findings,however, must await further prospective studies.26

Three thrombolytic agents have been approved bythe Federal Drug Administration (FDA) for thetreatment of acute PE. The dosage schedules are asfollows:

• streptokinase 250 000 units over 30 min, followedby 100 000 units h"1 for 24 h,

• urokinase 4400 units kg-1 over 10 min, followedby 4400 units kg-1 h-1 for 12-24 hours (nowwithdrawn from the market),

• recombinant tissue plasminogen activator (rt-PA)100 mg administered over 2 h.

Anticoagulation with heparin or LMWH is usuallycommenced when the aPTT is less than two timesthe control.

In weighing the risks and benefits of thrombolytictherapy, the main concern relates to bleeding. Datafrom five randomized trials indicate that the fre-quency of intracranial haemorrhage followingthrombolytic therapy is 1.9% (95% confidence inter-val, 0.7-4.1%).27 Diastolic hypertension was identi-fied as a risk factor. The incidence of major bleedinghas decreased, particularly with the use of bolus orshort-term infusions and of newer thrombolyticagents, but intracerebral haemorrhage continues tooccur more frequently than with heparin.

The role of thrombolytic agents in massive PEremains controversial. Until there is a clearly demon-strated benefit in terms of both morbidity andmortality, the question of risk/benefit will remain.In the meantime, the use of thrombolytic agentshas become simpler with the use of echocardiogra-phy28'29 or spiral CT to confirm the diagnosis, theuse of short-term or bolus infusion into peripheralveins, the elimination of monitoring by laboratorytests and treatment in high dependency units(HDUs) rather than in ICUs. The fact that a highpercentage of acute massive PE occurs followingsurgery indicates that greater efforts must be made toensure that prophylactic measures are applied.

PULMONARY EMBOLECTOMY

Pulmonary embolectomy is occasionally indicatedin the management of massive PE. It is usually only

Inferior vena caval interruption 259

considered if there is >50% obstruction of thepulmonary vasculature, the patient is in shock andthe PaO2 is <9.0 Kpa. In some centres, patientswho have contraindications to thrombolytic therapyare candidates for thrombectomy. On the otherhand, it could be argued that a patient who survivesthe first 2 hours will probably survive with adequatemedical management if no further PE occurs. It isunlikely that a randomized trial comparing throm-bolytic therapy with pulmonary embolectomy ispossible.

Early experience with the Trendelenburg procedurerevealed unacceptably high mortality (>50%). Withthe use of cardiopulmonary bypass support, mortalityrates between 16% and 57% have been reported. In areview of 651 patients undergoing emergency pul-monary embolectomy, the survival rate was 59.3%with cardiopulmonary bypass support and 47.7%without it.30 Patients' with chronic pulmonary hyper-tension, other medical disorders, or with symptoms ofmore than 7 days' duration have higher mortalityrates, as do patients who have sustained a cardiacarrest before embolectomy.31 Care to avoid vasodilata-tion at the initiation of anaesthesia is important.Pulmonary haemorrhagic infarction with reperfusionhas been reported. Pulmonary embolectomy is usual-ly accompanied by insertion of a vena cava filter.

The role of pulmonary embolectomy remainsunclear and will depend on the availability of a sur-gical team. Patients who are not candidates forthrombolysis (e.g. those who have had recentsurgery) or who have not responded to maximalmedical therapy may be candidates. However, arecent report of successful thrombolysis with intra-pulmonary urokinase in patients treated within10 days of surgery32 casts further doubt on the needfor this radical procedure.

PERCUTANEOUS CLOT EXTRACTION ORDISRUPTION

Pulmonary embolectomy via a catheter suctiondevice or mechanical disruption has been used in thetreatment of patients with acute massive PE whohave contraindications to anticoagulants or throm-bolysis. Mortality rates were 27% and 28%.33 Thecommonest cause of death is cardiac arrest fromventricular arrhythmia, right heart failure and

pulmonary haemorrhage. Some patients in whomclot extraction or disruption was not possible havegone on to successful pulmonary embolectomy onbypass. Attempts have been made to fragment PEusing conventional cardiac catheters or a catheterguide wire in conjunction with pulmonary thrombo-lytic therapy. Catheter clot extraction is currentlyconfined to a few centres and its role is unclear.

INFERIOR VENA CAVAL INTERRUPTION

Early approaches to inferior vena caval interruptionincluded ligation or plication using external clips.Both procedures were accompanied by an operativemortality rate of 12-14%, a recurrent pulmonaryemboli rate of 4-6% and an occlusion rate of67-69%.34 These complications gave rise to thedevelopment of catheter-inserted intraluminal fil-ters. An ideal filter is one that is easily and safelyplaced percutaneously, is biocompatible andmechanically stable, able to trap emboli withoutcausing occlusion of the vena cava, does not requireanticoagulation and is not ferromagnetic (i.e. doesnot cause artefacts on MRI). Although there is as yetno ideal filter, several types are available. Theseinclude the Greenfield stainless-steel filter, titaniumGreenfield filter, bird's nest filter, Vena Tech filter andSimon-Nitinol filter. In experienced hands, thesedevices can be quickly and safely inserted underfluoroscopic control. One filter (for example,Antheor™Tu 50-125 Medi-Tech, Boston ScientificCorp.) can be inserted temporarily in conjunctionwith thrombolytic therapy and then removed. Thefollow-up data available to date show that theGreenfield filter has had the best performance recordand any future comparative studies should use thisfilter as the standard.34

The main indications for caval filters are con-traindications to anticoagulants, recurrent PEdespite adequate anticoagulation and prophylacticplacement in high-risk patients.34 In the last cate-gory are patients with cor pulmonale or a previoushistory of PE who are placed in high-risk situationssuch as acetabular fracture or who have cancer.More controversial indications for the prophylacticinsertion of a filter include emergency surgeryoccurring within the first 4 weeks of commencinganticoagulant therapy following thrombolytictherapy.34


In the past, the detection of a free-floating throm-bus by ultrasound examination has been consideredan indication for either thrombectomy or insertionof an inferior vena cava filter. An important studycompared the clinical outcomes of patients who hadeither the presence or absence of a free-floatingthrombus in a proximal leg vein.35 There was no dif-ference in the incidence of PE or death between thetwo groups. The authors conclude that the routineinsertion of inferior vena cava filters in patients withfree-floating thrombi cannot be supported. This is inkeeping with an earlier observation36 that free-floatingthrombi become attached to the vein wall rather thanembolizing. In a further study, patients with proxi-mal DVT were randomized to receive an inferiorvena cava filter or anticoagulant treatment alone.37

All patients were treated with heparin, followed byoral anticoagulant therapy for 3 months. At 10 days,there was a significant difference in the incidence ofPE, but no difference in mortality. Extended follow-up at 1-2 years showed a (non-significant) increasein the incidence of PE in the control group, but ahigher incidence of recurrent DVT in the vena cavafilter group and no difference in mortality.

CHRONIC THROMBOEMBOLICPULMONARY HYPERTENSION

In a small number of patients, the emboli fail toresolve and undergo fibrovascular organization toproduce chronic obstruction to pulmonary arterialblood flow and progressive right ventricular failure,hypoxaemia and death. Many of these patients havebeen found to have thrombophilia such as the factor VLeiden mutation or the antiphospholipid antibodysyndrome and, in some, decreased fibrinolytic activ-ity can be demonstrated at the endothelial cell level.At the present time, there is no unifying concept forthe pathophysiology of chronic thromboembolicpulmonary hypertension. The clinical presentation isvariable. Patients will be short of breath on exertion.Other symptoms relate to decreased cardiac output,fatigue, syncope on exercise or in other situationsleading to a fall in systemic vascular resistance, suchas taking a hot shower. Chest pain, due to pleuriticinvolvement or to cardiac ischaemia, may beexperienced. There may be a history of previousDVT and/or PE, but frequently this is missing.Ultimately, patients develop severe hypoxia and right

heart failure with acute decompensation often fol-lowing trivial insult, e.g. urinary tract infection.

The physical findings are those of severe pul-monary hypertension. The diagnosis can be made bya combination of echocardiography, right heartcatheterization and pulmonary angiography and canbe further assessed with optic angioscopy, ultra-fastCT scanning or MRI. Supportive therapy consists ofanticoagulation, insertion of a vena cava filter, O2

and the judicious use of diuretics. Medical therapyaims to produce pulmonary vascular dilatation usingintravenous prostacyclin, oral calcium channelblockers or inhaled nitric oxide. These may be ofbenefit in preventing some of the secondary changessuch as endothelial cell dysfunction, acceleratedgrowth of pulmonary vascular smooth muscle cellsand vascular remodelling. The only definitive manage-ment available is pulmonary thromboendarterectomy(or transplantation). This former procedure iscompli-cated and carries a high risk for post-operative complications.38

CONCLUSION

Over the past 20 years, a large number of trials havebeen carried out on the diagnosis, prevention andtreatment of venous thromboembolism. Clinicalpractice has dramatically changed in response. Anumber of studies are currently underway to explorenew approaches to diagnosis and treatment. Theseinclude the use of pre-test clinical probabilities, therole of spiral CT and MRI and trials to determine theoptimal management using LMWH and newer anti-coagulants such as pentasaccharide and specific anti-thrombin agents.5 Longer duration of anticoagulanttherapy with LMWH or warfarin is aimed at decreas-ing recurrent DVT and PE and decreasing the inci-dence of the post-thrombotic syndrome. Clinicaltrials aimed at the prevention of DVT in high-riskmedical and surgical patients continue. Many of theremaining questions regarding management shouldbe answered in the near future.

REFERENCES

1. Anderson, FA, Wheeler, HB, Goldberg, RJ, et al. A

population-based perspective of the hospital incidence

and case-fatality rates of deep vein thrombosis within a

References 261

defined urban population. The Worcester DVT Study.

J Intern Med 1992; 232: 155-60.

2. Prandoni, P, Polistena, P, Bernard!, E, et al. Upper

extremity deep-vein thrombosis risk factors,

diagnosis, and complications. Arch Intern Med

1997; 157: 57-62.

3. Lane, DA, Mannucci, PM, Bauer, KA, et al. Inherited

thrombophilia: Part 1. Thromb Haemost 1996; 76(5):

651-62.

4. Lane, DA, Mannucci, PM, Bauer, KA, et al. Inherited

thrombophilia: Part 2. Thromb Haemost 1996; 76(6):

824-34.

5. Hull, RD, Pineo, GF. Clinical features of deep venous

thrombosis. In Venous thmmboemholism in an evidence-

based atlas, ed. RD Hull, G Raskob, GF Pineo. Armonk,

New York: Futura Publishing Co, 1996; 87-100.

6. American Thoracic Society. The diagnostic approach to

acute venous thromboembolism - clinical practice

guideline. Am J Respir Care Med 1999; 160: 1043-66.

7. Huisman, M, Buller, H, ten Cate, JW. Utility of

impedance plethysmography in the diagnosis of

recurrent deep-vein thrombosis. Arch Intern Med

1988; 148: 681-3.

8. Perrier, A, Desmarais, S, Miron, MJ, et al. Non-invasive

diagnosis of venous thromboembolism in outpatients.

Lancet 1999; 353(9148): 190-5.

9. Birdwell, BG, Raskob, GE, Whitsett, TL, et al. Predictive

value of compression ultrasonography for deep vein

thrombosis in symptomatic outpatients: clinical

implications of the site of vein noncompressibility.

Arch Intern Med 2000; 160(3): 309-13.

10. Rathburn, SW, Raskob, GE, Whitsett, TL Sensitivity and

specificity of helical computed tomography in the

diagnosis of pulmonary embolism: a systematic

review. Ann Intern Med 2000; 132(3): 227-32.

11. Meaney, JFM, Weg, JG, Chenevert, TL, et al. Diagnosis

of pulmonary embolism with magnetic resonance

angiography. N Engl J Med 1997; 336: 1422-7.

12. Hirsh, J, Warkentin, TE, Shaughnessy, SG, et al. Heparin

and low-molecular-weight heparin: mechanisms of

action, pharmacokinetics, dosing considerations,

monitoring, efficacy, and safety. Chest 2001;

119(Suppl. 1): 64S-94S.

13. Hull, RD, Raskob, GE, Brant, RF, et al. The importance

of initial heparin treatment on long-term clinical

outcomes of antithrombotic therapy. Arch Intern Med

1997; 157(10): 2317-21.

14. Campbell, NR, Hull, RD, Brant, R, et al. Different

effects of heparin in males and females. Clin Invest

Med 1998; 21(2): 71-8.

15. Campbell, NR, Hull, RD, Chang, SC, et al. Aging and

heparin-related bleeding. Arch Intern Med 1996;

156(8): 857-60.

16. Wheeler, A, Powell, L, Jaquiss, RD, Newman, JH.

Physician practices in the treatment of pulmonary

embolism and deep-venous thrombosis. Arch Intern

Med 1988; 148:1321-5.

17. Hull, RD, Raskob, GE, Rosenbloom, DR, et al. Optimal

therapeutic level of heparin therapy in patients with

venous thrombosis. Arch Intern Med 1992; 152:

1589-95.

18. Raschke, RA, Reilly, BM, Guidry, JR, et al. The weight-

based heparin dosing nomogram compared with a

'standard care' nomogram. Ann Intern Med 1993; 119:

874-81.19. Warkentin, TE. Heparin-induced thrombocytopenia

and its treatment./ Thromb Thrombolysis 2000; Suppl.

1:S29-35.

20. Hull, RD, Raskob, GE, Pineo, GF, et al. Subcutaneous

low-molecular-weight heparin compared with

continuous intravenous heparin in the treatment of

proximal-vein thrombosis. N Engl J Med 1992; 326:

975-82.

21. Gould, MK, Dembitzer, AD, Doyle, RL, Hastie, TJ,

Garber, AM. Low-molecular-weight heparins compared

with unfractionated heparin for treatment of acute

deep vein thrombosis - a meta-analysis of random-

ized, controlled trials. Ann Intern Med 1999; 130:

800-9.

22. Hyers, TM, Agnelli, G, Hull, RD, et al. Antithrombotic

therapy for venous thromboembolic disease. Chest

2001;119(Suppl. 1):176S-93S.

23. Schulman, S, Rhedin, AS, Lindmarker, P, et al. A

comparison of six months of oral anticoagulation

therapy after a first episode of venous thromboem-

bolism. N EnglJ Med 1995; 332: 1661-5.

24. Schulman, S, Granqvist, S, Holmstrom, M, et al. The

duration of oral anticoagulant therapy after a second

episode of venous thromboembolism. N Engl j Med

1997; 336: 393-8.

25. A Collaborative Study by the PIOPED Investigators.

Value of the ventilation/perfusion scan in acute

pulmonary embolism diagnosis (PIOPED). JAMA 1990;

263: 2753-9.

26. Arcasoy, SM, Kreit, JW. Thrombolytic therapy of

pulmonary embolism. Chest 1999; 155(6): 1695-707.

27. Kanter, DS, Mikkola, KM, Patel, SR, et al. Thrombolytic

therapy for pulmonary embolism. Frequency of

intracranial hemorrhage and associated risk factors.

Chest 1997; 111:1241-5.


28. Goldhaber, SZ, Haire, WD, Feldstein, ML, et al. Alteplase

versus heparin in acute pulmonary embolism: ran-

domized trial assessing right ventricular function and

pulmonary perfusion. Lancet 1993; 341: 507-11.

29. Come, PC, Kim, D, Parker, JA, et al. and participating

investigators. Early reversal of right ventricular

dysfunction in patients with acute pulmonary

embolism after treatment with intravenous tissue

plasminogen activator. J Am Coll Cardiol 1987; 10: 971-8.

30. del Campo, C. Pulmonary embolectomy: a review. Can

7Surg1985;-28:11.31. Dehring, DJ, Arens, JF. Pulmonary thromboembolism:

disease recognition and patient management.

Anesthesiology 1990; 73: 146.

32. Molina, JE, Hunter, DW, Yedlicka, JW, et al.

Thrombolytic therapy for postoperative pulmonary

embolism.Am J Surg 1992; 163: 375.

33. Timsit, JF, Reynaud, P, Meyer, G, et al. Pulmonary

embolectomy by catheter device in massive

pulmonary embolism. Chest 1991; 100: 655.

34. Greenfield, LJ. Evolution of venous interruption for

pulmonary thromboembolism. Arch Surg 1992;

127: 622.

35. Pacouret, G, Alison, D, Pettier, JM, et al. Free

floating thrombus and embolic risk in patients

with angiographically confirmed proximal deep

venous thrombosis. Arch Intern Med 1997;

157:305.

36. Baldridge, ED, Martin, MA, Welling, RE. Clinical

significance of free floating venous thrombi.

7 Vosc Surg1990; 11: 62-9.

37. Decousus, H, Leizorovitcz, A, Parent, F, et al. A

clinical trial of vena caval filters in the prevention

of pulmonary embolism in patients with proximal

deep-vein thrombosis. N Engl J Med 1998; 338(7):

409-15.

38. Jamieson, SW, Nomura, K. Indications for and the

results of pulmonary thromboendarterectomy for

thromboembolic pulmonary hypertension. Semin Vase

Surg2000; 13(3): 236-44.

20Organizational issues in respiratorycritical careADRIAN J WILLIAMS

Introduction 263

Definitions 264

Admission, discharge and re-admission policies 265

Organization and activity

Multidisciplinary critical care

References

266

268

269

INTRODUCTION

Intensive care is the first part of a continuum ofprogressive care (followed by intermediate care, self-care, long-term care and home care) that providesservices for patients with potentially recoverableconditions who can benefit from a higher level ofobservation and treatment than is available on a gen-eral hospital ward. This approach is usually consid-ered to have originated with the polio epidemic ofthe 1950s,1 during which patients who developedrespiratory failure were managed with negative-pressure ventilation using 'iron lungs'.2 Initially, theseventilators were located throughout the hospital and,amongst those with concomitant pharyngeal paraly-sis, mortality was over 90% in those early days, large-ly due to problems with aspiration, but also due tothe lack of expertise with assisted ventilation and thelack of monitoring. Ibsen measured arterial bloodgases and showed that the patients were dying fromventilatory failure. The improved understanding ofthe ventilatory status of patients, and the need forairway protection, led to the use of cuffed tra-cheostomy tubes and positive-pressure ventilation.The increased invasiveness of this treatment promptedcentralized care and, with this, mortality fell below

40%. Respiratory intensive care was established. Toquote Gilbertson:3

Ibsen's contribution in 1952, leading to a 50% reduc-

tion in mortality from combined respiratory and

pharyngeal paralysis, was a great step forward. Its

particular importance was that it revolutionised the

treatment of respiratory failure caused by the

scourge of the day, paralytic polio. The earlier prac-

titioners, with their 'Iron Lungs' and cuirasses, saved

a large number, if not a large proportion, of patients

with respiratory paralysis due to polio, and many

patients with previously fatal respiratory depression

from other causes. Intensive therapy has since

moved forward from simply providing respiratory

support to the management of multi-system failure,

but there can be no doubt that the modern specialty

owes its origins to the demonstration that patients

with failure of a vital organ could be kept alive by

mechanical support with skilled round-the-clock

nursing and medical care and could recover.

In subsequent years, intensive care remainedlargely defined by the need to provide invasive assist-ed ventilation against a backdrop of variable andoften multiple organ system failures, occurringas complications of severe sepsis, trauma andmajor surgery. Different units were managed either

264 Organizational issues in respiratory critical care

independently, often the case with the specialistsurgical units, or by a team with both generalphysicians and anaesthetists. In the USA in the 1980s,however, respiratory physicians became ever moreinvolved with the clinical practice of critical care, andtraining programmes rapidly evolved to encompassthe specialties of both respiratory medicine andcritical care. This development was paralleled bythe expansion of the critical care assembly of theAmerican Thoracic Society to become the largestof eight assemblies, and in January 1994, under theeditorial guidance of Robert Klock, the AmericanReview of Respiratory Diseases became the AmericanJournal of Respiratory and Critical Care Medicine. Thejournal welcomed manuscripts dealing with allaspects of critical care medicine, not just those relatedto the lungs. The respiratory physician with a back-ground in ventilation is increasingly well placed tounderstand the critically ill patient and, in particular,to provide respiratory critical care.

This chapter makes the case that the respiratoryphysician should be involved in the delivery of alllevels of critical care. With the increasing use ofacute non-invasive ventilation (NIV), and with out-reach teams identifying respiratory and circulatoryfailure much earlier on the wards, there is a growingneed for the respiratory physician to be trained andinvolved in critical care. Respiratory medicineskills, particularly in NIV and bronchoscopy, can besupplemented with training in airway managementand the circulation through training programmessuch as those existing in the USA. The respiratoryphysician will then be well placed to provide abridge between acute medicine and critical care.

DEFINITIONS

In the UK, the Department of Health has outlinedcategories of organ system monitoring and supportin the following way, all of which should be availablein intensive care.

• Advanced respiratory support: the provision ofmechanical ventilation via endotracheal tube ortracheostomy, excluding mask ventilation unlesscomplicated by other significant problems, or thepotential need for intubation such as withimpending respiratory failure and/or retainedbronchial secretions.

• Patients who require FiO2 > 0.4, those who maydeteriorate and require increasing FiO2 and/or2-hourly chest physiotherapy, those recentlyextubated and those requiring respiratory sup-port with continuous positive airways pressure(CPAP) or NIV by facemask. Some patients intub-ated for airway protection and otherwise stablewould also require this form of monitoring andsupport.

• Circulatory support: the use of vasoactive drugs,the monitoring of circulatory instability due tohypovolaemia not responding to modest volumereplacement, and following cardiac arrest.

• Neurologic monitoring and support: in the set-ting of central nervous system (CNS) depressionand with the capability of invasive intracranialmonitoring.

• Renal support: acute renal replacement treatmentvia various routes such as peritoneal dialysis andhaemofiltration.

The intensive care unit (ICU) is characterized by adesignated area where the minimum nurse:patientratio is 1:1 (not including the nurse in charge of theshift) and with 24-h resident medical cover. There isthe ability to provide advanced respiratory supportand to support other organ system failures such ascirculatory and renal failure as well as the manage-ment of co-morbidities.

In contrast, a high dependency unit (HDU) isdefined as an area capable of providing support ofone organ system excluding invasive ventilatorysupport, a level of care intermediate between thatof a general ward and the ICU. The HDU is charac-terized by a designated area with a minimumnurse:patient ratio of 1:2, appropriate monitoringfacilities and the continuous availability of medicalstaff. This concept of specialized high dependencycare had already been applied with the developmentof coronary care units (CCUs) and renal dialysisunits in the 1950s in the USA and in the 1960s in theUK. The ubiquitous presence of CCUs has meantthat cardiac disease in general (and myocardialinfarction in particular) is dealt with separately inalmost all hospitals. Renal dialysis units are lesscommon and, in their absence, this procedure usu-ally falls under the aegis of the ICU rather than theHDU. Another specialized unit that has existed inthe USA for more than 25 years is the respiratoryintensive care unit (RICU) or the pulmonary acute

Admission, discharge and re-admission policies 265

care unit. Here, complex respiratory support,including invasive ventilation, can be provided forcases of isolated respiratory failure and also limitedcirculatory support. The advantages of such an envir-onment include the ability to cultivate respiratoryexpertise in ancillary staff and to develop and utilizeweaning strategies. The existence of these units is inpart due to the fact that lone respiratory failureaccounts for 5-15% of a typical medical service'semergency work. They have also spawned regionalweaning centres where patients who have proveddifficult to liberate from assisted ventilation con-tinue to be treated outside the more expensive ICUand with an improved chance of weaning.4

ADMISSION, DISCHARGE ANDRE-ADMISSION POLICIES

Standards, policies and practice guidelines defineexpectations for clinical practice, create consistencyand promote continuity. Standards are defined as thelevel of performance that can or should be expected.Policies are a non-negotiable requirement represent-ing conditions that must exist to facilitate care of thestandard expected. They are directions for the unit'soperation. Practice guidelines (also referred to as clin-ical protocols) represent appropriate pathways forthe management of a specific clinical problem. TheAmerican Society of Critical Care Medicine providesexpert consensus recommendations on which 'bestpractice' may be based. The needs for such directivesare many, not least to reduce unnecessary expend-iture by providing a defence for limiting diagnostictesting and management strategies. By promotingthe use of best evidence, they should be viewed as animportant part of a total quality managementprogramme and not simply a cookbook.

Admission, discharge and re-admission policiesare crucial to the efficient running of the ICU,HDU and RICU. The aim of admission guidelinesis to exclude patients who are well enough to beadequately cared for in a general ward, or thosewho are hopelessly ill. Suggested admission anddischarge criteria were first published in the USAin 1988. In 1996, the Department of Health NHSExecutive published Guidelines on admission toand discharge from intensive care and high depen-dency units.5

Admission criteria

Referral to the ICU for specialist care may be madefrom many clinical areas, including the accident andemergency department, general medicine and surgicalwards, high dependency areas and occasionally fromother ICUs. It is usually considered appropriate forpatients requiring advanced respiratory support alone,or with combinations of two or more other organ sys-tem failures. Sufficient information must be madeavailable to allow an assessment of potential reversibil-ity, always erring on the side of optimism, and ofcomorbidities. Chronic impairment of one or moreorgans sufficient to restrict daily activities will weighagainst admission. However, patients may acceptlimitations that others may not. The need or potentialneed for ventilatory support or for other complexorgan system monitoring and/or support will beevident. High-risk surgical patients are another groupdeserving of intensive care, particularly because thereare ample data indicating that ensuring appropriateintravascular volume resuscitation and maintainingblood pressure and O2 delivery reduce morbidity andmortality in the early postoperative period.6 Periopera-tive optimization has also been shown to improveoutcome in a number of recent studies.

Admission to the HDU is appropriate for support ofa single organ system, excluding invasive ventilation,although non-invasive positive-pressure ventilatorysupport (NIPPV) can be instituted provided suitablemonitoring is available. However, the limited availabil-ity of such beds, in the UK in particular, means thatmany patients who may benefit from the services of theHDU may have to be admitted to a general ward. Thiswidespread practice in the UK is reflected by the factthat ethically approved studies have been conducted onthe use of NIV on the general ward in the treatmentof patients with acute respiratory failure.7 The adventof NIPPV has revolutionized the management ofrespiratory failure, but has produced a problem inidentifying suitable locations for treating patients inthis way and has provided added impetus to theconcept of the RICU. One-third of acute admissions tohospital are with respiratory problems and one-thirdof these are in patients with acute or acute-on-chronicrespiratory failure. Those patients not requiringimminent intubation and ventilation may benefitfrom early NIPPV, and an environment with expertisein this therapy may limit progression of the condition,thereby preventing the need for ICU admission. If


respiratory failure remains the only major problem,then such RICUs can provide invasive ventilation withthe advantages of a focused group of carers, nurses andtherapists who are familiar with NIV, and thisminimizes the period of intubation, partly due to theuse of therapist-supervised protocols. The experiencewhen both ICUs and RICUs are available is frequentlythat NIPPV is used to a greater extent in RICUs,perhaps for the reasons stated and due to the thoracictraining of the medical personnel.

Discharge criteria

Patients must be discharged sensibly from scarceICU beds to make room for more severely illpatients. The reported mortality after discharge fromintensive care is 9-27%.8,9 It may be assumed,although it is as yet unproven, that the dominant fac-tor determining the risk of hospital death is the phys-iological risk score, and a reduction to acceptablelevels should form the basis of a discharge. Theimportance of this has recently been reported byDaly et a/.,10 who developed a predictive model (withreference to patients' age, chronic health points,acute physiology points at discharge from the unit,length of stay in the unit and whether or not cardiacsurgery had been performed) with a cut-off of 0.6predicting subsequent mortality at 14% versus 1.5%.Those identified as at risk have a mortality of 25%and, if the model is valid, mortality after dischargefrom intensive care would be reduced by more thanone-third if those patients stayed an extra 2 days inintensive care.

Re-admission criteria

ICU outcome studies have identified re-admissionrates ranging from 4% to 10%, with 7.9% in the UK.It is important to identify patients at high risk of re-admission to allow them longer ICU stays or transferto an appropriate HDU (or RICU). The originaldiagnoses most frequently associated with a re-admission are hypoxic respiratory failure, inadequatepulmonary toilet and gastrointestinal bleeding. Thefirst two again raise the question of whether some ofthese patients should be managed from the outset ina specialized RICU.

ORGANIZATION AND ACTIVITY

The design and organisation should reflect thespecial needs of each specific group of critically illpatients. Current guidelines from national societiesare available, but it is recognized that these maychange. Open ward bays of four to eight beds aremost common, with additional single rooms in anappropriate ratio to provide facilities for barriernursing. The special HDU or single-discipline ICUmay be best placed close to the general ICU to enablesharing of facilities (such as laboratory and technicalstaff). The patient area needs to be spacious enoughto allow bedside services, hand-wash sinks should beprolific, and X-ray viewing boxes and emergencytrolleys evident. Nursing stations will be integral, andother staff areas, as well as support and storage facil-ities, are necessary.

Activity will in large part determine staffing, andthe relationship between the two is linked throughlevels of care. At a conference at Bethesda in 1983,levels of care were defined on the basis of averagenursing workload measured by means of aTherapeutic Intervention Scoring System (TISS).11

This provides the following levels of care:

• 40-45 TISS = 1 patient:! nurse = ICU,• 20-39 TISS = 1.5 patients:2 nurses = HDU,• 10-19 TISS = medium/high level care.

The TISS concept remains valuable and was rede-fined by Reis Miranda et al.12 as 1 point = 10.8 minof nurse work, therefore one nurse shift = 46 TISSpoints, a revalidation of the ICU/HDU nursingguidelines. Although each patientnurse ratio repre-sents a level of care, it is possible to aggregrate thevarious levels of care to three: level I: ratio >3; levelII: ratio average 2.5; level III: ratio <1.6.

Units might therefore be labelled not as ICU/HDUetc., but as in Table 20.1.

The actual activities involved in respiratory inten-sive care consist of regular and repetitive assessmentdesigned to detect events that might result in patientharm. They invariably include:

• arterial blood gases for respiratory and metabolicacid-base status,

• pulse oximetry for arterial O2saturation,• transcutaneous monitoring of CO2,• aspects of ventilation, respiratory muscle function

and respiratory mechanics.

Organization and Activity 267

Table 20.1 Levels of care defined according to nursedependence, resource usage and Acute Physiology Score

Grade IIIGrade IIGrade I

<1.61.7-2.9

>3

>2827-16

<15

>3837-25

<24

P/N, patient/nurse ratio; TISS, Therapeutic Intervention Scoring System;SAPS, Simplified Acute Physiology Score.

Pulmonary artery catheterization may result inuseful information at the expense of complications.When this should be performed is a balance of rea-sonable expectation that an event or change mightoccur and need to monitor therapy. Unnecessarymonitoring can create patient 'harm', with attendantcosts, device-related injury and, importantly, inappro-priate clinical response.

Funding of intensive care is variable across the UKand Europe and inevitably this has led to rationing,with a consequent increase in the severity of illness ofpatients admitted to the ICU. Admission of patientsto the general wards until their condition deterioratesto the point that intensive care is mandatory createsa situation in which too little is done too late andwhere the chance of reversing organ dysfunction isremote. This leads to inefficient use of resources fortreating those who are no longer able to benefit. Partof the reason for insufficient funding may be themultiple-specialty status that exists in Europe, whichleads to intensive care being a part of other specialtybudgets. However, changes are afoot, both in conti-nental Europe and in the UK, to allow intensive careto develop as a multidisciplinary subspecialty as inthe USA. The variability in organizational structureswithin European intensive care was evident from theEURICUS study,13 which also underscored the lackof independent specialty status for intensive care.

Partly as a result of these funding issues, import-ant modifications in the practice of respiratoryintensive care have recently emerged. The first ofthese is non-invasive ventilation in the acute care set-ting. The recent increase in the use of NIPPV in theacute care setting has been driven by the hope ofreducing the complications of invasive ventilationand the costs. In patients with exacerbations ofchronic obstructive airways disease, the largest single

diagnostic category, an average success rate in avoid-ing intubation with survival of 83% has beendemonstrated in five randomised, controlled tri-als,14-18 compared with 61% in controls. Predictorsof outcome have been assessed,19 and secretions, anedentulous state and cardiovascular instability havebeen found to be negative influences. However, ifassisted ventilation is indicated, it should first beattempted non-invasively using a facemask or nasalmask. The use of NIPPV has been extended to othercauses of acute respiratory failure, and an interna-tional consensus conference20 felt that 'NIPPV hasthe potential for reducing the morbidity and possiblythe mortality, associated with hypercarbic or hypox-aemic respiratory failure'; also that, 'available studiesindicate that NIPPV can be initiated outside theintensive care unit' and 'shortening weaning timeand avoiding reintubation represent promising indi-cations for NIPPV (see also Chapters 13 and 22).

In a recent meta-analysis, Keenan et al.21 concludedthat the evidence supports the use of NIPPV inchronic obstructive pulmonary disease (COPD).A remaining question is where such therapy mightbe instituted. Some studies have used the ICU,22

which limits the potential value in the 'real world',and another has examined the use of the general res-piratory ward.7 In the last-mentioned study, 232patients with COPD and a respiratory acidosis (pH7.25-7.35) in 14 UK hospitals were randomized toeither NIPPV or 'usual' treatment and a reducedneed for ventilation was demonstrated in the inter-vention group (15% versus 27%).

An alternative approach is to use specialized highdependency areas (the RICU) such as exist in Italy,23

where more than one-half of the patients admitted to26 units countrywide are treated with NIPPV as afirst line of treatment. RICUs have not been widelydeveloped in the UK, but in our experience they dohave a valuable part to play.

This changing clinical scene provides both theneed and the opportunity for the respiratory phys-ician to be trained in critical care and to provide abridge between acute medicine and intensive care.Indeed, the potential benefits of respiratory supportoutside the ICU include earlier intervention toprevent further respiratory deterioration and accessto support for those who would not otherwise beadmitted to the ICU. The selection of such patientsrequires a skilled team, including a respiratory phys-ician, and adequate monitoring.


New indications for NIPPV may include help withweaning from mechanical ventilation and the avoid-ance of re-intubation. The respiratory physicianpractised in the art of NIPPV, if also competent inthe management of the ventilated patient, is ideallyplaced to provide this. Although the majority ofstudies have been conducted in ICUs, NIPPVprovides an opportunity for delivering ventilatorysupport in RICUs or HDUs, where other acute physi-cians would also be involved.

Another development has been the creation ofcentres for weaning from mechanical ventilation.Ventilatory dependency in the ICU occurs in 20% ofpatients. Resolution of the disease process preventingweaning from respiratory support may still be possible,but the economic pressure to transfer the patients fromthe ICU as soon as possible is ever present. Choosingthe appropriate facility to which ICU patients shouldbe discharged will depend on the patient's clinical con-dition, the resources of the transfer destination andwhether weaning attempts will continue. Experiencewith continued attempts at liberation in long-termacute care facilities is accumulating.24,25

Weaning centres in the USA such as exist in LosAngeles at Barlow Hospital have demonstratedeconomic viability, with the same benefits beingsuggested in Italy, where these patients constitute12% of admissions to the RICUs.23

MULTIDISCIPLINARY CRITICAL CARE

Effective and efficient care needs the complementaryknowledge and skills of physicians, nurses andadministrators. For the ICU, there should be a singlemedical director as well as intensivists helped bysenior trainees and residents. The nursing com-plement also needs an identified lead nurse, withaccredited intensive care nurses, nurses in trainingand nursing assistants. Pharmacists, physiotherapistsand other therapists also have an important part toplay. The background of the medical and supportteams varies across hospitals and across countries. Inthe UK, most critical care facilities are staffed byanaesthetists. In many, though not all, instances theresponsible consultants continue to have regularsessions in the operating theatre, which prevent themfrom providing full-time ICU cover. At present, fewunits have full-time respiratory physician intensivists.However, as the training of respiratory physicians

comes to include intensive care, these physicians,when appointed to consultant posts, may have linkswith or input into the intensive care. Some will alsoextend their ICU training to include a period ofanaesthesia, currently recommended in the UK to bea minimum of 6 months, which will allow accredita-tion in intensive care medicine and the opportunityto direct an ICU.

In contrast, in the USA, 80-85% of medical ICUsare staffed by respiratory physicians with additionalcredentials (Boards in Critical Care Medicine). Inlarger (regional and/or teaching) hospitals, intensivecare medicine is subdivided by specialty (medical,coronary, obstetric, paediatric, surgical, with furthersubdivisions of surgery into cardiac versus general,and even organ transplantation specific), so that therespiratory critical care physician will usually only beresponsible for strictly medical problems, beinginvited to consult on medical issues in the other units.The training of respiratory critical care physicianstakes account of the limited experience of the medicalICU by including rotations through other surgicaland specialist critical care areas. In smaller hospitals,intensive care patients may be cared for in the sameunit, although by different specialists. Such units willusually have a respiratory medical care physician asmedical director, with powers to override the man-agement decisions of the specialists as clinical needdictates.

Nursing expertise in critical care environmentsincludes familiarity with mechanical ventilation andinvasive monitoring. The one-to-one or one-to-twonurse:patient ratio permits the nurse the time toprovide a wide range of treatments, includingrespiratory therapy. This has tended to limit thedevelopment of the physiotherapy practitioner inthe ICU. In some institutions, specialization as a res-piratory physiotherapist is encouraged, such thera-pists being more acquainted with mechanicalventilation and bronchoscopy, although stillwithout authority to alter ventilation management(see Chapter 6). The further specialization of suchindividuals into respiratory therapists has nothappened in the UK as it has in the USA. In theUSA, respiratory therapists supervise the ventilatorymanagement of patients under the direction of aphysician and are permitted to alter ventilatorystrategies according to strict protocols. Such'therapist-implemented protocols' have been used toextubate or wean patients in the ICU setting, with

References 269

evidence of a significant reduction in the periodof dependence or mechanical ventilation.12,24

At the other extreme, recent reviews of adult criticalcare services in the UK have re-emphasized theconcept of the 'patient at risk', for whom delay in therecognition of their condition and transfer to anappropriate area result in a poorer outcome. Bettertraining of junior medical and nursing staff, early-warning scoring systems and 'outreach' critical carehas been advocated.19 Goldhill has shown that physi-ologic abnormalities are common in the 24 hourspreceding admission to the ICU.23 Inadequate inputfrom experienced clinicians has been noted,24 as hasthe inadequate training in acute medicine that physi-cians in general have.25 The respiratory physician withintensive care training is ideally placed to address theseissues and to be part of, or even to direct, intensivecare teams that provide outreach services.

REFERENCES

1. Lassen, HCA. A preliminary report of the 1952

epidemic of polio in Copenhagen with special

reference to the treatment of acute respiratory

insufficiency. Lancet 1953; I: 37-41.

2. Williams, AJ. Iron lung. In The Oxford companion to

the body, ed. C Blakemore. Oxford: Oxford University

Press, 2001; 406-7.

3. Gilbertson, AA. Before intensive therapy? J R Soc

Med 1995; 88: 459-63.

4. Sirio, CA, Angus, DC, Rosenthal, GE. Cleveland Health

Quality Choice [CHQC] - an ongoing collaborative,

community-based outcomes assessment program.

NewHoriz 1994; 2: 321-5.

5. Department of Health. Guidelines on admission to and

discharge from intensive care and high dependency

units. London: NHS Executive, 1996.

6. Bishop, MH, Shoemaker, WC, Appel, PL, et al.

Prospective randomised trial of survivor values of

cardiac index, oxygen delivery and oxygen consumption

as resuscitation endpoints in severe trauma. 7 Trauma

1995; 38: 780-7.

7. Plant, PK, Owen, JL, Elliott, MW. Early use of non-

invasive ventilation for acute exacerbations of COPD

on general respiratory wards. A multicentre randomised

controlled trial. Lancet 2000; 355: 1931-5.

8. Munn, J, Willatts, SM, Tooley, MA. Health and

activity after intensive care. Anaesthesia 1995;

50: 1017-21.

9. Goodhill, DR, Sumner, A. Outcome of intensive care

patients in a group of British intensive care units. Crit

Care Med 1998; 26: 1337-45.

10. Daly, K, Beale, R, Chang, RWS. Reduction in mortality

after inappropriate early discharge from intensive care

unit: logistic regression triage model. fiM/2001;

322:1274-6.11. Cullen, DJ, Ferrara, 1C, Briggs, BA, et al. Survival,

hospitalisation charges and follow-up results in

critically ill patients. N Engl J Med 1976; 294:

982-7.

12. Reis Miranda, D, Ryan, DW, Schaufeli, WB, et al.

Organisation and management of intensive care. A

prospective study in 72 European countries. Berlin:

Springer, 1997.

13. Soohoo, G, Santiago, S, Williams, AJ. Nasal mechanical

ventilation for hypercapnoeic respiratory failure in

COPD. Crit Care 1994; 22: 1253-61.

14. Keenan, SP, Kernerman, PD, Cook, DJ, et al. The

effect of non-invasive positive pressure ventilation

on mortality in patients with acute respiratory

failure: a metaanalysis. Crit Care Med 1997;

25: 1685-92.15. Jasmer, RM, Luce, JM, Matthay, JA. Non-invasive

positive pressure ventilation with standard medical

therapy in hypercapnoeic acute respiratory failure.

Chest 1997; 111: 1672-8.

16. Confalonieri, M, Gorini, M, Ambrerino, C, et al.

Respiratory intensive care units in Italy: a national

census and prospective cohort study. Thorax 2001;

56: 373-8.

17. Scheinhorn, DJ, Chao, DC, Stearn-Hassenpflug, M,

et al. Outcomes in post ICU mechanical ventilation.

A therapist-implemented weaning protocol. Chest

2001; 119: 236-42.

18. Hamid, S, Noonan, YM, Williams, AJ, Davidson, AC. An

audit of weaning from mechanical ventilation in a UK

weaning centre. Thorax 1999; 54: 86.

19. Jaeschke, RZ, Meade, MO, Guyatt, GH, et al. How to

use diagnostic tests in the ICU; diagnosing weanability

using f/Vt. Crit Care Med 1997; 25: 1514-21.

20. Department of Health. Comprehensive critical care - a

review of adult critical care services. Report. London:

Department of Health, May 2000.

21. Department of Health and NHS Executive. Health

Service Circular [HSCj 2000/17. Modernising critical

care services. London: Department of Health, 23 May

2000.

22. Audit Commission. Critical to success. Report. London:

HMSO, October 1999.


23. Goldhill, DR, White, SA, Sumner, A. Physiologic values 25. Federation of Medical Royal Colleges Acute

and procedures in the 24 hours before ICU admission Medicine. The physician's role. Proposals for

from the ward. Anaesthesia 1999; 54: 529-34. the future. Report. London: Federation of

24. Neale, G. Risk management in the care of medical Royal Colleges, Publications Unit RCP,

emergencies. J R Coll Phys Lond 1998; 32: 125-9. February 2000.

21Ethical issues in the intensive care unitSEAN P KEENAN AND WILLIAM J SIBBALD

Introduction 271

Should this patient be admitted to theintensive care unit? 272

Once in the ICU, should life-support measuresbe continued or extended? 273

What role do advance directives play in decisionsabout ICU admission and care? 275

Under what circumstances is consent required in theICU and how should it be obtained? 276

Summary 276

References 277

INTRODUCTION

The purpose of this chapter is to summarize ethicalissues experienced on a regular basis by healthcareprofessionals working in the intensive care unit(ICU). Ethics has been defined as the discipline thatdeals with what is good and what is poor practiceand includes moral duty or obligation. While thisdefinition is open to interpretation, there is an addedfactor, generally recognized within medicine, of theobligation to conform to accepted professional stand-ards of conduct. The interpretation of good andpoor practices will vary among healthcare systemsand will partially depend upon the values of thesociety in which they exist. As a starting point, theethical principles that operate and influence clinicaldecision making are:

• patient autonomy, or the respect for a patients'right to self-direction, regardless of the opinion ofhealthcare providers,

• beneficence, or the obligation of healthcare work-ers to make decisions in the patient's best interest,

• non-maleficence, or the moral obligation to do noharm to the patient,

• justice, a concept originally attributed toAristotle, that reminds us that everyone should

be treated equally and not be discriminatedagainst.

Other chapters in this book provide informationthat respiratory critical care physicians need to makeimportant patient care decisions. To make gooddecisions, one requires both the knowledge of theunderlying pathophysiology of the critically ill andthe evidence (from the medical literature) regardingthe potential benefit or harm of diagnostic and thera-peutic options. Many decisions made in caring for thecritically ill do not require the clinician to spend muchtime considering ethical principles, for example in thedecision whether or not to perform a bronchoscopyfor suspected ventilator-associated pneumonia orsimply to initiate antibiotics. This chapter focuses onthose decisions that do call for consideration of ethicalprinciples and, therefore, require a different expertise.The following questions are covered.

• Should this patient be admitted to the ICU?• Once in the ICU, should life-support measures be

continued or extended?• What role do advance directives play in decisions

about ICU admission and care?• Under what circumstances is consent required in

the ICU and how should it be obtained?

272 Ethical issues in the intensive care unit

SHOULD THIS PATIENT BE ADMITTED TOTHE INTENSIVE CARE UNIT?

The ethics of intensive care unitadmission

Various societies, including the Society of CriticalCare Medicine, have addressed triaging criticallyill patients to the level of care that best meets theirneeds.1 Decisions regarding admission to the ICUincorporate all four ethical principles. One couldreasonably include the principles of beneficence andnon-maleficence in all decisions that medical practi-tioners make because our aim is always to help, notharm, our patients. We also respect patient autono-my, doing our best to follow a patient's wishes orthose expressed by his or her surrogate decisionmaker. At times, however, the principle of justice mayappear to be at odds with the others. ICU is a limit-ed resource and, as such, there are times when wefind ourselves acting not only as patient advocatesbut also as administrators who must 'manage' thisscarce resource. From the perspective of a physicianmanager, therefore, it may be appropriate to denyadmission to the ICU, despite the requests of familyor fellow clinicians, if it appears that a patient willnot benefit from this admission. But how do we know?

What is the role of the intensive careunit?

Although this may seem an odd question, it is a rea-sonable place to start when trying to define whoshould be admitted to an intensive care unit.Generally, intensive care units exist:

• to provide life support,• to provide intensive monitoring for patients who

do not currently require life support but whorequire interventions that may avoid the need forlife support, for example a patient with an activegastrointestinal bleed or a brittle diabetic withsevere ketoacidosis.

Each ICU admission comes with its own uniqueset of ethical issues. While it is often apparent whichpatient requires life support to continue to survive inthe short term, whether life support should beprovided for a specific patient may require serious

ethical deliberation. Prior to instituting life-supportmeasures, the following factors should, ideally, besatisfied.

• Is the institution of life support consistent withthe patient's wishes (patient autonomy)?

• Will the patient benefit from the institution of lifesupport in the short or long term (beneficence,non-maleficence)?

To answer the first question, a clinician (usuallynot an intensivist) should have the opportunity todiscuss life support with the patient, either duringa previous clinical encounter or at the time ofpresentation. In the latter case, patient competencemust be considered. Critically ill patients, even whileconscious, may not be able to give truly informedconsent; there may be impaired cognitive functionas a result of the critical illness or of medications,such as analgesics and sedatives. In these circum-stances, a surrogate, such as a family member orgeneral physician, may act on the patient's behalf. Itis important to note that the person who can act assurrogate varies amongst different healthcaresystems. Each practitioner must review with localauthorities the legislation that defines the appropri-ate surrogate for their patients. When decisions haveto be made quickly in the absence of explicitdirections, the default is to provide resuscitation andlife support until appropriate discussions can takeplace, i.e. presume 'life'.

The second question of patient benefit from lifesupport is more difficult and often cannot be fullyassessed prior to instituting life-support measures.Resuscitation may be required immediately (to sus-tain life) and before the clinician has a full under-standing of the process underlying the patient's acutepresentation. In other cases, however, there will betime for discussion of the benefits and detrimentsof initiating life support. A patient's prognosis maybe so poor that the initiation of life support may beinappropriate; the patient may therefore be consid-ered too ill for an ICU admission. Deciding not toadmit a patient to the ICU on the grounds that he orshe is too ill to benefit from admission is an increas-ingly common scenario for the critical care physician.The factors to take into account in considering admis-sion to an ICU include the following:1

• the likelihood of a successful outcome,• the patient's life expectancy due to the disease(s),

Once in the ICU, should life-support measures be continued or extended? 273

• the anticipated health status of the patient,• the wishes of the patient (or surrogate),• the burdens for those affected, including financial

and psychological,• missed opportunities to treat other patients,• health and other needs of the community,• individual and institutional moral and religious

values.

Another group of patients considered foradmission to the ICU are those who require intensivemonitoring (versus life-support) measures. Thesepatients are increasingly considered for triaging toa level of care other than the ICU - an example isa high dependency unit (HDU). In the past decade,we have seen the introduction of intermediate careunits designed specifically to care for patients requir-ing more intensive monitoring than is available onthe general ward. When such a resource is notavailable, decision making becomes more complex.The availability of ICU beds, alternate areas of careand the flexibility of the nurse:patient ratio on thegeneral ward will often dictate local practice. Whatmay be an inappropriate admission to the ICU inone institution may be justifiable in another.

Today's ICU physician, therefore, needs to be con-stantly alert to the ethical principles of patientautonomy, beneficence and non-maleficence whenadmitting patients. However, as the role of the ICUphysician has expanded from being a care providerto 'administrating' or 'managing' critical care beds,we find ourselves forced to consider the fourth prin-ciple of justice to a greater degree. This is not neces-sarily a bad thing: realizing that we cannot admiteveryone to the ICU, we have been forced to learnmore about the prognosis of different patientgroups. In the end, knowledge of prognosis, drivenby the principle of justice, should lead to decisionsthat are also in our patients' best interests.

ONCE IN THE ICU, SHOULD LIFE-SUPPORTMEASURES BE CONTINUED OREXTENDED?

Withholding and withdrawal of lifesupport

A major ethical decision confronting critical careprofessionals is whether, at some point in the

patient's ICU stay, life-support measures should belimited or withdrawn. Just as with an ICU admis-sion, this decision also incorporates the four ethicalprinciples. Care for the critically ill patient isinvasive, stressful and at times uncomfortable andpainful. In some circumstances, continuing full lifesupport could be considered contrary to the prin-ciple of non-maleficence - but it could be justifiedif the eventual outcome is considered to be good(beneficence). With an increasing understandingof the natural history and prognosis of criticalillness, critical care professionals may find them-selves caring for patients they know to have littlechance of survival.

Some patients die early in their ICU stay despiteaggressive therapy and before there is any need toreach a consensus that the outlook is hopeless, butthese are a minority. More often in the ICU webecome aware that a patient's outcome is extremelypoor after the institution of life support, and thissame patient may require addition of further lifesupport measures to 'survive' in the short term. Atthis time a decision must be made about whetherthe process of adding or continuing life-supportmeasures is appropriate. lust as before, we act in thepatient's best interest (beneficence), while avoidingharm (non-maleficence), and are guided by thepatient's wishes (patient autonomy). Possible deci-sions in this setting include:

• continuing full aggressive management,• continuing all measures, but limiting resuscita-

tion in the event of a cardiac arrest,• maintaining, but not escalating, the current level

of life support (withholding life support),• withdrawal of some, or all, current life support.

In most cases, discussion with other clinicians andfamily members leads to a consensus that is consistentwith autonomy, beneficence and non-maleficence.At times, however, differences of opinion arise and clin-icians find themselves in a situation in which patientautonomy, expressed by surrogate decision makers,is at odds with perceived principles of beneficence,non-maleficence and justice. Justice could then beviolated when resources needed by other patients arededicated to someone who has no chance of recovery.Sometimes, the pressure to continue care comes fromother clinicians rather than from family members. Thismay particularly be the case following iatrogenic comp-lications or critical illness following routine surgery.


What is happening now?

In a landmark paper, Smedira and colleagues docu-mented end of life care for two San Francisco ICUs inthe late 1980s.2 At that time, 51% of deaths followedthe decision either to withhold or withdraw life sup-port (although this high figure was partly due to thefact that brain-dead patients were included in thestudy). In a follow-up study 5 years later, the samegroup documented a significant increase to 90% inthe proportion of patients who died following with-holding or withdrawal of life support.3 Reports frommany centres over the past decade note that approx-imately 70-80% of patients die following limiting orwithdrawal of life support. Because such studies donot provide a complete view of practice in general,Cook and colleagues used hypothetical clinical casescenarios to demonstrate the variation amonghealthcare professionals in decision making aboutlife-support limitation.4 Factors considered mostimportant included: the likelihood of the patientsurviving the current episode, the likelihood of long-term survival, pre-morbid cognitive function andpatient age.

In a related study by the same group, healthcareworkers choosing either extreme levels of care (fullaggressive management) or comfort measures with inthe same patient scenario were equally confident intheir decision.5 This scenario-based study reportedwhat clinicians say they would do (rather than whatthey actually do), but it was consistent with anotherstudy that included data from 131 different ICUsfrom 38 US states.6 After excluding brain-deadpatients, this study reported that 26% of patientsdied despite full aggressive therapy, 24% diedwithout cardiopulmonary resuscitation, 14% died fol-lowing limitation of life support and 36% diedfollowing the withdrawal of life support. The mostinteresting findings were the differences amongstinstitutions. For example, death following the with-drawal of life support varied from 0% to 79%. It isinteresting that the rate of withdrawal of life supportwas lowest in the two US states with strict legal stand-ards for surrogate involvement in the decision-making process (New York and Missouri). Themedico-legal environment may therefore lead topractices that might be considered not to be in thepatients' best interest.

Another US survey also found variation in prac-tice among respondents. In this survey, 12% stated

that they had withdrawn life support without theknowledge of the family and 3% had withdrawn lifesupport against the objections of the family.7 Thesefindings highlight the potential ethical dilemma thatarises when physicians feel that their obligation tothe principles of beneficence and non-maleficencerun contrary to the wishes of the family. Whereas itis generally believed that physicians are not obligatedto provide futile care, many are probably uncomfort-able acting in opposition to family requests.However, this survey demonstrates that physicianswill act against family wishes if they believe both thatfurther care is futile and that they are supported bytheir healthcare system.

The withholding and withdrawal of life supporthave been deemed to be ethically the same by theUnited States President's Commission, a positionendorsed by most professional societies (includingthe Society of Critical Care Medicine, the BritishMedical Society and the American College of ChestPhysicians). Despite these endorsements, a signifi-cant number of healthcare workers do not view thesetwo actions as being equivalent. In a survey by theEthics Committee of the Society of Critical CareMedicine, 43% of respondents believed that with-holding life support was more acceptable than with-drawing life support, and 26% admitted to beingmore disturbed by withdrawing life support than bysimply withholding it. We found that withholdinglife support was more common in communityhospitals than in teaching centres, consistent withfindings that greater experience with end of lifedecision making is associated with a higher level ofwithdrawal of life support.8

If one considers the withholding and withdrawalof life support to be ethically the same, then continu-ing some life-support measures (but not others)could be perceived as acting contrary to the patient'sbest interests, violating the principles of beneficenceand non-maleficence.

Once a decision has been reached to withdraw lifesupport, which modalities should be withdrawn andin what order? We have noted that official statementsof professional societies and ethicists suggest allforms of withdrawal of life support are ethically thesame. Despite this, analysis of current practice clear-ly shows that there is more widespread comfort inwithdrawing haemofiltration or inotropic supportversus ventilation, nutrition and fluid therapy. Somepractitioners may feel that the withdrawal of fluid

What role do advance directives play in decisions about ICU admission and care? 275

therapy is cruel, but there is no evidence that patientssuffer as a result if the process is accompaniedby careful attention to the use of sedation andnarcotics. In the UK, the British Medical Associationaffirms that withholding fluids and nutrition doesnot constitute the removal of basic medical care(which should always be provided) as nasogastricfeeding is seen as artificial, i.e. a therapy.Nevertheless, it recommends resort to the courts, forthe time being, before such removal in cases ofpersistent vegetative state and stresses the need formulti-professional agreement to such discontinua-tion of therapy.9

The approach used to withdrawing life supporthas also been found to vary. In general, patientsreceive adequate medication in the form of opiates,with or without sedation, and life-support modal-ities are withdrawn at varying rates, sequentially orall together. While the ethically most correctapproach continues to be debated, we support anapproach that is as rapid as patient (and sometimesfamily) comfort will allow.

The approach to limiting or withdrawing lifesupport also varies among healthcare systems. Thissection has focused primarily upon practice asreported in North American centers. Compared tothe USA, critical care physicians in both the UK andAustralia historically seem to have adopted a morepaternalistic approach to decision making. This maynow change as a result of implementation of theEuropean Human Rights Act. Guidance from theBritish Medical Association on withdrawing andwithholding9 provides a clear analysis of the ethicalissues involved, especially for incompetent individu-als (through illness or lack of understanding) and,compared with previous practice, is more patientcentred.

In a study examining the attitudes of criticalcare physicians, ethical questions were approacheddifferently in 16 different European countries.10

Differences were partially related to geography, withsouthern countries tending to be more conservativethan the northern ones. This may reflect differencesin religious bias, with southern countries comprisinga greater proportion of Catholics compared toProtestants. 'Do not resuscitate' orders after a cardiacarrest are more common in the north, with theNetherlands leading the way at 91%, compared to alow of 8% in Italy. Similar differences in geographywere found concerning the frequency of withholding

and withdrawal of life support and the use of drugadministration to precipitate death. It should benoted that although physicians may claim a specificreligious affiliation, they may or may not adhere toits beliefs. Both withdrawal of life support and theadministration of drugs to hasten death were lesscommon among physicians viewing themselves asreligious.

WHAT ROLE DO ADVANCE DIRECTIVESPLAY IN DECISIONS ABOUT ICUADMISSION AND CARE?

An important development is the increased preva-lence of advanced care planning and the documentsthat have been developed around this issue.11 Thelatter are referred to as 'advanced directives' and arecomprised of an individual's wishes, in writing,regarding the use of life-support measures thatshould/should not be provided in the event that heor she is no longer able to participate in such deci-sions. In the interests of the ethical principle ofpatient autonomy, the development of advanceddirectives provides a means for patients to discusspotential health scenarios and to express their wishesregarding the use of life-support measures. In gener-al, advanced directives designate a surrogate decisionmaker for a patient who is either unconscious ordeemed incompetent.

In theory, advanced directives are a good idea andensure patient autonomy, but problems still remain.An advanced directive is only useful if it is extensiveenough to cover all potential scenarios. For anadvanced directive to be effective requires bothconsiderable time on the part of the physician (oftenmore than is available) and physician knowledge ofthe clinical scenarios (which may be beyond his orher area of expertise).

Whereas some advanced directives are compre-hensive, others provide less specific statements, suchas 'I do not wish to be kept alive on life support ifthere is no hope of recovery'. (Perhaps it is a signof how far we have progressed that most ICU physi-cians would find such a statement to be unhelpful!)The days of protracted life support for patientswithout hope of recovery have largely disappearedin North America and Europe. ICU physicians aregenerally comfortable with advising the withdrawal


of care under such circumstances, and society isgenerally supportive. However, a statement ofpatient's wishes when there is no hope of recoveryis only a starting point. In reality, critical careprofessionals work in a setting that deals in prob-abilities rather than certainties. The question is: inthe scenario with such-and-such a probability ofsurviving hospital stay, with such-and-such a rangeof potential health status, and knowing the priorquality of life, would the patient want life supportcontinued, resuscitation, etc? To make the situationmore difficult, the scenario changes with time andalters the prognosis.

Although there are potential pitfalls in advanceddirectives, they are clearly important and willbecome increasingly useful as those involved (physi-cians, lawyers, family and patients) work together.There is clearly a desire among both physiciansand the general public for their development. ASUPPORT study12 underlined the fact that physi-cians are often not aware of their patients' wishesregarding resuscitation and life-support measures.One helpful development in advanced directives isthe introduction of decision aids. These consist ofwell-thought-out scenarios that allow future patientsto make decisions in the setting of uncertainty thatreflects the reality of medicine. Although still notperfect, they add information about how patientsfeel about the use of life-support measures.

UNDER WHAT CIRCUMSTANCES ISCONSENT REQUIRED IN THE ICU ANDHOW SHOULD IT BE OBTAINED?

Consent in medicine is generally sought for twodifferent reasons:

• to administer a therapy or to conduct a diag-nostic test for clinical reasons,

• to enroll a patient in a research protocol.

In both cases, permission is ideally sought from acompetent patient. However, critically ill patientswho are not conscious or are clearly disoriented andconfused are appropriately considered incompetentand cannot give informed consent. Other critically illpatients who may appear alert and oriented, and thuscompetent, may not recall any events from their ICUstay. It is not possible to foresee which patients will or

will not recall conversations. Although these patientsare usually considered competent to give consent, itmay be best to involve family members in the process(assent) to ensure that decisions are felt to be consis-tent with the patient's pre-morbid character.

Patients who are considered incompetent require asurrogate decision maker. In some US states andCanadian provinces, legislation exists that clearlydefines how to identify an appropriate surrogate. Asmentioned previously, advanced directives can helpto identify the patient's chosen surrogate. In the set-ting in which no family exists, these states orprovinces will appoint a surrogate to act on thepatient's behalf. Where there are no formal proced-ures for the designation of a surrogate decisionmaker, the issue of consent is approached in a varietyof ways, but usually involves physicians acting on thepatient's behalf.

The issue of consent to participate in a researchstudy requires even greater care, as there is no evi-dence that critically ill patients will benefit from theirinvolvement in the study, and they may even beharmed. The principles of patient autonomy (respectfor patient choice, obtaining informed consent andmaintaining confidentiality), beneficence (favourablebalance of potential benefit and harm) and justice(all must be equally likely to be involved) must beapplied.11 In addition, the study must be scientificallyvalid, address a question of sufficient value, beconducted honestly and be reported accurately andpromptly.11

SUMMARY

This chapter highlights the main ethical issues thatare involved in caring for the acutely unwell. Clearly,on-going study of practice and how we makethese sometimes difficult decisions is required.Researchers in this area have provided us with aframework, but further work is needed to under-stand how we can best serve our patients. Our deci-sions will become easier as the public becomes moreaware of what current medical technology can andcannot achieve. Although practice will always varyamong healthcare systems, and even amongst phys-icians, it is hoped that greater attention to ethicalissues will make these decisions less difficult for allin the future.

References 277

REFERENCES

1. Society of Critical Care Medicine Ethics Committee.

Consensus statement on the triage of critically ill

patients. JAMA 1994; 271: 1200-3.

2. Smedira, NG, Evans, BH, Grais, LS, et al. Withholding

and withdrawal of life support from the critically ill.

N EnglJ Med 1990; 322: 309-15.

3. Prendergast, TJ, Luce, JM. Increasing incidence of

withholding and withdrawal of life support from the

critically \\\. Am J Respir Crit Care Med 1997; 155:15-20.

4. Cook, DJ, Guyatt, GH, Jaeschke, R, et al. Determinants

in Canadian health care workers of the decision to

withdraw life support from the critically \\\.JAMA

1995;273:703-8.

5. Walter, SD, Cook, DJ, Guyatt, GH, et al. Confidence in life-

support decisions in the intensive care unit: a survey of

healthcare workers. Crit Care Med 1998; 26: 44-9.

6. Prendergast, TJ, Claessens, MT, Luce, JM. A national

survey of end-of-life care for critically ill patients. Am J


7. Asch, DA, Hansen-Flaschen, J, Lanken, PN. Decisions to

limit or continue life-sustaining treatment by critical

care physicians in the United States: conflicts between

physicians' practices and patients' wishes. Am J Respir

Crit Care Med 1995; 151: 288-92.

8. Keenan, SP, Busche, KD, Chen, LM, et al. Withdrawal of

withholding of life support in the intensive care unit:

a comparison of teaching and community hospitals.

Crit Care Med 1998; 26: 245-51.

9. Withholding and withdrawing life-prolonging medicaltreatment, guidance for decision making, 2nd edition.London: BMJ Books, 2001.

10. Vincent, JL. Forgoing life support in western European

intensive care units: the results of an ethical

questionnaire. Crit Care Med 1999; 27: 1626-33.

11. Singer, PA. Bioethics at the bedside: a clinician's guide.

Ottawa: Canadian Medical Association, 1999.

12. The SUPPORT Principal Investigators. A controlled trial

to improve care for seriously ill hospitalized patients:

the study to understand prognoses and preferences

for outcomes and risks of treatments (SUPPORT).JAMA

1995; 274: 1591-8.13. The Acute Respiratory Distress Syndrome Network.

Ventilation with lower tidal volumes for acute lung

injury and the acute respiratory distress syndrome. N

Engl j Med 2000; 342: 1301-8.

22Respiratory failure: new horizons,new challengesA CRAIG DAVIDSON AND DAVID F TREACHER

Introduction 278

Recognizing respiratory failure 278

Recognizing the need for assisted ventilation inchronic obstructive pulmonary disease 279

Determining appropriateness of assisted ventilation 280

Weaning in Type 2 respiratory failure 281

Refractory hypoxaemia 281

References 291

INTRODUCTION

This final chapter reviews the current managementof acute respiratory failure and discusses some of theareas of controversy identified in the preceding chap-ters. The management of acute respiratory failure is,at present, undergoing re-evaluation in the light ofrecent research and technological developments.Similar to the way in which 'clot-busting' therapy hasimpacted on cardiology, non-invasive ventilation(NIV) is acting as the catalyst for this change. 'Patientat-risk teams' (PART)1 and 'medical emergencyteams'2 have developed in response to the evidencethat the care of the ward patient whose condition isdeteriorating may be inadequate and the delay inreferral to the intensive care unit (ICU) may be fatal.An important aspect of the experience of PARTteams is that respiratory distress is recognized as afrequent feature of the deteriorating ward patient.3'4

Its presence should trigger medical review andprompt a decision regarding the need for respiratorysupport and whether the use of non-invasive respira-tory support in the high dependency area is appro-priate or whether referral to intensive care isnecessary. Recommendations for 'seamless' care anddissolution of the historical barriers between ICUs

and general wards are compelling.5 The details ofdelivery will inevitably vary amongst hospitals andamongst different countries, but, at the interfacebetween the admitting clinician and the ICU, a sys-tem needs to be in place that identifies 'at-risk'patients, monitors their well-being, appropriately'steps up' care when necessary and carefully observespatients following their discharge from the ICU.

RECOGNIZING RESPIRATORY FAILURE

Classical medical teaching divides respiratory failureinto two forms (Table 22.1). Both are characterizedby a low .PaO2 (by definition, < 8 Kpa or 60 mmHg).In type 1 there is alveolar hyperventilation and intype 2 alveolar hypoventilation. The patient withtype 1 respiratory failure will usually present withobvious respiratory distress, whereas recognition oftype 2 failure may be more difficult. Assessment ofseverity, which will be necessary to determine care,depends on both the response to an increase in theinspired oxygen concentration (FiO2), because thisreflects the, size of the disturbance of gas exchange(the A-a gradient), and the underlying cause ofrespiratory distress along with any other associated

Recognizing the need for assisted ventilation in chronic obstructive pulmonary disease 279

Table 22.1 Causes of respiratory failure

Type 1 respiratory failure (primary abnormality: ventilation/perfusion mismatch)Acute lung injury/acute respiratory distress syndromeCardiogenic pulmonary oedemaPneumoniaPulmonary embolismPneumothoraxAsthma (moderate)Interstitial lung disease

Type 2 respiratory failure (primary abnormality: respiratory pump failure)Central nervous system encephalopathy, e.g. cerebrovascular accident, opiates, head injuryPeripheral neurological causes, e.g. myasthenia gravis, Guillain-Barre, cervical spine injury, poliomyelitisMuscle disease, e.g. muscular dystrophy, critical illness myopathyChest-wall disease, e.g. scoliosis, thoracoplasty, traumaSevere left ventricular failure or asthmaChronic obstructive airways disease

chronic illnesses. For instance, if the young man withpneumonia requires an FiO2 of 60% to maintainSaO2 >90% and he has associated organ dys-function, e.g. hypotension, his care demands ICUadmission. On the other hand, in the patient witha malignant pleural effusion, a similar degree ofrespiratory distress would indicate the urgent needfor pleurocentesis rather than transfer to a higherdependency area.

The patient with type 2 respiratory failure isfrequently more difficult to identify and it can beparticularly difficult to judge the risk of deterior-ation with the final outcome of respiratory arrest.In patients with chronic neuromuscular or chestwall disease, the onset of respiratory failure may beinsidious. Life-threatening episodes can be precipi-tated by minor illness, yet the presentation is oftenwith confusion and hypersomnolence rather thanbreathlessness. Examination of arterial blood gasesis therefore critical in evaluation. In the patientwith chronic respiratory failure, renal compensationwill be evident by an increase in buffer capacity.Back titration to a normal pH then providesa guide to the 'normal' PCO2. As a rough guide, anacute increase in PaCO2 of 1 kPa decreases pHby 0.06 units, whereas in chronic hypercapnia,with renal adjustment, it decreases pH by 0.02units.6 In our experience, confusion often ariseswhen there is a co-existent circulatory or septiccause for acidosis - hence the importance ofcalculating the anion gap, measuring serum lactateand testing for urinary ketones to exclude diabeticketoacidosis.

If respiratory failure develops as a result of centralnervous system (encephalopathy, trauma, infectionetc.) or peripheral nerve disease (Guillain-Barre,myasthenia gravis), hypoxaemia develops at a laterstage in the presence of a normal A-a gradient. It iscaused by marked hypoventilation as the vital capac-ity (VC) falls below 1 L and is reflected by progressivehypercapnia. Clinical suspicion and monitoring ofserial VC and arterial blood gases are thereforeimportant aspects of the care of such patients. It is inthe 'yet to be diagnosed' patient that the risk of sud-den deterioration is greatest. Here, the unsupervised,or unthinking, use of supplemental O2 may makeappropriate assessment less likely as the nurse, orjunior doctor, is falsely reassured by a 'normal'oximeter reading.

RECOGNIZING THE NEED FOR ASSISTEDVENTILATION IN CHRONIC OBSTRUCTIVEPULMONARY DISEASE

Confusion arising from the use of inappropriateO2 therapy is more commonly a feature ofadvanced respiratory failure resulting from chronicobstructive pulmonary disease (COPD). Themechanisms of O2 toxicity in these circumstancesremain controversial. Three factors are implicated:a reduction in central drive, a deterioration ingas exchange resulting from increased shunt,secondary to relief of hypoxic vasoconstriction,and a change to a less fatiguing but excessively

280 Respiratory failure: new horizons, new challenges

rapid breathing pattern all contribute.7- 9 Thepathophysiological mechanisms of CO2 retentionin COPD include hyperinflation impairing pumpefficiency, an abnormal respiratory pattern, withadditional dynamic hyperinflation as expiratorytime shortens, and resistive work resulting from airflow obstruction. Although minute ventilation isincreased, alveolar ventilation falls as load exceedsthe capacity of the ventilatory pump. Worseningrespiratory failure disturbs sleep so that in extremisit is difficult to distinguish central fatigue fromelectrophysiological respiratory muscle failure.Untreated, the tachypnoeic agitated patient lapsesinto confusion, with a fall in respiratory effort andterminal bradycardia and asystolic cardiac arrest.At what stage in this process should the clinicianinitiate ventilatory support?

The risk of death or of reaching the point ofunequivocal need for endotracheal intubation isbetter predicted by pH and PCO2 than by thedegree of hypoxaemia For instance, in trials assess-ing the impact of NIV, pH has been found to bepredictive. In the Brochard et al. study, 10 74% ofpatients with a mean pH 7.26 who were managedconventionally required intubation and the subse-quent in-hospital mortality was 29%. In the Plantet al. study, 11 27% of those with pH 7.25-7.35reached the intubation criteria and had a hospitalmortality of 20%. In another survey of outcome,Soo Hoo et al. reported a 54% intubation rate,which increased to 72% when pH <7.2.12

Subsequent multivariate analysis from the Plantstudy revealed that pH and PCO2 both contributeto risk, although the sensitivity and specificity ofthese factors alone do not allow sufficiently accur-ate prediction for application on an individualbasis.13 For instance, the odds ratio for a patientwith pH 7.30 and PCO2 8 kPa was 3.84, comparedto 16.8 for pH 7.25 and PCO2 10 kPa. Data fromthis study also showed that pH often improvesbetween the accident and emergency departmentand ward admission with conventional non-venti-lator management. In the absence of a clear needfor intubation, such as a Glasgow Coma Score <8,respiratory rate >40 or <10 min-1, conservativetherapy should therefore be initially employed. It isthe failure to improve, or further deterioration,that signals the need for assisted ventilation.

In Chapter 5, Dr Schonhofer reviewed the evi-dence for using NIV in respiratory failure for both

COPD and non-COPD causes. What is becomingapparent is that the patient whose condition stabil-izes or improves with NIV has a better prognosis.14

Part of the benefit clearly arises from avoidance ofthe complications of intubation and the risksinvolved with intensive care admission. Failure withNIV is an adverse prognostic sign, presumablybecause it indicates more serious underlying patho-physiology. However, there is the danger, however,that inappropriate use of NIV, and delayed recourseto intubation when it is failing, might also be respon-sible. One of the recommendations made by theBritish Thoracic Society in its guidelines for acuteNIV, 15 and in a consensus document on NIV, 16

is that clear limits should be set to define whatconstitutes a 'trial of therapy'. Explicit in theserecommendations is the requirement to determinethe appropriateness of aggressive management,including intubation and admission to the ICU.

DETERMINING APPROPRIATENESS OFASSISTED VENTILATION

Guidelines have been published to assist clinicians inthe UK in the difficult area of limiting or withholdingtherapy.17 It is an area, as described in Chapter 21, thatis culturally and religiously determined and that is fastchanging. Despite the considerable investment in timeand training required, a more open discussion withpatients to determine their wishes is to be encour-aged.18 In many cases, however, this may not be possi-ble because the patients' condition renders them'non-competent'. In the absence of advance directives,decision-making then rests with the clinician. Whatguidance is available? In a large European and USstudy, involving 1426 patients, survival followingmechanical ventilation for respiratory failure (all causes)was 55%.19 Depending upon the severity of lunginjury, survival ranged from 18% to 67%; multi-organfailure carried a 90% mortality rate. Outcome is crit-ically dependent on the population of patients includ-ed in studies and this is particularly the case in COPD.Of those admitted to hospital, and therefore includingmilder cases, approximately 10% will die and the1 year survival is reported at 58%.20 Of those notrequiring ventilatory support, 57-88% are reported tosurvive to 1 year,21 falling to 34-56% following intub-ation and a period of mechanical ventilation.22, 23


A better outcome (58-87%) has been reported fol-lowing mechanical ventilation with NIV.13,24, 25 Age,severity of airflow obstruction and presence ofchronic respiratory failure largely determine survivalin COPD. It remains to be seen whether prognosis isreally affected by avoidance of intubation by NPV andwhether prognosis can be further improved withdomiciliary NIV in these at-risk patients.26,27

WEANING IN TYPE 2 RESPIRATORYFAILURE

In Chapter 13, Dr Nava and colleagues consideredthe weaning process through which patients are lib-erated from mechanical ventilation. Although proto-cols allow earlier identification of the patient readyto breath spontaneously, and possibly speed theprocess,28, 29 there is no evidence to indicate that anyone method is indisputably superior to another inspeeding the process - although synchronized inter-mittent mandatory ventilation (SIMV) seems to bethe least favoured. Again, NIV is changing practice.For COPD patients, there is often a 'window ofopportunity' early in their ICU stay, when the acutephysiology has been corrected, sleep restored anddynamic hyperinflation and airway obstructionreduced and before sedative drugs have accumulated,when they may be extubated onto NIV. Although theevidence is at present limited, NIV may allow speed-ier extubation at a time when more conventionalindices of 'weanability' are not satisfied.30,31

Physiological studies certainly show that NIV off-loads the respiratory muscles as effectively as invasivepressure support and is preferred by patients.32 Evenif extubation criteria are met, patients with underly-ing COPD have a an incidence of post-extubationfailure of over 20%, and again NIV may be useful inpreventing re-intubation.33 Sometimes, glottic orsupraglottic narrowing contributes to post-extuba-tion respiratory distress and these aspects will not bedetected by conventional extubation criteria . Inaddition to treating upper airway narrowing, withsteroids or nebulized adrenaline, helium/O2 mix-tures may be useful.34 Certainly, avoiding re-intuba-tion is important. Such patients have a highermortality and this is, to some extent, explained byexposure to the risks of further nosocomial infec-tion. Re-intubation almost invariably leads to theinsertion of a tracheostomy and, with this, poten-

tially a more prolonged ICU stay. In these circum-stances, the availability of a coordinated weaningservice is helpful and, if available, a respiratoryintensive care unit (RICU) or a high dependency unit(HDU) would be the most suitable location forcontinued weaning and rehabilitation.

Although severe COPD is a common reason fordelayed weaning, underlying neuromuscular orchest-wall disease, and co-morbidity with cardiovas-cular or central nervous system problems complicat-ing the initial reason for ICU admission, are alsofrequent reasons for patients remaining ventilatordependent.35 In 153 weaning referrals to our special-ist respiratory ICU between 1998 and 2000, there wasa 73% overall survival to discharge, with 1-year and2-year survivals of 63% and 50%, respectively.Interestingly, 54% of patients referred from outsidehospitals required ventilatory support at discharge,compared with 22% for 'in-house' referrals, reflect-ing both the longer duration of ICU stay beforereferral and the higher incidence of neuromusculardisease among outside referrals.36 It would be usefulif there were reliable prognostic features for predict-ing eventual success at weaning in this type ofpatient. The obvious factors, such as age, presence ofco-morbidity and pre-morbid level of functioning,are not sufficiently predictive and we therefore placemore importance on patient motivation and familysupport. In the UK, if weaning eventually provesunsuccessful, the option of continued invasivemechanical ventilation in an institution is limited,and domiciliary care is critically dependent upon theprovision of suitable carers and acquiring sufficientfunding. Long-term care facilities are available in theUSA,37 and it would appear that support in the restof Europe for these patients is better and more cen-trally co-ordinated than in the UK. Provision variesconsiderably amongst countries, with few COPDpatients receiving mechanical ventilation at home inthe Netherlands but contributing a rapidly risingproportion to the total home-care population inFrance and Germany.

REFRACTORY HYPOXAEMIA

For the purpose of this discussion, refractory hypo-xaemia is defined as an arterial oxygen tension thatis <10 kPa (75 mmHg), with an inspired oxygentension (FiO2) of at least 0.6, which has failed to


respond to routine respiratory therapy. These levelscorrespond to a hypoxaemia index (PaO2/FiO2) of^125 mmHg (16.7 kPa). The conventional defini-tion of hypoxaemia as PaO2 <8 kPa when breathingair (FiO2 of 0.21) represents a hypoxaemia index<285 mmHg (38 kPa).

It is important to note that the currently usedAmerican European Consensus Guidelines for defin-ing hypoxaemia in acute lung injury and ARDS38

take no account of the positive end-expiratory pres-sure (PEEP) level or other set ventilatory parameterswhen the hypoxaemia index is derived. It seems bothmisleading and likely to cause heterogeneity in studypopulations if the level of PEEP is not standardizedbefore the calculation of this index. In this discussionit is assumed that optimum ventilator managementincludes a PEEP level of at least 10 cmH2O and thatother general aspects of management, includingdrainage of large pleural air or fluid collections, havebeen addressed.

The following issues related to the management ofrefractory hypoxaemia are discussed:

• 'appropriate' target levels for arterial oxygenationand inspired O2 concentration,

• 'appropriate' fluid balance targets and circulatorymanagement - the attempt to identify the 'correct'balance between an adequate intravascular vol-ume and increased lung oedema,

• additional therapies that should be consideredwhen the optimum ventilatory, fluid balance andcirculatory strategies have failed to improve oxy-genation.

ments, it is important to consider this componentof the O2 cascade.

The delivery of O2 from the external environ-ment via the lungs to the mitochondria withinindividual cells is summarized in Figure 22.1, withvalues quoted for a 75-kg individual undertakingnormal activity. The delivery of O2 from the envi-ronment to the systemic circulation is the part ofthe O2 cascade that is best understood, but thereremains much to be learnt about the factors thatcontrol regional delivery, diffusion and the cellularuptake and utilization of O2.

39

Global O2 delivery (DO2) is the product of cardiacoutput and the arterial O2 content (CaO2), which isitself determined by the O2 saturation and haemo-globin concentration of the arterial blood. Overthe past two decades, there has been considerabledebate over what constitutes an 'adequate' globalDO2. Maintaining an 'appropriate' DO2 by ensuring'adequate' intravascular volume replacement fol-lowed by the 'judicious' use of vasoactive agents isundoubtedly important in preventing organ failure,particularly in the peri-operative period40 and in theearly stages of critical illness.41 These issues havebeen reviewed elsewhere, 39 but the level of FiO2 andcorresponding value of PaO2 that are accepted areparticularly relevant to the management of thepatient with refractory hypoxaemia, because thisrepresents a balance between the risks of pulmonaryO2 toxicity and permissive hypoxaemia.

Of the four phases in the transport of O2 from theexternal environment to the tissue cells, two areconvective and two diffusive:

What are the appropriate target levelsfor arterial Pa02 and Fi02: pulmonaryoxygen toxicity or permissivehypoxaemia?

Appropriately for a book on respiratory critical care,considerable space has been devoted to the causesand management of patients with primary lungpathologies that result in severe hypoxaemia.However, the issue of oxygen delivery has not beenaddressed.

Because the prime function of the lungs is tooxygenate the blood and, together with the circula-tion, to ensure that the oxygen supply to the tissuesmeets their individual and fluctuating require-

(i) The convective or 'bulk flow' phases involvethe movement of O2 from the environment toalveoli by ventilation and its transport from thepulmonary to the systemic tissue capillaries:these are the active, energy-requiring stagesthat rely on work performed by the respiratoryand cardiac 'pumps'.

(ii) The diffusive phases involve the passage of O2

from alveolus to pulmonary capillary and fromsystemic capillary to tissue cell: these stagesare passive and depend on the gradient of O2

partial pressures, the tissue capillary density(which determines diffusion distance), theposition of the O2 dissociation curve (P50) andthe ability of the cell to take up and use O2

(Fig. 22.2).


Figure 22.1 Oxygen transport from atmosphere to mitochondria. Values in parentheses for a normal 75-kg individual (body surface area:

1.7 m2) breathing air (Fi02: 0.21) at standard atmospheric pressure (PB: 101 kPa). Partial pressures of 02, C02 in kPa; saturation in %;

contents (Ca02, Cv02) in mLL- 7; Hb in g L- 7; blood/gas flows (Qt, Vile) in L min ~ 1; oxygen delivery return (D02,02R), V02 and VC02 in ml

min - 1. P50 defines position of 02-haemoglobin dissociation curve; it is the P02 at which 50% of haemoglobin is saturated (normally 3.5 kPa).

S02, 02 saturation (%); P02, 02 partial pressure (kPa); PI02, inspired P02; PE02, mixed expired P02; PEC02, mixed expired PC02; PA02, alveolar

P02; Pa02, arterial P02; Sa02, arterial S02; Sv02, mixed venous S02; Qt , cardiac output; Hb, haemoglobin; Ca02, arterial 02 content; Cv02,

mixed venous 02 content; V02, 02 consumption; VC02, C02 production; 02R, 02 return; D02, 02 , 02 delivery; Vi/e, insp/exp minute volume; LV,

left ventricle; LA, left atrium; RV, right ventricle; LV, left ventricle.

Table 22.2 provides a practical illustration of theimpact on global O2 delivery (DO2) of correctinghypoxaemia, anaemia and a low cardiac output. Thisemphasizes that:

• global DO2 may be reduced by anaemia, O2 desat-uration or a low cardiac output, either singly or incombination;

• global DO2 depends on O2 saturation rather thanpartial pressure and, due to the sigmoid shape ofthe oxyhaemoglobin dissociation curve, littleextra benefit derives from increasing -PaO2 abovethe value (~9 kPa) that ensures that over 90% ofhaemoglobin is saturated with O2. However, thisdoes not apply to the diffusive component of O2

transport that does depend on the gradient of O2

partial pressure (Fig. 22.3).

Blood transfusion to polycythaemic levels mayseem an appropriate way to increase DO2 and miti-gate the impact of a low PaO2 because this is theresponse seen in other conditions associated withchronic hypoxaemia such as cyanotic heart disease

and COPD. However, blood viscosity increasesmarkedly above 100 g L-1 , which adversely affectsregional and microcirculatory blood flow, particu-larly if perfusion pressure is reduced, resulting in anexacerbation of tissue hypoxia.42 Recent evidencesuggests that even the traditionally acceptedhaemoglobin concentration for critically ill patientsof approximately 100 g L-1 may be too high,because an improved outcome was observed ifhaemoglobin was maintained between 70 and 90 g L-1,with the exception of patients with coronary arterydisease in whom a level of 100 g L-1 remains appro-priate.43 With the chosen haemoglobin concentra-tion achieved by transfusion and with an O2

saturation (SaO2) maintained at >90%, furtherincreases in PaO2 will have little impact on globalDO2, which will then be determined by the cardiacoutput.

However, increased levels of global DO2 cannotcompensate for disordered regional distribution ofDO2, for impaired diffusion between capillary andcell or for primary metabolic failure within the cellas occurs in sepsis and in cyanide poisoning.


Factors influencing oxygen transportfrom capillary blood to individual cells

The delivery of O2 from capillary blood to the celldepends upon:

• factors that influence diffusion: capillary PO2,tissue oedema, capillary density,

• the rate of O2 delivery to the capillary (DO2),• the position of the O2 - haemoglobin dissocia-

tion relationship (P50),• the rate of cellular O2 utilization and uptake (VO2).

Figure 22.2 Diagram to show the

importance of local capillary

oxygen tension and diffusion

distance from capillary to cell in

determining the rate of oxygen

delivery and the intracellular and

mitochondrial pa02. On the left

there is a low capillary P02 with a

reduced pressure gradient for

oxygen diffusion and an increased

diffusion distance resulting in a low

intracellular and mitochondrial

Pa02. By contrast, on the right the

higher Pa02 partial pressure

gradient and the shorter diffusion

distance results in significantly

higher intracellular Pa02 values.

The position of the O2-haemoglobin dissociationcurve - defined as the faO2 at which 50% of thehaemoglobin is saturated (P50) and which is normal-ly -3.5 kPa - is influenced by various physicochemi-cal factors. An increase in P50 or rightward shift ofthis relationship reduces the haemoglobin saturation(SaO2) for any given PaO2, thereby increasing tissueO2 availability. This is caused by pyrexia, acidosis,hypercapnia and an increase in intracellular phos-phate, notably 2, 3-diphosphoglycerate (2, 3-DPG),and explains in part the benefit derived from hyper-

Table 22.2 Relative effects of changes in PaO2, haemoglobin and cardiac output on oxygen delivery

Normal3

Patientb

Fi02

Fi02

Hb

Qt

0.210.210.350.600.600.60

13.06.09.0

16.516.516.5

967592989898

130 3.0707070

105105

1.42.13.83.83.8

170728896

142142

5.34.04.04.04.06.0

900288352384568852

0-68+22+9

+48+ 50

aNormal 75-kg subject at rest.bPatient with hypoxaemia, anaemia and reduced cardiac output and evidence of global tissue hypoxia.D02 = Ca02 x Qt ml min-1.Ca02 = (Hb X Sa02 X 1.34) + (Pa02 X 0.23) mL L-1.Where Fi02 is the fractional inspired 02 concentration, Pa02, Sa02, Ca02 are, respectively, the partial pressure, saturation and content of 02 in arterial blood,and Q, is the cardiac output.1.34 ml is the volume of 02 carried by 1 g of 100%-saturated haemoglobin and Pa02 (kPa) X 0.23 is the amount of 02 in physical solution in 1 L blood,which is less than 3% of total Ca02 for normal Pa02 (i.e. < 14 kPa).


Figure 22.3 Diagram illustrating the influence of intercapillary distance on the effects of hypoxia, anaemia and low flow on the relationship

between oxygen delivery and consumption. With a normal intercapillary distance, as shown in (a), the D02/V02 relationship is the same,

irrespective of whether the reduction in D02 is produced by progressive hypoxia, anaemia or reduction in cardiac output. However, in (b) there

is an increased intercapillary distance as occurs with tissue oedema, and in this situation progressive reduction in arterial Pa02 does alter the

D02/V02 relationship, with V02 falling at much higher levels of global D02 This effect is not seen when D02 is reduced by progressive

reduction in haemoglobin concentration or cardiac output.

capnia and why it is important to correct hypophos-phataemia.

Mathematical models of tissue hypoxia demon-strate that the fall in intracellular PO2 resultingfrom an increase in intercapillary distance is moresevere if the reduction in tissue DO2 is due to'hypoxic' hypoxia (a fall in PaO2) rather than dueto 'stagnant' (a fall in flow) or 'anaemic' hypoxia(Fig. 22.3). Thus, severe arterial hypoxaemia, par-ticularly in the presence of increased tissue oedema,will result in reduced O2 diffusion and cellularhypoxia. These observations suggest that the extentof tissue oedema and the level of PaO2, indepen-dent of DO2, will affect tissue oxygenation andpotentially the development of organ dysfunctionin critically ill patients.

Therefore, the following strategies directed atimproving global O2 delivery and preventing pul-

monary O2 toxicity are not without potential risk ofexacerbating tissue hypoxia:

• reducing FiO2 to prevent O2 toxicity,• giving fluids to increase intravascular volume,

particularly in patients with increased endothelialpermeability,

• giving vasoactive agents that alter the regional dis-tribution of DO0.

Improvement in the prognosis for the patientwith severe hypoxaemia and incipient or estab-lished organ failure awaits the development oftechnologies and therapies that allow the meas-urement and manipulation of both the regionaldistribution of blood flow and other 'down-stream' factors in the O2 cascade.44


How tolerant are the tissues ofhypoxia?

There is considerable variation in the tolerance ofindividual organs and cells to hypoxia.45 Corticalneurons are exquisitely sensitive to sudden reduc-tions in O2 supply and do not survive sustainedperiods of hypoxia. Following complete cessation ofcerebral perfusion, nuclear magnetic resonance(NMR) studies have shown a 50% decrease in neur-onal adenosine triphosphate (ATP) within 30 s andirreversible damage occurring within 3 min. Othertissues can survive anoxia for longer periods: kidneysand liver for approximately 20 min, skelet al musclefor about 75 min and vascular smooth muscle for upto 72 hours. Hair and nails provide the most extremeexample of anoxic tolerance because they continueto grow for several days after death.

There is a difference in the tolerance to severehypoxia and the response to complete anoxia andthis also differs in health and disease. In sepsis, inhi-bition of enzyme systems and O2 utilization reduceshypoxic tolerance.46 Methods aimed at enhancingmetabolic performance including the use of alterna-tive substrates, techniques to inhibit endotoxin-induced cellular damage and drugs to reduce oxidantintracellular damage are all currently under investi-gation. Progressive or repeated exposure to hypoxiaenhances tissue tolerance to O2 deprivation in a simi-lar way to altitude acclimatization. An acclimatizedmountaineer at the peak of Mount Everest can toler-ate a PaO2 of 4-4.5 kPa for several hours, andpatients with severe COPD can survive with a PaO2

<5 kPa for several years, both of which are degrees ofhypoxaemia that, if produced acutely, would result inconfusion and a reduced level of consciousness with-in a few minutes in a normal subject.

So, what is the critical level of tissue oxygenationbelow which cellular damage will occur? The answerdepends predominantly on the co-morbid factorsand the duration of hypoxia. For example, young,healthy individuals with the acute respiratory dis-tress syndrome (ARDS) can make a complete recov-ery following prolonged periods of severehypoxaemia with PaO2 as low as 6 kPa and O2 satu-rations below 85%. The older patient with wide-spread atheroma may not survive prolongedhypoxaemia at such levels.

The foregoing analysis might suggest that it wouldbe beneficial to increase the level of FiO2 to improve

tissue oxygenation and so potentially reduce organfailure and lead to an improved outcome. Currentpractice is to reduce FiO2 as far as 'reasonably' pos-sible, with the aim of limiting the risk of pulmonaryO2 toxicity. This assumes that the risk is significant,but should we be unduly concerned about pul-monary O2 toxicity?

Pulmonary oxygen toxicity

The demonstration that ventilation with low tidalvolumes reduces ventilator-induced lung injury andimproves outcome makes it important to considerwhether other lung-protective strategies such asavoiding high levels of inspired O2 may produceadditional benefit. Hyperoxia increases the levels ofreactive radicals, hydrogen peroxide and hydroxyl andsuperoxide ions, which inactivate sulphydryl enzymesthereby disrupting DNA synthesis. This damages thepulmonary capillary membranes and increases inter-stitial oedema, which reduces lung compliance andfurther impairs gas exchange. Decreased surfactantlevels have also been attributed to high inspired O2

levels.47 (See also Chapter 18, p. 239.)The risk of O2 toxicity is related to the absolute

O2 tension rather than to FiO2, as demonstrated byexperiments varying the atmospheric pressure atwhich the gas mixture is delivered. Astronauts suf-fer no ill-effects from 100% O2 at one-third atmos-phere pressure. Compared to other species, such asrats, humans appear relatively resistant to O2 toxic-ity and there is little evidence of significant damagefrom long-term ventilation with inspired PO2

(PiO2) of 50 kPa (FiO2 of 0.5) at standard atmos-pheric pressure. However, there is evidence thatsuch levels for long periods may be harmful whenthere is pre-existing lung damage.48 One study hasshown that an inspired O2 above 0.6 for over 24hours does produce reduced diffusion capacitywhen measured many months after an episode ofARDS.49 In general, there is an inverse relationshipbetween PO2 and the safe duration of exposure, i.e.the potential for injury relates to the area under thePO2- time curve.

It is, of course, difficult in the clinical setting todistinguish lung injury due to O2 toxicity from thatcaused by the underlying process and other aspectsof management. Nonetheless, on current evidence,prolonged ventilation with HO2 60 kPa (FiO2 > 0.6


at 1 atmosphere) may contribute to lung injury andadversely affect both the short-term and long-termoutcomes.

Our current practice in selecting the FiO2 (assum-ing ventilation at 1 atmosphere or 100 kPa pressure)and corresponding PaO2 levels that we would con-sider acceptable can be summarized as follows.

• If PaO2 > 10 kPa with FiO2 < 0.6, use lowest FiO2

up to maximum 0.6 to maintain PaO2 at 10-12kPa and continue conventional treatment.

• If PaO2>10 kPa with FiO2>0.6, target lowestFiO2 between 0.6 and 0.8 to maintain PaO2 at8-10 kPa and start further measures to improvegas exchange, e.g. recruitment manoeuvre.

• If PaO2<8 kPa with FiO2>0.8, set FiO2

0.8-1.0 to maintain PaO2 at 7-8 kPa andurgently start further measures; consider short-term nitric oxide 'rescue' until gas exchange hasimproved.

It remains uncertain but distinctly possible thatthe prolonged use of nitric oxide, a highly reactivechemical species, may itself cause lung injury andcompound the problems of O2 toxicity.50 Becausetwo multicentre trials of the use of nitric oxide inARDS have demonstrated no survival benefit andalso indicate that even the gas exchange benefitmay only last for 24 hours, we reduce and stop thenitric oxide as soon as tolerable gas exchange isachieved.

The urgency of correcting severe hypoxaemiarequiring FiO2 of >0.8 is greatly increased if thepatient is known to have peripheral vascular diseasewith coronary and cerebral atheroma, and one wouldaim to maintain the PaO2 > 8 kPa, even if thisrequired an FiO2 of 1.0, while the further strategiesto improve oxygenation were urgently implemented.

Both FiO2 >0.6 (PiO2 >60 kPa) and PaO2 <8kPa for prolonged (>6 hours) periods mayadversely affect outcome in patients with severelung injury. The need for increasing levels of FiO2

above 0,6 to maintain PaO2 > 9 kPa shouldprompt consideration of further measures toimprove gas exchange and, if necessary, the use ofshort-term 'rescue' nitric oxide as well as investi-gations to determine the cause of the decline.

Metabolic considerations in patientswith refractory hypoxaemia

Reduction in metabolic rate will reduce tissue O2

consumption and, if O2 supply remains constant,will improve cellular oxygenation. This alternativestrategy to increasing cellular O2 delivery may beachieved by:

• ensuring adequate analgesia and sedation to con-trol sympathetic activation from pain, agitation,shivering and various interventions (nursingprocedures, physiotherapy, visitors),

• instituting active cooling measures if core tempera-ture exceeds 38 °C: for each °C rise in tempera-ture, O2 consumption increases by 10-15%,

• avoiding drugs that increase metabolic rate,particularly inotropes such as adrenaline anddobutamine, and other beta-agonists,

• abolishing spontaneous respiratory effort, ifnecessary using muscle relaxants and therebyeliminating the metabolic costs of breathing.

Fluid management in refractoryhypoxaemia

Deciding on fluid therapy for a patient with severehypoxaemia represents a difficult balance betweenthe risk of increasing extravascular lung water, andfurther exacerbating the hypoxaemia and potentiallythe lung injury, and the risk of providing inadequateintravascular volume replacement, resulting inimpaired peripheral perfusion and the developmentof other organ failure. In Chapter 12, 'appropriate'fluid therapy in acute lung injury was summarized asthe provision of the minimum intravenous fluid thatensured 'adequate' cardiac output and tissue perfu-sion. While agreeing with this and that the failure tolimit pulmonary vascular pressure and the extent ofpulmonary oedema may contribute to ventilator-induced lung injury, such a broad consensus state-ment leaves the clinician uncertain as to whatprecisely is meant by 'adequate' cardiac output andtissue perfusion. Conventionally, this decision isbased on the atrial filling pressures, which assumesthat both pulmonary capillary permeability and theintrathoracic pressure are normal. These assump-tions are often not valid in the critically ill, particu-larly the patient with severe lung injury, in whom


atrial filling pressures in excess of 15 or even20 cmH2O, do not necessarily preclude intravasculardepletion.

Measurement of volume pre-load of the leftventricle is more relevant, but has not been possibleas part of routine clinical practice, althoughechocardiography can provide valuable informa-tion on the adequacy of filling of the left ventricleand, if performed serially, on the response tovolume loading. Techniques for determiningintrathoracic blood volume and extravascular lungwater are discussed in Chapter 8, but are still beingrefined and, as yet, are not widely available. How-ever, there are useful bedside signs that should alertthe clinician to the possibility of inadequateintrathoracic blood volume:

• Hypotension precipitated by sedation, analgesiaor postural change.

• Fluctuation in the arterial pressure trace duringpositive pressure ventilation and hypotensionwhen increasing levels of PEEP are applied.Indeed, a formal recruitment manoeuvre with theapplication of PEEP to over 30 cmH2O representsan imposed Valsalva manoeuvre. Intrathoracicvolume depletion is indicated by a marked reduc-tion in pulse pressure and mean arterial pressure(as illustrated in Fig. 22.3a), whereas a raisedintrathoracic blood volume produces a 'square-wave' response (Fig. 22.3b).

• Brief disconnection from the ventilator causes theblood pressure to rise and venous pressure to fall:the atrial pressure measurement 'off' the ventila-tor more accurately reflects the ventricular end-diastolic transmural pressure. This manoeuvre is,however, relatively contra-indicated in patientswith ARDS because any procedure that reducestracheal pressure will cause de-recruitment, withalveolar collapse and a further exacerbation of thehypoxaemia.

The difficulty of interpreting the absolute levels ofthe atrial filling pressures (RAP/LAP) may also beresolved by a fluid challenge, although this must beperformed with caution in the patient with radiologi-cal evidence of increased lung water. We would give amaximum of 200 ml of colloid and observe theimpact on blood pressure, flow and atrial filling pres-sures. In the volume-depleted patient, blood pressureand flow will increase with only a small, transientincrease in filling pressures. While pulmonary gas

exchange remains satisfactory, there is less anxietyabout giving further colloid. Sufficient volume willhave been given when either the target pressures areachieved and the evidence of poor peripheral perfu-sion and organ dysfunction has resolved, or whenthere is a sustained rise in filling pressures with eitherdeterioration in gas exchange or chest X-rayappearances.

SETTING THE TARGET FLUID BALANCE

The crystalloid and colloid balance over the previous24 hours and the intravascular and extravascularcompartments should be individually reviewed twicea day. The crystalloid balance should include theplanned enteral intake, fluid for central lines anddrug infusions, urine output and correction for both'insensible' losses (sweat, diarrhoea) and the state ofhydration of the extravascular tissue space. A dailytarget balance from—1.5 L to > +1 L may beappropriate, but, typically in the patient with severelung injury, it will be between —1.0 L and +0.5 L.The more robust the circulation and the greater theevidence of tissue oedema, the more negative thecrystalloid balance should be.

If the assessment of the extravascular spaceshows gross peripheral oedema with dense alveolarfluid infiltrates and the presence of pleuraleffusions, then, while the circulation remainsrobust, negative fluid balances of 2 L or more per24 hours would be appropriate, using either forcednaturesis if renal function is maintained orhaemofiltration if necessary. The potential benefitsof achieving such a negative crystalloid balanceapply not only to lung function and improvingoxygenation, but also to the peripheral tissues,which are frequently oedematous in this setting.Resolution of this tissue oedema will improve O2

diffusion from capillary to cell, as already discussed.During the worst excesses of goal-directed therapy,in grossly oedematous patients referred to our unitwith severe acute lung injury, negative crystalloidbalances of over 10 L were achieved within 72 hourswithout any apparent adverse effects on the circula-tion or critical organ perfusion, but resulted ina marked improvement in lung compliance and gasexchange. It should be remembered that bothpulmonary and chest-wall compliance improvewith the resolution of pulmonary interstitial andalveolar oedema and the reduction of oedema inthe tissues of the chest wall.


If the intravascular space is under-filled, andparticularly if the patient is already on vasopressoragents (which increase the risk of compromisedregional perfusion, particularly to the splanchnicbed), fluid should be given. The rate of crystalloidinfusion should be increased but, acutely, the extravolume required to reach the pre-load target shouldbe given as colloid, as described. After appropriateblood transfusion, synthetic colloid rather than albu-min should be used. A much-debated meta-analysiscomparing the use of albumin with crystalloid orsynthetic colloid concluded that, in the critically ill,albumin was associated with an increased mortal-ity.51 Certainly, attempting to correct a low serumalbumin in such patients with a significant inflam-matory response is futile because their vascularendothelium will be freely permeable to albumin.There is relatively little evidence on which to base thechoice of synthetic colloid (starch or gelatin), but theincrease in intravascular volume is sustained forlonger with the starch solutions, which also providea wider range of molecular weight products and asodium-free option.

Three final observations are relevant with regardto the cardiorespiratory interactions in this groupof patients with refractory hypoxaemia. The first isthat, if excessive fluid is once given and a markedincrease in lung water occurs, it requires a fargreater negative crystalloid balance to remove thatfluid than the volume originally given. Secondly,correction of intravascular volume depletion withsubsequent improvement in cardiac output and O2

delivery will increase mixed venous O2 saturation,which, in the context of a large pulmonary shunt,must increase the arterial O2 tension if the shuntremains unchanged. Finally, further advances influid management in patients with severe lunginjury will depend on the development of bedsidetechnology that provides reliable measurements ofboth extravascular lung water and regional tissueperfusion.44

Further strategies in the treatment ofrefractory hypoxaemia

The range of further strategies that have been evalu-ated in the treatment of severe acute lung injurywith refractory hypoxaemia are discussed inChapter 12. Before embarking on any escalation of

treatment, it is important to be sure that the basicventilatory strategy and general management areoptimum. If such measures improve oxygenationand the hypoxaemia index is above 125 mmHg(16.7 kPa), most clinicians would not feel it neces-sary to introduce additional strategies. However, ifthe hypoxaemia index does fall below this level,which strategies should be considered? Daily chestX-rays will be performed on such patients andobviously any gas or fluid collection should bedrained. It can be difficult to identify the loculatedor anterior pneumothorax on antero-posterior chestX-ray and, as discussed in Chapter 10, a computedtomography scan should be considered beforeembarking on further therapy.

RECRUITMENT MANOEUVRES

The different lung recruitment manoeuvres that canbe used are described in Chapter 12, and Figure 8.1(see Chapter 8) gives an example of such a man-oeuvre using a high-frequency oscillator.

The manoeuvre is usually performed in aparalysed patient who is ventilated using low tidalvolume (Vt 5-7 mL kg-1) , inverse ratiopressure-controlled ventilation with at least 10cmH2O PEEP already applied. By increasing thePEEP level, the patient is 'held' at end-inspiration atan inflation pressure of 30-45 cmH2O for 30 s to 2min, after which ventilation is resumed with PEEPset at 15-18 cmH2O, depending on the circulatoryresponse. The blood pressure will frequently fallduring the manoeuvre due to the increase inintrathoracic pressure and this will determine thepressure level and the length of time for which it isapplied. Extra fluid administration should be avoid-ed and, if necessary, a temporary increase in vaso-pressor therapy should be used to support the bloodpressure.

It is important to remember that de-recruitmentreadily occurs in these patients secondary to anycause of a fall in tracheal pressure, which include:

• disconnection from the ventilator,• suctioning down the endotracheal tube, particu-

larly if high pressures are used and despite using a'closed' suction system,

• bronchoscopy,• spontaneous inspiratory effort opposing the effect

of PEEP and reducing end-expiratory volume,


• chest-wall compression during physiotherapy -this partly explains the hypoxaemia that is oftenseen in these patients after physiotherapy, whichshould only be given by an experienced practitionerafter careful discussion with the clinical team.

HIGH-FREQUENCY OSCILLATION

This is a mode of ventilation that is conceptuallysimple and that achieves an 'open lung' and, byeliminating the biphasic swing of tidal breathing,avoids the potential problem with conventionalventilation of cyclical over-inflation and collapsein different parts of the lung. Figure 22.2 illus-trates the principles involved. A continuousdistending pressure of up to 55 cmH2O is appliedby adjusting the gas flow and expiratory valve;this effectively applies super-PEEP and, as it isincreased, alveoli are recruited and the functionalresidual capacity increases. The frequency(180-900 min-1) and amplitude or power of theoscillator determine the resulting tidal volume.This is only in the range 0.1-5 ml kg-1, barelyexceeding the physiological dead space, andproduces alveolar pressure swings of less than5 cmH2O. The enhanced gas exchange (bothoxygenation and CO2 clearance) that results relieson a combination of different gas transportmechanisms, including Taylor dispersion, molec-ular diffusion and cardiogenic mixing in additionto convective ventilation.52 It differs fromhigh-frequency jet ventilation in the mechanismof delivery (oscillator versus jet), and in thatexpiration is active rather than passive and thealveolar distending pressure is directly set. Unlikejet ventilation, a standard humidifier can be usedwith high-frequency oscillatory ventilation(HFOV) and nitric oxide can be deliveredthrough the system.

We resort to using HFOV when gas exchange hasnot responded to other therapies, and haveobserved striking improvements in gas exchange inmost cases. It is generally well tolerated, but prob-lems can arise with hypotension, which mayrequire an increase in vasopressor therapy: anyextra fluid should be given with great caution.There is obviously a risk of pneumothorax, but ourexperience is that it is no more frequent than withconventional ventilation.

Outcome benefit has yet to be demonstrated, but apilot study involving ARDS patients showedimproved oxygenation, 53 and a recent prospectivestudy involving patients with severe ARDS (meanhypoxaemia index <100 mmHg) concluded that itwas a safe and effective rescue therapy in severe oxy-genation failure.54

NITRIC OXIDE

Most clinicians now try to avoid the long-term use ofnitric oxide because two large multicentre, random-ized studies failed to demonstrate any outcome bene-fit and suggested that the improvement inoxygenation only lasted for 24 hours. Our practice isto use it as a short-term agent for the transport ofpatients with otherwise refractory hypoxaemia andalso as a rescue therapy while alternative therapies(prone, oscillation) are instituted. There is evidencethat it may produce additive benefit when used withalmitrine and prone positioning.55 Concerns remainthat longer term use may exacerbate the lung injuryand, if used in further randomized studies, it will beimportant to collect long-term follow-up data onlung function.

POSTURAL THERAPY

Prone positioning produces an impressive improve-ment in gas exchange in about two-thirds of patientswith refractory hypoxaemia and it has become anincreasingly popular therapy despite the logisticproblems presented in the heavily instrumentedpatient. A multicentre Italian study confimed suchan improvement in gas exchange without anyincrease in complications in the treatment group, butno outcome benefit could be demonstrated.56

However, the numbers enrolled (304) may have beeninadequate, the treatment group was only 'proned'for 7 hours per day, for a total of 10 days, and therewere varying degrees of non-compliance with thestudy protocol due to staffing shortage.

We believe that it remains a promising therapyand that the imminent production of a 'bed' that willautomatically rotate the patient into the prone posi-tion with all lines attached promises to eliminate thelabour-intensive issues for the staff. It will also allowgreater standardization of the precise positioning ofthe patient when prone, as well as the opportunity to

References 291

intervene earlier in the disease process and to inves-tigate the impact of different 'proning' intervals andeven the effect of continuous rotation in the proneposition.

There are a number of other therapies, includingsurfactant therapy, liquid ventilation and extracor-poreal techniques, that are still under investigationbut are not, at this stage, either proven or practicalalternatives to the therapies discussed.

Future advances will depend on carefullydesigned, randomized, controlled, multicentrestudies that characterize patients precisely anddemonstrably execute the agreed protocol, as theARDSnet group showed in their study of tidal vol-umes in acute lung injury.

REFERENCES

1. Goldhill, DR, Worthington, L, Mulcahy, M, Sumner, A.

The patient-at-riskteam: identifying and managing

seriously ill ward patients. Anaesthesia 1999; 54:

853-60.

2. Hourihan, F, Bishop, K, Hillman, KM, Daffurn, K, Lee, A.

The medical emergency team: a strategy to identify

and intervene in high risk patients. Clin Int Care 1995;

6: 269-72.

3. Schein, MH, Hazday, N, Pena, M, Ruben, BH, Sprung, CL

Clinical antecedents to in-hospital cardiac arrest. Chest

1990; 98:1388-92.

4. Goldhill, DR, White, A, Sumner, A. Physiological values

and procedures in the 24 hours before ICU admission

from the ward. Anaesthesia 1999; 54: 529-34.

5. Comprehensive critical care. London: Department of

Health, 2000.

6. Tobin, MJ. Essentials of critical care medicine. New

York: Churchill Livingstone, 1989.

7. Aubier, M, Muriciano, D, Fournier, M, et al. Central

respiratory drive in acute respiratory failure of

patients with chronic obstructive pulmonary disease.


8. Dunn, WF, Nelson, SB, Hubmayr, RD. Oxygen induced

hypercarbia in obstructive pulmonary disease. Am Rev

Respir Dis 1991; 144: 526-30.

9. Mador, MJ, Tobin, MJ. Acute respiratory failure. In

Chronic obstructive pulmonary disease. Calverley, P,

Pride N, eds. London: Chapman & Hall, 1995; 461-94.

10. Brochard, L, Moncebo, J, Wysocki, M, et al. Non inva-

sive ventilation for acute exacerbations of chronic

obstructive pulmonary disease. N EnglJ Med 1995;

333: 817-22.

11. Plant, P, Owen, J, Elliott, M. A multi centre

randomised control trial of the early use of non

invasive ventilation for acute exacerbations of

chronic obstructive pulmonary disease. Lancet 2000;

355: 1931-5.

12. Soo Hoo, GW, Hakiman, I, Santiago, SM. Hypercapnic

respiratory failure in COPD patients. Response to

therapy. Chest 2000; 117: 169-77.

13. Plant PK, Owen JL, Elliott, MW. Non-invasive

ventilation in acute exacerbations of chronic

obstructive pulmonary disease: long term survival

and predictors of in-hospital outcome. Thorax 2001;

56: 708-12.

14. Carlucci, A, Richard, JC, Wysocki, M, et al. Noninvasive

versus conventional mechanical ventilation. An

epidemiological survey. Am J Respir Crit Care Med

2001; 163:874-80.

15. British Thoracic guidelines on the use of acute

non-invasive ventilation. Thorax 2001.

16. Evans, TW. International Consensus Conference in

Intensive Care Medicine: non-invasive positive

pressure ventilation in acute respiratory failure.

Int Care Med 2001; 27: 166-78.

17. Withholding and withdrawing life prolonging

treatment. Guidance for decision-making. London:

BMJ Publishing, 2001.

18. Dales, RE, O'Connor, A, Hebert, P, et al. Intubation and

mechanical ventilation for COPD: development of an

instrument to elicit patient preferences. Chest 1999;

116: 792-800.

19. Vasilyev, S, Schaap, RN, Mortensen, JD. Hospital

survival rates of patients with acute respiratory

failure in modern respiratory intensive care units.

C7?«M995; 107:1083-8.

20. Connors, AF, Dawson, NV, Thomas, C. Outcomes

following acute exacerbations of severe chronic

obstructive pulmonary disease. Am J Respir Crit

Care Med 1996; 154: 959-67.

21. Wagner, MG, Wagner, DP, et al. Hospital and 1 year

survival of patients admitted to intensive care units

with acute exacerbations of chronic obstructive

disease. JAMA 1995; 274: 1852-7.

22. Stauffer, JL, Fayter, N, Graves, B, et al. Survival

following mechanical ventilation for acute respiratory

failure in adult men. Chest 1993; 104: 1222-9.

23. Gracey, DR, Naessens, JM, Krishan, I, et al. Hospital

and post-hospital survival among patients with COPD

who require mechanical ventilation for acute

respiratory failure. Chest 1989; 101: 211-14.


24. Bardi, G, Pierotello, R, Desideri, M, et al. Nasal

ventilation in COPD exacerbations: early and late

results of a prospective controlled study. Ear Respir

72000; 15: 98-104.

25. Confalonieri, M, Perigi, P, Scartabellati, A, et al. Non

invasive mechanical ventilation improves the immediate

and long term outcome of COPD patients with acute

respiratory failure. Eur RespirJ 1996; 9: 422-30.

26. Meecham-Jones, DJ, Paul, ESA, Jones, PW, et al. Nasal

pressure support ventilation plus oxygen compared

with oxygen therapy alone in hypercapnic COPD. Am J


27. Jones, SE, Packham, S, Hebden, M, Smith, AP.

Domicilary nocturnal intermittent positive pressure

ventilation in patients with respiratory failure due to

severe COPD: long term follow up and effect on

survival. Tftorax 1998; 53: 495-8.

28. Ely, EW, Baker, A, Dunagan, D, et al. Effect of the

duration of mechanical ventilation of identifying

patients capable of breathing spontaneously. N Engl

7 Med 1996; 335: 1864-9.

29. Vitacca, M, Vianello, A, Colombo, D, et al. Comparison

of two methods for weaning patients with chronic

obstructive pulmonary disease requiring mechanical

ventilation for more than 15 days. Am J Respir Crit

Care Med 2001; 164: 225-30.

30. Nava, S, Ambrosino, N, Clini, E, et al. Non invasive



pulmonary disease. A randomised controlled trial.

Ann Intern Med 1998; 128: 721-8.

31. Girault, C, Daudenthoni, S, Chevron, V, et al.

Non invasive ventilation as a systematic extubation

and weaning technique in acute on chronic

respiratory failure. Am J Respir Crit Care Med 1999;

160: 86-92.

32. Vitacca, M, Ambrosino, N, Clini, E, et al. Physiological

response to pressure support ventilation delivered

before and after extubation in patients not capable of

totally spontaneous autonomous breathing. Am J


33. Carlucci, A, Gregoretti, C, Squadrone, V, et al.

Preventative use of non-invasive mechanical

ventilation (NIMV) to avoid post-extubation respiratory

failure: a randomised controlled trial. Eur Respir

72001: 18 (Suppl. 33): 29s.

34. Jaber, S, Carlucci, A, Bourssarsar, M, et al. Helium -

oxygen in the post extubation period decreases

inspiratory effort. Am J Respir Crit Care Med 2001;

164: 633-7.

35. Hamid, S, Noonan, YM, Williams, AJ, Davidson, AC. An

audit of weaning from mechanical ventilation in a UK

weaning centre. Thorax 1999; 54: 86.

36. Pilcher, D, Hamid, S, Williams, AJ, Davidson, AC.

Outcomes, cost and long term survival of patients

referred to regional centres for weaning from

mechanical ventilation. Presentation at World

Congress ICM, Sydney, 2001.

37. Scheinhorn, DJ, Chao, DC, Stearn-Hassenpflug, M,

et al. Outcomes in post ICU mechanical ventilation.

Treatment of 1123 patients at a regional weaning

centre. Chest 1997; 111: 1654-9.

38. Bernard, GR, Artigas, A, Brigham, KL, et al. and the

Consensus Committee of the American European

Conference on ARDS. Definitions, mechanisms,

relevant outcomes and clinical trial coordination.


39. Leach, RM, Treacher, DF. The pulmonary physician in

critical care 2. Oxygen delivery and consumption in

the critically ill. Thorax 2002; 57: 1-7.

40. Wilson, J, Woods, I, et al. Reducing the risk of major

elective surgery: randomized, controlled trial of

preoperative optimisation of oxygen delivery. BMJ

1999; 318: 1099-103.

41. Rivers, E, Nguyen, B, Havstad, S, et al. Early

goal-directed therapy in the treatment of severe sepsis

and septic shock. N Engl J Med 2001; 345:1368-77.

42. Harrison, MJG, Kendall, BE, Pollock, S, Marshall, F.

Effect of haematocrit on carotid stenosis and cerebral

infarction. Lancet 1981; ii: 114-15.

43. Hebert, PC, Wells, G, Blajchman, M, et al. A multicenter,

randomized, controlled clinical trial of transfusion

requirements in critical care. N Engl J Med 1999; 340:

409-17.

44. Consensus Conference. Tissue hypoxia: how to detect,

how to correct, how to prevent. Am j Respir Crit Care

Med 1996; 154: 1573-8.

45. Robin, ED. Of men and mitochondria: coping with

hypoxic dysoxia. Am Rev Respir Dis 1980; 122: 517-31.

46. Bradley, SG. Cellular and molecular mechanisms of

action of bacterial endotoxins. Ann Rev Microbiol

1979; 33: 67-94.

47. O'Connor, BS, Vender, JS. Oxygen therapy. Crit Care

C///71995; 11: 67-78.

48. Denke, SM, Fanburg, BL Normobaric oxygen toxicity

of the lung. N Engl J Med 1980; 303: 726.

49. Elliott, CG, Rasmusson, BY, Crapo, RO, et al. Prediction

of pulmonary function abnormalities after adult

respiratory distress syndrome (ARDS) Am Rev Respir

D/S1987; 135: 634.

References 293

50. Pryor, WA. The chemistry of peroxynitrite: product

from reaction of nitric oxide with superoxide. Am J

Physiol 1995; 268: 699-722.

51. Cochrane Injuries Group Albumin Reviewers. Human

albumin administration in critically ill patients:

systematic review of randomised controlled trials.

BMJ 1998; 317: 235-40.

52. Chang, HK Mechanisms of gas transport during

ventilation by HFOV: a brief review. J Appl Physiol

1984.

53. Fort, P, Pharma, C, Westerman, J, et al. High frequency

oscillatory ventilation for adult respiratory distress

syndrome, a pilot study. Crit Care Med 1997; 25: 937.

54. Mehta, S, Lapinsky, SE, Hallett, DC, et al. Prospective

trial of high-frequency oscillation in adults with acute

respiratory distress syndrome. Crit Care Med 2001; 29:

1360-9.

55. Borelli, M, Lampati, L, Vascotti, E, Fumagalli, R,

Pesenti, A. Hemodynamic and gas exchange response

to inhaled nitric oxide and prone positioning in acute

respiratory distress syndrome patients. Crit Care Med

2000; 28: 2707-12.

56. Gattinoni, L, Tognoni, G, Pesenti, A, et al. and the

Prone-Supine Study Group. Effect of prone positioning

on the survival of patients with acute respiratory

failure. N EnglJ Med 2001; 345(8): 568-73.

Index

Abbreviations: ALI, acute lung injury; ARDS, acute respiratory distress syndrome; BMT, bone marrow transplantation; COPD, chronic obstructivepulmonary disease; DVT, deep venous thrombosis.

Abdominal muscles 2-3Abscess, lung, CT 227Accessory muscle use, breathing pattern and 18Acid-base balance see pHAcquired immunodeficiency syndrome see HIV diseaseAction potentials, phrenic nerve 7Activated partial thromboplastin time, heparin dose and 254Activated protein C resistance 250Active cycle of breathing 71Active movements in bed 76Activities, ICU 266-7Acute lung injury see InjuryAcute respiratory distress syndrome (ARDS) 138-9

at-risk patients, investigating 151definitions 138, 139epidemiology 139Hamman-Rich syndrome and its relationship to 239histological changes 142management 153-69

clinical trials 140mechanical ventilation see subheading below

mechanical ventilation in 34, 156-7case study 35-6volume-controlled 25, 34, 157, 161

pathogenesis 143-8pathophysiology 32,138-69, 155-6precipitating conditions 139-40primary vs secondary 32sepsis and multiple organ dysfunction syndrome in 140—41systemic and pulmonary markers 148-51traumatic brain injury and, case study 43-4X-ray 81see also Injury (non-traumatic), acute lung

Acute respiratory failure see Respiratory failureAdhesion molecules and ARDS/ALI 142, 145, 151Admission

in community-acquired pneumonia 182criterialpolicies 265, 265-6, 272-3, 275-6

ethical aspects 272-3, 275-6to HDU see High dependency unit

Adrenaline, anaphylaxis 117-18Adult respiratory distress syndrome see Acute respiratory distress syndromeAdvance directives 275-6Advanced respiratory support 264Afferent, pulmonary 17Age, community-acquired pneumonia and 182AIDS see HIV diseaseAir embolism, chest trauma 125-6Air leaks 130-34, 229-33

aetiology 130,229-31

Airflow see FlowAirway(s)

chronic obstructive disease see Chronic obstructivepulmonary disease

clearance techniques 71humidification 30, 71large, obstruction 116-19resistance see Resistance

Airway occlusion (in rapid airway occlusion technique) 9,11Airway occlusion pressure measurement during

weaning from ventilation 56-7Airway pressure

in mechanical ventilation 23-4limiting 33measurement 96-8

positive see Positive pressure ventilationAirway pressure contour, ventilated patient 14Airway pressure release ventilation 34-6Alimentary tract see Gastrointestinal tractAllergic granulomatous angiitis (Churg-Strauss syndrome) 246Allergy, interstitial lung disease 237Allogeneic bone marrow transplantation, infections 204, 204-5Almitrine 165Alveolar damage, diffuse (DAD) 236Alveolar haemorrhage, diffuse 246-7Alveolar hyperventilation in type I respiratory failure 278, 279Alveolar hypoventilation in type II respiratory failure 278, 279Alveolar ventilation (VA) 94Alveolitis, cryptogenic fibrosing 236Ambulatory ventilation 77Amiodarone, pulmonary toxicity 245Amphotericin B 214

neutropenic patients 212Anaesthesia for mechanical ventilation 29Anaphylaxis 117-18Angioedema 117-18Angiography, haemoptysis 109, 110Antacids 196Antibiotics (antibacterial)

in community-acquired pneumonia 186-7with empyema 228resistance 183-4, 188

in gastric aspiration 114in immunocompromised hosts 214

neutropenic patients 212in nosocomial pneumonia 196-8

with empyema 228in selective decontamination of digestive tract 196in water aspiration 115

Anticoagulants, DVT 254-7

Index 295

Antifungal drugs 214neutropenic patients 212

Anti-inflammatory drugs 166-7community-acquired pneumonia 188

Antineutrophil cytoplasmic antibodies (ANCA) 246Antinuclear antibodies (ANA) 246Anti-oxidant molecules/activity 150Antiparasitic drugs 214Antiviral drugs 214Aortic body chemoreceptors 16Aortic injury 128-9Aortic sinus baroreceptors 17Apneustic centre 15Apnoea, sleep 5ARDS see Acute respiratory distress syndromeArterial blood gases 89-90

chemoreceptors andcentral 16peripheral 16

measurementprinciples 89temperature correction 89-90

in ventilatory drive assessment 17see also Carbon dioxide; Oximetry; Oxygen

Aspergillosis 205diagnosis 213immunocompromised hosts 111

solid organ recipients 208treatment 214

Aspiration (technique)chest see Chest drainageof secretions in ventilated patients 196

Aspiration (unintentional inhalation) 110-16of fluid 112-16

gastric contents see Gastric contentsof foreign body/solid particles 112

bronchoscopy 83interstitial lung disease due to 237of toxic substances see Toxic substances

Assist control ventilation 24Assisted spontaneous breathing see Pressure support ventilationAsthma (bronchial asthma) 36-40

mechanical ventilation 37-40goals 38indications 37-8, 38

pathophysiology 36-7Atelectasis see CollapseAtracurium, ventilated patients 29Atrophy, diaphragmatic, ventilated patient 171Autologous bone marrow transplantation, infections 204, 204-5Autonomy, patient 271, 272AutoPEEP see Positive end-expiratory pressure

Bacterial infections/pneumonia, immunocompromised hosts 203, 204-5diagnostic tests 213HIV-infected patients 208, 215post-transplant patients

BMT patients 205, 206solid organs 207

see also Antibiotics; MicrobiologyBaroreceptors, aortic and carotid sinus 17Barotrauma (ventilator-related), pneumothorax risk 132

preventive strategies 233Beneficence 271, 272Benzodiazepines, ventilated patients 29Biochemical studies

interstitial lung disease 237-8pleural effusions 221, 227

Biofeedback, weaning difficulties 177-8Biopsy

bronchial/lungamiodarone toxicity 245interstitial lung disease 237

pleural 222see also Histology

Bleeding see HaemorrhageBleomycin, pulmonary toxicity 244Blood

aspiration 116coughing-up see Haemoptysis

Blood clots, large airways 116Blood gases, arterial see Arterial blood gases; Venous blood,

oxygen saturationBlood volume, intrathoracic (ITBV) 103, 104

fluid management and 288Blunt chest trauma 126-30

mechanisms 126pneumothorax 132

Bone marrow transplantation, infections 203, 204, 204-5Botulism 120Brain injury

hypoxic, in near-drowning 116traumatic 40-43

mechanical ventilation 40-43Brainstem, respiratory control and the 14-15Breathing

active cycle of 71patterns 8

and accessory muscle use 18in mechanical ventilation 23

spontaneousassisted/supplementary see Pressure support ventilationlung distension during, estimation 157-8

work of see WorkBritish Thoracic Society, definition of severe

community-acquired pneumonia 182Bronchial angiography, haemoptysis 109, 110Bronchial asthma see AsthmaBronchial injuries 127Bronchiolitis obliterans organizing pneumonia (BOOP) 236, 240Bronchoalveolar lavage 84

infections/pneumoniaimmunocompromised patients 211, 212nosocomial 193, 194

Broncnopleural fistula 229-33Bronchoscopy 82-6

haemoptysis 83, 108infections

empyema 227immunocompromised patients 211-12nosocomial 84

Brush specimen, protected, nosocomial pneumonia 193, 194Burkholderia pseudomallei 184Burns 83

chemical see Chemicalsthermal, bronchoscopy 83

Caeruloplasmin 150Cancer see MalignancyCandidiasis, pulmonary, BMT patients 205Capnography 94—6Carbon dioxide

arterial tension/partial pressure (PaCO2)central chemoreceptors and 15in COPD, and risk of death and need for intubation 280peripheral chemoreceptors and 16

296 Index

Carbon dioxide cont.principles of measurement 89raised see Hypercapniatemperature correction in measurement 89transcutaneous measurement 95-6in ventilatory drive assessment 17

expired, analysis 94-5extracorporeal removal 162-3rebreathing in non-invasive ventilation 60retention during O2 supplementation, hypercapnic patients 16-17

Carbon monoxide poisoning 119Carboxyhaemoglobin (COHb) 119

pulse oximetry with 92Cardiovascular system

injury 128-9in mechanical ventilation 52-3

in intermittent positive pressure ventilation 30-31see also Circulatory support; Heart

Carotid body chemoreceptors 16Carotid sinus baroreceptors 17Catalase 150Catheters

chest X-ray in monitoring of 81pulmonary artery see Pulmonary artery catheters

Caval filters 259-60Cavitation, X-ray, diagnostic value 211CD 14 and ALI/ARDS 148Cell adhesion molecules and ARDS/ALI 142, 145, 151Cell-mediated immunodeficiency, infections in 202Central nervous system 14-15

chemoreceptors 15respiratory control and the 14-15

Central respiratory depression 119Central sleep apnoea 5Central venous lines, chest X-ray monitoring 81Central venous pressure measurement 102-3Cephalosporins 187Chemical inhalation (and injuries incl. burns) 118-19

bronchoscopy 83children 116

Chemokines and ARDS/ALI 149Chemoreceptors

central 15peripheral 16

Chemotherapeutic drugs, pulmonary toxicity 243-4Chest

CT see Computed tomographyflail 126-7trauma 124-30

blunt see Blunt chest traumabronchoscopy 83penetrating see Penetrating chest traumapneumothorax in 131

X-ray see X-rayChest drainage/aspiration (thoracostomy)

air collections (incl. pneumothorax) 130-31, 134, 232pleural effusions 223-4, 224, 228-9

indications 228X-ray monitoring of tubes 81

Chest physiotherapy 77-8Chest wall

compliance, abnormal, end-expiratory plateaupressure with 157

injuries 126-7Cheyne-Stokes respiration 16Children, accidental poisoning 116Chlamydia spp. causing pneumonia 183

Chronic obstructive pulmonary disease 36-40mechanical/assisted ventilation 37-40, 279-80

goals 38indications 38, 38-9non-invasive 267, 280, 281recognizing and determining need or appropriateness 279-80,280, 281

pathophysiology 36-7respiratory failure 36, 279-80

Churg-Strauss syndrome 246Chylothorax219-20Circulatory support 264Clostridium spp.

botulinum 120tetani 120-21

Clotting system and ARDS 150CMV see Cytomegalovirus infectionCoagulation cascade and ARDS 150Cold-water immersion 115Collagen, in ARDS/ALI 142, 147,149-50Collagen vascular disease see Connective tissue diseaseCollapse, lung (incl. atelectasis)

bronchoscopy 84lobar collapse, X-ray 81

Colloid balance/administration 288, 289Colony-stimulating factor see Granulocyte colony-stimulating factorCommunication, tracheostomy patient 74Community-acquired infections/pneumonia 181-91

admittance to ICU 182age and 182co-morbidity 182epidemiology 181management 186-9

with empyema 228non-improving patient 188-9

mixed infections 184pathogens causing 182-5patient characteristics 181prognosis 185, 186

Compliance 8-11in mechanical ventilation 23-4, 96

end-expiratory plateau pressure with abnormalchest-wall compliance 157

Compressive phase of cough 74Computed tomography of chest

haemoptysis 108-9infections 210interstitial lung disease 236

acute interstitial pneumonia 239cryptogenic organizing pneumonia 241

pleuraeffusions 220-21empyema (vs lung abscess) 226-7

pneumothorax 231portable 86pulmonary embolism 254

Conduction time, phrenic nerve 7Connective tissue (collagen vascular) disease

diffuse alveolar haemorrhage in 247immunodeficiency in 209pleural effusions in 219

Consent 276Consolidation, X-ray, diagnostic value 211Constant flow inflation (=pulse method) 10, 11Continuous lateral rotation therapy see Rotational therapyContinuous positive airway pressure (CPAP) 28, 71

in gastric aspiration 113in heart failure (and associated pulmonary oedema) 44, 65

Index 297

Controlled mandatory ventilation 24Controlled mechanical ventilation (in general) 61

in asthma or COPD 39in brain injury 42

Contusionheart 128lung 127

Co-oximeters 91Cor pulmonale in COPD 36Coronary care units 264Coronary heart disease (ischemic heart disease),

mechanical ventilation 44Corticosteroids see SteroidsCorynebacterium diphtheria infection 121Cough 74-5

physiology 74-5tracheostomy patient 74-5

augmentation 75evaluation 75

see also HaemoptysisCrush injuries, chest 126Cryptococcosis neoformans 209

diagnosis 213treatment 214

Cryptogenic fibrosing alveolitis 236Cryptogenic organizing pneumonia 236, 240-41Crystalloid balance/administration 288, 289CSF see Granulocyte colony-stimulating factorCuffed tracheostomy tubes 73

cuff deflation 74Cuffless tracheostomy tubes 74Cuirass ventilation 73Cyanide poisoning 119Cyclo-oxygenase inhibitors see Non-steroidal anti-inflammatory drugsCyclophosphamide, pulmonary toxicity 244Cytokines (in ARDS/ALI) 147, 148-9

antagonism 166-7Cytology, pleural effusions 221Cytomegalovirus infection

diagnosis 213transplant recipients

bone marrow 206solid organ 207

treatment 214Cytotoxic drugs, pulmonary toxicity 243-4

D-dimer test 251,252Dead space ventilation (VD) 17, 94Deaths (mortalities)

in ARDS/ALIcauses 140epidemiology 140, 153

in community-acquired pneumonia 181in ICU, factors associated with 185

in COPD, pH and PCO2 and risk of 280discharge and risk of 266

Decision aids 276Decontamination of digestive tract, selective 196Demand valves 47, 48Denver shunt 224-6Desquamative interstitial pneumonia 236Diagnostic methods 80-87Diaphragm 1-2

atrophy (ventilated patient) 171EMG 14, 18, 51-2fatigue 4injury 130

length, optimum 4strength measurement 6

Differential lung ventilation in pneumonia 187-8Diffuse alveolar damage (DAD) 236Diffuse alveolar haemorrhage 246-7Digestive tract see Gastrointestinal tractDiphtheria 121Discharge criteria 265, 266Doctors see PhysiciansDomiciliary ventilation 66, 281Double-lumen chest tubes, empyema 228Double-lumen tracheostomy tubes 74Doxycycline pleurodesis 224Drug-induced interstitial lung disease 237, 243-5Drug therapy

ARDS 165-7community-acquired pneumonia 188see also specific (types of) drugs

Dyesinterfering with pulse oximetry 92lung water measurement 103

Dynamic compliance in mechanical ventilation, measurement 10, 11Dynamic hyperinflation see HyperinflationDynamic PEEPp measurement 12, 97-8

E-selectin 151Echocardiography

pulmonary embolism 253-4transoesophageal, bedside 86

Economic issues, intensive care funding 267Effusions see Parapneumonic effusions; PleuraElastance 96

reciprocal of see ComplianceElderly, community-acquired pneumonia and 182Electrical stimulation, phrenic nerve 6Electromyography (EMG) 49

diaphragmatic 14, 18, 51-2oesophageal, sensing inspiration 49

Electrophysiological tests 6-7Embolism 253-4

air, in chest trauma 125-6fat 128pulmonary 253-4, 257-60

clinical features 250, 251diagnosis/differential diagnosis 86, 253-4massive 257-8pleural effusions with 219prevention 259-60sources 249treatment 258-9

Emergencies 105-37bronchoscopy 82medical 105-23

EMG see ElectromyographyEmphysema, subcutaneous 134Empyema, pleural (pyothorax) 189, 226-9

chronic 226definition 226diagnosis 226-8pathogenesis/clinical features 226-7

End-expiratory plateau pressure and abnormal chest-wall compliance 157End-expiratory pressure 97-8

positive see Positive end-expiratory pressure (PEEP)End-inspiratory pressure measurement 97Endocrine disorders 120Endoscopy see Bronchoscopy; ThoracoscopyEndothelium in ARDS/ALI 141, 142, 144-5

298 Index

Endotracheal intubation 29-30, 134-5airflow resistance (in tube) 13anaesthesia for, risks 29avoiding 70-73in brain injury 41bronchoscopy and 82-3chest X-ray of tube position 80-81complications 171

pneumonia 171-2in COPD, pH and PCO2 and need for 280in gastric aspiration 114in haemoptysis 108management 134-5removal of tube see Extubationin status asthmaticus 38

Eosinophil(s), ARDS and 147Eosinophilia

pleural, associated conditions 222pulmonary (and eosinophilic pneumonia) 242-3

Eosinophilic pneumonia 242-3Epithelium

injury in ARDS 145-6irritant receptors 17

Equipmentin infection control 195ventilation see Ventilation, mechanical

Ethical issues 271-7Expectoration, tracheostomy patient 74-5Expiratory area (medulla) 15Expiratory flow, decreased resistance to, ventilated patients 39-40Expiratory muscles 1Expiratory phase of cough 74-5Expiratory time, increase in mechanical ventilation 39

see also Inspiratory: expiratory timeExtracorporeal CO2 removal 162-3Extracorporeal membrane oxygenation 162

community-acquired pneumonia 188Extravascular lung water (EVLW) 103, 104

fluid management and 288Extubation 173, 174-5

criteria 173failure following 176

non-invasive ventilation for 66, 73unplanned 175

Exudative effusions 219-20chronic refractory, management 224—6with empyema 226transudative vs 217, 221

Exudative stage of ALI 142

FactorV Leiden 250Fat embolism 127Fatigue, respiratory muscle 4Fibrinolytics see ThrombolyticsFibrino-purulent stage, empyema 226Fibrosing alveolitis, cryptogenic 236Fibrosis, pulmonary

in ARDS/ALI 142,146-7idiopathic 236

Fibrothorax 229Fistula

bronchopleural 229-33tracheo-oesophageal 116

Flail chest 126-7Flow (airflow)

deprivation, ventilated patients 14

laminar, resistance in 13in mechanical ventilation

expiratory, decreased resistance to 39-40inspiratory flow as trigger 26, 47-8, 49rate 22relationship with airway pressure and

compliance 23-4see also Constant flow inflation

Flow-time diagrams 98-9Flow-volume curves 102Fluid aspiration see AspirationFluid management 167

in refractory hypoxaemia 287-9Food aspiration 112Forced expiratory volume in one second (FEVj) 8Foreign body, aspiration see AspirationFractures, rib 127Free radicals and lung injury 144, 150-51Freshwater aspiration 114Functional residual capacity (FRC) 8, 27

low, physiological effects 27Funding of intensive care 267Fungal pneumonia, transplant recipients

bone marrow 205-6, 206diagnosis 213drug therapy see Antifungal drugssolid organ 208

Gallium-67 scans, interstitial lung disease 236Gas(es)

arterial blood see Arterial blood gasesexchange of, impairment in asthmalCOPD 36harmful, inhalation 118-19

Gas embolism (air embolism), chest trauma 125-6Gastric contents, aspiration 112-14, 172

bronchoscopy 83Gastric pH, low, maintenance 196Gastrointestinal tract (digestive tract; gut)

disorders, pleural effusions with 219selective decontamination 196

G-CSF see Granulocyte colony-stimulating factorGlucocorticoids see SteroidsGranulocyte colony-stimulating factor (G-CSF)

community-acquired pneumonia 188neutropenic patients 212nosocomial pneumonia 198

Granulomatosis, Wegener's 209, 246Granulomatous angiitis, allergic (Churg-Strauss syndrome) 246Great vessel injury 128-9Growth factors and ARDS/ALI 147, 149-50Guillain-Barre syndrome 119, 122Gunshot wounds, high-velocity 125Gut see Gastrointestinal tract

H2-blockers 196Haematological malignancy 209Haemodynamic measurements 102-3Haemoglobin

abnormal species, pulse oximetry with 92oxygen delivery and concentration of 283, 284oxygen saturation (SaO2), measurement 90-91

Haemoptysis, massive 105-10clinical evaluation 105-6management 106-10

bronchoscopy 83, 108site and cause, determination 108, 108-9

Index 299

Haemorrhage/bleedingheparin-related 256pulmonary

diffuse alveolar haemorrhage 246-7interstitial lung disease due to 237

Hamman-Rich syndrome (acute interstitial pneumonia)236, 239-40

Heartinjury 128, 129intermittent positive pressure ventilation effects on 30-31ischaemic disease, mechanical ventilation 44lung and, interactions in mechanical ventilation 52-4output, oxygen delivery and changes in 283, 284

Heart failure 44acute right, in COPD 36congestive 44

pleural effusions in 218mechanical ventilation 44

delayed weaning 172non-invasive 44, 65

Heat and moisture exchangers and nosocomial pneumonia 195Helical CT see Spiral CTHelium dilution technique 8Heparin 254-7

complications 255-7thromboembolic disease

DVT 254-7pulmonary embolism 258

Hepatic dysfunction, pleural effusions 218Herpes simplex virus, transplant recipients


High dependency unit (HDU) 264admission

with community-acquired pneumonia 182criteria 265

High-frequency oscillation ventilation 78, 290High-frequency ventilation 161High-resolution CT, pleural effusions 220-21High-velocity chest wounds

blunt 126penetrating 125

Hilar node X-ray, diagnostic value 211Histamine receptor type 2 blockers 196Histology

amiodarone toxicity 245ARDS/ALI 142cryptogenic organizing pneumonia 241diffuse alveolar haemorrhage 247vasculitis 246

Histoplasma capsulatum treatment 214HIV disease/AIDS 201, 208-9

infections in 203, 204, 208-9treatment 215

Home, ventilation at 66, 281Hospital-acquired pneumonia see Nosocomial infectionsHSV see Herpes simplex virusHuman immunodeficiency virus see HIVHumidification, airway 30, 71Humoral immunodeficiency, infections in 202, 209Hydrocarbon aspiration 116Hydrogen peroxide 150Hypercapnia 161-2

CO2 retention during O2 supplementation in 16-17paralysis allowing adaptation to 158permissive 33, 154, 161-2

inARDS 154,161-2asthmalCOPD patients 39

therapeutic 162ventilatery response in 17

Hypercapnic acute respiratory failurecauses 59non-invasive ventilation 63-4

outcome 65Hyperinflation

dynamic 98in asthma or COPD 36, 37reducing/minimizing 39, 40

manual 77-8in mechanical ventilation 29

in weaning phase 56-7Hypertension, pulmonary, chronic thromboembolic 260Hyperventilation

alveolar, in type I respiratory failure 278, 279therapeutic, brain-injured patients 42

Hypothermia, cold-water immersion 115Hypoventilation, alveolar, in type II respiratory failure 278, 279Hypoxaemia281-91

in asthma 39permissive 282-3refractory 281-91

metabolic considerations 287treatment 287-91

Hypoxaemic acute respiratory failurecauses 59non-invasive ventilation 64

Hypoxiafollowing brain injury 40-41tissue 286

tolerance to 286ventilation strategies and risk of 285

Hypoxic acute respiratory failure see Hypoxaemic acute respiratory failureHypoxic brain injury in near-drowning 116Hypoxic pulmonary vasoconstriction see Pulmonary vasoconstrictionHypoxic ventilatory drive 16

latrogenic damagepleural effusions due to 219pneumothorax due to 131 230-31to respiratory muscle pump 6

Iced saline, haemoptysis 109Imaging/radiology

infections 210interstitial lung disease 236

acute interstitial pneumonia 239-40cryptogenic organizing pneumonia 241

pleuraeffusions see Pleural effusionsempyema 226-7

pneumothorax 82, 231thromboembolic disease

DVT 252pulmonary embolism 254

see also specific modalitiesImmunocompromised host (incl. immunodeficiency diseases), respiratory

infections (incl. pneumonia) 199, 201-16causes, and associated pathogens 202clinical features 210diagnosis 210-12, 213epidemiology 204-9outcome predictors and prognosis 204pathophysiology 201-4

300 Index

Immunocompromised host, respiratory infections cont.preventive strategies 204-9treatment 212-15

Immunological markersinterstitial lung disease 237pleural effusions 221-2

Immunomodulatory therapy, nosocomial pneumonia 198Immunosuppressive drugs, interstitial lung disease 239Impedance plethysmography, DVT 252Index of Rapid Shallow Breathing 173Indocyanine green

lung water measurement 103pulse oximetry and 92

Indwelling device monitoring via chest X-ray 80-81Infections (respiratory/lung incl. pneumonia) 120, 120-21

community-acquired see Community-acquired infectionscontrol of 195immunocompromised host see Immunocompromised hostinterstitial lung disease with 237nosocomial see Nosocomial infectionspleural effusions with 219pulmonary infiltrates with 83-4

Inflammatory conditions, pleural effusions with 219Inhalation

of pulmonary vasodilatory drugs 165unintentional see Aspiration

Injury (non-traumatic), acute lung 138-69at-risk patients, investigating 151brain-injured patients 42definitions 138, 139epidemiology 139-42histological changes 142oxygen-induced see Oxygen, toxicitypathogenesis 143-8precipitating conditions 139-40scoring 138, 139sepsis and multiple organ dysfunction syndrome in 140-41systemic and pulmonary markers 148-51'therapeutic interventions 153-69

clinical trials 140ventilatory strategies 32-6, 99-100

ventilator-associated see Ventilation, mechanicalsee also Acute respiratory distress syndrome

Injury (traumatic/mechanical) see TraumaInspiration, oesophageal EMG in sensing of 49Inspiratory area (medulla) 15Inspiratory: expiratory time (I:E) ratios 33-4, 161

air leaks and 133Inspiratory flow triggering mechanical ventilation 26, 47-8, 49Inspiratory hold button 80Inspiratory muscles 1, 2Inspiratory phase of cough 74Inspiratory pressure-volume curve 9Inspiratory work, measurement 51-2Inspired oxygen concentration (FiO2)

appropriate target levels in hypoxaemia 282-3, 286, 287in recognition of respiratory failure 278

Interferon-[gamma], nosocomial pneumonia 198Interleukins and ARDS/ALI

IL-1P 148-9IL-6 149IL-8 149IL-10 149

Intermittent controlled ventilation, respiratorymuscle function restored by 40

Intermittent mandatory ventilation (IMV) 26synchronized see Synchronized intermittent mandatory ventilation

Intermittent non-invasive ventilation 176Intermittent positive pressure ventilation

(IPPV; IPP breathing) 24, 71, 71-2heart-lung interactions 52sedation 29

Interstitial lung disease 235-48classification 235-6

by cause 237clinical features 236definition 235diagnosis 236-7terminology 235-6therapy 238-9

Intrathoracic blood volume see Blood volumeIntrathoracic thermal volume (ITTV) 103, 104Intrinsic positive end-expiratory pressure see Positive

end-expiratory pressureIntubation see Endotracheal intubationInverse ratio ventilation (IRV) 34, 161Iron-binding anti-oxidant activity 150'Iron lungs' 263Irritant receptors, epithelial 17Ischaemic heart disease, mechanical ventilation 44Isoflurane, ventilated patients 29Isotope (radionuclide) scans, interstitial lung disease 236

Joint receptors 17Justice 271Juxta-capillary receptors 17

Ketoconazole 166Kidney (renal) support 264KL-6 and amiodarone toxicity 245Klebsiella pneumoniae 184

L-selectin 151Laminar airflow, resistance in 13Laryngeal oedema 117,118Laryngospasm 116-17Lavage, bronchoalveolar see Bronchoalveolar lavageLegionella pneumonia (Legionnaire's disease)

community-acquired 183diagnosis 185


Leukaemia 209Life support (ethical issues regarding) 272-3

continuing/withholding/withdrawing 273-5, 280Light, ambient, pulse oximetry and 92Lipopolysaccharide and ALI/ARDS 148Liquid ventilation 163Lisofylline 167Liver dysfunction, pleural effusions 218Lobar collapse, X-ray 81Loeffler's syndrome 242Low-molecular weight heparin, DVT 257Low-velocity chest wounds

blunt 126penetrating 125

Lung(s)abscess, CT 227in ARDS/ALI 155-6

histology 142protective ventilation strategies 34, 156-63

chronic obstructive disease see Chronic obstructive pulmonary diseasecollapse see Collapsedistension during spontaneous breathing, estimation 157-8

Index 301

drug-induced disease 237, 243-5heart and, interactions in mechanical ventilation 52-4infections see Infectionsinfiltrates

with infections 83^1non-infectious causes 209-10

injurynon-traumatic see Injurytraumatic 127-8

interstitial disease see Interstitial lung diseasemechanics 7-14open-lung approach 33, 156-7, 158-60protective ventilation strategies 34

inARDS 154, 156-63in interstitial lung disease 240

recruitment manoeuvres (and de-recruitment) 156-7, 289-90re-expansion

with pleural effusions 224-6with pneumothorax 232-3

resection, bronchoscopy after 85-6surfactant see Surfactantvolume see Volumewater, measurement 103-4X-rays of abnormalities in 81-2

Lung function tests, cryptogenic organizing pneumonia 241Lung transplantation

bronchoscopy after 85-6for failed weaning 177-8in interstitial lung disease 240

Lupus erythematosus, systemic 246Lymph nodes, X-ray, diagnostic value 211Lymphocytosis, pleural, associated conditions 222Lymphoproliferative diseases 209

Macrolides 186Macrophage(s), ARDS/ALI and 147, 149Macrophage inhibitory factor and ARDS 149Magnetic resonance imaging, pleural effusions 221Magnetic stimulation, phrenic nerve 6Malignancy

haematological 209pleural effusions 219

Malnutritionnosocomial pneumonia and 196weaning and 177

Mandatory ventilationcontrolled 24intermittent see Intermittent mandatory ventilation

Manganese superoxide dismutase 150Manual hyperinflation 77-8Manually-assisted cough 75Masks in non-invasive ventilation 59Mast cells and ARDS 147Mean airway pressure measurement 97Mechanical in/exsufflator 75Mechanical ventilation see VentilationMechanics, lung 7-14Mediastinal node X-ray, diagnostic value 211Medulla (brainstem), respiratory control and the 14, 15Melioidosis (Burkholderia pseudomallei) 184Metabolic considerations, refractory hypoxaemia 287Methaemoglobin (MetHb), pulse oximetry with 92Methicillin-resistant S. aureus 183-4Methotrexate, pulmonary toxicity 244Methylene blue, pulse oximetry and 92Microbiology (incl. bacteriology)

community-acquired pneumonia 182-5

empyema 228immunocompromised-host infections, related to cause of

immunodeficiency/immunosuppression 202, 203nosocomial pneumonia 192-3

Mini-tracheostomy 74Minute ventilation 17

decrease in mechanical ventilation (asthma or COPD patients) 39total expired (VE) 94

Missile wounds, chest 125Mobilizing patient 177

in bed 76out of bed 76-7, 177

Monitoring 88-104indwelling device, via chest X-ray 80-81intensive, ethical aspects of patients requiring 273in mechanical ventilation 96-102

in non-invasive ventilation 73neurological 264

Mortalities see DeathsMotion artefact, pulse oximetry 92Movement(s), active, in bed 76Movement artefact, pulse oximetry 92Multidisciplinary team 268-9Multiple organ dysfunction in ARDS/ALI 140-41

ventilator-induced lung injury and 154Multiple trauma with ARDS, airway pressure release ventilation 35-6Murray lung injury score 139Muscle(s)

receptors 17respiratory see Respiratory muscles

Muscle relaxants (neuromuscular blocking agents)for ventilation 29, 158

Muscle spindles 17Myasthenia gravis 3, 122Mycobacterium tuberculosis see Tuberculosis

Near-drowning 114-15Negative pressure ventilation 73, 263Neurogenic pulmonary oedema 42Neurological monitoring 264Neurological problems/disorders 119-20

mechanical ventilation with 22respiratory disease in 3

Neuromuscular abnormalities 121-2critical illness 5-6mechanical ventilation with 22

Neuromuscular blocking agents for ventilation 29, 158Neutropenia, infections in 202, 203, 204-5

treatment 212Neutrophils

antineutrophil cytoplasmic antibodies (ANCA) 246lung injury and 144

Nitric oxide inhalation 165, 290community-acquired pneumonia 188interstitial lung disease 239

Nitrofurantoin, pulmonary toxicity 244-5Nitrogen dioxide inhalation 118-19Nitrogen wash-out technique 8Nocardiosis

diagnosis, 213post-transplant, 207treatment 214

Nodular infiltrates, X-ray, diagnostic value 211Non-invasive ventilation (NIV) in acute respiratory

failure 28, 58-69, 72-3, 172aspiration risk 110-12in cardiogenic pulmonary oedema 44, 65

302 Index

Non-invasive ventilation (NIV) in acute respiratory failure cont.case studies 63-4in COPD 267, 280, 281duration 65historical background 58implementation 62—4interface 59intermittent/sequential use 176invasive vs 58-9modes 59-61monitoring 73outcome 65-6pneumonia (community-acquired) 187practical aspects 72-3spectrum of acute respiratory failure treated 59success, factors influencing 63in type II respiratory failure 281ventilator settings/setting up patient 72weaning with 66-7, 175, 176see also Positive pressure ventilation, non-invasive

Non-maleficence 271, 272Non-steroidal anti-inflammatory drugs (cyclo-oxygenase inhibitors;

prostaglandin synthase inhibitors) 167community-acquired pneumonia 188

Nosocomial infections (incl. pneumonia) 192-200bronchoscopy 84complications 188definitions and epidemiology 192-3diagnosis 193-4pathogenesis 193prevention 194-6risk factors 194treatment 196-8

with empyema 228in ventilated patients 30, 59, 171-2

aspiration of secretions and 196delaying weaning 171-2patient position and 195-6

Nuclear antigen, antibodies to (ANA) 246Nursing expertise 268Nutritional status

nosocomial pneumonia and 196weaning and 177

Obesity, massive 44Obstruction, large airways 116-19Obstructive pulmonary disease, chronic see Chronic obstructive

pulmonary diseaseObstructive sleep apnoea 5Oedema

Iaryngealll7, 118pulmonary 103

cardiogenic, in acute heart failure 44interstitial lung disease due to 237neurogenic 42

see also AngioedemaOesophageal EMG in sensing of inspiration 49Oesophageal pressure measurement 13-14, 51Oesophageal rupture 129-30Oesophagotracheal fistula 116Open-lung approach 33, 156-7,158-60Oral anticoagulants 257Organizational issues 263-70Organizing pneumonia (OP) 236, 240-41

cryptogenic 236, 240-41radiography 240

Osteoporosis, heparin-induced 256-7Oximetry 90-94Oxygen

consumption in respiration (by respiratory muscles), measurement 13, 52content, arterial, calculation 93delivery (DO2) and transport 282-5

calculation of systemic DO2 93factors affecting 284-5maintaining appropriate levels 282-3

inspired, concentration see Inspired oxygen concentrationreactive species of, ALI/ARDS and 144, 150-51saturation

arterial (SaO2), measurement 90-91venous (SvO2), measurement 93-4

tension/partial pressure in arteries (PaO2) 282-3appropriate target levels in hypoxaemia 282-3, 287continuous analysis 90low see Hypoxaemia; Hypoxaemic acute respiratory failuremeasurement principles 89oxygen delivery affected by changes in 284pulse oximetry at high values of 92temperature correction in measurement 89

therapy/supplementation (oxygenation)brain-injured patients 41CO2 retention during 16-17extracorporeal see Extracorporeal membrane oxygenation

toxicity (to lungs) 282-3, 286-7prevention, and risk of tissue hypoxia 285

see also Hypoxia

Paralysis for ventilation 29, 158Paraneoplastic effusions 219Parapneumonic effusions 189

pleural fluid analysis 227Parasitic pneumonia


bone marrow 206solid organ 207-8

treatment 214PDGF and ARDS/ALI 147, 150Peak airway (peak inflation) pressure measurement 97Penetrating chest trauma 124-6

injury pneumothorax 132Pentoxifylline 167Percussion, therapeutic 77Percutaneous procedures

pulmonary embolectomy 259tracheostomy 85 134

Perfluorocarbons 163Perfusion, tissue, and pulse oximetry 92

see also Ventilation-perfusionPericardial injury 129Peripheral chemoreceptors 16pH, arterial blood

in COPD, and risk of death and need for intubation 280measurement 89

temperature correction 89-90Pharmacotherapy see Drug therapyPhlebography (venography), DVT 252Phlegmasia cerulea dolens 250Phrenic nerve

damage 1-2iatrogenic 6

electrophysiological measurements 6-7stimulation 6, 7

Index 303

Physicians (doctors)ICU, role 273respiratory 268

Physiotherapy 70-79Plain films see X-rayPlateau pressure 97

end-expiratory, and abnormal chest-wall compliance 157Platelet(s), ARDS/ALI and 147Platelet-derived growth factor and ARDS/ALI 147, 150Plethysmography, impedance, DVT 252Pleura 217-34

biopsy 222empyema see Empyemafluid analysis, empyema 227-8pressure measurement 18X-rays of abnormalities in 81-2

Pleural effusions 217-26bronchoscopy 84-5causes 218-20clinical findings 220diagnostic procedures 221-2imaging 220-21

X-rays 211, 220management 223-6

Pleurodesis 224-6pneumothorax 134

Pneumococcus see Streptococcus pneumoniaePneumocystis carinii

diagnosis 213in HIV disease/AIDS 208

prognosis 204transplant patients

bone marrow 205-6solid organ 208

treatment 214Pneumocytes and ARDS/ALI 145-6, 146Pneumomediastinum 134Pneumonectomy (lung resection), bronchoscopy after 85-6Pneumonia 181-200

aspiration 113community-acquired see Community-acquired infectionseosinophilic 242-3immunocompromised host see Immunocompromised hostinterstitial

acute 236, 239-40desquamative 236non-specific 236usual 236

nosocomial see Nosocomial infectionsorganizing see Organizing pneumoniapleural effusions with 219

Pneumonitis, amiodarone 245Pneumotaxic centre 15Pneumothorax 131-4, 229-33

diagnosis/investigations 231-2X-ray 82, 231

management 132-4, 232-4primary spontaneous 131,229-30secondary spontaneous 131,132, 232

recurrence and its prevention 233traumatic/iatrogenic 131,230-31

Poiseuille's law 13Poisoning, children, accidental 116Poliomyelitis 120Polyneuritis, acute ascending (Guillain-Barre syndrome)

119, 122

Polytrauma (multiple trauma) with ARDS, airway pressure releaseventilation 35-6

Pons, respiratory control and the 14, 15Positioning (posture), patient 76, 290-91

in community-acquired pneumonia treatment 187enhancing respiratory muscle function 71reducing gastric aspiration 172reducing respiratory load 71in refractory hypoxaemia 290-91in ventilated patients, nosocomial pneumonia and 195

Positive airway pressure see Positive pressure ventilationPositive end-expiratory pressure (PEEP) 27-8, 33, 33-4, 97-8, 158-60

external (and in general or unspecified) 282in acute interstitial pneumonia 240added to autoPEEP 53-4in ALI/ARDS 33, 155, 155-6, 156-7, 158-60asthmalCOPD patients 40in brain injury 41in haemoptysis 108pneumothorax risk 231

intrinsic (autoPEEP) 11-13, 29, 33-4, 53-4, 97-8clinical implications 53-4dynamic 12, 97-8external PEEP added to 53-4measurement 12, 97-8in weaning phase 56-7

'liquid' 163in 'open-lung' approach 33, 156, 158-60selection 158-60

Positive pressure ventilation (positive airway pressure) 71-2continuous see Continuous positive airway pressurein gastric aspiration 114intermittent see Intermittent positive pressure ventilationnon-invasive (NIPPV) 71-2, 265, 267

indications 267, 268Postoperative procedures

bronchoscopy 85-6non-invasive ventilation 65-6

Posture see PositioningPressure

airway see Airway pressureairway occlusion, measurement during weaning from

ventilation 56-7central venous, measurement 102-3measurements in ICU 6oesophageal, measurement 13-14, 51pleural, measurement 18pulmonary artery occlusion, measurement 102-3see also Airway pressure contour; Positive end-expiratory pressure

Pressure-controlled ventilation (PCV) 25, 61, 161in ARDS/ALI 34, 157, 161in asthma or COPD 39in brain injury 42

Pressure support ventilation (PSV; assisted/supplementaryspontaneous breathing) 26-7, 34, 60, 172

inALI 157asthmalCOPD patients 40CPAP plus, in cardiogenic pulmonary oedema 65triggering 61in weaning process 172, 173

Pressure-time diagrams 98-9Pressure-time product 14, 51Pressure-volume curve 158-60

inspiratory 9Pressure-volume curves 99-102Procollagen (collagen precursors) and ARDS/ALI 142, 147, 150

304 Index

Proliferative stage of ALI 142Prone positioning 76

in refractory hypoxaemia 290Propofol, ventilated patients 29Proportional assist ventilation (PAV) 27, 61Prostacyclin

inhaled 165community-acquired pneumonia 188

intravenous 166Prostaglandin E1 use 166Prostaglandin synthase inhibitors see Non-steroidal

anti-inflammatory drugsProtected brush specimen, nosocomial pneumonia 193, 194Protein C, activated, resistance 250Pseudomonas aeruginosa

community-acquired 184HIV disease 208transplant recipients 207

Psychological factors affecting weaning 177Pulmonary acute care unit (respiratory ICU; RICU),

264-5, 267Pulmonary angiography, haemoptysis 109Pulmonary artery catheters 267

pulmonary capillary wedge catheter monitoring by X-ray 81Pulmonary artery occlusion pressure, measurement 102-3Pulmonary embolism, diagnosis 86Pulmonary hypertension, chronic thromboembolic 260Pulmonary non-vascular aspects/problems/disorders etc. see LungPulmonary vasculature

manipulation 165resistance in mechanical ventilation 52

Pulmonary vasoconstriction, hypoxic 165drugs influencing 165in hypercapnic patients 16-17in pneumonia 187

Pulse method 10, 11Pulse oximetry, principles 91-2Pyogenic bacterial infection


Pyothorax see Empyema

Quinolones 186

Radiograph, chest see X-rayRadiology see Imaging and specific modalitiesRadionuclide scans, interstitial lung disease 236Rapid airway occlusion technique 9,11Rapid Shallow Breathing Index 173Reactive oxygen species and ALI/ARDS 144,150-51Re-admission criterialpolicies 265, 266Receptors in respiratory control 15-16, 17Recoils, lungs/chest wall, passive 15Re-expansion see LungsReflex responses in intermittent positive pressure ventilation 31Rehabilitation 75-8, 177

in-bed strategies 76out-of-bed strategies 76-7

Relaxation phase of cough 75Renal support 264Research, consent for participation in 276Resistance

airway/respiratory system 12-13decreasing, asthmalCOPD patients 39-40measurement 12-13, 96

pulmonary vascular, in mechanical ventilation 52Resolution in ALI/ARDS 146

Respiration, Cheyne-Stokes 16Respiratory centres 14-15Respiratory control 14-18Respiratory depression, central 119Respiratory distress syndrome, acute see Acute respiratory

distress syndromeRespiratory drive 14-18, 54-7

measurement 56-7in mechanical ventilation 54-7

Respiratory failure, acute 58-69, 278-93in COPD 26, 279-80management 278-93

practical guidelines 61-2non-invasive ventilation see Non-invasive ventilationrecognizing 278-9type I 278type II 278, 279

weaning 281Respiratory ICU (RICA) 264-5, 267Respiratory insufficiency in immunocompromised

patients, treatment 212Respiratory load, positioning reducing 71Respiratory muscles 1-7

accessory, breathing pattern and use of 18acquired damage 5-6assessment in ICU 6-7disease processes 3failure 4-5

in asthma or COPD 36fatigue 4intermittent controlled ventilation restoring function 40O2 consumption, measurement 13, 52positioning enhancing function 71shortening 4strength

impairment in ventilated patients 171measurement 6

training 171work see Work

Respiratory syncytial virus (RSV)diagnosis 213treatment 214

Resuscitation, in near-drowning 115Re-warming, cold-water immersion 115Rhythm of ventilation see VentilationRib fractures 127Rifampicin 186, 187Rotational therapy (incl. continuous lateral rotation therapy) 76, 290-91

interstitial lung disease 239in refractory hypoxaemia 290-91

RSV see Respiratory syncytial virus

Saline wateraspiration (seawater) 114iced, in haemoptysis 109

Scintigraphy (radionuclide scans), interstitial lung disease 236Seawater aspiration 114Secretions, aspiration, ventilated patients 196Sedation

and nosocomial pneumonia risk 196for ventilation 29

Seldinger technique 135Selectins 151Selective decontamination of digestive tract 196Self-extubation 175-6Self-poisoning 119Semi-recumbent positioning 76

Index 305

SepsisinARDS/ALI 140-41severe, immune dysfunction in 209

Shunt fraction, calculation 93-4Single-lumen tracheostomy tubes 73-4, 74Sleep 5Sleep apnoea 5Sleep deprivation in mechanical ventilation 60Smoke inhalation 119Spasm, laryngeal 116-17Speech, tracheostomy patient 74Spiral (helical) CT

haemoptysis 108-9pulmonary embolism 254

Sputum plugs 116Sputum samples, infection diagnosis 210-11Staphylococcus aureus causing community-acquired pneumonia 183-4Static compliance in mechanical ventilation, measurement 9-10Static PEEP;, measurement 12Status asthmaticus, mechanical ventilation 37, 38Steroids (corticosteroids; glucocorticoids) 166

community-acquired pneumonia 188interstitial lung disease 239

amiodarone pneumonitis 245cryptogenic organizing pneumonia 241eosinophilic pneumonia 242

Stomach see entries under GastricStreptococcus pneumoniae (pneumococcus),

community-acquired infection 183antibiotics 186co-infection with other pathogens 184diagnosis 185

Streptokinaseintrapleural 189, 228, 229pulmonary embolism 258

Stretch receptors, pulmonary 17Strongyloides stercoralis

diagnosis 213transplant recipients 208

Subcutaneous emphysema 134Super-syringe method 9, 10Superoxide dismutase, manganese 150Surfactant 164-5

compositional changes in ARDS 164replacement therapy 164-5

Surgeryempyema 229haemoptysis 110lung transplant see Lung transplantationlung volume-reduction 177-8procedures following see Postoperative procedurespulmonary embolism 258-9

Synchronized intermittent mandatory ventilation 26, 172in weaning process 172, 173

Systemic inflammatory response syndrome 140^11ventilator-induced lung injury and 154

Systemic lupus erythematosus 246

T-piece 172,173Talc poudrage/pleurodesis 224Tamponade, traumatic 129Team, multidisciplinary 268-9Technetium-99m scans, interstitial lung disease 236Temperature correction, arterial blood gas/pH analysis 89-90Tetanus 120-21Tetracycline pleurodesis 224TGFs and ARDS/ALI 147

Therapeutic Intervention Scoring System 266Thermal burns, bronchoscopy 83Thermodilution method, lung water measurement 103-4Thoracentesis 221-2Thoracoscopy 222, 224

medicalempyema 229pneumothorax 231-2

video-assisted see Video-assisted thoracoscopic surgeryThoracostomy see Chest drainageThoracotomy, chest trauma 125Thorax see ChestThrombocytopenia, heparin-induced 256Thromboembolism, venous see Embolism, pulmonary;

ThrombosisThrombolytics/fibrinolytics

pleural instillation 189, 228-9pulmonary embolism 258

Thrombophilias, inherited 250Thromboplastin time, activated partial, heparin dose and 254Thrombosis, deep venous (DVT) 249-52

aetiology/pathogenesis 249-50clinical features 250—51diagnosis 251-4laboratory features 251predisposing factors 250prognostic categories 249treatment 254-7

Tidal volume 17in mechanical ventilation 157

limiting 33Tilt table 77Tissue hypoxia see HypoxiaTissue perfusion and pulse oximetry 92Tissue plasminogen activator, pulmonary embolism 258TNF-a see Tumour necrosis factor-aTobin's Index 173Toxic substances (and poisons), inhalation/aspiration 118-19

children 116self-poisoning 119see also Drug-induced interstitial lung disease

Toxoplasmosis (T. gondi)diagnosis 213transplant recipients

bone marrow 206solid organ 207-8

treatment 214Trachea

injuries 127intubation see Endotracheal intubation

Tracheo-oesophageal fistula 116Tracheostomy 73-5, 134-5

management 134-5percutaneous 85, 135

Transcutaneous CO2 tension measurement 95-6Transferrin 150Transforming growth factors and ARDS/ALI 147Transoesophageal echocardiography, bedside 86Transplantation

bone marrow, infections associated with 203, 204solid organ

infections associated with 203lung see Lung transplantation

Transudative effusions 218exudative vs 217, 221therapy 223

of refractory chronic effusions 224—6

306 Index

Trauma 118brain see Brain injurychest see Chest, traumaimmune dysfunction in 209lung 127-8multiple, with ARDS, airway pressure release

ventilation 35-6see also Burns

Triggering mechanical ventilation 25-6, 40, 47-9clinical problems with 49in pressure support ventilation 61

Tuberculosis (M. tuberculosis infection) 121diagnosis 213in HIV disease 208-9in transplant recipients 207treatment 214

Tumour necrosis factor-a (and ARDS/ALI) 148antibody to 167

Ultrasound (diagnostic)bedside 86cardiac see EchocardiographyDVT 252pleural effusions 220

United States see USAUrokinase

intrapleural 228pulmonary embolism 258

USAmultidisciplinary critical care 268weaning centres 268withholding/withdrawal of life support 274

Vagus nerve 17Valves, demand 47,48Varicella-zoster virus

post-BMT 206treatment 214

Vascular injury 128-9Vasculitis 237Vasoconstriction, pulmonary see Pulmonary vasoconstrictionVecuronium, ventilated patients 29Vena caval filters 259-60Venography, DVT 252Venous blood, oxygen saturation (SvO2), measurement 93-4Venous lines, central, chest X-ray monitoring 81Venous pressure measurement, central 102-3Venous thromboembolism see Embolism, pulmonary;

ThrombosisVentilation (natural)

intrinsic rhythm 14-15modification 15

minute see Minute ventilationVentilation, mechanical 21-69, 153-5, 156-63

assisted mode 25-7asthmalCOPD patients 40brain-injured patients 42

avoiding 70-73complications (incl. lung injury) 32, 38, 100-101, 147-8, 153-5

infections see Nosocomial infectionspneumothorax (and its management) 132, 132-3, 231, 233prevention 156-63

in COPD see Chronic obstructive pulmonary diseaseequipment

failure 29and infection control 195

in flail chest 127in gastric aspiration 114in haemoptysis 108historical background 21, 58at home 66, 281immunocompromised patients 212indications 22in interstitial lung disease 238-9modes/methods 24

newer 161non-invasive ventilation 59-61

monitoring see Monitoringin near-drowning 115non-invasive see Non-invasive ventilationpatient interactions 47-57in pneumonia 187-8practical aspects 28-31recent advances, 27-8static compliance during, measurement 9-10strategies 32-5

in ALI 32-6, 99-100, 156-63protective see Lung, protective ventilation strategies

triggering see Triggeringweaning see Weaningwork of breathing 13-14, 49-52see also specific methods

Ventilation-perfusion (V7Q) relationships 94Ventilatory drive

assessment 17-18hypoxic 16

Ventricular failure, left, in ischaemic heart disease 44Video-assisted thoracoscopic surgery (VATS) 222

empyema 229Viral pneumonia



treatment 214Vital capacity 8Volume(s)

intrathoracic blood see Blood volumeintrathoracic thermal (ITTV) 103, 104lung 7-8

cardiovascular impact of 52-3surgical reduction 177-8see also Flow-volume curves; Pressure-volume curves;

Tidal volumeVolume-controlled ventilation (VCV) 24-5, 61,161

in ARDS 25, 34, 157,161in brain injury 42

VZV see Varicella-zoster virus

Warfarin 257Water

aspiration 114-15lung

extravascular see Extravascular lung watermeasurement 103-4

Weaning from mechanical ventilation 56-7, 66-7, 170-80, 268airway occlusion pressure measurement during 56-7brain-injured patients 42centres for 268definition of weaned 170difficulties/delays 5, 177-8

causes 170-71, 177-8,281

Index 307

failure 176-7modality 172-3

weaning time 173with non-invasive ventilation 66-7, 175, 176protocols 173-5requirements 172in type II respiratory failure 281

Wegener's granulomatosis 209, 246Work, respiratory muscle (work of breathing) 13-14,

49-52measurement 13-14, 51-2

clinical applications 52in mechanical ventilation 13-14, 49-52

X-ray radiograph (plain films), chesthaemoptysis (massive) 106infections 210interstitial lung disease 236

acute interstitial pneumonia 239cryptogenic organizing pneumonia 241

in Murray Lung Injury Score 139pleura

effusions 211, 220empyema 226

pneumothorax 82, 231portable 80-82

Xanthine oxidase 150

Respiratory Critical Care

Documents

Transcript of Respiratory Critical Care