Sepsis and Organ Dysfunction: The Challenge Continues

Sepsis and Organ Dysfunction The Challenge Continues

Springer-Verlag Italia Srl. Springer-Verlag Italia Srl.

A.E. Baue G. Berlot A.Gullo J.-L. Vincent (Eds)

Sepsis and Organ Dysfunction The Challenge Continues

ORGAN FAILURE ACADEMY

Springer

A.E. BAUE, M.D. Department of Surgery, Saint Louis University, Health Sciences Center, St. Louis - USA

G. BERLOT, M.D. Department of Anaesthesia, Intensive Care and Pain Therapy University of Trieste, Cattinara Hospital, Trieste - Italy

A. GULLO, M.D. Department of Anaesthesia, Intensive Care and Pain Therapy University of Trieste, Cattinara Hospital, Trieste - Italy

J.-L. VINCENT, M.D. Department of Intensive Care, Erasme University Hospital Free University of Brussels - Belgium

O.F.A. - ORGAN FAILURE A CADEMY, VIA B ATIISTI , 1 - 34125 TRIESTE (ITALY)

Steering Committee A.E. Baue, M.D., Department of Surgery, Saint Louis University Health Sciences Center, St. Louis - USA

G. Berlot, M.D., Department of Anaesthesia, Intensive Care and Pain Therapy University ofTrieste, Cattinara Hospital , Trieste . Italy

A. Gullo, M.D., Department of Anaesthesia, Intensive Care and Pain Therapy University of Trieste, Cattinara Hospital, Trieste - Italy

L. Silvestri, M.D., Department of Anaesthesia, Intensive Care and Pain Therapy University of Trieste, Cattinara Hospital , Trieste - Italy

G. Sganga, M.D. , Department of Surgery, and C.N.R. Shock Centre Catholic University of Sacro Cuore, Rome . Italy

© Springer-Verlag Italia 2000 Originally published by Springer-Verlag Italia, Milano 2000

ISBN 978-88-470-0096-4 ISBN 978-88-470-2284-3 (eBook) DOI 10.1007/978-88-470-2284-3

Library of Congress Cataloging-in-Publication Data: Applied for

This work is subject to copyright. AII rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in other ways, and storage in data banks. Duplication of this publication or parts thereof is only permitted under the provisions of the Italian Copyright Law in its current version and permission for use must always be obtained from Springer-Verlag. Violations are liable for prosecution under the Italian Copyright Law.

The use of general descriptive names, registered names, trademarks, etc., in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protecti ve laws and regulations and therefore free for general use.

Product liability: the publishers cannot guarantee the accuracy of any information about dosage and application contained in this book. In every individual case the user must check such infonnation by consulting lhe relevant literature.

Cover design: Simona Colombo, Milan

SPIN: 10753508

Table of Contents

Introduction A.E. BAUE.................................................................................................................................... 11

Sepsis and Organ Dysfunction. The Challenge Continues G. BERLOT, U. LUCANGELO, AND A. GULLO ................................................................................ 17

OXYGEN TRANSPORT IN SEPSIS

Biochemical Regulation of the Microcirculation G.P. NOVELLI ............................................................................................................................... 37

Oxygen Supply and Consumption in Tissues A. MAYR, W. PAIK, AND W. HASIBEDER ........ ............ ...... .................. .................. ........ ...... .......... 43

Ischaemia-Reperfusion in Sepsis C. ADEMBRI, A.R. DE GAUDIO, AND G.P. NOVELLI ..................................................................... 49

Mechanism of Oxygen Extraction Defect in Septic Shock W. PAlK, H. KNOTZER, AND W. HASIBEDER.. .......................... ...... ...... ........ ...... .................... ....... 57

ORGAN DYSFUNCTION AND BIOHUMORAL MISMATCH IN SEPSIS

Gut Perfusion in Sepsis and Shock J.F. PALIZAS ................................................................................................................................. 67

Pathophysiology of Encephalopathy N. LATRONICO, G.F. BUSSI, AND A. CANDIANI............................................................................. 77

Lung Dysfunction in the Early Phase of Sepsis P. NEUMANN ..................................... '" ................... , ... ... ... ... ......... ... ......... ... ...... ... ... ... ... ... ..... ... .... 85

The Kidney in Sepsis J.A. KELLUM................................................................................................................................ 91

Pathophysiology of Liver Dysfunction in Sepsis N. BRIENZA .................................................................................................................................. 103

Inflammatory Cells in Septic Shock H. ZHANG, C. HSIA, AND G. PORRO ............................................................................................. 107

SEPSIS TRIAL

"Revised Terminology on Sepsis J.-L. VINCENT ............................................................................................................................... 115

VI

The Epidemiology and Outcome of Patients with Sepsis: Clear as Mud R.S. WAX, AND D.e. ANGUS ....................................................................................................... 123

Are there Useful New Markers of Sepsis? M. MEISNER, AND K. REINHART ................................................................................................... 137

A Paradigm Shift: The Bidirectional Effect of Inflammation on Bacterial Growth G.U. MEDURI ............................................................................................................................... 145

Is the Dosing and Timing of the Intervention Adequate? A.F. SUFFREDINI ........................................................................................................................... 155

Clinical Trials of Mediator-Targeted Therapy in Sepsis J.e. MARSHALL ............................................................................................................................ 161

Index ........................................................................................................................................... 175

Authors Index

Adembri C. Department of Anaesthesiology and Intensive Care, Florence University School of Medicine, Florence (Italy)

Angus D.C. Department of Anaesthesiology and Critical Care Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania (U.S.A.)

BaueA.E. Department of Surgery, St. Louis University Medical Centre, St. Louis, Missouri (U.S.A.)

Berlot G. Department of Anaesthesiology and Intensive Care, Trieste University School of Medicine, Trieste (Italy)

Brienza N. Department of Anaesthesia and Intensive Care, Bari University School of Medicine, Bari (Italy)

Bussi G.F. Department of Anaesthesia and Intensive Care, University of Brescia, Brescia (Italy)

CandianiA. Department of Anaesthesia and Intensive Care, University of Brescia, Brescia (Italy)

De Gaudio A.R. Department of Anaesthesiology and Intensive Care, Florence University School of Medicine, Florence (Italy)

GulloA. Department of Anaesthesiology and Intensive Care, Trieste University School of Medicine, Trieste (Italy)

Hasibeder W. Department of Anaesthesia and General Critical Care Medicine, The Leopold Franzens University of Innsbruck, Innsbruck (Austria)

Hsia C. Division of Respiratory Medicine, Mount Sinai Hospital, Toronto University, Toronto, Ontario (Canada)

Kellum J.A. Department of Anaesthesiology and Critical Care Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania (U.S.A.)

Knotzer H. Department of Anaesthesia and General Critical Care Medicine, The Leopold Franzens University of Innsbruck, Innsbruck (Austria)

Latronico N. Department of Anaesthesia and Intensive Care, University of Brescia, Brescia (Italy)

Lucangelo U. Department of Anaesthesiology and Intensive Care, Trieste University School of Medicine, Trieste (Italy)

Marshall J.C. Department of Surgery, The University of Toronto, Toronto General Hospital, University Health Network, Toronto, Ontario (Canada)

VIII

MayrA. Department of Anaesthesia and General Critical Care Medicine, The Leopold Franzens University of Innsbruck, Innsbruck (Austria)

Meduri G.D. Department of Medicine, Pulmonary and Critical Care Division, University of Tennessee Medical Centre, Memphis, Tennessee (U.S.A.)

Meisner M. Department of Anaesthesiology and Intensive Care Therapy, University Hospital, lena (Germany)

Neumann P. Department of Anaesthesiology, Emergency and Intensive Care Medicine, University Hospital, Giittingen (Germany)

Novelli G.P. Department of Anaesthesiology and Intensive Care, Florence University School of Medicine, Florence (Italy)

Pajk W. Department of Anaesthesia and General Critical Care Medicine, The Leopold Franzens University of Innsbruck, Innsbruck (Austria)

Palizas J.F. Department of Intensive Care, Bazterrica Clinic, Buenos Aires (Argentina)

Porro G. Division of Respiratory Medicine, Mount Sinai Hospital, Toronto University, Toronto, Ontario (Canada)

Reinhart K. Department of Anaesthesiology and Intensive Care Therapy, University Hospital, lena (Germany)

Suffredini A.F. Department of Critical Care Medicine, National Institute of Health, Bethesda, Maryland (U.S.A.)

Vincent J.-L. Department of Intensive Care, Free University of Bruxelles, Erasme Hospital, Bruxelles (Belgium)

Wax R.S. Department of Anaesthesiology and Critical Care Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania (U.S.A.)

Zhang H. Division of Respiratory Medicine, Mount Sinai Hospital, Toronto University, Toronto, Ontario (Canada)

Abbreviations

AAA, aromatic amino acids

ALT, alanine aminotransferase

ARDS, acute respiratory distress syndrome

ARF, acute respiratory failure

AST, aspartate aminotransferase

ATN, acute tubular necrosis

ATP, adenosine-triphosphate

BAL, bronchoalveolar lavage

BALF, bronchoalveolar lavage fluid

BBB, blood-brain barrier

BCAA, branched chain amino acid

BPI, bactericidal permeability increasing pro

tein

CARS, compensatory anti-inflammatory re-

sponse syndrome

CBF, cerebral blood flow

CMR02, cerebral metabolic rate of oxygen

CNS, central nervous system

eNOS, constitutive NO synthase

CRP, C-reactive protein

CSF, cerebrospinal fluid

DIC, disseminated intravascular coagulation

EAA, excitatory amino acids

ENDOCAB, endotoxin core antibody

FOR, free oxygen radicals

GABA, y-aminobutyric acid

G-CSF, granulocyte colony stimulating factor

GFR, glomerular filtration rate

GM-CSF, granulocyte-macrophage colony

stimulating factor

HDR, host defense response

HMG-l, high mobility group-l protein

HNP-1-3, human neutrophil peptide

IL-l, interleukin-l

INF -')', interferon-y

iNOS, inducible NO synthase

I-R, ischaemia-reperfusion

LBP, lipopolysaccharide binding protein

LOCM, low-osmolality contrast media

LPS, lipopolysaccharide

MAP, mean arterial pressure

MARS, mixed inflammatory and anti-inflam-

matory response syndrome

MIP-2, macrophage inflammatory protein-2

MODS, multiple organ dysfunction syndrome

MOF, multiple organ failure

NMMA, N-monomethyl arginine

NO, nitric oxide

OR, oxygen radical

PAF, platelet activating factor

PA02, alveolar oxygen tension

PE, plasmapheresis

PMNs, polymorphonuclear granulocytes!cells

RBF, renal blood flow

RVR, renal vascular resistance

SAO/R, splanchnic artery occlusion and reperfusion

SE, septic encephalopathy

SIRS, systemic inflammatory response syn

drome

TFPI, tissue factor pathway inhibition

TNF, tumour necrosis factor

YAP, ventilator-associated pneumonia

XD, xanthine dehydrogenase

XO, xanthine oxidase

Introduction

A.E. BADE

What is new about sepsis and organ failure? In a few moments we will hear about sepsis terminology, the incidence and mortality of sepsis, new mediators and reasons for failure in therapy. I will not comment on the areas where our speakers are the experts. I look forward to learning from them.

However, is there anything else which complicates the scene of sepsis and organ failure? I will raise several questions some of which may be answered in the presentations to follow and some may come up in the discussion at the end of the presentations. The first question for the presenters is: How do you treat sepsis? Next, there are a number of new (and confusing) concepts of organ damage and variations in patients that have often not been recognised. Are they valid? How important are they? I will review several of them with you.

Some of the new concepts of organ damage which are interesting but also confusing include cytokine pleiotropy and redundancy. These have been reviewed recently by Sanchez-Cuenca et al. Their concept of cytokine functional pleiotropy and redundancy is that a specific cytokine might play several biological roles in various tissues and cells and several different cytokines may exert similar and overlapping functions on certain cells. They believe that this can be explained on a molecular level by the small gp130 group of cytokines [1]. The complexities of cytokines are brought out by the association of tumor necrosis factor microsatellites with both the incidence and outcome of severe sepsis. Shu et al. described a variety of polymorphisms within the TNF locus and found that certain micro satellites were associates with a fatal outcome of severe sepsis [2]. Circulating cell interactions are also very important and are difficult to study. Recently, Aziz et al. reviewed the neutrophil-platelet interactions in infection. They found that platelets added to neutrophils increase the resistance to cell filtration and thus the stickiness of these cells. Such interactions may impair microvascular blood flow in sepsis [3].

The question can then be asked as to whether all of SIRS or most of sepsis is due to micro-organisms, viruses or bacteria which have not been detected by the usual methods of isolation, culture and transmission. The detection of microbial DNA in the blood by polymerase chain reaction (PCR) methodology is much more sensitive and will help to define new aspects of sepsis and infection

12 A.E. Baue

[4]. PCR can detect bacterial fragments in seven minutes [5]. This is a new frontier [6]. There is now a specialised pathogen laboratory in California working with the Unexplained Illness Working Group. This is a network of infectious disease experts co-ordinated by the Centers for Disease Control and Prevention (CDC) in Atlanta [7]. Their techniques have led to identification of the enteroviruses, the Ebola and Hanta viruses, new herpes viruses, the virus which causes Whipple's disease and other diseases. There are other techniques using PCR including consensus PCR, micro-array techniques and pathogen detection slips. Much more will be learned about the usefulness of these techniques in the near future [8].

How many more mediators are there? The more we learn, the more complex these agents become. Another new and confusing concept of organ damage is based on mesenteric lymph - right or wrong? Is it important or is it irrelevant? The evidence is confusing. Upperman with Deitch and their group found in an animal model that post-haemorrhagic shock mesenteric lymph is cytotoxic to endothelial cells and activates neutrophils [9]. Moore in Denver immediately jumped on the band wagon and announced that mesenteric lymph must be the critical bridge between dysfunctional gut and multiple organ failure [l0]. However, Lemaire with Stoutenbeek et al. found that thoracic duct lymph from patients with multiple organ failure did not have a high endotoxin content. Levels of pro-inflammatory cytokines were low but those of anti-inflammatory cytokines very high. This suggests that the thoracic duct is not a major route of bacterial translocation in patients with MOF [11].

Recently a post implantation syndrome has been documented. When devices are implanted in the body, there is an immune inflammatory response. Zimmer et al. found that, when aortic stents were inserted for arterial aneurysm repair, there is an exaggerated inflammatory response to such devices [12]. This type of reaction has also been observed by Spanier et al. with implanted left ventricular assist devices. These are believed to act as implanted immune-inflammatory organs [13]. Replacement rather than blockade of various markers and substances has recently received active consideration. There is increased mortality in individual with a low protein C concentration. Thus, activated protein C, rhAPC is now being tested in phase I and phase II trials in patients with sepsis [14]. Transgenic and knockout models of surgical disease have been proposed by Arbeit and Hirose [15]. These authors suggest that genetically manipulated mammals may add considerably to our knowledge. I think it is important to learn about these things. They will help to determine what is important; however, can such models explain the redundancy, the overlap and the interactions that occur in a complex biological system, since they are based on one knockout or transgenic change in each animal system?

Another area of new and confusing concepts that contribute to organ damage is that of arteriosclerosis and the question of whether it is due to infection. There have been many recent bacteriological surprises including the importance of C. difficile in various forms of diarrhoea and other large bowel problems and

Introduction 13

the involvement of H. pylori in both gastric and duodenal ulceration and also as a factor in the development of gastric cancer. The bacteria Chlamydia pneumoniae has been found in 70% of plaques of individuals with arteriosclerosis that have been autopsied. Other agents that have been implicated are the herpes simplex cytomegaloviruses and cytomegaly. Are these causes or just associations? This has been reviewed thoroughly by Hatch recently [16]. The evidence is suggestive but a cause and effect relationship remains to be established. How this will be done also remains to be determined.

There have also been studies recently showing genetic differences in the cytokine response to sepsis. This involves DNA polymorphism for TNF, IL-l and IL-l RA. Some of these factors protect and some produce a poor prognosis with variations in patients. This evidence has been provided by Zehnbauer, Freeman and Buchman [17]. Other evidence for differences in immunity between patients has been laid out by Bennett-Guerrero et al. They found that pre-operative, preillness differences in immunity and prior immune history contributed to the outcome after cardiac surgery. They measured the endotoxin core antibody (ENDOCAB) in patients having cardiac operations. Low pre-operative levels predicted adverse post-operative outcomes. Patients with high levels did much better, had fewer infections and other problems. This suggests the need for exposure to Gram negative bacteria and other agents before cardiac operations in order to increase immunological sufficiency [18].

Angele, Chaudry and their group showed that after injury male animals do less well than females. They have studied this phenomenon extensively and found that it is related to testosterone. Testosterone receptor blockade improves the outcome for males. When testosterone secretion decreases in elderly male animals the differences between males and females decrease [19]. In another study in animals Sam et al. found that sepsis produced depression of testosterone and the steroidogenic acute regulatory protein [20]. Recently Offner et al. found that males had dramatically more major infections following trauma than women [21]. However, in septic patients Eachempati et al. found that there was a higher mortality in females than in men [22].

With autoimmune diseases, however, the reverse is true. Prolactin and estrogen seem to be problems. Multiple sclerosis occurs twice as frequently in females, rheumatoid arthritis is three times as frequent and lupus erythematosus is nine times more frequent in females than in males. This has been described recently by Whitacre et al. [23]. Other variations in patients include age, prior illness such as chronic obstructive pulmonary disease, lifestyle, obesity, nutrition and lack of exercise. Thus, a tremendous amount of scientific information has become available which must be assimilated before we can put it into use.

This exciting science has led to much excitement by some investigators. For example Forceville et al. stated that: "In severely ill ICU patients with SIRS we observed an early 40% decrease in plasma selenium. This could explain the three-fold increase in morbidity and mortality rates in these patients" [24]. Isn't that incredible? Wong, in Critical Care Medicine, stated that "In molecular

14 A.E. Baue

terms, MODS may occur when endothelial cells undergo disparate patterns of gene expression either simultaneously or consecutively" [25]. This is another fascinating statement.

Other major problems that we face are infections in intensive care units (leU), resistant organisms, and so on. Our intensive care units tend to be cesspools of organisms with considerable transfer of organisms from one bed to another and resistant organisms arriving with patients from nursing homes. There is great need to control the bacteriology of our intensive care units and our patients with sepsis.

Finally, the extraordinary explosion of knowledge about inflammation, mediators and critically ill patients with sepsis has greatly increased our knowledge of these processes. Therapy is improving slowly and gradually but primarily as a result of better monitoring and support of organ function.

References 1. Sanchez-Cuenca J, Martin JC, Pellicer A, Simon C (1999) Cytokine pleiotropy and redundan

cy - gp130 cytokines in human implantation. Immunology Today 20:57-59 2. Shu Q, Fang XM, Daufeldt S et al (1999) Association of tumor necrosis factor microsatellites

with incidence and outcome of severe sepsis. Crit Care Med 27:A49 3. Aziz M, Kirschenbaum LA, Astiz ME et al (1999) Neutrophil-platelet interactions in sepsis.

Crit Care Med 27:A49 4. Cursons RTM, Jeyerajah EM, Sleigh JW (1999) The use of polymerase chain reaction to de

tect septicemia in critically ill patients. Crit Care Med 27:937-940 5. Belgrader P, Benett W, Hadley D et al (1999) PCR detection of bacteria in seven minutes. Sci

ence 284:449-450 6. Wilmore DW (1998) Polymerase chain reaction surveillance of microbial DNA in critically

ill patients: exploring another new frontier. Ann Surg 227: 10-11 7. Balter M (1998) Molecular methods fire up the hunt for emerging pathogens. Science 282:

219-221 8. Kane TD, Alexander JW, Johannigman JA (1998) The detection of microbial DNA in the

blood. Ann Surg 227:1-9 9. Upperman JS, Deitch EA, Guo W et al (J 998) Post-hemorrhagic shock mesenteric lymph is

cytotoxic to endothelial cells and activates neutrophils. Shock 10:407-414 10. Moore EE (1998) Mesenteric lymph: The critical bridge between dysfunctional gut and mul

tiple organ failure. Shock 10:415-416 11. Lemaire L, van Lanschot J, Stoutenbeek CP et al (1999) Thoracic duct in patients with multi

ple organ failure: no major route of bacterial translocation. Ann Surg 229: 128-136 12. Zimmer S, Heiss M, Schardey H et al (1998) Inflammatory syndrome after endovascular aor

tic prothesis - A comparative study. Langenbecks Arch Chir I: 13-17 13. Spanier TB, Oz MC, Stern DM et al. Long term implanted left ventricular assist devices func

tion as immune-inflammatory organs. Am J Thor and Cardiovasc Surg (in press) 14. Naturally occurring anticoagulants and negative outcomes in trauma patients. Soc Crit Care

Med Symposium Highlights, 1998 15. Arbeit JM, Hirose R (1999) Murine mentors: transgenic and knockout models of surgical dis

ease. Ann Surg 229:21-40 16. Hatch T (1998) Chlamydia: old ideas crushed, new mysteries bared. Science 282:638-639

Introduction 15

17. Zehnbauer B, Freeman, Buchman T (1996) Clinical molecular genetics and critical care medicines. Crit Care Med 24(3):373-375

18. Bennett-Guerrero E, Ayuso L, Hamilton-Davies C et al (1997) Relationships of preoperative anti-endotoxic care antibodies and adverse outcomes following cardiac surgery. JAMA 277 (8):646-650

19. Angele MK, Wichmann MD, Ayala A et al (1997) Testosterone receptor blockade after hemorrhage in males, restoration of the depressed immune function and improved survival following subsequent sepsis. Arch Surg 132:1207-1214

20. Sam AD II, Sharma AC, Lee LY et al (1999) Sepsis produces depression of testosterone and steroidogenic acute regulatory (StAR) protein. Shock 11:298-301

21. Offner PJ, Moore EE, Biffl WL (1999) Male gender is a risk factor for post-injury major infections. Arch Surg (in press)

22. Eachempati SR, Hydo LJ, Barie PS (1999) Is there a gender-based difference in outcome for septic patients admitted to the surgical ICU? Arch Surg (in press)

23. Whitacre CC, Reingold SC, O'Looney PA et al (1999) A gender gap in autoimmunity. Science 283:1277-1278

24. Forceville X, Vitoux D, Remy G et al (1998) Selenium systemic immune response syndrome, sepsis and outcome in critically ill patients. Crit Care Med 26: 1536-1544

25. Wong HR (1998) Nuclear factor-Kappa B and nitric oxide regulating life and death: Nonsense or harsh reality? Crit Care Med 26: 1470-1471

Sepsis and Organ Dysfunction. The Challenge Continues

G. BERLOT, U. LUCANGELO, A. GULLO

The simultaneous poor functioning of more than one organ or system is an extremely common occurrence in patients admitted to intensive care units, and indeed is one of the main causes of death of such patients. This condition, originally indicated by the acronym MOFS (Multi Organ Failure Syndrome), has recently been renamed as MODS (Multi Organ Dysfunction Syndrome) [1]. This change of nomenclature was principally due to a) the need to express the concept of evolution, contained in the term dysfunction, which reflects the spectrum of intermediate situations existing between full function and full-blown failure and b) the lack of uniformity over a definition of organ failure, which according to the author ranges from a variation in a given parameter to the need for supportive therapy [I]. Nevertheless, beyond the semantic differences, in the light of the most recent animal and clinical studies it appears clear that a) in the vast majority of cases the changes which cause an organ to dysfunction can also cause its failure and that, b) once established the dysfunction can progress towards failure even after the removal of its cause. Thus, in the last analysis, MODS seems to be the final common pathway of a heterogeneous series of clinical events, including shock, mechanical or heat trauma, infection, sepsis, acute pancreatitis and rupture of an aortic aneurysm. In other words, independently of the cause, MODS is generally preceded by a situation associated with a period of cardiovascular instability and/or release of mediators (q. v.) [2].

Some factors, active both systemically and in a more limited environment, capable of facilitating the development of MODS can be identified. Of these former factors, there are some which are particularly important e.g. physiopathologic changes caused by the interaction between the current disease and pre-existing conditions, age over 65-years old and the presence of non-surgically related sepsis or cardiac arrest [2]. The particularly relevant factors among the second group seem to be Adult Respiratory Distress Syndrome (ARDS), which in its tum is predisposed to by sepsis, aspiration of gastric contents and pulmonary bruising [2]. Independently of its causes, recent studies have shown that for MODS, as for other disorders, there is an individual, genetically-determined susceptibility to triggering stimuli which, for an equivalent level of initial insult, could explain the varying degrees of severity of the clinical evolution in apparently similar patients [3].

18 G. Berlot, U. Lucangelo, A. Gullo

Both the number of malfunctioning organs and the duration of the dysfunction have been compared with prognosis: in one study carried out by Knaus et al. [4], the mortality rate rose from 22% to 41 % in patients in whom the dysfunction of a single organ lasted from one to seven days, while it was virtually 100% when three or more organs were malfunctioning for a period equal to or greater than seven days. It is worth remembering that in the study just cited, the criteria used to define organ failure were values markedly outside the normal range. In contrast, it is not clear whether, and if so, to what extent each single dysfunction affects prognosis [1]. There has been an apparent decrease in mortality due to septic shock [5] and ARDS [6] over the last few years: this finding could be ascribed to more aggressive management [7, 8] from the very earliest stages of the condition or to the use of support techniques (e.g. mechanical ventilation) with fewer destabilising effects on both the cardiovascular system and the lung parenchyma itself [9-12].

The role of mediators in the pathogenesis of sepsis and MODS Numerous experimental and clinical studies have shown that the considerable haemodynamic, metabolic and respiratory changes that characterise sepsis and shock, derive from the production of a heterogeneous series of endogenous substances, produced in the course of the interaction between infecting micro-organism and host, which together give rise to a generalised inflammatory reaction [13]. MODS can, therefore, be considered as a final common pathway of a series of heterogeneous harmful factors which are characterised by the activation of a complex system of endogenous mediators such as tumour necrosis factor (TNF), an ever increasing number of interleukins (IL) which according to the situation have pro- or anti-inflammatory activity, platelet activating factor (PAF), numerous derivatives of arachidonic acid, nitric oxide (NO), and free oxygen radicals (FOR). The combined action of these substances, in isolation or in association with other factors (e.g. hypoxia) causes widespread tissue damage to various areas [3], with subsequent development of MODS. In many cases the production of these substances is accompanied by that of their specific inhibitors (e.g. soluble receptors with chelating activity or agents blocking the receptor of the target cell), with the teleologic aim of limiting the inflammatory response and preventing damage to the target tissue [14]. The situation is further complicated by the presence of numerous interactions with other biological systems, e.g. the coagulation cascade, the production of nitric oxide (NO), the synthesis of acute phase proteins, and the complement system [13].

The concentration of these mediators in the biological fluids of subjects exposed has been the subject of various studies which had the aims of a) identifying markers of the systemic inflammatory process, b) providing prognostic information and c) evaluating the effects of antagonism of the mediators (q.v.). These studies have allowed several important conclusions to be drawn. First,

Sepsis and Organ Dysfunction. The Challenge Continues 19

there are marked differences between experimental and clinical conditions. In contrast to that reported for healthy subjects, in whom intravenous administration of endotoxins is followed quickly by a blood peak of TFN which also disappears quickly [15], persistent levels of this mediator are found both in the blood during septic shock [16, 17] and in tracheal secretions during ARDS [18].

Second, in septic patients, the persistence of high levels of these mediators is associated with the development of MOF and increased mortality [16, 17]. Third, there is the possibility of an iatrogenic increase in the production of the mediators of sepsis; such an increase has been found both after administration of some classes of antibiotics, with the consequent liberation of endotoxins from the microbes leading to activation of the mediator network [19], and as a consequence of high intrapulmonary pressures caused by an excessive flow volume in patients with ARDS [11].

Finally, there are differences between the results of various clinical studies which are probably linked to the heterogeneity of the patients studied or the sampling time: for example, high levels of IL-6 were found by some authors in the blood of septic patients [20, 21], but another study did not confirm this finding [22]. In effect, despite the sound biological premises, in clinical practice there are significant difficulties in measuring levels of these mediators because of the high costs of the method used and the fluctuations over time in their secretion [23]. Finally, despite the theoretical importance of their titre, it should be remembered that a) the circulatory levels of a given mediator represent only a part (similar to the point of an iceberg) of the total amount produced, and b) the biologically active fraction is that already bound to the specific cell receptors, so to a great extent escapes detection [24]. The clinical consequences of these limitations are substantial. First, the administration of an antibody against a circulating mediator could be too late if it is based on the blood concentrations of that particular substance since the time needed for blood sampling and measuring could prevent quick treatment. Second, as demonstrated by Damas et al. [25], in a group of patients with septic shock, the profile of the production and thus the appearance of mediators of sepsis in the blood could differ according to the speed of onset of haemodynamic instability. It follows that, given that both the types of mediator and their relative effect undergo temporal changes, the search for such factors should be repeated over time and in any case extended to a whole variety of substances, with extremely high costs which probably cannot be sustained outside the specific setting of research projects. At present it, therefore, seems improbable that the level of mediators responsible for sepsis and MODS can have a practical role in the management of all patients at risk.

Furthermore, it has been hypothesised that in the course of the same episode of sepsis, two different and contrasting processes can exist, one characterised by the production of inflammatory agents and related inhibitors (mixed inflammatory and anti-inflammatory response syndrome, MARS) and the other by a prevalence of mediators with marked anti-inflammatory characteristics (compensatory anti-inflammatory response syndrome, CARS) [26]. This phenome-


non could explain the apparent paradox of patients mounting a generalised inflammatory response despite presenting a clinical picture of immunodepression. This hypothesis can also be used to explain, at least in part, the negative results of various clinical trials aimed at evaluating substances intended to counteract some mediators of sepsis (e.g. anti-TNF antibodies, or cell receptor blockers, IL-1 and PAF) which, albeit in experimental conditions, have been shown to be extremely effective: in fact it has been hypothesised that if administered during the CARS phase, they would not produce any therapeutic benefit [27].

The role of systemic and regional haemodynamic alterations All the components of the cardiovascular apparatus are profoundly involved in the course of sepsis and MODS. Despite the hyperkinetic haemodynamic profile normally present in these patients, depression in cardiac activity has been demonstrated both in vitro and in vivo and has been attributed to negative inotropic effects caused by numerous mediators produced during sepsis [19]. There are both experimental and clinical findings that highlight the role of these mediators, demonstrating how their neutralisation (e.g. by administration of anti-TNF antibodies) or their elimination (e.g by methods of extracorporeal depuration) is associated with better cardiovascular function [28, 29].

The microcirculation is also heavily affected during sepsis and MODS; in these circumstances the changes in the microcirculation derive principally from a) widespread obstruction of the microvasculature by microemb01i, b) the loss of peripheral vasoregulation and c) the increase of capillary permeability [19]. The consequent tissue oedema further reduces diffusion of oxygen from capillaries to the periphery. These mechanisms have numerous interconnections and it thus difficult to separate their effects from each other. The role played by reperfusion following a sustained period of hypoxaemia plays a particularly important role in the pathogenesis of MODS. Indeed, even if a sufficiently prolonged period of ischaemia is clearly able to cause irreversible cell lesions, the main damage often occurs after reperfusion because of the formation of free oxygen radicals (FOR) and the lesions they induce [30]. The most extensively studied FOR are the superoxide radical (0-), hydrogen peroxide (H20 2) and the hydroxyl radical (OH-). In conditions of hypoxia, adenosine triphosphate (ATP) is transformed into adenosine diphosphate (ADP), adenosine monophosphate (AMP) and, in the end, into adenosine, which is converted into hypoxanthine. At the same time the enzyme xanthine dehydrogenase (XD), which is present in significant quantities in enterocytes, is proteolytic ally converted into xanthine oxidase (XO) [30]. This last enzyme transforms hypoxanthine into xanthine and the xanthine into uric acid. By-products of both of these reactions are FOR. During the reperfusion stage more oxygen is available and thus more FOR are produced because of the greater availability of substrate for XO. The cells normally produce a variety of FOR inhibitors, including glutathione peroxidase and


superoxide dismutase. Other exogenous inhibitors are superoxide dismutase and vitamin C [30]. FOR can damage the cells by various mechanisms, including hyperoxidation of the membrane, changes in DNA and cytoplasmic proteins [30]. The differing susceptibilities that different tissues have to ischaemia can be explained, at least in part, by the time necessary for XD to be transformed into XO, ranging from a few seconds in the intestine to 30 minutes in the spleen, kidneys and lungs [30].

The mediators of sepsis are able to activate endothelial cells, inducing them to participate in the generalised inflammatory reaction and in the procoagulant state of sepsis and MODS. This activation takes place by the expression of receptors (ELAM -1, ICAM -1) on the cell surface which facilitate leukocyte adherence and the secretion of other mediators, among which TNF, PAF, and IL-I [31]. Leukocyte adhesion, mediated by contact between the leukocyte receptors CDII and CDl8 and the endothelial receptors ELAM-I and ICAM-I, therefore seems to be a prerequisite both for subsequent tissue damage and the oxygenation disorders seen during the course of sepsis. Inevitably these changes affect tissue oxygenation and cell energy metabolism, which is strongly modified during sepsis and other critical pathologies. In experimental conditions of gradually reduced oxygen availability, oxygen consumption (V02) remains largely independent of oxygen transport (D02); only when the value of this latter parameter falls below a given threshold (D02crit), does the V02 begin to fall and become dependent on the D02; this threshold is considered the beginning of anaerobic metabolism and is associated with an increase in blood lactates [32]. Over the same period several authors [33-36] demonstrated that septic, trauma and postoperative patients had a better survival if their D02 and V02 were kept above normal levels. It was, therefore, hypothesised a) that the phenomenon of VOiD02 dependence indicated a condition of tissue hypoxia which was alone, or in part, responsible for the setting up and maintenance of MODS and b) that this could therefore be prevented by an increase in V02 in its tum produced by increasing the D02: theoretically this last variable should be increased until a stable V02 is reached.

However, other authors were not able to find this dependence when the V02 was measured directly rather than calculated by the Fick equation [37, 38], which gave rise to the suggestion that the VOiD02 dependence was attributable to "mathematical pairing" caused by the simultaneous rise in the same variable (cardiac output), used in the calculation of both parameters. Furthermore, more recent studies have shown that in different groups of patients treated with different therapeutic goals [39] (supramaximal values of V02 and 002 vs. a SV02 > than the control group), the mortality did not change or indeed there was a higher rate of mortality in the group of patients who were treated particularly aggressively in order to reach supramaximallevels of V02 and 002 [40]. Because of these controversies it is still not clear a) whether the supramaximal haemodynamic values should be reached always and in any case, and b) what is the best way of improving tissue oxygenation and thus preventing and correcting any hy-


poxaemic tissue stress. In the light of these last considerations, indiscriminate raising of the D02 is no longer considered a therapeutic objective per se, but is conditioned by the presence of dysoxia, that is, the signs of cell stress caused by reduced availability of oxygen at a regional level.

The gastrointestinal tract (GIT) seems particularly sensitive to hypoxia, and in experimental conditions, biochemical signs of hypoxic stress to the intestinal mucosa are manifested earlier than in other regions [41]. This phenomenon can be explained by the particular architecture of the microvasculature of the intestinal villi, which acts as a countercurrent exchanger between the arteriolar and venular extremities. The result is a reduced O2 tension at the extremity of the villus [42]. The particular angle (90°) at which the arterioles branch off from the main vessels also produces a lower haematocrit of the blood within them, thus further reducing the O2 content [43].

The splanchnic oxygen supply can be compromised by many factors, but in this respect, it is not always easy to translate experimentally obtained information into clinical practice. In effect, the response of the GIT to some cardiocirculatory disorders seems to be extremely heterogeneous, with wide interspecies variations. For example in some cardiogenic and haemorrhagic models of shock, regional blood flow to the GIT is significantly reduced, both by the activation of the angiotensin-aldosterone axis and by adrenergic stimulation [44,45]. However, in other experimental situations (burns), conflicting results have been found. In some models there was a reduction in blood flow to the GIT, while in others blood flow remained normal [46]. The results obtained from models of sepsis and septic shock are also controversial, once again probably because of interspecies differences; once again, in some experiments oxygen transport (D02)

was increased [47,48] while in others it was reduced [49]. Fewer data are available about intestinal blood flow in the critically ill patient. In one group of patients with burns, increased blood flow to the GIT was found associated, however, with increased vascular permeability and a decrease in oxygen extraction [50]. In septic, but not shocked, patients mesenteric blood flow was increased [5\]. Independently of these results, it should be noted that overall measurement of blood flow to the GIT is a (relatively) crude index of its perfusion. In one model of sepsis, the intestinal \102 decreased for values of D02 greater than those at which the overall \102 decreased [52]. Recent studies on patients with septic shock and low output states secondary to cardiac surgery showed how the regional blood flow to the GIT (measured by indocyanine green infusion) was substantially independent of systemic blood flow both at baseline conditions and after pharmacological treatment [53, 54]. Thus it seems that areas of regional hypoperfusion, probably in the GIT, can exist despite satisfactory systemic levels of D02 and \1°2 , Furthermore, the epithelial cells of the intestine seem to be a preferred target for many mediators of sepsis and FOR. Experimentally the administration of endotoxin or TNF has been associated with the appearance of areas of mucosal necrosis whose formation could be prevented by an antagonist of PAF [55]. The same lesions occurred after administrating PAF [561. In vitro,


gamma-interferon is able to cause an increase in the permeability of intestinal mucosa [57]. In addition to these direct effects, some mediators of sepsis cause a substantial decrease in intestinal D02 by a direct action on myocardium and/or vasoparesis induced in the mesenteric vessels [58].

In the light of the above, it is worth considering in detail the indications and limitations of some methods of monitoring risk of dysoxia commonly used in clinical practice. Measurement of the concentrations of blood lactates has been and remains a useful, cheap way of measuring the adequacy of D02 at a systemic level. In fact, lactic acid production increases in relation to D02crit and, in patients who demonstrate the VOiD02 dependency phenomenon, it remains raised for the entire duration of the dysoxia [32]. Besides providing immediate information about haemodynamic compensation, the repeated measurements of lactate concentrations have a better predictive capacity than TNF and IL-6 [59]. It should, however, be remembered that in conditions of sepsis, hypoxia is not the only cause of hyperlactacidaemia [60]: increased lactic acid concentrations can also be the result of an increased use of glucose in muscles and/or a block at the level of pyruvate dehydrogenase induced by endotoxin. Furthermore it has been hypothesised that there is not an absolute energy deficit in septic subjects [61] and that in consequence, most of the lactate is not produced but rather derives from an increase of Na+K+ ATPase under the stimulation of catecholamines [62]. Independently of the foregoing, it is possible that apparently normal levels of lactates can co-exist with hypoxic stress of the GIT.

The adequacy of mesenteric perfusion can be indirectly assessed by monitoring the gastric intramucosal pH (pHi) [63], which can be calculated from the Henderson-Hasselbalch formula once the data from simultaneous measurements of mucosal CO2 (measured via a modified nasogastric tube) and systemic arterial bicarbonate concentration are known. In critically ill patients this index has been shown to be useful from both a prognostic point of view and that of titrating inotropic support since it allows the inotropic dose to be calibrated against the region most vunerable to therapeutic interventions [64, 65]. It should, however, be remembered that interventions directed at increasing the overall D02 can have negative effects on intestinal V02. In fact, cardiovascular drugs can affect splanchnic V02 in a variety of ways. First, a-antagonists can cause vasoconstriction, and thus ischaemia, of the intestinal mucosa [66]. Second, inotropic agents or vasodilators can reduce mesenteric blood flow by a volume-dependent reduction in cardiac output or by a "steal" effect in favour of other vascular beds [41]. The main criticism of pHi measurements is based on the fact that the systemic arterial bicarbonate concentration may be different from its concentration in the blood which perfuses the gastric mucosa. In order to overcome this problem, some authors have proposed the use of the endogastric-systemic CO2 gradient; this strategy seems to have some advantages [67]. Besides the aspects of monitoring and the prognostic significance, the assessment of changes in pHi following the administration of various vasoactive drugs has shown that the effect of these drugs on splanchnic perfusion is largerly unpredictable and that

24 G. Serlot, U. Lucangelo, A. Gullo

some substances with a-stimulating activity such as noradrenaline can in reality improve splanchnic perfusion [68].

Therapeutic possibilities: mediator antagonism The mediators of sepsis can be antagonised in various ways [69]. First, both the endotoxin and the substances produced in response to interactions between them and the immunocompetent cells can be blocked by specific antibodies. Second, receptors present on the cell surface can be made unavailable for interaction with the mediators by specific antagonists without activating capacity. Third, circulating soluble receptors, identical to those on the cell surface, can be used in order to form complexes with the specific mediator, thus preventing the mediator's adhesion to the target cells.

Essentially these strategies mimic more or less those put into action by the organism in order to limit the spread of an inflammatory response, however it has been generated. Nevertheless, despite the firm biological and physiological bases, almost all the controlled, randomised, double-blind studies of the use of anti-mediator substances so far published have demonstrated only partial and limited effects in some subgroups of patients [2].

Anti-endotoxin antibodies were the first substances of this class to be used in clinical trials. At the beginning of the '80s Ziegler et al. demonstrated that the administration of antisera with a high titre of polyclonal antibodies against lipid A of endotoxin (15) was associated with a better survival [70]. This initial result was subsequently confirmed by two other studies that used polyclonal anti-endotoxin antibodies obtained from pooled sera [71, 72]. Despite these favourable results, the administration of polyclonal sera was associated with a series of potentially dangerous limitations, such as the instability of the solution to be injected, the difficulty in titering antibody activity precisely, and the risk of disease transmission associated with the pooled sera from which the antibodies were obtained [73]. These problems were overcome thanks to the development of genetic engineering, and, ten years later, the administration of monoclonal anti-endotoxin antibodies was associated with better survival in two groups of patients enrolled in the same number of trials. The first study was of 486 patients with suspected gram-negative sepsis who were treated with E5, an antilipidA IgM, or a placebo [74]. Although mortality in the treated and placebo groups was substantially the same, the mortality rate was significantly lower in treated patients who had a documented gram-negative sepsis. At the same time another study was undertaken using anti-endotoxin monoclonal IgM (HA-I A) given as a single intravenous dose of 100 mg [75]. The results showed that the survival rate was better in the treated group, and especially in patients with high blood levels of endotoxin [73]. However, a confirmatory study with E5 evidenced the lack of effects on survival although the administration of E5 was associated with significant improvement in organ dysfunction [76]. Doubts were


also raised about the trial used to study HA-IA [77]. The perplexities were particularly concentrated on the randomisation of the patients and the effect of concomitant treatments: in brief, the patients in the placebo group were older, had a higher APACHE II score and their antibiotic treatment was in many cases inadequate. In order to lay to rest these controversies a definitive confirmatory study of the efficacy of HA-IA was undertaken, but the intermediate analysis of the data demonstrated a slight increase in the mortality of a subgroup of treated patients (those who were not bacteraemic), so the study was stopped, and HA-IA was withdrawn from clinical experimentation causing, apart from anything else, irreparable damage to the drug company that produced it [73].

A series of studies were directed towards evaluating the effects of anti-TNF antibodies. Experimentally these substances had a considerable protective effect on the cardiovascular apparatus, mitigating and reducing the entity of septic shock secondary to endotoxin [78]. Initial clinical studies on patients with septic shock showed that there was a transitory increase in arterial blood pressure and other haemodynamic parameters [79, 80]. A preliminary study on anti-TNF antibodies carried out on a small number of patients demonstrated that this substance had no harmful side-effects and was associated with a substantial decrease in the level of circulating TNF [81]. Unfortunately, two larger, controlled, randomised, double-blind studies failed to demonstrate any improvement in the treated patients [82, 83]. This having been said, one of the studies did show an effect on survival in the subgroup of patients with high blood levels of IL-6 [83]. A confirmatory study was, therefore, started, in which only those patients with an IL-6 > 1000 pg/ml were randomised for treatment. This study, too, was suspended when an intermediate analysis of the results revealed an excess of mortality in the group treated with the active agent compared to the mortality in the group given a placebo.

Another anti-mediator strategy is that of preventing the mediator interacting with receptors on the target cell. This objective can be reached by a) blocking cell receptors specific for a given substance, or b) intercepting the molecules of the mediator itself before they reach the cell. The strategy of cell receptor blockade has been studied by using an antagonist to the receptors for IL-I (IL-lra), obtained by genetic engineering techniques. Under natural conditions this substance is produced at the same time as IL-l [84] and was isolated from the serum of healthy volunteers treated with endotoxin [85]. IL-lra has been studied in various clinical trials. In the first, there was a dose-dependent reduction in the mortality rate, which was 44% in the placebo group compared to 32%, 25% and 16% in the groups treated with increasing doses of the active principle [86]. On the basis of these results a larger, randomised, double-blind study was started but this second study failed to confirm the results of the first [87]. Subsequent retrospective analyses of subgroups showed a possible effect on patients with dysfunction of two or more organs and/or a high risk of death (> 24% calculated according to APACHE II criteria) [88]. This led to a definitive study being carried out, aimed at demonstrating the possible efficacy of the molecule under ex-


amination at least in those subgroups of patients. Unexpectedly IL-lra was demonstrated to be ineffective even in this last trial [89].

Other, conceptually similar studies concentrated on the use of soluble receptors of TNF. The effects of this cytokine are, in fact, mediated by two classes of receptors of different molecular weights (TNFR I and TNFR II with molecular weights of 55 kD and 75 kD respectively) on the target cells' surface [90]. The extracellular component of these receptors is released by the cells themselves during septic states and these molecules bind to the TNF thus preventing it from binding with the receptor on the cell membrane [90]. Under experimental conditions, the administration of these soluble receptors was associated with an attenuation of the increased lung capillary permeability and neutrophil capture induced by damage from intestinal ischaemia-reperfusion [91]. However, as hypothesised by Van Zee et al. [92], the clinical usefulness of soluble TNF receptors could be limited by their short half-life (in the order of a few minutes) and by the fact that the active molecules of TNF circulate in a trimeric form and at least two of its components need to be blocked before biological inactivation can occur. In order to overcome this drawback, recombinant molecules of rTNF-a were bound to molecules of IgG (TNF:Fc) [93]. The clinical effects of administration of these substances were studied in two clinical trials. In the first, the TNFR:Fc was given to patients with septic shock using three different dose regimens and placebo [94]. Although a dose-effect relationship with respect to survival was noted, this was not statistically significant between the groups: furthermore, there was an excess of mortality in the patients with grampositive infections given the highest dose of the receptor. Another study using TNFR:Fc was therefore undertaken, again comparing the effect of three different doses against placebo [95]. The study enrolled 486 patients with a diagnosis of severe sepsis and septic shock. At the intermediate analysis an excess of mortality was found in the group treated with the lowest dose and so this arm of the study was closed. Overall, in the remaining groups there was a non-significant trend to a reduction in mortality rate at day 28. Subsequent analysis revealed a significant reduction in mortality of patients who had high levels of IL-6 and who had received the highest dose of TNFR:Fc [95]. These inconclusive results can, to some extent, be interpreted in the light of data produced by Goldie et al. [96], who measured the circulating levels of mediators and their natural antagonists (including IL-lra, TNFR I and II) in a group of septic patients. In their study the initial serum levels of soluble receptors of TNF but not IL-l were higher in patients who died, and markedly higher in the patients who went on to develop septic shock. Furthermore, the concentrations of the various antagonists studied were from 30 to 100,000 times higher than the corresponding cytokine. Taken together, these results seem to indicate that a) in situations associated with a systemic inflammatory response, cytokine antagonists are produced in enormous quantities, such as to overwhelm completely the number of target molecules and that b) the concentration of mediators is higher in patients who die and/or in those destined to develop severe complications such as MOF.


In consequence it is unlikely that the administration of further synthetic antagonists could confer any sort of protective effect. Despite these results, research is being directed towards the effects of administrating PAF antagonists [97]. Different antagonists of PAF have been identified and/or synthesised, and some of these have been used in both clinical and experimental contexts. In rats treated with endotoxin, the administration of a specific antagonist for PAF reduced intestinal mucosal lesions [98]. In a large, randomised study administration of the antagonist BN 52021 caused a non-significant increase in survival compared to placebo; this increase became statistically significant in patients with Gramnegative infections, with or without shock [99]. A more recent study did not, however, confirm these encouraging results, and the adminstration of BN 52021 was not shown to have a significant advantage in terms of organ dysfunction or mortality [100].

From the foregoing it would seem that despite the solid biological and physiopathological bases, the results of the many clinical trials in which a variety of substances with anti-mediator activity have been tested, have all (or almost all) been disappointing, or at best, much less encouraging than might reasonably have been expected.

We do, however, believe that it is worthwhile examining the results in more detail, bearing in mind that, as for the very recent results on the use of corticosteroids in sepsis, further analysis can lead to different conclusions [101]. It is possible to identify some factors which could have negatively influenced the results. First there is a problem of a time-window beyond which there is no sense in using anti-mediator treatments. Unfortunately, as laid out above, at present there are no analyses which allow this window to be identified early, and the majority of the clinical indicators of the generalised inflammatory response associated with sepsis, such as tachycardia, hypotension, and leukocytosis, are non-specific and in any case not linked to the action of a particular mediator. These limitations are compounded by the difficulty in titering because of high cost, the titration method, time required to carry out the research and the irregular production and the notable fluctuations in their release [23]. Second, the postulated existence of different and opposite phases of immune behaviour would make it inappropriate to administer, for example, an anti-inflammatory substance (e.g. anti-TNF antibodies) in the CARS phase, characterised by global depression of the inflammatory response. Finally, given that the various mediators of sepsis form a network characterised by many positive and negative feedbacks, blocking only one of these is probably unlikely to have much effect on the overall response. In these situations, analogous to that which occurs in leukocytosis, it would be logical to use a cocktail of different substances rather than the administration of a single agent.

An alternative strategy for trying to antagonise the action of mediators of sepsis is that of removing them from the host by methods commonly used for the treatment of critically ill patients with acute renal failure [102]. There are both experimental and clinical results which indicate that these methods may


help to eliminate mediators of sepsis and thus lighten the severity of the physiopathological changes induced by these and/or even improve the prognosis of the patients treated.

The main mechanisms put forward to explain these results are elimination from the organism and/or adsorption on the fibers used in the treatment [103]. Recently in order to obtain a faster and more efficient removal of the mediators, plasmapheresis (PE) either alone or in association with other techniques of depuration, has been used. From studying the clinical evolution of four groups of septic patients with MODS, Barzilay et al. [l 04] observed a significant decrease in mortality in patients treated with PE associated with other methods, although the authors attribute the improvement in survival in the group treated with PE to an increased elimination of mediators. More recently, PE carried out in a group of septic patients caused significant improvement in all tested haemodynamic parameters, but not a parallel improvement in mortality rate; this result was attributed to the elimination of mediators and substances with myocardiodepressant activity, which did not however affect the prognosis [105]. Thus, at present extracorporeal depuration seems promising, despite it is not yet being possible to identify with precision the groups of patients who would gain most benefit.

Conclusions

Sepsis and the generalised inflammatory states associated with it involve all the systems of the host and, profoundly disturbing the compensatory systems, lead to a substantial change in homeostasis. In effect, although notable progress has been made in the knowledge about individual mediators and/or the disorders these provoke, the interactions and feed-back mechanisms which occur during the pathogenesis of sepsis and the generalised inflammatory states are yet to be clarified.

References 1. Members of the American College of Chest Physician/Society of Critical Care Medicine

Consensus Committee (1992) Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 20:864-890

2. Bernard GR (1995) Sepsis trials. Am J Resp Crit Care Med 102:4-10 3. Kumar A, Short J, Parrillo JE (1999) Genetic factors in septic shock. JAMA 292:579-581 4. Knaus WA, Draper EA, Wagner DPO, Zimmermann JE (1985) Prognosis in acute organ sys

tem failure. Ann Surg 202:685-693 5. Friedman G, Silva E, Vincent JL (1998) Has the mortality of septic shock changed with time'!

Crit Care Med 26:2078-2086 6. Jardin F, Fellahi JL, Beauchet A et al (1999) Improved prognosis of acutc respiratory distress

syndrome 15 years on. Intcnsivc Carc Med 25 :936-941


7. Lewandowsky K (1999) Epidemiological data challenge ARDS/ALI definition. Intensive Care Med 25:884-886

8. Steltzer H, Krafft P (1999) Improved outcome of ARDS patients. Are we really performing better? Intensive Care Med 25:887-889

9. Stewart TE, Meade OM, Cook DJ, Granton JT et al (1998) Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med 338:355-361

10. Brochard L and the Multicenter Trial Group on tidal volume reduction in ARDS (1998) Tidal volume reduction for the prevention of ventilator-induced lung injury in ARDS. Am Rev Respir Crit Care Med 158:1831-1838

II. Amato M, Barbas CSV, Medeiros DM et al (1998) Effects of a protective ventilatory strategy on mortality in ARDS. N Engl J Med 338:347-354

12. Ranieri VM, Suter PM, Tortorella C et al (1999) Effect of mechanical ventilation on inflammatory mediators in patients with acute respiratory distress syndrome. A randomized controlled study. JAMA 282:54-61

13. Davies MG, Hagen PO (1997) Systemic inflammatory response syndrome. Br J Surg 84: 920-935

14. Moldawer LL (1994) Biology of proinflammatory cytokines and their antagonist. Crit Care Med 22:S3-S7

15. Michie HR, Manogue KR, Springs DR et al (1988) Detection of circulating tumour necrosis factor after endotoxin administration. N Engl J Med 318;23: 1481-1486

16. Pinsky MR, Vincent JL, Deviere J et al (1993) Serum cytokine levels in human septic shock: relation to multiple system organ failure and mortality. Chest 103:565-55

17. Calandra T, Baumgartner JD, Grau GE et al (1990) Prognostic values of tumour necrosis factor/cachectin, interleukin-I, interferon-alpha, and interferon gamma in the serum of patients with septic shock. J Infect Dis 161 :982-987

18. Millar AB, Singer M, Meager et al (1989) Tumour necrosis factor in bronchopulmonary secretions of patients with adult respiratory distress syndrome. Lancet 23 :712-714

19. Berlot G, Vincent JL (1992) Cardiovascular effects of cytokines. Clin Intens Care 3:199-205 20. Saladino R, Erikson M, Levy N et al (1992) Utility of serum interleukin-6 for diagnosis of in

vasive bacterial disease in children. Ann Emerg Med 21: 1413-1417 21. Fassbender K, Pargger K, Muller W, Zimmerli W (1993) Interleukin-6 and acute phase pro

tein concentrations in surgical intensive care patients: diagnostic signs in nosocomial infections. Crit Care Med 21: 1175-1180

22. Calandra T, Gerain J, Heumann D et al (1991) High circulating levels of interleukin 6 in patients with septic shock: evolution during sepsis, prognostic value and interplay with other cytokines. Am J Med 91:23-29

23. Lefer A (1989) Significance of lipid mediators in shock states. Circ Shock 27:3-12 24. Cavaillon JM, Munoz C, Fitting C et al (1992) Circulating cytokines: the tip of the iceberg?

Circ Shock 38: 145-152 25. Damas P, Canivet JL, De Groote D et al (1997) Sepsis and serum cytokine concentrations.

Crit Care Med 25:405-412 26. Bone RC (1996) Sir Isaac Newton, Sepsis, SIRS and CARS. Crit Care Med 24: 1125-1136 27. Cain BS, Meldrun DR, Harken AH, McIntyre RC (1998) The physiologic basis for anticy

tokine clinical trials in the treatment of sepsis. J Am Coil Surgeons 186:337-350 28. Vincent JL, Bakker J, Marecaux G et al (1992) Administration of anti TNF antibodies im

proves left ventricular function in septic shock patients: results of a pilot study. Chest 101: 810-815

29. Gomez A, Wang R, Unruh H et al (1990) Hemofiltration reverses left ventricular dysfunction during sepsis in dogs. Anesthesiology 73:671-785

30. McCord JM (1985) Oxygen derived free radicals in post-ischemic tissue injury. N Engl J Med 312:159-163

31. Cotran RS (1990) Cytokine and endothelial cell biology. Physiol Rev 70:427-451


32. Bakker J, Vincent JL (1991) The oxygen supply dependency phenomenon is associated with increased blood lactate levels. J Crit Care 6: 152-159

33. Shoemaker WC, Appel PL, Kram HB (1988) Prospective trial of supranormal values of survivors as therapeutic goals in high risk surgical patients. Chest 94: 1176-1182

34. Moore FA, Haenel JB, Moore EE, Whitehill TA (1992) Incommensurate oxygen consumption in response to maximal oxygen availability predicts postinjury multiple organ failure. J Trauma 33:58-65

35. Fleming A, Bishop M, Appel P et al (1992) Prospective trial of supranormal values as goals of resuscitation in severe trauma. Arch Surg 127: 1175-1181

36. Tuchshmidt J, Fried J, Astiz M, Rackow E (1992) Elevation of cardiac output and oxygen delivery improves outcome in septic shock. Chest 102:216-220

37. Bartlett RH, Dechert RE (1990) Oxygen kinetics: pitfalls in clinical research. J Crit Care 5:77-80

38. Ronco JJ, Fenwick JC, Wiggs BR et al (1993) Oxygen consumption is independent of increases in oxygen delivery by dobutamine in septic patients who have normal or increased plasma lactate. Am Rev Resp Dis 147:25-31

39. Gattinoni L, Brazzi L, Pelosi P et al (1995) A trial of goal oriented hemodynamic therapy in critically ill patients. N Engl J Med 333: 1025-1032

40. Hayes MA, Timmins AC, Yau EHS et al (1994) Elevations of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med 330: 1717 -1722

41. Fink MP (1991) Gastrointestinal mucosal injury in experimental models of shock, trauma and sepsis. Crit Care Med 19:627-641

42. Jacobson LF, Noer RF (1952) The vascular pattern of the intestinal villi in various laboratory animals and in man. Anat Rev 114:85-90

43. Jodal M, Lundgren M (1970) Plasma skimming in the intestinal tract. Acta Physiol Scand 80:50-55

44. Porter JM, Sussman MS, Bulkley GB (1989) Splanchnic vasoconstriction in circulatory shock. In: Marston A, Bulkley GB, Fiddian-Green RG et al (eds) Splanchnic ischemia and multi-organ failure. Arnold, London, pp 73-88

45. Adar R, Frankin A, Spark RF et al (1976) Effect of dehydration and cardiac tamponade on SMA flow: role of vasoactive substances. Surgery 79:534-538

46. Jones WG II, Minei JP, Barber AE et al (1990) Bacterial translocation and intestinal atrophy after thermal injury and bum wound sepsis. Ann Surg 211 :399-403

47. Lang CH, Bagby GJ, Ferguson JL et al (1984) Cardiac output and redistribution of organ blood flow in hypermetabolic sepsis. Am J PhysioI246:R331-334

48. Nxumalo JL, Teranaka M, Shenk WG (1978) Hepatic blood flow measurement III. Hepatic blood flow measured by ICG clearance and electromagnetic flow meters in a canine septic shock model. Ann Surg 187:299-304

49. Whithwoth PW, Cryer HM, Garrison RN et al (1989) Hypoperfusion of the intestinal microcirculation without decreased cardiac output during live Escherichia coli sepsis in rats. Circ Shock 27:111-115

50. Aulick LH, Goodwin CW, Becker RA et al (1981) Visceral blood flow following injury. Ann Surg 193:112-116

51. Dahn MS, Lange P, Lobdel K et al (1987) Splanchnic and total body oxygen consumption differences in septic and injured patients. Surgery 100:69-74

52. Nelson DP, Samsel RW, Wood LDH et al (1988) Pathological supply dependence of systemic and intestinal oxygen uptake during endotoxemia. J Appl Physiol 64:2410-2418

53. Ruokonen E, Takala J, Kari A et al (1993) Regional blood flow and oxygen transport in septic shock. Crit Care Med 21:1296-1303

54. Ruokonen E, Takala J, Kari A (1993) Regional blood flow and oxygen transport in patients with the low cardiac output syndrome after cardiac surgery. Crit Care Med 21: 1304-1311


55. Hsueh W, Gonzalez-Crussi F, Arroyave JL et al (1986) Platelet activating factor induced ischemic bowel necrosis. The role of platelet activating factor antagonists. Am J Pharmacol 123:79-85

56. Gonzalez-Crussi F, Hsueh W (1983) Experimental models of ischemic bowel necrosis. The role of platelet activating factor and endotoxin. Am J PathoII12:127-132

57. Madara JL, Stafford J (1989) Interferon-gamma directly affects barrier function of cultured intestinal epithelial monolayers. J Clin Invest 83:724-728

58. Rackow EC, Astiz ME (1991) Pathophysiology and treatment of septic shock. JAMA 266:548-554

59. Marecaux G, Pinsky MR, Dupont E et al (1996) Blood lactate levels are better prognostic indicators than TNF and IL-6 levels in patients with septic shock. Intensive Care Med 22: 404-408

60. Gutierrez G, Wulf ME (1996) Lactic acidosis in sepsis: a commentary. Intensive Care Med 22:6-16

61. Hotchkiss RS, Karl IE (1992) Reevaluation of the role of cellular hypoxia and bioenergetic failure in sepsis. JAMA 267:1503-1510

62. James Howard J, McCarter FD, Fischer JF (1999) Lactate is an unreliable indicator of tissue hypoxia in injury or sepsis. Lancet 354:505-508

63. Fiddian-Green RG (1989) Studies in splanchnic ischemia and multiple organ failure. In: Marston A, Bulkley GB, Fiddian-Green RG et al (eds) Splanchnic ischemia and multi-organ failure. Arnold, London, pp 349-363

64. Doglio GR, Pusajo JF, Egurrola MA et al (1991) Gastric mucosal pH as a prognostic index of mortality in critically ill patients.Crit Care Med 19:1037-1040

65. Gutierrez G, Palizas F, Doglio G et al (1992) Gastric intramucosal pH as a therapeutic index oftissue oxygenation in critically ill patients. Lancet 339:195-199

66. Greenway CV, Stark RD (1971) Hepatic vascular bed. Physiol Rev 51:23-28 67. Friedman G, Berlot G, Kahn RJ, Vincent JL (1994) Combination of blood lactate levels and

pHi in severe sepsis. Crit Care Med 22;1 :112 68. Silva E, DeBacker D, Creteur J, Vincent JL (1998) Effects of vasoactive drugs on gastric in

tramucosal pH. Crit Care Med 26: 1749-1758 69. Christman JW, Holden EP, Blackwell TS (1995) Strategies for blocking the systemic effects

of cytokines in the sepsis syndrome. Crit Care Med 23:955-963 70. Ziegler EJ, McCuchan JA, Fierer J et al (1982) Treatment of gram-bacteremia and shock with

human antiserum to a mutant Escherichia coli. N Engl J Med 307:1225-1230 71. Lachman E, Pitsoe SB, Gaffin SL (1984) Antilipolysaccharide immunotherapy in the man

agement of septic shock of obstetric and gynaecological origin. Lancet 1 :981-983 72. Fomsgaard A, Baek L, Fomsgaard JS et al (1988) Preliminary study in treatment of septic

shock patients with antilipopolysaccharide IgG from blood donors. Scand J Infect Dis 21: 697-708

73. Talan DA (1993) Recent developments in our understanding of sepsis: evaluation of antiendotoxin antibodies and biological response modifiers. Ann Emer Med 22: 1871-1990

74. Greenman RL, Schein RMH, Martin MA et al (1991) A controlled clinical trial ofE5 murine monoclonal IgM antibody to endotoxin in the treatment of Gram-sepsis. JAMA 266:1097-1102

75. Ziegler EJ, McCutchan JA, Fierer J et al (1991) Treatment of Gram-bacteremia and septic shock with HA-IA human monoclonal antibody against endotoxin. A randomized, double blind, placebo controlled trial. N Engl J Med 324:429-436

76. Bone RC, Balk RA, Fein AM et al (1995) A second large controlled clinical study of E5, a monoclonal antibody to endotoxin: results of a prospective, multicenter, randomized, controlled study. Crit Care Med 1995; 23:994-1006

77. Cunnion RE (1992) Clinical trials of immunotherapy for sepsis. Crit Care Med 20:721-723 78. Silva AT, Bayston KF, Cohen J (1990) Prophylactic and therapeutic effects of a monoclonal

antibody to tumour necrosis factor-alfa in experimental Gram-shock. J Inf Dis 162:421-427


79. Exley AR, Cohen J, Buunnan WA et al (1990) Monoclonal antibody to TNF in severe septic shock. Lancet 335:1275-1277

80. Vincent JL, Bakker J, Marecaux G et al (1992) Administration of anti TNF antibodies improves left ventricular function in septic shock patients: results of a pilot study. Chest 101:810-815

81. Dhainaut JFA, Vincent JL, Richard C et al (1995) DP 571, a humanized antibody to tumor necrosis factor-alpha: safety, pharmacokinetics, immune response, and influence of the antibody on cytokine concentrations in patients with septic shock. Crit Care Med 23:1461-1469

82. Cohen J, Carlet J for the INTERSEPT group (1996) INTERSEPT: an international, multicenter, placebo-controlled trial of monoclonal antibody to human tumor necrosis factor-a in patients with sepsis. Crit Care Med 24: 1431-1440

83. Rheinhart K, Wiegand-Lohnert C, Grimminger F et al (1996) Assessment of the safety and efficacy of the monoclonal anti-tumour necrosis factor antibody fragment, MAK 195F, in patients with sepsis and septic shock: a multicenter, randomized, placebo-controlled, dose ranging study. Crit Care Med 24:733-742

84. Arend WP (1991) Interleukin 1 receptor antagonist: a new member of interleukin 1 family. J Clin Invest 88: 1445-1451

85. Granowitz EV, Santos AA, Pouts aka DD et al (1991) Production of interleukin-I receptor antagonist during experimental endotoxemia. Lancet 1338: 1423-1424

86. Fisher CJ, Siotman GJ, Opal SM et al (1994) Initial evaluation of human recombinant interleukin 1 receptor antagonist in the treatment of sepsis syndrome: a randomized, open label, placebo-controlled multicentre trial. Crit Care Med 22: 12-21

87. Fisher CJ, Dhainaut JF, Opal SM et al (1994) Recombinant human interleukin 1 receptor antagonist in the treatment of patients with sepsis syndrome. Results from a randomized, double blind, placebo-controlled trial. JAMA 271: 1836-1843

88. Knaus WA, Harrell FE, LeBreque JF et al (1996) Use of predicted risk of mortality to evaluate the efficacy of anticytokine therapy in sepsis. Crit Care Med 24:46-56

89. Opal SM, Fisher CJ, Dhainaut JFA et al (1997) Confinnatory interleukin-I receptor antagonist trial in severe sepsis: a phase III, randomized, double blind, placebo-controlled, multicenter trial. Crit Care Med 25:1115-1124

90. Bazzoni F, Beutler B (1996) Seminars in Medicine at the Beth Israel Hospital, Boston: the tumour necrosis factor ligand and receptor families. N Engl J Med 34: 1717 -1725

91. Sorkine P, Setton A, Halpern P et al (1995) Soluble tumour necrosis factor receptors reduce bowel ischemia-induced lung permeability and neutrophil sequestration. Crit Care Med 23:1377-1381

92. Van Zee KJ, Kohno T, Fisher E et al (1992) Tumour necrosis factor soluble receptors circulate during experimental and clinical inflammation and can protect against excessive tumour necrosis factor alpha in vitro and in vivo. Proc Natl Acad Sci USA 89:4845-4849

93. Mohler KM, Torrance DS, Smith CA et al (19939 Soluble tumour necrosis factor (TNF) receptors are effective therapeutic agents in lethal endotoxemia and function simultaneously as both TNF carriers and TNF antagonists. J Immunol151: 1548-1561

94. Fisher CJ, Agosti JA, Opal SM et al (1996) Treatment of septic shock with the tumor necrosis factor receptor:Fc fusion protein. N Engl J Med 334: 1697 -1702

95. Abraham E, Glauser MP, Butler T et al (1997) p55 tumor necrosis factor receptor fusion protein in the treatment of patients with severe sepsis and septic shock. A randomized controlled multicenter trial. JAMA 277:1531-1538

96. Goldie AS, Fearon KC, Ross JA et al (1995) Natural cytokine antagonists and endogenous antiendotoxin core antibodies in sepsis syndrome. JAMA 274: 172-177

97. Bone RC (1992) Phospholipids and their inhibitors: a critical evaluation of their role in the treatment of sepsis. Crit Care Med 20:884-890

98. Sun X, Hsueh W (1992) Bowel necrosis induced by tumour necrosis factor in rats is mediated by platelet activating factor. J Clin Invest 81: 1328-1331


99. Dhainaut JFA, Tenaillon A, Le Tulzo Yet al (1994) Platelet-activating factor antagonist BN 52021 in the treatment of severe sepsis: a randomized, double-blind, placebo-controlled, multicenter clinical trial. Crit Care Med 22: 1720-1728

100. Dhainaut JFA, Tenaillon A, Hemmer Met al (1998) Confirmatory platelet-activating factor receptor antagonist trial in patients with severe Gram-negative bacterial sepsis: A phase III, randomized double-blind, placebo-controlled, multicenter trial. Crit Care Med 26: 1963-1971

101. Bollaert PE, Charpentier C, Levy B et al (1998) Reversal of late septic shock with supraphysiologic doses of hydrocortisone. Crit Care Med 26:645-650

102. Ronco C, Bellomo R (1999) Continuous renal replacement therapy in the intensive care unit. Intensive Care Med 781-789

103. De Vriese AS, Vanholder RC, Pascual M et al (1999) Can inflammatory cytokines be removed efficiently by continuous renal replacement therapies? Intensive Care Med 25: 903-910

104. Barzilay E, Kessler D, Berlot G et al (1989) The use of extracorporeal supportive techniques as additional treatment for sepsis-induced MOF patients. Crit Care Med 17:634-637

105. Berlot G, Gullo A, Fasiolo S et al (1997) Hemodynamic effects of plasma exchange in septic patients: preliminary report. Blood Purification 15:45-53

I OXYGEN TRANSPORT IN SEPSIS I

Biochemical Regulation of the Microcirculation

G.P. NOVELLI

Microcirculation is the collective name for the smallest peripheral section of the circulatory system; it comprehends arterioles, capillaries and venules, each with its own structural and functional characteristics.

The main role of microcirculation is to mantain the physiological matching of oxygen supply to the metabolic demands of cells and tissues [1, 2].

The concept that oxygen exchanges only take place in true capillaries is well established but the mechanisms of regulation are debated [3]. In fact oxygen is exchanged between the intravascular and extravascular compartments without any significant restriction due to its high solubility and diffusibility: as a consequence tissue oxygenation strongly depends on mechanisms which regulate the blood flow through the microvessels and microcirculation.

To understand the function of the microcirculation in the maintenance of tissue oxygenation, the cylinder of Krogh must be taken in account. Krogh's cylinder is the volume of tissue reached by oxygen diffusing from a single capillary (assumed to be straight). When all the available capillaries are open each cylinder has a small diameter and the oxygen pressure in the dependent cells is high and uniform. On the other hand, when the number of functioning capillaries decreases the residual ones assume a compensatory function: therefore the diameter of the cylinders largely increases but also the gradient of oxygen pressure within the center and the periphery increases. The consequence is that oxygen concentration in the cells becomes very low and O2 consumption becomes "flow dependent". The causes of microcirculatory dysfunction during sepsis include mechanical obstruction (by adherent granulocytes, stiff erythrocytes, cell debris, etc.) or edema (of the endothelium or of the interstitium).

Therefore, the prognostic value of increasing 02 delivery and consumption is obvious: it means that the microcirculatory network is able to increase its exchange area and to supply adequate oxygen to the cells.

The alterations of microcirculation induced by endotoxin in animals were long since described by Zweifach [4] and by Thomas [5] who described reduced vasomotion, stagnation and increased capillary permeability resulting in interstitial edema. The impairment of microcirculation is worsened by intravascular coagulation (DIe) and further reduction in microvascular perfusion [6]. In addi-

38 G.P. Novelli

tion, red blood cells become less deformable [7] so creating another obstacle to blood flowing due to accumulation of red cells behind the white cells adhering to the endothelium: the therapeutic goal is the recruitment of functioning capillaries.

The poor function of microcirculation in septic patients was demonstrated in septic patients by absent or highly reduced post-ischemic reactive hyperemia [6].

Mediators of microcirculatory function The regulation of microcirculation is quite complex and in healthy conditions involves nerves, catecholamines, acetylcholine, serotonin, histamine, and kinins all together modulating constriction, dilatation and permeability 18].

However, many other mediators are active on the microcirculation, mostly as affected by endotoxin.

Neuropeptide Y is a 36-aminoacid adrenal peptide that has been reported to restore the vascular responsiveness to catecholamines in endotoxemic rats [9].

Restoration by cyclooxygenase inhibitors of the responsiveness to catecholamines has been demonstrated in many animal models of sepsis. The simplest explanation would be that the endotoxin-induced impairment of vascular responsiveness is consequent to some metabolite of arachidonic acid. However, prostanoids "per se" blunt the response to catecholamines, so that the conclusion is to attribute to prostaglandin some actions of modulation of adrenoreceptors or of an indirect effect of some intermediates like oxygen radicals [10].

The transcription nuclear factor NF-kB is a ubiquitous DNA binding protein necessary for directing high level transcription of many pro-inflammatory genes. The NF-kB is activated (in absence of the inhibitory factor I-kBs) by endotoxin, by TNFs, by IL-6, by hemorrhage, by ischemia-reperfusion, by oxygen radicals and by barotrauma [1 1].

In sepsis and septic shock the NF-kB locally regulates the production of cytokines, chemokines, adhesion proteins, NOSs and COX-2.

A relationship between high levels of the nuclear factor and poor prognosis in men has been suggested. Inhibition of the nuclear factor has been obtained with anti-oxidants, N-acetylcysteine, corticosteroids or dugmentation of the activity of the inhibitor factor I-kB [12].

Recently experiments have been performed on endotoxaemic rats treated with dithiocarbamate, a substance that inhibits I-kB degradation, so preventing the NF-kB activation [13]. The microvascular damage due to endotoxin was largely reduced, as were neutrophil transmigration and increased permeability.

The granulocytes play a key role in the regulation of the microcirculation through rheological and biochemical mechanisms. Granulocytes activated by

Biochemical Regulation of the Microcirculation 39

hypotension, endotoxin, fibrin fragments, complement, etc. stop rolling along the surface of endothelium, start to form pseudopods, become stiffer and adhere to capillaries and post-capillary venules, so also obstructing the flow of abnormallly rigid erythrocytes. Inhibition of granulocyte activation with a Cl-esterase inhibitor prevents plasma extravasation [14].

The activated granulocytes produce lysosomal enzymes and mostly oxygen radicals ("respiratory burst") that are associated with lipid peroxidation, increased capillary permeability and interstitial edema; they also promote further adherence of granulocytes [15]. At the same time there is the release of significant amounts of TNF-a, IL-l, and IL-8 [16] whose specific actions on the microcirculation are known but quite indefinite. TNF-a causes persistant vasodilation [17] but at the same time it favors adherence of platelets [18] and granulocytes and microendothelial damage.

The adhesion of granulocytes to the endothelial lining is dependent on the interaction between adhesion molecules on the cell surface (CD II-CD 18) and molecules (ICAMl, E-selectin, P-selectin) on the endothelial surface [19].

Nitric oxide inhibition promotes granulocyte adhesion in the mesenteric microvessels [20] and oxygen radicals promote expression of adhesion molecules.

This scenario is confirmed by experiments in which antibodies against antiadhesion molecules prevented the increase in microcirculatory permeability after ischemia [21, 22] as did an anti-neutrophil serum [23].

The endothelium has an important role in the regulation of microcirculation, at least unless deeply damaged by granulocytes and their toxic products.

The main active product of endothelium is nitric oxide (previously identified as EDRF) but an EDHF (endothelium derived hyperpolarizing factor) and endothelins must also be mentioned. The first seems to produce relaxation of the underlying vascular muscles by a mechanism of hyperpolarization [24]. The endothelins provoke contraction of blood vessels under stimulation of thrombin, hemoglobin and TNF [25]. The role of EDHF and endothelins in the regulation of microcirculation has not been clarified except as an indirect one.

The synthesis of nitric oxide by vascular endothelium is responsible for the vasodilator tone that is essential for the regulation of blood flow [26].

NO is synthetized from L-arginine by NO-synthases (NOSs). The constitutive NOS (c-NOS) produces NO in small quantities and within seconds. The inducible NOS (i-NOS) produces large amounts of NO in a sustained fashion after stimulation with endotoxin, TNF, ILs and interferon. The synthesis of NO is inhibited by a series of analogues of arginine (L-NAME, L-NMMA, L-NNA) that lack specificity for one of the NOSs. Studies on the direct effects of NO on the normal microcirculation seem to be lacking although it might be accepted that the vasodilation of muscular vessels increases the flow in the capillary network (at least until perfusion pressure is adequate).

40 G.P. Novelli

However, NO reduces or prevents platelet aggregation [27, 28], the adhesion of granulocytes to the endothelium [20] and inhibits the release of oxygen radicals from activated granulocytes [29]. Inhibition of NOS results in a profound increase in oxidant generation in the mesentery of rats [21], so favoring adhesion of granulocytes [30, 31], increases venular permeability and alters endothelial actin cytoskeleton [32].

A potential role for NO in the pathogenesis of sepsis in man has been proposed [33] and therapy of sepsis in men with NOS inhibitors has been attempted.

A study of the intestinal microcirculation of septic rats performed by Spain [34] demonstrated that NOS inhibition exacerbates the constriction and hypoperfusion.

The inhibition of nitric oxide synthesis increases the adhesion of granulocytes to the endothelium [35], possibly involving leukotrienes [36].

The regulation of microcirculation is equivalent to the regulation of tissue oxygenation: many mediators are active at the same time, but at present our ability to manipulate them is very poor.

References I. Zweifach BW (1961) Functional behavior of the microcirculation. Thomas, Springfield 2. Mortillaro NA (1983) The physiology and pharmacology of the microcirculation. Academic

Press, New York 3. Gaehtgen P (1990) Microcirculatory control of tissue oxygenation. In: Vincent JL (ed) Up

date in intensive care and emergency medicine. Springer, Berlin, pp 44-52 4. Zweifach BW, Thomas L (1957) The relationship between the vascular manifestations of

shock induced by endotoxin, trauma and hemorrhage. J Exper Med 106:385-40 I 5. Thomas L (1954) The physiological disturbances produced by endotoxin. Rev Physiol 16:

467-490 6. Astiz ME, De Gent ME, Lin R (1995) Microvascular function and rheologic changes in hy

perdynamic sepsis. Crit Care Med 23:265-271 7. Rogers F, Dunn R, Barrett J et al (1985) Alterations of capillary blood flow during sepsis.

Circ Shock 15: I 05-110 8. Greenberg S, Curro FA, Tanaka PT (1983) Regulation of vascular smooth muscles of the mi

crocirculation. In: Mortillaro NA (ed) The physiology and pharmacology of the microcirculation. Academic Press, New York, pp 39-141

9. Evequoz D, Waeber B, Corder R et al (1987) Markedly reduced blood pressure responsiveness in endotoxemic rats: reversal by neuropeptide Y. Life Sci 41 :2573-2580

10. Parratt JR (1989) Alterations in vascular reactivity in sepsis and endotoxemia. In: Vincent JL (ed) Update in intensive care and emergency medicine. Springer, Berlin, pp 27-40

II. Christmas JW, Lancaster LH, Blackwell TS (1988) Nuclear factor kB: a pivotal role in the systemic inflammatory response syndrome and new target for therapy. Int Care Med 24: 1131-1138

12. Essani NA, Fisher MA, Jaeschke H (1997) Inhibition of NF-kB activation by dimethylsulfoxide correlates with the suppression of TNF-a formation, reduced ICAM gene transcription and protection against endotoxin-induced liver injury. Shock 7:90-96

Biochemical Regulation of the Microcirculation 41

13. Shu FL, Xiaobing YE, Malik AB (1999) Pirrolidine dithiocarbamate prevents I-kB degradation and reduces microvascular injury induced by lipopolysaccharide in multiple organs. Molec Pharmacol 55:658-677

14. Schmid W, Stenzel K, Gebhard MM et al (1999) Cl-esterase inhibitor and its effects on endotoxin induced leukocyte adherence and plasma extravasation in postcapillary veins. Surgery 125:280-287

15. Suzuki YI, Packer RL (1993) Inhibition of NF-kB activation by vitamin E derivatives. Biochem Biophys Res Commun 193:277-283

16. Scannell G (1996) Leukocyte responses to hypoxic/ischemic conditions. New Horiz 4: 179-183

17. Minghini A, Britt LD, Hill MA (1998) Interleukin I and interleukin 6 mediate muscle arterial vasodilation: "in vitro" versus "in vivo" studies. Shock 9:210-215

18. Lou J, Donati YAR, Juillard P et al (1997) Platelets play an important role in TNF-induced microvascular endothelial cell pathology. Am J Pathol 151: 1397 -1405

19. Maekawa K, Futami S, Nishida M et al (1998) Effects of trauma and sepsis on soluble L-selectin and cell surface expression of L-selectin and CD 116. J Trauma 44:460-467

20. Kubes P, Suzuki K, Granger DN (1991) Nitric oxide: an endogenous modulator of leukocyte adhesion. Proc Natl Acad Sci USA 88:4651-4655

21. Kurose T, Anderson DC, Miyasaka M et al (1994) Molecular determinants of reperfusion induced leukocyte adhesion and vascular protein leakage. Circ Res 74:336-343

22. Kurose I, Kubes S, Wolf Ret al (1993) Inhibition of nitric oxide production. Mechanisms of vascular albumin leakage. Circ Res 73: 164-171

23. Granger DN (1988) Role of xanthine oxidase and granulocytes in ischemia-reperfusion injury. Am J PhysioI255:HI269-HI275

24. Cohen RA, Vanhoutte PM (1995) Endothelium dependent hyperpolarization. Beyond nitric oxide and cGMP. Circulation 92:3337-3349

25. Rubanyi GM, PolokoffMA (1994) Endothelins: molecular biology, biochemistry, pharmacology, physiology and pathophysiology. Pharmacol Rev 46:325-415

26. Moncada S, Higgs A (1993) The L-arginine-nitric oxide pathway. N Engl J Med 329:2002-2012

27. Radomski MW, Palmer RMJ, Moncada S (1987) Endogenous NO inhibits human platelet adhesion to vascular endothelium. Lancet 2: 1057 -1058

28. Radomski MW, Palmer RMJ, Moncada S (1990) Characterization of the L-arginine-nitric oxide pathway in human platelet. Brit J PharmacollOl :325-328

29. Clancy RM, Leszczynska-Piziak L, Abramson SH (1992) Nitric oxide, an endothelial cell relaxing factor, inhibits neutrophil superoxide anion production a direct action on the NADPH oxidase. J Clin Invest 90: 1116-1121

30. Suzuki M, Asako H, Kubes Pet al (1991) Neutrophil-derived oxidants promote leukocyte adherence in post-capillary venu1es. Microvasc Res 42: 125-138

31. Mitchell DJ, Jingcheng YU, Kabel T (1998) Local L-NAME decreases blood flow and increases leukocyte adhesion via CD 18. Am J PhysioI274:HI264-HI268

32. Baldwin AL, Thurston G, Al Naemi H (1998) Inhibition of nitric oxide synthesis increases venular permeability and alters endothelial actyn cytoskeleton. Am J Physiol 274:H 1776-H1798

33. Kirkeboen KA, Strand OA (1999) The role of nitric oxide in sepsis: an overview. Acta Anaesth Scand 43:275-288

34. Spain DA, Wilson MA, Bar-Natan MF et al (1994) Role of nitric oxide in the small intestinal microcirculation during bacteremia. Shock 2:41-46

35. Niu X, Smith W, Kubes P (1994) Intracellular oxidative stress induced by nitric oxide synthesis inhibition increases endothelial cell adhesion to neutrophils. Circ Res 74: 1133-1140

36. Arndt H, Russell JB, Kurose I et al (1993) Mediators of leukocyte adhesion in rat mesenteric venules elicited by inhibition of nitric oxide synthesis. Gastroenterology 105:675-680

Oxygen Supply and Consumption in Tissues

A. MAYR, W. PAJK, W. HASIBEDER

To maintain a sufficient tissue oxygen tension for adequate aerobic metabolism mammals developed complex oxygen transport systems. Oxygen must diffuse from capillaries to intracellular compartments, where oxygen consumption takes place. The oxygen tension necessary to keep intra- and extramitochondrial oxygen consuming enzymatic processes saturated differs markedly. In addition a significant difference in the oxygen dependence of isolated mitochondria and in vivo oxidative phosphorylation exists. Cellular diffusion barriers, differences in oxygen solubility within the cytoplasm, mitochondrial clustering and differences in oxygen affinity of enzymatic processes may account for these observations. It is likely that in vivo several tissues already live at the margin of hypoxia. In addition extramitochondrial oxygen consumption may be limited at tissue oxygen tensions above values where substrate limitation of oxidative phosphorylation starts. Evidence suggests that inhibition of extramitochondrial oxidase reactions may result in significant organ dysfunction without changes in intracellular energy charge. Manipulation of systemic and organ oxygen consumption to gain "biological time" during episodes of decreased oxygen delivery or greatly enhanced energy needs is everyday clinical practice in intensive care units.

This article focuses on determinants of cellular oxygen supply and consumption. In addition examples of effective clinical manipulation of oxygen transport parameters in order to gain biological time are given.

Basal metabolic rate for oxygen For any given organism and any specific tissue there is a minimum metabolic rate and oxygen consumption required to assure cellular integrity and function under certain defined conditions [1]. For example in healthy humans with normal body composition, under resting conditions and after at least 12 hours fasting' oxygen consumption is fairly constant at a rate of 0.17 mmolxkg- I xmin- I

under thermoneutral conditions. Approximately 90-95% of oxygen is consumed by mitochondria to produce cellular energy in the form of adenosine-triphosphate (ATP), the main cellular energetic fuel for enzymatic processes. Approxi-

44 A. Mayr, W Pajk, W Hasibeder

mately 5-10% of whole body oxygen consumption results from extramitochondrial oxidases involved in production of neurotransmitters and detoxification processes [2]. Most of the ATP fuels active electrolyte transport across cellular membranes to maintain membrane potentials. For example it is estimated that 30-60% of ATP is required for proper sodium potassium transport across the cell membrane [3]. Protein biosynthesis requires approximately 5% of ATP produced. However, in certain specific cells, e.g. reticulocytes, protein biosynthesis may account for up to 30% of ATP consumption [4]. The remainder of ATP is required to fuel vital organ function of heart, respiratory muscles, brain or to maintain energy requiring transport processes, e.g. within the liver and gut.

In mammals whole body oxygen consumption can be kept constant over a wide range of systemic oxygen supply by virtue of several compensatory physiologic mechanisms involved in systemic oxygen transport [5]. During altitude exposure (decreased inspired oxygen concentration) minute ventilation increases in order to improve alveolar oxygen partial pressure, cardiac output increases to maintain systemic oxygen delivery and with prolonged exposure to hypoxia haematocrit increases to augment arterial oxygen carrying capacity. The main adaptive mechanism to moderate to severe anaemia is an increase in cardiac output while during low cardiac output states systemic oxygen extraction ratio is augmented in order to maintain systemic oxygen supply. Additionally under all three conditions a rightwardshift of the oxyhaemoglobin dissociation curve has been described, facilitating oxygen release from haemoglobin at capillary oxygen tensions.

If these compensatory mechanisms are exhausted or fail tissue hypoxia develops and cellular injury leading to organ dysfunction, failure and death may result. Although increased anaerobic metabolism may prolong hypoxia tolerance, the efficiency of this defensive strategy in men is not very effective. In contrast to aerobic metabolism of glucose which results in the production of 36 mol ATP/mol glucose consumed, anaerobic metabolism of glucose only yields 2 mol ATP/mol glucose [6]. Development of progressive intracellular acidosis during increased anaerobic metabolism leads to inhibition of cell metabolism and deterioration of mitochondrial function further decreasing cell viability [7].

Factors governing cellular oxygen consumption According to Fick's law of diffusion oxygen transport within a given tissue will depend on the oxygen tension gradient from capillary to cell, the available capillary surface area for gas exchange, the diffusion properties of tissue and the distance between capillary and a specific tissue point of interest. Furthermore, the oxygen tension gradient depends on tissue oxygen consumption and will increase with augmented aerobic metabolism, as long as available capillary surface area and capillary oxygen tension remain constant. In most tissues investi-

Oxygen Supply and Consumption in Tissues 45

gated measured values of oxygen tension are therefore quite heterogeneous ranging from values between 3-5 mmHg to values approaching arterial P02.

The oxygen concentration necessary to maintain aerobic metabolism in isolated mitochondria has been found to be very low and oxygen supply limitation has been reported to occur at oxygen concentrations of 0.02-0.3 fllllol, roughly equivalent to oxygen tensions between 0.1-0.5 mmHg [2, 8]. However, in intact cells mitochondrial aerobic function is usually limited at oxygen concentrations 10 to 100 times higher when compared with isolated mitochondria, increasing the risk for oxygen supply limitation during periods of disturbed systemic oxygen transport. Several factors account for this difference: the cell membrane and cytoplasm represent a significant diffusion barrier for oxygen [8, 9]. In isolated liver cells partial destruction of the cell membrane significantly decreases the oxygen concentration necessary for half-maximal oxidation of cytochrome c [9]. In experiments using fluorescence dye, which changes fluorescence intensity in dependence of oxygen concentration, significant intracellular heterogeneity concerning oxygen solubility within cytoplasm has been demonstrated [10]. Oxygen solubility may vary by a factor of 3-5 within the cytoplasm and in general the oxygen diffusion coefficient of cytoplasm is only about 1170 of that measured for normal saline [10, 11]. Therefore, oxygen consumption may be limited differently according to intracellular location of oxygen consuming metabolic processes. Finally, in highly metabolising tissues, e.g. liver cells, epithelial cells and working muscle mitochondria are clustered near locations of high ATP demand [8, 9]. In cardiac muscle mitochondria are concentrated in columns between myofibrils, in epithelial cells mitochondria are mostly found near the basement membrane and in liver cells mitochondria are packed in clusters approaching 2 fllll in diameter at locations of high energy metabolism. Mitochondrial clustering increases local oxygen consumption by adding the oxygen consumption of individual mitochondria thereby creating an oxygen sink responsible for significant intracellular oxygen gradients within cells and from the cell membrane to mitochondria.

It is likely that even in humans some tissue compartments work at the nadir of hypoxia or under conditions of oxygen supply limitation. For example, by using whole organ transmission spectroscopy in the isolated perfused rat kidney, it was found that 25-40% of cytochrome aa3 appeared to be in its reduced form [12]. Agents that altered metabolism within the renal cortex did not affect cytochrome aa3 oxidation. However, blocking active chloride transport within the renal medulla by application of loop diuretics significantly increased cytochrome oxidation demonstrating that tissue within the medulla already worked under hypoxic conditions. Other authors reported cytochrome reductions approaching 20% in the heart and brain of different mammal species in isolated organ perfusion experiments [7].

In several animal species oxygen deprivation leads to adaptive changes in cell metabolism significantly prolonging tissue hypoxia tolerance [6, 13]. Recent experimental work in mammal cells demonstrated that cytochrome oxidase func-


tion may adapt to prolonged periods of hypoxia in order to maintain oxygen consumption [14, 15]. In addition, during hypoxia mitochondria might generate reactive oxygen species, which may participate in a variety of intracellular signalling processes, probably involved in cell adaptation to hypoxia, In cardiomyocytes progressing hypoxia leads to a stepwise increase in the production of superoxyd anion and hydrogen peroxide associated with a concomitant decrease in contractile function [16]. Antioxidant treatment abolished the decrease in contractile function during hypoxia. Therefore, it is conceivable that significant adaptation processes to hypoxia at cellular level may also exist in human tissues.

Extramitochondrial oxygen consumption and cell function Oxygen is substrate for more than 100 different cellular oxidase reactions in mammal cells. The half maximum saturation values of these enzymatic reactions differ markedly and a large number of metabolic changes occur at oxygen concentrations higher than the critical value of mitochondrial oxygen consumption [2]. Within the nervous system oxygen is substrate for synthesis and degradation of dopamine, norepinephrine and serotonine. In addition, synthesis of amino acid neurotransmitters (aspartate, glutamate, gamma-aminobutyrate, glycine) and acetylcholine depends upon oxidation of glucose for their carbon skeletons [17].

Even mild to moderate hypoxia causes some central nervous system dysfunction, e.g. impaired dark adaptation of retinal rots and cones occurring at an alveolar oxygen tension (PA02) of approximately 85 mmHg. At PA02 values of 45-60 mmHg humans demonstrate impairments in concentration and short-term memory. Below PA02 values of 45 mmHg lethargy, hallucinations and muscular incoordination occurs which progress to unconsciousness at PA02 values of 35 mmHg [16]. It is interesting to note that brain high energy phosphates only decline at PA02 values below 35 mmHg [18].

Within the liver several extramitochondrial metabolic processes depend on molecular oxygen [2]. Cytochrome P-450 function decreases under moderate hypoxia. Monoamine oxidase, an enzyme that oxidises biogenic amines appears to be oxygen dependent in vivo under normoxic conditions. Studies of drug conjugation reactions have shown that these metabolic processes, although not directly dependent upon oxygen as substrate, are also selectively affected by hypoxia. For example rates of sulphate and glycine conjugation parallel the decrease in cellular energy charge during hypoxia. Glucuronidation reactions are sensitive to hypoxia and sensitivity increases during conditions of starvation.

Clinical significant organ dysfunction may occur under conditions of oxygen deprivation without any deterioration in oxidative phosphorylation. It still remains to be determined whether prolonged decreases of extramitochondrial oxygen consuming metabolism may be harmful in humans.

Oxygen Supply and Consumption in Tissues 47

Clinical manipulation of the oxygen transport system Manipulation of systemic oxygen delivery and systemic oxygen consumption is daily practice in intensive care units. Clinicians are often not aware that most of their actions will actively manipulate parts of the oxygen transport system in order to gain biological time by reducing the metabolic rate of organ systems or whole body. For example, analgosedation in the intensive care unit significantly reduces the stress response to critical illness protecting the patient from consequences of uncontrolled sympathetic and hormonal output [19]. Assisted and controlled ventilation will decrease or eliminate workload of respiratory muscles significantly attenuating cardiovascular stress especially in patients with limited cardiac reserve [20]. In the elderly flair patient with moderate to high degree fever, reduction of systemic oxygen consumption by aggressive cooling either externally or sometimes with the aid of continuous venovenous haemofiltration may be life-saving due to a significant reduction in systemic oxygen consumption. Quite similar procedures to reduce cardiac workload after complicated cardiac surgery or during periods of cardiac failure from other aetiologies, e.g. left ventricular assist devices, intraaortic balloon counterpulsation are often life-saving in patients by their ability to significantly reduce myocardial oxygen consumption and to selectively improve myocardial oxygen supply during critical periods of illness. These examples clearly demonstrate that the rational of most activities in the intensive care unit is based on their ability to manipulate biological time of a certain patient.

Conclusion Oxygen supply dependency of isolated mitochondria and cells differ markedly. Diffusion barriers for oxygen in vivo, mitochondrial clustering and heterogeneities of intracellular oxygen solubility may account for these differences. Extramitochondrial oxygen consuming metabolism, although only representing 5-10% of whole body oxygen consumption, is important for maintaining vital organ functions especially in the central nervous system. Exhaustion of adaptive mechanisms of the oxygen transport system leads to organ dysfunction, organ failure and death. Manipulation of oxygen transport parameters in the intensive care unit is everyday practice and significantly affects outcome by prolonging biological time.


References 1. Hochachka PW, Guppy M (1987) The time extension factor. In: Hochachka PW, Guppy M

(eds) Metabolic arrest and the control of biological time. Harvard University Press, Cambridge Massachusetts London, pp 1-9

2. Jones DP, Kennedy FG, Andersson BS et al (1985) When is a mammalian cell hypoxic? Insights from studies of cells versus mitochondria. Molecular Physiology 8:473-474

3. Hulbert AJ, Else PL (1981) Comparison of the "mammal machine" and the "reptile machine": Energy use and thyroid activity. Am J PhysioI241:R350-R356

4. Rapoport SM (1985) Mechanisms of the maturation of reticulocytes. In: Gilles R (ed) Circulation, respiration and metabolism. Springer, Berlin, pp 333-342

5. Robin ED (1980) Of men and mitochondria: Coping with hypoxic dysoxia. Am Rev Respir Dis 122:517-531

6. Hochachka PW, Mommsen TP (1983) Protons and anaerobiosis. Science 219:1391-1397 7. Gutierrez G (1991) Cellular energy metabolism during hypoxia. Crit Care Med 19:619-626 8. Kennedy FG, Jones DP (1986) Oxygen dependence of mitochondrial function in isolated rat

cardiac myocytes. Am J Physiol 250:C374-C383 9. Jones DP (1984) Effect of mitochondrial clustering on O2 supply in hepatocytes. Am J Physi-

01 247:C83-C89 10. Benson DM, Knopp JA, Longmuir IS (1980) Intracellular oxygen measurements of mouse liv

er cells using quantitative fluorescence video microscopy. Biochim Biophys Acta 591: 187 -197 11. Mastro AM, Babich MA, Taylor WD et al (1984) Diffusion of a small molecule in the cyto

plasm of mammalian cells. Proc Natl Acad Sci USA 81:3414-3418 12. Epstein FH, Balaban RS, Ross BD (1982) Redox state of cytochrome aa3 in isolated perfused

rat kidney. Am J PhysioI243:F356-F363 13. Hochachka PW, Guppy M (1987) Animal anaerobes. In: Hochachka PW, Guppy M (eds)

Metabolic arrest and the control of biological time. Harvard University Press, Cambridge Massachusetts London, pp 10-35

14. Arai AE, Pantely GA (1991) Active downregulation of myocardial energy requirements during prolonged moderate ischemia in swine. Circ Res 69: 1458-1469

15. Arai AE, Grauer SE, Anselone CG et al (1995) Metabolic adaptation to a gradual reduction in myocardial blood flow. Circulation 92:244-252

16. Duranteau J, Chande1 NS, Kulish A et al (1998) Intracellular signalling by reactive oxygen species during hypoxia in cardiomyocytes. J Bioi Chem 273: 11619-11624

17. Gibson GE, Pulsinelli W, Blass JP et al (1981) Brain dysfunction in mild to moderate hypoxia. Am J Med 70:1247-1254

18. Siesjo BK (1984) Cerebral circulation and metabolism. J Neurosurg 60:883-908 19. Mangano DT, Siliciano D, Hollenberg M et al (1992) Postoperative myocardial ischemia -

Therapeutic trials using intensive analgesia following surgery. Anesthesiology 76:342-353 20. Hurford WE, Lynch KE, Strauss HW et al (1991) Myocardial perfusion as assessed by thalli

um-201 scintigraphy during the discontinuation of mechanical ventilation in ventilator-dependent patients. Anesthesiology 74: 1007 -10 16

Ischaemia-Reperfusion in Sepsis

C. ADEMBRI, A.R. DE GAUDIO, G.P. NOVELLI

It has been over 20 years since Hearse described the "oxygen paradox" and the "calcium paradox" in hearts undergoing ischaemia-reperfusion (I-R) [1]. The concept that it is reperfusion itself that increases and potentiates ischaemia-induced damage has progressively gained ground, and it is now one of the main factors taken into account in the treatment of various pathological states, from crush injury to transplantation [2, 3]. Even when reperfusion itself is the goal of therapy, such as during thrombolysis for infarcted myocardium, reperfusion-associated dysfunctions, which range from arrhythmias to stunning, must be considered as they significantly affect morbidity and mortality rates [4, 5]. Regardless of the cause and the modality that have provoked it, the sequence of ischaemia and reperfusion actually induces a typical inflammatory response which is not restricted to the injured tissue but frequently has a systemic recoil [6].

Recently, it has been proposed that I-R plays a critical role in initiating sepsis or multiple organ failure (MOF) [7]. According to this hypothesis, defects in tissue perfusion (which may range from low flow or total ischaemia to partial or complete restoration of flow), are no longer viewed as the natural consequence of the impairment in micro and macrocirculation that is a late sepsis or MOF-associated event. Indeed, such episodes of I-R may actually trigger a chain of events that is typical of the inflammatory response which promotes, maintains and characterises septic states [8, 9]. The inflammatory response may immediately follow on I-R when this is sufficiently severe as to activate the inflammatory system directly (the one-hit model). Otherwise, it may explode subsequently if the initial I-R has only primed the ischaemic-reperfused tissue so that it will respond in an exaggerated way when exposed to any another minimal insult (the two-hit model) [10].

Although any organ can undergo I-R in septic patients, particular attention has recently been focused on the gut for two main reasons, one anatomical and the other functional: 1) The anatomical configuration of the microvillar circulation (a single central artery surrounded by a network of capillaries and venules which can easily shunt oxygenated blood) predisposes intestinal mucosa to ischaemic insult whenever perfusion pressure fails [11, 12]. This is extremely frequent in septic patients, in whom blood is diverted from gut to more "noble or-

50 C. Adembri, A.R. De Gaudio, G.P. Novelli

gans" as a consequence both of an endogenous protective mechanism and of pharmacological interventions [13, 14]. 2) The gastrointestinal tract, which is continuously and naturally exposed to many different antigenic, mutagenic, toxic stimuli, is an organ which is extremely rich in immunocompetent and inflammatory cells, and even under normal conditions leukocytes are present in the intraepithelial and subepithelial compartments [15]. It thus follows that even a minimal ischaemic insult or episode of I-R finds here the conditions for promoting an intense inflammatory reaction that is likely to have systemic effects, even if these are initially not detected because they have no immediate clinical consequences.

Mechanisms of 1-R damage

The relative contribution of the ischaemic and reperfusion phases in determining tissue damage remains unknown since it depends on many factors.

Ischaemia causes damage which is proportional to: its duration (the amount of conversion of xanthine dehydrogenase to xanthine oxidase is proportional to the duration of ischaemia itself); its degree (complete or partial); the modality of its appearance (acute or chronic, in the latter case allowing protective mechanisms to be activated, such as K+ ATP channels in ischaemic myocardium); the oxygen-dependence of the organ; and a much-neglected fact, the coexistence of concomitant diseases affecting the endothelium (such as hypertension or diabetes) [16-19].

Reperfusion damage itself depends on many factors; a role in it has been demonstrated for oxygen tension of reperfusion flow (with high oxygen content, the massive generation of oxygen radicals (OR) that ensues overwhelms the organs' scavenger activity) for pH of the cells, and if it is totally or partially effective (the "no-reflow" phenomenon, that is the failure of blood to flow at reperfusion can aggravate ischaemia and severely influence the evolution of the damage) [20-22]. As is easily perceived, ischaemic and reperfusion injuries are not two independent entities: some alterations are generated during ischaemia or are triggered by it even if they are not evident until reperfusion ensues. Thus, the intensity of OR production in the myocardium has been shown to be proportional to the severity of antecedent ischaemia or hypoperfusion [23].

It is also very difficult to distinguish the relative importance of all the cell types or mediators reported to have a role in promoting or sustaining the inflammatory response to I-R. However, an increasing body of evidence indicates that it is the endothelium that has a major role in initiating the ischaemic-reperfusion sequences, acting as trigger of the immune/inflammatory system [24]. At the same time, it is now evident that polymorphonuclear cells (PMNs) are the cells which are mainly involved in producing damage ("the leukocyte amplification phase"), because of their capacity for infiltrating tissues and generating toxic

Ischaemia-Reperfusion in Sepsis 51

products [25, 26]. Therefore, understanding of leukocyte/endothelium interactions is fundamental to the knowledge of the determinants of I-R-induced microvascular injury and dysfunction and also to their treatment or prevention.

Of the humoral mediators that have a role in causing I-R damage, OR and nitric oxide (NO) appear to be those mainly involved. These molecules, alone or reacting together, can have a direct toxic action on exposed cells or can modify important cellular pathways, such as those regulating the expression of pro-inflammatory genes. Moreover, a growing body of evidence suggests that OR and NO are very important physiological modulators of the interactions between leukocytes and endothelium that characterise the inflammatory response [26].

OR, I-R and inflammation Oxygen radicals (such as O2-, OH) derive from oxygen during its reduction to water. They possess an unpaired electron in the external orbital, rendering them unstable from an energy point of view, and thus extremely reactive. Biologically speaking, even a molecule that does not strictly have a radical chemical configuration can be called an oxygen radical (such as H20 2, HCIO), because it shares with the true radicals the same origin from oxygen and the same facility to react to produce more stable end-products [27].

There are many potential sources of OR in the biological set. These are: xanthine oxidase (XO); the membrane-bound NADPH oxidase of PMNs; the activation of the arachidonic cascade; and the mitochondrial respiratory chain. The relative importance of each single source depends on the species and tissue considered. For example, while the XO system seems to be unimportant in human skeletal muscle [2, 28], this is not true for intestinal villi. In the gut, xanthine dehydrogenase (XD) may be transformed into XO not only by protease activation during ischaemia but also directly by pancreatic enzyme proteolysis. Moreover, in the gut XO is found exclusively in the mucosal layer, with increasing content from the base to the tip of the villus. Thus, the distribution of the enzyme is very similar to that of the I-R associated damage [29], further supporting its role in I-R-associated gut damage.

OR have no single cell target. They react with many cell components, from lipid membranes to the sugar portion of DNA, causing profound functional and even structural changes. Morphologically, the hallmark of peroxidation can be considered to be the appearance of cell or organelle swelling as a consequence of changes in membrane permeability due to lipid peroxidation. Disruption of cell integrity may be evident in more severe cases [30].

In addition to a "direct" cytotoxic action, an important role for OR-associated damage during I-R is linked to their ability to interact with the nuclear factor kB (NF-kB). NF-kB is one of the most important transcription factors involved in the expression of many inflammatory and immune response genes. The acti-


vation of NF-kB leads to a co-ordinated increase in the expression of many genes whose products mediate inflammatory and immune responses, such as the endothelial adhesion molecules involved in PMN cell adhesion to endothelium, as well as cytokines and growth factors [31].

Thus OR, by promoting the activation of this factor, playa crucial role in the amplification of I-R-induced inflammatory response.

In contrast to their proinflammatory action, OR may exert a "protective" role. According to recent data, low intracellular levels of OR, such as those produced by mitochondria in hypoxic conditions, might activate intracellular signalling pathways that allow the cell to adapt to unfavourable conditions, by reversible suppression of ATP utilisation (hypoxic adaptation) [32].

OR, I-R and the gut At intestinal level, two major effects of I-R-induced OR production are changes in intestinal permeability and the bacterial translocation phenomenon [14]. Indeed, increased intestinal permeability occurs both in experimental models of 1-R and in humans after severe trauma, haemorrhagic shock or sepsis, all of these conditions being characterised by changes in organ perfusion. On the other hand, bacterial translocation, i.e. the passage of bacteria through the intestinal mucosal barrier to reach the systemic circulation, has been clearly demonstrated only in experimental models but has not clearly been identified in humans. In fact, although bacterial translocation has been reported in humans in different clinical conditions, it does not involve the systemic circulation but appears to be restricted to regional lymph nodes. Thus it cannot be directly considered as a cause of sepsis or MOE It is very likely that in humans the increase in mucosal permeability and translocation are two independent processes [14].

Thus the relation between gut I-R and sepsis appears to lie rather in the ability to promote an inflammatory response which may have systemic effects than to the causing of direct invasion of the bloodstream by pathogenic bacteria.

NO, I-R and inflammation NO is a lipid-soluble, free radical molecule with a quite weak reactivity, exerted almost exclusively with oxygen, transition metals and other radicals. NO can act as a signalling molecule and, by triggering the production of a second messenger, GMP, it has direct regulatory and/or anti-inflammatory functions, such as the regulation of vascular tone, the inhibition of activation, adhesion and aggregation of platelets and PMNs [33-35]. NO can also exert indirect effects through NO-derived molecules, and these effects, almost all proinflammatory and cyotoxic, mainly consist of the inhibition of enzyme function, the promotion of


lipid peroxidation and the depletion of antioxidant molecules [35]. As a consequence of this dual functional aspect, it is difficult to define exactly the role of NO and its derivative in I-R and sepsis, in which NO has been considered alternately as beneficial and deleterious for the evolution of clinical conditions [33].

One of the major causes of these conflicting opinions may be the tendency to consider NO alone. It is now evident that some biological actions previously attributed to the radical form of nitric oxide (NO.) are actually to be ascribed to other "reactive nitrogen species" which are formed by NO. Only some of these molecules are radicals; they have different chemical structures and therefore different behaviour patterns. For example, peroxynitrite, ONOO-, the product of the reaction between O2- and NO, while not a radical molecule is highly toxic for cells. Moreover, the chemical attitude of NO in its radical configuration differs greatly from that of NO+ and NO-. Thus, the different chemical configurations of NO, together with changes in the availability of cellular substrates, may completely shift the expected reactions in opposite directions, with consequent different, perhaps even opposite, physiological effects [35]. For example, for ONOO- to be generated, a precise ratio between NO and O2- concentrations is necessary. Since the cell concentrations of O2- and NO under normal conditions have been estimated to be 1nM and 1-10 ~M, respectively, it seems likely that it is the O2 concentration that sets off the reaction between these radicals ONOO[26]. The concentration of NO influences where peroxynitrite is formed. NO preferentially diffuses through membranes before reacting and then, where it finds an adequate O2- concentration, it reacts with it to produce peroxynitrite. Thus, peroxynitrite is rarely formed where NO is produced, and it can act without damaging the NO- producing cell [26].

Conflicting results indicating the ambiguous role of NO come particularly from experimental studies on splanchnic I-R.

NO as an anti-inflammatory molecule in intestinal J-R

In the splanchnic artery occlusion and reperfusion (SAO/R) experimental model, within a few minutes of reperfusion NO production is reduced to about 40-60%, and reaches about 75-85% after 2 hours. Accordingly, P-selectin expression is enhanced, neutrophils adhere to the endothelium and infiltrate tissues, following all the steps that characterise the acute inflammation process [25]. In this condition, the cause of impaired NO production is very probably the interaction of NO with O2-, at first generated not by activated neutrophils but by the ischaemic/reperfused endothelium. In fact, when molecular oxygen is reintroduced into the ischaernic tissue, the endothelial XO (formed during ischaemia by a proteolytic process from XD in the presence of high levels of intracellular calcium) acts on accumulated hypoxanthine to form uric acid, and O2- as a byproduct [24]. O2- greatly favours leukocyte-endothelium interactions, promoting adhesion and subsequently infiltration of PMNs into ischaemic-reperfused tissue [26].


In this setting, exogenously-administered NO is protective and NO blockade enhances damage [33, 36], because NO loss acts as an endothelial trigger, allowing the sequence of events previously described and the subsequent involvement of neutrophils (the leukocyte amplification phase).

NO as a pro-inflammatory molecule The pro-inflammatory role that NO has been seen to play in experimental models of inflammation or I-R of the gut may be due not only to the prevalence of production of highly-toxic NO derivatives, but also to the observation time of the experiments [33, 37, 38]. In acute inflammation models (observation time within 3 hours), only the effects of NO produced by constitutive NOS (cNOS), which are pre-eminently regulatory effects, have been revealed. Constitutive NOS is calcium-dependent and therefore produces NO in response to stimuli that modify intracellular calcium level, but as the calcium levels rapidly return to normal, NO production is not maintained [37]. On the other hand, the production of NO from inducible nitric oxide synthase (iNOS) is independent of intracellular calcium levels, and depends primarily on the regulation of iNOS transcription. Once it is expressed, iNOS produces NO independently of any feedback control. The expression of iNOS takes some hours to complete. Thus, for example, experiments in which LPS is employed to provoke gut injury but have an observation time limited to 3 hours have failed to clarify the contribution of NO production by iNOS to the pathogenesis of sepsis. When the experimental protocols have been extended to 6 hours, the pathological, pro-inflammatory role of NO and the beneficial effect of blocking its production in terms, for example, of decreased bacterial translocation and mucosal permeability and improved mitochondrial function, have been clearly shown [37, 39, 40]. At that observation time, the production of large amounts of NO coincides with the production of large quantities of O2-, rendering possible their reaction to produce peroxynitrite.

The pro-inflammatory action of NO is partly sustained by the activation of nuclear factor kB. In unstimulated cells NF-kB is found in cytoplasm, where it is bound to IkBa and IkB~, which prevent it from entering the nuclei. When the cell is stimulated, IKE is phosphorylated and thus degraded and NF-kB freely enters the nucleus, where it binds to promote regions of targeted genes. The activation of NF-kB leads to a co-ordinated increase in the expression of many genes whose products mediate inflammatory and immune response, including iNOS [31]. Many stimuli may activate NF-kB, the most important of these being TNFa, PAF, LPS and oxidants such as H20 2. Thus, OR can enhance the production of NO by inducing the expression of iNOS. Glucocorticoids are potent inhibitors of NF-kB activation, and this may account for most of their antiinflammatory action and renew the interest for these therapeutic agents in sepsis [31].


Conclusions Emerging data seem to suggest the paramount importance of gut I-R for the deleterious effects it can have not only on intestinal function but also, and perhaps more importantly, on triggering an inflammatory response which may directly cause or favour the development of sepsis. As a consequence, attempts to "resuscitate" the gut together with the other organs are now considered a priority in critically ill patients, with the aim of preventing the appearance of I-R. In fact, the complexity of the interrelations between mediators and cells involved in I-R, interactions that may also vary with time, have so far rendered the pharmacological approach incomplete or inappropriate, hence unable to modify the clinical course significantly.

References 1. Hearse OJ, Humphrey SM, Bullock GR (1978) The oxygen paradox and the calcium paradox:

two facets of the same problem? J Mol Cell Cardiol 10:641-668 2. Odeh M (1991) The role of reperfusion-induced injury in the pathogenesis of the crush syn

drome. N Eng J Med 324: 1417-1422 3. Pegg DE (1986) Organ preservation. Surg Clin North Am 66:617-632 4. Grech ED, Malcolm 11, Ramsdale DR (1995) Reperfusion injury after acute myocardial in

farction. Br Med J 310:477-478 5. Bolli R, Marban E (1999) Molecular and cellular mechanisms of myocardial stunning. Physi-

01 Rev 79:609-634 6. Granger ON, Korthuis RJ (1995) Physiologic mechanisms of postischemic tissue injury. Ann

Rev PhysioI57:311-332 7. Meakins JL, Marshall JC (1986) The gastrointestinal tract: the "motor" of multiple organ fail

ure. Arch Surg 121:197-201 8. Biffl WL, Moore EE (1996) Splanchnic ischaemia and mUltiple organ failure. Br J Anaesth

77:59-70 9. Kirton OC, Civetta JM (1999) Ischemia-reperfusion injury in the critically ill: a progenitor of

multiple organ failure. New Horizons 7:87-95 10. Moore F, Sauaia A, Moore E et al (1996) Post-injury multiple organ failure: a bimodel phe

nomenon. J Trauma 40:501-512 II. Takala J (1996) Determinants of splanchnic blood flow. Br J Anaesth 77:50-58 12. Siegemund M, Studer W, Ince C (1998) Ischemiaireperfusion of the gut. In: Vincent JL (ed)

Yearbook of Intensive Care and Emergency Medicine. Springer, Berlin, pp 637-648 13. Marik PE (1999) Total splanchnic resuscitation, SIRS, and MODS. Crit Care Med 27:257-258 14. Vallet B, Lebuffe G (1999) The role of the gut in multiple organ failure. In: Vincent JL (ed)

Yearbook of Intensive Care and Emergency Medicine. Springer, Berlin, pp 539-546 15. Fiocchi C (1997) Intestinal inflammation: a complex interplay of immune and nonimmune

cell interactions. Am J Physiol 273:G769-G775 16. Korthuis RJ, Smith JK, Carden DL (1989) Hypoxic reperfusion attenuates postischemic mi

crovascular injury. Am J PhysioI256:H315-H319 17. Parks DA, Williams TK, Beckman JS (1988) Conversion of xanthine dehydrogenase to oxi

dase in ischemic rat intestine: a reevaluation. Am J Physiol 254:G678-G774 18. Kersten JR, Waltier DC (1999) Modulation of the adaptive response to myocardial ischemia

by coexisting disease. Am J Physiol 276:H2268-H2270 19. Gross GJ, Fryer RM (1999) Sarcolemmal versus mitochondrial ATP-sensitive K+ channels

and myocardial preconditioning. Circ Res 84:973-979


20. Grace PA (1994) Ischemia-reperfusion injury. Br J Surg 81:637-647 21. Murphy E, Cross H, Steenbergen C (1999) Sodium regulation during ischemia versus reper

fusion and its role in injury. Circ Res 84: 1469-1470 22. Menger MD, Rucker M, Vollmar B (1997) Capillary dysfunction in striated muscle is

chemiaireperfusion: on the mechanisms of capillary "no-reflow". Shock 8:2-7 23. Bolli R, Patel BS, Jeroudi MO et al (1988) Demonstration of free radical generation in

"stunned" myocardium of intact dogs with the use of the spin trap alphaphenyl N-ter-butyl nitrone. J Clin Invest 82:476-485

24. Bulkley GB (1994) Reactive oxygen metabolites and reperfusion injury: aberrant triggering of reticuloendothelial function. Lancet 344:943-936

25. Lefer AM, Lefer DJ (1999) Nitric oxide II. Nitric oxide protects in intestinal inflammation. Am J Physiol G572-G575

26. Grisham MB, Granger DN, Lefer DJ (1998) Modulation of leukocyte-endothelial interactions by reactive metabolites of oxyge n and nitrogen: relevance to ischemic heart disease. Free Rad BioI Med 25:404-433

27. Southom PA, Powis G (1988) Free radicals in medicine. Chemical nature and biological reactions. Mayo Clin Proc 63:381-389

28. Novelli GP, Adembri C, Gandini E et al (1997) Vitamin E protects human skeletal muscle from damage during surgical ischemia-reperfusion. Am J Surg 172:206-209

29. AR'Rayab, Granger DN, Hollwarth et al (1986) Ischemia-reperfusion injury: role of oxygenderived freee radicals. Acta Physiol Scand 548[Suppl]:47-63

30. Formigli L, Domenici-Lombardo L, Adembri C et al (1992) Ischemia-reperfusion of human skeletal muscle: a role for neutrophils. Hum Path 23:627-634

31. Barnes PJ, Karin M (1997) Nuclear factor-kB. A pivotal transcription factor in chronic inflammatory disease. N Eng J Med 336: 1066-1071

32. Duranteau J, Chandel NS, Schumacker PT (1999) Intracellular signaling by reactive oxygen species during hypoxia. In: Vincent JL (ed) Yearbook of Intensive Care and Emergency Medicine. Springer, Berlin Heidelberg New York, pp 386-394

33. Fink MP (1999) Nitric oxide and the gut: one more piece in the puzzle. Crit Care Med 7: 248-249

34. Kirkeboen KA, Strand OA (1999) The role of nitric oxide in sepsis-an overview. Acta Anaesthesiol Scand 43:275-288

35. Grisham MB, Jourd'heuil D, Wink DA (1999) Nitric oxide 1. Physiological chemistry of nitric oxide and its metabolites: implications in inflammation. Am J Physiol 276:G3l5-G32l

36. Kubes P, Wallace JL (1995) Nitric oxide as a mediator of gastrointestinal injury? - say ain't so. Mediators Inflammation 4:397-405

37. Miller MJS, Sandoval M (1999) Nitric oxide III. A molecular prelude to intestinal inflammation. Am J Physiol G795-G799

38. Miller MJS, Grisham MB (1995) Nitric oxide as a mediator of inflammation? - you had better believe it. Mediators Inflammation 4:387-396

39. Unno N, Wang H, Menconi MJ et al (1997) Inhibition of nitric oxide synthase ameliorates lipopolysaccharide-induced gut mucosal barrier dysfunction in rats. Gastroenterology 113: 1246-1257

40. Sorrells DL, Friend C, Koltukutsuz U et al (1996) Inhibition of nitric oxide with aminoguanidine reduces bacterial translocation after endotoxin challenge in vivo. Arch Surg 131: 1155-1163

Mechanism of Oxygen Extraction Defect in Septic Shock

w. PAJK, H. KNOTZER, W. HASIBEDER

Critically ill patients suffering from septic shock most often present in a high output, low peripheral resistance cardiovascular status. In some patients cardiac output may increase to values above 15 Ixmin- 1 and mixed venous oxygen saturation may exceed normal values demonstrating decreased systemic oxygen extraction. Despite high systemic blood flow, progressive lactic acidosis and irreversible shock develops, suggesting progressive deterioration of oxygen supply to tissues. In these patients the major defect in organ oxygen supply is thought to be located within the microcirculation.

Although the existence of microcirculatory failure as a major cause of organ dysfunction, failure and death in septic shock is still a matter of debate, microcirculatory abnormalities have repeatedly been demonstrated in different animal models and in critical ill patients [1-4]. One problem encountered with studies investigating effects of sepsis and septic shock on microcirculatory function is, that animal models merely mimic the clinical situation. In patients septic shock usually develops gradually over several days. In contrast in animal models septic shock is acutely induced either by infusion of live bacteria or endotoxin, by caecal ligation and puncture or by infecting certain body compartments with large numbers of bacteria. These models ignore gradually developing host responses observed during the clinical course of disease which may affect microvascular responses differently. In addition, differences in the pathophysiologic response of the microcirculation to bacteraemia or endotoxaemia may exist and possible differences between species further complicate the problem.

This article summarises evidence for an oxygen extraction defect in sepsis and septic shock. In addition, the mechanisms of microcirculatory failure and possible therapeutic strategies to improve microcirculatory function are discussed.

Observation of an oxygen extraction deficit in sepsis In 1988 Nelson reported in a dog model of Pseudomonas aeruginosa bacteraemia that systemic critical oxygen delivery was increased and systemic critical

58 W Pajk, H. Knotzer, W Hasibeder

oxygen extraction ratio, the oxygen extraction ratio where oxygen supply dependency starts, was significantly decreased in comparison with control animals [5]. Subsequently the same authors observed that intestinal oxygen extraction was impaired, while the ability of skeletal muscle to extract oxygen was unchanged in an endotoxin shock model in dogs [6, 7]. In addition, the reactive hyperaemia response to short time arterial occlusion of intestinal vessels was absent in most endotoxic animals, demonstrating an inability of the microcirculation to adequately respond to a short time oxygen depth. Morphological analysis of capillary surface density in mucosal villi and crypts showed significantly higher perfused capillary density in control animals when compared with endotoxin treated animals [8]. In contrast to control animals no relationship between intestinal venous oxygen tension, intestinal oxygen extraction ratio and number of perfused mucosal capillaries could be observed. In 1996 we investigated mucosal and serosal tissue oxygen supply in a pig model of short time endotoxaemia. In animals infused with E. Coli endotoxin via the superior mesenteric artery a significant decrease in mucosal tissue oxygen tension and mucosal microvascular haemoglobin oxygen saturation occurred when compared with control animals [9]. Despite the fact that 46.3% of measured mucosal P02 values ranged within 0-5 mmHg, representing tissue areas either already hypoxic or at the nadir of hypoxia, there were no differences in intestinal oxygen extraction ratio or mesenteric venous P02 between control and endotoxin animals. Histological analysis of mucosal tissue demonstrated heterogeneous mucosal injury presenting as tissue oedema, swelling of vascular endothelial cells, tissue haemorrhage, tissue infiltration with leukocytes and destruction of the villus architecture. In critically ill patients, most of them suffering from SIRS and SIRS with sepsis, we observed a significantly diminished reactive hyperaemia response in the forearm skin to arterial occlusion of 5 min in comparison to an age matched healthy control group [4]. Diminished reactive hyperaemia was not related to a specific diagnosis but correlated with degree of physiologic derangement.

Therefore, current data indicate that disturbed microcirculatory function is present under conditions of bacteraemia and endotoxaemia in animals and in humans. There may be significant differences with regard to severity of microcirculatory disturbances between different organs or even tissue compartments in a specific organ. Microcirculatory abnormalities may lead to oxygen extraction failure at least under conditions of reduced systemic oxygen supply.

Mechanisms of microcirculatory failure In patients with septic shock increasing demand of vasopressor agents and fluids to maintain adequate systemic perfusion pressure, development of pulmonary and peripheral oedema and deterioration of organ functions are suggestive of progressive alterations within the microcirculation. In addition abnormal-

Mechanism of Oxygen Extraction Defect in Septic Shock 59

ities in the rheologic properties of erythrocytes, leukocytes, the activation of leukocytes and the coagulation system may significantly add to microcirculatory failure. Microvascular abnormalities observed in experimental studies include arteriovenous shunts, inhomogeneity of organ blood flow and loss of capillary cross sectional area due to deposition of microaggregates.

Archie studied shunting of radionuclide labeled micro spheres by the systemic circulation in six body regions in endotoxic shock and peritonitis in dogs. He reported significant regional arterial-venous shunting in the splanchnic circulation in the peritonitis model and within the kidney in the endotoxin model [10].

Using in vivo videomicroscopy a redistribution of hepatic microvascular blood flow within the liver lobule with increased perfusion of certain microvascular segments and decreased perfusion of others was reported during bacteraemia in rats [11]. Using the same technology several authors observed intense vasoconstriction of mucosal villus arterioles in the small intestine during bacteraemia and endotoxaemia [12, 13]. Rai reported early vasoconstriction in the brain, lungs, heart, kidneys and gastrointestinal tract, followed by vasodilatation and stagnation of blood flow, particularly in the venous bed of lungs and heart in a dog endotoxin shock model [14]. Alterations in microvascular tone most likely result from uncontrolled release of a variety of vasoactive substances from endothelial cells, leukocytes and injured tissue [15, 16]. The concentration of an individual vasoactive agent may significantly differ within tissue regions and organ compartments and may account for heterogeneity in tissue blood flow observed in some experiments. This may also explain differences in severity of tissue injury observed in histological examinations of organs.

The development of tissue oedema may aggravate microcirculatory failure by increasing diffusion distances and by mechanical compression of capillaries and venules. Causes of oedema formation include protein leakage, separation of tight junctions between endothelial cells, loss of negative surface charges on the membrane of capillary endothelial cells and direct endothelial cell injury and disruption [17-19].

Multiple studies have demonstrated abnormalities in red blood cell deformability responsible for decreased red blood cell flux through capillary vessels within the microcirculation during sepsis and septic shock. Intravital microscopic examinations have shown restricted capillary red blood cell flux due to cell aggregation, vasoconstriction and use of longer alternate vascular circuits [20]. Marked conformational alteration in the cytoskeletal proteins of erythrocytes may explain, in part, the marked reduction in red blood cell deformability [21]. Current evidence suggests that cytokines and activated leukocytes trigger the pathophysiologic events responsible for decreased red blood cell deformability [21,22].

Activation of endothelial cells and leukocytes leading to leukocyte rolling and sticking on the intraluminal surface of venular and probably capillary en-


dothelium is reported to contribute to microcirculatory failure. The interaction of leukocytes and the endothelium is triggered by the expression of surface receptors on both the endothelial cell and the leukocyte and is mediated in part by low shear stress, release of reactive surface proteins, histamine, cytokines and arachidonic acid metabolites [23]. It has been calculated that leukocyte activation may increase microvascular flow resistance by approximately 15% [24].

Finally the activation of the coagulation system may significantly alter microvascular blood flow [25]. Disseminated fibrin depositions in the microcirculation of various organ systems leading to microvascular thrombosis is a frequent finding in sepsis models and in patients suffering from septic shock. This microcirculatory damage seems to be especially pronounced in the lungs and kidneys. The mechanisms of disseminated intravascular coagulation are complex and involve direct activation of factor XII, e.g. by endotoxin, activation of platelets with subsequent procoagulant release, direct endothelial damage with exposure of clot-promoting surfaces, release of granulocyte procoagulant material and release of tissue factor [25].

Therapeutic strategies to improve microcirculatory failure

During the last years clinicians tried to improve microcirculatory blood flow and thus tissue oxygen supply by pushing the cardiovascular system with catecholamines in order to achieve supranormal values of cardiac index and systemic oxygen delivery [26, 27]. It was believed that this strategy will decrease incidence and severity of organ failure and subsequent mortality in high risk surgical patients, patients with ARDS, sepsis and septic shock. However, recent studies have cast doubt on the effectiveness of this strategy and one study reported increased mortality and morbidity in patients treated to achieve supranormal oxygen transport goals [28, 29]. Nevertheless, use of catecholamines may increase tissue oxygen supply in specific organs [30, 31]. In addition, catecholamines and phosphodiesterase inhibitors may exert antiinflammatory actions and may interfere with endothelial leukocyte interaction, thus reducing microcirculatory damage [32, 33].

In an endotoxin model we observed that DA-J agonists dopamine and dopexamine improved mucosal oxygen supply in the small intestine in pigs [30]. However, intravenous application of dobutamine failed to improve mucosal tissue oxygen tension, despite a significant increase in systemic oxygen delivery. In contrast, in patients with sepsis and septic shock dobutamine has been reported to increase gastric blood flow and pHi [31]. In vitro experiments have shown that the phosphodiesterase inhibitor amrinone decreases endothelial cell expression of the leukocyte adhesion molecules ICAM-I and E-selectin at therapeutic concentrations [32]. Dopamine inhibited endothelial E-selectin expression in human umbilical endothelial cells exposed to human interleukin-l ~ [32].


Agents improving rheological properties of blood cells or preventing uncontrolled coagulation may exert beneficial effects on the microcirculation during sepsis and septic shock [34-36]. The 21-aminosteroid tirilazad mesylate has been shown to inhibit leukocyte adhesion and emigration in postcapillary venules of the mesenteric microcirculation during endotoxaemia [34]. Pentoxifylline decreases platelet aggregation and improves filterability of whole blood and polymorphonuclear leukocytes in in-vitro studies [35]. Treatment of animals with a polyclonal antibody against tissue factor significantly attenuated tissue factor mediated coagulation and improved survival in a murine endotoxin shock model [36].

In animal experiments multiple interventions are able to improve microcirculatory function by affecting vasoregulation, rheologic properties of blood cells, by modulating the interaction of the endothelium with leukocytes or by interacting with the coagulation system.

Conclusion During sepsis and septic shock microcirculatory abnormalities have repeatedly been demonstrated in animal models and in humans. In addition, in animal models a significant oxygen extraction defect, especially within the gastrointestinal tract, has been demonstrated. However, the clinical significance of microcirculatory disturbances for the development of organ injury is still a matter of debate. The mechanisms of microcirculatory failure are complex and involve the dysregulation of microvascular tone, activation of the coagulation system and decrease in rheologic properties of blood cells. In animal models various pharmacological interventions improve microcirculatory failure. However, differences within species in response to pharmacological intervention, lack of monitoring systems of regional tissue oxygen supply still prevent rational clinical application.


References I. Hotchkiss RS, Karl IE (1992) Reevaluation of the role of cellular hypoxia and bioenergetic

failure in sepsis. JAMA 267: 1503-1510 2. Sair M, Etherington PJ, Curzen NP et al (1996) Tissue oxygenation and perfusion in endotox

emia. Am J PhysioI271:HI620-HI625 3. Garrison RN, Spain DA, Wilson MA et al (1998) Microvascular changes explain the "two

hit" theory of multiple organ failure. Ann Surg 227:851-860 4. Haisjackl M, Hasibeder W, Klaunzer S et al (1990) Diminished reactive hyperemia in the skin

of critically ill patients. Crit Care Med 18:813-818 5. Nelson DP, Beyer C, Samsel RW et al (1987) Pathological supply dependence of O2 uptake

during bacteremia in dogs. J Appl Physiol 63: 1487-1492 6. Samsel RW, Nelson DP, Sanders WM et al (1988) Effect of endotoxin on systemic and skele

tal muscle O2 extraction. J Appl Physiol65: 1377-1382 7. Nelson DP, Samsel RW, Wood LD et al (1988) Pathological supply dependence of systemic

and intestinal O2 uptake during endotoxemia. J Appl Physiol 64:2410-2419 8. Drazenovic R, Samsel RW, Wylam ME et al (1992) Regulation of perfused capillary density

in canine intestinal mucosa during endotoxemia. J Appl Physiol 72:259-265 9. Hasibeder W, Germann R, Wolf HJ et al (1996) The effects of short-term endotoxemia and

dopamine on mucosal oxygenation in the porcine jejunum. Am J Physiol 270:G667 -G675 10. Archie JP (J 977) Anatomic arterial-venous shunting in endotoxic and septic shock in dogs.

AnnSurg 186:171-176 II. Unger LS, Cryer HM, Garrison RN (1989) Differential response of the microvasculature in

the liver during bacteremia. Circ Shock 29:335-344 12. Whitworth PW, Cryer HM, Garrison RN et al (1989) Hypoperfusion of the intestinal micro

circulation without decreased cardiac output during live Escherichia coli sepsis in rats. Circ Shock 27: 111-122

13. Schmidt H, Secchi A, Wellmann R et al (1996) Effect of endotoxemia on intestinal villus microcirculation in rats. J Surg Res 61:521-526

14. Rai DK, Gupta LP, Singh RH et al (1974) A study of microcirculation in endotoxin shock. Surg Gynecol Obstet 139: 11-16

15. Tepperman BL, Brown JF, Whittle BJ (1993) Nitric oxide synthetase induction and intestinal epithelial cell viability in rats. Am J Physiol 265:G214-G218

16. Baudry N, Rasetti C, Vicaut E (1996) Differences between cytokine effects in the microcirculation of the rat. Am J Physiol 271 :HI186-H1192

17. Schutzer KM, Larsson A, Risberg B (1993) Lung protein leakage in feline septic shock. Am Rev Respir Dis 147: 1380-1385

18. Gotloib L, Shostak A, Galdi P et al (1992) Loss of microvascular negative charges accompanied by interstitial edema in septic rats' heart. Circ Shock 36:45-56

19. Hinshaw LB (1996) Sepsis/septic shock: participation of the microcirculation: an abbreviated review. Crit Care Med 24: I 072-1 078

20. Baker CH, Wilmoth FR, Sutton ET (1986) Reduced RBC versus plasma microvascular flow due to endotoxin. Circ Shock 20: 127 -139

21. Bellary SS, Anderson KW, Arden WA (J 995) Effect of lipopolysaccharide on the physical conformation of the erythrocyte cytoskeletal proteins. Life Sci 56:91-98

22. Betticher DC, Keller H, Maly FE (1993) The effect of endotoxin and tumour necrosis factor on erythrocyte and leucocyte deformability in vitro. Bf J Haematol 83: 130-137

23. Bienvenu K, Granger DN (1993) Molecular determinants of shear rate-dependent leukocyte adhesion in postcapillary venules. Am J Physiol 264:H 1504-H 1508

24. Harris AG, Skalak TC (1993) Effects of leukocyte activation on capillary hemodynamics in skeletal muscle. Am J Physiol 264:H909-H916

25. Mammen EF (1998) The haematological manifestations of sepsis . .J Antimicrob Chemother 41:17-24


26. Shoemaker WC, Appel PL, Kram HB et al (1988) Prospective trial of supranormal values of survivors as therapeutic goals in high-risk surgical patients. Chest 94: 1176-1186

27. Tuchschmidt J, Fried J, Astiz M et al (1992) Elevation of cardiac output and oxygen delivery improves outcome in septic shock. Chest 102:216-220

28. Gattinoni L, Brazzi L, Pelosi P et al (1995) A trial of goal-oriented hemodynamic therapy in critically ill patients. N Engl J Med 333: 1025-1032

29. Hayes MA, Timmins AC, Yau EHS et al (1994) Elevation of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med 330: 1717 -1722

30. Germann R, Haisjackl M, Schwarz Bet al (1997) Inotropic treatment and intestinal mucosal tissue oxygenation in a model of porcine endotoxemia. Crit Care Med 25: 1191-1197

31. Silva E, DeBacker 0, Creteur J et al (1998) Effects of vasoactive drugs on gastric intramucosal pH. Crit Care Med 26: 1749-1758

32. Fortenberry JD, Huber AR, Owens ML (1997) Inotropes inhibit endothelial cell surface adhesion molecules induced by interleukin-l beta. Crit Care Med 25:303-308

33. Tighe 0, Moss R, Haywood G et al (1993) Dopexamine hydrochloride maintains portal blood flow and attenuates hepatic ultrastructural changes in a porcine peritonitis model of multiple system organ failure. Circ Shock 39: 199-206

34. Schmidt H, Schmidt W, Muller T et al (1997) Effect of the 21-aminosteroid tirilazad mesylate on leukocyte adhesion and macromolecular leakage during endotoxemia. Surgery 121: 328-334

35. Ward A, Clissold SP (1987) Pentoxifylline. A review of its pharmacodynamic and pharmacokinetic properties, and its therapeutic efficacy. Drugs 34:50-97

36. Dackiw AP, McGilvray 10, Woodside M (1996) Prevention of endotoxin-induced mortality by antitissue factor immunization. Arch Surg 131: 1273-1278

ORGAN DYSFUNCTION AND BIOHUMORAL MISMATCH IN SEPSIS

Gut Perfusion in Sepsis and Shock

J. F. P ALIZAS

The splanchnic circulation contains about 30% of total blood volume and receives about 30% of the cardiac output. These proportions reveal the great influence the splanchnic vasculature and its regulation have on the systemic vascular behaviour in normal physiological conditions and even more so in states of haemodynamic instability and shock [I].

In fact, a typical response of the splanchnic circulation to different situations, characterised by hypovolaemic and/or low cardiac output states, is an intense vascular spasm disproportionate to that of the systemic vasculature. This vasoconstriction results in an increase of mesenteric arterial resistance and a decrease in mesenteric flow two to five fold greater than that in the peripheral circulation. These changes reflect selective splanchnic vasospasm.

This profound splanchnic vasoconstriction is mediated by an increase in sympathetic nervous system activity, vasopressin and activation of the renin-angiotensin system [2, 3].

The distinctive behaviour of splanchnic circulation is also characterised by its precocity. Splanchnic vasospasm is one of the first phenomena that can be seen when hypovolaemic or cardiogenic shock occurs. In septic states, although pinan early charge, the behaviour of the splanchnic vasculature is quite different. Vasodilatation and increase in blood flow are common findings. However, even though blood flow to the whole intestine may be increased, mucosal flow is in most cases impaired with subsequent mucosal ischaemia [4].

As has already been stated, intestinal ischaemia is a precocious manifestation of different forms of haemodynamic disturbances. Some authors have christened this phenomenon as the "canary" of the body (D. Dantzker), resembling the employment of canaries in coal mines to detect toxic gases before any harm could occur to miners. Therefore, detection of intestinal ischaemia would be of great importance in prognosis and probably in treatment in the critical care setting.

Measurement of pHi Gastric tonometry is a method designed to monitor gut mucosa oxygenation status.

68 J.F. Palizas

Experiments published several years ago [5, 6] showed that gas tension in the lumen of a hollow viscus (alimentary tract, gall bladder, and urinary bladder) is the same as that in the superficial layers of their mucosa. This knowledge was employed for the first time in 1959 by Boda et al. [7]. The authors used gastric tonometry to estimate arterial PC02, measuring PC02 in saline maintained for a certain equilibration period in a gastric balloon. They found a close correlation between values of gastric PC02 and end tidal PC02 with the exception of the values in patients with "haemodynamic instability"!? In them, tonometered PC02 was much higher than end tidal PC02 values.

Based upon information available [8, 9], it can be considered that PC02 in gastric air or fluid is in eqUilibrium with PC02 in the internal lining of the stomach (gastric mucosa). If luminal PC02 is measured by tonometry and mucosal bicarbonate concentration estimated (assuming that bicarbonate concentrations in mucosa and arterial blood are the same), the intramucosal pH (pHi) can be calculated using the Henderson-Hasselbalch equation.

To measure gastric tonometered PC02 (PtC02), a gastric tonometer is necessary. Although gastric juice has been used [10] further validation is needed. A device commonly used consists of a modified nasogastric tube (TRIP-NGS catheter, Tonometrics Inc, Worcester, MA, USA). It has two common sumping ports and a third one, which is connected through a gas impermeable tube to a silicone balloon placed on the end of the nasogastric device. This third port allows the balloon to be filled with saline solution and for samples to be taken for PtC02 measurements. The balloon is gas permeable and allows a time dependent equilibration of PC02 between saline solution and surrounding tissue and fluid.

The first step towards getting reliable pHi measurements is the X-ray confirmation of the balloon's position in the lumen of the stomach. Then, 2.5 cc of 0.9% saline solution should be injected into the tonometer. After an equilibration period (not less than 20 minutes), the saline solution is sampled, having discarded the first 1 cc that fills the dead space of the tube. The PC02 of saline sampled is measured in a blood gas machine together with a simultaneous arterial blood sample. Finally, pHi is calculated as follows:

pHi = 6.1 + log [arterial -HCOi(F * 0.03 * tonometered saline PC02)]

(Where F is a time dependent factor for fully or even partially equilibrated samples and is provided by the manufacturer).

One theoretical concern about the tonometric pHi calculation is the assumption that bicarbonate concentration in gastric mucosa is in equilibrium with bicarbonate concentration in arterial blood. However, the eqUilibrium is dependent on dynamic conditions and concentration differences could be related to the rate of change. For instance, fast administration of sodium bicarbonate could invalidate tonometric pHi calculation because bicarbonate equilibration between blood and gastric mucosa takes more time than that employed to make the measurement [11-13].

Gut Perfusion in Sepsis and Shock 69

Nevertheless, animal studies have validated tonometry. There is a very good correlation between measurements of intramucosal pH made by tonometry and those directly via electrodes in different situations, including endotoxic shock. In low or non-flow states the correlation is not so good and tonometer values tend to underestimate the fall in pHi. However, the direction of changes in pHi is predictable and reproducible [9].

The meaning of gastric intramucosal acidosis Some statements about the meaning of pHi are well supported by experimental and clinical evidence. First of all, tonometered measured pHi is a reliable reflection of tissue pH and a good method for monitoring regional tissue oxygenation. In the second place, pHi has been shown to be an important prognostic tool in terms of complications and outcome in critically ill and high-risk surgical patients. Last but not least, pHi normalisation is a therapeutic goal that could contribute to reducing ICU mortality.

However, considerable uncertainties still remain about the actual meaning of intramucosal acidosis, particularly in the course of severe sepsis and septic shock. The underlying mechanism more frequently advocated to explain intramucosal acidosis is tissue hypoxia. Decrements in oxygen transport unable to maintain systemic oxygen consumption may compromise mucosal oxygenation. As was mentioned before (Fig. 1), hydrolysis of adenosine triphosphate without complete resynthesis generates hydrogen ions (H+). These protons are reutilised in mitochondrial respiration under aerobic conditions. During anaerobic metabolism this process is strongly inhibited and proton accumulation results in cellular metabolic acidosis. Neutralisation of hydrogen ions by interstitial C03Hgenerates CO2, which can then be measured by tonometry.

There is good experimental correlation between changes in gut oxygen uptake and the development of intramucosal acidosis. Grum et al. [14] showed that reductions in oxygen delivery to the intestine due to hypoxia, ischemia or a combination of both, cause a fall in pHi only below a critical point at which oxygen uptake becomes supply dependent.

It has also been shown that hypoperfusion results in CO2 accumulation in tissues and in the venous circulation [15]. Reductions in cardiac output are followed by progressive increases in venous-arterial CO2 gradient due to a diminished clearance of CO2. This mechanism explains the mixed venous respiratory acidosis of low cardiac output states. If the decrease in blood flow is restricted selectively to the mucosa, selective intramucosal acidosis will develop because of reduced elimination of CO2. This increase in CO2 levels could be aerobically produced in the absence of tissue hypoxia.

Unpublished data extracted from experiments in our laboratory show that upper mesenteric artery blood flow was reduced by haemorrhage in a group of

70

650

500

350

200

mUMin/M2

002

Blood Loss +

• • __ 0 -.~:- ~----- --- -- -- - --- -- --- ---- ----1---- --- -- ------ -- ;,----- --- -- -----~~. pHi --- --- . + . " " -----

1. F. Palizas

pHi 7.5

7.4

7,3

7.2

V02 * ~1

50

Preop.

* . *

Surgery

Fig. 1.

* *

7

leu

dogs. The logarithmic increase in the gradient of CO2 concentration between venous gut drainage and jejunal mucosa indicates that the effects of low blood flow are progressively greater in the mucosa layer than in the rest of the gut

It is not possible to define clearly the relative impacts that tissue hypoxia, metabolic alterations, and reduced clearance of aerobically generated CO2 have on measured pHi. Nevertheless, pHi behaviour signals regional changes that the monitoring of systemic oxygenation parameters is unable to detect

Low pHi is a characteristic finding in septic patients and experimental endotoxaemia, Evidence coming from animal experiments supports the presence of alterations in gut perfusion in sepsis, both in macro and microcirculation. Vallet et al. [16] showed that hyperdynamic endotoxaemia in dogs produces decreased intestinal oxygen uptake, low mucosal P02 and intramucosal acidosis, The occurrence of this phenomenon despite high blood flow and normal serosal P02 suggests disturbed distribution of microvascular blood flow with regional tissue hypoxia in the mucosal layer, On the other hand, in another study [171, reductions in trans mesenteric oxygen uptake could not be documented; however, intramucosal acidosis was present Van der Meer et al. [18] showed that endotoxaemia in pigs causes ileal mucosal acidosis in the absence of mucosal hypoxia


and without changes in mucosal perfusion. Actually, if intramucosal acidosis develops without mucosal hypoxia and ischemia, other mechanisms could be responsible.

Different investigators have shown different alterations in intermediary metabolism in sepsis. These alterations include derangement in mitochondrial respiration [19] and inactivation of pyruvate dehydrogenase [20]. Both defective mechanisms could induce mucosal acidosis in sepsis unrelated to low tissue P02.

Based upon the information already shown, pHi behaviour in low and normal or high flow states (sepsis syndrome) appears to be quite different.

In hypovolaemia and cardiogenic shock a selective decrease in mesenteric blood flow and increase in mesenteric arterial resistance have been shown. These variations in flow and resistance are secondary to a selective action of neurogenic stimulation and vasoconstrictor agents (renin-angiotensin system) on mesenteric vasculature. This vasoconstriction is two- to five-fold greater than the rise in total peripheral resistance, thus reflecting selective splanchnic vasospasm. This drop in mesenteric flow is followed very rapidly by a drop in intestinal pHi; the restoration of normal flow conditions returns pHi values to normal. Fig. I shows the evolution of a high-risk surgical patient during anaesthesia monitoring. A drop in cardiac output secondary to blood loss is swiftly followed by a drop in pHi, which returns to normal after plasma volume expansion.

On the other hand, a steep drop in pHi and increases in cardiac output are common in patients with severe sepsis or septic shock. Possible causative mechanisms of gut mucosal acidosis without local hypoxia have been already discussed. However other questions arise: What is the treatment to be employed to raise pHi back to normal? Is it reasonable to push cardiac output even higher? Is it worthy doing? Is pHi drop in sepsis with high cardiac output a warning indicating that an infectious disease is not under control? If this is true, is a low pHi an indication to change to different antibiotics or to perform surgical drainage of an occult focus?

Available evidence is not adequate to answer to all these questions. It appears that a drop in pHi in high output sepsis warns the clinical practitioner that the patient is not doing well [21, 22], that a careful evaluation of treatment is mandatory and that a decision aimed at raising cardiac output is probably justified[23,24].

In Fig. 2, a severely septic patient under anaesthesia monitoring is shown (unpublished observation). During surgical drainage of acute cholecystitis it could be seen that a deep drop in pHi occurred at the same time as an acute increase in cardiac output. Immediately after surgery, overt septic shock developed; after blood pressure had been normalised by volume expansion and dopamine infusion, further increases of the dopamine infusion rate (up to 40 Ilg/kg/min) were needed to raise the pHi back to normal levels.

mL/MinjM2 PHt 7.5

10SO t pHi Septic

+ \ Shock 7.4

900 /

! 7.3

7SO \ + 7.2

soo + \

-+- \ + - + 7.1

+ 450 V02 7

30 42

300 * • - * 6.9

* • • • * • • • 150 6.8

Preop. Surgery leu


cal ICU patients. Doglio et al. [21] were able to determine the effect of therapeutic interventions in the early evolution of gastric pHi. They found important differences in patient mortality according to the admission and 12 hour gastric pHi values. Patients in whom therapeutic manoeuvres failed to correct a low admission pHi had the greatest mortality rate (86.7%), whereas survival improved (mortality rate 36.4%) in those patients whose gastric mucosal pH returned to normal during the first 12 hours of their ICU stay. Those patients with normal pHi on admission and again 12 hours later had the lowest mortality rate (28.6%). The authors stated "the greater survival noted in patients with normal gastric intramucosal pH during the first 12 hours in the ICU makes a strong argument for using gastric intramucosal pH as a monitor of local tissue hypoxia".

The association of survival with increases in gastric pHi showed by Doglio et al. suggested that gastric pHi could become an index to guide therapeutic interventions in critically ill patients.

Consequently, Gutierrez et al. [24] designed a multicentre trial to evaluate the efficacy of a resuscitation protocol which had the therapeutic goal of achieving a normal gastric pHi (> 7.35). Two hundred and sixty critically ill patients, with Apache II scores between 15 to 25 were randomised into control or active treatment groups. Control and active treatment groups were divided into two additional groups, according to gastric pHi on admission. Thus, there were four groups after randomisation: a) control and b) active treatment, normal admission gastric pHi and c) control and d) active treatment, low admission gastric pHi. Control groups were resuscitated using conventional clinical and haemodynamic parameters as therapeutic goals. In the active treatment groups, if gastric pHi was lower than 7.35 after conventional resuscitation, a special protocol of volume expansion and dobutamine infusion aimed at raising gastric pHi to normal was followed.

In patients admitted to the ICU with normal (> 7.35) gastric pHi the outcome was significantly improved when the active resuscitation protocol was followed. As a drop in gastric pHi was seen in about 90% of these patients, the improvement in outcome could be explained by the early recognition of intramucosal acidosis followed by its prompt reversal by increases in global oxygen delivery. However, in patients admitted with a low gastric pHi « 7.35), further increases in oxygen delivery did not improve the outcome in the active treatment group. These findings could be explained assuming that intramucosal acidosis in these groups (low pHi on admission) had been present for a prolonged period before admission, long enough to result in secondary irreversible tissue damage in some of the patients. It is also possible that the protocol used to increase oxygen delivery was not appropriate to solve the severe metabolic disturbances of these patients.

In summary, it appears that early recognition of gastric mucosal acidosis in critically ill patients and its rapid reversal result in an encouraging improvement in outcome.

74 J.E Palizas

In a recent publication, My then and Webb [30] used plasma volume expansion to preserve gut mucosal perfusion during elective cardiac surgery in a comparative, randornised active vs. control protocol. The incidence of gut mucosal hypoperfusion at the end of surgery was reduced in the active treatment group, as were the number of patients in whom major complications developed, mean number of days spent in the intensive care unit and mean number of days spent in the hospital.

Recently, our group organised a multicentre protocol to compare the efficacy of pHi against cardiac index as a therapeutic goal in patients with severe sepsis and septic shock needing the insertion of a pulmonary artery catheter. A pHi of 7.32 was compared with a cardiac index of 3 litfminlm2 as therapeutic goals of fluid and vasoactive drug resuscitation manoeuvres. The unpublished preliminary data show two main results:

a) prognOSIS:

- 14 out of 16 non-survivors had pHi values < 7.32 (therapeutic goal not achieved) after 24 hours of treatment aimed to increase cardiac output;

14 out of 18 non-survivors had cardiac index> 3 litfminlm2 (therapeutic goal achieved) after 24 hours of treatment aimed to increase cardiac output;

b) pulmonary artery catheter days:

the pulmonary artery catheter remained inserted for a shorter time in patients monitored by a gastric tonometer (3.85 ± 2.06 days vs. 5.37 ± 4.06 days).

Conclusion After the first years of basic and clinical research using tonometric monitoring of gut mucosa, it can be concluded that this technique appears to be a useful tool for assessing mesenteric oxygenation status.

It is still not known whether it will become a major advance in the evaluation and treatment of critically ill patients. What seems to be clear is that tonometry has led clinicians, for the first time, to think about oxygenation from a different point of view.


References 1. Banks RO, Gallavan RH, Zinner MJ et al (1985) Vasoactive control on the mesenteric circula

tion. Fed Proc 44:2743-2749 2. McNeill JR (1970) Intestinal vasoconstriction after hemorrhage: roles of vasopressin and an

giotensin. Am J PhysioI219:1342-1347 3. Bailey RW, Oshima A, O'Roark WA, Bulkley GB (1987) A reproducible, quantitatable and

rapidly reversible model of cardiogenic shock in swine. In: Tumbleson ME (ed) Swine in Biomedical Research. Plenum, New York, pp 363-372

4. Systemic and splanchnic hemodynamic derangement in the sepsis syndrome (1989) In: Marston A, Bulkley GB, Fiddian-Green RG, Haglund Vh (ed) Splanchnic Ischemia and Multiple Organ Failure. Mosby, St. Louis, pp 10 1-1 06

5. Bergofsky EH (1964) Determination of tissue oxygen tension by hollow visceral tonometers: effects of breathing enriched oxygen mixtures. J Clin Invest 43: 193-200

6. Dawson AM, Trencharrh D, Guz A (1965) Small bowel tonometry: Assessment of small gut mucosal oxygen tension in dog and man. Nature 206:943-944

7. Boda D, Muninyi L (1959) Gastrotonometry. An aid to the control of ventilation during artificial respiration. Lancet Vol I: 181-182

8. Fiddian-Green RG, Pittenger G, Whitehouse WM (1982) Back-diffusion of CO2 and its influence on the intramural pH in gastric mucosa. J Surg Res 33:39-48

9. Antonsson JB, Boyle CC, Kruithoff KL et al (1990) Validation of tonometric measurement of gut intramucosal pH during endotoxaemia and mesenteric occlusion in pigs. Am J Physiol 259:G519-523

10. Gys T, Hubens A, Neels H et al (1988) The prognostic value of gastric intramural pH in surgical intensive care patients. Crit Care Med 16: 1222-1224

11. Takala JV, Parviainen I, Silohao M et al (1994) Saline PCOz is an important source of error in the assessment of gastric intramucosal pH. Crit Care Med 22: 1877-1879

12. Heard SO, Helmsmoortel CM, Kent JC et al (1991) Gastric tonometry in healthy volunteers: effects of ranitidine on calculated intramural pH. Crit Care Med 19:271-274

13. Maynard N, Atkinson S, Mason R et al (1994) Influence of intravenous ranitidine on gastric intramucosal pH in critically ill patients. Crit Care Med 22:A 79

14. Grum CM, Fiddian-Green RG, Pittenger GL et al (1984) Adequacy of tissue oxygenation in intact dog intestine. J Appl Physiol 56: 1065-1069

15. Weil MH, Rackow EC, Trevino R et al (1986) Difference in acid-base state between venous and arterial blood during cardiopulmonary resuscitation. N Engl J Med 315: 153-156

16. Vallet B, Lund N, Curtis SE et al (1994) Gut and muscle p02 in endotoxemic dogs during shock and resuscitation. J Appl Physiol 76:793-800

17. Antonsson JB, Engstrom L, Rasmussenn I et al (1995) Changes in gut intramucosal pH and gut oxygen extraction ratio in a porcine model of peritonitis and hemorrhage. Crit Care Med 23: 1872-1881

18. Van der Meer TJ, Wang H, Fink MP (1995) Endotoxemia causes ileal mucosal acidosis in the absence of mucosal hypoxia in a normodynamic porcine model of septic shock. Crit Care Med 23: 1217-1226

19. Mela L, Bacalzo LV, Miller LD (1971) Defective oxidative metabolism of rat liver mitochondria in hemorrhagic and endotoxic shock. Am J PhysioI220:571-576

20. Vary TC, Siegel JH, Nakatani T et al (1986) Effect of sepsis on activity of PDH complex in skeletal muscle and liver. Am J Physiol 250:E634-640

21. Doglio GR, Pusajo JF, Egurrola MA et al (1991) Gastric mucosal pH as a prognostic index of mortality in critically ill patients. Crit Care Med 19: 1037-1040

22. Marik PE (1993) Gastric intramucosal pH: A better predictor of multiorgan dysfunction syndrome and death than oxygen-derived variables in patients with sepsis. Chest 104:225-229

23. Gutierrez G, Clark C, Brown SD et al (1994) Effects of dobutamine on oxygen consumption and gastric mucosal pH in septic patients. Am J Respir Crit Care Med 150:324-329

76 J.F. Palizas

24. Gutierrez G, Palizas F, Doglio G et al (1992) Gastric intramucosal pH as a therapeutic index of tissue oxygenation in critically ill patients. Lancet 339: 195-199

25. Fiddian-Green RG, Amelin PM, Herrmann JB et al (1986) Prediction of the development of sigmoid ischemia on the day of aortic operations. Arch Surg 121 :654-660

26. Bjorck M, Hedberg B (1994) Early detection of major complications after abdominal aortic surgery: Predictive value of sigmoid colon and gastric intramucosal pH monitoring. Br J of Surg 81:25-30

27. Fiddian-Green RG, Baker S (1987) Predictive value of measurement of the stomach wall pH for complications after cardiac operations: Comparison with other monitoring. Crit Care Med 15:153-156

28. Roumen RMH, Vreugde JPC, Goris RJA (1994) Gastric tonometry in multiple trauma patients. J Trauma 36:313-316

29. Mohsenifar Z, Hay A, Hay Jet al (1993) Gastric intramural pH as a predictor of success or failure in weaning patients from mechanical ventilation. Ann Intern Med 119:794-798

30. My then MG, Webb AR (1995) Peri operative plasma volume expansion reduces the incidence of gut mucosal hypoperfusion during cardiac surgery. Arch Surg 130:423-429

Pathophysiology of Encephalopathy

N. LATRONICO, G.F. BussI, A. CANDIANI

The metabolic complexity of the central nervous system (CNS) makes it dependent upon the functional integrity of other body systems for the adequate provision of essential nutrients and elimination of toxins. It is therefore not surprising that various metabolic effects on the CNS are secondary to systemic diseases. These are situations in which a diffuse brain malfunction is clinically evident, despite the evidence of structural brain alteration is lacking. It is only when the metabolic disorder has been profound that structural changes occur, thus accounting for the permanent neurological deficits that some patients exhibit.

Attention and cognitive functions such as perception, thinking and memory are affected early. Alertness tends to fluctuate between agitation and lethargy, but in severe cases stupor and coma may ensue. Noteworthy, focal motor signs, abnormalities of pupillary reactivity and of ocular movements are exceedingly rare even in deep coma [1].

Numerous endogenous conditions, including fluid and electrolytes disorders, hypoglycaemia, diabetes and pancreatitis, liver and renal failure, cancer, hyperand hypothermia, nutritional and hypoxic disorders may be responsible for a metabolic encephalopathy in the critically ill patient [1, 2].

Septic encephalopathy (SE) defines a clinical picture characterised by the impairment of the alertness and cognitive function, which is described in patients with sepsis. A structural brain alteration is not evident on neuroradiological investigations, and the clinical signs may be fully reversible, provided the systemic infection is timely dominated and complications are avoided. So defined, the SE fulfils the definition of metabolic encephalopathy, although other authors classify it among the inflammatory CNS disorders [1]. Systemic inflammation has a key role in promoting the multiple organ dysfunction syndrome and a systemic inflammatory response syndrome, either due to infection (sepsis) or not, is a recognised [3-5] although still a debated entity [6]. However, local signs of inflammation within the CNS are lacking [7].

78 N. Latronico, G.F. Bussi, A. Candiani

The neuronal function, not strictly the neuron, is involved The loss of the highest integrative cerebral functions in SE is compatible with a diffuse dysfunction of neurons. Also the fact that electroencephalogram (EEG) correlates well with the severity of SE [8] points to the functional involvement of the neuron, the EEG-generating cell. The available evidence indicates that the major determinants of the neuronal function - the cerebral blood flow, the blood-brain barrier, the cerebral environment and the neuron itself - are altered in SE.

The supply of oxygen and nutrients Cerebral blood flow (CBF) is reduced in dogs [9] and humans [10, 11] with severe sepsis. The cerebral metabolic rate of oxygen (CMR02) is increased in dogs probably as a consequence of gross endotoxin blood-brain barrier damage and passage of hydrogen ions and catecholamines into the brain. This imbalance between demand (increased) and supply (reduced) makes the brain vulnerable to hypoxia, and can be prevented by propanolol pretreatment [9]. In humans, the CMR02 seems reduced, however only one study actually measured it in 6 patients [11]. In this study, decreased oxygen consumption was paralleled by EEG slowing, suggesting normal coupling between metabolic activity and blood flow. Microthrombosis is also reported [1], however this seems a finding of advanced, complicated disease, while encephalopathy is an early event [12].

Wijdicks and Stevens in a retrospective series of 84 patients found that the development of SE was associated with hypotension [13]. Two out of 14 patients developing SE had widespread ischaemic cortical damage on neuropathological examination. However, criteria for hypotension (timing, severity, duration) are not presented, and those for SE were only clinical, a low sensitive method compared to EEG [1, 8]. Eidelman et al. in a prospective study could not find an association between hypotension and SE, which was instead associated with increased mortality, bacteraemia, and renal and hepatic dysfunction [2].

Hypoxic-ischaemic damage by a number of causes is likely to aggravate an already altered brain. However, no convincing evidence supports its role in initiating SE.

The blood-brain barrier (BBB) A breakdown of the BBB has since long been postulated after studies showing an increased permeability with the passage of colloidal iron oxide [14], 14C_ amino acids [15], horseradish peroxidase [16] and 125I-albumin [17] in septic animals, as well as the increased cerebro-spinal fluid level of proteins in septic humans [8].

Pathophysiology of Encephalopathy 79

Recently, Papadopoulos et al. showed increased oedema around cortical microvessels, but not around larger vessels [18]. On the contrary, the endothelial cells appeared normal with morphologically intact mitochondria and intercellular tight junctions.

To summarise, the mechanism(s) by which BBB permeability is increased remains obscure: increased pinocytosis or molecular alteration of the occludins, the tight-junction constituent proteins, are possible explanation [18]. The perivascular oedema may affect the supply of oxygen and nutrients to the brain, and may contribute to the symptoms of SE.

The cerebral environment

Astrocytes are not just framework cells

The neuronal function is strictly dependent on the maintenance of appropriate chemical and physical cerebrospinal fluid (CSF) properties. For example, the reduction of CSF osmolality greatly increases the neuronal excitability, and seizure is a common complication of acute hyponatraemia [19]. The BBB has a central role in the maintenance of cerebral homeostasis, and so the astrocytes. However, it is now clear that astrocytes also have intense metabolic activity. Studies over the last 15 years have buried the image of astrocytes as "stodgy, merely supportive, and unresponsive ugly sister of neurons" [20]. Their importance in removing glutamate and aspartate, the two major excitatory amino acids (EAA), is crucial in preventing neurotoxicity [20, 21] (Fig. 1), a role that Lugaro had intuited at the beginning of this century:

"Every nervous termination suffers a chemical modification and this chemical modification in tum gives stimulus to another neuron. If this is true, the interneuronal articulation would be at the center of the chemical exchange, and therefore would comprise in all the most proximal, vacant interstitial spaces, a region for infiltration of the protoplasmatic prolongations of feathery extensions of the neuroglia [the astrocytes], perhaps with the purpose of collecting and instantly processing the smallest amount of waste product" (Lugaro, 1907) [22].

Astrocyte swelling was demonstrated by Clawson et al. in endotoxaemic rabbits [14]. This result has been recently confirmed in Papadopoulos' study, in which a substantial astrocyte damage with gross swelling of perivascular endfeet and their rupture and detachment from microvessel endothelial cell wall was clearly present [18]. Disruption of BBB is a likely explanation for these alterations [23], which in tum may profoundly alter CSF composition by increasing the extracellular glutamate concentration. This latter has been estimated to be approximately 0.6 fJ,mollL, and substantial excitotoxic damage is expected to occur when the glutamate concentration reaches 2 to 5 fJ,mollL. Since the astrocytes (and neurons) contain 10 mmol of glutamate per litre, the potential for disaster is obvious [21] (Fig. 1).

80 N. Latronico, G.P. Bussi, A. Candiani

Ammonia

Glutamate

Glutamate 0.6 lJllloVL

Glutamate

Ammonia

Fig. 1. Glutamate-glutamine shuttle: the presynaptic neuron (pre) stimulates the post-synaptic neuron (post) by means of glutamate. Extracellular glutamate concentration must be kept below 2-5 ~mol/L, and it is therefore aminated to glutamine by glutamine synthetase (GS) in the astrocyte (A). Glutamine is then partly deaminated to glutamate by glutaminase (GLM) to restore the neuronal pool, and partly conveyed to the blood. The glutamine passage across the blood-brain barrier (BBB) is coupled with entrance of neutral amino acids

In summary, the alteration of neuronal environment, particularly EAA cytotoxicity, is a conceivable operating mechanism in SE. To date, however, only indirect evidence is available to support this view.

Neurotransmitter imbalance

Aromatic (AAA: phenylalanine, tyrosin, tryptophan) and sulfur-containing amino acids (taurine, cysteine, methionine) are increased in plasma from septic animals and humans. Conversely, the branched chain amino acids (BCAA: valine, leucine, isoleucine) and y-aminobutyric acid (GABA) are reduced. This pattern is similar to that seen in hepatic encephalopathy, and is thought to be due to enhanced muscle proteolysis coupled with a hepatic failure to handle these breakdown products [15, 24-28].

AAA are precursors for brain neurotransmitters, and their brain uptake is increased during sepsis [26-28] (Fig. I). GABA is a major inhibitory neurotrans-


mitter in the CNS, and an increase in GABA-A receptors density has been demonstrated in septic rodents [29].

The result is a neurotransmitter imbalance with concomitant activation of the GAB A and serotoninergic inhibitory systems and inhibition of the catecholaminergic activating system [28-30] (Fig. 2).

BLOOD

BRAIN

Tryptophan

+~

Muscle ProteolYSitand liver dysfunction

Increased plasmatic levels of AAA

Phenylalanine

I Tyrosine

~l GABA Serotonin Pheny lethilamine, phenylethanolamine

Octopamine Dopamine

Norepinephrine

+ Inhibitory neurotransm itters

Tyramine

+ Weak (false) activating neurotransmitters

Sensorium depression

- Activating neurotransmitters

Impaired immunomediated systemic inflammation

Fig. 2. Alterations of plasma and brain aromatic amino acids (AAA, phenylalanine, tyrosin, tryptophan), and their supposed effect on consciousness [24-30) and immunity [31)

There is general agreement that plasma amino acids alterations are caused by subtle liver dysfunction, probably an early as well as underestimated problem due to the complexity of diagnostic approach [15, 25-28]. It is worth citing that the brain amino acids derangement, particularly the noradrenergic system, may in tum profoundly affect the immune-mediated systemic inflammatory response [31].

The alteration of plasma and brain amino acids profile may explain some of the consciousness alterations seen in SE patients, and BCAA infusion has been shown to restore brain amino acids and neurotransmitter profile towards normal. However, direct proof that BCAA ameliorate the symptoms of SE is lacking.

82 N. Latronico, G.F. Bussi, A. Candiani

The neuron In Papadopoulos' study dark, shrunken neurons were seen in the frontal cortex of pigs after 8 hours of sepsis, They were not present in control sham-operated animals, and therefore they were not an artifact [23], Furthermore, the cytoplasm of other neurons appeared dilute, The authors speculate that a cytotoxic event could be operating. Taken together with the described astrocyte alterations, it is possible that astrocytes fail to keep the extracellular excitatory amino acids concentration low. When excitotoxic levels are reached, cellular (neurons and astrocytes) necrosis ensues, thus liberating massive amount of excitatory amino acids, and creating a vicious, self-propagating circle.

As already cited, at present cytotoxicity is only a sound hypothesis still to be tested. Whatever the mechanism(s) of neuronal damage, the Papadopoulos' results are important, since they demonstrate that neurons are directly damaged in an early phase of sepsis.

Conclusions

The available evidence suggests that the blood-brain barrier, the brain environment, the neuron itself, and in advanced state the brain energy supply - that is, all the major determinants of the neuronal function - are altered in SE. Ultrastructural and functional investigations are needed to validate the on-going hypotheses and put them in a rank order.

References I. Young BG, Ropper AH, Bolton CF (eds) (1998) Coma and impaired consciousness. A clini

cal perspective. McGraw-Hill, New York 2. Eidelman LA, Putterman D, Putterman C, Sprung CL (1996) The spectrum of septic en

cephalopathy. Definitions, etiologies, and mortalities. JAMA 275:470-473 3. Bone RC, Balk RA, Cerra FB et al (1992) Definitions for sepsis and organ failure and guide

lines for the use of innovative therapies in sepsis. Crit Care Med 2:864-890 4. Latronico N, Fenzi F, Recupero D et al (1996) Critical illness myopathy and neuropathy.

Lancet 347: 1579-1582 5. Latronico N (1997) Acute myopathy of intensive care. Ann Neurol 42: 131-132 6. Vincent JL (1998) Search for effective immunomodulating strategies against sepsis. Lancet

351 :922-923 7. Perry VH, Andersson PB, Gordon S (1993) Macrophages and inflammation in the central

nervous system. Trends Neurosci 16:268-273 8. Young BG, Bolton CF, Archibald YM et al (1992) The electroencephalogram in sepsis-asso

ciated encephalopathy. J Clin Neurophysiol 9: 145-152 9. Westerlind A, Larsson LE, Haggendal J, Ekstrom-Jodal B (1991) Prevention of endotoxin-in

duced increase of cerebral oxygen consumption in dogs by propanoJol pretreatment. Acta Anaesthesiol Scand 35:745-749


10. Bowton DL, Bertels NH, Prough DS et al (1989) Cerebral blood flow is reduced in patients with sepsis syndrome. Crit Care Med 17:399-403

11. Maekawa T, Fujii Y, Sadamitsu D et al (1991) Cerebral circulation and metabolism in patients with septic encephalopathy. Am J Emerg Med 9: 139-143

12. Bolton CF, Young GB, Zochodne DW (1993) The neurological complications of sepsis. Ann NeuroI33:94-100

13. Wijdicks EFM, Stevens M (1992) The role of hypotension in septic encephalopathy following surgical procedures. Arch Neurol 49:653-656

14. Clawson CC, Hartmann JF, Vernier RL (1966) Electron microscopy of the effect of gramnegative endotoxin on the blood-brain barrier. J Comp Neurol 127: 183-198

15. Jeppson B, Freund HR, Gimmon Z et al (1981) Blood-brain barrier derangement in sepsis: cause of septic encephalopathy? Am J Surg 141: 136-141

16. du Moulin GC, Paterson D, Hedley-White J, Broitman SA (1985) E. coli peritonitis and bacteremia cause increased blood-brain barrier permeability. Brain Res 340:261-268

17. Deng X, Wang X, Andersson R (1995) Endothelial barrier resistance in multiple organs after septic and non septic challenges in the rat. J Appl Physiol 78:2052-2061

18. Papadopoulos MC, Lamb FJ, Moss RF et a1 (1999) Faecal peritonitis causes oedema and neuronal injury in pig cerebral cortex. Clin Sci 96:461-466

19. Latronico N, Zappa S, Antonini S, Candiani A (1995) Osmolalita e sistema nervoso centrale. In: Spandrio L (ed) L'osmola1ita. Laboratorio e clinica. Ed Sorbona, Milano, pp 113-128

20. Murphy S (ed) (1993) Astrocytes. Pharmacology and function. Academic Press, San Diego, CA 21. Lipton SA, Rosenberg PA (1994) Excitatory amino acids as a final common pathway for neu

rologic disorders. New Engl J Med 330:613-622 22. Lugaro E (1907) Sulle funzioni della neuroglia. Riv Patol Nerv Ment 12:225-233 23. Noremberg MD (1994) Astrocyte responses to CNS injury. J Neuropathol Exp Neurol 53:

213-220 24. James JH, Jeppson B, Ziparo V, Fisher JE (1979) Hyperammonaemia, plasma amino acid im

balance, and blood-brain amino acid transport: a unified theory of portal-systemic encephalopathy. Lancet 2:772-775

25. Hasselgren PO, Fisher JE (1986) Septic encephalopathy. Etiology and management. Intensive CareMed 12:13-16

26. Mizock BA, Sabelli HC, Dublin A et al (1990) Septic encephalopathy. Evidence for altered phenylalanine metabolism and comparison with hepatic encephalopathy. Arch Int Med 150: 443-449

27. Sprung CL, Cerra FB, Freund HR et al (1991) Amino acid alterations and encephalopathy in the sepsis syndrome. Crit Care Med 19:753-757

28. Freund HR, Muggia-Sullam M, LaFrance R et al (1986) Regional brain amino acid and neurotransmitter derangements during abdominal sepsis in the rat: the effect of amino acids infusions. Arch Surg 121 :209-216

29. Kadoi Y, Saito S (1996) An alteration in the y-aminobutyric acid receptor system in experimentally induced septic shock in rats. Crit Care Med 24:298-305

30. Soejima Y, Fujii Y, Ishikawa T et al (1990) Local cerebral glucose utilization in septic rats. Crit Care Med 18:423-427

31. Chrousos GP (1995) The hypothalamic-pituitary-adrenal axis and immune-mediated inflammation. New Eng1 J Med 332: 1351-1362

Lung Dysfunction in the Early Phase of Sepsis

P. NEUMANN

Sepsis is defined as the systemic inflammatory response to infection and is frequently complicated by organ dysfunction (severe sepsis according to the recent consensus definition) [1]. The lungs seem to be particularly vulnerable to the septic inflammatory response, and sepsis is the underlying cause in about 40% of all patients with acute respiratory failure [2]. Thus, sepsis is the most frequently encountered risk factor for the development of acute respiratory failure. However, only 30 to 40% of septic patients will eventually develop an acute respiratory distress syndrome (ARDS) [3, 4]. Other risk factors for lung dysfunction, such as pancreatitis and multiple trauma [5] are commonly associated with a systemic inflammatory response syndrome (SIRS) [1] which is clinically indistinguishable from sepsis, although no infectious agents can be isolated. SIRS demonstrates that lung dysfunction may develop primarily due to an immunological response rather than a direct pulmonary damage caused by an infectious agent. Experimentally, lung injury can be induced by injection or infusion of lipopolysaccharide components of gram-negative bacteria (endotoxin), and elevated plasma endotoxin levels have been shown to correlate with the development of ARDS in patients at risk for ARDS [6]. Endotoxin-induced lung injury has been studied extensively in different species in an attempt to gain insight into the pathogenesis of lung dysfunction in sepsis and in order to develop therapeutic strategies which may improve the poor prognosis of respiratory failure associated with sepsis [2]. It should be emphasised, however, that severe sepsis with lung dysfunction may also develop due to gram-positive, viral or fungal antigens.

Clinical parameters in endotoxin-induced lung injury Infusion of endotoxin may significantly impair gas exchange and cause structural changes within the lungs, but the severity of changes is modified by the type of endotoxin, dose administered and the species studied [7]. In pigs, intravenous administration of endotoxin from E. coli typically results in a prompt and dramatic increase of the pulmonary artery pressure (MPAP) with initially only moderate disturbances of gas exchange [8, 9]. Then, MPAP gradually declines

86 P. Neumann

but remains elevated above baseline levels throughout the following hours, while gas exchange tends to deteriorate further [8]. Extravascular lung water increases significantly in endotoxin-induced porcine lung injury [8, 10], but infusion of 10 f.lg • kg-I. h- I endotoxin did neither cause intraalveolar oedema nor macroatelectasis [8]. Thus, fluid accumulation was mainly restricted to the interstitial space with thickened alveolar walls, perivascular and peri bronchiolar cuffing and dilated lymph vessels [8]. In contrast, 4 h after an increased dose of 20 f.lg • kg-I. h- 1 endotoxin about 20-30% of lung parenchyma is collapsed or flooded in a gravity dependent manner [10], which is similar to what has been described in ARDS patients [ 11-13]. In line with these morphological findings, hypoxaemia was mainly caused by perfusion of poorly ventilated lung areas (Low V AlQ) when macroatelectasis was absent (after 10 f.lg • kg-I • h -I endotoxin) [14], which may be explained by loss of the hypoxic pulmonary vasoconstriction after endotoxin infusion [15]. In contrast, true shunt predominated when atelectasis was present with a dose of 20 f.lg • kg-I. h-I endotoxin, and six to eight hours after induction of lung injury respiratory failure was occasionally so severe that hypoxaemia developed despite the use of mechanical ventilation with PEEP and pure oxygen (unpublished data from our group).

Radiologically, endotoxin-induced lung injury complies with the recent ARDS definitions of bilateral infiltrates seen on a frontal chest radiograph [5]. However, the radiological appearance of other types of experimental lung injury like oleic acid injection or repeated lung lavage, which do not require an inflammatory response to damage the lungs [16, 17], is quite similar to the endotoxin model. Even with computed tomography, endotoxin-induced lung injury can not be distinguished by its appearance from lung injury caused by repeated lung lavage or oleic acid injection [10].

Pathophysiology The pulmonary response initiated by the administration of endotoxin has been divided into two phases [7, 18]:

The first hours of endotoxin infusion are characterised by the marked increase in mean pulmonary artery pressure and pulmonary vascular resistance (see above). Presumably because of the increased hydrostatic pressure gradient across the endothelium, pulmonary lymph flow with a low protein content increases and extravascular lung water accumulates [8, 19]. Pretreatment with a cyclooxygenase inhibitor like ibuprofen or indomethacin attenuates the pulmonary arterial hypertension, the degree of hypoxaemia as well as the increase of extravascular lung water [19, 20]. Thus, arachidonic acid metabolites like the potent vasoconstrictors TXA2 seem to mediate some of the early effects of lung failure in sepsis.

In the second phase (> 2-4 h after induction of lung injury), the protein content of pulmonary lymph increases due to increased endothelial permeability

Lung Dysfunction in the Early Phase of Sepsis 87

[15]. The accumulation of protein rich fluid in the interstitium increases the interstitial oncotic pressure and thus promotes further extravasation of fluid. If the alveolocapillary membrane is severely damaged, intraalveolar oedema may develop (see above). Flooding of alveoli, which is a central feature of ARDS [21], is likely to cause additional lung damage by surfactant dilution and inactivation. This increases alveolar surface tension and thereby facilitates lung collapse and transendothelial fluid flux. However, the administration of aerosolised surfactant in patients with sepsis-induced ARDS did not improve survival in a large prospective multicenter study [22].

Mediators of lung injury Accumulation of polymorphnuclear granulocytes in the pulmonary microcirculation is the most prominent finding after endotoxin infusion [8, 15]. Accumulation of lymphocytes, microvascular stasis, endothelial cell damage, a thickened alveolar interstitium due to interstitial oedema and finally intraalveolar oedema may also be present (for review see [15]).

In vitro studies have shown that endotoxin can damage endothelial cells directly [15], but numerous proinflammatory cells and substances, referred to as mediators, are usually involved during the pathogenesis of lung injury in vivo. As an early responses to endotoxin, human alveolar macrophages produce tumor necrosis factor (TNFa) [23], interleukin 1 (IL-I) [24] as well as interleukin 8 (IL-8) [25], and higher cytokine levels can be detected in plasma- and bronchoalveolar fluid samples (BAL) of ARDS patients as compared to ventilated controls [26]. The concentration of IL-8 in the bronchoalveolar lavage fluid of patients at risk for an ARDS was higher in patients who subsequently developed an ARDS [25], and elevated IL-I levels in the BAL at day 7 after the onset of ARDS were associated with an increased mortality rate [27]. Furthermore, cytokine levels of TNF and IL-8 in the blood as well as bronchoalveolar fluid have been shown to correlate positively with the severity of lung injury [26]. Thus TNF, IL-8 and IL-I may be used as predictive markers in at-risk patients or during the course of ARDS.

Of pathophysiological importance is the capability of these cytokines to activate polymorphnuclear granulocytes (PMN) [28]. PMNs are widely thought to playa central role during the pathogenesis of lung injury [29] because PMNs incubated with endotoxin can induce acute lung injury when reinjected into animals [30], and PMNs may cause endothelial damage by a variety of mediators such as the oxygen radicals -OH, 0 3 and H20 2 and different lysosomal proteolytic enzymes (for review see [28]). In ARDS patients the number of activated PMNs, which express adhesion molecules involved in process of endothelial adhesion and transendothelial migration, are increased in bronchoalveolar fluid [26]. However, occasionally ARDS develops in neutropenic patients [31], and lung injury can also be induced in granulocyte depleted animals [32]. These

88 P. Neumann

findings demonstrate that neutrophil independent mechanisms, as direct endothelial damage caused by endotoxin, may also be of considerable importance during the pathogenesis of lung dysfunction in sepsis.

Activation of the coagulation system with disseminated intravascular coagulation (DIC) is a common finding in sepsis, and nearly 25 years ago Bone and co-workers reported an association between DIC and ARDS [33]. Endothelial cell damage in the early phase of sepsis can expose collagen to the blood stream and thereby activate platelet adhesion as well as the intrinsic and extrinsic pathways of coagulation. In bronchoalveolar fluid of ARDS patients, procoagulant activity as well as fibrin degradation are enhanced compared to at-risk patients who do not develop ARDS [34]. Activated platelets may release thromboxane A2 and serotonin which can cause pulmonary vasoconstriction. Microthrombi may in addition occlude constricted vessels. This would cause further ventilation-perfusion mismatch (besides the impaired hypoxic pulmonary vasoconstriction induced by endotoxin) and thus contribute to hypoxaemia. In endotoxin-induced lung injury, pretreatment with the specific thrombin inhibitor hirudin followed by hirudin infusion, decreased the pulmonary vascular resistance, the accumulation of extravascular lung water and tended to reduce the alveolar-arterial oxygen difference [35].

In ARDS patients the level of circulating fragments of complement factor 3 (C3f) is significantly higher than in patients with ARDS risk factors who do not develop acute lung failure [6]. Inhibition of the complement cascade by administration of soluble complement receptor I is able to ameliorate the increased pulmonary microvascular permeability and to reduce the retention of PMNs within the lungs in an animal model with intestinal ischaemia-reperfusion lung injury [36]. Since activation of the complement cascade activates PMNs and macrophages, it is likely to playa role during the pathogenesis of lung failure in sepsls.

Conclusion Acute respiratory failure is a typical complication of sepsis associated with a poor prognosis. Infusion of endotoxin can mimic many aspects of early ARDS in sepsis and is a useful model to study structural and functional lung changes associated with a systemic inflammatory response. Mediators like TNF, IL-I, factors of the complement- and coagulation cascades, metabolites of arachidonic acid, polymorphnuclear granulocytes producing lysosomal enzymes and toxic oxygen radicals have all been identified as contributors to lung injury in sepsis. However, it is not possible to point out one central mediator which is most important for the sequence of events in lung injury and sepsis. Inflammatory mediators, some of which are highlighted above, interact in a rather complex and still not completely understood manner. Therefore it is not surprising that so far no

Lung Dysfunction in the Early Phase of Sepsis 89

specific immunomodulatory agent has been able to convincingly improve the outcome of acute lung failure in sepsis.

References l. The ACCP/SCCM Consensus Conference Committee, Bone RC, Balk RA et al (1992) Defin

itions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest 101: 1644-1655

2. Vasilyev S, Schaap R, Mortensen J (1995) Hospital survival rates of patients with acute respiratory failure in modem respiratory intensive care units. Chest 107: 1083-1094

3. Pepe PE, Potkin RT, Reus DH et al (1982) Clinical predictors of the adult respiratory distress syndrome. Am J Surg 144:124-130

4. Kollef MH, Schuster DP (1995) The acute respiratory distress syndrome. N Engl J Med 332: 27-37

5. Bernard GR, Artigas A, Brigham KL et al (1994) Report of the American-European consensus conference on ARDS: definitions, mechanisms, relevant outcomes and clinical trial coordination. Intensive Care Med 20:225-232

6. Parsons PE, Worthen GS, Moore EE et al (1989) The association of circulating endotoxin with the development of the adult respiratory distress syndrome. Am Rev Respir Dis 140: 294-301

7. Frevert CWF, Warner AE (1999) Respiratory distress resulting from acute lung injury in the veterinary patient. J Vet Int Med 6: 154-165

8. Borg T, Alvfors A, Gerdin B et al (1985) A porcine model of early adult respiratory distress syndrome induced by endotoxaemia. Acta Anaesthesiol Scand 29:814-830

9. Rosenthal C, Caronia R, Quinn C et al (1998) A comparison among models of acute lung injury. Crit Care Med 26:912-916

10. Neumann P, Berglund JE, Mondejar EF et al (1998) Dynamics of lung collapse and recruitment during prolonged breathing in porcine lung injury. J Appl Physiol 85: 1533-1543

11. Gattinoni L, Pre senti A, Torresin A et al (1986) Adult respiratory distress syndrome profiles by computed tomography. J Thorac Imaging 1:25-30

12. Gattinoni L, Pesenti A, Bombino M et a1 (1988) Relationship between lung computed tomographic density, gas exchange, and PEEP in acute respiratory failure. Anesthesiology 69: 824-832

13. Pelosi P, D' Andrea L, Vitale G et al (1994) Vertical gradient of regional lung inflation in adult respiratory distress syndrome. Am J Respir Crit Care Med 149:8-13

14. Forsgren P, lakobson S, Modig 1 (1989) True shunt in relation to venous admixture in an experimental porcine model of early ARDS. Acta Anaesthesiol Scand 33:621-628

15. Brigham KL, Meyrick B (1986) Endotoxin and lung injury. Am Rev Respir Dis 133:913-927 16. Schuster DP (1994) ARDS: clinical lessons from the oleic acid model of acute lung injury.

Am 1 Respir Crit Care Med 149:245-260 17. Nielsen JB, Sjostrand UH, Edgren EL et al (1991) An experimental study of different ventila

tory modes in piglets in severe respiratory distress induced by surfactant depletion. Intensive Care Med 17:225-233

18. Ghosh S, Latimer RD, Gray BM et al (1993) Endotoxin-induced organ injury. Crit Care Med 21:S19-S24

19. Jenkins JK, Carey PD, Byrne K et al (1991) Sepsis-induced lung injury and the effects of ibuprofen pretreatment. Am Rev Respir Dis 143: 155-161

20. Price S, Harlan 1, Carrico CJ et al (1986) Indomethacin, dazoxiben and extravascular lung water after Escherichia coli infusion. J Surg Res 41: 189-197

21. Schuster DP (1995) What is acute lung injury') What is ARDS? Chest 107: 1721-1726

90 P. Neumann

22. Anzueto A, Baughman RP, Guntupalli KK et al (1996) Aerosolized surfactant in adult with sepsis-induced acute respiratory distress syndrome. N Engl J Med 334:4l7-1421

23. Van Nhieu JT, Misset B, Lebargy F et al (1993) Expression of tumor necrosis factor a gene in alveolar macrophages from patients with the adult respiratory distress syndrome. Am Rev RespirDis 147:1585-1589

24. Jacobs RF, Tabor DR, Burks AW et al (1989) Elevated interleukin-I release by human alveolar macrophages during the adult respiratory distress syndrome. Am Rev Respir Dis 140: 1686-1692

25. Donnelly SC, Strieter RM, Kunkel SL et al (1993) Interleukin-8 and the development of adult respiratory distress syndrome in at-risk patient groups. Lancet 341:643-647

26. Chollet-Martin S, Jourdain B, Gibert C et al (1996) Interaction between neutrophils and cytokines in blood and alveolar spaces during ARDS. Am J Respir Crit Care Med 153:594-60 I

27. Goodman RB, Strieter RM, Martin DP et al (1996) Inflammatory cytokines in patients with persistence of the acute respiratory distress syndrome. Am J Respir Crit Care Med 154: 602-611

28. Bone RC (1991) The pathogenesis of sepsis. Ann Intern Med 115:457-469 29. Marini JJ, Evans TW (1998) Round table conference: acute lung injury. 15th-17th March

1997 Brussels, Belgium. Intensive Care Med 24:878-883 30. Welsh CH, Lein EC, Worthen GS et al (1989) Endotoxin-pretreated neutrophils increase pul

monar vascular permeability in dogs. J Appl Physiol 66: 112-119 31. Ognibene FP, Martin SE, Parker MM et al (1986) Adult respiratory distress syndrome in pa

tients with severe neutropenia. N Engl J Med 315:547-551 32. Winn R, Maunder R, Chi E et al (1987) Neutrophil depletion does not prevent lung edema af

ter endotoxin infusion in goats. J Appl Physiol 62: 116-121 33. Bone RC, Francis PB, Pierce AK (1976) Intravascular coagulation associated with the adult

respiratory distress syndrome. Am J Med 61 :585-589 34. Fuchs-Buder T, de Moerloose P, Ricou B et al (1996) Time course of procoagulant activity

and D dimer in bronchoalveolar fluid of patients at risk for or with acute respiratory distress syndrome. Am J Respir Crit Care Med 153: 163-167

35. Hoffmann H, Siebeck M, Spannagl M et al (1990) EtTect of recombinant hirudin, a specific inhibitor of thrombin, on endotoxin-induced intravascular coagulation and acute lung injury in pigs. Am Rev Respir Dis 142:782-788

36. Xiao F, Eppihimer MJ, Willis BH et al (1997) Complement-mediated lung injury and neutrophil retention after intestinal ischemia-reperfusion. J Appl Physiol 82: 1459-1465

The Kidney in Sepsis

J.A. KELLUM

The kidney is a significant target organ for the sepsis syndrome and, depending on the definitions used, acute renal failure (ARF) occurs in as many as 15% of patients with critical illness [1] and sepsis appears to be a contributing factor in as many as 43% [2]. In addition, the kidney is usually the organ that fails first [3]. While the mortality rate of isolated acute renal failure is approximately 10 to 15%, ARF in the setting of sepsis carries a mortality rate between 50 to 90% according to published series. In a recent study [4], the mortality rate among 253 cases of ARF treated in the ICU was 71.5% whereas it was 31.5% among the 495 cases of ARF treated in a non-ICU setting (p = 0.001).

The kidney is particularly ill-suited to tolerate sepsis in part because both the inflammatory response as well the haemodynamic consequences of sepsis have harmful effects on the kidney. Furthermore, and perhaps more importantly, the diagnostic and therapeutic interventions that patients with sepsis receive are frequently associated with renal injury. Although new pharmacologic agents are being developed and tested, therapy for established renal injury has been generally discouraging and preventive therapy is therefore encouraged [5]. The options available to reduce renal injury are limited but effective measures have been developed and ineffective or potentially harmful strategies should be avoided [6].

Direct vs. secondary renal injury It may be important to distinguish the direct effects of sepsis, those due to the primary inflammatory process from the effects secondary to alterations in systemic haemodynamics or as a consequence of therapy. This is because we have therapeutic options available to address systemic haemodynamics while our ability to modulate local or systemic inflammation is limited and, to date, has not been associated with improved survival or renal function. Furthermore, there is a significant potential to produce additional organ injury from our supportive therapy and the capacity to limit this injury exists through the use of improved supportive techniques [7].

92 1.A. Kellum

However, it is difficult if not impossible to separate these primary and secondary effects in human sepsis. Hence, animal models have been used extensively for this purpose, but unfortunately, these models can be misleading for a variety of reasons [8]. First, in an effort to limit the variability associated with sepsis models that require infection, researchers have often relied on endotoxin given that many gram-negative organisms have very similar cell wall composition. Even gram-positive exotoxin shares significant amino acid homology with endotoxin. Clinically, patients with these infections can appear quite similar. Healthy animals given endotoxin provide a much more uniform population to examine than do patients with sepsis. Unfortunately, this insult may not be as uniform as once thought and important differences exist between types of organisms studied and strains of endotoxin used.

First, it is important to consider endotoxin, not as a universal mechanism of sepsis, but rather as a trigger for the complex machinery of the systemic inflammatory response. There are other triggers just as capable of inducing this response. In one study of patients with septic shock only 58% of patients with documented gram-negative bacteraemia were found to have endotoxaemia [91. Thus, endotoxin merely serves as a relatively reproducible trigger. Second, important differences also exist among patients in their ability to generate this response as well as differences associated with the type and dose of endotoxin and the conditions under which it is studied. Endotoxin binds to specific proteins (lipopolysaccharide binding proteins) in the blood and this binding results in the release of various inflammatory mediators [10]. While a complete discussion of the mediators involved and their effects is beyond the scope of this chapter, it is important to note that endotoxin can induce mediators that have opposing effects on the vasculature. For example, both prostacyclin (PGI2), a vasodilator, and endothelin-l, a vasoconstrictor, are stimulated by endotoxin. In addition to the pro-inflammatory effects of TNF-a, interleukin-l (lL-I) and the various phospholipids, endotoxin also induces anti-inflammatory mediators such as IL-4, IL-IO and IL-l receptor antagonist [11]. The degree to which each of these mediators is stimulated will in large part determine the resulting physiologic response and therefore explains some of the heterogeneity seen in experiments that examine the haemodynamic effects of endotoxin.

Unfortunately, there are other limitations to the endotoxin model. Significant heterogeneity has also been described between endotoxins derived from various species of gram-negative bacteria. There is a tendency to think of endotoxin as a single compound, which is ubiquitous amongst gram-negative bacteria. Text books and review articles often show the molecular structure of Lipid A (the active component of endotoxin) as if to suggest that all endotoxin is the same. In fact, it is actually a family of compounds which, although they possess remarkable homology, are not chemically identical. These subtle differences in chemical structure may translate to significant differences in biological effect. The affinity of each type of endotoxin for the various types of binding proteins and their effects on individual cells may be quite different depending on the type of

The Kidney in Sepsis 93

endotoxin and type of host. A dramatic example of this was seen by Lugon and coworkers who compared the effects of endotoxin from two different strains of E. coli [12]. Even though both endotoxins were derived from the same species of bacteria, the haemodynamic effects they produced were quite different; one strain decreased systemic mean arterial pressure (MAP) without effect on the renal vasculature, while the other increased effective renal vascular resistance (RVR) without effecting MAP. The effects observed with intact organisms are even more striking. Animals given similar intraperitoneal doses of E. coli or P. aeruginosa exhibited different clinical effects [13]. Compared with E. coli, P. aeruginosa produced lower levels of endotoxaemia, but greater cardiovascular dysfunction and higher mortality. Even different strains of E. coli have been shown to affect disproportionate levels of endotoxaemia and haemodynamic instability [14].

Thus it may be impossible to completely separate the effects of sepsis on the kidney into primary and secondary effects. Endotoxin itself has little direct effect on the kidney. When it is perfused directly into an isolated kidney, there are no significant effects on renal blood flow (RBF), glomerular filtration rate (GFR) or tubular function [15]. The effects of endotoxin on the kidney are produced through secondary mediators which exert their effects directly on the renal vasculature and indirectly by effecting parallel circuits. Most studies have found that with endotoxaemia there is a profound decrease in RBF even when blood pressure and cardiac output are maintained [12, 15-17]. Lugon et al. [12] estimated effective RVR by PAH clearance and found that one strain of E. coli endotoxin (LPS 0111 :B4) produced a 74% increase in total RVR, although another strain did not. Examination of the renal pressure-flow relationships during experimental endotoxaemia provides further evidence of these effects. For example, we [18] have shown that endotoxin produces a decrease in RBF while also decreasing RVR and critical closing pressure suggesting that RBF is affected more by the systemic than renal haemodynamics. A modest decrease in blood pressure also occurred with endotoxin (MAP: 120 vs. 94 mmHg); however the pressure-independent effects of endotoxin were 8 times greater based on the parameter estimates from the general linear model (P < 0.001) [18]. Hence, endotoxin infusion resulted in a decrease in RBF, apparently due to blood flow redistribution and a proportionally smaller decrease in perfusion pressure.

Renal injury as a consequence of hypotension The most common type of renal dysfunction in the lCU is pre-renal azotaemia. However, in critically ill patients, renal injury is most frequently due to acute tubular necrosis (ATN) resulting from an ischaemic or nephrotoxic injury, or both [19]. lschaemic ATN may result from selective blood flow reduction to the kidneys and indeed, RBF diminishes as a consequence of endotoxaemia. However, more commonly, ischaemic ATN occurs in the setting of systemic arterial hy-

94 l.A. Kellum

potension. In this regard, there are several aspects of the RBF that require particular attention. The first is the renal outflow pressure and the second is the autoregulatory threshold. Flow through any vascular bed is defined by the perfusion pressure divided by the resistance. Perfusion pressure is defined as the difference between the inflow and outflow pressures. For practical purposes, because outflow pressure is normally small and difficult to measure, the inflow pressure is taken as the perfusion pressure. However, several factors may influence the renal outflow pressure and thus alter flow irrespective of the actual resistance or the inflow pressure. Tense ascites is such an example. In this condition it is quite possible to increase the outflow pressure above the critical closing pressure for the kidney (usually around 30 mmHg) and hence reduce the renal perfusion pressure. Indeed, some investigators have found renal effects (decreased GFR and urine output) with intraabdominal pressures as low as 12 mmHg during laparoscopy [20].

The point at which RBF becomes entirely dependent on renal perfusion pressure is known as the renal autoregulatory threshold. The autoregulatory threshold for the kidney has been determined to be about 80 mmHg [21], and human studies of healthy volunteers support this conclusion [22]. Furthermore, some investigators have argued that the autoregulatory threshold for GFR is even higher than that for RBF [23, 24]. Accordingly, even modest decreases in mean arterial pressure may affect RBF. Moreover, a loss of autoregulation altogether may occur in certain situations such as acute renal failure [25] and has even been described with low-dose endotoxin infusion [23]. Finally, RBF is not determined by oxygen demand as it is for other organs. Adenosine actually produces afferent arteriolar vasoconstriction and a reduction in GFR [25]. Thus, renal oxygen demand influences GFR through this tubuloglomerular feedback mechanism.

Renal injury as a consequence of therapy Several diagnostic and therapeutic interventions commonly used in patients with sepsis are potentially dangerous to the kidneys. One of the most commonly anticipated etiologies of ATN is the use of intravenous contrast agents for imaging studies [26]. Although the pathogenesis of renal injury secondary to radiocontrast agents is not entirely understood, it appears to be due to medullary ischaemia [27, 28]. For some time, it has been postulated that this ischaemic injury occurs on the basis of decreased renal blood flow secondary to renal vasoconstriction. It is therefore surprising that studies have shown that renal blood flow actually increases with radiocontrast [29]. This has lead some investigators to hypothesize that mesenteric ischaemia is a demand-side phenomenon. In other words, the ionic load leads to medullary ischaemia because the medullary cellular oxygen demand becomes greater than the supply [30, 311.


Radiocontrast induced ATN is rare in patients without underlying renal, cardiac or hepatic dysfunction and occurs most commonly in patients with diabetic nephropathy [26]. In this group the incidence approaches 50% depending on the degree of baseline renal function and the use of ionic vs. non-ionic contrast media [32]. Several forms of therapy have been proposed to prevent or treat radiocontrast induced ATN, including saline, furosemide, mannitol, calcium channel blockers, dopamine, atrial natriuretic peptide and theophylline [32]. There are no placebo controlled trials testing the effectiveness of any of these therapies. Virtually all studies have used hydration (usually with 0.45% saline) in addition to the agent being tested and most authors recommend its use. Even then, little comparative data exist for these potential treatments. The exception is a recent study by Solomon and coworkers [27] which compared furosemide plus saline to mannitol plus saline to saline alone; again, 0.45% saline was used. This randomized trial in 78 "high risk" patients found that both diuretic regimens were less effective in preventing ATN than saline alone. Another study by Weinstein et al. found that renal function significantly deteriorated in patients pretreated with furosemide [33]. So for this indication we can safely conclude that volume expansion with 0.45% saline is unproven but potentially helpful while diuretics are clearly not and may even be harmful.

A variety of other drugs commonly used in sepsis may produce renal injury. Aminoglycosides, amphotericin B, immunosuppressive agents, non-steroidal anti-inflammatory drugs, vasoactive medications and ACE-inhibitors are perhaps the most common offenders. For patients with sepsis, the most important of these are antimicrobial agents. Aminoglycosides are associated with a high incidence of ARF, 8-26% according to published reports [34] whereas ampotericin B is associated with 60% overall incidence of nephrotoxicity and 88% incidence when total dosages exceed 5 g [35].

The risks to the kidney from vasoactive mediations are difficult to assess. Vasopressors are used in patients to treat hypotension and hypotension is a major risk factor for ATN. Studies using high-dose vasoactive drugs or using these agents in healthy humans or animals may not be applicable to patients with sepsis in which the agents are titrated to treat arterial hypotension. One animal study [18], mentioned above, utilized endotoxin to produce a hyperdynamic sepsis-like state and used norepinephrine to treat hypotension. In this study we demonstrated that norepinephrine infused at clinically relevant doses increases mean arterial blood pressure and induces a decrease in ohmic resistance but an increase in the downstream critical closing pressure in the renal artery of the dog. Furthermore, this study found that norepinephrine infusion in acute endotoxaemia at similar doses reverses systemic hypotension and improves RBF independent of perfusion pressure using a general linear model. Therefore, norepinephrine appears to have two beneficial effects on RBF during endotoxic shock, one related to its effects on renal vascular resistance and the other related to its effect on renal perfusion pressure. These findings are consistent with the results of similar investigations [36-39] and provide a physiologic basis for the admin-

96 l.A. Kellum

istration of norepinephrine during septic shock. Other agents, such as phenylephrine or epinephrine, may have different effects on RBF due to their effects on blood flow redistribution. Until studies confirm that these agents can be safely interchanged, the recommendation that norepinephrine be the vasopressor of choice in sepsis seems appropriate.

Another major potential source of injury to the kidney is the treatment of ARF. Haemodialysis may be injurious to the kidney in a variety of ways [7]. For example, haemodialysis has long been known to produce oliguria in some patients. This is because the removal of both excess volume and urea increases the fractional reabsorption in the remaining nephrons, decreasing the tubular flow and predisposing to tubular obstruction [40]. A more significant problem in the ICU is one of haemodynamic instability. Patients with active inflammatory states are usually vasodilated and hypotensive and have little cardiovascular reserve. Intermittent haemodialysis frequently results in hypotension even if no volume is removed. These conditions often prompt increases in vasopressor doses or fluid resuscitation. Such therapies may themselves cause further morbidity. Even fluid resuscitation may be detrimental as it may result in increased extravascular lung water and prolong mechanical ventilation.

However, perhaps the most significant risk associated with haemodialysis comes from the bio-interaction of the immune system with the artificial circuit. This interaction results in complement activation as well as neutrophil infiltration into the kidney and other organs [41, 42]. Immune effector cell activation and cytokine production appear to be particularly influenced by the characteristics of the filter membrane. Synthetic membranes appear to be more "biocompatible" in this regard. Himmelfarb et al. have reported the results of 153 patients, confirming a survival benefit for patients treated with biocompatible compared to cellulose membranes (64% vs. 43%, p = 0.03) [43]. This advantage would almost certainly be greater in patients in active inflammatory states such as sepsis. Such findings are not surprising given that a large observational study (n = 2410) has shown that in chronic renal failure, there is a 25% greater mortality in patients using non-biocompatible rather than biocompatible membranes (p < 0.001) [44].

At present it remains to be determined just how biocompatible a dialysis membrane should be.

Biocompatibility can be assessed by a variety of criteria. Traditionally, biocompatibility has been assessed on the basis of complement activation (especially C3a levels). When this approach is used, cellulose membranes such as cuprophane are judged as non-biocompatible, semi-synthetic materials (e.g. hemophane) are considered "semi-compatible" and the synthetic polymers are all treated the same (all biocompatible) [41]. However, when biocompatibility is assessed on the basis of activation of leukocytes exposed to the membrane fibers, differences between synthetic polymers emerge [45] and even so-called biocompatible materials such as hemophane appear to cause problems r46]. Assessing biocompatibility by leukocyte activation may be particularly important


when dealing with patients with SIRS because the progression to MODS appears to be related to the duration of inflammation. Thus, the prolongation of inflammation by renal replacement therapy may lead to MODS.

Treatment and prevention options To date there is no therapy that can reverse renal failure once it has developed. Therapeutic options are therefore focused on prevention and limiting ongoing organ injury. Given that inadequate perfusion is a considerable risk factor for the development of ATN, strict attention must be paid to the maintenance of both cardiac output and perfusion pressure. Of note, there is no evidence that a mean arterial pressure of 60 mmHg, a commonly used "magic number", is sufficient to provide an acceptable perfusion pressure in all patients. Particularly in patients with a known history of hypertension, who may have renal vascular disease, a mean arterial pressure of 70 or even 80 mmHg may be inadequate. In all cases, these pressures are likely to be below the autoregulatory threshold and thus, increases in arterial pressure may be useful in restoring renal perfusion to some seemingly normotensive patients, particularly those whose history suggests that they normally have a much higher blood pressure.

Ischaemic ATN implies that RBF is reduced and thus, there has been considerable interest in augmenting RBF with vasodilators, especially since some studies have shown that RVR is increased. However, given that some studies have shown that RVR is not increased, indeed our study showed a decrease in RVR [18], the appropriateness of vasodilator therapy is uncertain. Although animal models have generated conflicting results, human studies of ARF have demonstrated that diuretics or dopamine are ineffective and potentially harmful in the prevention and/or treatment of ARF in multiple subsets of patients. Indeed, a systematic review of randomized trials comparing fluids alone versus diuretics (7 trials) or dopamine (18 trials, > 700 patients) in patients at risk of ATN from various causes found no evidence of benefit associated with either class of agent (level I studies of diuretics and level II studies of dopamine) [6]. Thus the common clinical practice of using so-called "renal-dose" dopamine should be discontinued.

A more effective strategy for the prevention of ARF is to limit the effects of exposure to nephrotoxic agents. Three main strategies have been investigated: 1. single daily dosing of aminoglycosides; 2. lipid complex amphotericin B; and 3. low-osmolality contrast media. Compared to multiple dosing, single daily doses of Gentamicin may reduce the incidence of nephrotoxicity in patients with normal renal function but data are inconclusive. One meta-analysis (not limited to ICU patients) showed equal efficacy and no difference in nephrotoxicity (relative risk 0.78 (0.31-1.94» [47]. One randomized trial (n = 95) comparing once daily vs. thrice daily dosing found equal efficacy but less nephrotoxicity (de-

98 J.A. Kellum

fined as an increase in serum creatinine of 45 /-lolll or more). Risk was nearly five times higher with multidosing than with once daily dosing (11/45 (24%) vs. 2/40 (5%), relative risk 4.89 (1.15-20.74, p = 0.02) [48]. Lipid complex amphotericin B preparations appear to cause less nephrotoxicity compared to standard amphotericin B, but direct comparisons of long term safety are lacking. Preliminary evidence from a phase II trial of lipid complex amphotericin B in 556 patients has found an incidence of renal toxicity (defined by any increase in serum creatinine) of 24% (compared to over 60-80% with amphotericin B). Furthermore, those patients with baseline serum creatinine> 2.5 mgldl (221 /-l01l1) on standard amphotericin B had a significant decrease in serum creatinine on lipid complex (p < 0.001) [49]. It is still important to stress that renal toxicity is still quite common (20-30%) with lipid complex preparations. Furthermore, lipid complex amphotericin is much more expensive than amphotericin B and formal cost effectiveness studies have not been done. If lipid complex preparations are safer, this advantage does not appear to extend to simply infusing amphotericin in intralipid. This approach resulted in no benefit and may be associated with pulmonary side effects [50]. Low-osmolality contrast media (LOCM) do not appear to be warranted in patients without underlying renal disease, in whom radiocontrast induced ATN is rare. A systematic review of 31 RCTs (n = 5146) comparing low osmolality with normal contrast media [51] revealed that LOCM did not influence the development of ARF or need for dialysis (rare events) but there were benefits (less nephrotoxicity) with LOCM in terms of serum creatinine. The overall benefit was small and was greatest in patients with underlying renal impairment (odds ratio 0.5 (0.36-0.7) favoring low-osmolality contrast).

Finally, the modality of renal replacement therapy and dialyzer membranes used appear to have a significant impact on survival. BiocompatibJe membranes are associated with improved survival in ARF [43] and chronic renal failure [44] as mentioned earlier. Preliminary evidence from combining several small studies further suggests that irrespective of membrane type, continuous renal replacement therapy is associated with a higher survival in patients with ARF compared to intermittent haemodialysis (relative risk 0.72 (0.60-0.87), p < 0.0 I) [51].

Conclusion The effects of sepsis on the kidney are complex and include direct effects of inflammation, effects related to systemic haemodynamic alterations and the effects of therapy. The most modifiable of these effects appears to be the last. Strategies that employ volume loading rather diuretics or dopamine and those that limit the exposure to nephrotoxins appear to be most effective. Preservation of renal perfusion is best achieved by supporting blood pressure and cardiac output rather than by use of renal vasodilators. Biocompatible membranes improve outcome in both acute and chronic renal failure. Additional benefit


may be realized by use of continuous rather than intermittent renal replacement therapy.

References 1. Brivet FG, Kleinknecht OJ, Loirat P et al (1996) Acute renal failure in intensive care units -

causes, outcomes and prognostic factors of hospital mortality: a prospective multicenter study. Crit Care Med 24: 192-198

2. Myers BD, Moran SM (1986) Hemodynamically mediated acute renal failure. N Engl J Med 314:97-100

3. Tran DO, Oe PL, de Fijter CWH et al (1993) Acute renal failure in patients with acute pancreatitis: Prevalence, risk factors, and outcome. Nephrol Dial Transplant 8: 1079-1084

4. McCullough PA, Wolyn R, Rocher LL et al (1997) Acute renal failure after coronary intervention: incidence, risk factors, and relationship to mortality. Am J Med 103:368-375

5. Dishart MK, Kellum JA (1999) An Evaluation of Pharmacologic Strategies for the Prevention and Treatment of Acute Renal Failure (ARF). Drugs (in press)

6. Kellum JA (1997) The use of diuretics and dopamine in acute renal failure: a systematic review of the evidence. Crit Care 1 :53-59

7. Kellum JA (1998) Primum non nocere and the meaning of modern critical care. Curr Op Crit Care 4:400-405

8. Kellum JA (1997) Endotoxin and renal blood flow. Blood Purif 15:286-291 9. Danner RL, Elin RJ, Hosseini JM et al (1991) Endotoxemia in human septic shock. Chest

99: 169-175 10. Hoffman WD, Natanson C (1993) Endotoxin in septic shock. Anesth Analg 77:613-624 II. Zivot 18, Hoffman WD (1995) Pathogenic effects of endotoxin. New Horizons 3:267-275 12. Lugon JR, Boim MA, Ramos OL et al (1989) Renal function and glomerular hemodynamics

in male endoxemic rats. Kidney lnt 36:570-575 13. Danner RL, Natanson C, E1in RJ et al (1990) P. aeruginosa compared with E. coli produces

less endotoxemia but more cardiovascular dysfunction and mortality in a canine model of septic shock. Chest 98: 1480-1487

14. Hoffman WD, Danner RL, Quezado ZM (1996) Role of endotoxemia in cardiovascular dysfunction and lethality: virulent and nonvirulent Escherichia coli challenges in a canine model of septic shock. Infect Immun 64:406-412

15. Cohen n, Black AJ, Wertheim SJ (1990) Direct effects of endotoxin on the function of the isolated perfused rat kidney. Kidney Int 37: 1219-1226

16. Burnier M, Waeber B, Aubert JF et al (1988) Effects of nonhypotensive endotoxemia in conscious rats: Role of prostaglandins. Am J Physiol 254:H509-H516

17. Walker JF, Cumming AD, Lindsay RM et al (1986) The renal response produced by non hypotensive sepsis in a large animal model. Am J Kidney Dis 8:88-97

18. Bellomo R, Kellum JA, Wisniewski S, Pinsky MR (1999) Effects of norepinephrine on renal vascular resistance and renal blood flow in normal and endotoxemic dogs. Am J Resp Crit CareMed 159:1186-1192

19. Hou SH, Bushinsky DA, Wish 18 et al (1983) Hospital-acquired renal insufficiency: A prospective study. Am J Med 74:243-248

20. Iwase K, Takenaka H, Ishizaka I et al (1993) Serial changes in renal function during laparoscopic cholecystectomy. Eur Surg Res 25:203-212

21. Shipley RE, Study RS (1951) Changes in renal blood flow, extraction of insulin glomerular filtration rate, tissue pressure, and urine flow with acute alterations in renal artery pressure. Am J Physiol 167:676-688

22. Stone AM, Stahl WM (1970) Renal effects of hemorrhage in normal man. Ann Surg 172: 825-836

100 J.A. Kellum

23. Bersten AD, Holt AW (1995) Vasoactive drugs and the importance of renal perfusion pressure. New Horizons 1995;3(4):650-661

24. Kircheim HR, Ehmke H, Hackenthal E et al (1987) Autoregulation of renal blood flow, glomerular filtration rate and renin release in conscious dogs. Pflugers Arch 410: 441-449

25. Kelleher SP, Robinette JB, Conger JD (1984) Sympathetic nervous system in the loss of autoregulation in acute renal failure. Am J Physiol 246:F379-F386

26. Barret BJ (1994) Contrast nephrotoxicity. J Am Soc Nephrol 5: 125-137 27. Solomon R, Werner C, Mann D et al (1994) Effects of saline, mannitol, and furosemide to

prevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med 331: 1416-1420

28. Heyman SN, Brezis M, Epstien FH et al (1991) Early renal medullary hypoxic injury from radiocontrast and indomethacin. Kidney Int 40:632-642

29. Weisberg LS, Kurnik PB, Kurnik BR (1993) Dopamine and renal blood flow in radiocontrastinduced nephropathy in humans. Renal Failure 15:61-68

30. Heyman SN, Brezis M (1996) The Renal Medulla: Life on the edge of hypoxia. The Third Annual Symposium on Applied Physiology of the peripheral circulation. 96:9

31. Heyman SN, Fuchs S, Brezis M (1995) The role of medullary ischemia in Acute Renal Failure. New Horizons 3(4):597-607

32. Barrett BJ, Parfrey PS (1994) Prevention of nephrotoxicity induced by radiocontrast agents. N EnglJ Med 331:1449-1450

33. Weinstein JM, Heyman S, Brezis M (1992) Potential deleterious effect of furosemide in radiocontrast nephropathy. Nephron 62(4):413-415

34. Kahlmater G, Dahlager 11 (1984) Aminoglycoside toxicity - a review of clinical studies published between 1975 and 1982. J Antimicrob Chemother 13[Suppl A]:9-22

35. Butler WT, Bennett JE, Alling DW et al (1964) Nephrotoxicity of amphotericin B, early and late events in 81 patients. Ann Intern Med 61: 175-187

36. Anderson WP, Korner PI, Selig SE (1981) Mechanisms involved in the renal responses to intravenous and renal artery infusions of noradrenaline in conscious dogs. J Physiol 321 :21-30

37. Zhang H, Smail N, Cabral A et al (1997) Effects of norepinephrine on regional blood flow and oxygen extraction capabilities during endotoxic shock. Am J Resp Crit Care Med 155: 1965-1971

38. Martin C, Eon B, Saux P et al (1990) Renal effects of norepinephrine used to treat septic shock patients. Crit Care Med 18:282-285

39. Desjars P, Pinaud M, Bugnon D, Tasseau F (1990) Norepinephrine therapy has no deleterious renal effects in human septic shock. Crit Care Med 18: 1048-1049

40. Yeh BP, Tomki DJ, Stacy WK et al (1975) Factors influencing sodium and water excretion in uremic man. Kidney Int 7:103-110

41. Pastan S, Bailey J (1998) Dialysis therapy. N Engl J Med 338:1428-1437 42. Yagi N, Paganini EP (1998) Acute dialysis and continuous renal replacement: Emergence of a

new technology involving the nephrologist in the intensive care setting. Seminars Nephrol 117:306-320

43. Himmelfarb J, Tolkoff-Rubin N, Chandran P et al (1998) A multicenter comparison of dialysis membranes in the treatment of acute renal failure requiring dialysis. J Amer Soc Nephrol 9:257-266

44. Hakim RM, Held PJ, Stannard DC et al (1996) Effect of the dialysis membrane on mortality of chronic hemodialysis patients. Kidney Int 50:566-570

45. Carreno MP, Stuard S, Bonomini M et al (1996) Cell-associated adhesion molecules as early markers of bioincompatibility. Nephrol Dial Transplant 11 :2248-2257

46. Thylen P, Fernvik E, Lundahl J et al (1996) Modulation of CD 11 b/CDl8 on monocytes and granulocytes following hemodialysis membrane interaction in vitro. Internat J Artif Org 19: 156-163

47. Hatala R, Dinh TT, Cook DJ (1997) Single daily dosing of aminoglycosides in immunocompromised adults: a systematic review. Clin Infectious Dis 24:810-815


48. Prins JM, Buller HR, Kuijper EJ et al (1993) Once versus thrice daily gentamicin in patients with serious infections. Lancet 341:335-339

49. Walsh TJ, Hiemenz JW, Seibel NL et al (1998) Amphotericin B lipid complex for invasive fungal infections: Analysis of safety and efficacy in 556 cases. Clinical Infectious Diseases 26:1383-1396

50. Schoffski P, Freund M, Wunder R et al (1998) Safety and toxicity of amphotericin B in glucose 5% or intralipid 20% in neutropenic patients with pneumonia or fever of unknown origin: randomised study. Brit Med J 317:379-384

51. Kellum JA, Leblanc M, Angus DC et al (1998) Continuous vs. intermittent renal replacement therapy: Is there a difference in survival? (abstract). Crit Care Med 27:A63

Pathophysiology of Liver Dysfunction in Sepsis

N. BRIENZA

Liver dysfunction is commonly associated with critical illness. Up to 54% of patients admitted to intensive care units present abnormal liver function tests. Severe liver dysfunction is reported to occur in 12-95% of adult respiratory distress syndrome patients and liver blood tests on the first day of diagnosis are predictive of survival [1]. Liver failure is a critical determinant of mortality in trauma [2], in intra-abdominal sepsis [2], and after cardiopulmonary bypass surgery [3].

There are two main clinical manifestations of hepatic involvement in the critical setting:

a) Ischaemic hepatitis

b) Jaundice

Ischaemic hepatitis occurs in the early stage of the disease after severe shock states and is probably due to inadequate oxygen delivery to the liver with consequent centrolobular hepatocellular necrosis. The typical cell target of ischaemic hepatitis confirms the primary role of hypoxia, since the centrolobular zones are the most sensitive to hypoxic damage. The hallmark of ischaemic hepatitis is the elevation of plasma aspartate aminotransferase (AST) and alanine aminotransferase (ALT) concentrations (usually greater than 1000 lUll), with normal or slightly elevated bilirubin level. This is a self-limited process and resolves if shock is reversed. However, it predisposes to the other major syndrome of liver dysfunction, i.e. jaundice.

Jaundice represents the classical liver alteration observed in sepsis and multiple organ failure syndrome and is associated with high mortality. It has been documented in as many as 44% of patients admitted to the leu because of severe trauma or sepsis [2]. The association of jaundice with sepsis has been known since 1837, with early description of "pneumonia biliosa" [4], but only recently has the key role of liver in prevention and integration of the systemic response to sepsis become apparent.

Although jaundice is the most obvious manifestation of liver involvement in sepsis, there is a wide range of more subtle changes in hepatocellular function that occur in sepsis, and hepatic dysfunction ranges from poor hepatic clearance of drugs to frank hyperbilirubinaemia and intrahepatic cholestasis.

104 N. Brienza

The causes of the hepatic dysfunction remain controversial. The two main mechanisms, acting either independently or together, seem to be the reduction and/or heterogeneity of hepatic blood flow and the hepatotoxic action of sepsis or inflammation mediators involved in the systemic response of multiple organ system failure.

While in some experimental models, an early decrease in hepatic blood flow has been reported, with reduction in liver oxygen consumption occurring early and before other metabolic alterations, most studies have observed that global liver blood flow may be preserved or even increased during sepsis, due to pronounced vasodilation of the liver arterial inflow, without significant consistent changes in portal blood flow [5]. Moreover, human studies have observed an increase in total liver flow and oxygen consumption during the early hyperdynamic phase of sepsis [6]. Nevertheless, during sepsis in humans, the liver seems to be characterized by a state of relative ischaemia, in that the increased metabolic requirements are not matched by an adequate increase of blood flow [6].

Even in a state of relative preservation of total liver flow, heterogeneity of sinusoidal perfusion may be a critical determinant of liver dysfunction resulting in reduction of oxygen and nutrition delivery to some areas of the liver. This hypothesis is supported by observations at the macro- and micro-circulatory level.

Endotoxin administration, while reproducing the typical pattern of hyperdynamic septic shock, causes an upward shift of the portal vein pressure-flow (PQ) relationship over the whole range of flow analysed [5, 7]. The upward shift of P-Q relationship is related to an increase in both portal vein critical closing pressure and resistance. The increase in closing pressure (similar to the increase observed in the pulmonary circulation with sepsis) is probably due to changes at the sinusoidal level. It has been reported that inflammatory states such as sepsis induce significant sinusoidal alterations with cell swelling (Kupffer cells), haemorrhage, congestion, and extravasation outside the sinusoidal endothelial barrier [5]. Ito cells contain filaments, tubules, and contractile proteins suggestive of contractile capability. Pathological states may activate these cells, inducing them to act as functional sinusoidal sphincters with consequent sinusoidal constriction and contribution to the global increase in portal vein closing pressure. This effect seems dependent on endothelin, an endothelium-derived contracting factor, whose levels are elevated during endotoxaemia. Indeed, when administered directly in the portal vein, endothelin is able to reproduce the increases in portal vein closing pressure observed after endotoxaemia [8]. Moreover, following endotoxin exposure, the responsiveness of the portal bed to endothelin increases, thus further enhancing both presinusoidal and sinusoidal resistances [9].

The functional activation of the sinusoidal cells by stress conditions may be responsible for the heterogeneity of sinusoidal perfusion as demonstrated by numerous studies. Liver perfusion is heterogeneous at some levels under normal conditions and the degree of heterogeneity is increased during the response to stress [10]. The distribution of hepatic arterial flow becomes more heteroge-

Pathophysiology of Liver Dysfunction in Sepsis 105

neous when perfusion pressure decreases [11]. A stepwise decrease in arterial pressure results in an increase in the heterogeneity of liver surface P02 and sinusoidal distribution of a fluorescent indicator such that dilated sinusoids filled with fluorescence coexist with most sinusoids with no fluorescence at all [12]. During reperfusion after ischaemia a patchy distribution of blood flow is evident as the result of microvascular failure at the level of lobules as well as sinusoids [13]. In summary, most of the studies demonstrate that, independent of global liver blood flow, heterogeneity of sinusoidal perfusion is increased during stress states, demonstrating microcirculatory failure in many areas of the liver.

The second hypothesis explaining hepatocellular dysfunction in sepsis involves the consequence of the release of inflammatory mediators. Activation of PMN cells by complement factors, gut-derived platelet activating factor or LPS may promote their adhesion to hepatic endothelium by selectins and intercellular adhesion molecules (IeAM-I), upregulated by cytokines or PAP. Subsequent amplification of this response impairs Kupffer cell function and may promote production of oxygen free radicals, cytokines and nitric oxide, contributing to hepatic injury. Hepatocytes will be exposed to the toxic effect of endotoxin and these inflammatory mediators with reduction in cytochrome-P450 drug metabolism [14] and depression of specific hepatocyte metabolic pathways. Endotoxin, interleukin-l and tumour necrosis factor cause inhibition of albumin synthesis and increased synthesis of acute phase proteins by hepatocytes [15]. Moreover, liver dysfunction itself may promote further hepatocyte injury because of the association with both increased bacterial translocation and Kupffer cell phagocytic depression.

Liver dysfunction predisposes to failure of other organ systems [16]. Due to the depression of Kupffer cell phagocytic capacity, gut-derived bacteria, endotoxins and the mediators may spill over into the systemic circulation with well documented effects resulting in dysfunction of other organ systems.

Thus, due to its central role, liver dysfunction must be viewed not only as the failure of a single organ in the MOFS setting, but as the failure of the "leading actor" in the integration of the systemic responses to inflammatory/septic stress.

References 1. Schwartz DB, Bone RC, Balk RA et al (1989) Hepatic dysfunction in the adult respiratory

distress syndrome. Chest 95:871-875 2. Te Boekhorst T, Urlus M, Doesburg W et al (1988) Etiologic factors of jaundice in severely ill

patients. J Hepatol 7: 111-117 3. Collins JD, Bassendine MF, Ferner R et al (1983) Incidence and prognostic importance of

jaundice after cardiopulmonary bypass surgery. Lancet i: 1119-1123 4. Garvin IP (1837) Remarks on pneumonia biliosa. S Med and Surg 1:536-544 5. Ayuse T, Brienza N, O'Donnell CP et al (1995) Alterations in liver hemodynamics in an in

tact porcine model of endotoxin shock. Am J Physiol268 HI106-H1114 6. Dahn M, Lange P, Lobdell K et al (1987) Splanchnic and total body oxygen consumption dif

ferences in septic and injured patients. Surgery 10 1:69-80

106 N. Brienza

7. Brienza N, Ayuse T, Revelly JP et al (1995) Effects of endotoxin on the isolated porcine liver: Pressure-flow analysis. J Appl Physiol 78:784-792

8. Brienza N, Ayuse T, Revelly JP et al (1998) Peripheral control of venous return in critical illness: role of the splanchnic vascular compartment. In: Dantzker DR, Scharf SM (eds) Cardiopulmonary critical care. Saunders, Philadelphia, pp 93-114

9. Pannen BHJ, Bauer M, Zhang JX et al (1996) A time-dependent balance between endothelins and nitric oxide regulating portal resistance after endotoxin pretreatment. Am J Physiol 271: H1953-H1961

10. Clemens MG, Bauer M, Gingalewski C et al (1994) Hepatic intercellular communication in shock and inflammation. Shock 2: 1-9

II. Watanabe Y, Puschel GP, Gardemann A et al (1994) Presinusoidal and proximal intrasinusoidal cont1uence of hepatic artery and portal vein in rat liver: Functional evidence by orthograde and retrograde bivascular perfusion. Hepatology 19: 1198-1207

12. Schywalsky M, Metzger HP (1990) Redistribution of local hepatic blood tlow during acute bleeding and prolonged hemorrhagic hypotension studied using fluorochromed plasma proteins and surface P02 measurements In: Piiper J (ed) Oxygen transport to tissue. Plenum Press, New York, pp 697 -703

13. Clemens MG, Mc Donagh PF, Chaundry IH et al (1985) Hepatic microcirculatory failure following ischemia and reperfusion: Improvement with ATP-MgCI2 treatment. Am J Physiol 248 :H804-H811

14. Ghezzi P, Saccardo B, Villa P et al (1986) Role of interleukin-I in the depression of liver drug metabolism by endotoxin. Infect Immun 54:837-840

15. Hawker F (1991) Liver dysfunction in critical illness. Anaesth Intens Carc 19: 165-181 16. Matuschak GM, Rinaldo IE (1988) Organ interactions in the adult respiratory distress syn

drome during sepsis. Role of the liver in host defense. Chest 94:400-406

Inflammatory Cells in Septic Shock

H. ZHANG, C. HSIA, G. PORRO

Septic shock is a complex pathophysiological syndrome, initiated by microbial products, and associated with a systemic activation of inflammatory responses, which can lead to multiple organ failure and the suppression of immune responses. The effector cells play a pivotal role in the pathogenesis of this complex condition. For example, macrophages are a principal source of the key mediators of septic shock that produce pro-inflammatory cytokines such as tumor necrosis factor-a (TNF-a) and chemokines like macrophage inflammatory protein-2 (MIP-2). Whereas optimal levels of these cytokines and chemokines are important for a sufficient defense, at gradually higher concentrations they mediate stronger local and finally systemic responses, with predominantly destructive rather than protective effects on the host. Macrophage release chemokines that attract neutrophils (PMNs). The letter effector cells playa central role in some of these responses by accumulating in tissues and releasing reactive oxygen species, cytokines and proteases that injure host structures. To block the deleterious effects of these mediators, an anti-body strategy in the treatment of sepsis may be not practical because that sepsis is involved in an inflammatory cascade in which mediators interact each other, so that neutralization of a single inflammatory molecule may be not sufficient to shape the overall inflammatory response. Whereas the effector cells produce mediators upon stimulation, they can also modulate the host response in sepsis through the release of enzymes and soluble receptors, which can act to neutralize some toxic mediators. Advances in molecular and cellular biology have provided insights into the pathogenesis of septic shock. This article briefly summarizes some of the roles of alveolar macrophages and PMNs in lung injury during this complex condition of septic shock.

Macrophages Alveolar macrophages playa central role in maintaining normal lung structure and function through their capacity to scavenge particulates, remove macromolecular debri, kill microorganisms, function as an accessory cell in immune responses, recruit and activate other inflammatory cells, maintain and repair lung

108 H. Zhang, C. Hsia, G. Porro

parenchyma, provide surveillance against neoplasms and modulate normal lung physiology. Alveolar macrophages serve this role by a variety of mechanisms, including their ability to phagocytize, to express specific cell-surface receptors for immunoglobulin, complement and granulocyte-macrophage colony-stimulating factor, etc., and to synthesize and release a broad armamentarium of mediators. Alveolar macrophages can also be a danger to the lung, since they possess the capacity to injure normal structures. In this regard, alveolar macrophages are a "two-edged sword" that serves to defend the lung but that can also render it harm.

Probably the most important role of the macrophages in sepsis is the production of a variety of pro-inflammatory cytokines, such as TNF-a, interleukin-l (IL-l) and interferon-y (INF-y). TNF-a may be central to the pathophysiology of sepsis. Macrophages produce large quantities of TNF-a following lipopolysaccharide (LPS) administration. In addition, macrophages can respond to LPS and the cytokines to produce nitric oxide (NO) by enhancing the transcription of the inducible NO gene, which leads to increased iNOS. iNO itself is capable of mediating cytotoxic effects [1]. NO can also react with PMN-released superoxide to form peroxynitrite. The latter has highly cytotoxic and microbicidal actions [2].

A major complication in sepsis is progressively lung injury and susceptibility to lung infection. It is known that septic lung injury is associated with increased accumulation of PMNs in lung and enhanced production of exe chemokines in bronchoalveolar lavage fluids. The elevated exe chemokines may be due to the sepsis-induced alveolar macrophage activation that produce a large quantity of chemokines attracting PMN into the lung. When stimulated in vitro, bronchoalveolar macrophages from septic animals had greatly enhanced exe chemokine responses as compared with macrophages from sham-operated animals [3].

Recently a technique was developed to selectively eliminate alveolar macrophages from the lungs of experimental animals in vivo, by the use of the cytotoxic drug dichloromethylene diphosphanate encapsulated in liposomes [4]. After intratracheal administration of these liposomes they are avidly phagocytosed by alveolar macrophages, which causes accumulation of the drug in alveolar macrophage. The alveolar macrophages are subsequently selectively killed without damage to the surrounding tissue or causing any change in cell populations in the lungs [5].

Some investigators used macrophage depletion technique to prevent rejection and infection during small intestinal transplantation since when rejection compromises normal intestinal barrier mechanisms and bacterial translocation results. Macrophages playa role in controlling the intestinal luminal bacteria and implicating allograft rejection. In a model of rat small bowel transplantation, Fryer et al. [6] demonstrated that macrophage depletion in the donor resulted in increased translocation of bacteria to the peritoneal cavity if recipient macrophages were present. With macrophage depletion in both the donor and

Inflammatory Cells in Septic Shock 109

the recipient in a model of rat small bowel transplantation, graft survival was prolonged significantly compared with non-macrophage-dependent controls [6]. This indicates that strategies depleting recipient macrophages may be useful in controlling small bowel allograft rejection without increasing bacterial translocation. However, as mentioned above macrophages participate significantly physiological host defense system, whether macrophage depletion may also result in immuno-depression remains to be elucidated.

We hypothesized that instead of repletion, fresh alveolar macrophage transplantation may down regulate cytokines such as TNF-a and CXC chemokines such as MIP-2 by its physiological host defense mechanism in lung tissues following LPS stimulation. Using a model of rat lung explants, we recently found that co-culture of alveolar macrophage with LPS resulted in a significant increase in the CXC chemokine MIP-2. Interestingly, replacement of the activated alveolar macrophages with fresh alveolar macrophages could attenuate the release of MIP-2 compared to a group without fresh macrophage transplantation. The replacement with activated macrophages primed by LPS further increased production of MIP-2. This observation suggests that fresh alveolar macrophages exerts significantly host defense action attenuating the production of MIP-2 in lung tissues following LPS administration. On the other hand, the fresh alveolar macrophages dose-dependently produces high levels of TNF-a simultaneously in both the control and the LPS-treated lung explants. This observation may lead to several suggestions: 1) The regulating mechanisms between TNF-a and MIP-2 are different, 2) The production of MIP-2 is not only secondarily due to stimulation by TNF-a but may be largely related to other factors, 3) Macrophages may secret products that selectively inhibit CXC chemokines [7] but not TNF-a. These findings indicate that selective regulation of chemokine and cytokine expression result from different response of cells to macrophages, suggesting an unique role of macrophages during sepsis cascade. A simple macrophage depletion may interrupt its auto-regulation function on host defense.

PMNs PMNs are blood-derived inflammatory cells with oxidative and proteolytic potential that are usually the first line of defense against invading pathogens. The ability of PMN s to leave out of blood vessels and migrate to extravascular beds in tissues is crucial for elimination of bacterial infection. In addition to killing pathogens, cytotoxic substances produced by activated PMN s are deleterious for the host. Activation of PMN s is frequently implicated in the promotion of severe inflammatory processes, including tissue injury associated with LPS challenge.

It has been demonstrated that after endotoxin administration to humans and experimental animals, there is an increased oxidant production by PMNs as determined by the detection of superoxide anion, hydrogen peroxide, or chemiluminescence [8, 9]. Tissue injury during sepsis is often associated with PMN-de-

110 H. Zhang, C. Hsia, G. Porro

rived oxygen free radicals, and treatment with antioxidants or blocking PMN function can prevent endotoxin-induced liver and lung injuries [10-12]. Some investigators showed that the oxidative response of PMNs in endotoxic rat is LPS dose- and time-dependent [13]. LPS exposure of PMNs can also cause production of PMN chemoattractants such as IL-S and cytokines [14, 15].

PMN releases a large quantity of broad spectrum of mediators in defense against infections. One of the targeted mediators is human neutrophil peptide (HNP-1-3). Human neutrophils contain large amount of HNP-1-3 which comprise between 30 and 50% of the protein in its "azurophil" granules and 5 to 7% of its total cellular protein [16]. High concentrations of HNP-1-3 ranging from 0.9 to 170 /lg/ml are found in the plasma of patients with sepsis [17]. A recent study [1S] reported that the mean plasma concentrations of HNP-1-3 in the patients at the onset of bacterial infection, nonbacterial infection, and pulmonary tuberculosis were 4.2, 3.2, and 1.S times the means for healthy volunteers (plasma: 0.255 ± 0.007 /lg/ml). Plasma concentrations of HNP-I-3 correlated with peripheral blood neutrophil counts [1S]. HNP-1-3 concentrations in the pleural fluid, bronchoalveolar lavage fluid (BALF), urine, and cerebrospinal fluid were also elevated in patients with infections [IS]. High levels of plasma (1.7 ± 1.1 /lg/ml in ARDS vs. 0.3 ± 0.3 /lg/ml in control) and bronchoalveolar lavage fluid (0.S94 ± 1.06S /lg/ml vs. 0.016 ± 0.015 /lg/ml) defensins were also observed in patients with acute respiratory distress syndrome (ARDS) [19]. High HNP-1-3 concentrations in the plasma of patients with bacteraemia, sepsis and ARDS may reflect the number and activity of neutrophils in the circulation [17, 19]. Because HNP-1-3 at> 20 to 30 /lg/ml are cytotoxic for normal mammalian cells in vitro [20], the elevated concentration of HNP-l-3 in bacteraemia, sepsis and ARDS could be injurious if the peptides were present in a free form. Moreover, local HNP-l-3 concentrations at the site of release must be much higher and therefore might contribute to local tissue injury.

Large body of evidence supports that neutrophils play an important role in mediating acute lung injury, and that neutrophils are sources of HNP-1-3 [21, 22]. During endotoxaemia in rats and rabbits, the peptides have been shown to increase release of inflammatory cytokines such as IL-1 [23]. In vitro exposure of human lung epithelial cells to HNP-1-3 at a concentration of 100 /lg/ml for three hours resulted in a significant increase in the production of interleukin-S (lL-S) [24], a pivotal chemokine, and cytotoxicity as assessed by chromium release [25]. HNP-1-3 modulate TNF-a production by stimulated human monocytes [26]. The effect of HNP-1-3 on IL-S production was comparable to that produced by TNF-a in a human epithelial cell line and primary bronchial epithelial cells (62 and 73% of TNF-a-induced IL-S levels, respectively) [24]. Human HNP-1-3 also display chemoattractant activity for monocytes [27].

We recently purified HNP-1-3 from patients with cystic fibrosis, because this population of patients has full of PMN in the airway. We investigated the effects of HNP-1-3 on several strains of bacteria, red blood cells, human lung epithelial cells, and mouse lung explants. We were able to show that HNP-l-3 killed sev-

Inflammatory Cells in Septic Shock III

eral strains of LPS and Pseudomonas aeruginosa at a dose of 2-50 /-lg/ml without damaging human red blood cells. However, human lung epithelial cells were damaged by HNP-1-3. We also studied the effects of HNP-I-3 in lung explants following LPS administration and measured TNF-a and MIP-2 concentrations. We found that low concentration of HNP-I-3 did not produce and attenuated LPS-induced increase in TNF-a and MIP-2 concentrations. High concentrations of HNP-1-3 itself could produce and further enhanced LPS-induced increase in TNF-a and MIP-2 concentrations from lung explants. This suggests that uncontrolled accumulation of HNP-1-3 in inflammatory sites may be injurious to the host and thus interventions aimed at inactivating HNP-1-3, may significantly reduce the untoward effects of the inflammatory response.

Conclusion Although much has been learnt about the development, a more comprehensive and integrated approach that used both cellular and whole animal system is required to a) understand the molecular, cellular, and physiological mechanisms that underlie the macrophage and PMN responses of sepsis, b) determine the contributions of microorganisms, fresh and activated macrophages and PMNs to the pathophysiological sequelae of patients with sepsis and c) develop therapies that are based on this next knowledge.

Acknowledgements. This study was supported by MRC grant (8558). Dr. Haibo Zhang is a Fellow of Medical Research Council of Canada.

References 1. Nathan C (J 997) Inducible nitric oxide synthase: what difference does it make? 1 Clin Invest

100:2417-2423 2. Hogg N, Kalyanaraman B (J 999) Nitric oxide and lipid peroxidation. Biochim Biophys Acta

1411:378-384 3. Czermak Bl. Breckwoldt M. Ravage ZB et al (1999) Mechanisms of enhanced lung injury

during sepsis. Am 1 Pathol154: 1057-1065 4. van Rooijen N, Sanders A (1997) Elimination, blocking, and activation of macrophages: three

of a kind? 1 Leukoc BioI 62:702-709 5. Thepen T, Kraal G, Holt PG (1994) The role of alveolar macrophages in regulation of lung in

flammation. Ann NY Acad Sci 725:200-206 6. Fryer J, Grant 0, liang 1 et al (1996) Influence of macrophage depletion on bacterial translo

cation and rejection in small bowel transplantation. Transplantation 62:553-559 7. Kopydlowski KM, Salkowski CA, Cody Ml et al (1999) Regulation of macrophage

chemokine expression by lipopolysaccharide in vitro and in vivo. 1 Immunol 163: 1537-1544 8. Kharazmi A, Andersen LW, Baek L et al (1989) Endotoxemia and enhanced generation of

oxygen radicals by neutrophils from patients undergoing cardiopulmonary bypass. 1 Thorac Cardiovasc Surg 98:381-385

ll2 H. Zhang, C. Hsia, G. Porro

9. Cerasoli F Jr, McKenna PJ, Rosolia DL et al (l990) Superoxide anion release from blood and bone marrow neutrophils is altered by endotoxemia. Circ Res 67:154-165

10. Zhang H, Spapen H, Manikis Pet al (1995) Tirilazad mesylate (U-74006F) inhibits effects of endotoxin in dogs. Am J Physiol 268:H I 847-H 1855

II. Spapen H, Zhang H, Wisse E et al (1999) The 21-aminosteroid U74389G enhances hepatic blood flow and preserves sinusoidal endothelial cell function and structure in endotoxinshocked dogs. J Surg Res (in press)

12. Blackwell TS, Blackwell TR, Holden EP et al (1996) In vivo antioxidant treatment suppresses nuclear factor-kappa B activation and neutrophilic lung inflammation. J Immunol 157:1630-1637

13. Kajdacsy-Balla A, Doi EM, Lerner MR et al (1996) Dose-response effect of in vivo administration of endotoxin on polymorphonuclear leukocytes oxidative burst. Shock 5:357-361

14. Haziot A, Tsuberi BZ, Goyert SM (1993) Neutrophil CDI4: biochemical properties and role in the secretion of tumor necrosis factor-alpha in response to lipopolysaccharide. J Immunol 150:5556-5565

15. Yamamoto T, Kajikawa 0, Martin TR et al (1998) The role of leukocyte emigration and IL-8 on the development of lipopolysaccharide-induced lung injury in rabbits. J Immunol 161: 5704-5709

16. Rice WG, Ganz T, Kinkade JM et al (1987) Defensin-rich dense granules of human neutrophils. Blood 70:757-765

17. Panyutich AV, Panyutich EA, Krapivin VA et al (1993) Plasma defensin concentrations are elevated in patients with septicemia or bacterial meningitis. J Lab Clin Med 122:202-207

18. Ihi T, Nakazato M, Mukae H, Matsukura S (1997) Elevated concentrations of human neutrophil peptides in plasma, blood, and body fluids from patients with infections. Clin Infec Dis 25:1134-1140

19. Ashitani J, Mukae H, Ihiboshi H et al (1996) Defensin in plasma and in bronchoalveolar lavage fluid from patients with acute respiratory distress syndrome. Nippon Kyobu Shikkan Gakkai Zasshi 34:1349-1353

20. Lehrer RI, Lichtenstein AK, Ganz T (1993) Defensins: antimicrobial and cytotoxic peptides of mammalian cells. Annu Rev Immunol II: 105-128

21. Barnathan ES, Raghunath PN, Tomaszewski JE et al (1997) Immunohistochemicallocalization of defensin in human coronary vessels. Am J Pathol 150: 1009-1020

22. Mukaida N, Matsumoto T, Yokoi K et al (1998) Inhibition of neutrophil-mediated acute inflammation injury by an antibody against interleukin-8 (IL-8). Inflamm Res 47:S 151-S 157

23. Korneva EA, Rybakina EG, Orlov DS et al (1997) Interleukin-I and defensins in thermoregulation, stress, and immunity. Ann NY Acad Sci 813:465-473

24. Van Wetering S, Mannesse-Lazeroms SPG, Van Sterkenburg MAJA et al (1997) Effect of defensins on interleukin-8 synthesis in airway epithelial cells. Am J Physiol 272:L888-L896

25. Okrent DG, Lichtenstein AK, Ganz T (1990) Direct cytotoxicity of polymorphonuclear leukocytes granule proteins to human lung-derived cells and endothelial cells. Am Rev Respir Dis 141: 179-185

26. Misuno NI, Kolesnikova TS, Lehrer RI et al (1992) Effect of defensin HNP-l of human neutrophils on production of tumor necrosis factor a by human blood monocytes in vitro. Bull Exp Bioi Med 113:709-712

27. Territo MC, Ganz T, Selsted ME, Lehrer R (1989) Monocyte-chemotactic activity of defensins from human neutrophils. J Clin Invest 84:2017-2020

I SEPSIS TRIAL I

Revised Terminology on Sepsis

J.-L. VINCENT

Over recent years, major advances have been made in our understanding of the sepsis response, its origins, mediators, effects, and consequences. Ongoing research continues to add to our knowledge of this complex and complicated process. Sepsis is the systemic response to infection, but as it emerged that the inflammatory response which is aroused in sepsis can also occur in conditions in which infection is not apparent, attempts were made to introduce new terminology to define and label these conditions. Terms such as the sepsis syndrome [1] and the systemic inflammatory response syndrome (SIRS) [2] became popular and have been used widely in describing patients both individually and in larger groups for clinical trial purposes.

However, as we will discuss below such terms have, in fact, not been helpful, and have further complicated a field already burdened with excess terminology. Indeed more words are unnecessary, what is needed is a better basic understanding of the pathophysiology of sepsis and the development of reliable markers of sepsis which will aid the physician in the often difficult diagnosis of sepsis in the critically ill patient.

Infection versus inflammation The words sepsis and infection are often used as synonyms, although they are separate entities. Infection refers to an insult by a foreign organism (bacteria, fungus, virus, etc.) while sepsis is the inflammatory response that occurs as a result of the infectious invasion. Infection can certainly occur without sepsis, and similarly, it has become apparent that a "sepsis-like" inflammatory response can be elicited in conditions where there is no evidence of infection, such as pancreatitis, myocardial infarction, and trauma. The similarity between the "non-septic" inflammatory response and the septic response has raised the possibility that the inflammatory response in these conditions is in fact triggered by bacterial invasion via the gastrointestinal tract occurring following damage to the intestinal mucosal barrier [3, 4] but this hypothesis remains controversial at present. For the clinician, the challenge is to separate infectious from non-infectious causes of the inflammatory response, in order to institute appropriate antibiotic

116 J.-L. Vincent

therapy and/or surgical drainage when an infection is demonstrated. In the clinical context, terms such as SIRS are not helpful and even potentially dangerous as they may reduce the incentive to search for infection [5]. The criteria for SIRS are met when a patient exhibits two or more of the following: fever, tachycardia, tachypnea, leukocytosis. However, the majority of critically ill patients will fit the SIRS criteria [6, 7]. Using the same consensus definitions, SIRS with evidence of infection equals sepsis. Hence, a patient with a fever, leukocytosis, and a positive urine culture could be said to be septic by this definition, and yet for most clinicians such a patient would probably not be considered "septic" [8]. This example merely highlights the difficulties in defining sepsis; such overly sensitive and non-specific terms add nothing to aid the clinician in his diagnosis. Indeed, with no cure available for sepsis per se, treatment hinges on the identification and eradication of infection. To achieve this, a complete and repeated search for an infectious cause and origin must be conducted in all "septic" patients, with full bacteriological culture and radiographic imaging as indicated. The assessment of signs and markers of sepsis may be of assistance in determining the presence of sepsis.

Signs of sepsis The most commonly quoted clinical signs of sepsis are fever, tachycardia, tachypnea, and leucocytosis, and these are indeed present in many septic patients. However, these signs are often present in many critically ill non-septic patients. For example, a patient with asthma may have tachypnea, tachycardia and a raised white cell count but is not necessarily septic; patients with myocardial infection or cardiogenic pulmonary oedema may have all four classical signs of sepsis and yet not be septic. In addition, some patients who are septic may not have any of the signs of sepsis. For example, fever is perhaps the sign most typically associated with sepsis, and yet failure to develop fever or the presence of hypothermia is actually associated with higher mortality rates in septic patients [9]. Tachycardia may similarly be absent, particularly in patients on antiarrhythmic medication, and tachypnea is difficult to assess in the patient on mechanical ventilation. In fact, only a few patients with sepsis will present the typical signs of sepsis, and equally, not all patients who present signs of sepsis will in fact have sepsis (Fig. 1).

Markers of sepsis As the mediators involved in the inflammatory sepsis cascade were identified, it was proposed that some of these mediators could be useful markers of sepsis. Raised levels of many cytokines including tumor necrosis factor-a (TNF-a), interleukin (IL)-6, IL-8, and IL-IO are indeed present in many septic patients and

Revised Tenninology on Sepsis

Signs of sepsis

Trauma, pancreatitis, etc.

117

Clinical and bacteriological parameters

Fig. 1. Diagrammatic representation of the relationship between signs of sepsis and infection. Only those patients with signs of sepsis and infection can be said to have true sepsis

some have been associated with a worse outcome [10, 11]. IL-6 has been used as an entry criterion for a clinical trial of an anti-sepsis therapy [12]. However, raised cytokine levels are not consistently present and vary according to the time course of the disease process. In addition, they may be present in other conditions causing an inflammatory response such as trauma [13] or pancreatitis [14].

Alternative biological markers have therefore been suggested. Of these, Creactive protein (CRP) is the most well-known and commonly used marker. CRP is an acute phase protein produced by the liver, and plasma level rise with infection in response to circulating cytokines. CRP levels have been reported to be a useful marker of the presence of sepsis by several authors [15, 16] and Yentis et al. [17] noted a fall in CRP levels with successful treatment. Povoa et al. [16] found CRP levels to be more sensitive for infection than white cell count or body temperature.

Procalcitonin, the precursor of calcitonin, has also been suggested as a marker of sepsis. The source and function of procalcitonin in sepsis are unclear, although the liver may be an important source [18] and recent evidence suggests that procalcitonin is an active pro-inflammatory mediator in sepsis [19, 20]. Plasma procalcitonin levels can rise several hundred fold in patients with severe sepsis, and have been used to differentiate infectious from other causes of inflammation [21, 22]. Muller et al. [23] reported that procalcitonin levels had a better predictive value for sepsis than CRP or IL-6, and Oberhoffer et al. [24]

118 J.-L. Vincent

recently noted that procalcitonin had greater specificity and sensitivity for raised TNF and IL-6 levels in septic patients than CRP, leukocyte count or body temperature. Procalcitonin levels may be particularly useful as a marker of the severity of infection, and to confirm a diagnosis suggested by raised CRP levels [25, 26]. Other potential biological markers of sepsis include neopterin, elastase, and phospholipase A2, all of which have been reported to be present in the plasma of patients with sepsis [27-29].

Many other signs and markers of sepsis have been proposed (Table 1) and many more will be suggested as research in this field continues, but as yet none have been found to be consistently reliable, sensitive and specific for sepsis. Importantly, none of these signs should be taken in isolation; a complete picture can only be drawn by combining the clinical and haemodynamic evaluation with all other available indicators [30]. In addition, it is important not to base a diagnosis on a single value but rather to evaluate the trend in levels over time.

Table 1. Some of the many proposed signs or markers of sepsis

Fever (or hypothermia) Tachycardia Tachypnea Altered white cell count Raised C-reactive protein levels Raised cytokine levels - TNF, IL-6, IL-lO, etc. Raised procaicitonin levels Raised neopterin, elastase levels Unexplained hyperlactataemia Unexplained alteration in organ function

Degrees of immune response Many potential new anti-sepsis therapies have been proposed in recent years aimed both at limiting the immune response [31-331 and at augmenting it [34]. However, none have consistently been shown to improve outcome. There are many possible reasons for this apparent lack of efficacy [35], but one key factor is perhaps the fact that the patients studied are too heterogeneous. While sepsis is often seen as the enemy, it is in fact a normal and necessary response to an external threat, an invading microorganism. Sepsis involves both pro- and antiinflammatory mediators and it is the balance between the actions of these two interlinked groups which determines the degree of immune response and perhaps patient outcome. Different patients, and the same patient at different times, will exhibit different degrees of pro- and anti-inflammatory reaction [36]. Sim-

Revised Terminology on Sepsis 119

plistically, treating all septic patients with an anti-inflammatory agent may in fact worsen disease in those who already have a net anti-inflammatory balance, and similarly treating patients with a pro-inflammatory agent may only be beneficial in those with a net anti-inflammatory response. The ability to assess the status of the immune response at any time could therefore be of use in selecting patients for clinical trials of anti-sepsis treatments, and enabling adjustment of treatment to be made as the immune status in any patient alters. Immunologic monitoring by cellular stimulation techniques evaluating the production of cytokines by monocytes from septic patients after stimulation with endotoxin, is being developed and may help to characterize the degree of immune response in patients and follow their response to treatment [37]. However, these techniques are cumbersome and involve in vitro analysis; circulating cells may not show the same response as tissue cells. Further work is clearly needed to establish the full potentials of this approach.

Genetic predisposition

The possibility of genetic predisposition for sepsis is an exciting area of current research. The amount of TNF released during sepsis seems, at least in part, to be genetically determined with TNFB2 homozygous individuals producing higher levels of circulating TNF in sepsis and having higher mortality rates [38]. Similarly in meningococcal disease, a genetic predisposition to produce high concentrations of plasminogen activator inhibitor 1 (PAl-I), by carriage of the homozygous 40 deletion polymorphism in the PAI-I gene, has been reported to be associated with an increased mortality [39]. Identification of patients according to their genetic predisposition may enable more appropriate and focused targetting of anti-sepsis therapies.

Conclusion

As our understanding of the pathogenesis of sepsis grows, so does the terminology surrounding this often fatal condition. However, for the clinician, what matters is the ability to differentiate between true sepsis due to infection, and the similar inflammatory response of non-infectious origin. To this end, too much emphasis has been placed on the "traditional" signs of sepsis and more accurate markers of sepsis need to be developed and utilized in those patients in whom diagnosis is not clear, so that anti-microbial treatment can be instituted only where necessary. Immunologic monitoring may prove useful in defining the degree of immune response to facilitate the choice of anti-sepsis treatment, but this approach remains experimental at the present time.

120 J.-L. Vincent

References 1. Bone RC, Fisher CJJ, Clemmer TP et al (1989) Sepsis syndrome: a valid clinical entity.

Methylprednisolone Severe Sepsis Study Group. Crit Care Med 17:389-393 2. Anonymous (1992) 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 20:864-874

3. Ford EG, Baisden CE, Matteson ML, Picone AL (1991) Sepsis after coronary bypass grafting: evidence for loss of the gut mucosal barrier. Ann Thorac Surg 52:514-517

4. Brathwaite CE, Ross SE, Nagele R et al (1993) Bacterial translocation occurs in humans after traumatic injury: evidence using immunofluorescence. J Trauma 34:586-589

5. Vincent JL (1997) Dear SIRS, I'm sorry to say that I don't like you. Crit Care Med 25: 372-374

6. Rangel-Frausto MS, Pittet D, Costigan M et al (1995) The natural history of the systemic inflammatory response syndrome (SIRS). A prospective study. JAMA 273: 117 -123

7. Bossink AW, Groeneveld J, Hack CE, Thijs LG (1998) Prediction of mortality in febrile medical patients: How useful are systemic inflammatory response syndrome and sepsis criteria? Chest 113: 1533-1541

8. Opal SM (1998) The uncertain value of the definition for SIRS, Systemic inflammatory response syndrome. Chest 113:1442-1443

9. Clemmer TP, Fisher CJJ, Bone RC et al (1992) Hypothermia in the sepsis syndrome and clinical outcome. The Methylprednisolone Severe Sepsis Study Group. Crit Care Med 20: 1395-1401

10. Pinsky MR, Vincent JL, Deviere J et al (1993) Serum cytokine levels in human septic shock. Relation to multiple-system organ failure and mortality. Chest 103:565-575

11. Damas P, Canivet JL, de Groote D et al (1997) Sepsis and serum cytokine concentrations. Crit Care Med 25:405-412

12. Reinhart K, RAMSES Study Group (1998) Treatment of severe sepsis in patients with highly elevated IL-6 levels with anti-TNF monoclonal antibody MAK 195F. Crit Care 2:P18 (abstract)

13. Ertel W, Keel M, Bonaccio M et al (1995) Release of anti-inflammatory mediators after mechanical trauma correlates with severity of injury and clinical outcome. J Trauma 39:879-885

14. Brivet FG, Emilie D, Galanaud P (1999) Pro- and anti-inflammatory cytokines during acute severe pancreatitis: an early and sustained response, although unpredictable of death. Parisian Study Group on Acute Pancreatitis. Crit Care Med 27:749-755

15. Matson A, Soni N, Sheldon J (1991) C-reactive protein as a diagnostic test of sepsis in the critically ill. Anaesth Intensive Care 19: 182-186

16. Povoa P, Almeida E, Moreira P et al (1998) C-reactive protein as an indicator of sepsis. Intensive Care Med 24:1052-1056

17. Yentis SM, Soni N, Sheldon J (1995) C-reactive protein as an indicator of resolution of sepsis in the intensive care unit. Intensive Care Med 21:602-605

18. Nijsten MWN, Olinga P, The TH (1999) Procaicitonin behaves as a fast responding acute phase protein in vivo and in vitro. Crit Care Med (in press)

19. Whang KT, Vath SD, Nylen ES et al (1999) Procaicitonin and proinflammatory cytokine in interactions in sepsis. Shock 12:268-273

20. Nylen ES, Whang KT, Snider RHJ et al (1998) Mortality is increased by procaicitonin and decreased by an antiserum reactive to procaicitonin in experimental sepsis [see comments]. Crit Care Med 26:1001-1006

21. Hammer S, Meisner F, Dirschedl Pet al (J 998) Procaicitonin: a new marker for diagnosis of acute rejection and bacterial infection in patients after heart and lung transplantation. Transpl ImmunoI6:235-241

22. Rothenburger M, Markewitz A, Lenz T et al (1999) Detection of acute phase response and infection. The role of procaicitonin and C-reactive protein. Clin Chem Lab Med 37:275-279

Revised Terminology on Sepsis 121

23. Muller B, Becker KL, Schachinger H et al (1999) Procalcitonin peptides are reliable markers of sepsis in a medical intensive care unit. Crit Care Med (in press)

24. Oberhoffer M, Karzai W, Meier-Hellmann A et al (1999) Sensitivity and specificity of various markers of inflammation for the prediction of tumor necrosis factor-alpha and interleukin-6 in patients with sepsis. Crit Care Med 27: 1814-1818

25. Schroder J, Staubach KH, Zabel Pet al (1999) Procalcitonin as a marker of severity in septic shock. Langenbecks Arch Surg 384:33-38

26. Ugarte H, Silva E, Mercan D et al (1999) Procalcitonin used as a marker of infection in the intensive care unit. Crit Care Med 27:498-504

27. Pacher R, Redl H, Frass M et al (1989) Relationship between neopterin and granulocyte elastase plasma levels and the severity of multiple organ failure. Crit Care Med 17:221-226

28. Yao YM, Yu Y, Wang YP et al (1996) Elevated serum neopterin level: its relation to endotoxaemia and sepsis in patients with major burns. Eur J Clin Invest 26:224-230

29. Nyman KM, Uhl W, Forsstrom Jet al (1996) Serum phospholipase A2 in patients with multiple organ failure. J Surg Res 60:7-14

30. Vincent JL (1999) Procalcitonin: THE marker of sepsis? Crit Care Med (in press) 31. Opal SM, Fisher cn, Dhainaut JF et al (1997) Confirmatory interleukin-I receptor antagonist

trial in severe sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. The Interleukin-l Receptor Antagonist Sepsis Investigator Group. Crit Care Med 25: 1115-1124

32. Abraham E, Anzueto A, Gutierrez Get al (1998) Double-blind randomised controlled trial of monoclonal antibody to human tumour necrosis factor in treatment of septic shock. NORASEPT II Study Group. Lancet 351 :929-933

33. Dhainaut JF, Tenaillon A, Hemmer Met al (1998) Confirmatory platelet-activating factor receptor antagonist trial in patients with severe gram-negative bacterial sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. BN 52021 Sepsis Investigator Group. Crit Care Med 26:1963-1971

34. Wasserman D, loannovich JD, Hinzmann RD et al (1998) Interferon-gamma in the prevention of severe burn-related infections: a European phase III multicenter trial. The Severe Burns Study Group. Crit Care Med 26:434-439

35. Vincent JL (1997) New therapies in sepsis. Chest 112:330S-338S 36. Vincent JL (1999) The immune response in critical illness: Excessive, inadequate or dysregu

lated. In: Marshall JC, Cohen J (eds) Immune Response in the Critically Ill. Springer, Heidelberg, pp 12-21

37. Docke WD, Randow F, Syrbe U et al (1997) Monocyte deactivation in septic patients: restoration by IFN-gamma treatment. Nat Med 3:678-681

38. Stuber F, Petersen M, Bokelmann F, Schade U (1996) A genomic polymorphism within the tumor necrosis factor locus influences plasma tumor necrosis factor-alpha concentrations and outcome of patients with severe sepsis [see comments J. Crit Care Med 24: 381-384

39. Hermans PW, Hibberd ML, Booy R et al (1999) 4G/5G promoter polymorphism in the plasminogen-activator-inhibitor-I gene and outcome of meningococcal disease. Meningococcal Research Group. Lancet 354:556-560

The Epidemiology and Outcome of Patients with Sepsis: Clear as Mud

R.S. WAX, D.C. ANGUS

Considerable resources have been expended in recent years towards the care of patients with sepsis. Our lack of impact on the survival of septic patients, despite many therapeutic trials, suggests that a review of disease characteristics may help guide future efforts with greater success. In this chapter, we will review the case definition, incidence and occurrence rates of sepsis. Some studies have suggested that the epidemiology of sepsis has changed over time, including changes in incidence, mortality and microbiological etiology. The outcome of patients with sepsis will be reviewed, including the role of prognostic indices. Finally, we will summarize data regarding the impact of therapy on patient outcome. Throughout these various aspects of discussion of sepsis, we will illustrate concerns that contribute to some confusion regarding understanding of the condition of sepsis.

Case definition One of the principal tasks for researchers studying a disease is to generate a reliable and valid case definition. Unfortunately, "sepsis" is a syndrome describing a heterogeneous constellation of symptoms and signs, with no gold standard for comparison. The difficulty in developing clear case definitions has hampered our understanding of the pathophysiology of sepsis and impaired the development of successful therapy. Classically, the condition known as sepsis is thought of as a phenomenon related to host response to infection. However, many patients with sepsis do not have documented infection [1-3]. Other conditions, such as trauma or pancreatitis, can be associated with physiologic abnormalities and organ dysfunction similar to those seen in patients with severe infection. Thus, a wide spectrum of etiologic agents can trigger the host response seen in sepsis. Although infection plays an important role in sepsis, other factors must be considered to adequately define the sepsis syndrome. Inconsistent application of sepsis definition criteria contributes to confusion and variability in the literature [4]. Given similar inciting events or exposures, many patients manifest differences of physiologic dysfunction in sepsis. The American College of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM) Consensus Con-

124 R.S. Wax, D.C. Angus

ference paper published in 1991 attempted to introduce more explicit definitions for such terms as: systemic inflammatory response syndrome (SIRS), sepsis, severe sepsis, septic shock, sepsis-induced hypotension, and multiple organ dysfunction syndrome [4]. Although specific diagnostic criteria were provided for these subsets of septic patients, the considerable overlap of these definitions prevented resolution of the problems in case definition. For example, the classic editorial by Vincent criticized the legitimization of the concept of SIRS because of problems with over-sensitivity, lack of linkage with pathophysiology, loss of definition specificity, and unclear benefit for patient care and advancement in research [5].

Variability in the time course of sepsis can also introduce difficulty in case definition. Although some patients fulfill criteria for various labels in the spectrum of sepsis on admission to the ICU, other patients will only do so further into their hospital stay. In one prospective cohort analysis of septic patients admitted to the ICU, 18% of patients did not fulfill case definition criteria for sepsis at the time of admission, but did meet criteria within the first week of ICU stay [6]. Better understanding of the time course and outcome of sepsis may be attained by considering the impact of changing patient status.

Some attempts have been made to link case definition of sepsis with physiology or organ dysfunction scores, but the benefit of such linkage remains unclear. As discussed later in this chapter, the use of outcome predictive models may be helpful in stratifying mortality risk for patients, thus helping to focus efforts on those patients with the greatest potential for benefit from therapy. However, the introduction of factors into case definition that are associated with outcome but not the pathophysiology of disease (e.g., whether the patient came to the ICU from the ward or emergency department) may impede further understanding of sepsis and the development of potential therapy [5].

Our current ability to define the state of sepsis is suboptimal. Further efforts should be made to refine our case definitions to increase the likelihood for improved understanding and treatment of this heterogenous condition.

Incidence of sepsis An accurate estimate of the incidence of sepsis is obviously difficult given the above problems of case definition. However, even if we accept the ACCP/SCCM consensus criteria for case definition, there are still few studies that determine incidence, since incidence requires a general population denominator. One of the largest epidemiological studies describing patients with sepsis was conducted by the Center for Disease Control in 1990, using a stratified sample of United States hospital discharge data [7]. This study found that the incidence of sepsis increased from 73.6 per 100,000 patients in 1979 to 175.9 per 100,000 in 1989. The increased incidence in sepsis was thought to be related to the increased prevalence of patients with HIV / AIDS, prolonged survival of

The Epidemiology and Outcome of Patients with Sepsis: Clear as Mud 125

HIV/AIDS patients with a subsequent increased duration of risk, increased use of invasive devices such as central venous catheters, and increased ability to diagnose sepsis. However, interpretation of this data is limited by case definition (septicaemia instead of sepsis), provides no information on patient management or outcome, and is based on a limited survey of only 1 % of hospital discharges (based on the National Hospital Discharge Survey).

Rangel-Frausto and colleagues published a prospective observational study examining occurrence rates of sepsis in 3708 patients admitted to intensive care units or general medical wards in a tertiary medical teaching center. This study did use ACCP/SCCM consensus criteria but did not have any general population denominator data, and hence describe occurrence rates but not incidence rates. They found 68% of patients studied met criteria for SIRS. Of those patients with SIRS, 26% developed sepsis, 18% developed severe sepsis, and 4% developed septic shock [8]. A similar study by Sands et al. described a population at eight academic medical centers with an occurrence rate of sepsis of 2.8 patients per 1000 patient-days [9]. In this study, considerable variability in the incidence of sepsis was observed between different medical centers. Brun-Buisson and colleagues examined the occurrence rates for sepsis in a large patient cohort (n = 11,828) admitted to 170 adult ICU s in public hospitals in France. In this cohort, 9% met criteria for sepsis, and 6.3% were noted to have severe sepsis [10]. The Italian SEPSIS study [11], published in 1995, prospectively examined patients admitted to 99 ICUs. These patients were categorized into the spectrum of sepsis definitions used the ACCP/SCCM criteria [4] to prospectively examine the incidence of mortality in patients with sepsis admitted to 99 ICUs. Most of the patients identified with sepsis were classified on admission to the ICU as having SIRS (52%), with fewer patients meeting criteria for sepsis (4.5%), severe sepsis (2.1 %), and septic shock (3%).

As can be noted from the wide range of occurrence rates in these studies, there remains a great deal of variability in descriptions of the epidemiology of sepsis. Most of the data are derived from hospitalized patients in academic tertiary-care centers. As a result, one cannot generalize these observations to hospital populations at large, and a lack of knowledge regarding the size and demography of the underlying populations served by these hospitals precludes estimation of incidence rates. Differences in patient characteristics studies, such as geographic location [12] or hospital type [9], further confound comparison across studies. More population-based studies are required to better delineate the incidence of sepsis in the general population, and further analysis of such data will help us to understand factors that influence the epidemiology of sepsis.

Outcome of patients with sepsis One of the great difficulties in understanding the process of disease in sepsis is related to the choice of which outcome to measure in epidemiological and thera-


peutic studies. Is it adequate to follow a patient for 24 hours after admission and state whether or not the patient is alive? Is a 28-day follow-up adequate for evaluation of patients with sepsis? One study has suggested that the risk of death in patients with sepsis may be increased up to 5 years after the initial episode of sepsis [13]. Should we be examining the effects of sepsis on quality of life, rather than examining binary variables such as mortality? Many of these questions remain unanswered. Most of the studies of patients with sepsis have focused on mortality, typically over a 28-day period of observation. Despite the difficulties in defining appropriate outcome markers for sepsis, we shall proceed to describe some of the epidemiological data available.

Great variability has been noted in the mortality rates quoted in studies describing the outcome of patients with sepsis, with most information detailing mortality rates in specific populations. The 1990 study by the Center for Disease Control noted that the mortality rate for septicaemia decreased from 31.0% in 1979 to 25.3% in 1989 [7]. The improvement in survival was attributed to changes in the popUlation of patients who developed sepsis and, possibly, to improvements in supportive care and pharmacological therapy. A recent metaanalysis of 131 studies reported an overall mortality rate of 49.7%, with the majority of studies suggesting mortality rates between 40% and 80% [14]. This review also suggested that the mortality rate from sepsis decreased over time, although this decrease was less than that noted in the study by the Center for Disease Control. The authors of this meta-analysis speculated that lack of success in therapy and differences in patient demographics (e.g., higher baseline severity of illness) over time may have contributed to a less satisfying decrease in sepsis mortality. The identification of gram-positive organisms as a suspected infectious culprit in cases of sepsis was noted to be increasing. One of the major difficulties in interpreting this review was the marked heterogeneity of patients from different studies, including case definition criteria and severity of illness [15]. For example, mortality from sepsis differed significantly, depending on whether documented bacteraemia was required for study eligibility. Furthermore, the review was mainly limited to the results of interventional trials in severe sepsis, and the extent to which study entry criteria biased their estimates (selection bias) is unknown.

Other studies have attempted to analyze the influence of severity of illness or presence of multiple organ system failure on observed mortality in sepsis. In the Italian SEPSIS study, patients who met criteria for SIRS had similar mortality to patients who were admitted to the ICU for reasons other than sepsis (26.5% vs. 24%). Mortality rates increased as patients met criteria for categories of higher illness severity (36% for patients with sepsis, 52% for patients with severe sepsis, and 82% in those with septic shock). In the prospective study of 3708 patients admitted to general medical wards and ICUs conducted by Rangel-Frausto et aI., this stepwise progression in mortality based on severity was similarly noted (7% mortality from SIRS, 16% mortality from sepsis, 20% mortality from severe sepsis, and 46% mortality from septic shock) [8]. Given the marked dif-


ferences in mortality between subgroups categorized by application of the ACCP/SCCM case definition criteria, it appears that these case definitions must capture important differences in patient characteristics.

Predicting outcome in sepsis and related syndromes Multiple authors have attempted to use measures of physiologic derangement, organ dysfunction and comorbidity to predict mortality in patients with sepsis. Brun-Buisson and colleagues demonstrated statistically-significant associations between mortality and SAPS II (Simplified Acute Physiology Score), the presence of two or more acute organ failures at time of sepsis, shock or a low pH [10]. However, the clinical significance of some of these associations may be questioned. Some of the odds ratios calculated in this study had wide confidence intervals with limits closely approaching 1.0 (e.g., more than two organs failing had an odds ratio association of 3.5, with a 95% confidence interval of 1.01-12). Some odds ratio magnitudes were very close to 1.0, despite being statistically significant (e.g., SAPS II had an odds ratio for mortality association of 1.03). Perl et al. found that APACHE II scores did not independently predict mortality in a cohort of patients with suspected gram-negative sepsis, although this study only examined 100 patients [16]. A retrospective cohort study by Lundberg et al. claimed to demonstrate a statistically-significant predictive value of APACHE II scores, with an odds ratio for likelihood of death of 2.64 (95% confidence interval 1.20-6.59) [17]. Given that this paper did not provide the APACHE II cut-off score used to classify patients in their odds ratio calculation, it is difficult to interpret this result. In addition, controlling for yeast vs. nonyeast sepsis in their multivariate model eliminated the APACHE II score as a significant independent predictor of mortality.

Others have attempted to customize different probability models in an attempt to make them better predictors of mortality in sepsis. One paper developed new logistic regression models based on the Simplified Acute Physiology Score II (SAPS II) and the 24-hour Mortality Probability Model II (MAPS II 24) by segregating patients with sepsis from the original model databases [18]. Knaus and colleagues generated a specific mortality predictive model derived from placebo group data from a large-scale sepsis therapeutic trial [19]. Both the Knaus et al. model and the APACHE III model had statistically-significant abilities to predict mortality. However, APACHE II was again demonstrated to be ineffective. Further refinements in predictive models derived from sepsis-specific and generalized databases may be helpful in improving our ability to identify patients at high risk of death.

The difficulty in predicting mortality in patients with sepsis is worsened by the heterogeneity of mortality rates in different patient populations. Attack rates for sepsis have been shown to differ at least three-fold between various academic medical centers [9]. Because the positive predictive value of a test


changes depending on the prevalence of disease in the tested population, application of any predictive model must take into account the population of patients to be studied. A comparison of mortality rates in sepsis between Chile, the United States and Italy demonstrated marked differences in mortality rates despite efforts to standardize case definition of sepsis [12]. These results suggest either inconsistent application of case definition of sepsis and various subclassifications, or marked differences in the disease process or treatment between countries.

Using inflammatory mediator levels as markers of sepsis In an effort to find markers of sepsis that more reliably identify the disease or its prognosis, many studies have measured inflammatory mediator concentrations, and several have explored the relationship between serum concentrations and outcome. However, there are a number of concerns related to their potential prognostic application. TN Fa levels have been shown to be higher in non-survivors versus survivors of septic shock [20-26]. Higher IL-6 levels have also been associated with increased mortality [25]. However, most of these studies have had very small sample sizes, and the cytokine levels overlap widely between survivors and non-survivors. Some recent work has also suggested that some patients may have a genetic predisposition to sepsis and/or mortality from sepsis, such as the TNF2 polymorphism of the TNFa gene promoter [27]. Again, other studies have failed to demonstrate a difference in survival, and some have argued that the same genetic polymorphism may increase the risk of death in some forms of sepsis and decrease the risk in others [28].

Changes in the epidemiology of sepsis over time Intuitively, one might imagine that the epidemiology of sepsis is changing. For example, the number of patients receiving ICU care has increased over time, the technologies used in the ICU have changed, and the choice and use of antibiotics has changed. Predisposing factors, such as chemotherapeutic regimens, have also changed, and there have been marked changes in antibiotic resistance. Furthermore, there have been wide changes in the microbiologic etiologies of diseases such as pneumonia and acute exacerbations of chronic bronchitis [29]. However, because of the lack of good case definitions and true incidence studies, we can only make inferences about whether the epidemiology of sepsis is truly changing.

The CDC study did suggest both changes in incidence and mortality but relied only on reported rates of septicaemia [7]. Unfortunately, that study is ten years old. Preliminary data by Linde-Zwirble et al. [30] using ICD-9-CM based


criteria based on the ACCP/SCCM guidelines suggest the total number of cases in the mid 1990s is significantly higher than prior estimates but they did not look at changes over time. The recent systematic review by Friedman et al. discussed above reported a fall in mortality over time (Fig. 1), but selection bias may confound the interpretation of these data [14].

The identification of gram-positive organisms as suspected infectious culprits in cases of sepsis has become more common (Fig. 2) [14]. A recent review of clinical trials in sepsis has suggested that subgroups of patients with sepsis and gram-positive infection were less likely to benefit from anti-inflammatory therapy compared with patients with gram-negative sepsis (Fig. 3) [31]. Thus, the increase in gram-positive infection in septic patients may explain why mortality rates have only modestly decreased over time.

Impact of therapy on outcome Our inability to demonstrate therapeutic options that have clearly demonstrable impact on mortality remains a source of great frustration for clinicians and investigators. Many different immune and inflammatory modulation therapies have been studied in therapeutic trials, such as anti-endotoxin antibodies [32-35], anti-TNF antibodies and receptor antagonists [2, 36-42], and ibuprofen [43, 44], to name a few. Although some Phase II studies suggested a possible decrease in mortality with treatment, early enthusiasm has been quashed by later negative Phase III studies [45]. Many authors have explored the reasons for our inability to demonstrate successful therapeutic options, several papers have proposed possible pathophysiological explanations for the failed hunt for an effective therapy for sepsis [46-48], and still others have examined issues related to study design, particularly with regards to choices for study end-points for therapy and sample sizes [49-51]. The marked heterogeneity of patients with sepsis, difficulty with case definitions, differences in study design, and the lack of strong mortality prediction models may be hampering our ability to identify and study subgroups of patients who may benefit from certain therapies in sepsis.

Implications for the cost-effectiveness of anti-sepsis therapies The high cost of potential therapeutic agents for sepsis highlights the need for evaluation of the cost-effectiveness of therapy. Analysis of the impact of predictive factors that could better identify patients with higher salvage potential suggested that the costs of treatment with HA-IA per year of life saved could be as low as $5,200 for those patients with an anticipated gain in life expectancy of 20 years. In patients with an anticipated gain in life expectancy of less than one year, costs could be as high as $110,000 per year of life saved [52]. However,


3000 c:::::::I SURV

80

Ii"l1l8J.II NS 2500 -b- MORTALlTY(%)

70

60 ~

E 2000 0

50 ::I. ~

f/) - 1500 c: 40 ~ Q) .. nI 0- 1000

:::0 OJ

30 ... III .-.

20 ~ 0

500 10

0 58-62 63-67 68-72 73-77 78-82 83-87 88-92 93-97

Year

Fig_ 1. Trends in mortality over time as illustrated in a meta-analysis of sepsis studies. From: Friedman G, Silva E, Vincent JL (1998) Has the mortality of septic shock changed with time. Crit Care Med 26:2078-2086. With permission

~Gram -ve

100 o Gram +ve

~ • Gram +ve and ~ 80 E .!!! c: 60 nI Cl ... 0 40 ..... 0 Q) Q. 20 ?::

0

1958-1969 1970-1979 1980-1989 1990-1997

Period of Publication

Fig. 2. Distribution of organisms causing infection in septic patients as illustrated in a meta-analysis of sepsis studies. From: Friedman G, Silva E, Vincent JL (1998) Has the mortality of septic shock changed with time. Crit Care Mcd 26:2078-2086. With permission


CD CD C CQ .c o ';!.

40

20

o

-20

-40

40

CD 20 CD C CO 0 .c o ~ ·20

Treatment Better Than Placebo

Treatment Worse Than Placebo

PAF Antagonist

a)

Treatment Better Than Placebo

-40 Treatment Worse Than Placebo

-60 CB6 BAYx1351

b)

Bradykinin Corticosteroids Antagonist

sTNFr.Fc(II) sTNFr:Fc(l) _

Fig. 3. Differences in mortality between placebo and treatment groups depending on the infecting microorganism (gram-negative = solid, gram-positive = hatched), as illustrated in a summary of trial results using various anti-inflammatory compounds (miscellaneous trials in Fig. 3a, multiple anti-tumor necrosis factor compound trials in Fig. 3b). From: Opal SM, Cohen J (1999) Clinical gram-positive sepsis: does it fundamentally differ from gram-negative bacterial sepsis? Crit Care Med 27:1608-1616. With permission


cost-effectiveness estimates based on clinical trials may underestimate cost when therapies are used outside of the tightly controlled protocols found in clinical trials [53]. Re-evaluation of less expensive therapeutic options, such as the recent trials of steroids in sepsis [54, 55], may be associated with more favorable cost-effectiveness profiles. Further developments in the consideration of quality-of-life adjustments as measures of therapeutic benetit may be required in order to evaluate the cost-effectiveness of therapy in a manner that is meaningful to patients [56].

Conclusion Despite multiple strategies for treating patients with sepsis, the associated mortality remains high. Better understanding of the epidemiology and pathophysiology of sepsis may allow research efforts to be redirected towards more cost-effective therapeutic strategies. More accurate mortality prediction may facilitate discrimination of differences between patient subgroups that could identify those likely to benefit from specific forms of therapy. Although we have made considerable strides in our understanding of this condition, our lack of impact on mortality will hopefully stimulate our efforts to vigorously pursue novel and creative approaches to the care of patients with sepsis.

References I. Kieft H, Hoepelman AI, Zhou W et al (1993) The sepsis syndrome in a Dutch university hos

pital. Clinical observations. Arch Intern Med 153:2241-2247 2. Reinhart K, Wiegand-Lohnert C, Grimminger F et al (1996) Assessment of the safety and ef

ficacy of the monoclonal anti-tumor necrosis factor antibody-fragment, MAK 195F, in patients with sepsis and septic shock: a multicenter, randomized, placebo-controlled, dose-ranging study. Crit Care Med 24:733-742

3. Pittet D, Thievent B, Wenzel RP et al (1996) Bedside prediction of mortality from bacteremic sepsis. A dynamic analysis of ICU patients. Am 1 Respir Crit Care Med 153:684-693

4. Bone RC, Balk RA, Cerra FB et al (1992) Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 101: 1644-1655

5. Vincent lL (1997) Dear SIRS, I'm sorry to say that I don't like you. Crit Care Med 25: 372-374

6. Knaus WA, Sun X, Nystrom 0 et al (1992) Evaluation of definitions for sepsis. Chest 101: 1656-1662

7. Center for Disease Control (1990) Increase in national hospital discharge survey rates for septicemia - United States, 1979-1987. Morbid Mortal Weekly Rep 39:31-34

8. Rangel-Frausto MS, Pittet D, Costigan M et al (1995) The natural history of the systemic inflammatory response syndrome (SIRS). A prospective study. lAMA 273: 117-123

9. Sands KE, Bates DW, Lanken PN et al (1997) Epidemiology of sepsis syndrome in 8 academic medical centers. Academic Medical Center Consortium Sepsis Project Working Group. lAMA 278:234-240


10. Brun-Buisson C, Doyon F, Carlet J et al (1995) Incidence, risk factors, and outcome of severe sepsis and septic shock in adults: a multicenter prospective study in intensive care units. JAMA 274:968-974

11. Salvo I, de Cian W, Musicco M et al (1995) The Italian SEPSIS study: preliminary results on the incidence and evolution of SIRS, sepsis, severe sepsis and septic shock. Intensive Care Med 21 [SuppIJ2:S244-S249

12. Angus DC, Dougnac A, Hernandez G et al (1996) Sepsis and SIRS: Are we any nearer to consensus? Intensive Care Med 22:273

13. Quartin AA, Schein RM, Kett DH et al (1997) Magnitude and duration of the effect of sepsis on survival. JAMA 277:1058-1063

14. Friedman G, Silva E, Vincent JL (1998) Has the mortality of septic shock changed with time. Crit Care Med 26:2078-2086

15. Cohen NH (1998) Reduced mortality from septic shock - lessons for the future. Crit Care Med 26: 1956-1958

16. Perl TM, Dvorak L, Hwang T et al (1995) Long-term survival and function after suspected gram-negative sepsis. JAMA 274:338-345

17. Lundberg JS, Perl TM, Wiblin T et al (1998) Septic shock: an analysis of outcomes for patients with onset on hospital wards versus intensive care units. Crit Care Med 26: I 020-1 024

18. Le Gall JR, Lemeshow S, Leleu G et al (1995) Customized probability models for early severe sepsis in adult intensive care patients. Intensive Care Unit Scoring Group. JAMA 273: 644-650

19. Knaus WA, Harrell FEJ, LaBrecque JF et al (1996) Use of predicted risk of mortality to evaluate the efficacy of anticytokine therapy in sepsis. The rhIL-l ra Phase III Sepsis Syndrome Study Group. Crit Care Med 24:46-56

20. Calandra T, Gerain J, Heumann D et al (1991) High circulating levels of interleukin-6 in patients with septic shock: evolution during sepsis, prognostic value, and interplay with other cytokines. The Swiss-Dutch J5 Immunoglobulin Study Group. Am J Med 91 :23-29

21. Damas P, Reuter A, Gysen P et al (1989) Tumor necrosis factor and interleukin-l serum levels during severe sepsis in humans. Crit Care Med 17:975-978

22. Marks JD, Marks CB, Luce JM et al (1990) Plasma tumor necrosis factor in patients with septic shock. Mortality rate, incidence of adult respiratory distress syndrome, and effects of methylprednisolone administration. Am Rev Respir Dis 141:94-97

23. de Groote MA, Martin MA, Densen P et al (1989) Plasma tumor necrosis factor levels in patients with presumed sepsis. Results in those treated with anti lipid A antibody vs placebo. JAMA 262:249-251

24. Debets JM, Kampmeijer R, van der Linden MP et al (1989) Plasma tumor necrosis factor and mortality in critically ill septic patients. Crit Care Med 17:489-494

25. Calandra T, Baumgartner JD, Grau GE et al (1990) Prognostic values of tumor necrosis factor/cachectin, interleukin-1, interferon-alpha, and interferon-gamma in the serum of patients with septic shock. J Infect Dis 161 :982-987

26. Pinsky MR, Vincent JL, Deviere J et al (1993) Serum cytokine levels in human septic shock. Relation to multiple-system organ failure and mortality. Chest 103:565-575

27. Mira JP, Cariou A, Grall F et al (1999) Association of TNF2, a TNF-a promoter polymorphism, with septic shock susceptibility and mortality. A multicenter study. JAMA 282: 561-568

28. Kumar A, Short J, Parrillo JE (1999) Genetic factors in septic shock. JAMA 282:579-581 29. Marrie TJ (1998) Community-acquired pneumonia: epidemiology, etiology, treatment. Infect

Dis Clin North Am 12:723-740 30. Linde-Zwirble WT, Angus DC, Carcillo J et al (1999) Age-specific incidence and outcome of

sepsis in the US. Crit Care Med 27:A33 31. Opal SM, Cohen J (1999) Clinical gram-positive sepsis: does it fundamentally differ from

gram-negative bacterial sepsis? Crit Care Med 27: 1608-1616


32. McCloskey RV, Straube RC, Sanders C et al (1994) Treatment of septic shock with human monoclonal antibody HA-IA. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 121: 1-5

33. Ziegler El, Fisher Cl lr, Sprung CL et al (1991) Treatment of gram-negative bacteremia and septic shock with HA-I A human monoclonal antibody against endotoxin. A randomized, double-blind, placebo-controlled trial. N Engl 1 Med 324:429-436

34. Greenman RL, Schein RM, Martin MA et al (1991) A controlled clinical trial of E5 murine monoclonal IgM antibody to endotoxin in the treatment of gram-negative sepsis. lAMA 266: 1097-1102

35. Bone RC, Balk RA, Fein AM et al (1995) A second large controlled clinical study of E5, a monoclonal antibody to endotoxin: results of a prospective, multicenter, randomized, controlled trial. Crit Care Med 23 :994-1006

36. Exley AR, Cohen 1, Buurman W et al (1990) Monoclonal antibody to TNF in severe septic shock. Lancet 335:1275-1277

37. Fisher CJJ, Opal SM, Dhainaut IF et al (1993) Influence of an anti-tumor necrosis factor monoclonal antibody on cytokine levels in patients with sepsis. The CB0006 Sepsis Syndrome Study Group. Crit Care Med 21 :318-327

38. Fisher CJJ, Agosti 1M, Opal SM et al (1996) Treatment of septic shock with the tumor necrosis factor receptor:fc fusion protein. N Engl 1 Med 334: 1697-1702

39. Cohen 1, Carlet 1 (1996) INTERSEPT: an international, multicenter, placebo-controlled trial of monoclonal antibody to human tumor necrosis factor-alpha in patients with sepsis. International Sepsis Trial Study Group. Crit Care Med 24: 1431-1440

40. Dhainaut IF, Vincent lL, Richard C et al (1995) CDP571, a humanized antibody to human tumor necrosis factor-alpha: safety, pharmacokinetics, immune response, and influence of the antibody on cytokine concentrations in patients with septic shock. CPD571 Sepsis Study Group. Crit Care Med 23: 1461-1469

41. Abraham E, Glauser MP, Butler T et al (1997) p55 Tumor necrosis factor receptor fusion protein in the treatment of patients with severe sepsis and septic shock. A randomized controlled multicenter trial. Ro 45-2081 Study Group. lAMA 277: 1531-1538

42. Abraham E, Wunderink RG, Silverman H et al (1995) Efficacy and safety of monoclonal antibody to human tumor necrosis factor alpha in patients with sepsis syndrome. A randomized, controlled, double-blind, multicenter clinical trial. lAMA 273:934-941

43. Bernard GR, Wheeler AP, Russell lA et al (1997) The effects of ibuprofen on the physiology and survival of patients with sepsis. The Ibuprofen in Sepsis Study Group. N Engl 1 Med 336:912-918

44. Haupt MT, lastremski MS, Clemmer TP et al (1991) Effect of ibuprofen in patients with severe sepsis: a randomized, double-blind, multicenter study. The Ibuprofen Study Group. Crit CareMed 19:1339-1347

45. Natanson C, Esposito Cl, Banks SM (1998) The sirens' songs of confirmatory sepsis trials: selection bias and sampling error. Crit Care Med 26: 1927-1931

46. Bone RC (1996) Immunologic dissonance: a continuing evolution in our understanding of the systemic inflammatory response syndrome (SIRS) and the multiple organ dysfunction syndrome (MODS). Ann Intern Med 125:680-687

47. Vincent lL (1998) Search for effective immunomodulating strategies against sepsis. Lancet 351 :922-923

48. Bone RC (1996) Why sepsis trials fail. lAMA 276:565-566 49. Inman KJ, Martin CM, Sibbald WI (1992) Design and conduct of clinical trials in critical

care. 1 Crit Care 7: 118-128 50. Bone RC (1995) Sepsis and controlled clinical trials: the odyssey continues. Crit Care Med

23: 1313-1315 51. Abraham E (1999) Why immunomodulatory therapies have not worked in sepsis. Intensive

Care Med 25:556-566


52. Schulman KA, Glick HA, Rubin H et al (1991) Cost-effectiveness of HA-IA monoclonal antibody for gram-negative sepsis. Economic assessment of a new therapeutic agent. JAMA 266:3466-3471

53. Linden PK, Angus DC, Chelluri L et al (1995) The influence of clinical study design on costeffectiveness projections for the treatment of gram-negative sepsis with human anti-endotoxin antibody. J Crit Care 10: 154-164

54. Bollaert PE, Charpentier C, Levy B et al (1998) Reversal oflate septic shock with supraphysiologic doses of hydrocortisone. Crit Care Med 26:645-650

55. Briegel J, Forst H, Haller M et al (1999) Stress doses of hydrocortisone reverse hyperdynamic septic shock: a prospective, randomized, double-blind, single-center study. Crit Care Med 27: 723-732

56. Chalfin DB, Cohen IL, Lambrinos J (1995) The economics and cost-effectiveness of critical care medicine. Intensive Care Med 21:952-961

Are There Useful New Markers of Sepsis?

M. MEISNER, K. REINHART

Severe infections and sepsis are common causes of morbidity and mortality in intensive care units. Clinical and laboratory signs of systemic inflammation such as changes in body temperature, leukocytosis, and tachycardia are frequently used for the diagnosis of infection or sepsis [1]. However, these signs and symptoms are neither specific nor sensitive for infection or sepsis. Various other and non-microbial infection related aetiologies of systemic inflammation may induce these symptoms. For example, patients suffering from pancreatitis, major trauma or burns present with a similar inflammatory response even in the absence of infectious complications. On the other hand, bacteriological evidence of infection, though definitive and specific, may not develop concurrently with clinical signs of sepsis.

Early diagnosis of sepsis with better markers would allow early treatment, imperative in reducing mortality and morbidity. Markers may be helpful in conducting trials with immunomodulatory therapeutics. Markers of sepsis capable of predicting the immune status of the septic patient may help target the population most likely to benefit from such therapeutics. An ideal method or marker of infection should be cheap, easy to measure, be highly specific and sensitive, allowing early diagnosis of sepsis, and should correlate with the severity of infection and help gauge the efficacy of therapeutic measures. Although numerous parameters and methods have been proposed, none fulfils all the requirements.

A common feature of the presently used parameters is that they are related to the systemic inflammatory response secondary to infection. Humoral and cellular elements of the immune response are activated during the immune response and induce numerous mediators and inflammatory-related molecules e.g. cytokines, chemokines, acute phase proteins and various other metabolites. Some of these parameters are increased to a various degree in patients with sepsis and infection as compared to patients without systemic inflammation or infection. Recent studies intend to evaluate to what degree these parameters reflect the activation of the systemic inflammatory response in patients with or without microbial infection. Among other parameters, C-reactive protein (CRP), interleukin 6 (IL-6) and procalcitonin were focused on for their possible clinical use for the diagnosis of sepsis, severe sepsis and septic shock.

138 M. Meisner, K. Reinhart

Procalcitonin

Procalcitonin is the 13-kDa propeptide of calcitonin which is normally produced in the C-cells of the thyroid glands. In healthy individuals, procalcitonin levels are very low « 0.1 ng/mI). In patients with sepsis, however, procalcitonin levels increase dramatically, sometimes to more than several hundred ng/ml [3, 8]. The site of procalcitonin production during sepsis and its biological function is unknown. In a recent experimental study, inhibiting procalcitonin action during sepsis improved survival whereas treating animals with procalcitonin decreased survival [27]. The basic biology of procalcitonin during sepsis deserves further investigation.

A number of studies support procalcitonin as a marker of severe infections and of sepsis. Patients with procalcitonin levels below or equal to 0.5 ng/ml are unlikely to have severe sepsis or septic shock. Increases in procalcitonin levels may indicate the presence of infection; however, the cutoff point may differ depending on the setting. In the intensive care setting, only levels above 1.5 ng/dl may identify sepsis [10]. A localized focus of bacterial infection, failing to induce systemic inflammation, does not usually induce increased procalcitonin concentrations.

Procalcitonin levels reflect the severity of the inflammatory/infectious response. Infections accompanied by severe systemic reactions and signs of organ dysfunction or poor peripheral perfusion more profoundly increase procalcitonin levels than infections with only moderate systemic response. Therefore, therapies effective in controlling sepsis and reducing severity of disease may lead to reductions in procalcitonin levels [17]. In pediatric patients, procalcitonin levels fell after successful antibiotic treatment [3].

Procalcitonin values are of prognostic significance in patients with bacterial sepsis [29]. In critically ill patients, outcome was best predicted with procalcitonin as compared with TNF-a, IL-6, and C-reactive protein [23, 29]. Importantly, commonly used signs of infection such as body temperature and leukocyte count were poor predictors of outcome. Procalcitonin may also be helpful in the differentiation between bacterial and viral infections. In neonates and children, those with bacterial meningitis had significantly higher levels of procalcitonin than those with viral meningitis [7]. Preliminary results suggest that procalcitonin may help differentiate an infectious from a non-infectious cause of the acute respiratory distress syndrome (ARDS) in adults [5], and systemic fungal or bacterial infections from episodes of graft rejection [12, 19,20]. Procalcitonin can also be induced in immunocompromized patients [2, 19,20]. Also, in patients with necrotizing pancreatitis, procalcitonin was the best predictor of infection of the pancreatic necrosis as compared to C-reactive protein and IL-8. The predictive power of procalcitonin was almost equal to that of fine needle biopsy, the gold standard [31]. However, these studies are based on small numbers of patients and need to be reproduced in a larger patient popUlation.

Are There Useful New Markers of Sepsis? 139

Procalcitonin may not or may only slightly increase when infection remains confined to a tissue or organ with no systemic manifestations. In patients with severe infection, appropriate therapy may lead to very low levels of procalcitonin which may not indicate eradication of the infection. In these patients, antibiotics or other therapeutic interventions may be indicated despite normal or low procalcitonin levels.

Procalcitonin concentrations exceeding 10 ng/ml almost exclusively occur in patients with severe sepsis or septic shock. However, procalcitonin levels may increase (median 1-2 ng/ml, occasionally values up to 10 ng/ml) during non-infectious inflammation such as major trauma, major surgery, birth stress, or therapies with TNF-a or OKT3. In these situations, further increases or decreases in procalcitonin level may be more informative than the absolute value. Procalciton in levels also increase during cardiogenic shock; however, the levels are considerably lower than those during septic shock [10].

C-reactive protein C-reactive protein is an acute phase protein released by the liver after onset of inflammation or tissue damage. During infections, C-reactive protein has both pro- and anti-inflammatory effects. C-reactive protein may recognize and adhere to pathogens and to damaged cells and mediate their elimination through interaction with inflammatory cells and mediators. However, C-reactive protein also prevents adhesion of neutrophils to endothelial cells, inhibits superoxide production, and increases IL-l receptor antagonist production. Although IL-6 is the main stimulus for the induction of C-reactive protein, other cytokines also play a role in its production [6]. C-reactive protein is a frequently used clinical marker to assess the presence and severity of an inflammatory response. Some studies support C-reactive protein as a marker of infection or of sepsis [15]. C-reactive protein has been found to differentiate patients with pneumonia from those with endotracheal infections [9], to aid the diagnosis of appendicitis [14], to assess severity of sepsis [30] or to differentiate bacterial and viral infections [36]. Other studies, however, point to important properties of C-reactive protein which limit its usefulness as a marker of severe infection and sepsis. First, plasma levels of C-reactive protein increase up to 24 hours later than those of other markers such as cytokines or procalcitonin [6, 25]. Second, plasma concentrations of C-reactive protein may increase during minor infections and do not adequately reflect severity of infection, nor differentiate between survivors and non-survivors of sepsis [23, 37]. Third, plasma levels remain elevated for up to several days even when infection is eliminated [23]. Lastly, C-reactive protein is also elevated during inflammatory states of non-infectious etiologies, e.g. autoimmune and rheumatic disorders [11, 34], myocardial infarction, malignant tumors, or postoperatively [24]. Probably because of these reasons, the predic-

140 M. Meisner, K. Reinhart

tive values of C-reactive protein in various patient populations can be poor for the diagnosis of sepsis [13] and less so when assessing severity of sepsis.

Cytokines Cytokines are peptides that regulate the amplitude and duration of the host inflammatory response [13]. Cytokines are released from various cells (blood and endothelial cells, macrophages, etc.) in response to infectious stimuli and bind to specific receptors of other cells, changing their behavior and defining their role in the inflammatory response.

Mean serum levels of cytokines are increased in septic as compared to nonseptic patients. Persistently high or increasing levels of cytokines are mostly found in non-survivors, whereas low and decreasing levels are found in survivors of sepsis. Despite the important role cytokines play in the pathogenesis of sepsis, they do not fulfill many requirements of a good marker. First, some cytokines are only released sporadically during severe infections and may bind to receptor antagonists, and therefore have a very short circulating half life. Second, cytokines are induced by numerous diseases other than sepsis or infection. Third, assays to determine plasma cytokine levels are mostly expensive and time consuming. Lastly, cytokine levels may vary or may be undetectable depending on the assay used.

TNF-a, IL-l, IL-6, IL-8, and IL-I0 are cytokines often associated with the presence of sepsis. Cytokines such as TNF-a, IL-l~, and IL-lO are not always detectable in patients with sepsis. Because of this, they do not correlate well with the clinical course of the patients [38] and are therefore not very helpful as markers of infection despite their major role in the pathogenesis of sepsis.

Among cytokines, only IL-6 and IL-8 are most closely related to the severity and outcome of patients with sepsis [38, 18] and high levels of IL-6 have been proposed as an additional inclusion criterion for an immune modulatory sepsis trial with antibodies directed against TNF-a [32]. In neutropenic patients, IL-8 and IL-6, but not C-reactive protein, were significantly different between microbiologically documented infections and unexplained fevers [13]. In neonates, increased plasma levels of both IL-6 and IL-8 can predict early onset of sepsis with a high sensitivity and specificity [4]. During sepsis, high plasma levels of IL-6 and IL-8 suggest an increased risk of complications and poor outcome [ 13]. However, in patients with pancreatitis, increased IL-8 plasma levels do not indicate infection of pancreatic necrosis with high sensitivity [31]. Furthermore, IL-6 and IL-8 can also be induced to a variable degree after major surgery [35], after major trauma [26], during acute exacerbations of autoimmune disorders [33, 21], during viral infections and after transplant rejection [16, 22].

In summary, among cytokines, only IL-6 and IL-8 may have limited utility as a marker of the presence, severity and outcome of sepsis. However, clinical


use of these expensive parameters is questionable, and whether IL-6 can be used as a very early predictor of infection or of sepsis deserves further investigation.

Conclusion

Early diagnosis of infection in critically ill patients is important to prevent complications of microbial infection. Patient history and physical examination with or without routine laboratory parameters frequently suffice to establish the diagnosis of sepsis. However, in some patients these same symptoms and parameters may also occur during systemic inflammation of non-infectious etiology. Furthermore, defining the severity of sepsis on clinical data alone may be difficult. Recently cytokines (specially IL-6 and IL-8), C-reactive protein, and proca1citonin have been proposed as markers of sepsis and infection. All these parameters have some merit depending on the clinical circumstances. C-reactive protein and procalcitonin are parameters with a different profile during various degrees of systemic inflammation and sepsis. Proca1citonin is superior for the diagnosis and especially for the follow-up of patients with a bacterial focus complicated by symptoms of severe sepsis and septic shock. Especially during the more severe stages of systemic inflammation, plasma procalcitonin concentrations correlate better with the severity of inflammation than other parameters. Also, proca1citonin concentrations rapidly decline after successful elimination of the infectious focus and with the disappearance of the inflammatory response. Procalcitonin can also be used to assess the efficacy of therapeutic measures in controlling the source of sepsis. Furthermore, it is a very good predictor of increased TNF-a and IL-6 levels in septic patients [28]. This suggests that proca1citonin may reflect the immune status of septic patients and may be helpful in recruiting patients in immunomodulatory trials. C-reactive protein is a more sensitive parameter for the diagnosis of infection, but is less specific for the diagnosis of sepsis and bacterial infection. Various etiologies other than microbial infection, such as autoimmune disorders, tissue trauma or viral infections, easily induce this acute phase protein. C-reactive protein is induced at high concentrations already during less severe stages of the systemic inflammatory response; this means that C-reactive protein is not very useful for the discrimination between the different stages of severity, e.g., between severe sepsis and septic shock. Also the increase and especially the decrease of C-reactive protein plasma levels is delayed by 24-48 hours as compared to proca1citonin. Although cytokines such as IL-6 and IL-8 correlate to some degree with the severity of sepsis and patient outcome, they have not yet been adequately evaluated for diagnosis and clinical decision-making at the bedside. For the diagnosis of bacterial infection without signs of systemic inflammation, procalcitonin is not superior to C-reactive protein. However, in patients with severe sepsis and septic shock proca1citonin seems to be a useful parameter to improve diagnosis and therapy monitoring in these severely ill patients.

142 M. Meisner. K. Reinhart

References I. American College of Chest Physicians/Society of Critical Care Medicine Consensus Confer

ence (1992) Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 20:864-87

2. AI-Nawas B, Shah PM (1996) Procalcitonin in patients with and without immunosuppression and sepsis. Infection 24:434-436

3. Assicot M, Gendrel D, Cars in H et al (1993) High serum procaicitonin concentrations in patients with sepsis and infection. Lancet 341:515-518

4. Berner R, Niemeyer CM, Leititis JU et al (1998) Plasma levels and gene expression of granulocyte colony-stimulating factor, tumor necrosis factor-alpha, interleukin (IL)-l beta, IL-6, IL-8, and soluble intercellular adhesion molecule-I in neonatal early onset sepsis. Pediatr Res 44:469-477

5. Brunkhorst FM, Forycki ZF, Wagner J (1995) Discrimination of infectious and non-infectious etiologies of the adult respiratory distress syndrome (ARDS) with procaicitonin immunoreactivity. Clin Intens Care 6:3-3

6. Brunkhorst FM, Heinz U, Forycki ZF (1998) Kinetics of procaicitonin in iatrogenic sepsis. Intensive Care Med 24:888-889

7. Chiesa C, Panero A, Rossi Net al (1998) Reliability of procalcitonin concentrations for the diagnosis of sepsis in critically ill neonates. Clin Infect Dis 26:664-672

8. Dandona P, Nix D, Wilson MF et al (1994) Procaicitonin increase after endotoxin injection in normal subjects. J Clin Endocrinol Metab 79: 1605-1608

9. Dev D, Wallace E, Sankaran R et al (1998) Value of C-reactive protein measurements in exacerbations of chronic obstructive pulmonary disease. Respir Med 92:664-667

10. De Werra I, Jaccard C, Corradin SB et al (1997) Cytokines, nitrite/nitrate, soluble tumor necrosis factor receptors, and procaicitonin concentrations: comparisons in patients with septic shock, cardiogenic shock, and bacterial pneumonia. Crit Care Med 25 :607 -613

II. Eberhard OK, Haubitz M, Brunkhorst FM et al (1997) Usefulness of proca1citonin for differentiation between activity of systemic autoimmune disease (systemic lupus erythematosus/ systemic anti neutrophil cytoplasmatic antibody-associated vasculitis) and invasive bacterial infection. Arthritis Rheum 40:1250-1256

12. Eberhard OK, Langefeld I, Kuse E et al (1998) Procaicitonin in the early phase after renal transplantation - will it add to diagnostic accuracy? Clin Transpl 12:206-211

13. Engel A, Mack E, Kern P, Kern WV (1998) An analysis of IL-8, IL-6 and CRP serum concentrations to predict fever, gram-negative bacteremia and complicated infection in neutropenic cancer patients. Infection 26:213-221

14. Eriksson S, Granstrom L, Olander B, Wretlind B (1995) Sensitivity of interleukin-6 and C-reactive protein concentrations in the diagnosis of acute appendicitis. Eur J Surg 161 :41-45

15. Gabay C, Kushner I (1999) Acute-phase proteins and other systemic responses to inflammation. N Engl J Med 340:448-454

16. Gendrel D, Raymond J, Assicot M et al (1998) Procaicitonin, C-reactive protein and interleukin 6 in bacterial and viral meningitis in children. Presse Med 27: 1135-1139

17. Gramm H-J, Dollinger P, Beier W (1995) Procalcitonin - ein neuer Marker der intlammatorischen Wirtsantwort. Longitudinalstudien bei Patienten mit Sepsis und Peritonitis. Chir Gastroenterolll [Suppl 2]:51-54

18. Hamano K, Gohra H, Noda H et al (1998) Increased serum interleukin-8: correlation with poor prognosis in patients with postoperative mUltiple organ failure. Word J Surg 22:1077-1081

19. Hammer S, Meisner F, Dirschedl Pet al (1998) Procalcitonin: a new marker for diagnosis of acute rejection and bacterial infection in patients after heart and lung transplantation. Transpl ImmunoI6:235-241

20. Kuse ER, Langefeld I, Jaeger K, Ki.ilpmann WR (1999) Procalcitonin in fever of unkown origin (FUO) following liver transplantation - a parametcr to differentiate acute rejection from infection. Crit Care Med (in press)


21. Kutukculer N, Caglayan S, Aydogdu F (1998) Study of pro-inflammatory (TNF-alpha, IL-1 alpha, IL-6) and T-cell-derived (IL-2, IL-4) cytokines in plasma and synovial fluid of patients with juvenile chronic arthritis: correlations with clinical and laboratory parameters. Clin Rheumatol 17:288-292

22. Malaguarnera M, Di Fazio I, Romeo MA et al (1997) Elevation of interleukin 6 levels in patients with chronic hepatitis due to hepatitis C virus. J Gastroenterol 32:211-215

23. Meisner M, Tschaikowsky K, Palmaers T, Schmidt J (1999) Comparison of procalcitonin (PCT) and C-reactive protein (CRP) plasma concentrations at different SOFA scores during the course of sepsis and MODS. Crit Care 3:45-50

24. Meisner M, Tschaikowsky K, Hutzler A et al (1998) Postoperative plasma concentrations of procalcitonin after different types of surgery. Intensive Care Med 24:680-684

25. Monneret G, Labaune JM, Isaac C et al (1997) Procalcitonin and C-reactive protein levels in neonatal infections. Acta Paediatr 86:209-212

26. Nast-Kolb D, Waydhas C, Gippner-Steppert C et al (1997) Indicators of the posttraumatic inflammatory response correlate with organ failure in patients with multiple injuries. J Trauma 42:446-454

27. Nylen ES, Whang KT, Snider RH et al (1998) Mortality is increased by procalcitonin and decreased by an antiserum reactive to procalcitonin in experimental sepsis. Crit Care Med 26: 1001-1006

28. Oberhoffer M, Karzai W, Meier-Hellmann A et al (1999) Sensitivity and specifity of various markers of inflammation for the prediction of tumor necrosis factor and interleukin-6 in patients with sepsis. Crit Care Med (in press)

29. Oberhoffer M, Vogelsang H, Russwurm S et al (1999) Outcome prediction by traditional and new markers of inflammation in patients with sepsis. Clin Chern Lab Med 37:363-368

30. Povoa P, Almeida E, Moreira P et al (1998) C-reactive protein as an indicator of sepsis. Intensive Care Med 24: 1052-1056

31. Rau B, Steinbach G, Gansauge F et al (1997) The role of procalcitonin and interleukin-8 in the prediction of infected necrosis in acute pancreatitis. Gut 41:832-840

32. Reinhart K, Wiegand-Lohnert C, Grimminger F et al (1996) Assessment of the safety and efficacy of the monoclonal anti-tumor necrosis factor antibody-fragment, MAK 195F, in patients with sepsis and septic shock: a multicenter, randomized, placebo-controlled, dose-ranging study. Crit Care Med 24:733-742

33. Robak T, Gladalska A, Stepien H, Robak E (1998) Serum levels of interleukin-6 type cytokines and soluble interleukin-6 receptor in patients with rheumatoid arthritis. Mediators Inflamm 7:347-353

34. Schwenger V, Sis J, Breitbart A, Andrassy K (1998) CRP levels in autoimmune disease can be specified by measurement of procalcitonin. Infection 26:274-276

35. Shenkin A, Fraser WD, Series J et al (1989) The serum interleukin 6 response to elective surgery. Lymphokine Res 8:123-127

36. Shaw AC (1991) Serum C-reactive protein and neopterin concentrations in patients with viral or bacterial infection. J Clin PathoI44:596-599

37. Ugarte H, Silva E, Mercan D (1999) Procalcitonin used as a marker of infection in the intensive care unit. Crit Care Med 27:498-504

38. Wakefield CH, Barclay GR, Fearon KC et al (1998) Proinflammatory mediator activity, endogenous antagonists and the systemic inflammatory response in intra-abdominal sepsis. Scottish Sepsis Intervention Group. Br J Surg 85:818-825

A Paradigm Shift: The Bidirectional Effect of Inflammation on Bacterial Growth

G.D. MEDURI

The ability to generate and respond to signaling molecules establishes a mechanism for regulated cell-to-cell communication. Cells coordinate their growth and proliferation with autocrine and paracrine signaling by means of low molecular weight polypeptides called cytokines. Innate or natural immunity is a highly conserved defense mechanism against infections found in all multicellular organisms [I]. The inflammatory reaction is a fundamental component of the innate immune response, and its most proximal expression is characterized by the elaboration of proinflammatory cytokines - tumor necrosis factor (TNF)-a and interleukin (IL)-l~. Response to cytokines is generally viewed as exclusive to cells containing a defined nucleus, since cytokines are intended to work on welldefined eukaryotic cells with consequent signal transduction events.

When proinflammatory cytokines are present at optimal concentration, they recruit both specific and nonspecific immune cells, nonlymphoid leukocytes (monocytes/macrophages, neutrophils, basophils, and eosinophils), and lymphocytes to the site of assault and activate them, thereby helping to eradicate the assault and to restore homeostasis [2]. However, there are occasions when the host defense response (HDR), in terms of inflammation, is exaggerated and protracted. In such cases, this primary defense process may instead cause enhanced tissue injury and maladaptive repair, leading to vital organ dysfunction and failure [3]. Reduction in the effective concentration of proinflammatory mediators is an important component in the resolution of inflammation [4].

The relationship between bacteria and inflammation is traditionally viewed as unidirectional. Bacteria trigger inflammation, and inflammation - as part of the host innate immune response - destroys bacteria and localizes the spread of infection. Although correct, this simple relationship does not provide a complete picture of the pathogen-host interaction in acute life-threatening infections. This unchallenged (preconditioned) view of the pathogen-host interaction has influenced for years the interpretation of objective clinical data in critical care medicine. In the first part of this chapter, I will briefly review the clinical literature on nosocomial infections in acute respiratory distress syndrome (ARDS). Then I will present the results of a prospective study of ARDS patients that investigated longitudinally the relationship between circulatory and pulmonary proinflammatory cytokine levels, infections, and outcome. The findings of this study, as

146 G.V. Meduri

well as those of other groups, generated a novel hypothesis, suggesting that bacteria may grow in the presence of excessive cytokine levels. The second part of this chapter presents the results of recent in vitro studies in support of this new hypothesis.

Clinical observation in ARDS ARDS is a frequent form of hypoxemic respiratory failure characterized by the acute development of diffuse lung inflammation. In mortality data, after day 3 of ARDS, most patients die following a prolonged period of ventilatory support, during which they often develop fever and other criteria for systemic inflammatory response syndrome (SIRS) [5], clinical manifestations of infection [6-8], and multiple organ dysfunction syndrome (MODS) [9, 10]. In the medical literature, sepsis is associated with fatality in 36 to 90% of ARDS nonsurvivors [6, 7, 9, 10]. At necropsy, 69% of ARDS non survivors have histologic evidence of pneumonia [11]. These observations led to the hypothesis that in ARDS a direct correlation may exist between development of nosocomial infections, amplification of the systemic inflammatory response, and higher mortality [12]. Support for this hypothesis, however, relied only on clinical studies that did not use strict criteria for diagnosing nosocomial infections. Furthermore, this broadly accepted pathophysiological hypothesis (second hit hypothesis) was never tested prospectively in ARDS.

Nosocomial infections and inflammation

Nosocomial infections and systemic inflammatory response in ARDS

We conducted a prospective study to investigate, at the onset of ARDS and during the progression of the disease, the longitudinal relationship between circulatory proinflammatory cytokine levels, infections, and outcome [13]. In most patients, the etiology of ARDS was pulmonary or extrapulmonary sepsis. We reported that at the onset of ARDS, and over time, nonsurvivors (n = 17) had significantly (p < 0.001) higher plasma TNF-a, IL-l~, and IL-6 levels than survivors (n = 17) did [14]. During the first week of ARDS, plasma cytokine levels declined in all survivors, whereas they remained persistently elevated in all nonsurvivors. Nosocomial infections were more frequent in patients with persistent cytokine elevation over time. The rate of nosocomial infection per day of mechanical ventilation was I % in survivors and 8% in non survivors. Moreover, none of the proven (n = 36) or suspected (n = 55) nosocomial infections caused either a transient or a sustained increase in plasma TNF-a, IL-I~, IL-6, and IL-8 levels above preinfection values [13]. This latter finding is in agreement with the recent understanding of downregulation (also called lipopolysaccharide

A Paradigm Shift: The Bidirectional Effect of Inflammation on Bacterial Growth 147

(LPS) tolerance) of an activated system (see discussion in reference 13). In these patients, a plasma IL-l ~ > 400 pg/ml on day 7 of ARDS was 100% accurate in predicting outcome [13]. Sixty-seven percent of nosocomial infections developed after day 10 of ARDS, and among non survivors 15 of 18 nosocomial infections developed while plasma IL-l ~ was > 400 pg/ml. In addition to our work [13], two additional studies have described an association between high circulating IL-6levels and increased rate of infections [15, 16].

Ventilator-associated pneumonia and pUlmonary inflammation in ARDS

The relationship between ventilator-associated pneumonia and pulmonary inflammation was evaluated in a series of prospective studies. We evaluated with bilateral bronchoalveolar lavage (BAL) 94 ARDS patients with 172 episodes of suspected ventilator-associated pneumonia (VAP) and compared BAL results from contralateral sites [17]. Thirty-three of the 55 (60%) positive bronchoscopies had significant (> 104 CFU/ml) growth in only one side. Episodes with bilateral significant growth were more likely to be polymicrobial, to have a bacterial growth> 105 CFU/ml in the BAL, and to possess a higher percentage of polymorphonuclear (PMN) cells and intracellular microorganisms. These BAL findings indicated that episodes with a higher bacterial burden had cytological evidence of a more intense local inflammatory response and were more likely to be diffuse. Postmortem studies have also described a strong association between number of bacteria and severity of local inflammation [18-20]. The traditional interpretation of these data would suggest that the more severe inflammation was the result of a higher bacterial burden; however, this relationship was challenged by the results of our prospective study [21].

In a longitudinal study of patients with ARDS subjected to bilateral BAL weekly and when clinical manifestations of YAP developed [21], we reported that at the onset of ARDS and over time, nonsurvivors had significantly (p < 0.001) higher BAL TNF-a, IL-l~, and IL-6levels than survivors did [21]. Nonsurvivors had a higher rate of ventilator-associated pneumonias than survivors [13]. In 21 episodes of YAP, 16 unilateral and 5 bilateral cases had excellent agreement between right and left BAL TNF-a, IL-l~, IL-6, total protein, and albumin levels. In other words, patients with unilateral pneumonia had similar TNF-a, IL-l~, and IL-6 levels in the BAL obtained from the lung with significant bacterial growth compared to the BAL from the contralateral lung without growth [13]. Furthermore, YAPs were not associated with either a transient or a sustained increase in BAL TNF-a, IL-l~, IL-6, and IL-8 levels above preinfection values [13]. In agreement with our results, the findings of a recent experimental study of gram-negative pneumonia indicated that persistent elevation in BAL proinflammatory cytokines is associated with failure to clear intrapulmonary bacteria despite a large influx of PMN in the airspaces [22].

Experimental and human studies have shown that a lung affected by ARDS is impaired in its ability to clear a bacterial challenge. Several intrinsic defects

148 G.V. Meduri

have been previously implicated, primarily those related to changes in the alveolar environment and the function of phagocytic cells [22]. Polymorphonuclear cells recruited into the airspaces of patients with ARDS have shown evidence of impaired microbicidal activity [23, 24]; this mechanism partly explains the lung's inability to clear bacteria in spite of intense local inflammation. Furthermore, PMN clearing of bacteria is dose-dependent, and the efficiency of PMN bactericidal activity decreases with increasing bacterial load [25].

Recent understanding of bacteria and cytokine interaction In the interaction between a microorganism and its host, the host's defense does not go unchallenged [26]. Several reports have shown that DNA viruses have the ability to interfere with extracellular cytokines or inhibit cytokine synthesis [26]. Until recently, very little was known of the ability of bacteria to interfere with or to utilize extracellular cytokines secreted by the host cells or intracellular cytokines within phagocytic cells. Recent reports have shown that certain bacteria have receptors for cytokines IL-l ~ and TNF-a and that exposure of bacteria to these cytokines enhanced their growth [27-29].

Receptors The surfaces of gram-negative bacteria have receptors for proinflammatory cytokines TNF-a and IL-l~ [27-29], and the virulence property of the bacterium is altered as a consequence of cytokine binding [29]. Porat et al. [27] reported that virulent strains of E. coli express receptors for IL-1 ~ and demonstrated enhanced extracellular in vitro growth in the presence of biologically active recombinant IL-l~. Luo et al. [29] reported that TNF-a could bind efficiently to many strains of gram-negative bacteria and that TNF-a bacterium complexes can interact with TNF-a receptors present on eukaryotic cells. They also showed that TNF-a binding enhanced bacterial invasion of HeLa cells and phagocytosis by human and murine macrophages [29].

Enhanced bacterial growth with cytokines Enhanced bacterial growth in the presence of cytokines has been reported for E. coli, IL-1 ~ [27], interferon-y [30], IL-2, granulocyte macrophage colony stimulating factor (GM-CSF) [31], and Staphylococcus aureus (IL-4) [32]. Two studies reported that the intracellular growth of Mycobacterium avium intracellular complex was enhanced in human peripheral blood monocytes activated with the cytokines IL-3, IL-6, and GM-CSF [33, 34].

Antiinflammatory cytokines have also been reported to promote bacterial growth. Two studies have shown that IL-O and IL-4 can enhance the intracellular


replication of bacteria. Park and Skerrett [35] reported that priming of human monocytes with IL-10 significantly enhanced the intracellular growth of Legionella pneumophila. Hultgren et al. [32] reported reduced growth of S. aureus in the joints of an IL-4-deficient mouse and showed that exposure of macrophages to IL-4 reduced intracellular killing of S. aureus without impairing phagocytosis.

New hypothesis and hypothesis testing The findings from our studies described above [13, 14, 21] suggested that final outcome in patients with ARDS is related to the magnitude and duration of the host inflammatory response, and that intercurrent nosocomial infections might be an epiphenomenon of prolonged intense inflammation. The increased rate of nosocomial infections might be explained by impaired host defense response. However, we hypothesized that cytokines secreted by the host during ARDS may indeed favor the growth of bacteria and explain the association between an exaggerated and protracted release of cytokines and the frequent development of nosocomial infections. To test this hypothesis, we conducted in vitro studies evaluating the extracellular and intracellular growth response of three clinically relevant bacteria in response to graded concentrations of proinflammatory cytokines TNF-a, IL-1~, and IL-6 [36, 37]. The bacteria used were fresh isolates of Staphylococcus aureus, Pseudomonas aeruginosa, and Acinetobacter sps. obtained from patients with ARDS. The bacteria were grown in 3 ml of RPMIIDMEM medium without serum or antibiotics. Intracellular growth was tested in human monocytic cell line V937 and in blood monocytes of normal healthy volunteers.

In these studies, we identified a V-shaped response of bacterial growth to proinflammatory cytokines. When the tested bacteria were exposed in vitro to a lower concentration (10 pg to 250 pg) ofTNF-a, IL-l~, or IL-6 - similar to the plasma values detected in ARDS survivors [14] - extracellular and intracellular bacterial growth was not promoted, and human monocytic cells were efficient in killing the ingested bacteria [36, 37]. By contrast, when bacteria were exposed to higher concentrations of these proinflammatory cytokines - similar to the plasma values detected in ARDS non survivors [14] - intracellular and extracellular bacterial growth was enhanced in a dose-dependent manner [36, 37]. Blockade by specific neutralizing monoclonal antibodies significantly inhibited cytokine-induced extracellular and intracelullar bacterial growth [36, 37].

The effects of cytokines on extracellular bacterial growth were seen only with fresh isolates and were lost after six in vitro passages [36]. These findings indicate that, in the host milieu, Staphylococcus aureus, Pseudomonas aeruginosa, and Acinetobacter sps. may acquire a phenotypic ability to use cytokines as growth factors, and that subsequent removal of these pathogens from such milieu (after six in vitro passages) resulted in the loss of the acquired pheno-

150 G.U. Meduri

type. This phenomenon of loss of responsiveness to cytokines was also recorded by Porat et al. [38].

The intracellular growth of Staphylococcus aureus, Pseudomonas aeruginosa, and Acinetobacter sps. was also tested after exposure of U937 monocytic cells to graded concentrations of lipopolysaccharide (LPS). At low priming concentrations of LPS, we observed a significant reduction in intracellular bacterial growth in comparison to control. However, at a priming concentration of LPS equal to or greater than 100 ng, all three bacterial isolates had a significant growth enhancement in comparison to control (all p < 0.000 1, for all three bacteria). Taken together, our findings indicate that there may be a threshold of cellular activation at which phagocytic cells effectively kill ingested bacteria (Fig. 1). Above this threshold of cellular activation, however, the intracellular micromilieu becomes favorable to the survival and replication of the ingested bacteria. It is likely that bacteria that are internalized, and under selective pressure may adapt to an otherwise hostile microenvironment by switching on novel gene expression that enables them to utilize cytokines as their growth factors.

We then exposed U937 monocytic cells primed with the highest concentration of LPS (10 Ilg) to escalating concentrations (0 Ilg, 25 Ilg, 50 Ilg, 75 Ilg, 100 Ilg, 150 Ilg, and 250 Ilg) of methylprednisolone and quantified both intracellular bacterial growth and the intracellular transcription of TNF-a, IL-I~, and IL-6. We found that exposure of LPS-primed U937 monocytic cells to methylprednisolone prior to infection affected (in a dose-dependent manner) the mRNA expression of TNF-a, IL-l~, and IL-6, and the in vitro intracellular bacterial growth of internalized Staphylococcus aureus, Pseudomonas aeruginosa, and Acinetobacter sps.

The impairment in intracellular bacterial killing correlated with the increased expression of proinflammatory cytokines, while restoration of monocyte killing function upon exposure to methylprednisolone coincided with the downregulation of the expression of TNF-a, IL-l~, and IL-6. We found that, at the two highest concentrations of methylprednisolone (150 Ilg and 250 Ilg), the mRNA expression of all three cytokines was significantly blunted, irrespective of the LPS concentration. Hence, we presume that bacterial survival and replication within the phagocytic cells are functions of the cytokines expressed by such cells. In the presence of excessive activation, the intracellular environment appears to favor the emergence of new phenotypes of bacteria that are capable of utilizing cytokines for their growth. By showing that methylprednisolone can reduce (in a dose-dependent manner) the mRNA expression ofTNF-a, IL-l~, and IL-6, and the intracellular bacterial growth of the tested bacteria, we provide experimental evidence to suggest a cause-and-effect relationship between excessive inflammation and bacterial growth.

It is unclear how bacteria may use cytokines for their growth, since bacteria are prokaryotes without a defined nucleus and cytokines are intended to work on well-defined eukaryotic cells with consequent signal transduction events. However, in a host milieu, bacteria may adapt to eukaryotic cellular processes

A Paradigm Shift: The Bidirectional Effect of Inflammation on Bacterial Growth lSI

1800 S. allrells 1600

1400

~1200 ... ~ 1000

..J 800 E :3 600 LL 0 400

200

0 0 10 100 250 1000 10000

1800 Concentrations of cytokines (pg)

Ps. aeruginosa '600

1400

'" ~ 12,00

>< -1000 ..J E 800 -.. ::l LL GOO 0

400

200

0 0 10 100 250 1000 10000

2500 Acillelobacler sps. Concentrations of cytoklnes (pg)

_2000 '" 0

>< -1600 ..J .g ::l LL 1000

0

GOO

0 0 10 100 2 60 1000 10000

Concentrations of cytoklnes (pg)

Fig. 1. Intracellular bacterial growth of S. aureus, P. aeruginosa, and Acinetobacter sps. in U937 cells primed with graded concentrations of IL-l~, IL-6, and TNF-<x. U937 cells (2 x 106) were primed with 0, 10, 100, 250 pg, and I and lang IL-I~, IL-6, and TNF-a. The primed cells were mixed with 6 x 106 CFU of each tested bacterium and incubated for 2 h. Extracellular bacteria were killed with gentamicin. The cells with internalized bacteria were incubated at 37°C for 12 h under an atmosphere of 5% CO2, The cells were then lysed, serially diluted, and cultured onto Luria-Bertani (LB) agar (Difco, Detroit, Mich.) plates. The CFU per milliliter was estimated after 16 to 18 h of incubation. A concentration-dependent biphasic growth response was observed for all three bacteria, and this response was similar among the tested cytokines. The intracellular growth of three bacteria progressively decreased as the concentration of priming cytokines increased from 10 to 250 pg. However, this trend reversed when the priming concentration reached and exceeded 1,000 pg, and intracellular bacterial growth increased manyfold (p < 0.000 I, for all cytokines). From: Kanangat S, Meduri GU, Tolley EA et al (1999) Effects of cytokines and endotoxin on the intracellular growth of bacteria. Infect Immun 67(6):2834-2840. With permission

152 G.V. Meduri

[40]. Although, the subsequent sequence of intracellular events has not been delineated, it is possible that bacteria might use cytokines through receptor-mediated, signal-transduction-induced activities that would require the presence of biochemical processes akin to those seen in eukaryotic cells; cytokines may act on bacteria through a signaling process similar to that of eukaryotes, but involving different biochemical pathways; or bacteria may break down cytokines into biologically active fragments that are transported across the bacterial cell membranes and act on specific gene transcription and translation.

The bidirectional effects of proinflammatory cytokines on bacterial growth may help explain the frequent occurrence of nosocomial infections in patients with unresolving ARDS. Table 1 shows the traditional versus alternative interpretations of clinical findings in ARDS. If nosocomial infections in unresolving ARDS are indeed an epiphenomenon of exaggerated inflammation, it follows that treatment modalities that effectively decrease cytokine synthesis may reduce the incidence or severity of nosocomial infections.

Table 1. Traditional versus alternative interpretation of clinical data on nosocomial infections and inflammation in ARDS

Inflammation and bacteria - Inflammation kills bacteria, vs - Regulated inflammation kills bacteria while excessive (unregulated) inflammation may enhance

bacterial growth Nosocomial infections are more frequent - in nonsurvivors, vs - in patients with persistent cytokine elevation Nosocomial infections - amplify inflammation (second hit hypothesis) and worsen MODS, vs - do not amplify inflammation (downregulation, or LPS tolerance) The progression of systemic inflammation in ARDS - is amplified by nosocomial infections (2:: day 3 of ARDS), vs - is determined prior to day 3, by the success and/or failure of the host regulatory mechanisms Glucocorticoid treatment in patients with unregulated systemic inflammation - causes immunosuppression and enhances the risk for developing infections, vs - if given in low doses for a prolonged period (2:: 7 days) may have an important

immunomodulatory effect in regulating excessive inflammation and restoring homeostasis


References 1. Suffredini AF, Fantuzzi G, Badolato R et al (1999) New insights into the biology of the acute

phase response. J Clin ImmunoI19(4):203-214 2. Meduri GU, Estes RJ (1995) The pathogenesis of ventilator-associated pneumonia: II. The

lower respiratory tract. Intensive Care Med 21 (5):452-461 3. Meduri GU (1996) The role of the host defence response in the progression and outcome of

ARDS: pathophysiological correlations and response to glucocorticoid treatment. Eur Respir J 9(12):2650-2670

4. Rossi AG, Haslett C (1998) Inflammation, cell injury, and apoptosis. In: Said SI (ed) Proinflammatory and antiinflammatory peptides, lung biology in health and disease. Marcel Dekker, New York, pp 9-24

5. Bone RC, Balk RA, Cerra FB et al (1992) Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 101(6):1644-1655

6. Montgomery AB, Stager MA, Carrico CJ et al (1985) Causes of mortality in patients with the adult respiratory distress syndrome. Am Rev Respir Dis 132(3):485-489

7. Andrews CP, Coalson n, Smith JD et al (1981) Diagnosis of nosocomial bacterial pneumonia in acute, diffuse lung injury. Chest 80(3):254-258

8. Seidenfeld n, Pohl DF, Bell RC et al (1986) Incidence, site, and outcome of infections in patients with the adult respiratory distress syndrome. Am Rev Respir Dis 134(1):12-16

9. Bell RC, Coalson n, Smith JD et al (1983) Multiple organ system failure and infection in adult respiratory distress syndrome. Ann Intern Med 99:293-298

10. Suchyta MR, Clemmer TP, Elliott CG et al (1992) The adult respiratory distress syndrome. A report of survival and modifying factors [see comments]. Chest 101(4): 1074-1079

11. Meduri GU (1993) Late adult respiratory distress syndrome. New Horiz 1(4):563-577 12. Bone RC (1996) Toward a theory regarding the pathogenesis of the systemic inflammatory re

sponse syndrome: what we do and do not know about cytokine regulation. Crit Care Med 24(1):163-172

13. Headley AS, Tolley E, Meduri GU (1997) Infections and the inflammatory response in acute respiratory distress syndrome. Chest 111 (5): 1306-1321

14. Meduri GU, Headley S, Kohler G et al (1995) Persistent elevation of inflammatory cytokines predicts a poor outcome in ARDS. Plasma IL-l beta and IL-6 levels are consistent and efficient predictors of outcome over time. Chest 107(4):1062-1073

15. Romani L, Mencacci A, Cenci E et al (1996) Imparied neutrophil response and CD4+ T helper cell development in interleukin 6-deficient mice infected with Candida albicans. J Exp Med 183:1345-1355

16. Wagner HNJ, Bennett ILl, Lasagna L et al (1956) The effect of hydrocortisone upon the course of pneumococcal pneumonia treated with penicillin. Bull Johns Hopkins Hosp 98: 197-215

17. Meduri GU, Reddy RC, Stanley T et al (1998) Pneumonia in acute respiratory distress syndrome. A prospective evaluation of bilateral bronchoscopic sampling. Am J Respir Crit Care Med 158(3):870-875

18. Rouby n, Martin De Lassale E, Poete P et al (1992) Nosocomial bronchopneumonia in the critically ill. Histologic and bacteriologic aspects. Am Rev Respir Dis 146(4): 1059-1066

19. Johanson WG Jr, Seidenfeld 11, Gomez P et al (1988) Bacteriologic diagnosis of nosocomial pneumonia following prolonged mechanical ventilation. Am Rev Respir Dis 137(2):259-264

20. Chastre J, Fagon JY, Bornet-Lecso M et al (1995) Evaluation of bronchoscopic techniques for the diagnosis of nosocomial pneumonia. Am J Respir Crit Care Med 152(1):231-240

21. Meduri GU, Kohler G, Headley S et al (1995) Inflammatory cytokines in the BAL of patients with ARDS. Persistent elevation over time predicts poor outcome. Chest 108(5):1303-1314

154 G.U. Meduri

22. Fox-Dewhurst R, Alberts MK, Kajikawa 0 et al (1997) Pulmonary and systemic inflammatory responses in rabbits with gram-negative pneumonia. Am 1 Respir Crit Care Med 155 (6):2030-2040

23. Martin TR, Pistorese BP, Hudson LD et al (1991) The function of lung and blood neutrophils in patients with the adult respiratory distress syndrome. Implications for the pathogenesis of lung infections. Am Rev Respir Dis 144:254-262

24. Chollet-Martin S, 10urdain B, Gibert C et al (1996) Interactions between neutrophils and cytokines in blood and alveolar spaces during ARDS. Am 1 Respir Crit Care Med 154(3 Pt 1):594-601

25. Clawson CC, Repine JE (1976) Quantitation of maximal bactericidal capability in human neutrophils. 1 Lab Clin Med 88(2):316-327

26. Kotwal GJ (1997) Microorganisms and their interaction with the immune system. 1 Leukoc Bioi 62(4):415-429

27. Porat R, Clark BD, Wolff SM et al (1991) Enhancement of growth of virulent strains of Escherichia coli by interleukin-l. Science 254(5030):430-432

28. Zav'yalov VP, Chernovskaya TV, Navolotskaya EV et al (1995) Specific high affinity binding of human interleukin I beta by CaflA usher protein ofYersinia pestis. FEBS Lett 371 (1): 65-68

29. Luo G, Niesel DW, Shaban RA et al (1993) Tumor necrosis factor alpha binding to bacteria: evidence for a high-affinity receptor and alteration of bacterial virulence properties. Infect Immun 61 (3):830-835

30. Hogan lS, Todhunter DA, Smith KL et al (1993) Growth responses of coliform bacteria to recombinant bovine cytokines. J Dairy Sci 76(4):978-982

31. Denis M, Campbell D, Gregg EO (1991) Interleukin-2 and granulocyte-macrophage colonystimulating factor stimulate growth of a virulent strain of Escherichia coli. Infect Immun 59(5):1853-1856

32. Hultgren 0, Kopf M, Tarkowski A (1998) Staphylococcus aureus-induced septic arthritis and septic death is decreased in IL-4-deficient mice: role of IL-4 as promoter for bacterial growth. J Immunol 160(10):5082-5087

33. Denis M, and Gregg EO (1990) Recombinant tumour necrosis factor-alpha decreases whereas recombinant interleukin-6 increases growth of a virulent strain of Mycobacterium avium in human macrophages. Immunology 71 (I): 139-141

34. Shiratsuchi H, 10hnson 1L, Ellner 11 (1991) Bidirectional effects of cytokines on the growth of Mycobacterium avium within human monocytes. 1Immunol 146(9):3165-3170

35. Park DR, Skerrett Sl (1996) IL-IO enhances the growth of Legionella pneumophila in human mononuclear phagocytes and reverses the protective effect of IFN-y: differential responses of blood monocytes and alveolar macrophages. 1Immunol 157(6):2528-2538

36. Meduri GU, Kanangat S, Stefan 1 et al (1999) Cytokines IL-I~, IL-6, and TNF-a enhance in vitro growth of bacteria. Am 1 Respir Crit Care Med 161 :961-967

37. Kanangat S, Meduri GU, Tolley EA et al (1999) Effects of cytokines and endotoxin on the intracellular growth of bacteria. Infect Immun 67(6):2834-2840

38. Porat RB, Clark D, Wolf SM et al (1992) IL-I~ and Escherichia coli I Letterl. Science 258: 1562-1563

39. Meduri GU, Kanangat S, Tolley E et al (1999) Effects of glucocorticoids on the intracellular growth of bacteria in lipopolysaccharide-primed human monocytic cells. First international congress on cytokines/chemokines in infectious disease: final program and abstract book. National Institute of Allergy & Infectious Diseases and International Immunocompromised Host Society, Bethesda, Maryland, pp 131

40. Falkow S (1997) Perspectives series: host/pathogen interactions. Invasion and intracellular sorting of bacteria: searching for bacterial genes expressed during host/pathogen interactions. 1 Clin Invest 100(2):239-243

Is the Dosing and Timing of the Intervention Adequate?

A.F. SUFFREDINI

During the last decade, major advances have been made in defining mechanisms of shock and organ injury due to infection. Intact host inflammatory responses are critical to limit and resolve serious infections. However, when these responses are intense they contribute to tissue injury and organ failure. Attempts to suppress inflammatory responses using corticosteroids or mediator-specific antagonists (e.g. tumor necrosis factor (TNF) antibodies) have either not improved survival or have been harmful in septic patients with life-threatening infections [1-3]. The rationale for the use of anti-inflammatory agents in sepsis seems well founded. The proinflammatory cytokines TNF and IL-l are detected during infections and when administered to humans or animals result in manifestations of shock. In some animal models, antagonists to TNF or IL-l protect them from lethal bacterial challenge [41. However, the inability to translate these observations into clinical use has led to a reassessment of this therapeutic approach.

Recent meta-analyses have evaluated the treatment effects of anti-inflammatory therapy in sepsis. High dose corticosteroid administration (9 trials comprising 1297 patients) was associated with higher mortality rates than in placebo-treated patients [1, 5, 6]. Nonsteroidal anti-inflammatory agents (21 trials comprising 10276 patients) were evaluated and included trials of TNF antibodies, TNF soluble receptor p55 or p75 constructs, IL-l receptor antagonist (ILI ra), platelet activating factor antagonist, bradykinin antagonist and ibuprofen, a cyclooxygenase inhibitor. When the treatment effects rather than the individual patients from these trials were pooled, a small (3%) beneficial effect, equal to a 7% reduction in mortality, was found [7]. However, to demonstrate these modest effects for any single agent would require large trials (more than 6000 patients) [7].

In contrast to the failure of mediator-specific anti-inflammatory therapy to provide benefit in sepsis, TNF and IL-l antagonists have been used successfully in the treatment of chronic inflammatory conditions such as rheumatoid arthritis (e.g. TNFR:Fc, IL-lra) and Crohn's disease (e.g. anti-TNF antibody) [8, 9]. Why have the same agents been unsuccessful in improving outcomes from septic shock? It may be because septic patients are composed of a heterogeneous population due to differences in underlying disease, causative agents, and severity and sites of infection. Further, many patients with sepsis will recover with

156 A.F. Suffredini

conventional therapy alone and not require anti-inflammatory therapy. Therefore, identifying high risk patients and defining the clinical pharmacology (i.e. timing and dosing) of these agents is a major challenge.

Timing of administration of mediator-specific anti-inflammatory therapy

The earliest study that evaluated passive immunization of mice against the lethal effects of TNF showed that these responses were both dose- and time-dependent. For example, endotoxin administration resulted in a rapid (within 60 min) rise in blood TNF levels which returned to baseline values after 3 h. When antiTNF antibodies were administered 3 to 6 h prior to a lethal injection of endotoxin, survival was improved. However, when given simultaneously with or 3 to 6 h after the endotoxin challenge, survival was unchanged or worsened. Further, very high doses of endotoxin could overcome the protective effect of the antibodies [10]. Thus, in a non-infectious model of endotoxemia, anti-TNF therapy had to be given prior to the amplification of the inflammatory cascade by TNF to observe beneficial effects.

While many animal models of sepsis have shown beneficial effects with steroids or mediator-specific anti-inflammatory agents, the majority of these studies were performed using prophylactic or simultaneous administration of the anti-inflammatory agent. Of 1660 animals described in models of sepsis and anti-inflammatory therapy, 86% of the animals received treatments prior to or less than 2 h after the onset of sepsis [11]. In contrast, in clinical trials none of the agents were given prior to the onset of sepsis. In one of the high dose corticosteroid trials, therapy was administered to 80% of the patients within 4 h of the diagnosis of sepsis [12]. In a trial, which attempted to expedite treatment with an anti-TNF antibody, the interval from the diagnosis of sepsis to the antibody infusion was on average 7 ± 5 (SD) h [13]. In a later trial using the same anti-TNF antibody, the duration of shock prior to the treatment infusion was 8.3 ± 4.8 (SD) h [14]. Thus, delays between the onset of sepsis and the initiation of anti-inflammatory therapies are an important problem in clinical trials.

Dosing of anti-inflammatory therapy Dosing of anti-inflammatory agents for septic patients remains a substantial challenge. Phase 1 trials are used primarily to examine kinetics, tolerance, and safety of new therapeutic agents [15]. Yet only limited data exist from clinical trials in sepsis regarding dose-dependent responses using clinical end-points such as survival, blood pressure, or organ failure. Further, some of these agents have a narrow therapeutic dose range. When a dimeric TNF p75 receptor con-

Is the Dosing and Timing of the Intervention Adequate? 157

struct (TNFR:Fc) was given to septic patients, a dose-associated increase in mortality occurred (placebo -30%, TNFR:Fc 0.15 mg/kg -30%, 0.45 mg/kg -48%, and 1.5 mg/kg -53% mortality) [3].

Because of patient heterogeneity, a weight-based dosing may be inadequate to account for the wide diversity in inflammatory responses that are present during septic shock. When levels ofTNF and IL-l were measured in 146 septic patients, detectable levels were found in only 14% and 29% of the patients respectively and the median values were low « 15 pg/ml). Levels of cytokine antagonists (soluble TNF receptors, IL-lra, and IL-6) were 300 to 100 OOO-foid in excess of their respective cytokines [16]. These data demonstrate the wide variation in blood pro inflammatory cytokines early in sepsis. In addition, the relative abundance of endogenous anti-inflammatory molecules questions the need to administer further anti-inflammatory agents.

Several trials have attempted to use blood cytokine levels to identify patients who might respond to anti-inflammatory therapies. Plasma TNF concentrations (> 20 pg/ml) did not predict responsiveness to anti-TNF antibody therapy in septic shock [14]. IL-6, a potent counter-regulatory cytokine, appeared promising as a biomarker to reflect the magnitude of the inflammatory response [17]. Yet, in a large trial of an anti-TNF antibody for treatment of septic shock, IL-6 levels greater than 1000 pg/ml at time of study entry failed to define a subgroup of patients who responded to anti-TNF therapy [14]. Two other anti-TNF trials were unable to relate high IL-6 levels to patients who responded to anti -TNF therapy [18, 19].

Inflammatory suppressing agents may have a dose-associated loss of activity. Normal subjects were given TNFR:Fc in doses identical to those used in the clinical sepsis trial [3] and then given an i.v. endotoxin challenge [20]. Despite the absence of plasma TNF bioactivity after TNFR:Fc, high dose TNFR:Fc was less immunosuppressive than low dose TNFR:Fc as measured by secondary cytokine and stress hormone responses [20]. The mechanism that accounts for this dose-related loss of activity may relate to extravascular sources of TNF or inflammatory pathways that are independent of circulating TNF.

A similar example of dose-associated loss of anti-inflammatory activity was observed in healthy subjects given i.v. endotoxin following the administration of soluble type I IL-l receptor (sIL-IRI). The highest dose of sIL-IRI was associated with enhanced levels of TNF and IL-8 and a lack of effect on any acute phase reactants [21]. These pro inflammatory effects were due to sIL-1 Rl binding of both IL-I and IL-I ra. As the dose of the sIL-I R 1 was increased, levels of IL-lra, an endogenous inflammatory suppressing molecule, decreased and this was associated with enhanced release of cytokines [21].

The dose of corticosteroids used in sepsis is undergoing reassessment. High dose methylprednisolone (approximately 5 - 10 g over 24 h) was associated with increased mortality in septic shock (2, 12) [1]. However, three small-randomized clinical trials (125 total patients) suggest that "stress-dose" steroids

158 A.F. Suffredini

(e.g. 300 mg of hydrocortisone/day x 5 days equivalent to 60 mg of methylprednisolone/day) decrease the time for shock reversal and improve mortality [22-24]. These beneficial effects appear independent of occult adrenal insufficiency and may be due to alterations in down-regulated corticosteroid and ~-adrenergic receptors [25]. Larger clinical trials are needed to confirm these observations.

Severity of illness and effects of anti-inflammatory therapy Appropriate patient selection for anti-inflammatory therapy remains a significant problem in sepsis trials. Eichacker et al. have analyzed the association between control mortality and the effects of anti-inflammatory agents on survival [26]. Anti-inflammatory agents were most beneficial when used in experimental models with high mortality rates (70-90%). In contrast, when the same agents are used in models of infection with less severe mortality « 50%) the agents either had no benefit or were harmful. These data suggest one reason for the failure of anti-inflammatory therapies in clinical trials. Control mortality in human sepsis trials is approximately 40% and no overall clinical benefit was observed. Thus, anti-inflammatory agents would have the greatest benefit in the patients who have the highest likelihood of dying by blocking harmful inflammatory processes, whereas in patients with less severe infections, suppressing inflammation should be avoided as these agents may be detrimental to host defenses [26].

Conclusions The clinical experience to date suggests that the original hypothesis regarding inflammation and septic shock needs to be further refined. Biomarkers are needed to identify high-risk patients and to direct dosing and timing of novel therapies for future sepsis trials. It is unlikely that isolated blood levels of cytokines will be sufficient for this purpose as they do not reflect the net effect of pro- and anti-inflammatory signals on cells [27]. Therefore, broad approaches are needed to reevaluate and refine current biomarkers as well as identify new host or microbial-derived biomarkers.

The targets for anti-inflammatory therapy may also need to be revised. For example, new late mediators of lethal shock, procalcitonin [28] and high mobility group-I protein [29] have recently been described. In contrast to the early appearance of TNF and IL-l after an inflammatory stimulus, these molecules appear 8 to 12 h after an inflammatory stimulus. Increased levels of procalcitonin are detected in sepsis. Administration of exogenous procalcitonin to septic animals increases mortality whereas anti-procalcitonin antibodies are protective [28]. High mobility group-l protein (HMG-I) is lethal to normal mice and ele-

Is the Dosing and Timing of the Intervention Adequate? 159

vated levels are present in septic patients [29]. A single dose of anti-HMG-l given 30 min before a lethal dose of endotoxin did not prevent death but when given as three doses (- 30 min, 12 h, and 36 h) a protective effect was observed that lasted over two weeks [29]. These preliminary data suggest new pathways of inflammation independent of TNF and IL-I that contribute to sepsis and may represent new targets for intervention. Whether the time course of procaicitonin or HMG-I appearance during severe infections makes these targets more amenable to clinical intervention remains to be elucidated.

One area that holds promise for providing new insights into the pathogenesis of sepsis and organ failure is the use of cDNA microarrays [30, 31]. cDNA microarrays provide a means to understand functional gene expression of complex diseases and syndromes such as sepsis by interrogating thousands of genes simultaneously. These data will help to define gene products involved in sepsis and assist in the development of biomarkers. Further, the method can be used to detect genetic polymorphisms that affect outcome [32]. New functions and relationships between mediators can be defined and this may open up new approaches for therapies [28]. While the application of this technology to septic shock will be a major challenge to medical informatics because of the generation of huge data sets, the potential to gain insight into the biology of host-infection interactions appears tremendous.

References 1. Zeni F. Freeman B, Natanson C (1997) Anti-int1ammatory therapies to treat sepsis and septic

shock: a reassessment. Crit Care Med 25: 1095-1100 2. Bone RC, Fisher CJ Jr, Clemmer TP et al (1987) A controlled clinical trial of high-dose

methylprednisolone in the treatment of severe sepsis and septic shock. N Engl J Med 317:653-658

3. Fisher CJ Jr, Agosti JM, Opal SM et al (1996) Treatment of septic shock with the tumor necrosis factor receptor:Fc fusion protein. The Soluble TNF Receptor Sepsis Study Group. N Engl J Med 334: 1697 -1702

4. Natanson C, Hoffman WD, Suffredini AF et al (1994) Selected treatment strategies for septic shock based on proposed mechanisms of pathogenesis. Ann Intern Med 120:771-783

5. Cronin L, Cook DJ, Carlet J et al (1995) Corticosteroid treatment for sepsis: a critical appraisal and meta-analysis of the literature. Crit Care Med 23: 1430-1439

6. Lefering R, Neugebauer EA (1995) Steroid controversy in sepsis and septic shock: a metaanalysis. Crit Care Med 23:1294-1303

7. Natanson C, Esposito CJ, Banks SM (1998) The sirens' songs of confirmatory sepsis trials: selection bias and sampling error. Crit Care Med 26: 1927-1931

8. Moreland LW, Baumgartner SW, Schiff MH et al (1997) Treatment of rheumatoid arthritis with a recombinant human tumor necrosis factor receptor (p75)-Fc fusion protein. N Engl J Med 337:141-147

9. Present DH, Rutgeerts P, Targan S et al (1999) Int1iximab for the treatment of fistulas in patients with Crohn's disease. N Engl J Med 340: 1398-1405

10. Beutler B, Milsark IW, Cerami AC (1985) Passive immunization against cachectinltumor necrosis factor protects mice from lethal effect of endotoxin. Science 229:869-871

160 A.F. Suffredini

11. Eichacker PQ (1997) The relevance of animal models in sepsis. In: The Second International Symposium on Infections in the Critically III Patient. Feb 14, Barcelona, Spain

12. The Veterans Administration Systemic Sepsis Cooperative Study Group (1987) Effect of high-dose glucocorticoid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 317:659-665

13. Abraham E, Wunderink R, Silverman H et al (1995) Efficacy and safety of monoclonal antibody to human tumor necrosis factor a in patients with sepsis syndrome. JAMA 273:934-941

14. Abraham E, Anzueto A, Gutierrez G et al (1998) Double-blind randomised controlled trial of monoclonal antibody to human tumour necrosis factor in treatment of septic shock. NORASEPT II Study Group. Lancet 351 :929-933

15. Rolan P (1997) The contribution of clinical pharmacology surrogates and models to drug development - a critical appraisal. Br J Clin PharmacoI44:219-225

16. Goldie AS, Fearon KC, Ross JA et al (1995) Natural cytokine antagonists and endogenous antiendotoxin core antibodies in sepsis syndrome. The Sepsis Intervention Group. JAMA 274:172-177

17. Reinhart K, Wiegand-Lohnert C, Grimminger F et al (1996) Assessment of the safety and efficacy of the monoclonal anti-tumor necrosis factor antibody-fragment, MAK 195F, in patients with sepsis and septic shock: a multicenter, randomized, placebo-controlled, doseranging study. Crit Care Med 24:733-742

18. Abraham E, Glauser MP, Butler T et al (1997) p55 Tumor necrosis factor receptor fusion protein in the treatment of patients with severe sepsis and septic shock. A randomized controlled multicenter trial. Ro 45-2081 Study Group. JAMA 277: 1531-1538

19. Kay C (1996) Can better measures of cytokine responses be obtained to guide cytokine inhibition? In: Institutes' CH (ed) Designing Better Drugs and Clinical Trials for Sepsis/SIRS: Reducing Mortality to Patients and Suppliers. Feb 20-21, Washington, DC

20. Suffredini AF, Reda D, Banks SM et al (1995) Effects of recombinant dimeric TNF receptor on human inflammatory responses following intravenous endotoxin administration. J Immunol 155:5038-5045

21. Preas HL, Reda D, Tropea M et al (1996) Effects of recombinant soluble type I IL-l receptor on human inflammatory responses to endotoxin. Blood 88:2465-2472

22. Bollaert PE, Charpentier C, Levy B et al (1998) Reversal of late septic shock with supraphysiologic doses of hydrocortisone. Crit Care Med 26:645-650

23. Briegel J, Forst H, Haller M et al (1999) Stress doses of hydrocortisone reverse hyperdynamic septic shock: A prospective, randomized, double-blind, single center study. Crit Care Med 27:723-732

24. Chawla K, Kupfer Y, Goldman I, Tessler S (1999) Hydrocortisone reverses refractory septic shock (abstract). Crit Care Med 27[Suppl]:A33

25. Meduri GU, Kanangat S (1998) Glucocorticoid treatment of sepsis and acute respiratory distress syndrome: time for a critical reappraisal. Crit Care Med 26:630-633

26. Parent C, Natanson C, Cui X et al (\999) Control mortality and exclusion criteria: factors potentially altering the effects of anti-inflammatory therapies in clinical sepsis. Am J Resp Crit Care Med 159:A263

27. Brandtzaeg P, Osnes L, Ovstebo R et al (1996) Net inflammatory capacity of human septic shock plasma evaluated by a monocyte-based target cell assay: identification of interleukin-10 as a major functional deactivator of human monocytes. J Exp Med 184:51-60

28. Nylen ES, Whang KT, Snider RH Jr et al (1998) Mortality is increased by proca1citonin and decreased by an antiserum reactive to proca1citonin in experimental sepsis. Crit Care Med 26: 1001-1006

29. Wang H, Bloom 0, Zhang M et al (1999) HMG-1 as a late mediator of endotoxin lethality in mice. Science 285:248-251

30. The Chipping Forecast (1999) Nature Genetics January: 1-60 31. Iyer VR, Eisen MB, Ross DT et al (1999) The transcriptional program in the response of hu

man fibroblasts to serum. Science 283:83-87 32. Ramsay G (1998) DNA chips: state-of-the art. Nat BiotechnoI16:40-44

Clinical Trials of Mediator-Targeted Therapy in Sepsis

J.e. MARSHALL

Over the past two decades, more than five dozen phase II and phase III randomized controlled trials have been undertaken, testing the hypothesis that modulation of the host inflammatory response can improve survival in sepsis. Despite a compelling pre-clinical rationale, and promising preliminary data, none of these have led to the licensing of new therapies. It has even been suggested that the recurring failure of well-designed clinical studies to demonstrate improved survival for a variety of different therapeutic approaches should lead to a moratorium on clinical research, until the reasons for this failure are better understood [1]. Yet important insights can be gleaned from the body of work that has been performed to date. The biologic concept remains viable; the challenges involved in clinical investigation are formidable.

This chapter reviews the progress of clinical trials of mediator-directed therapies in sepsis, and the insights they provide to guide future work.

The biologic rationale A large body of experimental research has demonstrated conclusively that the lethality and morbidity of infection arise through the response of the host, rather than through the intrinsic toxicity of the infecting organism. Michalek and colleagues, for example, demonstrated that the detrimental consequences of endotoxin occur as a result of the activation of host immune cells. Using congenic mice differing in a single gene that confers sensitivity to endotoxin, these authors showed that endotoxin susceptibility could be transferred to resistant animals by infusion of bone marrow cells from sensitive animals, and that resistance was conferred by repopulating the sensitive strain with bone marrow-derived cells from the resistant animals [2]. Subsequent studies have identified a large number of endogenous mediators that are produced during acute inflammation, and that contribute to its deleterious sequelae. For example, even in the absence of antibiotic therapy, baboons could be protected against the effects of a lethal intravenous challenge with live E. coli if they were first treated with a neutralizing antibody to tumor necrosis factor [3]. Similarly rabbits can survive a live bacterial challenge if the cytokine, interleukin-l (IL-l) is neutralized with

162 J. C. Marshall

its naturally occurring inhibitor, the interleukin-l receptor antagonist (lL-1 ra) [4]. Literally hundreds of studies using dozens of animal models confirm a common theme: modulation of the host inflammatory response can alter survival, even in the absence of such traditional anti-infective measures as fluid resuscitation, antibiotics, and surgery [5].

The dramatic results seen in animal models suggested that a cure for human sepsis was just around the corner, and a number of strategies have been tested in randomized clinical trials. Unfortunately the situation in the human has proven to be much more complex.

The targets

Endotoxin (lipopolysaccharide)

Endotoxin or lipopolysaccharide comprises approximately 10% of the weight of a Gram negative bacterium [6]. It is a potent trigger for the release of host derived mediators of inflammation. Endotoxin, complexed to the plasma protein lipopolysaccharide binding protein (LBP), binds to the CD 14 receptor on macrophages and neutrophils [7], initiating the expression of a number of proinflammatory genes. Because Gram negative organisms have traditionally been a common cause of infection in the critically ill [8, 9], and because endotoxin can playa role in the pathogenesis of other infectious processes by virtue of its absorption from the gastrointestinal tract [10, 11], the endotoxin molecule is an attractive therapeutic target.

Several strategies to neutralize endotoxin have been studied. Monoclonal antibodies reactive against the common core polysaccharide of the endotoxin molecule have been produced using rough mutant strains of Gram negative organisms. Two of these - HA-l A and ES - have undergone evaluation in large clinical trials [12-15] (Table 1). The first study of 15 anti-serum, the forerunner of HA-IA, showed improved survival (78% for treated patients vs. 61 % for controls) [16]. A large multi-centre randomized trial of HA-l A also showed improved survival for patients with Gram negative infections, but found no evidence of benefit for patients who did not have Gram negative infection [12]. A subsequent large study, however, failed to reproduce the findings of benefit for patients with Gram negative infections, and more disturbingly, showed a trend towards harm in patients infected with Gram positive organisms [13]. Studies using ES showed no evidence of a survival benefit [14], but did suggest the possibility of reduced organ dysfunction for patients treated with the antibody [151-

Other approaches have been evaluated. An antibody against the enterobacterial common antigen, T88, failed to improve survival in a study of 836 patients (Panacek et al. unpublished). Similarly the amino acid, taurolidine, was without obvious benefit in a small study of one hundred patients r 17]. On the other hand, recombinant bactericidal permeability increasing protein (BPI), an en-

Clinical Trials of Mediator-Targeted Therapy in Sepsis 163

Table 1. Randomized trials of anti-endotoxin therapies

Agent No. of % mortality

Author patients Placebo Treated

J5 212 39 22 Ziegler 15 100 49 50 Calandra HA-IA 543 43 39 Ziegler HA-IA 2199 32 33 McCloskey E5 39 30 34 Greenberg E5 468 41 40 Greenman E5 847 Bone T88 826 34 37 Panacek Taurolidine 100 39 44 Willatts

dogenous protein with endotoxin-neutralizing capability, showed quite striking evidence of benefit in a phase II trial conducted in pediatric patients with meningococcaemia, reducing the mortality in this condition to less than 4% [18]. A phase III study of BPI in meningococcaemia has recently been completed, however results are not yet available.

Other strategies are currently undergoing evaluation. The antibiotic, polymyxin B, has endotoxin-neutralizing activity, but its intrinsic toxicity limits its clinical utility [19]. It has been complexed with dextran resulting in the elimination of its toxicity, but maintenance of biologic activity. This agent, termed PMX 622, produces striking abrogation of signs and symptoms following endotoxin administration to healthy human volunteers [20]; it is currently undergoing phase II studies. Endotoxin is also bound by HDL, and the potential utility of HDL as an anti-sepsis therapy is currently under study in Europe.

Studies of anti-endotoxin therapy illustrate a number of the pitfalls of sepsis clinical research. An agent that neutralizes endotoxin would only be predicted to benefit patients in whom endotoxaemia is present. The clinical criteria of sepsis syndrome [21] used to identify patients for these trials do not necessarily identify such a population of patients. Just over one third of patients enrolled in the multi-centre trial of HA-IA actually had Gram negative infections, and it was only in this group that benefit was evident [12]. Analysis of a small subgroup of patients in this trial showed that the greatest mortality reduction occurred in those patients in whom endotoxaemia could be documented [22]. The lack of a rapid assay for endotoxin that might identify patients who can benefit from therapy, has been a significant impediment to the evaluation of such therapies. The suggestion that endotoxin neutralization may be harmful for patients without Gram negative infections underlines the importance of identifying the appropriate population for study [13]. The preclinical literature clearly shows that benefit or harm is model dependent, and that an agent that improves survival in one model, may actually worsen it in another [5]. Finally, concerns have been ex-

164 J. C. Marshall

pressed that at least one of the agents used in these studies, HA-IA, did not adequately neutralize endotoxin activity and that its in vivo ability to bind endotoxin has never been adequately demonstrated [23].

Tumour necrosis factor (TNF)

Binding of endotoxin to the CD14 receptor on the macrophage triggers gene expression for a variety of inflammatory mediators, although their release can be elicited as well by other infectious and non-infectious stimuli. One of the most important of these is the cytokine tumour necrosis factor (TNF), detectable within 90 minutes of endotoxin stimulation both in vivo and in vitro [24]. TNF is a potent early mediator of systemic inflammation. Its administration to animals [25] or humans [26] reproduces the characteristic clinical manifestations of a septic response, while its neutralization in vivo can protect experimental animals against a lethal challenge with live bacteria or endotoxin [3, 25]. However the effects of TNF neutralization are not always beneficial: a number of experimental studies, generally those involving exposure to a large inoculum of live organisms, have shown increased mortality when TNF bioactivity is disrupted [27,28].

TNF is an early mediator of systemic inflammation. Following an intravenous bolus of endotoxin, its levels peak at 90 minutes then rapidly decline. It is variably detected in the circulation of septic critically ill patients [29, 30]. Indeed circulating levels of TNF do not correlate particularly well with the ultimate risk of mortality, and abnormalities of expression of cell-associated TNF [31], or of the monocyte receptor for TNF [32] are more reliable discriminators of patients at risk of dying.

Two strategies have been employed to neutralize TNF activity in vivo - neutralizing monoclonal antibodies, and soluble TNF receptor constructs composed of one of the two TNF receptor molecules complexed to an immunoglobulin.

Nine phase II or III clinical trials have evaluated the effects of neutralizing TNF using a monoclonal antibody (Table 2). While none of these has independently shown a significant reduction in mortality, pooled data show a small, but statistically significant mortality reduction from 42.4% in placebo treated patients to 38.9% in those receiving anti-TNF antibody (p = 0.03). A large phase III North American study involving 2,634 patients has recently completed patient accrual and results should be available by the end of the year.

In contrast to the results of studies using TNF monoclonal antibodies, those using TNF receptor constructs have not found the same beneficial effects (Table 3). This may in part be a consequence of a single study in which administration of a p75 TNF receptor construct resulted in a dose-dependent increase in mortality [33]. It was later demonstrated that this molecule bound TNF and delayed its clearance from the circulation, but did not neutralize its activity.

Clinical Trials of Mediator-Targeted Therapy in Sepsis

Table 2. Randomized trials of anti-TNF antibodies

Agent No. of % mortality patients Placebo Treated

CB006 80 32 44 CDP57I 42 60 63 cA2 56 39 36 Bay X 1351 971 33 30 Bay X 1351 553 40 37 Bay X 1351 1878 43 40 MAK 195F 122 41 47 MAK 195F 39 50 26 MAK 195F 446 57 54 TOTALS 4187 42.4 38.9 (*)

(*) Odds ratio 0.87 (0.76 - 0.98)

Table 3. Randomized trials of soluble TNF receptor constructs

Agent No. of % mortality patients Placebo Treated

p75 141 30 45 p55 498 39 38 p55 1362 28 27

165

Author

Fisher Dhainaut Clark Abraham Cohen Abraham Reinhart Fisher Reinhart

Reference

Fisher Abraham Abraham

There are a number of explanations for the limited success of neutralization of TNF in clinical sepsis. In the first place, TNF is only one of dozens of pro-inflammatory molecules, and, like endotoxin, it is not invariably present in the plasma of patients meeting criteria for sepsis syndrome [29]. Furthermore, it is an early mediator whose predominant biologic effects may have already occurred by the time a patient presents for treatment. Monoclonal antibodies are difficult to produce in large quantities, and their biologic activity may be lost. A study recruiting close to 1800 patients, the NORASEPT II Trial, demonstrated a statistically insignificant mortality reduction of 2.5%. Analysis of cytokine levels showed that the antibody effected only modest neutralization ofTNF when measured using an immunoassay, and no neutralization when measured by bioassay (unpublished). Finally, variability in concomitant care can have a significant impact on study outcomes. An analysis of data from a randomized trial of an antibody to TNF showed that the greatest evidence was seen in patients receiving optimal surgical, antimicrobial, and concomitant care [34]. These observations have been reproduced in an as yet unpublished European study of a different monoclonal antibody to TNF that found a mortality reduction of 10% for patients treated with the antibody who also received optimal surgical and antibiotic therapy; the overall mortality reduction on an intent to treat basis was only 3.7%.

166 J.C. Marshall

Interleukin-l (IL-l) Interleukin-l was one of the first endogenous mediators of inflammation to be identified [35]. Its presence had previously been inferred through studies showing that fever was mediated by a host-derived substance called endogenous pyrogen: endogenous pyrogen proved to be interleukin-l. Interleukin-l is a product of activated macrophages that exerts a number of physiologic effects including temperature elevation, hypotension, hypoferraemia, and activation of an acute phase response [35]. Release of interleukin-l is accompanied by the release of a naturally occurring competitive antagonist of IL- I, the interleukin- I receptor antagonist (IL-lra) [36]. IL-lra shares significant homology to IL-l and binds to the same cellular receptor, but fails to initiate signal transduction, and therefore fails to activate the cell. IL-lra improves outcome in a number of animal models of systemic challenge with endotoxin or live organisms. Like TNF, however, 11-1 itself is protective in certain models of uncontrolled infection [37].

Recombinant IL-lra has been evaluated in three randomized studies. The first of these, a small phase II trial, showed a dramatic dose dependent reduction in mortality for patients treated with IL-lra [38]. The mortality benefit seen in a subsequent multicenter trial was much smaller, and failed to achieved statistical significance on an intent to treat basis [33]. A second follow-up study that enrolled a sicker subset of patients was terminated at an interim analysis when the observed mortality benefit was much lower than expected [39], and investigation of IL-lra as a therapy in sepsis was discontinued. Pooled data from these three studies, however, do show a significant mortality reduction for the use of interleukin-l receptor antagonist in patients with sepsis syndrome (Table 4).

Table 4. Randomized trials of IL-lra

Agent

IL-lra IL-lra IL-lra Total

(*) Odds ratio 0.80 (0.65 - 0.99)

No. of patients

99 893 696 1688

% mortality

Placebo Treated

44 34 36 35.5

24 30 33 30.6 (*)

Author

Fisher Fisher Opal

Like TNF, IL-l is merely one of a number of overlapping and redundant mediators of inflammation, therefore the limited benefits of its neutralization may not be surprising. Moreover its effects vary with the inflammatory challenge. Subset analyses of the human data show the greatest evidence of benefit for sub-


groups of patients with polymicrobial infections or significant organ dysfunction. On the other hand, a trend towards harm was evident in patients with no risk factors [33].

Strategies to enhance the inflammatory response

The hypothesis of studies that neutralize endotoxin, tumour necrosis factor, or IL-I is that excessive inflammation is responsible for the morbidity of critical illness. However the alternative hypothesis - that the host response is deficient -has also been tested. Yolk and colleagues evaluated the effects of recombinant interferon-y in a small study of septic patients [40]. They showed that reduced expression of HLA-DR on the monocyte was associated with adverse outcome, and that administration of recombinant interferon-y could restore HLA-DR expression. However the study was too small to determine whether clinical outcome was improved. Granulocyte colony stimulating factor (G-CSF) is a myeloid cell growth factor that increases neutrophil numbers by stimulating their synthesis and release from bone marrow, and inhibiting the normal expression of programmed cell death or apoptosis [41]. G-CSF has been evaluated in a cohort of patients with community acquired pneumonia [42], and in a separate study of critically ill patients with ventilator associated pneumonia (unpublished). Neither of these studies showed significant mortality benefit.

Although the concept of augmenting an already activated inflammatory response may seem counter-intuitive, preclinical studies do show benefit in some models [43]. G-CSF consistently improves survival in experimental peritonitis, whether the intervention is given prior to or after the induction of peritonitis. On the other hand, the pre-clinical literature fails to show any benefit for G-CSF in experimental models of pneumonia, mirroring the results of human studies.

Platelet activating factor

Platelet activating factor (PAF) is a lipid autocoid that is a potent proinflammatory molecule [44]. Several different synthetic antagonists of the PAF receptor had been evaluated as therapeutic agents in inflammation.

Two randomized trials of the PAF receptor antagonist BN 52021, enrolling a total of 870 patients, showed a modest mortality reduction that failed to attain statistical significance [45, 46]. Lexipafant, another synthetic antagonist of the platelet activating factor receptor, has been evaluated as a therapeutic agent in acute pancreatitis. Phase two studies showed benefit for the use of Lexipafant, reflected in a reduction of the severity of organ dysfunction [47,48]. However a recent international phase III trial has failed to confirm this benefit (unpublished). An alternative approach using PAF acetylhydrolase to neutralize platelet activating factor is presently being evaluated in a phase II clinical trial.

168 J.e. Marshall

Ibuprofen has been evaluated as an antagonist of prostaglandins. Two trials have been performed, but failed, either individually or in aggregate, to show a significant impact on mortality [49, 50]. A bradykinin antagonist has been evaluated in further two studies, again with negative results [51].

Nitric oxide

Nitric oxide (NO) is a short-lived gas that is released in vivo through the action of nitric oxide synthase. NO synthase catalyses the conversion of the amino acid arginine to citrulline, in the process releasing a single molecule of nitric oxide. Nitric oxide is a potent vasodilator and the final mediator of the characteristic decrease systemic vascular resistance of sepsis [52]. It also plays an important role in microbicidal activity and in intracellular signalling.

Nitric oxide synthase can be inhibited using a methylated arginine compound, N-monomethyl arginine (NMMA), that serves as a competitive inhibitor and prevents the generation of nitric oxide. Its potential role as a therapy in sepsis has been evaluated in two clinical trials. The first of these demonstrated the agent to be a potent vasopressor that permitted the weaning of patients from conventional vasopressors [53]. Unfortunately an international phase III trial was stopped at an interim analysis when increased mortality was evident in the treated group (unpublished). The excess morbidity appeared to be a consequence of pulmonary hypertension and cardiac complications, again underlining the significant potential for harm associated with the improper use of these potent agents.

Corticosteroids

Corticosteroids have a long history as adjuvant therapy in the treatment of infection [54]. In the early days of antibiotic therapy, it was common to administer physiologic doses of corticosteroids to overcome a perceived state of adrenocortical insufficiency. Studies in animal models in the 1970s suggested a potential benefit for supraphysiologic doses of adrenal corticosteroids, and led to the performance of two large randomized trials of high dose methylprednisolone in sepsis and septic shock [55, 56]. These studies, notable for establishing many of the design principles that continue to guide current sepsis trials, failed to show any evidence of benefit, but rather suggested that the practise was associated with harm in the form of excessive infection and trend towards a higher mortality.

While the approach of dampening the inflammatory response with high doses of corticosteroids has been found wanting, the role of steroids as adjuvant therapy is making a comeback. Recent work suggests that adrenocortical insufficiency is common in the late stages of septic shock, on the basis of either increased production or reduced receptor sensitivity. A recent randomized trial of


pharmacologic doses of corticosteroids given to patients with late septic shock suggested benefit in this subset of patients [57].

The coagulation cascade

The activation of an inflammatory response induces a large number of downstream physiologic effects. Prominent amongst these is activation of the coagulation cascade with microvascular thrombosis, tissue ischaemia, and worsening of the inflammatory response [58, 59]. Thus several investigators have evaluated the potential role of anticoagulant strategies as therapeutic alternatives in septic shock. Administration of recombinant activated protein C has been evaluated in a small phase II study [60] and is currently the subject of a larger phase III trial. Similarly anti-thrombin III has shown some evidence of benefit in phase II studies [61, 62], and is currently undergoing evaluation in a phase III trial. Yet another strategy to modulate the coagulation cascade has been through the use of recombinant tissue factor pathway inhibitor (TFPI).

Other strategies

The therapeutic options for modulating the inflammatory response are many. Most work to date has focussed on individual mediators. However the biologic effects of these mediators are typically redundant and overlapping, leading a number of workers to suggest that combination therapy should be considered. The lack of clear cut efficacy for any single strategy, and the enormous cost of these agents, has precluded the evaluation of combination therapy to this point in time.

However there are a number of alternate strategies that may have the same biologic effects as combination therapy. Induction of gene transcription for proinflammatory mediators is accomplished through intracellular signalling pathways that can be blocked at a number of sites. The transmission of signals from the cell membrane to the nucleus occurs through a process known as tyrosine phosphorylation, in which enzymes are activated by the addition of a phosphate group to tyrosine molecules in the protein structure [63]. A number of synthetic specific and non-specific inhibitors of tyrosine kinases have been developed, and have been shown to yield benefit in preclinical models [64, 65]. Inflammatory gene expression is also dependent upon activation of the transcription factor NFKB. NFKB can be inhibited in vivo, and preclinical studies have suggested benefit for such a strategy [66].

Yet another potentially promising approach involves the identification of new mediators whose expression occurs later in the course of sepsis. Recently a protein called HMG1 (high mobility group 1) has been demonstrated to be a downstream mediator of the toxic effects of TNF. HMG 1 is released at least 8 hours after the inflammatory stimulus making it a feasible target for therapy [67].

170 J.e. Marshall

Finally strategies that focus on turning off the inflammatory response rather than preventing its initial activation may hold some promise. Anti-inflammatory molecules such as interleukin-lO have been evaluated in preclinical models. However the divergent effects of these strategies are particularly apparent: while IL-IO improves survival in models of endotoxin challenge, it worsens outcome when the challenge is with a live organism [68]. The cellular effectors of acute inflammation, in particular the neutrophil, are removed from the body through the induction of programmed cell death or apoptosis. The normal expression of apoptosis is inhibited in patients with systemic inflammation [69, 70], therefore restoration of the normal expression of apoptosis represents another strategy for hastening the resolution of acute inflammation.

Conclusions The disappointing results of trials of novel mediator-directed therapy that have been undertaken to date underline the enormous complexities of modulating the inflammatory response. We have learned a number of lessons.

First, arbitrary combinations of physiologic abnormalities such as those that define sepsis syndrome or SIRS are poor surrogates to identify patients who might benefit from a particular therapeutic strategy. They can be met, for example, by a 20 year old man following a laparotomy for a gunshot wound of the colon, and by an 87 year old woman in congestive heart failure with an Enterococcal urinary tract infection. Not only do they define a population of patients that is overly heterogenous with respect to underlying disease process and comorbidities, but they fail to differentiate patients who might benefit from differing therapeutic strategies. The effects of the limitations of entry criteria are two. First, heterogeneity in a study population increases the sample size needed to demonstrate an effect. Secondly, and perhaps more importantly, heterogenous populations may include some patients who might benefit from therapy, but others who might in fact be harmed. How best to identify candidates for enrollment into clinical trials remains to be defined; at a minimum, however, it would appear rational that there be objective evidence of dysregulated activity of the mediator that is being targeted.

Co-morbidities confound the problems of patient heterogeneity. Patients meeting the clinical criteria for sepsis do so as a result of other life-threatening conditions, conditions that may be intrinsically irreversible. It is not uncommon, for example, for patients to be enrolled in trials following cardiopulmonary resuscitation. The effect of significant comorbidity is to reduce the potential benefit that the experimental drug might offer, and thus to necessitate a significant increase in sample size.

A lack of reliable markers of biologic activity has been a well recognized shortcoming of sepsis trials [71]. It is uncommon for studies to evaluate reversal of the study entry criteria over the period of study drug administration. More-


over rapid assays of the study target, or of downstream mediators that might effect its activity are generally unavailable. In most areas of critical care, the intensivist titrates therapy to a physiologic or biochemical response. This has not been possible in the context of sepsis trials. The lack of reliable markers has also made it difficult to determine optimal doses or durations of therapy, and dosing has generally been established on arbitrary grounds based on phase II mortality data.

Multi-centre trials show significant variability from one centre to the next, and in particular from one country to the next. An obvious, but uncharacterized interaction between disease and therapy is evident, and likely has a large effect on trial outcome. It has been shown that the benefits of such therapies are greatest for patients in whom ICU care is optimal, suggesting, perhaps, a need for future studies to be done in selected specialty centres [34].

Finally, 28 day all cause mortality may not be the best measure of clinical benefit [72, 73]. The majority of deaths in a contemporary ICU result from discontinuation of therapy, and for the elderly or chronically ill patient, death is not necessarily the worst outcome imaginable. Mortality is an objective outcome, but an insensitive one, and necessitates the conduct of large clinical trials, increasing the variability discussed above. Alternate outcome measures including reversal or prevention of organ dysfunction may be more appropriate, particularly in the initial evaluation of these complex therapies.

Successful adjuvant therapy in cancer has been slow in developing, and there is every reason to believe that the situation in sepsis will be any different. Nonetheless, a burgeoning understanding of the complexities of the host inflammatory response and its role in the most common cause of ICU morbidity and mortality suggests that we will ultimately succeed.

References I. Nasraway SA (1999) Sepsis research: We must change course. Crit Care Med 27:427-430 2. Michalek SM, Moore RN, McGhee JR et al (1980) The primary role of Iymphoreticular cells

in the mediation of host responses to bacterial endotoxin. J Infect Dis 141:55-63 3. Tracey KJ, Fong Y, Hesse DG et al (1987) Anti-cachectinffNF monoclonal antibodies pre

vent septic shock during lethal bacteraemia. Nature 330:662-664 4. Arend WP (1991) Interleukin-I receptor antagonist. A new member of the interleukin-I fami

ly. J Clin Invest 88:1445-1451 5. Marshall JC, Creery 0 (1998) Pre-clinical models of sepsis. Sepsis 2: 187-197 6. Morrison DC, Ulevitch RJ (1978) The effects of bacterial endotoxins on host mediation sys

tems. Am J PathoI93:527-617 7. Wright SO, Ramos RA, Tobias PS et al (1990) CDI4, a receptor for complexes of lipopoly

saccharide (LPS) and LPS binding protein. Science 249: 1431-1433 8. McCabe WR, Jackson GG (1962) Gram negative bacteremia. Etiology and ecology. Arch In

tern Med 110:83-91 9. Maclean LD, Mulligan WG, Mclean APH, Duff JH (1967) Patterns of septic shock in man -

A detailed study of 56 patients. Ann Surg 166:543-562

172 J.e. Marshall

10. Walker RI (1978) The contribution of intestinal endotoxin to mortality in hosts with compromised resistance: a review. Exp Hematol 6: 172-184

11. Van Deventer SJH, Ten Cate JW, Tytgat GNJ (1988) Intestinal endotoxemia. Clinical Significance. Gastroenterology 94:825-831

12. Ziegler EJ, Fisher CJ, Sprung CL et al (1991) Treatment of gram-negative bacteremia and septic shock with HA-1A human monoclonal antibody against endotoxin. N Engl J Med 324: 429-436

13. McCloskey RV, Straube RC, Sanders C et al (1994) Treatment of septic shock with human monoclonal antibody HA-1A. A randomized double-blind, placebo-controlled trial. Ann InternMed 121:1-5

14. Greenman RL, Schein RMH, Martin MA et al (1991) A controlled clinical trial ofE5 murine monoclonal IgM antibody to endotoxin in the treatment of gram-negative sepsis. JAMA 266: 1097-1102

15. Bone RC, Balk RA, Fein AM et al (1995) A second large controlled clinical study of E5, a monoclonal antibody to endotoxin: Results of a prospective, multicenter, randomized, controlled trial. Crit Care Med 23:994-1006

16. Ziegler EJ, McCutchan JA, Fierer J et al (1982) Treatment of gram-negative bacteremia and shock with human antiserum to a mutant Escherichia coli. N Engl J Med 307: 1225-1230

17. Willatts SM, Radford S, Leitermann M (1995) Effect of the antiendotoxic agent, taurolidine, in the treatment of sepsis syndrome: A placebo-controlled, double-blind trial. Crit Care Med 23:1033-1039

18. Giroir BP, Quint PA, Barton P et al (1997) Preliminary evaluation of recombinant amino-terminal fragment of human bactericidal/permeability increasing protein in children with severe meningococcal sepsis. Lancet 350: 1439-1443

19. Flynn PM, Shenep JL, Stokes DC et al (1987) Polymyxin B Moderates acidosis and hypotension in established experimental gram-negative septicemia. J Inf Dis 156:706-712

20. Lin E, Coyle SM, Randhawa S et al (1998) Polymyxin-622 prevents endotoxin-induced inflammation in humans. Surg Forum 49:6-8

21. Bone RC, Fisher CJ, Clemmer TP et al (1989) Sepsis syndrome: a valid clinical entity. Crit <;are Med 17:389-393

22. Wortel CH, von der Mohlen MA, van Deventer SJ et al (1992) Effectiveness of a human monoclonal anti-endotoxin antibody (HA-IA) in gram-negative sepsis: relationship to endotoxin and cytokine levels. J Infect Dis 168: 1367-1374

23. Fink MP (1995) Another negative clinical trial of a new agent for the treatment of sepsis: Rethinking the process of developing adjuvant treatments for serious infections. Crit Care Med 23:989-991

24. Giroir BP (1993) Mediators of septic shock: New approaches for inten-upting the endogenous inflammatory cascade. Crit Care Med 21 :780-789

25. Beutler B, Milsark IW, Cerami AC (1985) Passive immunization against cachectin tumor necrosis factor protects mice from lethal effect of endotoxin. Science 229:869-871

26. Michie HR, Spriggs DR, Manogue KR et al (1988) Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 104:280-286

27. Takashima K, Tateda K, Matsumoto T et al (1997) Role of tumor necrosis factor alpha in pathogenesis of pneumococcal pneumonia in mice. Infect Immun 65:257-260

28. Deckert-Schluter M, B1uethmann H, Rang A et al (1998) Crucial role ofTNF receptor type 1 (p55), but not ofTNF receptor type 2 (p75), in murine toxoplasmosis. J Immunol 160:3427-3436

29. Casey LC, Balk RA, Bone RC (1993) Plasma cytokines and endotoxin levels con-elate with survival in patients with the sepsis syndrome. Ann Intern Med 119:771-778

30. Martin C, Boisson C, Haccoun M et al (1997) Patterns of cytokine evolution (tumor necrosis factor-a and interleukin-6) after septic shock, hemon-hagic shock, and severe trauma. Crit Care Med 25: 1813-1819

31. Pellegrini JD, Puyana JC, Lapchak PH et al (1996) A membrane TNFa/TNFR ratio con-elates to MODS score and mortality. Shock 6:389-396


32. Calvano SE, van der Poll T, Coyle SM et al (1996) Monocyte tumor necrosis factor receptor levels as a predictor of risk in human sepsis. Arch Surg 131 :434-437

33. Fisher CJ, Dhainaut J-FA, Opal SM et al (1994) Recombinant human interleukin I receptor antagonist in the treatment of patients with sepsis syndrome. Results from a randomized, double-blind, placebo-controlled trial. JAMA 271: 1836-1843

34. Sprung CL, Finch RG, Thijs LG, Glauser MP (1996) International sepsis trial (INTERSEPT): Role and impact of a clinical evaluation committee. Crit Care Med 24: 1441-1447

35. Dinarello CA (1996) Biological basis for interleukin-I in disease. Blood 87:2095-2147 36. Hannum CH, Wilcox CJ, Arend WP et al (1990) Interleukin-1 receptor antagonist activity of

a human interleukin-I inhibitor. Nature 343:336-340 37. Pecyk RA, Fraser-Smith EB, Matthews TR (1989) Efficacy ofinterleukin-lb against systemic

Candida albicans infections in normal and immunosuppressed mice. Infect Immun 57:3257-3258

38. Fisher CJ Jr, Siotman GJ, Opal SM et al (1994) Initial evaluation of human recombinant interleukin-l receptor antagonist in the treatment of sepsis syndrome: A randomized, open-label, placebo-controlled multicenter trial. Crit Care Med 22: 12-21

39. Opal SM, Fisher CJ Jr, Dhainaut JF et al (1997) Confirmatory interleukin-1 receptor antagonist trial in severe sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. Crit Care Med 25: 1115-1124

40. Docke WD, Randow F, Syrbe U et al (1997) Monocyte deactivation in septic patients: restoration by IFN-gamma treatment. Nature Med 3:678-681

41. Dale DC, Liles WC, Summer WR, Nelson S (1995) Review: Granulocyte colony-stimulating factor role and relationships in infectious diseases. J Infect Dis 172: 1075

42. Nelson S, Belknap SM, Carlson RW et al (1998) A randomized controlled trial of Filgrastim as an adjunct to antibiotics for treatment of hospitalized patients with community-acquired pneumonia. J Infect Dis 178: 1075-1080

43. Marshall JC (J 998) The effects of granulocyte colony-stimulating factor (G-CSF) in pre-clinical models of infection and acute inflammation. Sepsis 2:213-220

44. Anderson BO, Bensard DO, Harken AH (1991) The role of platelet activating factor and its antagonists in shock, sepsis, and mUltiple organ failure. Surg Gynecol Obstet 172:415-424

45. Dhainaut J-FA, Tenaillon A, Le Tulzo Yet al (1994) Platelet-activating factor receptor antagonist BN 52021 in the treatment of severe sepsis: A randomized, double-blind, placebo-controlled, multicenter clinical trial. Crit Care Med 22: 1720-1728

46. Dhainaut JF, Tenaillon A, Hemmer Met al (1998) Confirmatory platelet-activating factor receptor antagonist trial in patients with severe gram-negative bacterial sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. Crit Care Med 26: 1963-1971

47. Kingsnorth AN, Galloway SW, Formela LJ (\995) Randomized, double-blind phase II trial of lexipafant, a platelet -activating factor receptor antagonist in human acute pancreatitis. Br J Surg 82:1414-1420

48. Kingsnorth AN, for the British Acute Pancreatitis Study Group (1997) Early treatment with lexipafant, a platelet-activating factor antagonist reduces mortality in acute pancreatitis: a double-blind, randomized placebo controlled study. Gastroenterology 112[Suppl]:A453

49. Haupt MT, Jastremski MS, Clemmer TP et al (1991) Effect of ibuprofen in patients with severe sepsis: A randomized, double blind, multicenter study. Crit Care Med 19: 1339-1347

50. Bernard GR, Wheeler AP, Russell JA et al (1997) The effects of ibuprofen on the physiology and survival of patients with sepsis. N Engl J Med 336:912-918

51. Fein AM, Bernard GR, Criner GJ et al (1997) Treatment of severe systemic inflammatory response syndrome and sepsis with a novel bradykinin antagonist, deltibant (CP-0127): Results of a randomised, double-blind, placebo-controlled trial. JAMA 277:482-487

52. Billiar TR (1995) Nitric oxide: Novel biology with clinical relevance. Ann Surg 221 :339-349 53. Grover R, Lopez A, Lorente J et al (1999) Multicenter, randomized, placebo-controlled, dou

ble blind study of the nitric oxide synthase inhibitor 546C88: Effect on survival in patients with septic shock. Crit Care Med 27:A33

174 J.e. Marshall

54. Meduri GU (1999) An historical review of glucocorticoid treatment in sepsis: Disease pathophysiology and the design of treatment investigation. Sepsis 3:21-38

55. Bone RC. Fisher CJ, Clemmer TP et al (\ 987) A controlled clinical trial of high dose methylprednisolone in the treatment of severe sepsis and septic shock. N Engl J Med 317:654-658

56. Hinshaw et al The Veterans Administration Systemic Sepsis Cooperative Study Group (1987) Effect of high dose glucocorticoid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 317:659-665

57. Bollaert PE, Charpentier C, Levy B et al (1998) Reversal of late septic shock with supraphysiologic doses of hydrocortisone. Crit Care Med 26:627 -630

58. Spriggs DR, Sherman ML, Imamura K et al (1990) Phospholipase A2 activation and autoinduction of tumor necrosis factor gene expression by tumor necrosis factor. Cancer Res 50: 710 1-7107

59. McGilvray ro, Rotstein 00 (1998) The role of coagulation in systemic inflammation: A review of the experimental evidence. Sepsis 2: 199-208

60. Bernard GR, Hartman DL, Helterbrand JD, Fisher CJ (1999) Recombinant human activated protein C (rhAPC) produces a trend toward improvement in morbidity and 28 day survival in patients with severe sepsis. Crit Care Med 27:A33

61. Eisele B, Lamy M, Thijs LG et al (1998) Antithrombin III in patients with severe sepsis. Intensive Care Med 24:663-672

62. Baudo F, Caimi TM, deCataldo F et al (1998) Antithrombin III (ATIlI) replacement therapy in patients with sepsis and/or postsurgical complications: a controlled double-blind, randomized, multicenter study. Intensive Care Med 24:336-342

63. Foxwell BMJ, Barrett K, Feldmann M (1992) Cytokine receptors: structure and signal transduction. Clin Exp Immunol 90: 161-169

64. Novogrodsky A, Vanichkin A, Patya M et al (1994) Prevention of lipopolysaccharide-induced lethal toxicity by tyrosine kinase inhibitors. Science 264: 1319-1322

65. Sevransky JE, Shaked G, Novogrodsky A et al (1997) Tyrphostin AG 556 improves survival and reduces multi organ failure in canine Escherichia Coli peritonitis. J Clinic Invest 99: 1966-1973

66. Nathens AB, Bitar R, Bujard M et al (1997) Pyrrolidine dithiocarbamate attenuates endotoxin-induced lung injury. Am J Resp Cell Mol BioI 17:608-616

67. Wang H, Bloom 0, Zhang M et al (1999) HMG-1 as a late mediator of endotoxin lethality in mice. Science 284

68. Creery 0, Kumar A (1998) The potential for interleukin lO-directed therapies in human sepsis. Sepsis 2(3):209-212

69. Jimenez MF, Watson RWG, Parodo J et a1 (1997) Dysregulated expression of neutrophil apoptosis in the systemic inflammatory response syndrome (SIRS). Arch Surg 132: 1263-1270

70. Keel M, Ungethum U, Steckholzer U et al (1997) Interleukin-lO counterregulates prointlammatory cytokine-induced inhibition of neutrophil apoptosis during severe sepsis. Blood 90(9): 3356-3363

71. Dellinger RP, Opal SM, Rotrosen D et al (1997) From the bench to the bedside: The future of sepsis research. Chest 111:744-753

72. Petros AJ, Marshall JC, van Saene HKF (1995) Is mortality an appropriate endpoint for clinical trials in critical illness? Lancet 345:369-371

73. Hebert PC (1997) Mortality as an outcome in sepsis trials. Sepsis I (1 ):35-40

INDEX I

Acinetobacter 149

acute - renal failure 91 - respiratory distress syndrome 85, 145

adenosine - diphosphate 20 - monophosphate 20 - triphosphate 20

adhesion molecules 39

adult respiratory distress syndrome 17

amino acids 79

arninoglycosides 95

amphotericin B 95,98

antibiotics 162

anti-inflammatory - agents in sepsis 155 - therapy 155

aortic surgery 72

arachidonic acid metabolites 60

arginine 168

astrocytes 79

bacteraernia 58

bacteria 145

bacterial translocation 105

blood-brain barrier 78

bradykinin antagonist 168

cardiac -index 60 - output 57 - surgery 72

cardiomyocytes 46

catecholamines 60

central nervous system 77

cerebral blood flow 78

chemokines 107

chlamydia pneumoniae 13

citrulline 168

coagulation 169

compensatory anti-inflammatory response syndrome 19

corticosteroids 155, 168

C-reactive 139

critically ill patients 57

cuprophane 96

cyc100xygenase inhibitors 38

cytochrome oxidase 45

cytokines 11,59,60, 140, 158

E. coli 93

E. coli endotoxin 58

encephalopathy 77

endothelin 1,92,104

endothelium 39,60

endotoxaemia 58

endotoxin 85,92,104, 105, 161, 162 - core antibody 13 - induced lung injury 85

fluid resuscitation 162

free oxygen radicals 18

gamma-interferon 23

gastric tonometry 67

gentamicin 97

granulocytes 38

H. pylori 13

haemodialysis 96

high risk surgical patients 60

histamine 60

hydrogen peroxide 20

hydroxyl radical 20

hypoperfusion 69

hypovolaemia 72

ibuprofen 86, 168

indomethacin 86

inflammation 145

inflammatory response syndrome 85

inhibitors cic100xygenase - 38 phosphodiesterase - 60

intensive care 125

interferon-y 108

interleukin 8,18,87,145

interleukin-l 105,108,140,161,166 - receptor antagonist 162

ischaemic hepatitis 103

jaundice 103

kidney 91

Krogh's cylinder 37

Kupffer cell 105

178

Legionella pneumophila 149

leukocytes 60

lipopolysaccharide 162

liposomes 108

liver 104 - dysfunction 103, 105 - perfusion 104

lung injury 85

macrophage inflammatory protein-2 107

macrophages 107

methylprednisolone 157

microcirculation 20,37,57

mitochondria 43

mixed inflammatory and anti-inflammatory response syndrome 19

monoamine oxidase 46

multiple organ dysfunction syndrome 17

multiple organ failure syndrome 17, 103

neuron 78

neuropeptide Y 38

neurotransmitter 80

nitric oxide 18, 39, 108, 168 - synthase 168

nonsteroidal anti-inflammatory agents 155

organ - failure 11 - oxygen supply 57

outcome 123

oxygen 43,46 - supply 60 organ - supply 57 systemic - delivery 60 free - radicals 18

P. aeruginosa 93

pentoxifylline 61

phosphodiesterase inhibitors 60

plasmapheresis 28

platelet activating factor 18, 167

portal vein 104

procalcitonin 138, 158

prostacyclin 92

Pseudomonas aeruginosa 149

Index

sepsis 11,17,28,38,57,61,71,72,77,85, 91,103,104,108,123,137,141,161 severe - 124

septic - encephalopathy 77 - shock 57,61,92,107,124,141, 155

shock 72

splanchnic - circulation 67 - perfusion 23

Staphylococcus aureus 149

superoxide radical 20

surgery 162

syndrome acute respiratory distress - 85, 145 compensatory anti-inflammatory response - 19 inflammatory response - 85 mixed inflammatory and anti-inf1ammatory response - 19 mUltiple organ dysfunction - 17 multiple organ failure - 17, 103

systemic oxygen delivery 60

testosterone 13

thrombosis 60

TNF-ex 140

transcription nuclear factor NF-kB 38

trauma 72

tumor necrosis factor 18,87,105.145,164 - ex 107

xanthine dehydrogenase 20

xanthine oxidase 20

Sepsis and Organ Dysfunction: The Challenge Continues

Documents

Transcript of Sepsis and Organ Dysfunction: The Challenge Continues