MONITORAGE HEMODYNAMIQUE NON INVASIF AUX SOINS …
131
Année Académique 2005-2006 Département d’Anesthésiologie, Pharmacologie et Soins Intensifs Service des Soins Intensifs MONITORAGE HEMODYNAMIQUE NON INVASIF AUX SOINS INTENSIFS - INTERETS ET LIMITES - Thèse présentée pour l’obtention du titre de Privat-Docent de la Faculté de Médecine Université de Genève KARIM BENDJELID 2006 1
Transcript of MONITORAGE HEMODYNAMIQUE NON INVASIF AUX SOINS …
Année Académique 2005-2006
Département d’Anesthésiologie, Pharmacologie et Soins Intensifs Service des Soins Intensifs
MONITORAGE HEMODYNAMIQUE NON INVASIF AUX SOINS INTENSIFS
- INTERETS ET LIMITES -
Thèse présentée pour l’obtention du titre de
Privat-Docent de la Faculté de Médecine
Université de Genève
KARIM BENDJELID
2006
1
TABLE DES MATIERES :
ABREVIATIONS:.................................................................................................................... 8
LISTE DE TABLES ET DE FIGURES:................................................................................ 9
A. DEFINITIONS ET GENERALITES PHYSIOLOGIQUES...................................... 13
I. QU'EST-CE QUE LA VOLEMIE ? ........................................................................................ 13
II. QU'EST CE QUE LA PRECHARGE ? ................................................................................ 13
III. QU'EST-CE QUE LA RESERVE DE PRECHARGE ?........................................................... 14
IV. QU'EST-CE QUE LE RETOUR VEINEUX SYSTEMIQUE ? ................................................. 15
V. QUELS SONT LES DETERMINANTS DE LA PRESSION SYSTEMIQUE MOYENNE ? ............... 16
VI. QU'EST CE QU'UNE HYPOVOLEMIE ? ........................................................................... 16
VII. QUELS SONT LES DETERMINANTS DE LA PRESSION ARTERIELLE PERIPHERIQUE ?.... 17
B. PHYSIOPATHOLOGIE CARDIO-VASCULAIRE DU PATIENT VENTILE PAR
PRESSION POSITIVE.......................................................................................................... 18
I. INTRODUCTION................................................................................................................. 18
II. EFFETS DE LA VENTILATION SUR LE CŒUR DROIT ...................................................... 21
1. POST-CHARGE DU VENTRICULE DROIT .............................................................................. 23
III. EFFETS DE LA VENTILATION SUR LE CŒUR GAUCHE ........................................................ 24
1. LE RETOUR VEINEUX PULMONAIRE ................................................................................... 24
2. POST-CHARGE DU VENTRICULE GAUCHE ........................................................................... 24
2
IV. IMPORTANCE CLINIQUE DE L’INTERACTION CARDIO-PULMONAIRE SUR LE DÉBIT
CARDIAQUE SYSTÉMIQUE......................................................................................................... 25
V. EFFETS DE LA VENTILATION MÉCANIQUE SUR LES CONDITIONS DE CHARGE ET LES
ÉJECTIONS VENTRICULAIRES................................................................................................... 26
C. MONITORAGE HEMODYNAMIQUE DES DEFAILLANCES
CIRCULATOIRES : REVUE DE LA LITTERATURE.................................................... 28
I. TECHNIQUES NON INVASIVES MONITORANT QUALITATIVEMENT LES DEFAILLANCES
CIRCULATOIRES ....................................................................................................................... 30
1. LE MONITORAGE CLASSIQUE : FREQUENCE CARDIAQUE, PRESSION ARTERIELLE NON INVASIVE
ET PLETHYSMOGRAPHIE .......................................................................................................... 30
2. LA CAPNOGRAPHIE EXPIREE............................................................................................. 32
3. MESURE DE LA PRESSION TRANSCUTANEE EN CO2 (TCPCO2) ............................................. 32
II. TECHNIQUES NON INVASIVES MONITORANT QUANTITATIVEMENT LES DEFAILLANCES
CIRCULATOIRES ....................................................................................................................... 33
1. TECHNIQUES NON INVASIVES MONITORANT LA REPONSE AU REMPLISSAGE VASCULAIRE CHEZ
LES PATIENTS VENTILES MECANIQUEMENT. .............................................................................. 33
2. TECHNIQUES NON INVASIVES MONITORANT LA REPONSE AU REMPLISSAGE VASCULAIRE CHEZ
LES PATIENTS EN VENTILATION SPONTANEE: LE LEVER DE JAMBES PASSIF .................................. 33
3. TECHNIQUES NON INVASIVES MONITORANT LA MESURE DE DEBIT CARDIAQUE CHEZ LES
PATIENTS DES SOINS INTENSIFS. ............................................................................................... 34
a. Mesure du débit cardiaque moyennant la méthode Fick appliquée au CO2. ........... 34
b. Mesure du débit cardiaque moyennant la méthode d’impédancemetrie thoracique.35
c. Mesure du débit cardiaque moyennant la méthode du Doppler oesophagien.......... 36
d. Mesure du débit cardiaque moyennant l’échocardiographie-Doppler. .................... 39
3
e. Évaluation du débit cardiaque moyennant l’analyse de l'onde de pression artérielle
(méthode du « pulse contour »)........................................................................................ 41
INVESTIGATIONS PERSONNELLES
43
A. INDICES PREDICTIFS DE LA REPONSE AU REMPLISSAGE VASCULAIRE
CHEZ LES PATIENTS DES SOINS INTENSIFS VENTILES PAR PRESSION
POSITIVE OU RESPIRANT SPONTANEMENT :........................................................... 44
B. MONITORAGE NON INVASIF DE LA CIRCULATION CEREBRALE. ........... 48
AGENDA POUR LA RECHERCHE FUTURE 50
A. DEFINITION DE LA REPONSE AU REMPLISSAGE VASCULAIRE PAR LA
VARIATION DE LA SATURATION VEINEUSE EN OXYGENE................................. 50
B. IMPACT DU POIDS DU PATIENT SUR LES PARAMETRES
HEMODYNAMIQUES MESURES. .................................................................................... 51
C. MONITORAGE ET QUANTIFICATION NON INVASIVE DE LA
MICROCIRCULATION....................................................................................................... 53
D. INFLUENCE DU VOLUME COURANT ET DE LA FREQUENCE
RESPIRATOIRE SUR LES INDICES DYNAMIQUES DE REMPLISSAGE
VASCULAIRE. ETUDE ANIMALE. ................................................................................ 56
CONCLUSION 58
REFERENCES 60
4
REMERCIEMENTS:
Ma reconnaissance toute particulière est destinée au Prof P.M Suter pour l’opportunité
qu’il m’a offerte d'incorporer le Service des Soins Intensifs Chirurgicaux en Mai 2000,
sa contribution à mes travaux scientifiques ainsi que son soutien à ma candidature au
titre de Privat-Docent.
Je remercie en particulier le Dr J-A Romand pour son encouragement permanent au
développement de notre recherche en hémodynamique et pour sa critique de l’actuelle
thèse.
Mes remerciements aux Prof J.C Chevrolet, actuel chef de service des soins intensifs,
pour avoir soutenue ma candidature au titre de Privat-Docent et pour sa critique de
l’actuelle thèse.
Je tiens aussi à remercier le Prof J-L Teboul, cardiologue et professeur en réanimation
médicale à la faculté de médecine Paris Sud pour la critique de l’actuelle thèse.
Mes remerciements s’adressent également aux infirmières, infirmiers et médecins des
Soins Intensifs, pour avoir contribué au bon déroulement des protocoles de recherches
cliniques.
Enfin, je voudrais aussi remercier tous ceux qui m’ont entouré et soutenu sur le plan
humain, tout au long de ce difficile parcours scientifique.
5
Pour savoir où tu vas, sache d'abord où tu es.
Pour savoir où tu es, sache d'abord d'où tu viens.
Proverbe marin, anonyme.
6
SYNTHESE : L’objectif de la présente thèse est d’évaluer les intérêts et les limites du
monitorage hémodynamique non invasif chez les patients de soins intensifs qui présentent une
instabilité cardio-circulatoire. Comme un grand nombre de patients de soins intensifs sont
porteurs d’un cathéter artériel et d’une voie veineuse centrale, dans le présent travail, ces
abords vasculaires ne seront pas considérés comme des techniques invasives.
L’utilité de toute technique de monitorage hémodynamique aux soins intensifs est de
procurer des informations fiables et reproductibles sur l’état cardio-circulatoire d’un patient
en état de choc. Les données récoltées vont permettre au médecin de prendre des décisions
thérapeutiques plus éclairées afin d’optimiser l’état hémodynamique du patient et
secondairement d’améliorer son pronostic vital. Ce bénéfice pourrait s’expliquer par une
optimisation de l’état circulatoire avec une meilleur perfusion des tissus périphériques et par
la prévention de la survenue de complications (insuffisance rénale et mésentérique,
dysfonctions hépatiques et ventilatoires ...). Ces effets seront d’autant plus considérables que
la prise en charge du patient choqué se fera précocement dans les suites de l’installation de
l’instabilité circulatoire.
Le développement du monitorage hémodynamique non invasif aux soins intensifs est une
nécessité dans les années futures. En effet, ces techniques possèdent l’avantage d’être moins
agressives pour un patient déjà affaibli par une grave affection. Comparé aux anciennes
techniques de monitorage hémodynamique, elle auront l’avantage d’une iatrogénie moindre.
A ce jour, de nombreuses techniques non invasives de monitorage hémodynamique sont
utilisées aux soins intensifs. Ces techniques regroupent l’échocardiographie-Doppler, le
Doppler œsophagien, l’impédancemetrie, la réinhalation partielle du CO2 avec utilisation de
l'équation de Fick, la thermodilution transpulmonaire, la pléthysmographie, l’examen de la
courbe de la pression artérielle sanglante avec ses des dérivées et intégrales ...etc. Néanmoins,
ces techniques de monitorage se doivent dans l’avenir d’être aussi précises et plus
reproductibles que les anciennes méthodes invasives. En effet, utilisées pendant une
vingtaine d’années, les techniques invasives de monitorage hémodynamique n’ont fait la
preuve de leur utilité par la diminution de la morbi-mortalité que chez certains sous groupes
de patients en péri-opératoire. Une amélioration de la prise en charge des patients des soins
intensifs ne peut donc être attendue avec ces nouvelles techniques qu’au prix d’une exigence
en précision et en reproductibilité, tout en assurant un maximum d’innocuité.
7
ABREVIATIONS:
CEC : circulation extra corporelle
DC : débit cardiaque
DTI : Doppler tissulaire
PA : pression artérielle
PAS : pression artérielle systolique
PAD : pression artérielle diastolique
PAM : pression artérielle moyenne
PP : pression pulsée
PEP : pression expiratoire positive
RVS : résistance vasculaire systémique
SI : soins intensifs
VES : volume d’éjection systolique
VS : ventilation spontanée
VMPP : ventilation mécanique par pression positive
VD : ventricule droit
8
LISTE DE TABLES ET DE FIGURES:
Figure 1: La relation entre la précharge et le volume d'éjection systolique, appelée courbe de
fonction systolique de Starling.
Figure 2 : Concept des volumes sanguins circulants mobilisables et non-mobilisables
(« stressed-unstressed volume »).
Figure 3 : Enregistrement simultané de multiples paramètres hémodynamiques durant
ventilation spontanée (à gauche) et pression positive (à droite) utilisant le même volume
courant (modèle expérimental de chiens à thorax fermés).
Figure 4 : Echocardiographie bidimensionnelle par voie sous-costale objectivant un
collapsus inspiratoire de la veine cave inférieure.
Figure 5 : Courbes de retour veineux enregistrées sur quatorze chiens à thorax ouverts.
Figure 6 : Effet du volume pulmonaire sur la résistance vasculaire pulmonaire.
Figure 7 : Courbe sanglante de pression artérielle radiale.
Figure 8 : Principe de mesure du débit cardiaque moyennant la méthode Fick appliquée au
CO2.
Figure 9: Mesure du débit cardiaque par impédancemetrie après application d'un courant
électrique de faible amplitude et de haute fréquence entre deux paires d'électrodes, l'une
placée sur le cou et l'autre sur le thorax.
Figure 10 : Représentation schématique d’un Doppler œsophagien chez un patient en
objectivant le voisinage de l’œsophage et de l’aorte descendante
Figure 11 : Principe du calcul du volume d’éjection systolique à partir de la vélocité
aortique.
9
Figure 12 : Principe et exemple du calcul du débit cardiaque et du volume d’éjection
systolique à partir de la vélocité au niveau de l’anneau aortique en utilisant
l’échocardiographie-Doppler.
Figure 13 : Principe de la mesure du volume d’éjection systolique battement par battement
selon la méthode du « pulse contour ».
10
L'hémodynamique est l'étude des lois régissant le fonctionnement du système cardio-
vasculaire [1]. Pour interpréter correctement les signes cliniques que présente un
patient en insuffisance circulatoire et pour pouvoir en déduire une attitude
thérapeutique adaptée et efficiente, le médecin doit maîtriser les règles gouvernant
l'écoulement du sang dans les vaisseaux (pression, débit, tension,...) [1].
Particulièrement aux soins intensifs, le monitorage hémodynamique permet à
l’intensiviste de comprendre les effets induits par différentes thérapeutiques comme
les infusions de volume et/ou de médicaments administrés au patient qui présente une
instabilité hémodynamique. Ce mode de surveillance continu devient alors un outil
indispensable pour analyser aisément le fonctionnement de cette "plomberie"
complexe qu'est notre système cardio-vasculaire [2]. Comme exemple, le paragraphe
suivant reflète assez bien la complexité de la prise en charge thérapeutique d’un
patient admis à l’hôpital pour une instabilité circulatoire réfractaire à toute
thérapeutique.
« Un patient de 67 ans a présenté un malaise à l’aéroport de Genève. Lors de sa prise en
charge médicale initiale, l'examen retrouvait: un score de Glasgow à 3, un état de choc
avec une pression artérielle systolique (PAS) à 53 mmHg et une fréquence cardiaque à
137 b/minute. Par ailleurs, il présentait à l’examen clinique un ventre suspect. Un
remplissage vasculaire était débuté par des colloïdes et une perfusion continue
d'adrénaline était initiée devant l'instabilité hémodynamique et les épisodes de
bradycardie (PAS = 43 mmHg, fréquence cardiaque = 56 b/min). Le patient a ensuite été
transporté à l'hôpital avec une expansion volémique totale pré-hospitalière de 3000 ml de
cristalloïdes et une perfusion continue de 0.7 mg/h d'adrénaline. À l'arrivée en salle de
déchocage, l'état de choc persistait (PAS = 63 mmHg, fréquence cardiaque = 128 b/min).
L’intensiviste appelé au chevet du patient décide de l’instauration d’un monitorage
cardiovasculaire consistant en une échocardiographie-Doppler et la mise en place d’un
cathéter artériel pulmonaire pour la mesure du débit cardiaque et des pressions de
remplissages ventriculaires ».
Comme le reflète le cas clinique présenté précédemment, lors d’une défaillance
circulatoire aiguë plus communément appelé état de choc, la PAS sanguine chute au
11
dessous de 90 mmHg [3]. Dans ce cas bien précis, retrouvé fréquemment chez certains
patients de soins intensifs, le traitement de la détresse circulatoire repose sur la
détection de la cause de la baisse de cette pression artérielle systémique. En
conséquence, trois causes possibles doivent être recherchées : 1 - une baisse du
volume sanguin circulant (cas du choc hémorragique ou hypovolémique); 2 – une
baisse de la résistance vasculaire systémique (vasodilatation des artères, par exemple
dans le sepsis); 3 – une diminution de la fonction de la pompe cardiaque par baisse de
la contractilité (cas d’une dysfonction systolique secondaire à un infarctus du
myocarde), ce qui engendre une diminution du débit cardiaque.
La prise en charge des patients de soins intensifs présentant une défaillance
circulatoire aiguë repose donc sur une prise en charge minutieuse, comportant une
surveillance continue des paramètres hémodynamiques et des interventions
thérapeutiques ciblées afin de normaliser les variables obtenues en cas de grand
décalage avec les normes physiologiques [4]. Cette prise en charge est basée sur un
monitorage hémodynamique précis et reproductible afin d’éviter toute intervention
médicamenteuse préjudiciable à la survie du patient. En effet, il n’est pas rare que
l’infusion excessive de liquide intravasculaire ou d’amines inotropes et/ou de
vasopresseurs soit à l’origine d’une augmentation de la mortalité aux soins intensifs
[5-7].
12
A. DEFINITIONS ET GENERALITES PHYSIOLOGIQUES
Ce chapitre reprend les définitions des recommandations d'experts de la Société de Réanimation de
Langue Française (SRLF) sur les « Indicateurs du remplissage vasculaire au cours de
l'insuffisance circulatoire » [4].
I. QU'EST-CE QUE LA VOLEMIE ?
La volémie est le volume sanguin total de l'organisme (plasma et éléments figurés).
La valeur normale de la volémie est de 65 à 75 ml/kg.
La volémie ne peut être évaluée que par des techniques de dilution d'un indicateur qui reste
dans le secteur intravasculaire, techniques difficilement utilisables en réanimation.
II. QU'EST CE QUE LA PRECHARGE ?
Dans les études expérimentales, la précharge est la longueur de la fibre myocardique avant sa
contraction.
En clinique, il n'y a pas de consensus sur la définition de la précharge ventriculaire. Pour
chaque ventricule, la précharge peut être définie soit comme la dimension du ventricule en
télédiastole (diamètre, surface, volume), soit comme les conditions de charge du ventricule en
télédiastole (pression et contrainte transmurales).
La pression transmurale du ventricule est la pression de distension de ce ventricule, calculée
comme la pression intravasculaire moins la pression externe qui s'applique à ce ventricule.
Le volume télédiastolique est un déterminant majeur du volume d'éjection systolique
ventriculaire et donc du débit cardiaque ; en pratique clinique, la dimension du ventricule en
télédiastole est ainsi un bon indice de précharge.
La relation existant entre pression transmurale et volume télédiastolique du ventricule dépend
de l'élastance diastolique de ce ventricule.
13
III. QU'EST-CE QUE LA RESERVE DE PRECHARGE ?
La relation entre la précharge et le volume d'éjection systolique, appelée courbe de fonction
systolique, comprend deux parties (Figure 1) : une première partie dite de précharge
dépendance (portion ascendante) où un accroissement de précharge — par exemple par
l'expansion volémique — entraîne une faible augmentation de la pression transmurale, pour
une augmentation significative des dimensions diastoliques du ventricule et de son volume
d'éjection systolique (réserve de précharge) ; et une seconde partie dite de précharge
indépendance (plateau de la courbe) où l'augmentation de la précharge résulte en une
augmentation importante de la pression transmurale, sans augmentation significative des
dimensions diastoliques du ventricule et du volume d'éjection systolique.
Figure 1: La relation entre la précharge et le volume d'éjection systolique, appelée courbe de fonction
systolique de Starling [8].
La réserve de précharge d'un ventricule est d'autant plus marquée :
• que le ventricule travaille sur la portion ascendante de la courbe de fonction systolique ;
• que le ventricule travaille sur la partie initiale de la portion ascendante de la courbe, c'est-à-
dire à distance du plateau de la courbe ;
• et que la pente de cette portion ascendante est plus raide (fonction systolique conservée).
14
La courbe de fonction systolique s'applique aussi bien au ventricule gauche qu'au ventricule
droit. Pour qu'une expansion volémique induise une augmentation de volume d'éjection
systolique, il faut que les deux ventricules aient une réserve de précharge.
C’est la réserve de précharge qui détermine la réponse au remplissage vasculaire. Cette
réponse au remplissage est définie par le fait que soit le patient est répondeur au remplissage
(et donc il augmente son débit cardiaque d’un certain pourcentage, souvent défini à 15% [9]),
soit il n’augmente pas son débit cardiaque dans les suites de l’infusion de volume
intravasculaire. En accord avec cette définition, chez les patients placés en ventilation
mécanique, deux types d’indices de réponse au remplissage ont été proposés par les
hémodynamiciens [8, 10]. Premièrement, les indices statiques, qui consistent en une mesure
fixe au cours du temps (en général fin d’expiration), et secondairement les indices
dynamiques qui étudient une variation de ces indices au cours d’un cycle ventilatoire imposé
par le respirateur.
IV. QU'EST-CE QUE LE RETOUR VEINEUX SYSTEMIQUE ?
Le retour veineux systémique est le drainage par la pompe cardiaque du volume sanguin
depuis le réservoir veineux périphérique vers le réservoir central intrathoracique à travers de
grosses veines collapsibles.
Lorsqu'ils sont moyennés sur un cycle respiratoire, le retour veineux systémique et le débit
cardiaque sont égaux. Le gradient de pression gouvernant le retour veineux est la différence
entre la pression motrice d'amont qui règne dans le réservoir périphérique (c'est la pression
systémique moyenne) et la pression d'aval du retour veineux représentée par la pression
auriculaire droite intravasculaire.
Le retour veineux systémique est égal au rapport entre le gradient de pression de ce retour
veineux et la résistance au retour veineux systémique.
15
V. QUELS SONT LES DETERMINANTS DE LA PRESSION SYSTEMIQUE MOYENNE ?
La pression systémique moyenne dépend de la capacitance du réservoir veineux périphérique
et du volume sanguin hémodynamiquement actif (volume contraint ou mobilisable) qui y est
contenu. Il existe également un volume sanguin hémodynamiquement inactif (non contraint
ou non mobilisable), nécessaire au maintien « ouvert » des vaisseaux (Figure 2). Le rapport
des volumes contraint–non-contraint au sein du réservoir veineux périphérique dépend du
tonus veinoconstricteur [11].
Chez un homme sain, la pression systémique moyenne est de 7 à 12 mmHg. Elle n'est pas
mesurable en pratique clinique puisqu'elle ne peut être mesurée qu'à cœur arrêté.
stress Unstress
Figure 2 : Concept des volumes sanguins circulants mobilisables et non-mobilisables (« stressed-
unstressed volume ». [11]
VI. QU'EST CE QU'UNE HYPOVOLEMIE ?
L'hypovolémie absolue est définie comme une diminution du volume sanguin total circulant.
Cette diminution peut être liée à des pertes sanguines (hémorragie) ou à des pertes purement
plasmatiques (pertes digestives, rénales, cutanées, extravasation dans le tissu interstitiel).
16
L'hypovolémie, en diminuant la pression systémique moyenne, est responsable d'une
diminution du retour veineux systémique, de la précharge cardiaque, du volume sanguin
central et du débit cardiaque, en dépit de l'augmentation de la fréquence cardiaque (régulation
neuro-humorale).
L'hypovolémie relative est définie par une mauvaise répartition de la volémie entre les
compartiments centraux et périphériques. En effet, bien qu’il existe une volémie normale
voire augmentée, il existe une insuffisance de volume sanguin central comme lors de la
ventilation en pression positive ou lors d'une veinodilatation. La séquestration dans le
territoire veineux splanchnique y participe.
Le volume sanguin central est le volume de sang intrathoracique. Il représente classiquement
environ 20 % de la volémie, répartie grossièrement pour 50 % dans les cavités cardiaques et
pour 50 % dans la circulation pulmonaire.
VII. QUELS SONT LES DETERMINANTS DE LA PRESSION ARTERIELLE
PERIPHERIQUE ?
La pression artérielle moyenne (PAM) est le produit du débit cardiaque par la résistance
vasculaire systémique, auquel il convient théoriquement d'ajouter la pression systémique
moyenne. La PAM est une grandeur qu’il est important de prendre en compte en clinique car
elle est le déterminant principal de la pression de perfusion des organes.
La PAS dépend, pour une PAM donnée, de l'élasticité artérielle et des caractéristiques de
l'éjection ventriculaire gauche.
La pression artérielle diastolique (PAD) dépend de la résistance vasculaire systémique, de
l'élasticité artérielle et de la durée de la diastole.
La pression artérielle pulsée, appelée encore pression artérielle différentielle (PP = PAS –
PAD) dépend de la différence de volume sanguin présent dans une artère entre la systole et la
diastole et de la compliance de cette artère. En pratique, bien que ces grandeurs soient
souvent différentes, on assimile la pression pulsée périphérique à la pression pulsée centrale
17
et on considère que le volume d'éjection systolique est le principal déterminant de la
différence de volume sanguin présent dans une artère entre la systole et la diastole.
Finalement, il convient aussi de réaliser que les lois qui régissent l’écoulement sanguin sont
dépendantes du mode de ventilation des patients, à savoir une ventilation physiologique
spontanée ou artificielle en pression positive.
B. PHYSIOPATHOLOGIE CARDIO-VASCULAIRE DU PATIENT VENTILE
PAR PRESSION POSITIVE
Une grande proportion de patients des soins intensifs présente une défaillance respiratoire qui
nécessite une suppléance ventilatoire moyennant une ventilation mécanique par pression
positive (50 %) [12]. Chez ces patients ventilés grâce à une assistance, les interactions cardio-
pulmonaires sont affectées principalement par l'état de remplissage intra-vasculaire et son
impact sur le remplissage ventriculaire. En effet, une diminution du débit cardiaque par baisse
des remplissages ventriculaires peut être objectivée lors de l’instauration d’une ventilation
mécanique en pression positive. Ce phénomène est secondaire au fait que le cœur et les
poumons sont anatomiquement et fonctionnellement liés et que les variations de la pression
intra-thoracique et du volume pulmonaire, induites par la ventilation, modifient cycliquement
l’interaction cardio-pulmonaire. Les effets physiologiques et physiopathologiques de la
ventilation à pression positive sur les retours veineux ventriculaires ainsi que sur les
résistances vasculaires pulmonaire et systémique sont passés en revue dans ce chapitre.
I. INTRODUCTION
Situés dans la cage thoracique, qui est assez rigide, le cœur et les poumons sont
anatomiquement et fonctionnellement étroitement liés [13]. Il en résulte que toute
modification des caractéristiques physiques et physiologiques pulmonaires immanquablement
affecte le coeur et vice versa, d’où le terme d’interactions cardio-pulmonaires [14-16]. Le but
de ce chapitre est de revoir la physiopathologie cardio-pulmonaire lors d’une ventilation
18
mécanique à pression positive [17-19]. Rappelons brièvement qu’un ventilateur insuffle
périodiquement un mélange d’air et d’oxygène dans les voies aériennes supérieures afin de
pallier soit à une défaillance de l’échangeur gazeux, soit à l’insuffisance de ventilation
alvéolaire, par exemple lors d’une anesthésie générale avec relaxation musculaire ou en
présence d’une pathologie neuromusculaire. Alors qu’une pression inspiratoire sub-
atmosphérique (dite négative) pleurale et intra-thoracique génère le volume courant lors de la
respiration spontanée (VS), c’est en revanche une élévation périodique de la pression intra-
pulmonaire qui produit ce volume courant lors de la ventilation mécanique par pression
positive (VMPP), ce qui augmente également la pression pleurale [17]. Le volume courant
peut être identique entre la VS et la VMPP, mais les effets cardio-pulmonaires sont
diamétralement opposés. Cela s’explique par le fait que la VMPP est susceptible de modifier
les conditions d’éjection imposées au ventricule droit, par une augmentation de la pression
trans-pulmonaire (pression alvéolaire déduite de la pression pleurale; Figure 3) lors de la
délivrance du volume courant. Ce dernier augmente donc l’impédance à l’éjection du
ventricule droit (VD) en phase inspiratoire [20].
Les respirateurs actuels permettent aux patients de conserver une respiration spontanée à côté
de la ventilation mécanique proprement dite (aide inspiratoire, …). L’interaction cardio-
pulmonaire qui en résulte devient alors plus complexe, variant d’une « inspiration » à l’autre
[21]. Un exemple classique des effets cardio-circulatoires de la ventilation mécanique est
observé par l’anesthésiste lors de l’induction d’une anesthésie générale. En effet, lors de cette
situation, la pression artérielle s’abaisse pour plusieurs raisons : diminution du retour veineux
dans les suites d’une augmentation de la pression intra-thoracique [22] et effet inotrope
négatif ou vasodilatateur des médicaments neuro-analgésiques. Cependant, la pression
artérielle peut être immédiatement normalisée par l’administration intra-vasculaire rapide de
cristalloïdes ou de colloïdes [23].
Sous un régime de ventilation à pression positive, le volume de sang éjecté par les ventricules
dépend de plusieurs facteurs : premièrement du volume sanguin télé-diastolique [24] qui est
dépendant du retour veineux [25] et de la compliance ventriculaire [26, 27], ensuite de
l’inotropisme myocardique [28] et, enfin, de la pression qui règne dans les vaisseaux en aval
des ventricules [29]. Dans cette revue, nous allons principalement nous intéresser aux retours
veineux systémique et pulmonaire ainsi qu’à la post-charge des deux ventricules des patients
en ventilation spontanée et/ou sous ventilation mécanique.
19
Figure 3 : Enregistrement simultané de multiples paramètres hémodynamiques durant une ventilation
spontanée (à gauche) et pression positive (à droite) utilisant le même volume courant (modèle de
chiens à thorax fermé). Les lignes verticales discontinues objectivent le début et la fin de chaque
ventilation. SVRV : volume d’éjection du ventricule droit. SVLV : volume d’éjection du ventricule
gauche. PAO : pression aortique. Platm : pression transmurale de l’oreillette gauche. Ppatm : pression
transmurale de l’artère pulmonaire. Pratm : pression transmurale de l’oreillette droite. Paw : pression
aérienne. Ppl : pression pleurale. Pra : pression auriculaire droite.
20
II. EFFETS DE LA VENTILATION SUR LE CŒUR DROIT
Les effets de la ventilation sur les cavités cardiaques droites dépendent du gradient de
pression entre les veines extra-thoraciques et l’oreillette droite [30]. Si la pression veineuse
systémique varie peu au cours de la respiration tant spontanée que mécanique [31], il n’en est
pas de même de la pression auriculaire droite qui change en parallèle avec la pression qui
entoure cette cavité, soit la pression pleurale [25]. Lors de l’inspiration, en respiration
spontanée, la pression pleurale (pression extramurale) s’abaisse, induisant une baisse de la
pression auriculaire intramurale. Cette dépression augmente le gradient de pression
déterminant le retour veineux entre les veines extra-thoraciques et le cœur [25], ce qui conduit
à une augmentation du volume intra-auriculaire et de la pression auriculaire transmurale
(pression intramurale réduite de la pression extramurale) [18, 32]. La collapsibilité des veines
entrant dans le thorax limite l’augmentation de l’écoulement sanguin inspiratoire [33] comme
cela peut être observé échographiquement par un collapsus de la veine cave inférieure [34]
(Figure 4). En effet, lors d’une inspiration profonde, le flux sanguin augmente initialement
puis se stabilise et plafonne malgré la diminution progressive de la pression pleurale (concept
du « Starling resistor ») [35]. Ce « plafond » est déterminé par le collapsus des gros vaisseaux
veineux [25] (Figure 5).
L’afflux inspiratoire de sang dans l’oreillette droite et, respectivement, dans le ventricule droit
est, paradoxalement, un des facteurs responsables de la diminution de la pression aortique lors
de l’inspiration dans certaines conditions. En effet, une expansion ventriculaire droite
excessive peut gêner la fonction du ventricule gauche, parce que les ventricules partagent le
septum et tous deux sont situés à l’intérieur de la membrane péricardique, qui est peu
compliante [13]. Il en résulte que toute augmentation volémique d’un ventricule, en deçà d’un
minimum (« unstressed volume » ou volume non contraint ou non mobilisable ; Figure 2) [11]
se fera aux dépens du volume de l’autre, ceci étant nommé interdépendance bi-ventriculaire
[13, 20, 36]. Lors de l’inspiration le volume de remplissage du ventricule droit augmente,
faisant bomber le septum dans la cavité ventriculaire gauche. Le bombement limite le volume
télédiastolique du ventricule gauche et aboutit à une diminution du volume d’éjection
systémique et finalement abaisse la pression aortique [36]. Durant l’expiration, en VS, la
pression pleurale se rapproche de la pression atmosphérique, élevant relativement la pression
auriculaire droite par rapport à l’inspiration. Si cette dernière est plus élevée que la pression
dans les veines extra-thoraciques, le retour veineux systémique diminuera voir cessera jusqu’à
21
la prochaine inspiration. Lors de la VMPP en revanche, la pression pleurale reste positive
durant le cycle respiratoire, limitant le retour veineux sanguin surtout pendant l’insufflation
[37]. Une élévation de la pression veineuse systémique par l’administration intra-vasculaire
rapide de cristalloïdes ou de colloïdes permet de rétablir un gradient de pression entre les
veines extra-thoraciques et l’oreillette droite et ainsi d’améliorer le retour veineux [37].
Figure 4 : Echocardiographie bidimensionnelle par voie sous-costale objectivant un collapsus
inspiratoire de la veine cave inférieure.
22
Figure 5 : Courbes de retour veineux enregistrées dans un modèle expérimental chez quatorze chiens
à thorax ouvert. Cette figure objective clairement que le retour veineux augmente initialement (entre
+8 et -2 de pression auriculaire droite; coté droit de la courbe) puis se stabilise et plafonne malgré
une diminution progressive de la pression pleurale et de la pression auriculaire droite (concept du
« Starling resistor » ; coté gauche de la courbe) [25].
1. POST-CHARGE DU VENTRICULE DROIT
Au même titre que l’anatomie alvéolaire, l’anatomie vasculaire pulmonaire varie
cycliquement avec la respiration. L’état d’inflation des poumons est un déterminant important
d’une part, de la résistance qu’oppose l’arbre vasculaire pulmonaire à l’éjection du ventricule
droit et d’autre part de la capacitance du lit vasculaire pulmonaire. Une relation en «U»
caractérise la relation entre le volume d’insufflation des poumons et la résistance vasculaire
pulmonaire [38]; cette résistance est déterminée par l’anatomie vasculaire instantanée [39]. La
circulation pulmonaire peut être divisée en vaisseaux extra-alvéolaires et alvéolaires en
fonction de la pression qui les entoure [39]. Les vaisseaux extra-alvéolaires (gros calibre) sont
exposés à la pression interstitielle pulmonaire qui est approximativement égale à la pression
pleurale. Lorsque le volume pulmonaire est petit, ces vaisseaux ont tendance à collaber, ce
qui augmente la résistance vasculaire pulmonaire [38] (Figure 6). A l’inverse, lors d’une
grande insufflation pulmonaire dilatant les alvéoles, ce sont les vaisseaux alvéolaires
(capillaires, artérioles et veinules, selon la pression alvéolaire) qui sont collabés alors que les
vaisseaux extra-alvéolaires sont dilatés au maximum [40]. L’équilibre entre les vaisseaux
extra-alvéolaires et alvéolaires détermine la résistance vasculaire pulmonaire, mais celle-ci
peut aussi être modifiée par des facteurs neuro-humoraux (Figure 6) . En effet, la présence
d’une pression partielle d’oxygène alvéolaire basse ou d’une acidose active la
vasoconstriction hypoxique et augmente la résistance vasculaire pulmonaire [41].
L’augmentation de la postcharge du ventricule droit, alors que ce dernier est conçu pour
travailler contre une faible résistance vasculaire, limite l’éjection ventriculaire (concept de
« pression sensibilité du VD » décrit par Jardin ).
En résumé, sous ventilation à pression positive, la baisse de la pression artérielle moyenne et
du débit cardiaque est secondaire à la fois à la diminution du retour veineux systémique [31]
et à l’augmentation de la résistance vasculaire pulmonaire [28].
23
Figure 6 : Effet du volume pulmonaire sur la résistance vasculaire pulmonaire [38]
III. EFFETS DE LA VENTILATION SUR LE CŒUR GAUCHE
1. LE RETOUR VEINEUX PULMONAIRE
Contrairement à l’oreillette droite, le flux sanguin retournant vers l’oreillette gauche ne subit
pas de variation en relation avec les changements de la pression pleurale car l’oreillette
gauche et les veines pulmonaires sont affectées simultanément par le même régime de
pression pleurale. Néanmoins, elles sont soumises à des régimes différents de pression
péricardique (exemple de la péricardite constrictive). Cependant, l’inflation pulmonaire
détermine la capacitance du réservoir veineux pulmonaire, stockant ou expulsant plus ou
moins de sang dans les veines pulmonaires [42] ce qui affecte de manière cyclique le
remplissage auriculaire gauche [43, 44].
2. POST-CHARGE DU VENTRICULE GAUCHE
La résistance à l’éjection du ventricule gauche est classiquement approchée par la pression
artérielle systémique. Buda et coll. ont montré que la pression aortique (et de ce fait la
pression artérielle systémique) peut être grandement modifiée par de larges variations de la
pression pleurale qui affectent les performances cardiaques [45]. Un abaissement important
de la pression pleurale augmente la postcharge du ventricule gauche (par exemple lors d’une
manœuvre de Müller) et elle est diminuée lorsque la pression pleurale augmente (manœuvre
24
de Valsalva) [46-48]. L’analyse de la pression transmurale du ventricule gauche (pression
intraventriculaire gauche réduite de la pression intrapericardique) permet une meilleure
compréhension de la postcharge du ventricule gauche, plutôt que la seule mesure de la
pression aortique systolique [45]. Ce phénomène devient physiologiquement important
surtout lorsque la pression pleurale est très négative ou positive (asthme, exacerbation d’un
syndrome obstructif chronique, toux, etc.) ou lorsque le ventricule gauche est défaillant [40,
46].
IV. IMPORTANCE CLINIQUE DE L’INTERACTION CARDIO-PULMONAIRE SUR LE DÉBIT
CARDIAQUE SYSTÉMIQUE
La VMPP affecte différemment la circulation systémique, si l’on considère l’état de la
fonction systolique ventriculaire gauche. Premièrement, si cette dernière est normale, le
patient sera en «précharge dépendance» (Figure 1) [8, 49] par un effet de diminution du retour
veineux («saignée» mécanique). Si, en revanche, le ventricule gauche est défaillant, ce sera la
résultante des effets sur : le retour veineux vers le VD (1), le retour veineux pulmonaire (2), la
pression transmurale du ventricule gauche (c’est-à-dire la postcharge) (3) et finalement sur
l’oxygénation [50, 51] (4) qui déterminera les effets de cette ventilation sur le débit cardiaque
systémique. Si l’on considère le premier de ces facteurs, une chute de la pression artérielle
sera due à la diminution du retour veineux systémique, qui va limiter secondairement le retour
veineux pulmonaire et le volume télédiastolique du ventricule gauche, aboutissant à une chute
du débit cardiaque gauche et de la pression artérielle [18]. L’utilisation d’une pression
positive en fin d’expiration (PEP) voire l’apparition d’une auto-PEP [52-54], vont aggraver ce
phénomène par le biais de l’augmentation de la pression intrathoracique et pleurale [47]. Le
second facteur va jouer un rôle en cas de survenue d’une pression d’insufflation élevée,
surtout si la pathologie pulmonaire sous-jacente est non uniforme, et l’on peut assister à une
augmentation de la résistance vasculaire pulmonaire secondaire à la distension alvéolaire
induite par la VMPP [20, 55]. Cette augmentation de la résistance pulmonaire aura un effet
négatif sur le remplissage du ventricule gauche par la baisse du débit sanguin trans-
pulmonaire et par l’effet d’une augmentation de la contrainte septale exercée sur le ventricule
gauche (interdépendance ventriculaire) (1). Deux autres mécanismes ont un impact sur
l’hémodynamique cardiaque pendant la ventilation à pression positive. Lors de l’insufflation,
25
le remplissage ventriculaire gauche peut être augmenté (1) par un effet de chasse mécanique
du sang contenu dans les capillaires pulmonaires vers l’oreillette gauche (cas d’une
hypervolémie) [42], (2) par une augmentation inspiratoire de la pression intra-thoracique qui
abaisse la post-charge ventriculaire gauche, améliorant ainsi l’éjection du ventricule gauche
[47].
V. EFFETS DE LA VENTILATION MÉCANIQUE SUR LES CONDITIONS DE CHARGE ET LES
ÉJECTIONS VENTRICULAIRES
L’insufflation mécanique chez un patient profondément sédaté est à l’origine d’une
diminution de la pré-charge [18] et d’une augmentation de la post-charge ventriculaire droite
[20]. La réduction de la précharge ventriculaire droite est secondaire à la gêne au retour
veineux systémique [25] et l’augmentation de la post-charge ventriculaire droite est
secondaire à l’augmentation du volume pulmonaire [38],[55]. Ces deux phénomènes sont à
l’origine d’une baisse du volume d’éjection systolique ventriculaire droit par baisse de la
précharge et augmentation de la contrainte à l’éjection du ventricule droit. La baisse
inspiratoire du volume d’éjection du ventricule droit entraîne une diminution du remplissage
ventriculaire gauche après un délai de trois à quatre battements cardiaques en raison du temps
de transit pulmonaire prolongé [56]. Alors que la baisse de la précharge ventriculaire gauche
est, quant à elle, suivie d’une baisse du volume d’éjection du ventricule gauche qui va
coïncider avec la phase expiratoire de la ventilation mécanique. Le phénomène est inverse
lors de l’insufflation où le volume d’éjection ventriculaire gauche augmente du fait d’un
meilleur remplissage qui fait suite à une augmentation du volume systolique éjecté dans
l’artère pulmonaire survenu trois battements plus tôt lors de la phase expiratoire. Cette
interdépendance des deux ventricules qui sont montés en série [57] est à l’origine de
variations respiratoires des deux volumes d’éjections systoliques lorsque les chambres
ventriculaires sont “ précharges-dépendantes ” (Figure 1 ) [8, 49]. Cela va se traduire par une
variation ventilatoire de la pression artérielle systolique qui nous renseigne sur la « précharge-
dépendance » des ventricules [49] et sur le bénéfice d’un remplissage vasculaire [8]. Ce
concept intègre le fait qu’un patient qui augmente son débit cardiaque et plus ou moins sa
PAM dans les suites d’une perfusion de liquide intra-vasculaire est un répondeur à
l’augmentation de la précharge cardiaque. Ceci est d’autant plus important qu’au cours de
26
l’insuffisance circulatoire aiguë, l’analyse de la littérature révèle que seulement 40-70% des
patients sous ventilation mécanique sont répondeurs à une expansion volémique [8, 10], d’où
l’importance de pouvoir les distinguer des non-répondeurs. Ces variations cycliques des
volumes d’éjection systolique secondaires à la ventilation mécanique sont les déterminants de
ce qu’on appelle communément les predicteurs (indices) dynamiques (en opposition avec
indices statiques mesurés en fin d’expiration) de la réponse au remplissage vasculaire. Ces
indices dynamiques se traduisent par une variation ventilatoire de la pression artérielle
systolique, de la pression pulsée systémique, du volume d’éjection systolique, des vélocités
Doppler trans- aortiques et des temps de pré-éjection du ventricule gauche [23, 58-64]. Toutes
ces mesures nous renseignent sur la « précharge-dépendance » des ventricules avec une utilité
clinique pour quantifier les effets hémodynamiques d’une expansion volémique [23, 58-64].
Néanmoins, dans certaines conditions exceptionnelles comme le cœur pulmonaire aigu,
l’utilisation des variations ventilatoires de la pression pulsée systémique peut être trompeuse
et même dangereuse car à l’origine d’une infusion volumique délétère (faux positif) [28, 65].
Par ailleurs, il n’existe pas de consensus physiopathologique sur les déterminants majeurs de
cette variation ventilatoire des volumes d’éjection systolique (« précharge-dépendance »)
[66]. En effet, est-ce la variation du remplissage du ventricule droit ou est-ce la variation de
l’éjection du ventricule droit qui définit la modification cyclique de ces indices sous
ventilation mécanique? La cause prépondérante à l’origine de la variation ventilatoire des
volumes d’éjection systolique dépend, en fait, de l’état fonctionnel de ces deux derniers
organes. Lors de la ventilation mécanique d’un poumon sain (exemple: patient lors d’une
anesthésie générale élective), c’est la variation du retour veineux et du remplissage
ventriculaire droit qui détermine en grande partie le «swing» du volume d’éjection systolique
de ce dernier [67]. Au contraire, dans le cas de patients déjà porteurs d’une hypertension
artérielle pulmonaire ou d’une hyperinflation dynamique (auto-PEP) secondaire à une
maladie pulmonaire obstructive ou induite par la ventilation mécanique (SDRA) [28], c’est la
variation de l’éjection ventriculaire droite qui détermine en grande partie le «swing» du
volume d’éjection systolique de ce dernier [22, 65].
En prenant de la hauteur par rapport à ces données, on peut oser la conclusion suivante : les
interactions cardio-pulmonaires lors d’une ventilation mécanique à pression positive sont
caractérisées par une sensibilité extrême à l’état de remplissage intravasculaire du patient
27
ventilé. Cet effet est surtout identifiable chez le patient présentant des poumons avec une
compliance normale ou élevée (emphysème pulmonaire). En effet, une diminution du débit
cardiaque par baisse du remplissage ventriculaire peut être objectivée lors de l’instauration
d’une ventilation à pression positive. Ce phénomène est dû au fait que le cœur et les poumons
sont anatomiquement et fonctionnellement liés et que les variations de la pression intra-
thoracique et des volumes pulmonaires induits par la ventilation modifient cycliquement
l’interaction cardio-pulmonaire.
C. MONITORAGE HEMODYNAMIQUE DES DEFAILLANCES
CIRCULATOIRES : REVUE DE LA LITTERATURE
La prise en charge thérapeutique d’une défaillance circulatoire aiguë chez le patient de soins
intensifs (SI) comprend le plus souvent un remplissage vasculaire auquel on peut adjoindre
l’administration de catécholamines inotropes plus ou moins vasopressives. Au-delà de
l’intérêt d’un diagnostic précis de l’affection causale, la prise en charge adéquate de cette
insuffisance circulatoire aux conséquences souvent gravissimes repose principalement sur sa
mise en évidence et l’enregistrement continu des paramètres hémodynamiques tout au long de
son évolution. En anesthésie et aux SI, cet acte est désigné par le mot « monitorage » issu du
latin « monere » qui signifie avertir, alerter, prévenir...etc. Le diagnostic le plus précoce
possible des états de défaillance cardio-circulatoire constitue, aux SI, un objectif essentiel.
Une réanimation tardive même bien menée permettra rarement la restauration de la viabilité
des cellules mortes dans les différents organes du fait d’un état de choc prolongé [68]. La
prise en charge globale et précoce des états de choc aux urgences conditionne l'évolution
ultérieure [69]. Pour preuve, l'amélioration de l'état hémodynamique dès la réanimation
initiale faite aux urgences permet de réduire la mortalité et de diminuer la survenue de
défaillances viscérales secondaires [69].
Aux SI, le monitorage hémodynamique des insuffisances circulatoires repose principalement
sur la mesure continue de la fréquence cardiaque et la pression artérielle (PA) par méthode
oscillométrique [70]. Ces méthodes minimales sont utilisées afin d'évaluer l'état
hémodynamique en situation d'urgence. Secondairement, si la défaillance circulatoire se
28
précise, la mise en place d’un cathéter artériel sanglant dans un vaisseau périphérique permet
l’enregistrement en continue de l’onde de pression artérielle et, conséquemment, la mesure
des valeurs systolo-diastoliques ainsi que le calcul de la valeur de la pression artérielle
moyenne et de la PP (Figure 7). Dans le cas où l’origine d’une instabilité hémodynamique
n’est pas diagnostiquée, les déterminants principaux de la valeur de la pression artérielle
doivent être mesurés et certaines de leurs dérivées calculées. En effet, en appliquant
conjointement la loi d’Ohm et la loi de Poiseuille - qui gouvernent le débit moyen d’un fluide
newtonien (fluide à viscosité constante) circulant sur un tube cylindrique rigide – et en
assimilant la pression veineuse centrale à zéro, on peut décrire la pression aortique moyenne
(PAM) par l’équation suivante : PAM = DC x RVS, où DC représente le débit cardiaque et
RVS la résistance vasculaire systémique. Le DC étant dépendant de l’état de remplissage des
ventricules (précharge) et de la contractilité myocardique, ces deux paramètres doivent être
évalués. De ce fait, une évaluation précise du volume sanguin circulant effectif (ou contraint
« stressed ») ainsi que la mesure du DC doivent être envisagées afin de préciser le mécanisme
de l'état de choc et permettre sa prise en charge. Aux SI, le monitorage des précharges
ventriculaires et du débit cardiaque a longtemps été obtenu au moyen du cathéter artériel
pulmonaire (CAP) de Swan-Ganz [71]. En effet, ce dernier permet de mesurer le débit
cardiaque droit par thermodilution (équation de Stewart-Hamilton) et d’estimer la précharge
ventriculaire droite et gauche, via, respectivement, la mesure de la pression veineuse centrale
et celle de la pression artérielle pulmonaire occluse [71].
Figure 7 : Courbe sanglante de pression aortique (trait Plein) et périphérique (trait pointillé).
Longtemps considéré comme le « gold standard » du monitorage hémodynamique, le CAP est
de plus en plus vivement critiqué du fait des possibles effets délétères associés à son
29
utilisation [72]. De ce fait, des études prospectives randomisées contrôlées européennes et
effectuées outre-atlantique ont été menées ces dix dernières années et elles ont montré
l’absence de bénéfice de l’utilisation du dit cathéter pour la prise en charge hémodynamique
des patients dans les SI [73-75].
Récemment, des progrès techniques ont permis d'avoir accès de façon rapide et non invasive à
une exploration hémodynamique très exhaustive, ainsi qu'à des moyens diagnostiques
remarquables. Ces nouveaux procédés, sont de plus en plus utilisés pour la surveillance de la
défaillance circulatoire du patient des SI avec le bénéfice de présenter des mesures aussi
précises que la cathéter artériel pulmonaire, tout en étant des outils présentant un risque
moindre d’iatrogénicité une fois comparés aux techniques invasives sanglantes,
potentiellement dangereuses [76-82]. La présente revue de la littérature se fixe comme
objectif de résumer ces techniques et les concepts qui leurs sont sous-jacents, et de faire une
analyse scientifique et critique de ces techniques.
I. TECHNIQUES NON INVASIVES MONITORANT QUALITATIVEMENT LES DEFAILLANCES
CIRCULATOIRES
Dans ce paragraphe nous n’aborderons que les techniques de monitorage cardiovasculaire qui
ne permettent qu’une estimation qualitative de l’état cardio-circulatoire tout en ne nécessitant
aucun matériel invasif.
1. LE MONITORAGE CLASSIQUE : FREQUENCE CARDIAQUE, PRESSION ARTERIELLE NON
INVASIVE ET PLETHYSMOGRAPHIE
- La fréquence cardiaque est un très mauvais indice de monitorage hémodynamique du
fait de son manque de variation ou au contraire de sa très grande variabilité lors d’ une
détresse hémodynamique [83].
- La mesure automatique non invasive de la pression artérielle (PA) est aujourd'hui
facilement accessible grâce aux techniques oscillométriques, mais l'interprétation des
valeurs mesurées peut être délicate [84]. Cette technique mesure la PAM et calcule la
PAS et la PAD [70]. De plus, la mesure non invasive de PA par cette même technique
est discontinue et elle ne constitue donc pas stricto sensu un authentique moyen de
30
monitorage. Une autre limite de cette technique réside dans son emploi chez un patient
choqué en hypovolémie aiguë, situation où un défaut de corrélation entre la PAM et le
DC suggère que la PA constitue un mauvais indicateur de l’état circulatoire [85]. En
effet, la PAM est longtemps maintenue grâce aux mécanismes compensateurs
adrénergiques, puis elle chute brutalement et de façon non synchrone au DC. De plus,
en l’absence de mesure fiable du débit cardiaque et de la précharge ventriculaire, le
maintien d'une PA optimale est une condition indispensable mais non suffisante pour
le maintien d’une perfusion des organes des patients en état de choc. En effet, il est
difficile de définir avec certitude un chiffre critique minimal de PA chez ce type de
patients [86-88].
- La mesure continue de la saturation de l’hémoglobine en oxygène est considérée
comme une avancée notoire dans le monitorage du patient des soins intensifs depuis
l’avènement du l’oxymètre de pouls (SpO2). L'oxymètre de pouls (ou saturomètre)
permet de mesurer de façon simple, fiable, non invasive et continue la saturation
artérielle en oxygène de l'hémoglobine. Le principe repose sur l'émission de deux
lumières (rouge et infrarouge), avec une longueur d’onde, respectivement, de 660 et
940 nm, et de la mesure de leur absorption par le flux pulsatile. L'absorption de la
lumière rouge et infrarouge sera variable selon qu'elle rencontrera de l'hémoglobine
non oxygénée ou oxygénée. Le résultat est très bien corrélé à la saturation artérielle
mesurée par l’analyse des gaz du sang artériel (SaO2). L’oxymètrie de pouls n'étudie
que l’échantillon sanguin pulsatile de l'absorption des rayonnements, c'est-à-dire les
changements du volume sanguin induits par l'influx artériel dans l'échantillon de tissu
exploré. Aussi, de façon qualitative, la disparition du signal pulsatile de
pléthysmographie affiché sur la plupart des appareils au lit du patient doit constituer
un signal d'alerte d'une hypoperfusion des extrémités. Une étude récente réalisée chez
37 patients de soins intensifs a même objectivé qu’un index quantitatif de perfusion
périphérique découlant du signal de pléthysmograhie peut être utilisé cliniquement
pour déceler une défaillance circulatoire [89]. De plus, récemment il a été objectivé
que les variations de la courbe de pléthysmographie induites par la ventilation
mécanique sont corrélées aux variations de la courbe de PA induites par la ventilation
par pression positive [90, 91], ce qui rend ce monitorage non invasif un outil idéal
31
pour estimer la réponse au remplissage du patient ventilé artificiellement (voir
chapitre suivant).
2. LA CAPNOGRAPHIE EXPIREE
La mesure du CO2 dans l'air expiré informe sur la valeur de CO2 éliminée par les poumons
du patient. Par ce moyen, on peux analyser la production tissulaire de CO2 et son transport
des tissus vers les poumons via le système circulatoire. De ce fait, la mesure du CO2 expiré
(capnographie expirée) est classiquement monitorée en anesthésie sous la forme d’un recueil
d'une valeur ponctuelle de pression télé-expiratoire en CO2 (PetCO2). De nombreux
déterminants ont un impact sur cette mesure, car elle dépend non seulement de la pression
artérielle partielle en CO2 (PaCO2), mais aussi de la différence alvéolo-artérielle en CO2
(adéquation du rapport ventilation perfusion, espace mort, « mismatch »). En conséquence,
bien que cette valeur soit aussi dépendante du débit cardiaque et de l’état hémodynamique,
comme mesure, elle ne peut être considérée comme un bon indicateur hémodynamique [92-
94]. Plus encore, dans les états de choc sévères, la production tissulaire en CO2 diminue par
défaut d'apport en O2 aux cellules. Reste les situations extrêmes, comme l'arrêt cardio-
respiratoire [95], ou la capnographie expirée peut être utile pour refléter l'efficacité de la
réanimation entreprise [96, 97]. Enfin, cette technique n’est utilisable que chez le patient
intubé.
3. MESURE DE LA PRESSION TRANSCUTANEE EN CO2 (TCPCO2)
Le monitorage de la PaCO2 peut s’avérer être un indice important pour la prise en charge du
patient en état de choc. En effet, chez le patient ventilé, l'évolution de la TcPCO2 peut faire
également partie de l'évaluation hémodynamique sous certaine conditions [98]. Récemment,
des appareils de mesure transcutanée de la PaCO2 ont vu le jour et ils commencent à être
utilisés dans le domaine anesthésique [99] et aux soins intensifs [100, 101]. Ces appareils, qui
permettent une mesure non invasive, ont été validés quant à leurs précision et à leur innocuité
chez des patients insuffisants respiratoires [102]. La mesure de la TcPCO2 (à l'aide d'une
électrode posée sur la peau), comme c’est le cas pour la pléthysmographie, permet
l'évaluation de la perfusion cutanée. Le principal inconvénient de cette méthode est le temps
incompressible de chauffage et de calibration des électrodes. Récemment cette technique à été
proposée comme un moyen de monitorage continu pour réaliser les tests d’apnée afin
d’affirmer la mort cérébrale en vue d’une transplantation d’organes [103].
32
II. TECHNIQUES NON INVASIVES MONITORANT QUANTITATIVEMENT LES
DEFAILLANCES CIRCULATOIRES
1. TECHNIQUES NON INVASIVES MONITORANT LA REPONSE AU REMPLISSAGE VASCULAIRE CHEZ
LES PATIENTS VENTILES MECANIQUEMENT.
A ce jour, rares sont les indices non invasifs permettant de quantifier les précharges
ventriculaires droite ou gauche et/ou de prédire la réponse hémodynamique lors d’un
remplissage vasculaire par infusion de volume. En effet, en dehors de l’échocardiographie,
une seule technique à été développée aujourd’hui, par l’auteur, en collaboration avec une
équipe française [104]. Cette technique est basée sur la mesure d’un intervalle de temps
systolique qu’est le temps de contraction isovolumique (« pre-ejection period ») [59]. En
effet, au moyen de la courbe de pléthysmographie et de l’enregistrement
électrocardiographique, cet intervalle de temps systolique a été mesuré et sa variation
ventilatoire en pourcentage estimée chez les patients en ventilation mécanique [104]. Dans
cette étude récente, les auteurs ont montré qu’une variation respiratoire du temps de
contraction isovolumique, mesurée à l’aide d’ une courbe de pléthysmographie, supérieure à
4%, était capable de prédire la réponse positive au remplissage vasculaire avec une
sensibilité de 100% et une spécificité de 67% [104].
Par ailleurs, grâce à l’échocardiographie bidimensionnelle, Feissel et collaborateurs, ainsi que
Barbier et al., ont démontré que la variation respiratoire du diamètre de la veine cave
inférieure était un indice prédictif de la réponse au remplissage vasculaire chez les patients
profondément sédatés et ventilés en pression positive [58, 60]. Néanmoins, toutes ces
techniques sont inutilisables chez les patients qui ne seraient pas profondément sédatés, ou
qui sont en ventilation spontanée et/ou qui présentent des arythmies cardiaques.
2. TECHNIQUES NON INVASIVES MONITORANT LA REPONSE AU REMPLISSAGE VASCULAIRE CHEZ
LES PATIENTS EN VENTILATION SPONTANEE: LE LEVER DE JAMBES PASSIF
Récemment, une équipe parisienne a proposé la technique du lever de jambes passif pour
estimer la réponse au remplissage vasculaire [105]. Cette technique se basait sur une étude
plus ancienne qui avait montré que l’augmentation du retour veineux secondaire aux lever des
33
jambes faisait varier des indices de réponse au remplissage validés, comme la variation
respiratoire de la pression pulsée [106]. Dans ce récent travail, Monnet et collaborateurs ont
démontré qu’une augmentation supérieure à 10% du débit aortique (mesuré par Doppler
œsophagien [107]) lors d’un lever de jambes passif prédisait la réponse au remplissage
vasculaire avec une sensibilité de 97% et une spécificité de 94% [105].
3. TECHNIQUES NON INVASIVES MONITORANT LA MESURE DE DEBIT CARDIAQUE CHEZ LES
PATIENTS DES SOINS INTENSIFS.
Aucune méthode de mesure du débit cardiaque disponible en clinique n’est dépourvue de
limitation. Par conséquent, aucune d’entre elles n‘est très précise, voir a fait preuve de sa
supériorité par rapport à une autre. La thermodilution est la méthode la plus répandue et qui
fait office de référence. La mesure de débit cardiaque par des moyens non invasifs a été
largement développée durant les vingt dernières années. A ce jour, chercheurs et entreprises
biomédicales développent de plus en plus d’appareils pour permettre, grâce à des concepts
datant souvent du 19ème siècle, de mesurer de façon fiable et précise le débit cardiaque.
a. Mesure du débit cardiaque au moyen de la méthode Fick appliquée au CO2.
Une technique de ré-inhalation partielle du CO2 et l'utilisation de la version différentielle de
l'équation de Fick appliquée au CO2 permet la mesure non-invasive du débit cardiaque [98,
108-114]. Le système NICO® de Novametrix® est un moniteur de débit cardiaque qui repose
sur le principe que le débit cardiaque est égal à la production de CO2 indexée sur la différence
arterio-veineuse en CO2. De ce fait et grâce à la capnographie expirée, la production de CO2
peut être mesurée chez un patient intubé et ventilé en pression positive. Par ailleurs, le
contenu en CO2 artériel revient à mesurer PetCO2. Reste la question de mesurer le contenu
veineux en CO2. Pour cela, une valve est actionnée automatiquement toutes les 3 à 4 minutes,
avec l’installation d’une boucle de ré-inhalation pendant 50 secondes. Cela permet à l’alvéole
de se saturer en CO2 et mène à ce que le contenu alvéolaire en CO2 soit en équilibre avec le
contenu capillaire veineux. A cet instant même, la mesure du PetCO2 équivaut à la mesure du
contenu capillaire veineux en CO2 (Figure 8). La fraction de shunt est estimée par l'oxymétrie
de pouls et la concentration inspirée en oxygène. Néanmoins, cette technique de mesure
nécessite que les patients soient intubés et placés sous ventilation mécanique. Les premiers
34
résultats chez l'homme montrent une bonne corrélation avec la mesure du débit cardiaque par
thermodilution et par échocardiographie-Doppler [108-114]. Ce type de monitorage, peut
avoir un intérêt dans les prises en charge de patients en situation pré-hospitalière [98].
Néanmoins, certaines limites de cette technique doivent être connues, comme un changement
du métabolisme énergétique durant la mesure, une ventilation/minute instable et une variation
du shunt intra-pulmonaire. Par ailleurs, le praticien doit être conscient du risque
d’hypercapnie et d’acidose respiratoire chez les patients présentant des affections
respiratoires obstructives.
Figure 8 : Principe de mesure du débit cardiaque moyennant la méthode Fick appliquée au
CO2.
b. Mesure du débit cardiaque au moyen de la méthode d’impédancemetrie thoracique.
L'entrée et la sortie de sang dans le thorax à chaque systole provoquent des modifications des
propriétés électriques du thorax qui peuvent être mesurées par le calcul de l'impédance
thoracique. Le volume de la cavité thoracique étant estimé à partir du poids, de la taille et du
sexe du patient, l'impédance thoracique instantanée est calculée par l’application d'un courant
électrique de faible amplitude et de haute fréquence entre deux paires d'électrodes, l'une
35
placée sur le cou et l'autre sur l'abdomen (Figure 9) [115, 116]. Basé sur l’équation de
Kubicek [116], le traitement informatique des données obtenues pour chaque cycle permet
l'estimation du volume d’éjection systolique. Cette technique, élégante et non invasive, se
heurte toutefois à certaines difficultés techniques telles que l'acquisition du signal ou des
défauts de validité de la modélisation informatique et géométrique du système [117, 118].
Aussi, les nombreuses études ayant comparé cette technique aux autres méthodes de mesure
du débit cardiaque ont donné des résultats jugés suffisants mais variables d'une étude à l'autre.
Ceci étant expliqué par une qualité de mesure altérée chez les patients de réanimation [117-
119, 120 ]. Ces appareils sont coûteux (en moyenne 25000 à 30000 CHF), alors que leur
fiabilité reste encore à démontrer. Par ailleurs, la ventilation mécanique lorsqu’elle s’assortit à
des grandes variations des volumes d’air intrathoraciques, est une limite de la méthode
(fréquence respiratoire et volume/minute instables).
Figure 9 : Mesure du débit cardiaque par impédancemetrie après application d'un courant électrique
de faible amplitude et de haute fréquence entre deux paires d'électrodes, l'une placée sur le cou et
l'autre sur le thorax.
c. Mesure du débit cardiaque au moyen de la méthode du Doppler oesophagien.
La mesure semi-invasive (Doppler œsophagien) de la vitesse d'écoulement du sang dans
l'aorte thoracique ascendante ou descendante permet une mesure précise du débit cardiaque
[121]. Les principes techniques de la mesure du DC par effet Doppler consiste en l’obtention
d'un signal Doppler continu ou tous les écoulements rencontrés par le faisceau ultrasonore se
trouvent alignés sur le trajet étudié. Le positionnement d'une sonde Doppler à proximité de
36
l'aorte permet l'insonation de l'aorte thoracique descendante et le calcul de l’intégrale temps-
vitesse de l’éjection aortique (VTIAo) [122, 123]. Le volume d’éjection systolique aortique
pourra alors être obtenu selon la formule : VES = VTIAo X (π D2/4), où D est le diamètre du
vaisseaux aortique [121]. Les relations anatomiques étroites entre l'aorte thoracique
descendante et l'œsophage ont permis le développement de la mesure du VES par Doppler
transœsophagien. La sonde Doppler est rapidement et facilement insérée, et elle est descendue
de 35 à 40 cm dans l'œsophage. Le flux sanguin dans l'aorte thoracique descendante est
aisément identifié sur le profil de vélocité affiché à l'écran et surtout au moyen des
caractéristiques sonores typiques du flux aortique (Figures 10 & 11). La limitation technique
réside dans le fait que l’aorte et l’œsophage étant deux conduits parallèles, un angle entre la
sonde Doppler et le flux aortique de 45 à 60° (suivant les différents appareils) est une
condition sine qua non pour la fiabilité de la mesure [121]. Néanmoins, des variantes
anatomiques inter-individuelles peuvent être à l’origine d’un angle situé en dehors de cet
intervalle. De ce fait, dans ce cas bien précis, un écart de 10° pourra résulter sur une mesure
de débit erroné de –25% à +25% de la valeur réelle en fonction de la fréquence d’émission de
la sonde Doppler (4 ou 5 MHz) [121].
Figure 10 : Représentation schématique d’un Doppler œsophagien chez un patient, montrant le
voisinage de l’œsophage et de l’aorte descendante (a). Ondes Doppler obtenues avec mesure des
vélocités maximales (b).
L'obtention d'une mesure d’un signal de Doppler aortique ne prend habituellement que
quelques minutes. La sonde peut alors être laissée en place et elle permet l'enregistrement en
37
continu du flux sanguin aortique. Le déplacement secondaire de la sonde dans l'œsophage est
fréquent : une simple rotation de la sonde est alors souvent suffisante pour récupérer un signal
fiable. La mesure du VES peut se faire par deux moyens distincts : - soit la surface aortique
est mesurée échographiquement (diamètre mesuré ; Hemosonic®, Arrow®) ; - soit la surface
de l’aorte descendante est approchée à partir de nomogrammes intégrant l'âge, le poids, la
taille et la pression artérielle moyenne (CardioQ®, Deltex Medical Ltd®). Considérant que
30 % du VES est destiné aux troncs supra-aortiques, le VES mesuré dans l’aorte descendante
est rehaussé de la valeur correspondante à cette partie du corps. Néanmoins, cet index 30%-
70% entre le territoire céphalique et le territoire caudal peut varier dans des conditions
pathologiques et fausser les approximations du débit [124]. La corrélation entre la mesure du
DC par thermodilution et par Doppler œsophagien a été largement testée en anesthésie et en
réanimation [124, 125]. Les limites d'acceptation actuelles entres les deux techniques de
mesure vont de –1.5 l/min à + 2 l/min avec un coefficient de corrélation r=0.95 [126]. Cette
technique possède une fiabilité suffisante pour orienter la prise en charge thérapeutique et elle
est considérée comme un bon monitorage du débit cardiaque. Par ailleurs, la simple
visualisation du profil de vélocité aortique et le calcul de quelques paramètres mathématiques
simples permet d'orienter les décisions thérapeutiques vers un remplissage vasculaire ou un
traitement par des drogues inotropes [107].
Figure 11 : Principe du calcul du volume d’éjection systolique à partir de la vélocité aortique
(l’intégrale temps vitesse (VTI) représente la distance parcourue (« stroke distance »).
38
d. Mesure du débit cardiaque au moyen de l’échocardiographie-Doppler.
L'échocardiographie-Doppler transthoracique (ETT) est une technique non invasive,
réalisable au lit du patient, rapide et reproductible [127]. Depuis une vingtaine d'années, les
appareils d'échographie cardiaque-Doppler deviennent disponibles dans les unités de
réanimation grâce au développement d'appareils polyvalents permettant, en utilisant
différentes sondes, la réalisation d’un ensemble d’examens utiles aux patients de soins
intensifs. En effet, actuellement, outre que ces appareils permettent la réalisation d’examens
cardiaques, ils permettent d’investiguer le poumon [128, 129] et la circulation cérébrale [130].
L'échographie cardiaque réalisée par voie transthoracique ou par voie transœsophagienne
(ETO) apporte des informations morphologiques cardiaques et hémodynamiques exhaustives
et intéressantes: évaluation de la fonction systolique et diastolique du VG, mesure de la taille
des cavités, recherche d'anomalies valvulaires et péricardiques [131], visualisation d'un shunt
intracardiaque, diagnostic des pathologies de l'aorte thoracique et détection d'une d'embolie
pulmonaire [132, 133]. Cependant l'ETO, technique semi-invasive relève exclusivement du
recours à des échocardiographistes chevronnés [127].
Au cours des états de choc, l'ETT permet le plus souvent d'orienter la thérapeutique et de
l'adapter régulièrement du fait de sa reproductibilité [28, 65]. Le principal obstacle à
l'utilisation de l'échocardiographie en réanimation est l'obligation de disposer d'opérateurs
entraînés susceptibles de répondre aux demandes dans des délais très brefs et à toute heure
[127]. De plus, la qualité et l'intérêt des examens pratiqués dépendra du niveau de formation
et de l'entraînement de l’intensiviste [127]. Plus encore, celui-ci devra connaître ses limites et
recourir aux cardiologues spécialisés en cas de difficultés. Les sociétés cardiologiques
considèrent qu'une expérience de trois mois à plein temps et de 150 échocardiographies sont
nécessaires pour prétendre à une réelle compétence [134-137]. L’intensiviste pourrait se
limiter à une formation rigoureuse minimale, mais suffisante pour répondre à certaines
questions précises auxquelles l'échocardiographie apporte des réponses facilement accessibles
[137].
39
Le débit cardiaque peut être calculé moyennant l’échocardiographie-Doppler après la mesure
du volume d’éjection systolique au niveau de l’anneau aortique. A ce niveau et grâce à
l’échocardiographie bidimensionnelle, le diamètre (D) de l’anneau est mesuré pendant la
systole (repérage temporel grâce à l’onde T sur l’ECG) [138]. Par la suite, l’intégrale temps-
vitesse de l’éjection aortique (VTIAo) pourra alors être obtenu moyennant le Doppler pulsé au
niveau de l’anneau. Le VES = VTIAo X (π D2/4), avec le DC qui sera égal à cette valeur
multipliée par la fréquence cardiaque pendant la mesure (Figure 12).
Par ailleurs, l’échocardiographie permet aussi une estimation de la PVC grâce à la mesure du
diamètre télé-expiratoire de la veine cave inférieure, en fenêtre sous-costale, 2 cm avant son
abouchement dans l'oreillette droite [139 , 140, 141] ou grâce au recours au Doppler tissulaire
(DTI) de l’anneau tricuspidien [142]. Les pressions de remplissages du ventricule gauche
peuvent être estimées au moyen du DTI de l’anneau mitral [143-146].
Ainsi, l’échocardiographie représente actuellement une technique majeure qui permet dans la
plupart des défaillances cardio-circulatoires d'orienter le diagnostic et la prise en charge
thérapeutique d'un état de choc [65]. La répétition aisée de cet examen au chevet du patient
constitue un avantage important pour vérifier la réponse aux thérapeutiques [147, 148]. Reste,
qu’il faut garder à l’esprit que cette technique ne monitore pas de façon continue le malade,
mais qu’elle permet un « status » discontinu au cours de l’évaluation hémodynamique. De ce
fait, en cas d’une prise en charge de multiples patients instables sur le plan hémodynamique
lors d’une garde nocturne par exemple, la possibilité de monitorage des patients par cette
technique non invasive est assez limitée.
40
Figure 12 : Principe et exemple du calcul du débit cardiaque et du volume d’éjection systolique à
partir de la vélocité au niveau de l’anneau aortique en utilisant l’échocardiographie-Doppler.
e. Évaluation du débit cardiaque au moyen de l’analyse de l'onde de pression artérielle (méthode
du « pulse contour »).
La possibilité de déterminer le débit cardiaque en utilisant la méthode du « pulse contour » est
une méthode étudiée par les cliniciens et chercheurs depuis des décennies [149-151]. Des
succès préliminaires ont été obtenus en utilisant la méthode de surface sous la courbe ou par
l’analyse de divers composants de l’onde de pression artérielle [152-155]. Aux soins intensifs,
la mise en place d'un cathéter artériel permet la mesure en continu du volume d’éjection
systolique, battement par battement, selon la méthode du « pulse contour ». L'onde de PA est
séparée en deux parties successives par l'incisure dicrote (Figure 13). Une relation linéaire
entre l'aire sous la courbe de la première partie systolique de l'onde de PA et le VES du VG
est démontrée [155]. De ce fait, l’aire sous la courbe de la composante systolique de l’onde de
pression artérielle est déterminée par la relation volume d’éjection systolique-onde de
pression associée à cette éjection. Cette relation peut varier d’un individu à l’autre, et chez un
même individu en fonction des changements des conditions cliniques. Cette relation chez un
même individu, permet de calculer une constante « K » qui pourra être utilisée dans des
mesures ultérieures.
41
La mesure du débit cardiaque par une autre technique (thermodilution classique ou
transpulmonaire, dilution d’un colorant, ultrasons, Doppler, …) permet l'étalonnage de la
relation entre la surface sous la courbe de l'onde systolique de PA et le VES. Cet étalonnage,
en fait caractérise la valeur K de cette constante. Ainsi, la surveillance battement par
battement de l'onde de PA permet le calcul battement par battement du VES et ainsi du débit
cardiaque. Cette méthode a été validée en réanimation même si les variations de la résistance
vasculaire systémique altèrent la qualité des mesures et changent donc la valeur K. Aussi, les
variations de cette surface (calculable à chaque battement cardiaque) permettent d’acquérir
les variations respiratoires du VES battement par battement chez les patients placés en
ventilation mécanique. Ces variations respiratoires du volume d’éjection systolique chez les
patients sédatés profondément, sans arythmies et ventilés en pression positive, sont des
indices fiables de la réponse au remplissage intravasculaire [156].
Figure 13 : Principe de la mesure du volume d’éjection systolique battement par battement
selon la méthode du « pulse contour »
Actuellement, plusieurs compagnies possèdent des logiciels de mesure du débit cardiaque, en
utilisant les techniques du « pulse contour ». L’appareil PiCCO® de « Pulsion systems® »
utilise une calibration par thermodilution transpulmonaire [156]. L’appareil Pulsco® de
« LIDCO® » utilise une calibration par dilution d’un indicateur au lithium [157, 158].
D’autres systèmes plus récents n’utilisent pas de calibration et déterminent la constante K par
l’analyse de la courbe de pression artérielle [153, 159]. Néanmoins, à l’opposé des méthodes
« pulse contour » qui utilisent une calibration, les outils de mesure sans calibration n’ont pas
été encore validés par des études cliniques.
42
INVESTIGATIONS PERSONNELLES
Dans la seconde partie de ce travail, l’analyse des études réalisées par l’auteur sur le
monitorage des patients de soins intensifs sera présentée.
43
A. INDICES PREDICTIFS DE LA REPONSE AU REMPLISSAGE VASCULAIRE
CHEZ LES PATIENTS DES SOINS INTENSIFS VENTILES PAR PRESSION
POSITIVE OU RESPIRANT SPONTANEMENT :
L’estimation de la réponse à un remplissage vasculaire est une donnée qui doit être acquise
lors du monitorage cardiovasculaire d’un patient de soins intensifs. C’est pourquoi il convient
de proposer des paramètres qui sont à même de déterminer, avant l’administration
intraveineuse d’un soluté de remplissage, si un patient va augmenter ou non son débit
cardiaque, donc bénéficier ou non du remplissage intra-vasculaire.
Après la découverte du cathéter artériel pulmonaire de Swan et Ganz, des indices dits
statiques, comme la pression veineuse centrale et la pression artérielle pulmonaire occluse,
ont été proposés pour prédire quelle serait la réponse du cœur en terme de débit dans les
suites de l’infusion d’un volume intra-vasculaire chez les patients sédatés profondément et
ventilés mécaniquement [160-163]. Toutefois, à ce jour, ces indices statiques n’ont pas fait la
preuve de leur utilité pour véritablement prédire quelle pourrait être la réponse à un
remplissage vasculaire chez ce type de patients [8, 10]. De même, à ce jour, rare sont les
études qui ont investigué la prédiction de la réponse au remplissage chez une population
homogène de patients ventilant spontanément [105, 164-166].
A l’opposé, des indices dits dynamiques, comme la quantification des variations
respiratoires de la pression artérielle systémique, celle du temps de pré-éjection et/ou du
volume d’éjection systolique ont été validés pour prédire les effets hémodynamiques d'une
expansion volémique chez les patients ventilés mécaniquement sous sédation profonde [8,
10]. Ces résultats découlent du fait que la ventilation mécanique à pression positive induit des
variations cycliques mesurables du volume d'éjection ventriculaire gauche. L'amplitude de ces
variations respiratoires du volume d'éjection ventriculaire gauche témoigne ainsi de la
sensibilité des deux ventricules aux variations de précharge faisant suite à l'insufflation
mécanique. C’est pourquoi, en pratique clinique, chez les patients sédatés profondément et
ventilés mécaniquement, la variabilité respiratoire du volume d'éjection ventriculaire gauche
peut être évaluée, permettant ainsi de prédire la réponse au remplissage vasculaire [10].
Néanmoins, en dehors de la variation respiratoire de la pression veineuse centrale [164], ces
indices ne sont pas utiles pour prédire l’augmentation du débit cardiaque après un remplissage
44
vasculaire chez les patients qui ventilent spontanément ou qui déclenchent un respirateur
(mode en aide respiratoire) [105, 166]. Dans cette dernière étude investiguant des patients qui
ventilent spontanément, Heenen et collaborateurs ont objectivé que les indices statiques
(pression veineuse centrale et pression artérielle pulmonaire occluse) étaient meilleurs
prédicteurs de la réponse au remplissage que les indices dynamiques [166]. Le message
intéressant de cette dernière étude n’est pas le fait qu’un indice dynamique objectivant les
variations respiratoires du volume d’éjection systolique soit inutile dans ce contexte (ce qui
était pressenti depuis de nombreuses années et récemment étudié [105]), mais c’est le fait que
des indices statiques comme la pression veineuse centrale et/ou la pression artérielle
pulmonaire occluse se trouvent être meilleurs prédicteurs [166]. Concernant ce type de
patients qui ventilent spontanément, il est impératif qu’à l’avenir soient réalisées de nouvelles
études prospectives et mécanistiques pour faire la part de ces discordances [164, 166].
L’auteur de la présente thèse expose des données dans ce sens.
Tout d’abord, deux revues de la littérature objectivant que les indices dynamiques de réponse
au remplissage vasculaire sont d’une plus grande utilité que les indices statiques chez des
patients des soins intensifs profondément sédatés et placés sous ventilation mécanique [8].
Néanmoins, si chez les patients sédatés profondément la question est résolue, suite à une
étude récente et contradictoire cette affirmation reste à prouver chez les patients qui ventilent
spontanément [164, 166].
Concernant les indices statiques, l’auteur s’est attaché à discuter la question de la validité
de la mesure de la pression veineuse centrale (pression auriculaire droite) comme un index
valide ou non du retour veineux [25, 167-169] ? Cette question est débattue depuis longtemps
au sein de la communauté des physiologistes cardio-vasculaires, sans être vraiment résolue.
En particulier, la question de savoir si la pression veineuse centrale est un déterminant ou une
conséquence du retour veineux demeure largement discutée. Cette revue de la littérature
permet d’estimer physiopathologiquement la relevance clinique de la variation respiratoire de
la pression veineuse centrale comme indice de remplissage vasculaire chez les patients
ventilés mécaniquement [164]. Dans une autre étude clinique, prospective, nous avons montré
que la mesure par voie échocardiographique du diamètre de la veine cave inférieure était
corrélée à la pression veineuse centrale chez des patients placés en ventilation mécanique
45
[140]. Ce résultat avait déjà été montré par plusieurs autres auteurs [139, 170, 171]. Toutefois,
la robustesse de la relation entre ces deux paramètres, assurée par la mesure de leurs
coefficients de corrélation, était différente dans ces trois dernières études. L’originalité de
notre travail a été d’expliquer cette discordance en raison de différences dans les méthodes de
mesures échocardiographiques utilisées. De plus, nous avons objectivé l’importance du
moment de la mesure échocardiographique au niveau du cycle cardiaque. [140, 169].
Concernant les indices dynamiques, notre groupe a montré que l’ « intervalle de temps
systolique » représenté par le temps de prééjection (mesuré au moyen de
l’électrocardiogramme et d’une courbe de pression artérielle sanglante) pouvait varier de
façon dynamique au cours d’un cycle ventilatoire en respiration mécanique à pression
positive et que l’amplitude de cette variation était en relation directe avec la réponse au
remplissage vasculaire après une infusion de volume [59]. Utilisant la courbe de
pléthysmographie et l’enregistrement électrocardiographique, cet intervalle de temps
systolique a été mesuré. Sa variation ventilatoire a été estimée chez des patients placés en
ventilation mécanique [104]. Ce travail a permis de démontrer qu’une variation respiratoire de
la période qui précédait l’éjection ventriculaire gauche, mesurée de manière non invasive à
l’aide d’une courbe de pléthysmographie, quand elle était supérieure à 4%, prédisait une
réponse positive au remplissage, avec une sensibilité de 100% et une spécificité de 67%
[104]. Plus récemment, utilisant la courbe de plethysmograhie pulsée, nous avons objectivée
que la variation respiratoire de l’amplitude de l’onde de pouls plethysmographique
(plethysmographie pulsée) était un indice utile pour prédire la réponse au remplissage
vasculaire chez des patients septiques, profondément sédatés et placés en ventilation
mécanique [172]. Une variation respiratoire de l’amplitude de l’onde de pouls
plethysmographique supérieure à 14 %, prédisait une réponse positive au remplissage, avec
une sensibilité de 94% et une spécificité de 80% [172].
46
Reprints of :
• “Bendjelid K, Romand J-A. Fluid responsiveness in mechanically ventilated patients. A review of
indices used in intensive care. Intensive Care Med 2003 ; 29 : 352-360.” [8]
• “Coudray A, Romand J-A, Treggiari MM, Bendjelid K. Fluid responsiveness in spontaneous breathing
patients. A review of indices used in intensive care. Crit Care Med 2005; 33:2757-62.” [173]
• “Bendjelid K, Romand J-A, Walder B , Suter P.M, Fournier G. The correlation between measured
inferior vena cava diameter and right atrial pressure depends on the echocardiographic method used in
patients who are mecanichaly ventilated. J Am Soc Echocardiogr 2002; 15: 944-9.” [140]
• “Bendjelid K. Right Atrial Pressure: determinant or result of change in venous return ? Chest 2005,
128 : 3639-40.” [169]
• “Bendjelid K, Suter PM, Romand J-A. The respiratory change in preejection period: a new method to
predict fluid responsiveness. J Applied Physiol 2004; 96: 337-42.” [59]
• “Feissel M, Badie J, Merlani PG, Faller JP, Bendjelid K. Preejection period variations predict the fluid
responsiveness of septic ventilated patients. Crit Care Med 2005; 33 : 2534-9.” [104] • “Feissel M, Teboul JL, Merlani P, Badie J, Faller JP, Bendjelid K (2007) Plethysmographic dynamic
indices predict fluid responsiveness in septic ventilated patients. In press, Intensive Care Med” [172]
47
Received: 21 November 2002Accepted: 21 November 2002Published online: 21 January 2003© Springer-Verlag 2003
Abstract Objective: In mechanical-ly ventilated patients the indiceswhich assess preload are used withincreasing frequency to predict thehemodynamic response to volumeexpansion. We discuss the clinicalutility and accuracy of some indiceswhich were tested as bedside indica-tors of preload reserve and fluid re-sponsiveness in hypotensive patientsunder positive pressure ventilation.Results and conclusions: Although
preload assessment can be obtainedwith fair accuracy, the clinical utilityof volume responsiveness-guidedfluid therapy still needs to be dem-onstrated. Indeed, it is still not clearwhether any form of monitoring-guided fluid therapy improves sur-vival.
Keywords Positive pressure ventilation · Hypotension · Volumeexpansion · Cardiac index
Intensive Care Med (2003) 29:352–360DOI 10.1007/s00134-002-1615-9 R E V I E W
Karim BendjelidJacques-A. Romand
Fluid responsiveness in mechanically ventilated patients: a review of indices used in intensive care
Prediction is very difficult, especially about the future.Niels Bohr
Introduction
Hypotension is one of the most frequent clinical signsobserved in critically ill patients. To restore normalblood pressure, the cardiovascular filling (preload-defined as end-diastolic volume of both ventricularchambers), cardiac function (inotropism), and vascularresistance (afterload) must be assessed. Hemodynamicinstability secondary to effective or relative intravascularvolume depletion are very common, and intravascularfluid resuscitation or volume expansion (VE) allows res-toration of ventricular filling, cardiac output and ulti-mately arterial blood pressure [1, 2]. However, in theFrank-Starling curve (stroke volume as a function of pre-load) the slope presents on its early phase a steep portionwhich is followed by a plateau (Fig. 1). As a conse-quence, when the plateau is reached, vigorous fluid re-suscitation carries out the risk of generating volumeoverload and pulmonary edema and/or right-ventriculardysfunction. Thus in hypotensive patients methods ableto unmask decreased preload and to predict whether car-
diac output will increase or not with VE have beensought after for many years. Presently, as few methodsare able to assess ventricular volumes continuously anddirectly, static pressure measurements and echocardio-graphically measured ventricular end-diastolic areas areused as tools to monitor cardiovascular filling. Replacingstatic measurements, dynamic monitoring consisting inassessment of fluid responsiveness using changes in sys-tolic arterial pressure, and pulse pressure induced bypositive pressure ventilation have been proposed. Thepresent review analyses the current roles and limitationsof the most frequently used methods in clinical practiceto predict fluid responsiveness in patients undergoingmechanical ventilation (MV) (Table 1).
One method routinely used to evaluate intravascularvolume in hypotensive patients uses hemodynamic re-sponse to a fluid challenge [3]. This method consists ininfusing a defined amount of fluid over a brief period oftime. The response to the intravascular volume loadingcan be monitored clinically by heart rate, blood pressure,pulse pressure (systolic minus diastolic blood pressure),and urine output or by invasive monitoring with the mea-surements of the right atrial pressure (RAP), pulmonaryartery occlusion pressure (Ppao), and cardiac output.Such a fluid management protocol assumes that the in-
K. Bendjelid (✉) · J.-A. RomandSurgical Intensive Care Division,Geneva University Hospitals,1211 Geneva 14, Switzerlande-mail: [email protected].: +41-22-3827452Fax: +41-22-3827455
48
Verwendete Distiller 5.0.x Joboptions
Dieser Report wurde automatisch mit Hilfe der Adobe Acrobat Distiller Erweiterung "Distiller Secrets v1.0.5" der IMPRESSED GmbH erstellt. Sie koennen diese Startup-Datei für die Distiller Versionen 4.0.5 und 5.0.x kostenlos unter http://www.impressed.de herunterladen. ALLGEMEIN ---------------------------------------- Dateioptionen: Kompatibilität: PDF 1.2 Für schnelle Web-Anzeige optimieren: Ja Piktogramme einbetten: Ja Seiten automatisch drehen: Nein Seiten von: 1 Seiten bis: Alle Seiten Bund: Links Auflösung: [ 600 600 ] dpi Papierformat: [ 595 785 ] Punkt KOMPRIMIERUNG ---------------------------------------- Farbbilder: Downsampling: Ja Berechnungsmethode: Bikubische Neuberechnung Downsample-Auflösung: 150 dpi Downsampling für Bilder über: 225 dpi Komprimieren: Ja Automatische Bestimmung der Komprimierungsart: Ja JPEG-Qualität: Mittel Bitanzahl pro Pixel: Wie Original Bit Graustufenbilder: Downsampling: Ja Berechnungsmethode: Bikubische Neuberechnung Downsample-Auflösung: 150 dpi Downsampling für Bilder über: 225 dpi Komprimieren: Ja Automatische Bestimmung der Komprimierungsart: Ja JPEG-Qualität: Mittel Bitanzahl pro Pixel: Wie Original Bit Schwarzweiß-Bilder: Downsampling: Ja Berechnungsmethode: Bikubische Neuberechnung Downsample-Auflösung: 600 dpi Downsampling für Bilder über: 900 dpi Komprimieren: Ja Komprimierungsart: CCITT CCITT-Gruppe: 4 Graustufen glätten: Nein Text und Vektorgrafiken komprimieren: Ja SCHRIFTEN ---------------------------------------- Alle Schriften einbetten: Ja Untergruppen aller eingebetteten Schriften: Nein Wenn Einbetten fehlschlägt: Warnen und weiter Einbetten: Immer einbetten: [ ] Nie einbetten: [ ] FARBE(N) ---------------------------------------- Farbmanagement: Farbumrechnungsmethode: Alle Farben zu sRGB konvertieren Methode: Standard Arbeitsbereiche: Graustufen ICC-Profil: RGB ICC-Profil: sRGB IEC61966-2.1 CMYK ICC-Profil: U.S. Web Coated (SWOP) v2 Geräteabhängige Daten: Einstellungen für Überdrucken beibehalten: Ja Unterfarbreduktion und Schwarzaufbau beibehalten: Ja Transferfunktionen: Anwenden Rastereinstellungen beibehalten: Ja ERWEITERT ---------------------------------------- Optionen: Prolog/Epilog verwenden: Nein PostScript-Datei darf Einstellungen überschreiben: Ja Level 2 copypage-Semantik beibehalten: Ja Portable Job Ticket in PDF-Datei speichern: Nein Illustrator-Überdruckmodus: Ja Farbverläufe zu weichen Nuancen konvertieren: Nein ASCII-Format: Nein Document Structuring Conventions (DSC): DSC-Kommentare verarbeiten: Nein ANDERE ---------------------------------------- Distiller-Kern Version: 5000 ZIP-Komprimierung verwenden: Ja Optimierungen deaktivieren: Nein Bildspeicher: 524288 Byte Farbbilder glätten: Nein Graustufenbilder glätten: Nein Bilder (< 257 Farben) in indizierten Farbraum konvertieren: Ja sRGB ICC-Profil: sRGB IEC61966-2.1 ENDE DES REPORTS ---------------------------------------- IMPRESSED GmbH Bahrenfelder Chaussee 49 22761 Hamburg, Germany Tel. +49 40 897189-0 Fax +49 40 897189-71 Email: [email protected] Web: www.impressed.de
Adobe Acrobat Distiller 5.0.x Joboption Datei
<< /ColorSettingsFile () /AntiAliasMonoImages false /CannotEmbedFontPolicy /Warning /ParseDSCComments false /DoThumbnails true /CompressPages true /CalRGBProfile (sRGB IEC61966-2.1) /MaxSubsetPct 100 /EncodeColorImages true /GrayImageFilter /DCTEncode /Optimize true /ParseDSCCommentsForDocInfo false /EmitDSCWarnings false /CalGrayProfile () /NeverEmbed [ ] /GrayImageDownsampleThreshold 1.5 /UsePrologue false /GrayImageDict << /QFactor 0.9 /Blend 1 /HSamples [ 2 1 1 2 ] /VSamples [ 2 1 1 2 ] >> /AutoFilterColorImages true /sRGBProfile (sRGB IEC61966-2.1) /ColorImageDepth -1 /PreserveOverprintSettings true /AutoRotatePages /None /UCRandBGInfo /Preserve /EmbedAllFonts true /CompatibilityLevel 1.2 /StartPage 1 /AntiAliasColorImages false /CreateJobTicket false /ConvertImagesToIndexed true /ColorImageDownsampleType /Bicubic /ColorImageDownsampleThreshold 1.5 /MonoImageDownsampleType /Bicubic /DetectBlends false /GrayImageDownsampleType /Bicubic /PreserveEPSInfo false /GrayACSImageDict << /VSamples [ 2 1 1 2 ] /QFactor 0.76 /Blend 1 /HSamples [ 2 1 1 2 ] /ColorTransform 1 >> /ColorACSImageDict << /VSamples [ 2 1 1 2 ] /QFactor 0.76 /Blend 1 /HSamples [ 2 1 1 2 ] /ColorTransform 1 >> /PreserveCopyPage true /EncodeMonoImages true /ColorConversionStrategy /sRGB /PreserveOPIComments false /AntiAliasGrayImages false /GrayImageDepth -1 /ColorImageResolution 150 /EndPage -1 /AutoPositionEPSFiles false /MonoImageDepth -1 /TransferFunctionInfo /Apply /EncodeGrayImages true /DownsampleGrayImages true /DownsampleMonoImages true /DownsampleColorImages true /MonoImageDownsampleThreshold 1.5 /MonoImageDict << /K -1 >> /Binding /Left /CalCMYKProfile (U.S. Web Coated (SWOP) v2) /MonoImageResolution 600 /AutoFilterGrayImages true /AlwaysEmbed [ ] /ImageMemory 524288 /SubsetFonts false /DefaultRenderingIntent /Default /OPM 1 /MonoImageFilter /CCITTFaxEncode /GrayImageResolution 150 /ColorImageFilter /DCTEncode /PreserveHalftoneInfo true /ColorImageDict << /QFactor 0.9 /Blend 1 /HSamples [ 2 1 1 2 ] /VSamples [ 2 1 1 2 ] >> /ASCII85EncodePages false /LockDistillerParams false >> setdistillerparams << /PageSize [ 595.276 841.890 ] /HWResolution [ 600 600 ] >> setpagedevice
353
travascular volume of the critically ill patients can be de-fined by the relationship between preload and cardiacoutput, and that changing preload with volume infusionaffects cardiac output. Thus an increase in cardiac outputfollowing VE (patient responder) unmasks an hypovo-lemic state or preload dependency. On the other hand,lack of change or a decrease in cardiac output followingVE (nonresponding patient) is attributed to a normovo-lemic, to an overloaded, or to cardiac failure state.Therefore, as the fluid responsiveness defines the re-sponse of cardiac output to volume challenge, indiceswhich can predict the latter are necessary.
Static measurements for preload assessment
Measures of intracardiac pressures
According to the Frank-Starling law, left-ventricular pre-load is defined as the myocardial fiber length at the end
Table 1 Studies of indices used as bedside indicators of preloadreserve and fluid responsiveness in hypotensive patients underpositive-pressure ventilation (BMI body mass index, CO cardiacoutput, CI cardiac index, SV stroke volume, SVI stroke volume in-dex, IAC invasive arterial catheter, MV proportion of patients me-chanically ventilated, ↑ increase, ↓ decrease, PAC pulmonary ar-tery catheter, R responders, NR nonresponders, FC fluid challenge,
HES hydroxyethyl starch, RL Ringer’s lactate, Alb albumin,∆down delta down, ∆PP respiratory variation in pulse pressure,LVEDV left-ventricular end diastolic volume, SPV systolic pres-sure variation, SVV stroke volume variation, TEE transesophagealechocardiography, Ppao pulmonary artery occlusion pressure,RAP right atrial pressure, RVEDV right-ventricular-end diastolicvolume, FC fluid challenge)
Variable Tech- n MV Volume (ml) Duration Definition Definition p: Refer-measured nique (%) and type of of FC of R of NR difference ence
plasma substitute (min) in baseline values R vs. NR
Rap PAC 28 46 250 Alb 5% 20–30 ↑ SVI ↓ SVI or unchanged NS 37Rap PAC 41 76 300 Alb 4.5% 30 ↑ CI CI ↓ or unchanged NS 18Rap PAC 25 94.4 NaCl 9‰ + Until ↑Ppao ↑ SV ≥10% ↑ SV <10% 0.04 31
Alb 5% to ↑ PpaoRap PAC 40 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% NS 36Ppao PAC 28 46 250 Alb 5% 20–30 ↑ SVI ↓ SVI or unchanged NS 37Ppao PAC 41 76 300 Alb 4.5% 30 ↑ CI CI ↓ or unchanged NS 18Ppao PAC 29 69 300–500 RL ? bolus ↑ C0>10% C0 ↓ or unchanged <0.01 40Ppao PAC 32 84 300–500 RL ? ↑ CI >20% ↑ CI <20% NS 41Ppao PAC 16 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% 0.1 42Ppao PAC 41 100 500 pPentastarch 15 ↑ SV ≥20% ↑ SV <20% 0.003 25Ppao PAC 25 94.4 NaCl 9‰, Until ↑Ppao ↑ SV ≥10% ↑ SV <10% 0.001 31
Alb 5% to↑ PpaoPpao PAC 40 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% NS 36Ppao PAC 19 100 500–750 HES 6% 10 ↑ C0>10% ↑ SV <10% 0.0085 39RVEDV PAC 29 69 300–500 RL ? bolus ↑ C0>10% C0 ↓ or unchanged <0.001 40RVEDV PAC 32 84 300–500 RL ? ↑ CI >20% ↑ CI <20% <0.002 41RVEDV PAC 25 94.4 NaCl 9‰, Until ↑Ppao ↑ SV ≥10% ↑ SV <10% 0.22 31
Alb 5% to↑ PpaoLVEDV TEE 16 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% 0.005 42LVEDV TEE 41 100 500 Pentastarch 15 ↑ SV ≥20% ↑ SV <20% 0.012 25LVEDV TEE 19 100 8 ml/kg HES 6% 30 ↑ CI >15% ↑ CI <15% NS 79LVEDV TEE 19 100 500–750 HES 6% 10 ↑ C0>10% ↑ SV <10% NS 39SPV IAC 16 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% 0.0001 42SPV IAC 40 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% <0.001 36SPV IAC 19 100 500–750 HES 6% 10 ↑ C0>10% ↑ SV <10% 0.017 39∆down IAC 16 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% 0.0001 42∆down IAC 19 100 500–750 HES 6% 10 ↑ C0>10% ↑ SV <10% 0.025 39∆PP IAC 40 100 500 HES 6% 30 ↑ CI >15% ↑ CI <15% <0.001 36
Fig. 1 Representation of Frank-Starling curve with relationship be-tween ventricular preload and ventricular stroke volume in patientX. After volume expansion the same magnitude of change in pre-load recruit less stroke volume, because the plateau of the curve isreached which characterize a condition of preload independency
49
354
of the diastole. In clinical practice, the left-ventricularend-diastolic volume is used as a surrogate to define left-ventricular preload [4]. However, this volumetric param-eter is not easily assessed in critically ill patients. In nor-mal conditions, a fairly good correlation exists betweenventricular end-diastolic volumes and mean atrial pres-sures, and ventricular preloads are approximated by RAPand/or Ppao in patients breathing spontaneously [5, 6].Critically ill patients often require positive pressure ven-tilation, which modifies the pressure regimen in the tho-rax in comparison to spontaneous breathing. Indeed, dur-ing MV RAP and Ppao rise secondary to an increase inintrathoracic pressure which rises pericardial pressure.This pressure increase induces a decrease in venous re-turn [7, 8] with first a decrease in right and few heartbeats later in left-ventricular end-diastolic volumes, re-spectively [9, 10]. Under extreme conditions such asacute severe pulmonary emboli and/or marked hyperin-flation, RAP may also rise secondary to an increase af-terload of the right ventricle. Moreover, under positivepressure ventilation not only ventricular but also tho-racopulmonary compliances and abdominal pressurevariations are observed over time. Thus a variable rela-tionship between cardiac pressures and cardiac volumesis often observed [11, 12, 13, 14]. It has also been dem-onstrated that changes in intracardiac pressure (RAP,Ppao) no longer directly reflect changes in intravascularvolume [15]. Pinsky et al. [16, 17] have demonstratedthat changes in RAP do not follow changes in right-ven-tricular end-diastolic volume in postoperative cardiacsurgery patients under positive pressure ventilation. Reuse et al. [18] observed no correlation between RAPand right-ventricular end-diastolic volume calculatedfrom a thermodilution technique in hypovolemic patientsbefore and after fluid resuscitation. The discordance be-tween RAP and right-ventricular end-diastolic volumemeasurements may result from a systematic underesti-mation of the effect of positive-pressure ventilation onthe right heart [16, 17]. Nevertheless, the RAP valuemeasured either with a central venous catheter or a pul-monary artery catheter is still used to estimate preloadand to guide intravascular volume therapy in patient under positive pressure ventilation [19, 20].
On the left side, the MV-induced intrathoracic pres-sure changes, compared to spontaneously breathing, on-ly minimally alters the relationship between left atrialpressure and left-ventricular end-diastolic volume mea-surement in postoperative cardiac surgery patients [21].However, several other studies show no relationship be-tween Ppao and left-ventricular end-diastolic volumemeasured by either radionuclide angiography [12, 22],transthoracic echocardiography (TTE) [23], or trans-esophageal echocardiography (TEE) [24, 25, 26]. Thelatter findings may be related to the indirect pulmonaryartery catheter method for assessing left atrial pressure[27, 28], although several studies have demonstrated
that Ppao using PAC is a reliable indirect measurementof left atrial pressure [29, 30] in positive-pressure MVpatients.
Right atrial pressure used to predict fluid responsiveness
Wagner et al. [31] reported that RAP was significantlylower before volume challenge in responders than innonresponders (p=0.04) when patients were under posi-tive pressure ventilation. Jellinek et al. [32] found that aRAP lower than 10 mmHg predicts a decrease in cardiacindex higher than 20% when a transient 30 cm H2O in-crease in intrathoracic pressure is administrated. Presum-ing that the principle cause of decrease in cardiac outputin the latter study was due to a reduction in venous re-turn [9, 33, 34, 35], RAP predicts reverse VE hemody-namic effect. Nevertheless, some clinical investigationsstudying fluid responsiveness in MV patients have re-ported that RAP poorly predicts increased cardiac outputafter volume expansion [18, 36, 37]. Indeed, in thesestudies RAP did not differentiate patients whose cardiacoutput did or did not increase after VE (responders andnonresponders, respectively).
Ppao used to predict fluid responsiveness
Some studies have demonstrated that Ppao is a good pre-dictor of fluid responsiveness [13, 31, 38]. RecentlyBennett-Guerrero et al. [39] also found that Ppao was abetter predictor of response to VE than systolic pressurevariation (SPV) and left-ventricular end-diastolic areameasured by TEE. However, several other studies notedthat Ppao is unable to predict fluid responsiveness and todifferentiate between VE-responders and VE-nonre-sponders [18, 25, 36, 37, 40, 41, 42]. The discrepancybetween the results of these studies may partly reflectdifferences in patients’ baseline characteristics (e.g., de-mographic differences, medical history, chest and lungcompliances) at study entry. Furthermore, differences inlocation of the pulmonary artery catheter extremity rela-tive to the left atrium may be present [43]. Indeed, ac-cording to its position, pulmonary artery catheter maydisplay alveolar pressure instead of left atrial pressure(West zone I or II) [44]. The value of Ppao is also influ-enced by juxtacardiac pressure [45, 46] particularly ifpositive end-expiratory pressure (PEEP) is used [28]. To overcome the latter difficulty in MV patients whenPEEP is used, nadir Ppao (Ppao measured after airwaydisconnection) may be used [46]. However, as nadirPpao requires temporary disconnection from the ventila-tor, it might be deleterious to severely hypoxemic pa-tients. No study has yet evaluated the predictive value of nadir Ppao for estimating fluid responsiveness in MVpatients.
50
355
In brief, although static intracardiac pressure mea-surements such as RAP and Ppao have been studied andused for many years for hemodynamic monitoring, theirlow predictive value in estimating fluid responsivenessin MV patients must be underlined. Thus using only in-travascular static pressures to guide fluid therapy canlead to inappropriate therapeutic decisions [47].
Measures of ventricular end-diastolic volumes
Radionuclide angiography [48], cineangiocardiography[49], and thermodilution [50] have been used to estimateventricular volumes for one-half a century. In intensive careunits, various methods have been used to measure ventricu-lar end-diastolic volume at the bedside, such as radionu-clide angiography [51, 52], TTE [23, 53, 54], TEE [55],and a modified flow-directed pulmonary artery catheterwhich allows the measurement of cardiac output and right-ventricular ejection fraction (and the calculation of right-ventricular end-systolic and end-diastolic volume) [31, 41].
Right-ventricular end-diastolic volume measured by pulmonary artery catheter used to predict fluid responsiveness
During MV right-ventricular end-diastolic volume mea-sured with a pulmonary artery catheter is decreased byPEEP [56] but is still well correlated with cardiac index[57, 58] and is a more reliable predictor of fluid respon-siveness than Ppao [40, 41]. On the other hand, otherstudies have found no relationship between change inright-ventricular end-diastolic volume measured by pul-monary artery catheter and change in stroke volume intwo series of cardiac surgery patients [16, 18]. Similarly,Wagner et al. [31] found that right-ventricular end-diastol-ic volume measured by pulmonary artery catheter was nota reliable predictor of fluid responsiveness in patients un-der MV, and that Ppao and RAP were superior to right-ventricular end-diastolic volume. The discrepancy be-tween the results of these studies may partly reflect themeasurement errors of cardiac output due to the cyclicchange induced by positive pressure ventilation [59, 60,61, 62], the inaccuracy of cardiac output measurement ob-tained by pulmonary artery catheter when the flux is low[63], and the influence of tricuspid regurgitation on themeasurement of cardiac output [64]. Moreover, as right-ventricular end-diastolic volume is calculated (stroke vol-ume divided by right ejection fraction), cardiac output be-comes a shared variable in the calculation of both strokevolume and right-ventricular end-diastolic volume, and amathematical coupling may have contributed to the closecorrelation observed between these two variables. Never-theless, right-ventricular end-diastolic volume comparedto Ppao may be useful in a small group of patients with
high intra-abdominal pressure or when clinicians are re-luctant to obtain off-PEEP nadir Ppao measurements [65].
Right-ventricular end-diastolic volume measured by echocardiography used to predict fluid responsiveness
TTE has been shown to be a reliable method to assessright-ventricular dimensions in patients ventilated withcontinuous positive airway pressure or positive-pressureventilation [66, 67]. Using this approach, right-ventricu-lar end-diastolic area is obtained on the apical fourchambers view [68]. When no right-ventricular windowis available, TEE is preferred to monitor right-ventricu-lar end-volume in MV patients [53, 55, 69, 70, 71]. Thelatter method has become more popular in recent yearsdue to technical improvements [72]. Nevertheless, nostudy has evaluated right-ventricular size measurementsby TTE or TEE as a predictor of fluid responsiveness inMV patients.
Left-ventricular end-diastolic volume measured by echocardiography used to predict fluid responsiveness
TTE has been used in the past to measure left-ventricularend-diastolic volume and/or area [23, 67, 73, 74] in MVpatients. However, no study has evaluated the left-ven-tricular end-diastolic volume and/or area measured byTTE as predictors of fluid responsiveness in MV pa-tients. Due to its greater resolving power, TEE easily andaccurately assesses left-ventricular end-diastolic volumeand/or area in clinical practice [53, 75] except in patientsundergoing coronary artery bypass grafting [76]. Howev-er, different studies have reported conflicting resultsabout the usefulness of left-ventricular end-diastolic vol-ume and/or area measured by TEE to predict fluid re-sponsiveness in MV patients. Cheung et al. [26] haveshown that left-ventricular end-diastolic area measuredby TEE is an accurate method to predict the hemody-namic effects of acute blood loss. Other studies have re-ported either a modest [25, 42, 77] or a poor [78, 79]predictive value of left-ventricular end-diastolic volumeand area to predict the cardiac output response to fluidloading. Recent studies have also produced conflictingresults. Bennett-Guerrero et al. [39] measuring left-ven-tricular end-diastolic area with TEE before VE found nosignificant difference between responders and nonre-sponders. Paradoxically, Reuter et al. [80] found thatleft-ventricular end-diastolic area index assessed by TEEbefore VE predicts fluid responsiveness more accuratelythan RAP, Ppao, and stroke volume variation (SVV). Inthe future three-dimensional echocardiography couldsupplant other methods for measuring left-ventricularend-diastolic volume and their predictive value of fluidresponsiveness. In a word, although measurements of
51
356
ventricular volumes should theoretically reflect preloaddependence more accurately than other indices, conflict-ing results have been reported so far. These negativefindings may be related to the method used to estimateend-diastolic ventricular volumes which do not reflectthe geometric complexity of the right ventricle and to theinfluences of the positive intrathoracic pressure on left-ventricular preload, afterload and myocardial contractili-ty [81].
Dynamic measurements for preload assessment
Measure of respiratory changes in systolic pressure, pulse pressure, and stroke volume
Positive pressure breath decreases temporary right-ven-tricular end-diastolic volume secondary to a reduction invenous return [7, 82]. A decrease in right-ventricularstroke volume ensues which become minimal at end pos-itive pressure breath. This inspiratory reduction in right-ventricular stroke volume induces a decrease in left-ven-tricular end-diastolic volume after a phase lag of fewheart beats (due to the pulmonary vascular transit time[83]), which becomes evident during the expiratoryphase. This expiratory reduction in left-ventricular end-diastolic volume induces a decrease in left-ventricularstroke volume, determining the minimal value of systolicblood pressure observed during expiration. Conversely,the inspiratory increase in left-ventricular end-diastolicvolume determining the maximal value of systolic bloodpressure is observed secondary to the rise in left-ventric-ular preload reflecting the few heart beats earlier in-creased in right-ventricular preload during expiration.Furthermore, increasing lung volume during positivepressure ventilation may also contribute to the increasedpulmonary venous blood flow (related to the compres-sion of pulmonary blood vessels [84]) and/or to a de-crease in left-ventricular afterload [85, 86, 87], which together induce an increase in left-ventricular preload.Finally, a decrease in right-ventricular end-diastolic vol-ume during a positive pressure breath may increase left-ventricular compliance and then left-ventricular preload[88]. Thus due to heart-lung interaction during positivepressure ventilation the left-ventricular stroke volumevaries cyclically (maximal during inspiration and mini-mal during expiration).
These variations have been used clinically to assesspreload status and predict fluid responsiveness in deeplysedated patients under positive pressure ventilation. In1983 Coyle et al. [89] in a preliminary study demonstrat-ed that the SPV following one mechanical breath is in-creased in hypovolemic sedated patients and decreasedafter fluid resuscitation. This study defined SPV as thedifference between maximal and minimal values of sys-tolic blood pressure during one positive pressure me-
chanical breath. Using the systolic pressure at end expi-ration as a reference point or baseline the SPV was fur-ther divided into two components: an increase (∆up) anda decrease (∆down) in systolic pressure vs. baseline(Fig. 2). These authors concluded that in hypovolemicpatients ∆down was the main component of SPV. Thesepreliminary conclusions were confirmed in 1987 byPerel et al. [90] who demonstrated that SPV following apositive pressure breath is a sensitive indicator of hypo-volemia in ventilated dogs. Thereafter Coriat et al. [91]demonstrated that SPV and ∆down predict the responseof cardiac index to VE in a group of sedated MV patientsafter vascular surgery. Exploring another pathophysio-logical concept, Jardin et al. [92] found that pulse pres-sure (PP; defined as the difference between the systolicand the diastolic pressure) is related to left-ventricularstroke volume in MV patients. Using these findings, Michard et al. [35, 36,] have shown that respiratorychanges in PP [∆PP=maximal PP at inspiration (PPmax)minus minimal PP at expiration (PPmin); (Fig. 2) andcalculated as: ∆PP (%)=100 (PPmax-PPmin)/(Ppmax+PPmin)/2] predict the effect of VE on cardiac index inpatients with acute lung injury [35] or septic shock [36].The same team proposed another approach to assessSVV in MV patients and to predict cardiac responsive-ness to VE [79]. Using Doppler measurement of beat-to-beat aortic blood flow, they found that respiratorychange in aortic blood flow maximal velocity predictsfluid responsiveness in septic MV patients. MeasuringSVV during positive pressure ventilation by continuousarterial pulse contour analysis, Reuter et al. [80] have re-cently demonstrated that SVV accurately predicts fluidresponsiveness following volume infusion in ventilatedpatients after cardiac surgery.
Fig. 2 Systolic pressure variation (SPV) after one mechanicalbreath followed by an end-expiratory pause. Reference line per-mits the measurement of ∆up and ∆down. Bold Maximal and min-imal pulse pressure. AP Airway pressure; SAP systolic arterialpressure
52
357
Systolic pressure variation used to predict fluid responsiveness
The evaluation of fluid responsiveness by SPV is basedon cardiopulmonary interaction during MV [93, 94]. In1995 Rooke et al. [95] found that SPV is a useful moni-tor of volume status in healthy MV patients during anes-thesia. Coriat et al. [91] confirmed the usefulness of SPVfor estimating response to VE in MV patients after vas-cular surgery. Ornstein et al. [96] have also shown thatSPV and ∆down are correlated with decreased cardiacoutput after controlled hemorrhage in postoperative car-diac surgical patients. Furthermore, Tavernier et al. [42]found ∆down before VE to be an accurate index of thefluid responsiveness in septic patients, and that a ∆downvalue of 5 mmHg is the cutoff point for distinguishingresponders from nonresponders to VE. Finally, in septicpatients Michard et al. [36] found that SPV is correlatedwith volume expansion-induced change in cardiac out-put. However, Denault et al. [81] have demonstrated thatSPV is not correlated with changes in left-ventricularend-diastolic volume measured by TEE in cardiac sur-gery patients. Indeed, in this study, SPV was observeddespite no variation in left-ventricular stroke volume,suggesting that SPV involves processes independent ofchanges in the left-ventricular preload (airway pressure,pleural pressure, and its resultant afterload) [81].
Pulse pressure variation used to predict fluid responsiveness
Extending the concept elaborated by Jardin et al. [92],Michard et al. [36] found that ∆PP predicted the effect ofVE on cardiac output in 40 septic shock hypotensive pa-tients. These authors demonstrated that both ∆PP andSPV, when greater than 15%, are superior to RAP andPpao, for predicting fluid responsiveness. Moreover,∆PP was more accurate and with less bias than SPV.These authors proposed ∆PP as a surrogate for strokevolume variation concept [93], which has not been vali-dated yet. In another study these authors [35] includedVE in six MV patients with acute lung injury and foundthat ∆PP is a useful guide to predict fluid responsive-ness.
Stroke volume variation to predict fluid responsiveness
Using Doppler TEE, Feissel et al. [79] studied changesin left-ventricular stroke volume induced by the cyclicpositive pressure breathing. By measuring the respiratoryvariation in maximal aortic blood flow velocity these au-thors predicted fluid responsiveness in septic MV pa-tients. Left-ventricular stroke volume was obtained bymultiplying flow velocity time integral over aortic valve
by valve opening area during expiration. However, thisfinding may be biased, as expiratory flow velocity timeintegral is a shared variable in the calculation of bothcardiac output and expiratory maximal aortic blood flowvelocity and a mathematical coupling may contribute tothe observed correlation between changes in cardiac out-put and variation in maximal aortic blood flow velocity.Finally, Reuter et al. [80] used continuous arterial pulsecontour analysis and found that SVV during positivepressure breath accurately predicts fluid responsivenessfollowing VE in ventilated cardiac surgery patients [80].Using the receiver operating characteristics curve, theseauthors demonstrated that the area under the curve is sta-tistically greater for SVV (0.82; confidence interval:0.64–1) and SPV (0.81; confidence interval: 0.62–1)than for RAP (0.45; confidence interval: 0.17–0.74)(p<0.001) [97]. Concisely, dynamic indices have beenexplored to evaluate fluid responsiveness in critically illpatients. All of them have been validated in deeply se-dated patients under positive-pressure MV. Thus such in-dices are useless in spontaneously breathing intubatedpatients, a MV mode often used in ICU. Moreover, regu-lar cardiac rhythm is an obligatory condition to allowtheir use.
Conclusion
Positive pressure ventilation cyclically increases intra-thoracic pressure and lung volume, both of which de-crease venous return and alter stroke volume. Thus VEwhich rapidly restore cardiac output and arterial bloodpressure is an often used therapy in hypotensive MV pa-tients and indices which would predict fluid responsive-ness are necessary. RAP, Ppao, and right-ventricular end-diastolic volume, which are static measurements, havebeen studied but produced conflicting data in estimatingpreload and fluid responsiveness. On the other hand,SPV and ∆PP, which are dynamic measurements, havebeen shown to identify hypotension related to decreasein preload, to distinguish between responders and nonre-sponders to fluid challenge (Table 1), and to permit titra-tion of VE in various patient populations.
Although there is substantial literature on indices ofhypovolemia, only few studies have evaluated the cardi-ac output changes induced by VE in MV patients. More-over, therapeutic recommendations regarding unmaskedpreload dependency states without hypotension need fur-ther studies. Finally, another unanswered question is re-lated to patients outcome: does therapy guided by fluidresponsiveness indices improve patients survival?
Acknowledgements The authors thank Dr. M.R. Pinsky, Univer-sity of Pittsburgh Medical Center, for his helpful advice in thepreparation of this manuscript. The authors are also grateful forthe translation support provided by Angela Frei.
53
358
References
1. Guyton AC, Richardson TQ, LangstonJB (1964) Regulation of cardiac outputand venous return. Clin Anesth 3:1–34
2. Guyton AC (1967) Regulation of cardi-ac output. N Engl J Med 277:805–812
3. Horst HM, Obeid FN (1986) Hemody-namic response to fluid challenge: a means of assessing volume status inthe critically ill. Henry Ford Hosp MedJ 34:90–94
4. Suga H, Sagawa K (1974) Instanta-neous pressure-volume relationshipsand their ratio in the excised, supportedcanine left ventricle. Circ Res35:117–126
5. Crexells C, Chatterjee K, Forrester JS,Dikshit K, Swan HJ (1973) Optimallevel of filling pressure in the left sideof the heart in acute myocardial infarc-tion. N Engl J Med 289:1263–1266
6. Buchbinder N, Ganz W (1976) Hemo-dynamic monitoring: invasive tech-niques. Anesthesiology 45:146–155
7. Cournand A, Motley H, Werko L,Richards D (1948) Physiological stud-ies of the effect of intermittent positivepressure breathing on cardiac output inman. Am J Physiol 152:162–174
8. Fessler HE, Brower RG, Wise RA, Permutt S (1992) Effects of positiveend-expiratory pressure on the caninevenous return curve. Am Rev RespirDis 146:4–10
9. Potkin RT, Hudson LD, Weaver LJ,Trobaugh G (1987) Effect of positiveend-expiratory pressure on right andleft ventricular function in patientswith the adult respiratory distress syndrome. Am Rev Respir Dis135:307–311
10. Fewell JE, Abendschein DR, CarlsonCJ, Murray JF, Rapaport E (1980) Con-tinuous positive-pressure ventilationdecreases right and left ventricular end-diastolic volumes in the dog. Circ Res46:125–132
11. McKenney PA, Apstein CS, MendesLA, Connelly GP, Aldea GS, SheminRJ, Davidoff R (1994) Increased leftventricular diastolic chamber stiffnessimmediately after coronary artery by-pass surgery. J Am Coll Cardiol24:1189–1194
12. Calvin JE, Driedger AA, Sibbald WJ(1981) Does the pulmonary capillarywedge pressure predict left ventricularpreload in critically ill patients? Crit Care Med 9:437–443
13. Packman MI, Rackow EC (1983) Optimum left heart filling pressure during fluid resuscitation of patientswith hypovolemic and septic shock.Crit Care Med 11:165–169
14. Jardin F, Gueret P, Dubourg O, FarcotJC, Margairaz A, Bourdarias JP (1985)Right ventricular volumes by thermo-dilution in the adult respiratory distresssyndrome. A comparative study usingtwo-dimensional echocardiography asa reference method. Chest 88:34–39
15. Lichtwarck-Aschoff M, Zeravik J,Pfeiffer UJ (1992) Intrathoracic bloodvolume accurately reflects circulatoryvolume status in critically ill patientswith mechanical ventilation. IntensiveCare Med 18:142–147
16. Pinsky MR, Desmet JM, Vincent JL(1992) Effect of positive end-expirato-ry pressure on right ventricular func-tion in humans. Am Rev Respir Dis146:681–687
17. Pinsky MR (1994) Cardiovascular effects of ventilatory support and with-drawal. Anesth Analg 79:567–576
18. Reuse C, Vincent JL, Pinsky MR(1990) Measurements of right ventricu-lar volumes during fluid challenge.Chest 98:1450–1454
19. Saarela E, Kari A, Nikki P, Rauhala V,Iisalo E, Kaukinen L (1991) Currentpractice regarding invasive monitoringin intensive care units in Finland. A nationwide study of the uses of arte-rial, pulmonary artery and central venous catheters and their effect onoutcome. The Finnish Intensive CareStudy Group. Intensive Care Med17:264–271
20. Boldt J, Lenz M, Kumle B, Papsdorf M(1998) Volume replacement strategieson intensive care units: results from apostal survey. Intensive Care Med24:147–151
21. Guyton RA, Chiavarelli M, PadgettCA, Cheung EH, Staton GW, HatcherCR (1987) The influence of positiveend-expiratory pressure on intraperi-cardial pressure and cardiac functionafter coronary artery bypass surgery. J Cardiothorac Anesth 1:98–107
22. Lemaire F, Teboul JL, Cinotti L, GiottoG, Abrouk F, Steg G, Macquin-MavierI, Zapol WM (1988) Acute left ventric-ular dysfunction during unsuccessfulweaning from mechanical ventilation.Anesthesiology 69:171–179
23. Jardin F, Valtier B, Beauchet A, Dubourg O, Bourdarias JP (1994) Inva-sive monitoring combined with two-dimensional echocardiographic studyin septic shock. Intensive Care Med20:550–554
24. Hinder F, Poelaert JI, Schmidt C, HoeftA, Mollhoff T, Loick HM, Van Aken H(1998) Assessment of cardiovascularvolume status by transoesophagealechocardiography and dye dilution dur-ing cardiac surgery. Eur J Anaesthesiol15:633–640
25. Tousignant CP, Walsh F, Mazer CD(2000) The use of transoesophagealechocardiography for preload assess-ment in critically ill patients. AnesthAnalg 90:351–355
26. Cheung AT, Savino JS, Weiss SJ, Aukburg SJ, Berlin JA (1994) Echocar-diographic and hemodynamic indexesof left ventricular preload in patientswith normal and abnormal ventricularfunction. Anesthesiology 81:376–387
27. Swan HJ, Ganz W, Forrester J, MarcusH, Diamond G, Chonette D (1970)Catheterization of the heart in manwith use of a flow-directed balloon-tipped catheter. N Engl J Med283:447–451
28. Berryhill RE, Benumof JL (1979)PEEP-induced discrepancy betweenpulmonary arterial wedge pressure andleft atrial pressure: the effects of con-trolled vs. spontaneous ventilation andcompliant vs. noncompliant lungs inthe dog. Anesthesiology 51:303–308
29. Humphrey CB, Gibbons JA, FolkerthTL, Shapiro AR, Fosburg RG (1976)An analysis of direct and indirect mea-surements of left atrial filling pressure.J Thorac Cardiovasc Surg 71:643–647
30. Lozman J, Powers SR Jr, Older T, Dutton RE, Roy RJ, English M, Marco D, Eckert C (1974) Correlationof pulmonary wedge and left atrialpressures. A study in the patient receiv-ing positive end expiratory pressureventilation. Arch Surg 109:270–277
31. Wagner JG, Leatherman JW (1998)Right ventricular end-diastolic volumeas a predictor of the hemodynamic response to a fluid challenge. Chest113:1048–1054
32. Jellinek H, Krafft P, Fitzgerald RD,Schwarz S, Pinsky MR (2000) Rightatrial pressure predicts hemodynamicresponse to apneic positive airwaypressure. Crit Care Med 28:672–688
33. Van Trigt P, Spray TL, Pasque MK,Peyton RB, Pellom GL, Christian CM,Fagraeus L, Wechsler AS (1982) Theeffect of PEEP on left ventricular dia-stolic dimensions and systolic perfor-mance following myocardial revascu-larization. Ann Thorac Surg33:585–5892
34. Johnston WE, Vinten-Johansen J, Santamore WP, Case LD, Little WC(1989) Mechanism of reduced cardiacoutput during positive end-expiratorypressure in the dog. Am Rev RespirDis 140:1257–1264
54
359
35. Michard F, Chemla D, Richard C, Wysocki M, Pinsky MR, LecarpentierY, Teboul JL (1999) Clinical use of res-piratory changes in arterial pulse pres-sure to monitor the hemodynamic ef-fects of PEEP. Am J Respir Crit CareMed 159:935–939
36. Michard F, Boussat S, Chemla D, Anguel N, Mercat A, Lecarpentier Y,Richard C, Pinsky, MR, Teboul JL(2000) Relation between respiratorychanges in arterial pulse pressure and fluid responsiveness in septic pa-tients with acute circulatory failure.Am J Respir Crit Care Med162:134–138
37. Calvin JE, Driedger AA, Sibbald WJ(1981) The hemodynamic effect of rapid fluid infusion in critically ill patients. Surgery 90:61–76
38. Krausz MM, Perel A, Eimerl D, CotevS (1977) Cardiopulmonary effects ofvolume loading in patients in septicshock. Ann Surg 185:429–434
39. Bennett-Guerrero E, Kahn RA, Moskowitz DM, Falcucci O, BodianCA (2002) Comparison of arterial systolic pressure variation with otherclinical parameters to predict the response to fluid challenges during car-diac surgery. Mt Sinai J Med69:96–100
40. Diebel LN, Wilson RF, Tagett MG,Kline RA (1992) End-diastolic volume.A better indicator of preload in the crit-ically ill. Arch Surg 127:817–821
41. Diebel L, Wilson RF, Heins J, Larky H,Warsow K, Wilson S (1994) End-dia-stolic volume versus pulmonary arterywedge pressure in evaluating cardiacpreload in trauma patients. J Trauma37:950–955
42. Tavernier B, Makhotine O, Lebuffe G,Dupont J, Scherpereel P (1998) Systol-ic pressure variation as a guide to fluidtherapy in patients with sepsis-inducedhypotension. Anesthesiology89:1313–1321
43. Rajacich N, Burchard KW, Hasan FM,Singh AK (1989) Central venous pres-sure and pulmonary capillary wedgepressure as estimates of left atrial pres-sure: effects of positive end-expiratorypressure and catheter tip malposition.Crit Care Med 17:7–11
44. Shasby DM, Dauber IM, Pfister S(1981) Swan-Ganz catheter locationand left atrial pressure determine theaccuracy of the wedge pressure whenpositive end-expiratory pressure isused. Chest 80:666–670
45. Raper R, Sibbald WJ (1986) Misled bythe wedge? The Swan-Ganz catheterand left ventricular preload. Chest89:427–434
46. Pinsky M, Vincent JL, De Smet JM(1991) Estimating left ventricular fill-ing pressure during positive end-expi-ratory pressure in humans. Am RevRespir Dis 143:25–31
47. Swenson JD, Bull D, Stringham J(2001) Subjective assessment of leftventricular preload using transesopha-geal echocardiography: correspondingpulmonary artery occlusion pressures. J Cardiothorac Vasc Anesth 15:580–583
48. Mullins CB, Mason DT, Ashburn WL,Ross J (1969) Determination of ven-tricular volume by radioisotope-angi-ography. Am J Cardiol 24:72–78
49. Kasser IS, Kennedy JW (1969) Measurement of left ventricular vol-umes in man by single-plane cinean-giocardiography. Invest Radiol 4:83–90
50. Balcon R, Oram S (1968) Measurementof right ventricular end-systolic andend-diastolic volumes by the thermodi-lution technique. Br Heart J 30:690–695
51. Viquerat CE, Righetti A, Suter PM(1983) Biventricular volumes andfunction in patients with adult respira-tory distress syndrome ventilated withPEEP. Chest 83:509–514
52. Dhainaut JF, Devaux JY, Monsallier JF,Brunet F, Villemant D, HuyghebaertMF (1986) Mechanisms of decreasedleft ventricular preload during continu-ous positive pressure ventilation inARDS. Chest 90:74–80
53. Terai C, Uenishi M, Sugimoto H, Shimazu T, Yoshioka T, Sugimoto T(1985) Transesophageal echocardio-graphic dimensional analysis of fourcardiac chambers during positive end-expiratory pressure. Anesthesiology63:640–646
54. Jardin F, Brun-Ney D, Hardy A, Aegerter P, Beauchet A, Bourdarias JP(1991) Combined thermodilution andtwo-dimensional echocardiographicevaluation of right ventricular functionduring respiratory support with PEEP.Chest 99:162–168
55. Vieillard A, Schmitt JM, Beauchet A,Augarde R, Prin S, Page B, Jardin F(2001) Early preload adaptation in sep-tic shock? A transesophageal echocar-diographic study. Anesthesiology94:400–406
56. Jellinek H, Krafft P, Hiesmayr M,Steltzer H (1997) Measurement of rightventricular performance during apneain patients with acute lung injury. J Trauma 42:1062–1067
57. Diebel LN, Myers T, Dulchavsky S(1997) Effects of increasing airwaypressure and PEEP on the assessment ofcardiac preload. J Trauma 42:585–590
58. Cheatham ML, Nelson LD, Chang MC,Safcsak K (1998) Right ventricularend-diastolic volume index as a predic-tor of preload status in patients on posi-tive end-expiratory pressure. Crit CareMed 26:1801–186
59. Jansen JR, Schreuder JJ, Bogaard JM,van Rooyen W, Versprille A (1981)Thermodilution technique for measure-ment of cardiac output during artificialventilation. J Appl Physiol 51:584–591
60. Stevens JH, Raffin TA, Mihm FG, Rosenthal MH, Stetz CW (1985) Ther-modilution cardiac output measure-ment. Effects of the respiratory cycleon its reproducibility. JAMA253:2240–2242
61. Okamoto K, Komatsu T, Kumar V,Sanchala V, Kubal K, Bhalodia R,Shibutani K (1986) Effects of intermit-tent positive-pressure ventilation oncardiac output measurements by ther-modilution. Crit Care Med 14:977–980
62. Jansen JR, Schreuder JJ, Settels JJ,Kloek JJ, Versprille A (1990) An ade-quate strategy for the thermodilutiontechnique in patients during mechani-cal ventilation. Intensive Care Med16:422–425
63. Siniorakis EE, Nikolaou NI, Sarantopoulos CD, Sotirelos KT, Iliopoulos NE, Bonoris PE (1994) Volume loading in predominant rightventricular infarction: bedside haemo-dynamics using rapid response ther-mistors. Eur Heart J 15:1340–1347
64. Spinale FG, Mukherjee R, Tanaka R,Zile MR (1992) The effects of valvularregurgitation on thermodilution ejec-tion fraction measurements. Chest101:723–731
65. Yu M, Takiguchi S, Takanishi D, Myers S, McNamara JJ (1995) Evalua-tion of the clinical usefulness of ther-modilution volumetric catheters. Crit Care Med 23:681–686
66. Jardin F, Farcot JC, Gueret P, Prost JF,Ozier Y, Bourdarias JP (1984) Echo-cardiographic evaluation of ventriclesduring continuous positive airwaypressure breathing. J Appl Physiol56:619–627
67. Huemer G, Kolev N, Kurz A, ZimpferM (1994) Influence of positive end-expiratory pressure on right and leftventricular performance assessed byDoppler two-dimensional echocardiog-raphy. Chest 106:67–73
68. Jardin F, Dubourg O, Bourdarias JP(1997) Echocardiographic pattern ofacute cor pulmonale. Chest111:209–217
69. Mathru M, Kleinman B, Dries DJ, RaoT, Calandra D (1990) Effect of openingthe pericardium on right ventricular hemodynamics during cardiac surgery.Chest 98:120–123
55
360
70. Schuster S, Erbel R, Weilemann LS,Lu WY, Henkel B, Wellek S, SchinzelH, Meyer J (1990) Hemodynamics dur-ing PEEP ventilation in patients withsevere left ventricular failure studiedby transesophageal echocardiography.Chest 97:1181–1189
71. Le Tulzo Y, Seguin P, Gacouin A, Camus C, Suprin E, Jouannic I, Thomas R (1997) Effects of epineph-rine on right ventricular function in pa-tients with severe septic shock andright ventricular failure: a preliminarydescriptive study. Intensive Care Med23:664–670
72. Vignon P, Mentec H, Terre S, GastinneH, Gueret P, Lemaire F (1994) Diag-nostic accuracy and therapeutic impactof transthoracic and transesophagealechocardiography in mechanically ven-tilated patients in the ICU. Chest106:1829–1834
73. Jardin F, Brun-Ney D, Auvert B, Beauchet A, Bourdarias JP (1990) Sepsis-related cardiogenic shock. Crit Care Med 18:1055–1060
74. Jardin F, Fourme T, Page B, LoubieresY, Vieillard-Baron A, Beauchet A,Bourdarias JP (1999) Persistent pre-load defect in severe sepsis despite flu-id loading: a longitudinal echocardio-graphic study in patients with septicshock. Chest 116:1354–1359
75. Clements FM, Harpole DH, Quill T,Jones RH, McCann RL (1990) Estima-tion of left ventricular volume andejection fraction by two-dimensionaltransoesophageal echocardiography:comparison of short axis imaging andsimultaneous radionuclide angiogra-phy. Br J Anaesth 64:331–336
76. Urbanowicz JH, Shaaban MJ, CohenNH, Cahalan MK, Botvinick EH, Chatterjee K, Schiller NB, Dae MW,Matthay MA (1990) Comparison oftransesophageal echocardiographic andscintigraphic estimates of left ventricu-lar end-diastolic volume index andejection fraction in patients followingcoronary artery bypass grafting. Anesthesiology 72:607–612
77. van Daele ME, Trouwborst A, vanWoerkens LC, Tenbrinck R, FraserAG, Roelandt JR (1994) Transesopha-geal echocardiographic monitoring of preoperative acute hypervolemic hemodilution. Anesthesiology81:602–609
78. Greim CA, Roewer N, Apfel C, LauxG, Schulte am Esch J (1997) Relationof echocardiographic preload indices tostroke volume in critically ill patientswith normal and low cardiac index. Intensive Care Med 23:411–416
79. Feissel M, Michard F, Mangin I, RuyerO, Faller JP, Teboul JL (2001) Respira-tory changes in aortic blood velocity asan indicator of fluid responsiveness inventilated patients with septic shock.Chest 119:867–873
80. Reuter DA, Felbinger TW, Schmidt C,Kilger E, Goedje O, Lamm P, GoetzAE (2002) Stroke volume variationsfor assessment of cardiac responsive-ness to volume loading in mechanicallyventilated patients after cardiac sur-gery. Intensive Care Med 28:392
81. Denault AY, Gasior TA, Gorcsan J, 3rd,Mandarino WA, Deneault LG, PinskyMR (1999) Determinants of aorticpressure variation during positive-pres-sure ventilation in man. Chest116:176–186
82. Morgan BC, Martin WE, Hornbein TF,Crawford EW, Guntheroth WG (1966)Hemodynamic effects of intermittentpositive pressure respiration. Anesthesiology 27:584–590
83. Scharf SM, Brown R, Saunders N,Green LH (1980) Hemodynamic effects of positive-pressure inflation. J Appl Physiol 49:124–131
84. Brower R, Wise RA, HassapoyannesC, Bromberger-Barnea B, Permutt S(1985) Effect of lung inflation on lungblood volume and pulmonary venousflow. J Appl Physiol 58:954–963
85. Abel JG, Salerno TA, Panos A, Greyson ND, Rice TW, Teoh K, Lichtenstein SV (1987) Cardiovasculareffects of positive pressure ventilationin humans. Ann Thorac Surg43:198–206
86. Fessler HE, Brower RG, Wise RA, Permutt S (1988) Mechanism of re-duced LV afterload by systolic and dia-stolic positive pleural pressure. J ApplPhysiol 65:1244–1250
87. Pinsky MR, Matuschak GM, Klain M(1985) Determinants of cardiac aug-mentation by elevations in intrathorac-ic pressure. J Appl Physiol58:1189–1198
88. Taylor RR, Covell JW, SonnenblickEH, Ross J (1967) Dependence of ven-tricular distensibility on filling of theopposite ventricle. Am J Physiol213:711–718
89. Coyle JP, Teplick RS, Michael CL, Davison JK (1983) Respiratory varia-tions in systemic arterial pressure as anindicator of volume status. Anesthesi-ology 59:A53
90. Perel A, Pizov R, Cotev S (1987) Systolic blood pressure variation is asensitive indicator of hypovolemia inventilated dogs subjected to gradedhemorrhage. Anesthesiology67:498–502
91. Coriat P, Vrillon M, Perel A, Baron JF,Le Bret F, Saada M, Viars P (1994) A comparison of systolic blood pres-sure variations and echocardiographicestimates of end-diastolic left ventricu-lar size in patients after aortic surgery.Anesth Analg 78:46–53
92. Jardin F, Farcot JC, Gueret P, Prost JF,Ozier Y, Bourdarias JP (1983) Cyclicchanges in arterial pulse during respira-tory support. Circulation 68:266–274
93. Michard F, Teboul JL (2000) Usingheart-lung interactions to assess fluidresponsiveness during mechanical ven-tilation. Crit Care 4:282–289
94. Gunn SR, Pinsky MR (2001) Implica-tions of arterial pressure variation inpatients in the intensive care unit. Curr Opin Crit Care 7:212–217
95. Rooke GA, Schwid HA, Shapira Y(1995) The effect of graded hemor-rhage and intravascular volume re-placement on systolic pressure varia-tion in humans during mechanical andspontaneous ventilation. Anesth Analg80:925–932
96. Ornstein E, Eidelman LA, Drenger B,Elami A, Pizov R (1998) Systolic pres-sure variation predicts the response toacute blood loss. J Clin Anesth10:137–140
97. Reuter DA, Felbinger TW, Kilger E,Schmidt C, Lamm P, Goetz AE (2002)Optimising fluid therapy in mechani-cally ventilated patients after cardiacsurgery by on-line monitoring of leftventricular stroke volume variations.Comparison with aortic systolic pres-sure variations. Br J Anaesth88:124–126
56
Continuing Medical Education Article
Fluid responsiveness in spontaneously breathing patients: A reviewof indexes used in intensive care
Alice Coudray, MD; Jacques-André Romand, MD, FCCM; Miriam Treggiari, MD, MPH;Karim Bendjelid, MD, MS
I n critically ill patients, volume ex-pansion is frequently used to im-prove the hemodynamic profileand restore adequate blood pres-
sure when an absolute or relative hypo-volemia is suspected (1). Thus, in theintensive care setting, hemodynamic
measurements (right atrial pressure[Rap], pulmonary artery occlusion pres-sure [Ppao], and cardiac output [CO]) areroutinely used to evaluate preload depen-dency (1–3). However, several studies(1–3) demonstrated that these indexesare not reliable predictors of fluid respon-
siveness (FR). The results of these studieshighlight that accurate indexes areneeded to correctly predict the responseto volume therapy in critically ill patients(3).
According to the Starling law, whichdescribes a positive relation between car-diac muscle fiber length and contractility(4), fluid administration is thought toimprove CO by increasing preload. How-ever, beyond its ascending limb, the Star-ling curve reaches a plateau and furtherfluid administration can be deleterious,leading to right ventricular overloadand/or pulmonary edema (5, 6). Several
LEARNING OBJECTIVES
On completion of this article, the reader should be able to:
1. Define static and dynamic indexes.
2. List indexes that are valuable to predict fluid responsiveness in spontaneously breathing patients.
3. Use this information in a clinical setting.
All authors have disclosed that they have no financial relationships or interests in any commercial companies pertaining tothis educational activity.
Wolters Kluwer Health has identified and resolved all faculty conflicts of interest regarding this educational activity.
Visit the Critical Care Medicine Web site (www.ccmjournal.org) for information on obtaining continuing medical educationcredit.
Objective: In spontaneously breathing patients, indexes predictinghemodynamic response to volume expansion are very much needed.The present review discusses the clinical utility and accuracy ofindexes tested as bedside indicators of preload reserve and fluidresponsiveness in hypotensive, spontaneously breathing patients.
Data Source: We conducted a literature search of the MEDLINEdatabase and the trial register of the Cochrane Group.
Study Selection: Identification of reports investigating, prospec-tively, indexes of fluid responsiveness in spontaneously breathingcritically ill patients. All the studies defined the response to fluidtherapy after measuring cardiac output and stroke volume using thethermodilution technique. We did not score the methodological qual-ity of the included studies before the data analysis.
Data Extraction: A total of eight prospective clinical studies incritically ill patients were included. Only one publication evalu-ated cardiac output changes induced by fluid replacement in a
selected population of spontaneously breathing critically ill pa-tients.
Data Synthesis: Based on this review, we can only concludethat static indexes are valuable tools to confirm that the fluidvolume infused reaches the cardiac chambers, and thereforethese indexes inform about changes in cardiac preload. However,respiratory variation in right atrial pressure, which represents adynamic measurement, seems to identify hypotension related to adecrease in preload and to distinguish between responders andnonresponders to a fluid challenge.
Conclusions: Further studies should address the question ofthe role of static indexes in predicting cardiac output improve-ment following fluid infusion in spontaneously breathing patients.(Crit Care Med 2005; 33:2757–2762)
KEY WORDS: fluid resuscitation; static indexes; dynamic indexes;monitoring
Research Fellow, Surgical Intensive Care (AC),Lecturer (J-AR), Scientific Assistant Professor, Chef deClinique Scientifique (KB), Geneva University Hospital,Geneva, Switzerland; Associate Professor of Anesthe-siology, University of Washington, Seattle, WA (MT).
The authors declare no conflict of interest.Address requests for reprints to: Karim Bendjelid,
MD, MS, Surgical Intensive Care Division, Geneva Uni-versity Hospitals, CH-1211 Geneva 14, Switzerland.E-mail: [email protected]
Copyright © 2005 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000189942.24113.65
2757Crit Care Med 2005 Vol. 33, No. 12
57
indexes as Rap (7–10), Ppao (9–11), andventricular end-diastolic volumes (12–14) have already been reported to predictFR in critically ill patients receiving pos-itive pressure ventilation. When a fluidchallenge was administered to mechani-cally ventilated (MV) patients (15), somepatients increased CO (responders),whereas others did not (nonresponders)(16). Baseline filling pressures and vol-umes were compared among respondersand nonresponders to evaluate if theycould predict FR (1). Static indexes asRap, Ppao, right ventricular end-diastolicvolume index (RVEDVI), and left ventric-ular end-diastolic volume index (LVEDVI)were not found to be accurate FR predic-tors in MV patients (3). In contrast, dy-namic indexes (systolic pressure, pulsepressure, preejection period, and strokevolume variations secondary to respira-tory cycle) have been shown to be reliablepredictors of FR (17–20). Dynamic in-dexes, however, are valid only when mea-sured in deeply sedated MV patients insinus rhythm, since they result fromheart-lung interaction during well-defined, stable, positive-pressure ventila-tion cycles (3).
Spontaneous breathing (SB) condi-tions differ from positive pressure venti-lation in deeply sedated patients becauseintrathoracic pressure is negative during
inspiration (21, 22) and the respiratoryrate is variable. Moreover, the amplitudeof the intrathoracic pressure swings, inpatients with no heart or lung diseases, ismuch lower in SB than in MV patients(23). Does this situation improve the re-liability of static indexes to predict FR inSB patients? The present review analyzesthe strengths and limitations of the mostfrequently used methods in clinical prac-tice to predict FR in SB patients.
METHODS
Search Strategy. Combinations of keywords related to FR (e.g., fluid responsiveness,volume expansion, fluid challenge, preloadassessment) and SB adult subjects were usedto search the MEDLINE database and trialregister of the Cochrane Group. The lastsearch was December 23, 2004. We checkedthe bibliographies of retrieved reports and re-views. We did not consider data from ab-stracts, letters, and animal studies.
Inclusion Criteria, End Points, and Defini-tions. Studies investigating FR in a selecteddeeply sedated MV population were excluded.All the studies cited in this review defined theresponse to fluid therapy after measuring car-diac output and stroke volume using the ther-modilution technique (Table 1). When aftervolume infusion, cardiac output and/or strokevolume increased with at least a defined quan-tity, the patients were declared responders tofluid challenge. We did not score the method-
ological quality of the included studies beforethe data analysis.
RESULTS
Seven studies (7, 9–11, 13, 14, 24) andthree reviews (1, 3, 25) were identified.An additional study (12) was found in thereference list of one of the review articles.Therefore, a total of eight prospectiveclinical studies in critically ill patientswere included. Information from one ad-ditional study conducted in healthy vol-unteers was also included (26).
Static Measurements to PredictFluid Responsiveness in HealthyVolunteers
Only one FR study was conducted in12 healthy volunteers (26). Rap, Ppao,RVEDVI, LVEDVI, stroke volume index(SVI), and CO were assessed before andafter the infusion of 3 L of normal sa-line. Ventricular volumes were calcu-lated using ejection fraction assessed byradionuclide cineangiography, and COwas measured using the classic ther-modilution technique. After 3 L of sa-line infusion, the four static indexes(Rap, Ppao, RVEDVI, LVEDVI) increasedsignificantly. At baseline, no correlationwas found between Rap and RVEDVI or
Table 1. Summary study protocol and main results in eight studies identified in critically ill patients
Indexes Measured Technique Type of Patients References Nbre Pat SB (%)NbreFC
Rap PAC Mixed 7 28 54 28Rap PAC Mixed 9 41 24 82Rap PAC Mixed 10 25 5.6 36Rap PAC Mixed 11 33 36 31Rap PAC Septic shock 12 18 33 18Ppao PAC Mixed 7 28 54 28Ppao PAC Mixed 9 41 24 82Ppao PAC Mixed 14 15 31 22Ppao PAC Trauma � sepsis 13 26 16 65Ppao PAC Mixed 11 33 36 31Ppao PAC Mixed 10 25 5.6 36Ppao PAC Septic shock 12 18 33 18RVEDV PAC/cardiac scintigraphy Mixed 7 28 54 28RVEDVI PAC Mixed 14 15 31 22RVEDV PAC Trauma � sepsis 13 26 16 65RVEDVI PAC Mixed 10 25 5.6 36RVEDVI PAC/cardiac scintigraphy Septic shock 12 18 33 18RVEDVI PAC Mixed 9 41 24 82LVEDV PAC/cardiac scintigraphy Mixed 7 28 54 28LVEDVI PAC/cardiac scintigraphy Septic shock 12 18 33 18�Rap PAC Mixed 11 33 36 31�Rap PAC CABG 24 28 100 28
Eight studies 214 37.5 310
2758 Crit Care Med 2005 Vol. 33, No. 12
58
between Ppao and LVEDVI. Neither Rapnor Ppao was correlated with SVI atbaseline. Furthermore, none of thebaseline indexes (Rap, Ppao, RVEDVI,LVEDVI) were correlated with changein cardiac index after the fluid chal-lenge. Based on this study, it appearsthat static pressures and ventricularvolume measurements do not predictcardiac index increase after volume ex-pansion in healthy subjects (26). How-ever, changes in static indexes after vol-ume expansion were useful to confirmthat the fluid administered filled thecardiac chambers, as indicated by theirchange in size.
Static Measurements to PredictFluid Responsiveness inCritically Ill Patients
Measures of Intracardiac Pressures.The venous-return-Rap curve is used toillustrate that venous return decreaseswhen Rap (back-pressure) increases,given a constant mean systemic pressure(27–29). Furthermore, since venous re-turn equals CO (29), we may assume that
the intravascular volume status can bedefined by the relationship between pre-load and CO and that changing preloadwith volume infusion will affect CO (11).In physiologic conditions, a reasonablygood correlation exists between ventric-ular end-diastolic volumes and meanatrial pressure. Thus, ventricular preloadis approximated by Rap and/or Ppao in SBpatients (30, 31). However, under patho-logic conditions such as acute severe pul-monary embolism and/or marked hyper-inflation, Rap may also increasesecondary to an increased afterload of theright ventricle. Under these conditions,changes in intracardiac pressures (Rap,Ppao) no longer directly reflect changesin intravascular volume (3).
Regarding right atrial pressure as apredictor of fluid responsiveness, Wagneret al. (10) and Schneider et al. (12) re-ported that before a volume challenge,Rap was significantly lower in fluid re-sponders than in nonresponders. How-ever, the majority of patients were stud-ied during positive pressure ventilation(94.4% and 67%, respectively). In threeothers studies (7, 9, 11) combining SB
and MV patients (46%, 76%, and 64% ofMV patients), there was no relationshipbetween initial Rap and FR. Based onthese studies, the significance of Rap topredict FR in SB patients cannot be es-tablished definitively (Table 1). Interest-ingly, however, in all five studies (7,9–12), Rap increased significantly aftervolume infusion.
When we examined pulmonary arteryocclusion pressure as a predictor of fluidresponsiveness, five of seven studiesshowed no relationship between initialPpao and FR in a mixed population of SBand MV patients (7, 9, 11–13). An in-crease in Ppao after fluid infusion wasobserved in these five studies (7, 9, 11–13). The proportions of patients who werebreathing spontaneously were 16%, 24%,33%, 36%, and 54%. The other two stud-ies had discrepant results (10, 14). Thefirst one showed an initial Ppao higher inresponders than in nonresponders (14), aresult that is counterintuitive based oncommon physiologic knowledge. How-ever, only 31% of the patients were SB,and it is unknown if SB and MV patientswere equally distributed among respond-
Table 1. —continued
Volume (mL) andType of Plasma
Substitute Duration of FC Definition of R Definition of NR
Difference inIndexes’Baseline
Values R vs.NR
250 Alb 5% 20–30 mins 1 SVI 2 SVI or unchanged NS300 Alb 4.5% 30 mins 1 CI CI 2 or unchanged NS
NaCl 0.9% � Alb 5%,FFP to 1 Ppao
Until1 Ppao 3 mm Hg 1 SV �10% 1 SV �10% .04
NaCl 0.9% to 1 Rap Until1 Rap �2 mm Hg 1 CO �250 mL/min 1 CO �250 mL/min NS500 mL FFP 30 mins 1 SVI 2 SVI or unchanged �.05250 Alb 5% 20–30 mins 1 SVI 2 SVI or unchanged NS
300 Alb 4.5% 30 mins 1 CI CI 2 or unchanged NS300–500 RL ? (bolus) 1 C0 �10% C0 2 or unchanged �.01300–500 RL ? 1 CI �20% 1 CI �20% NS
NaCl 0.9% to 1 Rap Until 1 Rap �2 mm Hg 1 CO �250 mL/min 1 CO �250 mL/min NSNaCl 0.9%, Alb 5% Until 1 Ppao 3 mm Hg 1 SV �10% 1 SV �10% .001
500 mL FFP 30 mins 1 SVI 2 SVI or unchanged NS250 Alb 5% 20–30 mins 1 SVI 2 SVI or unchanged NS300–500 RL ? (bolus) 1 C0 �10% C0 2 or unchanged �.001300–500 RL ? 1 CI �20% 1 CI �20% �.05
NaCl 0.9%, Alb 5% Until 1 Ppao 3 mm Hg 1 SV �10% 1 SV �10% 0.22500 mL FFP 30 mins 1 SVI 2 SVI or unchanged NS300 Alb 4.5% 30 mins 1 CI CI 2 or unchanged NS250 Alb 5% 20–30 mins 1 SVI 2 SVI or unchanged NS500 mL FFP 30 mins 1 SVI 2 SVI or unchanged NS
NaCl 0.9% to 1 Rap Until 1 Rap �2 mm Hg 1 CO �250 mL/min 1 CO �250 mL/min �.05NaCl 0.9% to 1 Rap Until 1 Rap �2 mm Hg 1 CO �250 mL/min 1 CO �250 mL/min �.05
Eight studies: Nbre Pat, 214; SB, spontaneously breathing patients (37.5%); Nbre FC, 310; FC, fluid challenge; R, responders; NR, nonresponders; Rap,right atrial pressure; PAC, pulmonary artery catheter; Alb, albumin; SV, stroke volume; SVI, stroke volume index; NS, nonsignificant; FFP, fresh frozenplasma; CO, cardiac output; Ppao, pulmonary artery occlusion pressure; CI, cardiac index; RL, Ringer’s lactate; RVEDV, right ventricular end-diastolicvolume; RVEDVI, right ventricular end-diastolic volume index; LVEDV, left ventricular end-diastolic volume; LVEDVI, left ventricular end-diastolic volumeindex; �Rap, decrease in Rap �1 mm Hg following spontaneous inspiration; CABG, coronary artery bypass graft.
2759Crit Care Med 2005 Vol. 33, No. 12
59
ers and nonresponders (14). The secondstudy showed significantly lower Ppaovalues in responders compared with non-responders (10). However, in the latterstudy, only 5.6% of the patients were SB.Based on these reports, no conclusionscan be drawn regarding the usefulness ofPpao to predict FR when attempting torestrict to SB patients only (Table 1).
In summary, even if static intracardiacpressure measurements such as Rap andPpao have been used for many years forhemodynamic monitoring in critically illpatients, their predictive value in esti-mating FR in SB patients still needs to bedetermined. However, these indexes arecertainly a valuable tool to confirm thatthe volume infused reached the cardiacchambers and thus inform about cardiacpreload responsiveness.
Measures of Ventricular End-DiastolicVolumes. Radionuclide angiography (32),cineangiocardiography (33), and ther-modilution (34) techniques have beenused to estimate ventricular volumes forhalf a century. In the intensive care unit,different methods to assess ventricularend-diastolic volumes have been used atthe bedside such as radionuclide angiog-raphy (35, 36), transthoracic echocardi-ography (37–39), transesophageal echo-cardiography (40), and a modified flow-directed pulmonary artery catheter thatallows the measurement of CO and rightventricular ejection fraction (as well asthe calculation of right ventricular end-systolic and end-diastolic volumes) (10,13).
We examined right ventricular end-diastolic volume measured by pulmonaryartery catheter as a predictor of fluid re-sponsiveness. The pulmonary artery cathe-ter equipped with a special thermal fila-ment allows continuous CO measurementand RVEDVI calculation (10, 13). The latterindex has been shown to correlate well withan independently measured continuous CO(absence of a mathematical coupling) (41).In two clinical studies, a significant associ-ation was found between low baselineRVEDVI and FR (13, 14). These studies,performed by the same group, included16% and 31% of SB patients. In two othersstudies (9, 10), no correlation was foundbetween baseline RVEDVI and response tovolume challenge. These discrepanciescould in part be explained by differences inthe population studied. Indeed, Diebel et al.(13, 14) studied a majority of trauma pa-tients who were younger and more likely tohave healthy cardiovascular status (normalheart volumes at baseline), whereas the
other authors investigated elderly patients(higher prevalence of ischemic cardiomy-opathy, with enlarged ventricles and re-duced compliance) (9, 10) (Table 1). How-ever, thermodilution-derived RVEDVIchanges following volume infusion demon-strated that the volume infused reached thecardiac chambers (9, 14).
We next looked at right ventricularend-diastolic volume measured by pul-monary artery catheter and cardiac scin-tigraphy as a predictor of fluid respon-siveness. Using the classic pulmonaryartery catheter to measure CO and car-diac scintigraphy to estimate the rightventricular ejection fraction, the RVEDVImay be calculated (35). Two studies (7,12) showed no relationship between base-line RVEDVI and FR. Again, the popula-tion was mixed in terms of requirementfor respiratory support (33% and 54% ofSB patients). One study, using RVEDVIcalculated from cardiac scintigraphy, in-dicated that this index is a valuable toolto confirm that the volume infused filledthe cardiac chambers (12).
Finally, we examined left ventricularend-diastolic volume measured by pul-monary artery catheter and cardiac scin-tigraphy as a predictor of fluid respon-siveness. The same two studies (7, 12)measured LVEDVI as well (using the leftventricular ejection fraction). Both sug-gested that LVEDVI was not a good pre-dictor of FR (Table 1). However, in onestudy, LVEDVI was shown to be a valu-able tool to indicate that the volume in-fused reached the cardiac chambers (12).
Although measurement of ventricularvolumes should reflect preload depen-dence more accurately than other in-dexes, their predictive value in estimatingFR in critically ill patients is poor, withexception of two studies published by thesame investigators (13, 14). Unfortu-nately, no study evaluated the value ofRVEDVI and LVEDVI in predicting FR ina selected group of critically ill SB pa-tients. However, these static indexes arevaluable tools (9, 12, 14) to evaluate if thevolume infused reached the cardiacchambers and to test the cardiac preloadresponsiveness.
Dynamic Measurements toPredict Fluid Responsiveness inCritically Ill Patients
Respiratory Variations in Right AtrialPressure as a Predictor of Fluid Respon-siveness. During a normal inspiration, adecrease in pleural and pericardial pres-
sure is observed. This pressure changeleads to a decrease in Rap, ultimatelyincreasing venous return and right car-diac output. A few heartbeats later, theincrease in right CO results in an eleva-tion of the left CO.
Respiratory variation in Rap was in-vestigated by Magder et al. (11, 24) as apredictor of FR. The hypothesis was thatwhen the heart is not volume responsive(i.e., on the flat portion of the Starlingcurve), Rap will not fall during inspira-tion nor will the CO rise after volumeinfusion. In a first study (11), the authorsincluded both SB (36%) and MV (64%)patients. In SB MV patients, the inspira-tory fall in Rap was assessed during ashort disconnection from ventilator. Raphad to fall by �1 mm Hg during inspira-tion to be considered positive for respira-tory variations. Patients were excluded ifthe respiratory effort was not sufficient togenerate at least a 2 mm Hg decrease inPpao during inspiration. Respiratory vari-ation in Rap (�Rap) was a very good pre-dictor of FR. Importantly, the static Rapvalue before volume expansion was not areliable index to discriminate respondersfrom nonresponders to a fluid challenge(11) (Table).
The second study, by the same groupof investigators (24), focused on a popu-lation of SB patients where �Rap wasdefined using the criteria previously de-scribed. Again, a strong association wasfound between �Rap and FR (Table).
DISCUSSION
The potential increase in vascular ca-pacitance and capillary leakage may affectthe predictive value of FR indexes. In-deed, volume infusion might increase ei-ther intravascular blood volume (venouspooling) or interstitial space (capillaryleak) but not necessarily cardiac preload(42, 43). Unless some objective measurecan confirm that the fluid infused led toan increase in ventricular volumes, car-diac preload responsiveness cannot beevaluated (11). Therefore, the validationof an indicator of FR requires the dem-onstration of two simultaneous condi-tions: a) that the index changed in paral-lel with the fluid load and that thischange is associated with an increase instroke volume; and b) that the baselineindex (before volume expansion) is corre-lated with the change in stroke volumefollowing volume infusion.
Briefly, dynamic indexes such as sys-tolic pressure and pulse pressure varia-
2760 Crit Care Med 2005 Vol. 33, No. 12
60
tions have been investigated to evaluateFR in MV patients (3, 17–20). However,all of these indexes have been validated indeeply sedated patients under positivepressure ventilation. Therefore, their ap-plicability cannot be extrapolated to SBintubated patients, a mechanical ventila-tion mode often used in critically ill in-tensive care patients. Moreover, regularcardiac rhythm is required to interpretthese indexes.
Our primary goal when we plannedthis work was to perform a meta-analysisto respond to the question: Which indexof fluid responsiveness is a valuable toolin spontaneously breathing patient?However, at completion of our literaturesearch, it became apparent that conduct-ing a formal meta-analysis would nothave been appropriate, given the insuffi-cient data reported in the literature andthe heterogeneity of patient selection andtechniques used to assess fluid respon-siveness (e.g., volume [mL] and type ofplasma substitute, duration of FC, per-cent of spontaneously breathing patients,responders definitions, type of patients).Moreover, we may expect that the datanot presented may be of interest butwould be very difficult to obtain as thesestudies were performed between 1981and 1999.
Based on published literature, very lim-ited information is available about the va-lidity of static indexes (Rap, Ppao, RVEDVI,and LVEDVI) to predict FR in critically illSB patients. Indeed, there was no patientselection or stratification based on the typeof ventilatory support (i.e., both SB and MVpatients were pooled in the analyses re-ported). As static indexes are modified bypositive pressure ventilation, it is possiblethat the response to fluid challenge differswhen fluids are administered to SB or MV
patients. It is necessary to analyze FRamong SB and MV patients separately.Therefore, further studies are needed toevaluate if static indexes could be used aspredictors of FR in SB critically ill patients.
With regard to dynamic indexes, onlyone study addressed FR in a selectedgroup of SB critically ill patients (24)(Table). �Rap was evaluated and found tobe reliable in predicting FR. In this study,however, the significance of Ppao varia-tion (�2 mm Hg inspiratory decrease) asan indicator of a sufficient respiratoryeffort may be questioned. Indeed, we mayexpect that if left ventricular chambersare overfilled or the pericardial constraintis increased, no variation in Ppao willoccur even if the pleural pressure falls.Thus, further studies are also needed toconfirm �Rap reliability in predicting FRin SB patients. Moreover, another ques-tion should be addressed: Does fluid ther-apy guided by FR indexes improve patientoutcome? Indeed, it is still unclear if anyform of monitoring-guided fluid therapyimproves survival in SB patients.
CONCLUSIONS
Although the literature about indexesof hypovolemia is abundant, only one re-port investigated the CO changes inducedby volume infusion in a selected popula-tion of critically ill patients breathingspontaneously. Based on this review, wecan only conclude that static indexes arevaluable tools to demonstrate that thevolume load has affected the cardiac pre-load. It remains unclear whether staticindexes can predict CO improvement fol-lowing fluid infusion in critically ill SBpatients. Finally, respiratory variation inRap, which represents a dynamic mea-surement, seems to identify hypotensionrelated to a decrease in preload and todistinguish between responders and non-responders to a fluid challenge (Table 1).This index is probably the most promis-ing for the assessment of FR in criticallyill SB patients.
We believe that there are no convinc-ing data to claim that one index is supe-rior to another one. Our work is an up-date of fluid responsiveness indexes inspontaneously breathing critically ill pa-tients. It underlines that fluid responsive-ness indexes in spontaneously breathingpatients have not been studied exten-sively compared with mechanically venti-lated patients.
REFERENCES
1. Michard F, Teboul JL: Predicting fluid re-sponsiveness in ICU patients: A critical anal-ysis of the evidence. Chest 2002; 121:2000–2008
2. Hollenberg SM, Ahrens TS, Annane D, et al:Practice parameters for hemodynamic sup-port of sepsis in adult patients: 2004 update.Crit Care Med 2004; 32:1928–1948
3. Bendjelid K, Romand JA: Fluid responsive-ness in mechanically ventilated patients: Areview of indices used in intensive care. In-tensive Care Med 2003; 29:352–360
4. Starling E: The Linacre Lecture on the Lawof the Heart Given at Cambridge, 1915. Lon-don, UK, Longmans, Green and Co, 1918
5. Balogh Z, McKinley BA, Cocanour CS, et al:Supranormal trauma resuscitation causesmore cases of abdominal compartment syn-drome. Arch Surg 2003; 138:637–642
6. Boussat S, Jacques T, Levy B, et al: Intravas-cular volume monitoring and extravascularlung water in septic patients with pulmonaryedema. Intensive Care Med 2002; 28:712–718
7. Calvin JE, Driedger AA, Sibbald WJ: The he-modynamic effect of rapid fluid infusion incritically ill patients. Surgery 1981; 90:61–76
8. Calvin JE, Driedger AA, Sibbald WJ: Does thepulmonary capillary wedge pressure predictleft ventricular preload in critically ill pa-tients? Crit Care Med 1981; 9:437–443
9. Reuse C, Vincent JL, Pinsky MR: Measure-ments of right ventricular volumes duringfluid challenge. Chest 1990; 98:1450–1454
10. Wagner JG, Leatherman JW: Right ventricu-lar end-diastolic volume as a predictor of thehemodynamic response to a fluid challenge.Chest 1998; 113:1048–1054
11. Magder S, Georgiadis G, Cheong T: Respira-tory variations in right atrial pressure predictthe response to fluid challenge. J Crit Care1992; 7:76–85
12. Schneider AJ, Teule GJ, Groeneveld AB, et al:Biventricular performance during volumeloading in patients with early septic shock,with emphasis on the right ventricle: A com-bined hemodynamic and radionuclide study.Am Heart J 1988; 116:103–112
13. Diebel L, Wilson RF, Heins J, et al: End-diastolic volume versus pulmonary arterywedge pressure in evaluating cardiac preloadin trauma patients. J Trauma 1994; 37:950–955
14. Diebel LN, Wilson RF, Tagett MG, et al: End-diastolic volume. A better indicator of pre-load in the critically ill. Arch Surg 1992;127:817–821
15. Horst HM, Obeid FN: Hemodynamic re-sponse to fluid challenge: A means of assess-ing volume status in the critically ill. HenryFord Hosp Med J 1986; 34:90–94
16. Stetz CW, Miller RG, Kelly GE, et al: Reliabilityof the thermodilution method in the determi-nation of cardiac output in clinical practice.Am Rev Respir Dis 1982; 126:1001–1004
F urther studies
should address the
question of the
role of static indexes in pre-
dicting cardiac output im-
provement following fluid in-
fusion in spontaneously
breathing patients.
2761Crit Care Med 2005 Vol. 33, No. 12
61
17. Tavernier B, Makhotine O, Lebuffe G, et al:Systolic pressure variation as a guide to fluidtherapy in patients with sepsis-induced hypo-tension. Anesthesiology 1998; 89:1313–1321
18. Michard F, Boussat S, Chemla D, et al: Relationbetween respiratory changes in arterial pulsepressure and fluid responsiveness in septic pa-tients with acute circulatory failure. Am J Re-spir Crit Care Med 2000; 162:134–138
19. Reuter DA, Felbinger TW, Schmidt C, et al:Stroke volume variations for assessment of car-diac responsiveness to volume loading in me-chanically ventilated patients after cardiac sur-gery. Intensive Care Med 2002; 28:392–398
20. Bendjelid K, Suter PM, Romand JA: The re-spiratory change in preejection period: A newmethod to predict fluid responsiveness.J Appl Physiol 2004; 96:337–342
21. Magder S: More respect for the CVP. Inten-sive Care Med 1998; 24:651–653
22. Magder S: Clinical usefulness of respiratoryvariations in arterial pressure. Am J RespirCrit Care Med 2004; 169:151–155
23. Guyton AC: Regulation of cardiac output.N Engl J Med 1967; 277:805–812
24. Magder S, Lagonidis D: Effectiveness of albu-min versus normal saline as a test of volumeresponsiveness in post-cardiac surgery pa-tients. J Crit Care 1999; 14:164–171
25. Gunn SR, Pinsky MR: Implications of arterialpressure variation in patients in the intensivecare unit. Curr Opin Crit Care 2001;7:212–217
26. Kumar A, Anel R, Bunnell E, et al: Pulmo-nary artery occlusion pressure and centralvenous pressure fail to predict ventricularfilling volume, cardiac performance, or the
response to volume infusion in normal sub-jects. Crit Care Med 2004; 32:691–699
27. Guyton AC, Satterfield JH, Harris JW: Dy-namics of central venous resistance with ob-servations on static blood pressure. Am JPhysiol 1952; 169:691–699
28. Guyton AC, Lindsey AW, Abernathy B, et al:Venous return at various right atrial pres-sures and the normal venous return curve.Am J Physiol 1957; 189:609–615
29. Guyton AC: Determination of cardiac outputby equating venous return curves with car-diac response curves. Physiol Rev 1955; 35:123–129
30. Crexells C, Chatterjee K, Forrester JS, et al:Optimal level of filling pressure in the leftside of the heart in acute myocardial infarc-tion. N Engl J Med 1973; 289:1263–1266
31. Buchbinder N, Ganz W: Hemodynamic mon-itoring: Invasive techniques. Anesthesiology1976; 45:146–155
32. Mullins CB, Mason DT, Ashburn WL, et al:Determination of ventricular volume by ra-dioisotope-angiography. Am J Cardiol 1969;24:72–78
33. Kasser IS, Kennedy JW: Measurement of leftventricular volumes in man by single-planecineangiocardiography. Invest Radiol 1969;4:83–90
34. Balcon R, Oram S: Measurement of rightventricular end-systolic and end-diastolicvolumes by the thermodilution technique.Br Heart J 1968; 30:690–695
35. Viquerat CE, Righetti A, Suter PM: Biven-tricular volumes and function in patientswith adult respiratory distress syndrome ven-tilated with PEEP. Chest 1983; 83:509–514
36. Dhainaut JF, Devaux JY, Monsallier JF, et al:Mechanisms of decreased left ventricularpreload during continuous positive pressureventilation in ARDS. Chest 1986; 90:74–80
37. Terai C, Uenishi M, Sugimoto H, et al: Trans-esophageal echocardiographic dimensionalanalysis of four cardiac chambers during pos-itive end-expiratory pressure. Anesthesiology1985; 63:640–646
38. Jardin F, Brun-Ney D, Hardy A, et al: Com-bined thermodilution and two-dimensionalechocardiographic evaluation of right ven-tricular function during respiratory supportwith PEEP. Chest 1991; 99:162–168
39. Jardin F, Valtier B, Beauchet A, et al: Invasivemonitoring combined with two-dimensionalechocardiographic study in septic shock. In-tensive Care Med 1994; 20:550–554
40. Vieillard A, Schmitt JM, Beauchet A, et al:Early preload adaptation in septic shock? Atransesophageal echocardiographic study.Anesthesiology 2001; 94:400–406
41. Nelson LD, Safcsak K, Cheatham ML, et al:Mathematical coupling does not explain therelationship between right ventricular end-diastolic volume and cardiac output. CritCare Med 2001; 29:940–943
42. Kumar A, Anel R, Bunnell E, et al: Preload-independent mechanisms contribute to in-creased stroke volume following large vol-ume saline infusion in normal volunteers: aprospective interventional study. Crit Care2004; 8:R128–R136
43. Michard F, Reuter DA: Assessing cardiac pre-load or fluid responsiveness? It depends onthe question we want to answer. IntensiveCare Med 2003; 29:1396; author reply 1397
2762 Crit Care Med 2005 Vol. 33, No. 12
62
Correlation Between Measured Inferior VenaCava Diameter and Right Atrial Pressure
Depends on the Echocardiographic MethodUsed in Patients Who Are Mechanically
VentilatedKarim Bendjelid, MD, Jacques-A. Romand, MD, FCCM, Bernhard Walder, MD,
Peter M. Suter, MD, FCCM, FCCP, and Gerard Fournier, MD, Geneve, Switzerland; andLyon-Sud, France
In patients who are mechanically ventilated, thecorrelation between inferior vena cava diameter(IVCD) measurements and mean right atrial pres-sure (RAP) varies in the literature. The purpose ofthis study was to test if the correlation between IVCDand RAP measurement in patients who are criticallyill depends on the transthoracic echocardiography(TTE) methodology used. Twenty patients who werecritically ill, sedated, and required respiratory sup-port were prospectively studied by TTE during me-chanical ventilation in a controlled mode. The TTEmeasures of IVCD were made, using methods previ-ously cited. First, IVCD was measured at end-expira-tion and end-diastole, with ECG synchronization,using the M-mode, on short-axis view 2 cm belowthe right atrium. Second, IVCD was assessed at end-expiration, without ECG synchronization, using the2-dimensional long-axis view at the same location.RAP was measured simultaneously by using a cen-tral venous catheter positioned in the superior vena
cava. All measurements were taken in the supineposition. IVCD at end-expiration and end-diastole,with ECG synchronization, using the M-mode, andIVCD at end-expiration, without ECG synchroniza-tion, using the 2-dimensional long-axis view, corre-late linearly with RAP (0.81, P < .0001 and 0.71, P �.0004). Mean bias between the 2 TTE methods(Bland-Altman analysis) was 1.6 mm (SD � 2.03mm). In conclusion, this study confirms that varia-tion of correlation between TTE IVCD measurementand RAP depends on the ultrasonographic method-ology used and the timing of measurement duringthe cardiac cycle. IVCD at end-expiration and end-diastole, with ECG synchronization, using the M-mode (IVCD-MM) correlates more satisfactory withRAP than with IVCD at end-expiration, without ECGsynchronization, using the 2-dimensional long-axisview, in patients during mechanical ventilation. (JAm Soc Echocardiogr 2002;15:944-9.)
In patients with mechanical ventilation, meanright atrial pressure (RAP), measured by a centralvenous catheter, is often used to estimate preloadand to guide intravascular volume therapy.1,2 Thisinvasive procedure may cause immediate or de-layed complications such as arterial puncture,pneumothorax, bloodstream infection, and throm-bosis of central veins.3,4 Less invasive methods to
estimate preload and response to fluid therapysuch as echocardiography and transoesophagealDoppler have been tested for many years. Oneproposed method is the inferior vena cava diam-eter (IVCD) measurement by transthoracic echo-cardiography (TTE) that has been reported tocorrelate closely with RAP in patients who arespontaneously breathing.5-8 However, in patientwho are mechanically ventilated, only 3 studieshave evaluated the correlation between IVCD andRAP9-11 and their conclusions were contradictory.Lichtenstein et al9 concluded that during ventila-tory support RAP correlated with IVCD, whereasJue et al10 and Nagueh et al11 observed unsatisfac-tory correlations between IVCD and RAP. Onepossible explanation for these differences may berelated to the methods used to measure IVCD. Wedesigned a study, in patients who were mechani-cally ventilated and critically ill, to investigatewhether the correlation between RAP and IVCDdepended on the TTE methodology used.
From the Surgical Intensive Care Division, Geneva UniversityHospital, Switzerland, and Service de Reanimation medicale, Cen-tre hospitalier Lyon-Sud, Faculte de Medecine Lyon-Sud (G.F.).Preliminary data have been presented as oral presentation to theXIIIth International Congress of Echocardiography, June 11,1999, Paris, France.Reprint requests: Karim Bendjelid, MD, chef de clinique, Divisiondes soins intensifs chirurgicaux, Hopitaux universitaires de Gen-eve, CH-1211 Geneve 14, Switzerland (E-mail:[email protected]).Copyright 2002 by the American Society of Echocardiography.0894-7317/2002/$35.00�0 27/1/120701doi:10.1067/mje.2002.120701
944
63
PATIENTS AND METHODS
The protocol was approved by the local institutionalethical board. Over a 6-month period, the study wasperformed in a medical intensive care department on 21patients undergoing mechanical ventilation. All patientshad a central venous catheter (Multicath, Vygon, Ecouen,France) inserted by the jugular or subclavian route forclinical reasons and were hemodynamically stable, asdefined by “a variation in heart rate and blood pressure ofless than 10% over the 15-minute period before startingthe protocol.” Patients were sedated with a target Ram-say12 score of 5. The sedation drug, midazolam (5-10mg/h), and an opiate drug, sufentanil citrate (15-30 mg/h),were administered continuously. One patient requiredneuromuscular blockade with vecuronium bromide foradaptation of mechanical ventilation. The patients weremechanically ventilated (Evita 4, Drager, Lubeck, Ger-many) with the following initial settings: volume assistcontrol mode: tidal volume of 8 to 10 mL/kg�1 of bodyweight and a respiratory rate of 12 breaths per minute�1.Ventilation parameters were adjusted to maintain a PaCO2
at 40 � 5 mm Hg. The inspired oxygen fraction wasadapted for a PaO2 more than 90 mm Hg. The level ofpositive end-expiratory pressure (PEEP) was lower orequal to 5 cm H20 in all patients. Vasopressor, inotropicdrugs, or both were administrated when needed accord-ing to standard criteria and the administration rate was notchanged during the study period.
Pressures Measurements
The correct positioning of the central venous catheter inthe superior vena cava was evaluated by chest radiograph.
Only patients with correct positioning, defined as “distaltip close to the junction with the high right atrium,” wereenrolled in this study. RAP was measured in the supineposition with fluid-filled tubing connected to a gradedcolumn of water. Calibration was performed before pres-sure measurements; 0 was referenced at midthorax. RAPwas determined at end-expiration and an average of 3measurements were obtained and the mean value calcu-lated.
Ultrasonography Measurements
TTE measures were performed with an ultrasound system(Sonos 5500 digital-imaging system, Hewlett-Packard, An-dover, Mass). Two measurement methods were used: themethod of Lichtenstein et al9 and the technique of Jue atal.10 The method of Lichtenstein et al9 uses 2-dimensional(2D) subcostal views, with the patient in the supineposition; the inferior vena cava (IVC) is displayed in itslongitudinal long axis and its circular size is appreciatedusing short-axis views during its thoracic course. The IVCis studied longitudinally and then measured transversely atend-expiration and end-diastole, ie, during the R wave onthe ECG. The measurement is made proximally to thejunction with the hepatic vein that lies approximately 2cm before right atrium. On short-axis view, the M-modecursor of the 2D sector was used to generate a timemotion recording of the circular size of the IVC (IVCD-MM) (Figure 1, A). In addition, IVC measurement wasassessed, according to the methodology of Jue et al10 thatmeasured IVCD using the 2D long-axis view 2 cm distal tothe junction with the right atrium during end-expirationand without ECG synchronization (IVCD-2D) (Figure 1,B). Offline IVCD-MM measurements were performed with
Figure 1 Left, Combined M-mode and 2-dimensional (2D) subcostal views, patient No. 10. Inferior venacava (IVC) is displayed on short-axis. M-mode recording is obtained with simultaneous 2D echocardi-ography imaging. Measure of IVCD is made on the M-mode echocardiogram, at end-expiration,synchronously to end-diastole (Q wave on ECG). Right, 2D subcostal views, patient No. 10. IVC isdisplayed in its longitudinal long-axis. HV, hepatic vein; RA, right atrium. Measure of IVCD is made atend-expiration without ECG synchronization.
Journal of the American Society of EchocardiographyVolume 15 Number 9 Bendjelid et al 945
64
the short-axis method without synchronization to end-diastole and also using the 2D long-axis view synchronizedto end-diastole. Because measures of RAP were not dis-played on intensive care department monitors but madewith fluid-filled tubing connected to a graded column ofwater, they were obtained simultaneously to IVCD-MMand IVCD-2D measurement by 2 independent blindedinvestigators (echocardiographer and nurse).
Statistical Analysis
Results are expressed as mean � SD. TTE measures ofIVCD-MM and IVCD-2D were correlated with simulta-neous measures of RAP using linear regression analysis.Linear regression and Bland-Altman analyses13 were usedto evaluate the agreement between IVCD-MM and IVCD-2D. Statistical calculation were carried out using SPSS 6.1software. (SPSS Inc, Cary, NC). Statistical significance wasset at P � .05.
RESULTS
In 1 patient, the IVCD was not obtained because ofthe inability to get a TTE subcostal view of the IVC.Mean time to perform and interpret the echocardio-graphic measurement was about 25 minutes. Thus,the protocol was conducted in 20 patients (13 men,7 women; mean age 62 years old, range 34-78) whorequired mechanical ventilation for respiratory fail-ure of various causes (Table 1). Average RAP was 7
cm H2O with a range from 2 to 14 cm H2O. Displayedby the 2D long-axis view, changes in IVCD during therespiratory cycle, estimated visually, were small andalways less than 5%. All patients presented a circularsize of IVC on the short-axis view. Examples of TTEdata measurements are shown in Figure 1, A and B.The linear correlation between IVCD-MM and RAP was0.81 (r2 � 0.66, P � .0001) (Table 2, Figure 2, A) andthe correlation between IVCD-2D and RAP was 0.71(r2 � 0.51, P � .0004) (Table 2, Figure 2, B). Whennonsynchronous to end-diastole, IVCD-MM correlatedless to RAP (r � 0.76; P � .0001) than IVCD-MM toRAP. When synchronous to end-diastole, IVCD-2Dcorrelated less to RAP (r � 0.75; r2 � 0.57; P � .0001)than IVCD-MM to RAP but better than IVCD-2D toRAP. The correlation between IVCD-MM and IVCD-2Dwas 0.88 (r2 � 0.77, P � .0001). Bland-Altman analysisbetween the 2 TTE methods showed a mean bias of1.6 mm in IVCD and a 95% confidence limit ofagreement of �2.47 mm to 5.67 mm (Figure 3). Withthe use of the regression equation
RAP (cm H2O) � 0.63 � IVCD-MM (mm) � 0.83
DISCUSSION
The current study demonstrates a positive correla-tion between IVCD-MM and RAP in patients who aremechanically ventilated and critically ill. Further-more, in comparison with the IVCD-2D, the corre-
Table 1 Patients characteristics
Patient
No Diagnosis
Sex
M/F Age (y) Weight (kg) Height (cm) BMI
Myo-
relaxant CA SAPS
Outcome
S/D
1 Acute pulmonary embolism M 69 80 168 28.3 N Y 28 S2 Bacterial pneumonia F 66 56 162 21.3 N Y 34 S3 Septic shock M 54 63 170 21.8 N N 38 S4 Septic shock M 73 68 173 22.7 N Y 50 D5 Guillain-Barre syndrome M 34 77 182 23.2 N N 17 S6 Acute pancreatitis F 55 52 159 20.6 N N 43 S7 Bacterial pneumonia M 49 64 176 20.7 N Y 32 S8 Drug intoxication M 64 64 166 23.2 N N 53 S9 COPD pneumonia F 67 85 163 32.0 N N 34 S
10 Intracranial hemorrhage F 70 74 157 30.0 Y N 52 D11 Septic shock M 61 92 176 29.7 N Y 53 S12 COPD pneumonia M 65 55 157 22.3 N Y 31 S13 Drug intoxication M 51 70 179 21.8 N N 37 D14 Acute pancreatitis F 78 61 163 23.0 N N 49 D15 COPD pneumonia M 64 82 171 28.0 N N 38 S16 Septic shock F 45 72 170 24.9 N Y 39 S17 Acute heart failure M 72 77 176 24.9 N N 49 S18 COPD pneumonia M 59 89 169 31.2 N Y 38 S19 Septic shock F 68 50 155 20.8 N Y 37 S20 Acute heart failure M 66 77 173 25.7 N N 39 S
Mean � SD 61.5 � 10.7 70.4 � 12.2 168.2 � 7.8 24.8 � 3.7 39.5 � 9.4Ratio 13/7 19/1 11/9 16/4
BMI, Body mass index; CA, cathecolamine (vasopressor and/or inotropic) support; SAPS, simplified acute physiology score; S/D, survived/dead; COPD,chronic obstructive pulmonary disease.
Journal of the American Society of Echocardiography946 Bendjelid et al September 2002
65
lation between IVCD-MM and RAP is more satisfac-tory.
The final position of the IVC and the right atriumare within the thorax. Thus, intrathoracic pressurechange during mechanical positive pressure ventila-tion affects both RAP and IVCD.14 Indeed, raisingintrathoracic pressure by positive pressure ventila-tion decreases venous return to the right ventricleand increases pressure and size of this compliantvessel.15 This increase of IVC size has been alreadystudied using 2D TTE.16,17 On the basis of thisprevious work,16,17 IVCD has been proposed as anoninvasive method for estimating RAP in patientsunder mechanical ventilation.9-11 In these patients,mean RAP is used for estimation of right ventriclepreload. However, the invasiveness of the method-ology required and the complications caused bycentral venous access and catheters in situ hasmotivated a search for alternatives. TTE is one ofthese. However, only a few studies have prospec-tively evaluated the correlation between TTE mea-surement of IVCD and RAP in patients who aresedated and undergoing positive pressure ventila-tion.9-11
The current study confirms the satisfactorycorrelation (r � 0.78, P � .0001) observed byLichtenstein et al9 and the poorer correlationobserved by Jue et al10 (r � 0.58, P � .001) andNagueh et al11 (r � 0.4, P � .23) between TTEIVCD and RAP in patients undergoing positive
pressure ventilation. How can these differencesbe explained? First, Jue et al10 and Nagueh et al11
have used IVCD-2D for measurement. However, inthe supine position this method may be lessaccurate because of the geometry of the vessel,which is not always circular. Indeed, the normalconfiguration of the IVC may vary from round toelliptical on short-axis and slender to wide onlong-axis views.18 This inaccuracy of measure-ment is minimized with IVCD-MM methodology5
that could appreciate the circular size of IVC.Second, a high incidence of tricuspid regurgita-tion and vena cava backflow has been observed inpatients who are mechanically ventilated19 thatcan affect the size of IVC and IVCD measures.20
The impact of this regurgitation can be avoided ifTTE IVCD measurement is made at the end-diastole (R wave on ECG) period when vena cavabackflow19 is not possible. Indeed, even if RAPcorrelated with IVCD-MM without synchroniza-tion to end-diastole and IVCD-2D synchronized toend-diastole, 0.76 (r2 � 0.58, P � .0001) and 0.75(r2 � 0.57, P � .0001), respectively, these corre-lations were lower than the correlation observedbetween RAP and IVCD-MM. Thus, the superiorityof IVCD-MM over IVCD-2D is probably relatedboth to the timing of measurement and to thedifferent view of the IVCD measurement. Morenoet al5 examined the relation between IVCD-MMand IVCD-2D in 175 patients who were spontane-ously breathing. Their correlation was 0.84 be-tween M- and 2D-mode and 0.88 between long-and short-axis measurements. In our patients whowere mechanically ventilated, the same correla-tion of 0.88 was found between IVCD-MM andIVCD-2D. However, the mean bias between bothtechniques (Bland-Altman method13) was 1.6 � 2,demonstrating a limited agreement21 between the2 methods of measure.
Clinical Implications
The IVCD-MM-ultrasonographic assessment is a sim-ple technique even for an echocardiographer with-out extensive experience and may provide a wel-come alternative for estimating cardiac filling inpatients undergoing mechanical ventilation. It ispredictable that the M-mode approach should bemore strongly correlated to mean RAP because ofthe greater accuracy and resolving power of M-modeand the minimization of the effect of backflow fromtricuspid regurgitation. The importance of determi-nation of RAP is not only in monitoring preload butalso in noninvasive assessment of pulmonary arterialpressure using Doppler-echocardiography. Indeed,using the regression equation
RAP (cm H2O) � 0.63 � IVCD-MM (mm) � 0.83
Table 2 Simultaneous measure of right arterial pressureand inferior vena cava diameter
Patient No RAP (cm H2O)
IVCD-MM
(mm)
IVCD-2D
(mm)
1 14 23 182 10 13 123 3 7 54 5 9 75 7 10 86 11 17 187 6 16 148 5 14 139 12 20 20
10 8 17 1511 5 11 1012 9 12 1513 2 8 914 12 15 915 8 12 916 7 14 1117 4 9 718 5 8 819 9 12 1020 6 13 10
Mean�SD 7.4 � 3.2 13.1 � 4.1 11.4 � 4.1
RAP, right arterial pressure; IVCD-MM, inferior vena cava diameter usingM-mode echocardiogram; IVCD-2D, inferior vena cava diameter using2-dimensional echocardiogram.
Journal of the American Society of EchocardiographyVolume 15 Number 9 Bendjelid et al 947
66
the noninvasive assessment of systolic pulmonaryarterial pressure using maximal velocity of tricuspidregurgitation should be more accurate in patientswho are mechanically ventilated.
Study Limitations
The patients of this study had mean PEEP levels of 5cm H2O or less, thus, we cannot extend our resultsto situations where higher PEEP is used. Indeed,when analyzing the subgroup of patients (n �28/49) with low PEEP levels (�5 cm H2O) Jue et al10
observed a slightly improved correlation betweenIVCD-2D and RAP, 0.64 (PEEP � 5 cm H2O) versus
0.58 (PEEP � 6 � 4 cm H2O), respectively. Never-theless, in our study, 4 of 20 patients had chronicobstructive pulmonary diseases raising the possibil-ity that intrinsic PEEP was present, which mayexplain some of the discrepancy between IVCD andRAP. Another limitation of this study is that even if apositive correlation is found between IVCD-MM andRAP, the wide overlap of data points observedreduce the value of the IVCD measurements toaccurately predict RAP in the individual patient.
CONCLUSION
The correlation between RAP and the noninvasiveTTE measure of IVCD is affected by the ultrasono-graphic methodology used in patients who aremechanically ventilated and critically ill. Our datasuggests that IVCD-MM can be obtained easily at thebedside of patients, in the supine position, and maybe used as an alternative method to estimate rightventricular preload in patients under positive pres-sure ventilation. However, in individual patients,this measurement must be evaluated cautiously asIVCD value alone does not necessarily reflects RAP.
We are grateful for the translation support provided byAngela Frei.
REFERENCES
1. Jellinek H, Krafft P, Fitzgerald RD, Schwarz S, Pinsky MR.Right atrial pressure predicts hemodynamic response to apneicpositive airway pressure. Crit Care Med 2000;28:672-8.
2. Magder S. More respect for the CVP. Intensive Care Med1998;24:651-3.
3. Agee KR, Balk RA. Central venous catheterization in thecritically ill patient. Crit Care Clin 1992;8:677-86.
Figure 2 Left, Correlation between right atrial pressure (RAP) in cm H2O and inferior vena cava diameter(IVCD)(mm) measured using M-mode (IVCD-MM). Linear regression analysis and its 95% CI are shown.Regression equation is written mathematically on figure. Right, Correlation between RAP in cm H2O andIVCD (mm) measured using 2-dimensional view (IVCD-2D). Linear regression analysis and its 95% CIare indicated. RAPcal, calculate RAP.
Figure 3 Bland-Altman analysis of agreement betweeninferior vena cava diameter (IVCD) (mm) measured usingM-mode (IVCD-MM) and IVCD (mm) measured using2-dimensional view (IVCD-2D). Middle solid line indicatesaverage differences between 2 methods (1.6 mm), whereasouter dashed lines represent 2 SD. Circles correspond topatients.
Journal of the American Society of Echocardiography948 Bendjelid et al September 2002
67
4. Zarshenas Z, Sparschu RA. Catheter placement and misplace-ment. Crit Care Clin 1994;10:417-36.
5. Moreno FL, Hagan AD, Holmen JR, Pryor TA, StricklandRD, Castle CH. Evaluation of size and dynamics of theinferior vena cava as an index of right-sided cardiac function.Am J Cardiol 1984;53:579-85.
6. Mintz GS, Kotler MN, Parry WR, Iskandrian AS, Kane SA.Real-time inferior vena caval ultrasonography: normal andabnormal findings and its use in assessing right-heart function.Circulation 1981;64:1018-25.
7. Kircher BJ, Himelman RB, Schiller NB. Noninvasive estima-tion of right atrial pressure from the inspiratory collapse of theinferior vena cava. Am J Cardiol 1990;66:493-6.
8. Ommen SR, Nishimura RA, Hurrell DG, Klarich KW. Assess-ment of right atrial pressure with 2-dimensional and Dopplerechocardiography: a simultaneous catheterization and echo-cardiographic study. Mayo Clin Proc 2000;75:24-9.
9. Lichtenstein D, Jardin F. Appreciation non invasive de lapression veineuse centrale par la mesure echographique ducalibre de la veine cave inferieure en reanimation. Rean Urg1994;3:79-82.
10. Jue J, Chung W, Schiller NB. Does inferior vena cava sizepredict right atrial pressures in patients receiving mechanicalventilation? J Am Soc Echocardiogr 1992;5:613-9.
11. Nagueh SF, Kopelen HA, Zoghbi WA. Relation of mean rightatrial pressure to echocardiographic and Doppler parametersof right atrial and right ventricular function. Circulation 1996;93:1160-9.
12. Ramsay MA, Savege TM, Simpson BR, Goodwin R. Con-trolled sedation with alphaxalone-alphadolone. BMJ 1974;2:656-9.
13. Bland JM, Altman DG. Statistical methods for assessing agree-
ment between two methods of clinical measurement. Lancet1986;1:307-10.
14. Jellinek H, Krenn H, Oczenski W, Veit F, Schwarz S, Fitzger-ald RD. Influence of positive airway pressure on the pressuregradient for venous return in humans. J Appl Physiol 2000;88:926-32.
15. Pinsky MR. Cardiovascular effects of ventilatory support andwithdrawal. Anesth Analg 1994;79:567-76.
16. Mitaka C, Nagura T, Sakanishi N, Tsunoda Y, Amaha K.Two-dimensional echocardiographic evaluation of inferiorvena cava, right ventricle, and left ventricle during positive-pressure ventilation with varying levels of positive end-expira-tory pressure. Crit Care Med 1989;17:205-10.
17. Natori H, Tamaki S, Kira S. Ultrasonographic evaluation ofventilatory effect on inferior vena caval configuration. Am RevRespir Dis 1979;120:421-7.
18. Nakao S, Come PC, McKay RG, Ransil BJ. Effects of posi-tional changes on inferior vena caval size and dynamics andcorrelations with right-sided cardiac pressure. Am J Cardiol1987;59:125-32.
19. Jullien T, Valtier B, Hongnat JM, Dubourg O, Bourdarias JP,Jardin F. Incidence of tricuspid regurgitation and vena cavalbackward flow in mechanically ventilated patients: a colorDoppler and contrast echocardiographic study. Chest 1995;107:488-93.
20. Mohiaddin RH, Wann SL, Underwood R, Firmin DN, ReesS, Longmore DB. Vena caval flow: assessment with cine MRvelocity mapping. Radiology 1990;177:537-41.
21. LaMantia KR, O’Connor T, Barash PG. Comparing methodsof measurement: an alternative approach. Anesthesiology1990;72:781-3.
Journal of the American Society of EchocardiographyVolume 15 Number 9 Bendjelid et al 949
68
DOI: 10.1378/chest.128.5.3639 2005;128;3639-3640 ChestKarim Bendjelid
Right Atrial Pressure: Determinant or Result of Change in Venous Return?
This information is current as of November 24, 2005
http://www.chestjournal.org/cgi/content/full/128/5/3639located on the World Wide Web at:
The online version of this article, along with updated information and services, is
ISSN: 0012-3692. may be reproduced or distributed without the prior written permission of the copyright holder. 3300 Dundee Road, Northbrook IL 60062. All rights reserved. No part of this article or PDFpublished monthly since 1935. Copyright 2005 by the American College of Chest Physicians, CHEST is the official journal of the American College of Chest Physicians. It has been
at Bibliotheque Faculte Medecine Geneve on November 24, 2005 www.chestjournal.orgDownloaded from
69
Right Atrial Pressure*Determinant or Result of Change in VenousReturn?
Karim Bendjelid, MD, MS
According to the concept of Guyton, cardiac output is largely controlled by venous return, whichis determined by the difference between mean systemic venous pressure and right atrialpressure. In the analysis of the venous return curve, other authors have suggested that rightatrial pressure is the dependent variable and venous return is the independent variable (rightatrial pressure decreased because cardiac output increased). The present report analyzes thishistorical debate, which has already lasted > 50 years. (CHEST 2005; 128:3639–3640)
“There are two kinds of truth, small truth and great truth.You can recognize a small truth because its opposite is afalsehood. The opposite of a great truth is another truth”
Niels Bohr (1885–1962)
O ne of the definitions of physiology is that it is“the science of how the body works.” The key to
obtaining a full understanding of the human circu-latory function is the determination of its autoregu-lation. In the steady state, the traditional teaching isthat pulsatile BP is the result of cardiac function(output), and vascular structure and function (largevessel compliance and peripheral arterial resistance).As cardiac output must equal venous return, adecrease in cardiac output means a decrease invenous return.1 Since this has been taught to threegenerations of intensivists, venous return is definedas the result of a constant mean circulatory pressure(ie, pressure under the condition of no flow) and anindependently variable “back pressure” right atrialpressure.2 The venous return-right atrial pressureillustration was used to argue that venous returnincreases because right atrial pressure decreases,
given a constant mean circulatory pressure.3 Statedin this way, right atrial pressure-mean circulatorypressure is the gradient for venous return (drivingforce).3–5
After a theoretical analysis, with the developmentof a mathematical model, Levy6 questioned thestatement that venous return increases because of adecrease in right atrial pressure. Using a mathemat-ical model, he suggested that in the analysis of thevenous return curve, right atrial pressure is a depen-dent variable and venous return is an independentvariable (ie, right atrial pressure decreased becausecardiac output increased).6 Experimental animal re-sults have also come to the same conclusion.7 EvenGuyton4 noted that in the animal an inverse changein right atrial pressure was observed when he in-duced a change in the cardiac output by use of anartificial pump, in the absence of the collapsible tube(ie, Starling resistor). In the absence of cardiacdysfunction, venous return is more essential in de-termining cardiac output that the pump itself. How-ever, when venous return and cardiac output are notidentical (for short periods of time), and the totalvascular volume is fixed, the difference is made up bya reciprocal exchange of volume between compliantcompartments.8 In this situation, the illustration ofvenous return-right atrial pressure that was used toargue that venous return increases because rightatrial pressure decreases, given a constant meancirculatory pressure,3 could be questioned. Indeed,the switch in vascular volume between compliant
*From the Surgical Intensive Care Division, Geneva UniversityHospitals, Geneva, Switzerland.Manuscript received March 29, 2005; revision accepted March30, 2005.Reproduction of this article is prohibited without written permissionfrom the American College of Chest Physicians (www.chestjournal.org/misc/reprints.shtml).Correspondence to: Karim Bendjelid, MD, MS, Chef de CliniqueScientifique, Surgical Intensive Care Division, Geneva UniversityHospitals, CH-1211 Geneve 14, Switzerland; e-mail: [email protected]
critical care reviews
www.chestjournal.org CHEST / 128 / 5 / NOVEMBER, 2005 3639
at Bibliotheque Faculte Medecine Geneve on November 24, 2005 www.chestjournal.orgDownloaded from
70
compartments affects mean circulatory pressure,right atrial pressure, the resistance to venous return,and ultimately venous return.9 Stated in this way,right atrial pressure-mean circulatory pressureshould be the pressure gradient caused by flowrather than the gradient for venous return.6
When cardiac output increases, venous pressuredecreases because the venous reservoir is depleted.As stated by Levy,6 the development of a mathemat-ical model is an abstraction, and hence the assign-ment of dependent and independent variables maybe arbitrary. Thus, right atrial pressure may be theconsequence of the cardiac flow value around thecircuit. In a review article, Tyberg9 demonstratedhow changing venous tone modulates cardiac outputboth in physiologic conditions and in disease states.Preferring the interpretation of Levy,6 Tyberg9 hasextended his concepts and developed a modifiedpressure-volume mode of circulation.9 Also, in thehope of clarifying this issue, Brengelmann8 reexam-ined this question through a review of the originalexperiments on venous return. He emphasized thefact that Guyton et al5 did not record venous returnin dynamic states but that their data were all takenfrom steady states. In a different experimental prep-aration (from the one employed by Guyton et al5), hemaintained a fixed blood volume to illustrate theconsequences of and the differences between dy-namic and steady-state conditions.8 This experimentindicated that an increase in right atrial pressurecauses increased cardiac output in the cardiac sub-division and that an increase in cardiac output causesdecreased right atrial pressure in the vascular subdi-vision.
The role of the normal heart in regulating cardiacoutput is to lower right atrial pressure, allowingbetter drainage of blood from the compliant veinsand venules,1,10 which means that venous return andright atrial pressure are dependent variables. Thisstatement is in agreement with the relatively morerecent point of view given by Guyton when review-ing the article by Levy6 (see the editors’ note at theend of the article). For Guyton, the question “is rightatrial pressure the stimulus (independent variable)and the cardiac flow the response (dependent vari-able) or vice versa” is not a good question. Theindependent variables are such factors as heart rate,contractility, and the resistance and capacitance ofeach segment of the circulation.11 Considering re-
flexes and hormones, even the independent variablescited above become dependent variables. Therefore,both right atrial pressure and venous return aredependent variables,11 and both may be displayed onthe horizontal axis or the vertical axis.
When teaching, the senior lecturer has to choose aparticular way of describing the relationship betweentwo dependent variables in order to avoid confusingmedical students.12 We may expect that the questionof what would happen to cardiac output if the venouspressure value changed equals what would happen tothe venous pressure if the cardiac output changed, asthe two variables are dependent and interdependent.Therefore, we imagine that Guyton has not imposedplots of these open-loop relationships of two vari-ables. He has preferred one question in regard toanother. He has made the choice of a teacher.13
References1 Magder S. More respect for the CVP. Intensive Care Med
1998; 24:651–6532 Vignon P. Evaluation of fluid responsiveness in ventilated
septic patients: back to venous return. Intensive Care Med2004; 30:1699–1701
3 Guyton AC, Lindsey AW, Kaufmann BN. Effect of meancirculatory filling pressure and other peripheral circulatoryfactors on cardiac output. Am J Physiol 1955; 180:463–468
4 Guyton AC. Determination of cardiac output by equatingvenous return curves with cardiac response curves. PhysiolRev 1955; 35:123–129
5 Guyton AC, Lindsey AW, Abernathy B, et al. Venous returnat various right atrial pressures and the normal venous returncurve. Am J Physiol 1957; 189:609–615
6 Levy MN. The cardiac and vascular factors that determinesystemic blood flow. Circ Res 1979; 44:739–747
7 Grodins FS, Stuart WH, Veenstra RL. Performance charac-teristics of the right heart bypass preparation. Am J Physiol1960; 198:552–560
8 Brengelmann GL. A critical analysis of the view that rightatrial pressure determines venous return. J Appl Physiol 2003;94:849–859
9 Tyberg JV. How changes in venous capacitance modulatecardiac output. Pflugers Arch 2002; 445:10–17
10 Magder S, De Varennes B. Clinical death and the measure-ment of stressed vascular volume. Crit Care Med 1998;26:1061–1064
11 Guyton AC, Coleman TG, Granger HJ. Circulation: overallregulation. Annu Rev Physiol 1972; 34:13–46
12 Michael JA. Students’ misconceptions about perceived phys-iological responses. Am J Physiol 1998; 274:S90–98
13 Hall JE, Cowley AW Jr, Bishop VS, et al. In memoriam:Arthur C. Guyton (1919–2003). Physiologist 2003; 46:126–128
3640 Critical Care Reviews
at Bibliotheque Faculte Medecine Geneve on November 24, 2005 www.chestjournal.orgDownloaded from
71
DOI: 10.1378/chest.128.5.3639 2005;128;3639-3640 ChestKarim Bendjelid
Right Atrial Pressure: Determinant or Result of Change in Venous Return?
This information is current as of November 24, 2005
& ServicesUpdated Information
http://www.chestjournal.org/cgi/content/full/128/5/3639figures, can be found at: Updated information and services, including high-resolution
References
Lhttp://www.chestjournal.org/cgi/content/full/128/5/3639#BIBat: This article cites 13 articles, 6 of which you can access for free
Permissions & Licensing
http://www.chestjournal.org/misc/reprints.shtmltables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,
Reprints http://www.chestjournal.org/misc/reprints.shtml
Information about ordering reprints can be found online:
Email alerting servicesign up in the box at the top right corner of the online article. Receive free email alerts when new articles cite this article
Images in PowerPoint format
article figure for directions. teaching purposes in PowerPoint slide format. See any online Figures that appear in CHEST articles can be downloaded for
at Bibliotheque Faculte Medecine Geneve on November 24, 2005 www.chestjournal.orgDownloaded from
72
Innovative Methodology
The respiratory change in preejection period: a new methodto predict fluid responsiveness
Karim Bendjelid, Peter M. Suter, and Jacques A. RomandDivision of Surgical Intensive Care, Geneva University Hospitals, CH-1211 Geneva 14, Switzerland
Submitted 30 April 2003; accepted in final form 5 September 2003
Bendjelid, Karim, Peter M. Suter, and Jacques A. Romand.The respiratory change in preejection period: a new method to predictfluid responsiveness. J Appl Physiol 96: 337–342, 2004; 10.1152/japplphysiol.00435.2003.—The accuracy and clinical utility of pre-load indexes as bedside indicators of fluid responsiveness in patientsafter cardiac surgery is controversial. This study evaluates whetherrespiratory changes (�) in the preejection period (PEP; �PEP) predictfluid responsiveness in mechanically ventilated patients. Sixteen post-coronary artery bypass surgery patients, deeply sedated under me-chanical ventilation, were enrolled. PEP was defined as the timeinterval between the beginning of the Q wave on the electrocardio-gram and the upstroke of the radial arterial pressure. �PEP (%) wasdefined as the difference between expiratory and inspiratory PEPmeasured over one respiratory cycle. We also measured cardiacoutput, stroke volume index, right atrial pressure, pulmonary arterialocclusion pressure, respiratory change in pulse pressure, systolicpressure variation, and the �down component of SPV. Data weremeasured without positive end-expiratory pressure (PEEP) and afterapplication of a PEEP of 10 cmH2O (PEEP10). When PEEP10 induceda decrease of �15% in mean arterial pressure value, then measure-ments were re-performed before and after volume expansion. Volumeloading was done in eight patients. Right atrial pressure and pulmo-nary arterial occlusion pressure before volume expansion did notcorrelate with the change in stroke volume index after the fluidchallenge. Systolic pressure variation, �PEP, �down, and change inpulse pressure before volume expansion correlated with stroke vol-ume index change after fluid challenge (r2 � 0.52, 0.57, 0.68, and0.83, respectively). In deeply sedated, mechanically ventilated pa-tients after cardiac surgery, �PEP, a new method, can be used topredict fluid responsiveness and hemodynamic response to PEEP10.
fluid resuscitation; heart-lung interactions; monitoring
AFTER CARDIAC SURGERY, INTRAVENOUS fluid administration is auniversally accepted treatment for hypotension occurring dur-ing positive pressure ventilation. Nevertheless, vigorous fluidresuscitation carries the risk of generating volume overload andpulmonary edema. To prevent such complications, severalindexes have been used to assess preload (6). However, theaccuracy of filling pressures, such as right atrial (Pra) and/orpulmonary arterial occlusion pressure (Ppao) to estimate car-diac filling, have been questioned in patients after cardiacsurgery (28). Thus indexes able to unmask preload dependencyand to predict increase in cardiac output with volume expan-sion are actively searched (14). In deeply sedated, mechani-cally ventilated patients after cardiac surgery, dynamic in-dexes, such as systolic pressure variation (SPV) and strokevolume variations (SVV), have been demonstrated to be moreaccurate to predict fluid responsiveness than filling pressures(23, 27, 28).
The preejection period (PEP), the time from the onset ofventricular depolarization to the beginning of left ventricularejection, is a systolic time interval that allows assessment ofventricular function (36). More than 30 years ago, Weissler etal. (37) measured PEP with simultaneous electrocardiogram(ECG), phonocardiogram, and carotid arterial pulse tracing.Presently, PEP can be obtained by simultaneous ECG record-ing and arterial pressure wave tracing, which are often moni-tored in critically ill patients (2). Even if PEP depends slightlyon afterload and cardiac contractility, it always decreases witha greater preload (36). Interestingly, in mechanically ventilatedpatients, we recently observed that expiratory (PEPE) andinspiratory PEP (PEPI), measured at the lower and highersystolic pressure value on arterial pressure tracing over onerespiratory cycle, were of different values (3). By analogy withthe concept of positive pressure ventilation-induced SVV (16),we hypothesized that the respiratory change (�) in PEP (�PEP)depends predominantly on the change in ventricular preloadand is minimally influenced by contractility or afterload. Ac-cordingly, with this hypothesis, in a preliminary study, �PEPwas found to be a good predictor of fluid responsiveness (4).The aim of the present study was to test whether �PEP predictshemodynamic changes induced by positive end-expiratorypressure (PEEP) and volume infusion in patients after coronaryartery bypass graft. �PEP was also compared with otherclinically used preload indexes [Pra, Ppao, SPV, �down (com-ponent of SPV)] and respiratory changes in pulse pressure(�PP) (20).
METHODS
The study was approved by our institutional ethics committee, andall patients gave written consent the day before their surgery. Onlypatients scheduled for coronary artery bypass graft surgery wereprospectively screened. Exclusion criteria (for eligibility into thestudy) were decreased preoperative left ventricular ejection fraction(�45% assessed by echocardiography, isotopic or angiographic ven-triculography), valve diseases, other associated cardiac surgery,and/or history of chronic obstructive pulmonary disease. Postopera-tively, patients without a pulmonary artery catheter, without sinusrhythm, with hemodynamic instability, and/or bleeding �100 ml/hwere also excluded.
Perioperative management. Perioperative management was per-formed as previously described (29). After surgery, patients weretransferred to the Surgical Intensive Care Unit. On arrival, sedationand analgesia were provided by continuous infusion of midazolamand morphine, titrated for a Ramsay score of 6 (25). The patients wereplaced on mechanical ventilation (Evita 4, Drager, Lubeck, Germany).Initial ventilator settings in controlled mechanical ventilation modewere tidal volume of 8–10 ml/kg body wt and respiratory rate of 12
Address for reprint requests and other correspondence: K. Bendjelid, Divi-sion of Surgical Intensive Care, Univ. hospital of Geneva, CH-1211 Geneva 14(E-mail: [email protected]).
The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
J Appl Physiol 96: 337–342, 2004;10.1152/japplphysiol.00435.2003.
8750-7587/04 $5.00 Copyright © 2004 the American Physiological Societyhttp://www.jap.org 337
73
breaths/min. Both were adjusted to maintain an arterial PCO2 at 40 �5 Torr (5.3 � 0.5 kPa). The inspired oxygen fraction was adjusted foran arterial oxygen saturation of �92%. Inspiratory-to-expiratory ratiowas 1:2. Body temperature, lead II ECG, and urine output weremonitored throughout the postoperative period. The patients wereobserved for at least 2 h to confirm hemodynamic stability, which wasdefined as a �10% change in the hemodynamic (heart rate, meanarterial pressure, and cardiac output), no clinically relevant bleeding(�100 ml/h), and normal body temperature. A chest radiograph wasobtained and assessed before data collection to define the correctposition of the endotracheal tube, pulmonary artery, and centralvenous catheters and the position of the surgical drains and also toconfirm the absence of cardiopulmonary abnormalities (grossly en-larged mediastinal silhouette, pleural effusions, or pneumothorax). Allpatients received a continuous intravenous infusion of 0.9% NaCl ata rate of 65 ml/h.
Hemodynamic monitoring. All pressure transducers were refer-enced to midchest. The correct position of the pulmonary arterycatheter tip in West’s zone 3 was checked by using a methodpreviously described (33). Mean cardiac output was estimated byaveraging triplicate injections of 10 ml of 0.9% NaCl at roomtemperature delivered randomly during the respiratory cycle. MeanPra and mean Ppao were measured at end-expiration. In addition, fluidoutput (chest drains, urine, and nasogastric losses) was measuredduring the study protocol.
Respiratory change in PEP. PEP was defined as the time intervalbetween the beginning of the Q wave on the ECG and the upstroke ofthe invasive radial arterial pressure curve. PEP was measured by usingan electronic tool named Callipers (Agilent Technologies, M3150A)
before analysis of SPVs so as not to be influenced by the results.�PEP (in %) was defined as the difference between PEPE and PEPI
measured over one respiratory cycle. PEPE and PEPI were done,respectively, at the minimal and at the maximal systolic pressure valueon the arterial pressure trace over one respiratory cycle. Each PEPvalue was an average of three measurements. These measurementswere repeated during three different respiratory cycles (total of 9measurements). The measurements were done with a speed of record-ing of 50 mm/s, and the values were averaged. �PEP (%) wascalculated as 100 � (PEPE � PEPI)/[(PEPE � PEPI)/2] (5) (Figs. 1and 2). The interobserver variability in measuring �PEP has beendetermined by a “blinded” fashion, with an observer (Dr. ChristopheAbbeg) unaware of the particular significance of respiratory changesin arterial pressure.
SPV measurements. The systemic arterial blood pressure variationcurve, obtained from the radial artery catheter, was recorded (AgilentTechnologies, M3150A). Pressure waveform analysis was performedoffline on a paper chart with a speed of 12.5 mm/s and a pressure scaleadjusted to the systemic systolic pressure value. The SPV (24) wasdetermined from the chart. The value of the systolic blood pressuremeasured after an end-expiratory pause period of 6–10 s was used asa reference pressure to measure the �down (24, 32). Because thearterial pressure may have an additional nonrespiratory, low-fre-quency fluctuation (Mayer waves) (1), the �down was determinedduring the first three respiratory cycles that immediately preceded theapnea period. Pulse pressure (21) was measured offline with pressurescale adjusted to the systemic systolic pressure value. �PP (in %) wascalculated as previously described (21).
Fig. 1. Systolic pressure variation (SPV) af-ter 1 mechanical breath followed by an end-expiratory pause. Reference line permits themeasurement of change (�) in up (�up) and�down. Thick lines, maximal and minimalpulse pressure (PP). PEPI, inspiratorypreejection period (PEP); PEPE, expiratoryPEP; SAP, systolic arterial pressure; ECG,electrocardiogram.
Fig. 2. Measurement on positive end-expi-ratory pressure (PEEP) in hypotensive pa-tient with tachycardia. �PEP (%), using cal-ipers, equals 100 � (120 � 100)/[(120 �100)/2] � 18%. Note that SPV � 8 mmHg,�down � 6 mmHg, and �PP (%) � 19%.Bold lines, maximal and minimal PP. PA,arterial pressure; Resp, respiration. Scale: 1mm � 3 mmHg.
Innovative Methodology
338 �PEP AND FLUID RESPONSIVENESS
J Appl Physiol • VOL 96 • JANUARY 2004 • www.jap.org
74
Study protocol. All studies were performed in deeply sedated andnonspontaneously breathing patients (Ramsay 6) in the supine posi-tion. If patients were receiving vasoactive drugs, the rate of adminis-tration was not changed. The study protocol consisted of two sequen-tial ventilatory steps of 30 min each: controlled mechanical ventilationwith PEEP � 0 [zero end-expiratory pressure (ZEEP)], and controlledmechanical ventilation with PEEP � 10 cmH2O (PEEP10) withoutchanging any other ventilatory settings. When PEEP10 induced areduction in mean arterial pressure of �15% after the 25-min period,a hemodynamic measurement was then obtained, immediately fol-lowed by 0.9% NaCl (500 ml � total quantity of fluid of chest drains,urine, and nasogastric output in milliliters), given over 25 min, still onPEEP10. A third set of hemodynamic measurements was then ob-tained.
Data recording. At the end of each period, all of the followingvariables were recorded from the bedside monitor (Agilent Technol-ogies, M3150A): body temperature; mean dilution cardiac output;heart rate; calculated stroke volume; systolic, diastolic, and meanarterial and pulmonary arterial pressures; Pra; Ppao; SPV; �down;�PP; and �PEP. Mean dilution cardiac output, estimated by averagingtriplicate injections, was used for statistical analysis. Arterial andmixed-venous blood gases were also simultaneously measured (Stat-profile Ultra, Nova Biomedical, Waltham, MA), and standard calcu-lated variables were obtained from hemodynamic and blood-gas data.Ventilator settings (respiratory rate, tidal volume, inspiratory-to-expi-ratory ratio, inspired oxygen fraction, PEEP), peak and mean airwaypressures, and auto-PEEP were recorded.
Statistical analysis. The data were analyzed by using Graph PadPrism (Graph pad software version 3, San Diego, CA) for the personalcomputer. The nonparametric Mann-Whitney test was used to com-pare the effects of PEEP and volume expansion on hemodynamic andrespiratory parameters. Additionally, the same test was used to com-pare the interobserver variability in measuring �PEP. Correlationswere obtained by using regression analysis. All values are expressedas means � SD, and P � 0.05 was considered statistically significant.
RESULTS
Twenty-five patients gave informed consent preoperatively.Postoperatively, nine were excluded before starting the studyprotocol. Five patients presented postoperative atrial fibrilla-tion, three were not equipped with a pulmonary arterial cath-
eter, and one patient was hemodynamically unstable. Sixteenpatients were included in the final analysis, and all of themtolerated the experimental protocol well. Demographic andpreoperative characteristics are presented in Table 1. The meancardiopulmonary bypass duration was 107 � 42 min, and theaortic cross-clamping time was 73 � 33 min. Catecholamineinfusions were required in four patients for cardiopulmonarybypass weaning (dobutamine, n � 1, and/or norepinephrine,n � 3). However, during the study protocol, in the intensivecare unit, among these four patients, only two still requiredcatecholamine infusions (dobutamine: n � 1; 8 g�kg�1�min�1,and norepinephrine: n � 1; 0.1 g�kg�1�min�1). Hemodynamicvariables, respiratory, airway pressures, and gas exchangevalues are shown in Tables 2 and 3. In all patients, systolicarterial pressures were higher during the inspiratory than dur-ing the expiratory period. The interobserver variability inmeasuring �PEP was 8% (P � 0.89, Mann-Whitney test).
Only eight patients presented a �15% decrease in meanarterial pressure after PEEP10. Among these eight patients, onepatient was under norepinephrine infusion (thick solid line inFig. 3). Volume infusion produced an increase in stroke vol-ume index (SVI) from 29 � 4 to 34 � 4 ml/m2 (P � 0.0002),and �PEP decreased from 11 � 3 to 5 � 3% (P � 0.002) (Fig.3). No difference in response to volume infusion was seen inthe patient receiving norepinephrine compared with the otherpatients.
Static pressures predicting response to PEEP and fluidresponsiveness. The static pressures predicting the response toPEEP and the fluid responsiveness were measured in ZEEP,Pra, and Ppao and correlated with the PEEP10-induced changein cardiac index (in %) (r2 � 0.34, P � 0.02; r2 � 0.47, P �0.003; respectively). Pra and Ppao before volume expansiondid not correlate with changes in SVI after volume expansion(P � 0.7 and P � 0.3, respectively).
SPV and PP variation prediction of response to PEEP andfluid responsiveness. The SPV and PP variation prediction ofresponse to PEEP and the fluid responsiveness were measuredin ZEEP, SPV, �down (the component of SPV), and �PP and
Table 1. Patient characteristics
Patient No. Age, yrGender(M/F) BS, m2 LVEF, % CGB, no. PO2/FIO2 VT, ml/kg
CT,ml/cmH2O
Outcome(S/D)
1 66 F 1.81 80 3 433 6 29 S2 68 F 1.71 70 3 314 8 37 S3 56 M 2.09 55 2 343 5 32 S4 75 M 1.95 55 2 438 8 51 S5 69 F 1.73 56 1 436 8 43 S6 68 M 1.72 65 3 360 10 34 S7 59 M 2.09 53 1 425 7 41 S8 81 M 1.84 76 1 280 9 69 S9 62 M 2.15 67 3 400 6 31 S
10 53 M 1.97 52 2 343 7 27 S11 53 M 1.68 45 2 334 9 26 S12 52 M 1.96 60 3 272 8 33 S13 60 M 2.06 74 3 282 7 41 S14 63 M 1.77 75 4 328 9 30 S15 81 F 1.64 71 2 337 8 26 S16 68 M 1.93 60 4 285 7 34 S
Mean�SD 64.6�8.54 1.88�0.17 63.4�10.1 2.4�0.9 350�59 8�1 37�11Ratio 4/12 16/16
M, male; F, female; BS, body surface; LVEF, left ventricular ejection fraction; CGB, coronary grafts bypassed; PO2/FIO2, ratio of PO2 to inspired O2 fraction;VT, tidal volume; CT, static compliance of the respiratory system; S/D, survived/died.
Innovative Methodology
339�PEP AND FLUID RESPONSIVENESS
J Appl Physiol • VOL 96 • JANUARY 2004 • www.jap.org
75
correlated with the PEEP10-induced change in cardiac index (in%) (r2 � 0.28, P � 0.03; r2 � 0.52, P � 0.002; r2 � 0.63, P �0.0002, respectively). The correlations among SPV, �down,and �PP before volume expansion correlated with changes inSVI after volume expansion (P � 0.04, P � 0.01, and P �0.001, respectively; see Fig. 4).
Respiratory change in PEP-predicting response to PEEPand fluid responsiveness. The respiratory change in PEP-predicting response to PEEP and the fluid responsiveness weremeasured in ZEEP and �PEP and correlate with the PEEP10-induced change in cardiac index (in %) (r2 � 0.53, P � 0.001).�PEP before volume expansion correlated with change in SVIafter volume expansion (P � 0.03) (see Fig. 4).
DISCUSSION
The present study shows a good correlation between �PEPvalue before volume expansion and SVI increase after fluidchallenge in mechanically ventilated cardiac surgery patientson PEEP. In addition, �PEP at ZEEP predicts hemodynamicresponse to PEEP10. However, �down (the component of SPV)and �PP were better indexes to predict fluid responsivenessthan �PEP.
In mechanically, deeply sedated, ventilated patients, positivepressure ventilation cyclically increases intrathoracic pressureand lung volume. Both reduce venous return, alter cardiacpreload, and decrease stroke volume. Thus left ventricularstroke volume varies cyclically, being maximal during me-chanical breath and minimal during expiration. During hypo-volemia, the greatest mechanical breath-induced SVVs areobserved (16). The respiratory change in stroke volume resultsin SPV and �PP. These two indexes have been shown toidentify decreased preload hypotension and to distinguish be-tween responders and nonresponders to fluid challenge indifferent patient populations (12, 20, 21, 32).
Because PEP depends on preload, afterload, and contractility(36), it is related to stroke volume. Indeed, in an animal study,Wallace et al. (35) demonstrated that increasing stroke volumeshortens PEP. Several human studies have also found thatdecreased PEP after fluid challenge is associated with increasein stroke volume (10, 13, 18). Interestingly, Brundin et al. (9)demonstrated that intermittent positive pressure ventilationincreased PEP by the reduction of venous return and thusstroke volume. In the present study, �PEP was used as anindex of preload responsiveness, with the hypothesis that, as
Table 2. Effects of PEEP on hemodynamic and respiratoryparameters (16 patients)
ZEEP PEEP10 P
HR, beats/min 88�12 86�12 0.08MAP, mmHg 69�9 67�9 0.2MPAP, mmHg 21�4 23�3 0.004*Pra, mmHg 11�3 13�3 0.004*Ppao, mmHg 12�3 14�2 0.004*�PEP, % 8�4 7�5 0.4SPV, mmHg 10�4 9�4 0.4�Down, mmHg 5�3 5�5 0.9�PP, % 12�7 12�6 0.5CI, 1�min�1�m�2 2.9�0.3 2.6�0.4 �0.0001*SVRI, dyn�s�1�cm�5 1,574�246 1,661�282 0.08PVRI, dyn�s�1�cm�5 250�85 269�87 0.41DO2, ml/min 730�160 684�181 0.009*PaO2, Torr 93�22.5 99�15 0.1PaCO2, Torr 39.8�5.3 39.9�6 0.9Plateau, cmH2O 18�4 24�3 �0.0001*CT, ml/cmH2O 37�11 45�9 0.013*Auto-PEEP, cmH2O 1.7�1.4 0.8�0.8 0.06
Values are means � SD. ZEEP, zero end-expiratory pressure; PEEP10, 10cmH2O of positive end-expiratory pressure; HR, heart rate; MAP, meanarterial pressure; MPAP, mean pulmonary arterial pressure; Pra, right atrialpressure; Ppao, pulmonary arterial occlusion pressure; �PEP, respiratorychange in preejection period; SPV, systolic pressure variation; �down, deltadown; �PP, respiratory variation in pulse pressure; CI, cardiac index; SVRI,systemic vascular resistance index; PVRI, pulmonary vascular resistanceindex; DO2, oxygen delivery; PaO2, arterial PO2; PaCO2, arterial PCO2; Plateau,end-inspiratory airway pressure. *P � 0.05 (Mann-Whitney test).
Table 3. Effects of PEEP and volume infusion onhemodynamic and respiratory parameters in eight patients
ZEEP PEEP10
VE onPEEP10 P/P
HR, beats/min 90�10 87�11 82�9 0.6/0.3MAP, mmHg 69�12 57�6 69�5 0.02*/0.0007†MPAP, mmHg 19�4 22�3 23�3 0.11/0.5Pra, mmHg 10�3 12�3 13�3 0.2/0.5Ppao, mmHg 10�3 13�2 15�3 0.03*/0.1�PEP, % 8�2 11�3 5�3 0.02*/0.002†SPV, mmHg 11�3 10�4 7�2 0.6/0.07�Down, mmHg 6�3 7�4 3�2 0.6/0.02†�PP, % 18�4 14�4 7�3 0.06/0.001†CI,
1�min�1�m�2 2.9�0.3 2.5�0.4 2.9�0.2 0.04*/0.02†SVRI,
dyn�s�1�cm�5 1,615�309 1,600�349 1,620�361 0.9/0.9PVRI,
dyn�s�1�cm�5 248�73 305�95 248�99 0.2/0.2DO2, ml/min 691�143 596�159 611�120 0.2/0.8PaO2, Torr 96�28.6 98�13 112�15 0.8/0.06†PaCO2, Torr 42�5.7 42�6 41�3 0.9/0.7Plateau, cmH2O 16�4 24�3 25�3 0.0005*/0.5CT, ml/cmH2O 41�13 44�5 43�5 0.5/0.7Auto-PEEP,
cmH2O 1.5�0.9 1�1 0.7�0.5 0.3/0.5
Values are means � SD. VE, volume expansion. *P � 0.05, PEEP10/ZEEP(Mann-Whitney test); †P � 0.05, VE/PEEP10 (Mann-Whitney test).
Fig. 3. �PEP (%) on zero end-expiratory pressure (ZEEP), on 10-cmH2OPEEP (PEEP10) [before volume expansion (VE)], and after VE. Thick solidtrace, patient requiring norepinephrine infusion during VE. Mann-Whitney testwas found significant. Thick horizontal lines � mean.
Innovative Methodology
340 �PEP AND FLUID RESPONSIVENESS
J Appl Physiol • VOL 96 • JANUARY 2004 • www.jap.org
76
for SPV and �PP, �PEP is related to positive pressure breathinduces change in ventricular stroke volume related to changein ventricular preload (17). Thus the recorded lower PEP valueduring the mechanical inspiratory phase compared with thehigher PEP during expiratory phase is in accordance with thehypothesis that �PEP is related to the respiratory change in leftventricular stroke volume. Indeed, early after mechanicalbreath, capacitance pulmonary vessels discharge into the pul-monary veins (8). This would increase left ventricular preloadat that phase of the cycle. Furthermore, an inspiratory increaseof left ventricular stroke volume, thus determining the minimalvalue of PEP, is observed secondary to the rise in left ventric-ular preload, which reflects the three heartbeats that wereincreased earlier in right ventricular preload during expiration(22). Accordingly, �PEP was a good predictor of hemody-namic response to PEEP10 (decrease in preload) and to fluidchallenge (increase in preload).
The present study confirms that Pra and Ppao before volumeexpansion do not correlate with the volume expansion-inducedchange in SVI, as already demonstrated in different patientpopulations (20, 26, 32) and after cardiac surgery (15, 28).These results could be explained by the absence of correlationbetween cardiac filling pressures and cardiac volumes in pa-tients after coronary artery bypass surgery, as demonstrated byBuhre et al. (11). However, even if �PEP were found to be agood predictor of hemodynamic response to fluid challenge,�down and �PP were better indexes to predict fluid respon-siveness (Fig. 4). Nevertheless, �down data acquisitions ne-cessitate an expiratory pause of at least 5 s, and, in the absenceof automatic bedside measurements, �PP assessment is timeconsuming. In comparison, �PEP is easily assessed by usingcalipers of a central monitor, and its calculation is rapid.Moreover, in the future, �PEP could be measured automati-cally by using a personal computer (7) and/or assessed nonin-vasively at the bedside by the thoracic electrical bioimpedancetechnique (19).
As recently published (28), another message addressed bythe present study is that dynamic indexes such as SPV, �PEP,and �PP could be used as predictors of fluid responsivenessafter cardiac surgery, even if patients are equipped with a chestdrain. Indeed, application of thoracic drainage seems to perturbminimally the physiological change in pleural pressure inducedby positive pressure ventilation.
One limitation of the study is that PEP could also beminimally influenced by afterload variations induced bypositive pressure ventilation. Indeed, increased pleural pres-sure (accompanying a positive pressure breath) may de-crease left ventricular transmural pressure (afterload) andthus increase left ventricular stroke volume. In one-half ofthe patients, the cardiac index was not affected by PEEP.Hence, we can cautiously assume that these patients were onthe flat portion of the Starling left ventricular function curve(30). In this situation, SPV, SVV, �PEP, and �PP may bedue mainly to an augmentation of the stroke volume duringthe mechanical breath, which is related to decrease inafterload and expressed by the �up of the systolic pressure(31). However, this positive pressure effect is rarely ob-served in patients with normal cardiac function after cardio-pulmonary bypass (34). Indeed, Van Trigt et al. (34) havedemonstrated that, in patients after coronary artery bypasssurgery, PEEP10 or greater produces a significant fall incardiac output, due to a decrease in preload, without achange in left ventricular contractility and afterload.
In conclusion, in patients after coronary artery bypass sur-gery, this study found �PEP to be a good predictor of hemo-dynamic response to PEEP and a reliable preload parameter forpredicting an increase in cardiac output after volume infusion.In addition, our data confirm that Pra and Ppao are of littlevalue in predicting the hemodynamic effects of volume expan-sion in cardiac surgical patients with preserved left ventricularsystolic function.
Fig. 4. Linear correlation analysis of the re-lationship between preload parameters mea-sured before VE and change in stroke volumeafter VE. A: �PEP; B: SPV; C: �down; D:�PP. P � 0.05 was considered significant.
Innovative Methodology
341�PEP AND FLUID RESPONSIVENESS
J Appl Physiol • VOL 96 • JANUARY 2004 • www.jap.org
77
ACKNOWLEDGMENTS
The authors are grateful for the support provided by Dr. Christophe Abbegin �PEP measurement (blinded investigator). This work was performed in thedivision of Surgical Intensive Care, University Hospitals of Geneva, Switzer-land. The central monitor (Agilent Technologies M3150A) was lent byAgilent-Philipps (Switzerland) during the study period.
Preliminary data have been presented as an oral communication to the VIIIWorld Congress of Intensive and Critical Care Medicine, October 2001,Sydney, and as a poster presentation to the 31th Congress of the Society ofCritical Care Medicine, 26–30 January 2002, San Diego.
REFERENCES
1. Baujard C, Ponchon P, and Elghozi JL. Effects of graded hemorrhageon short-term variability of blood pressure in conscious rats. Fundam ClinPharmacol 10: 511–517, 1996.
2. Bendjelid K, Merlani PG, Fournier G, Suter PM, and Romand JA.Estimation of the responsiveness to volume challenge by the measure ofthe preejection period in mechanically ventilated patients (Abstract). CritCare Med 28: A155, 2000.
3. Bendjelid K and Romand JA. Ventilatory changes in preejection periodis correlated to cardiac output in patients following cardiac surgery(Abstract). Intensive Care Med 27: A126, 2001.
4. Bendjelid K and Romand JA. Inspiratory to expiratory variation ofpreejection period during intermittent positive pressure ventilation pre-dicts fluid responsiveness in critically ill patients (Abstract). IntensiveCare Med 27: A51, 2001.
5. Bendjelid K and Romand JA. Respiratory changes in preejection period(PEP) predicts hemodynamic response to PEEP (Abstract). Crit Care Med29: A222, 2001.
6. Bendjelid K and Romand JA. Fluid responsiveness in mechanicallyventilated patients: a review of indices used in intensive care. IntensiveCare Med 29: 352–360, 2003.
7. Bernardi L, Calciati A, Lumina C, Licci E, and Finardi G. Automaticon-line measurement of systolic time intervals using a personal computer.J Cardiogr 16: 133–139, 1986.
8. Brower R, Wise RA, Hassapoyannes C, Bromberger-Barnea B, andPermutt S. Effect of lung inflation on lung blood volume and pulmonaryvenous flow. J Appl Physiol 58: 954–963, 1985.
9. Brundin T, Hedenstierna G, and McCarthy G. Effect of intermittentpositive pressure ventilation on cardiac systolic time intervals. ActaAnaesthesiol Scand 20: 278–284, 1976.
10. Buch J, Egeblad H, Hansen PB, Kjaergard H, Waldorff S, andSteiness E. Correlation between changes in systolic time intervals and leftventricular end-diastolic diameter after preload reduction. Non-invasivemonitoring of pharmacological intervention. Br Heart J 44: 668–671,1980.
11. Buhre W, Weyland A, Schorn B, Scholz M, Kazmaier S, Hoeft A, andSonntag H. Changes in central venous pressure and pulmonary capillarywedge pressure do not indicate changes in right and left heart volume inpatients undergoing coronary artery bypass surgery. Eur J Anaesthesiol16: 11–17, 1999.
12. Coriat P, Vrillon M, Perel A, Baron JF, Le Bret F, Saada M, andViars P. A comparison of systolic blood pressure variations and echocar-diographic estimates of end-diastolic left ventricular size in patients afteraortic surgery. Anesth Analg 78: 46–53, 1994.
13. Engel UR, Burckhardt D, and Kloppenburg P. Response of systolictime intervals to fluid challenge in uncomplicated acute myocardial in-farction. Bibl Cardiol 00: 22–26, 1976.
14. Fontes ML, Bellows W, Ngo L, and Mangano DT. Assessment ofventricular function in critically ill patients: limitations of pulmonaryartery catheterization. Institutions of the McSPI Research Group. J Car-diothorac Vasc Anesth 13: 521–527, 1999.
15. Godje O, Peyerl M, Seebauer T, Lamm P, Mair H, and Reichart B.Central venous pressure, pulmonary capillary wedge pressure and in-trathoracic blood volumes as preload indicators in cardiac surgery patients.Eur J Cardiothorac Surg 13: 533–540, 1998.
16. Gunn SR and Pinsky MR. Implications of arterial pressure variation inpatients in the intensive care unit. Curr Opin Crit Care 7: 212–217, 2001.
17. Jardin F, Farcot JC, Gueret P, Prost JF, Ozier Y, and Bourdarias JP.Cyclic changes in arterial pulse during respiratory support. Circulation 68:266–274, 1983.
18. Matsuno Y, Morioka S, Murakami Y, Kobayashi S, and Moriyama K.Mechanism of prolongation of pre-ejection period in the hypertrophied leftventricle with normal systolic function in unanesthetized hypertensivedogs. Clin Cardiol 11: 702–706, 1988.
19. Mattar JA, Shoemaker WC, Diament D, Lomar A, Lopes AC, DeFreitas E, Stella FP, and Factore LA. Systolic and diastolic timeintervals in the critically ill patient. Crit Care Med 19: 1382–1386, 1991.
20. Michard F, Boussat S, Chemla D, Anguel N, Mercat A, LecarpentierY, Richard C, Pinsky MR, and Teboul JL. Relation between respiratorychanges in arterial pulse pressure and fluid responsiveness in septicpatients with acute circulatory failure. Am J Respir Crit Care Med 162:134–138, 2000.
21. Michard F, Chemla D, Richard C, Wysocki M, Pinsky MR, Lecar-pentier Y, and Teboul JL. Clinical use of respiratory changes in arterialpulse pressure to monitor the hemodynamic effects of PEEP. Am J RespirCrit Care Med 159: 935–939, 1999.
22. Michard F and Teboul JL. Using heart-lung interactions to assess fluidresponsiveness during mechanical ventilation. Crit Care 4: 282–289,2000.
23. Ornstein E, Eidelman LA, Drenger B, Elami A, and Pizov R. Systolicpressure variation predicts the response to acute blood loss. J Clin Anesth10: 137–140, 1998.
24. Perel A, Pizov R, and Cotev S. Systolic blood pressure variation is asensitive indicator of hypovolemia in ventilated dogs subjected to gradedhemorrhage. Anesthesiology 67: 498–502, 1987.
25. Ramsay MA, Savege TM, Simpson BR, and Goodwin R. Controlledsedation with alphaxalone-alphadolone. Br Med J 2: 656–659, 1974.
26. Reuse C, Vincent JL, and Pinsky MR. Measurements of right ventric-ular volumes during fluid challenge. Chest 98: 1450–1454, 1990.
27. Reuter DA, Felbinger TW, Kilger E, Schmidt C, Lamm P, and GoetzAE. Optimizing fluid therapy in mechanically ventilated patients aftercardiac surgery by on-line monitoring of left ventricular stroke volumevariations. Comparison with aortic systolic pressure variations. Br JAnaesth 88: 124–126, 2002.
28. Reuter DA, Felbinger TW, Schmidt C, Kilger E, Goedje O, Lamm P,and Goetz AE. Stroke volume variations for assessment of cardiacresponsiveness to volume loading in mechanically ventilated patients aftercardiac surgery. Intensive Care Med 28: 392–398, 2002.
29. Romand JA, Treggiari-Venzi MM, Bichel T, Suter PM, and PinskyMR. Hemodynamic effects of synchronized high-frequency jet ventilationcompared with low-frequency intermittent positive-pressure ventilationafter myocardial revascularization. Anesthesiology 92: 24–30, 2000.
30. Starling E. The Linacre Lecture on the Law of the Heart Given atCambridge. London: Longmans, Green, 1918.
31. Szold A, Pizov R, Segal E, and Perel A. The effect of tidal volume andintravascular volume state on systolic pressure variation in ventilateddogs. Intensive Care Med 15: 368–371, 1989.
32. Tavernier B, Makhotine O, Lebuffe G, Dupont J, and Scherpereel P.Systolic pressure variation as a guide to fluid therapy in patients withsepsis-induced hypotension. Anesthesiology 89: 1313–1321, 1998.
33. Teboul JL, Besbes M, Andrivet P, Axler O, Douguet D, Zelter M,Lemaire F, and Brun-Buisson C. A bedside index assessing the reliabil-ity of pulmonary occlusion pressure measurements during mechanicalventilation with positive end-expiratory pressure. J Crit Care 7: 22–29,1992.
34. Van Trigt P, Spray TL, Pasque MK, Peyton RB, Pellom GL, ChristianCM, Fagraeus L, and Wechsler AS. The effect of PEEP on leftventricular diastolic dimensions and systolic performance following myo-cardial revascularization. Ann Thorac Surg 33: 585–592, 1982.
35. Wallace AG, Mitchell JH, Sheldon Skinner N, and Sarnoff SJ. Dura-tion of the phases of left ventricular systole. Circ Res 12: 611–619, 1963.
36. Weissler AM. Current concepts in cardiology. Systolic-time intervals.N Engl J Med 296: 321–324, 1977.
37. Weissler AM, Harris WS, and Schoenfeld CD. Bedside technics for theevaluation of ventricular function in man. Am J Cardiol 23: 577–583,1969.
Innovative Methodology
342 �PEP AND FLUID RESPONSIVENESS
J Appl Physiol • VOL 96 • JANUARY 2004 • www.jap.org
78
Pre-ejection period variations predict the fluid responsiveness ofseptic ventilated patients
Marc Feissel, MD; Julio Badie, MD; Paolo G. Merlani, MD; Jean-Pierre Faller, MD;Karim Bendjelid, MD, MS.
I n septic shock, prompt correctionof vascular volume deficits canimprove survival (1, 2), whereasexcessive volume expansion (VE)
can be deleterious (3, 4). Moreover, arecent study demonstrated that increas-ing fluid overload is associated with de-creased survival in a pediatric populationwith sepsis (5, 6).
Therefore, in septic patients withacute circulatory failure, reliable predic-tors of fluid responsiveness are needed atthe bedside (7, 8). Baseline values ofstatic variables (central venous pressure,
pulmonary artery occlusion pressure, andechocardiographic ventricular diastolicdimensions) have successively beenfound to be inaccurate predictors of vol-ume responsiveness (9–11). In contrast,dynamic variables related to positivepressure ventilation (systolic pressurevariation, respiratory changes in arterialpulse, and superior vena cava collapsibil-ity) have been demonstrated as accuratepredictors of fluid responsiveness in me-chanically ventilated septic patients (9,10, 12).
The pre-ejection period (PEP), thetime from the onset of ventricular depo-larization to the beginning of left ventric-ular ejection, is a systolic time intervalthat allows assessment of ventricularfunction (13). Currently, PEP can be ob-tained by simultaneous electrocardio-gram (ECG) recording and arterial pres-sure waves tracing. Both of these areoften monitored in critically ill patients(14). Recently we have demonstrated thatin cardiac surgical patients, the respira-
tory change in pre-ejection period(�PEP) is a reliable dynamic index for theprediction of increase in cardiac outputafter volume infusion (15).
The aim of the present study was to testwhether �PEP predicts hemodynamicchanges in response to volume infusion inpatients with septic shock. �PEP was com-pared with the respiratory changes in pulsepressure (�PP) (9). Moreover, �PEP wasmeasured using both tracing of invasivearterial pressure waves (�PEPKT) and non-invasive pulse plethysmographic wave-forms (�PEPPLET).
MATERIALS AND METHODS
The ethical committee waived the need fora written informed consent, considering theprotocol (transthoracic echocardiography andarterial pressure monitoring) to be a part ofroutine clinical practice. Only patients withseptic shock as defined by the InternationalSepsis Definitions Conference (16), equippedwith systemic arterial catheters and poten-tially needing a volume challenge, were pro-
From the Intensive Care Unit, Centre Hospitalier,Belfort, France (MF, JB, J-PF); and the Surgical Inten-sive Care Unit, Department of Anesthesiology, Phar-macology and Surgical Intensive Care, Geneva Univer-sity Hospitals, Switzerland (PGM, KB).
The authors declare no conflict of interest.This article has an online data supplement.Copyright © 2005 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000186415.43713.2F
Objectives: In septic patients with acute circulatory failure,reliable predictors of fluid responsiveness are needed at thebedside. We hypothesized that the respiratory change in pre-ejection period (�PEP) would allow the prediction of changes incardiac index following volume administration in mechanicallyventilated septic patients.
Design: Prospective clinical investigation.Setting: A ten-bed hospital intensive care unit.Patients: Patients admitted after septic shock equipped with
an arterial catheter.Interventions: Pre-ejection period (PEP)—defined as the time
interval between the beginning of the R wave on the electrocar-diogram and the upstroke of the radial arterial pressure curve(PEPKT) or the pulse plethysmographic waveforms (PEPPLET)—andcardiac index (transthoracic echocardiography-Doppler) were de-termined before and after volume infusion of colloid (8 mL·kg�1).�PEP (%) was defined as the difference between expiratory andinspiratory PEP divided by the mean of expiratory and inspiratoryvalues. Respiratory changes in pulse pressure (�PP) was alsomeasured.
Measurements and Main Results: Twenty-two volume chal-lenges were done in 20 deeply sedated patients. �PEPKT,�PEPPLET, and �PP (measured in all patients) before volumeexpansion were correlated with cardiac index change afterfluid challenge (r2 � .73, r2 � .67, and r2 � .70, respectively,p < .0001). Patients with a cardiac index increase induced byvolume expansion >15% and <15% were classified as re-sponders and nonresponders, respectively. Receiver operatingcharacteristic curves showed that the threshold �PP value of17% allowed discrimination between responder/nonresponderpatients with a sensitivity of 85% and a specificity of 100%. Forboth �PEPKT and �PEPPLET, the best threshold value was 4%with a sensitivity-specificity of 92%– 89% and 100%– 67%,respectively.
Conclusions: The present study found �PEPKT and �PEPPLET
to be as accurate as �PP in the prediction of fluid responsive-ness in mechanically ventilated septic patients. (Crit Care Med2005; 33[Suppl.]:E2534)
KEY WORDS: fluid resuscitation; heart-lung interactions; moni-toring.
E2534 Crit Care Med 2005 Vol. 33, No. 11
79
spectively screened. Exclusion criteria (for el-igibility for the study) were the presence ofmoderate to severe valve disease, contraindi-cation to a fluid challenge (hypoxemia, highpulmonary artery occlusion pressure), lack ofdeep sedation (presence of voluntary effort),absence of sinus rhythm, and hemodynamicinstability (defined as variation in mean arte-rial pressure and cardiac output of �10% overthe 15-min period before starting the proto-col).
Patient Management. Sedation and anal-gesia were provided by continuous infusion ofmidazolam and remifentanyl titrated for aRamsay score of 6 (17). Some patients weretherapeutically paralyzed (cisatracurium) ifthe attending physician deemed this appropri-ate. All patients were ventilated with positivepressure ventilation (tidal volume, 8 –10mL·kg�1 of body weight). The respiratory ratewas set to obtain a PaCO2 between 35 and 45mm Hg. The inspired fraction of oxygen wasadjusted to obtain an arterial oxygen satura-tion �92%. Inspiratory to expiratory ratio wasapproximately 1:2 in all patients.
Hemodynamic Monitoring. All pressuretransducers were referenced to mid-chest. Allpatients were monitored using a pulse oxime-try sensor with plethysmography (SpO2/Pleth,M3150A technology, Philips Medical Systems,Andover, MA) attached to the patient’s finger(phalanx) with a clip.
Cardiac Output Measurements. All pa-tients had an echocardiography-color-Dopplerinvestigation shortly before and after volumeinfusion. Complete two-dimensional echocar-diograms and color-Doppler ultrasound exam-inations were performed using a commerciallyavailable echocardiographic system (Sonos4500, Philips Medical Systems) in a semire-cumbent position with head at 45°. All trac-ings were recorded by one investigator, andeach value represented the average of five trac-ings. Echocardiography-Doppler traces wereanalyzed off-line. The cardiac output was mea-sured at the level of the aortic annulus. Aorticannulus diameter (DAo) was measured at mid-systole (T wave on ECG) and during the expi-ratory phase of the respiratory cycle, from azoomed two-dimensional image in theparasternal long axis view. From an apicalfive-chamber view, aortic flow (at the annuluslevel) was recorded using pulsed Doppler. Ve-locity time integral was measured for aorticflow (VTIAo) at the end of the expiratory pe-riod. With the use of these measurements,stroke volume could be calculated using thefollowing formula: (DAo)2 � 3.14 � VTIAo/4.To obtain cardiac output, stroke volume wasmultiplied by heart rate. Stroke volume andcardiac output were divided by the body sur-face area to obtain the stroke volume indexand cardiac index.
Respiratory Change in Pre-Ejection Pe-riod. PEP was defined as the time intervalbetween the beginning of the R wave on theECG and the upstroke of the invasive radial
arterial pressure curve (PEPKt). PEP was alsodefined as the time interval between the be-ginning of the R wave on the ECG and theupstroke of the pulse plethysmographic curve(PEPplet) (online data supplement figure). Ar-terial blood pressure-time, pulse plethysmog-raphy-time, ECG-time, and airway pressure-time curves were digitized at 500 Hz andsampled using an analogic/numeric system(Biopac Systems, Goleta, CA). Recording wasassessed using an MP100wsw Starter systemfor PC/Windows (AcqKnowledge software,Biopac Systems, Santa Barbara, CA). The dataacquired online were stored on a laptop com-puter for subsequent analysis. PEPKt and�PEPplet values in percentages were defined asthe difference between expiratory PEP (PEPE)and inspiratory PEP (PEPI) measured over onerespiratory cycle divided by the mean of expi-ratory and inspiratory values (15). PEPE andPEPI were done respectively at the minimaland the maximal wave value on the arterialpressure and pulse plethysmography tracesover one respiratory cycle. Each PEP valuerecorded was a mean of three measurements.These measurements were repeated over threedifferent respiratory cycles (total of nine mea-surements). Variation of samples allowed mea-surements on the most obvious curve illustra-tions and values were averaged. �PEP (%) wascalculated as 100 � (PEPE � PEPI)/[(PEPE �PEPI)/2] (15). The interobserver variability inmeasuring �PEP was determined in a blindedfashion, with a second observer (MF, JB). Allmeasurements were made before analysis ofpulse pressure variations (�PP) in order not tobe influenced by the results.
Respiratory Change in Pulse Pressure. Thevariation of the arterial blood pressure curve,obtained from the radial artery catheter, wasrecorded. The analysis of pressure waveformswas performed off-line on the laptop (Acq-Knowledge software, Biopac Systems). Pulsepressure was measured with a pressure scaleadjusted to systemic systolic pressure value.Respiratory change in pulse pressure (�PP in%) was calculated as previously described (18).
Study Protocol. All studies were performedin patients in a semirecumbent position withhead at 45° position. If patients were receivingvasoactive drugs, the rate of administrationwas not changed. Measurements were per-formed in duplicate, first before VE and then30 mins after VE using 8 mL·kg�1 6% hy-droxyethylstarch (Hesteril; Fresenius Kabi,Sèvres, France). Ventilatory settings and dos-ages of inotropic drugs were held constantduring the study protocol.
Statistical Analysis. For the statisticalanalysis, Stata Statistical Software, release 8.0(Stata Corporation, College Station, TX) wasused. Data were compared using Student’s t-test for continuous variables and Fisher’s ex-act test for categorical variables. Ordinal dataor nonnormally distributed continuous datawere compared by the Mann-Whitney U test orthe nonparametric Wilcoxon’s rank sum test
for paired observations. Correlations were de-termined using the Spearman rank analysis.
Patients were divided into two groups ac-cording to the percentage increase in cardiacindex (CI) in response to VE. In accord withprevious work by Stetz et al. (19), we assumedthat a 15% change in CI was needed for clin-ical significance. Patients with a CI increaseinduced by VE �15% and �15% were classi-fied as responders and nonresponders, respec-tively. We compared hemodynamic variablesbefore and after VE in responder and nonre-sponder patients. Receiver operating charac-teristic curves for responders/nonresponderswere generated for �PP, �PEPKT, and�PEPPLET, varying the discriminating thresh-old of each variable. The areas under the re-ceiver operating characteristic curves (�SE)were calculated for each variable and com-pared (20).
The sample size was planned by design todetect a significant difference in the �PP value(in %) between responders and nonrespondersequivalent to a previous study (i.e., 15% � 5%vs. 6% � 3%) (9) with the level of significancefixed at � .05 and a power of 90%. Thecalculation resulted in at least five patients ineach group. As in this previous study aboutonly 40% of patients increased their CI �15%(9), and to be sure of having at least fiveresponders and five nonresponders in thepresent study, we decided to include 20 pa-tients, as was practically done in another sim-ilar study by our group (10).
Results are expressed as mean � SD, if notspecified otherwise. All tests were two tailed,and p � .05 was considered statistically sig-nificant.
RESULTS
Twenty patients, all of whom toleratedthe experimental protocol, were includedin the final analysis. Demographic char-acteristics are presented in Table 1. Cat-echolamine infusions were required in allpatients (dobutamine n 11, and/or nor-epinephrine n 14, and/or dopamine n 6). In all patients, systolic arterial pres-sures were higher during the inspiratoryperiod than during the expiratory period.Hemodynamic variables before and aftervolume infusion are shown in Table 2.Volume infusion produced an increase incardiac index from 2.3 � 0.9 to 2.8 � 1L·min�1·m�2 (p � .0001). Thirteen pa-tients were classified as responders (car-diac index increase �15%) and nine pa-tients as nonresponders. Intraobserverand interobserver variabilities, deter-mined by repeating measurements in tenrandomized patients, were expressed asthe mean percent error (i.e., the differ-ence between two observations, dividedby the mean of the two observed values).
E2534 Crit Care Med 2005 Vol. 33, No. 11
80
For measurements of PEPPLET (millisec-onds) and PEPKT (milliseconds), the val-ues were 7 � 3% and 5 � 3% by the sameobserver and 8 � 4% and 6 � 4% be-tween the two different observers (MF,JB), respectively. For measurements ofcardiac output (L·min�1), measured us-ing Doppler, the value was 7 � 4% by thesame observer and 9 � 5% between thetwo different observers (MF, JB).
�PEPKT, �PEPPLET, and �PP beforevolume expansion correlated withchanges in CI after volume expansion (p� .01, Fig. 1). The decreases in �PEPKT,�PEPPLET, and �PP were significantlycorrelated with the fluid infusion-in-duced increase in cardiac index (r2 .64,r2 .53, r2 .69, respectively; p � .01),in such a way that the bigger the decreasein the index was, the more significant wasthe increase in CI. Before volume expan-
sion, �PEPKT, �PEPPLET, and �PP wereall higher in responders than in nonre-sponders (p � .01, Fig. 2).
The areas under the receiver operatingcharacteristic curves (�SE) were as fol-lows: 0.96 � 0.03 for �PP, 0.97 � 0.03for �PEPKT, and 0.94 � 0.05 for �PEP-PLET (Fig. 3). The area for �PP was notsignificantly different from the area for�PEPKT (p .82) or �PEPPLET (p .69).The threshold �PP value of 17% alloweddiscrimination between responder andnonresponder patients with a sensitivityof 85% (95% confidence interval, 55–98%) and a specificity of 100% (95% con-fidence interval, 66–100%). If the thresh-old reported in previous studies (13%)was used (9), the sensitivity was 92%(95% confidence interval, 64–100%) andthe specificity 78% (95% confidence in-terval, 40–97%). For both �PEPKT and
�PEPPLET, the best threshold value was4% with a sensitivity and a specificity of92% (95% confidence interval, 64–100%)and 89% (95% confidence interval, 51–100%) for �PEPKT, and 100% (95% con-fidence interval, 75–100%) and 67%(95% confidence interval, 30–93%) for�PEPPLET.
DISCUSSION
The present study shows a good cor-relation between �PEP value before VEand CI increase after fluid challenge inmechanically ventilated septic patients.Furthermore, with a threshold value of4%, �PEPKT and �PEPPLET were shownto predict fluid responsiveness with a sen-sitivity-specificity of 92%– 89% and100%– 82%, respectively. In conse-quence, we found �PEP to be as accurateas �PP in monitoring the response tovolume infusion in septic patients.
Many studies have shown that staticpressure measurements and echocardio-graphically measured ventricular areasare poor predictors of fluid responsive-ness in septic patients (9–11). In con-trast, dynamic monitoring consisting ofassessment of fluid responsiveness usingrespiratory changes in systolic arterialpressure, pulse pressure, aortic blood ve-locity, and vena cava diameter has beenfound to be accurate in septic patients(9–12, 21, 22).
The pathophysiologic concept of dy-namic variables is very robust (23, 24).Indeed, in mechanically, deeply sedated,ventilated patients, positive pressure ven-tilation cyclically increases intrathoracicpressure and lung volume. Both reducevenous return, alter cardiac preload, anddecrease stroke volume. Thus, left ven-tricular stroke volume varies cyclically,being maximal during mechanical inspi-ration and minimal during expiration.The greatest mechanical breath-inducedstroke volume variations are observed inthe context of hypovolemia (25). Theserespiratory changes in stroke volume re-sult in systolic pressure variation (21),�PP (9), �PEP (15), and vena cava diam-eter modifications (11, 12, 22). Every oneof these indicators has been shown toidentify decreased preload hypotensionand to distinguish between respondersand nonresponders to fluid challenge indifferent patient populations.
In a recent clinical commentary, basedon previous studies (26, 27), Magder (28)
Table 2. Effects of volume infusion on patients’ hemodynamic variables
Before VE After VE p
HR, beats � min�1 114 � 27 105 � 24 a
MAP, mm Hg 69 � 17 85 � 20 b
�PEPKT, % 6 � 4 3 � 2 a
�PEPPLET, % 6 � 3 3 � 2 a
�PP, % 19 � 13 5 � 4 a
CI, L/min/m2 2.3 � 0.9 2.8 � 1 a
VE, volume expansion; HR, heart rate; MAP, mean arterial pressure; �PEPKT, respiratory changein pre-ejection period (with catheter); �PEPPLET, respiratory change in pre-ejection period (withplethysmography); �PP, respiratory variation in pulse pressure; CI, cardiac index.
ap � .001, b p � .01, before VE/after VE (Wilcoxon’s rank sum test for paired observations).
Table 1. Patient characteristics
PatientNo. Cause of Sepsis Age, yrs
Gender,Male/Female BS, m2 Outcome
1 Pneumonia 69 F 1.95 D2 Pneumonia 38 M 1.8 S3 Pneumonia 77 F 1.75 S4 Pneumonia 66 M 1.95 S5 Aspiration pneumonia 58 M 1.9 D6 Pneumonia 50 F 1.7 S7 Pneumonia 21 M 1.9 S8 Septicemia of unknown origin 73 M 1.75 D9 Peritonitis 66 F 1.7 S
10 Peritonitis 76 M 1.7 D11 Peritonitis 70 M 2.1 S12 Peritonitis 86 F 1.7 D13 Pneumonia 80 M 1.7 S14 Pneumonia 64 M 1.5 S15 Aspiration pneumonia 69 M 1.5 S16 Peritonitis 37 M 1.95 S17 Aspiration pneumonia 38 M 1.9 S18 Mediastinitis 63 M 1.95 D19 Peritonitis 67 F 1.45 S20 Aspiration pneumonia 75 M 1.85 D
Mean � SD 62 � 17 1.8 � 0.2Ratio 6/14 7/13
BS, body surface; D, died; S, survived.
E2534 Crit Care Med 2005 Vol. 33, No. 11
81
questioned the usefulness of dynamicvariables in quantitatively predicting thefluid responsiveness. The present work inseptic mechanically ventilated patientsdoes not support the theoretical analysisof Magder (28) and is in line with thework of Vieillard-Baron et al. (12), where66 critically ill septic patients were stud-ied. In this last study, respiratory changesin the superior vena cava diameter werefound to be as accurate as �PP for pre-dicting fluid responsiveness (12). Thiswork, which studied a right ventriculardynamic echocardiographic variable,confirmed a concept that had alreadybeen observed by another team (29).
The pressure curves in the aorta andat peripheral sites are of a similar com-position. After the opening of the aorticvalve, the increase in pressure representsthe beginning of ventricular ejection.Thus, the time from the onset of theventricular depolarization (R on theECG) to the beginning of left ventricularejection may be assessed with radial arte-rial pressure (15) and/or noninvasivelywith plethysmography (30). Since Star-ling’s (31) work, it has been well knownthat increasing the initial length of themuscle fiber by increasing the heart vol-ume enhances the cardiac contractilityand output. His predecessor Frank (32)established another important principle:the greater the initial heart volume, themore rapid the rate of increase and thelarger the peak pressure reached. Frank(32) was, therefore, able to show that anincreasing heart volume stimulated theventricle to contract more rapidly. Thus,physiologically, it seems coherent thatthis time should be increased when theleft ventricle preload is decreased andvice versa.
In a recent study, Bendjelid et al. (15)demonstrated that �PEP was a useful in-dex of preload responsiveness in cardiacsurgical patients. However, in that study,this variable was not as accurate as �PP(15). In the present study the measured�PEP value using the invasive radial ar-terial pressure or the digital plethysmog-raphy wave was found to be as accurate as�PP for predicting fluid responsiveness.The difference between these studiescould be related to the general patientpopulation and/or the number of patientsstudied. However, it is possible that themethodology used to measure PEP(Biopac Systems) in the present studymight be responsible for the better re-sults for �PEP. Indeed, in the first study,the resolution (the shortest time that can
Figure 1. Linear correlation analysis of the relationship between dynamic preload variables measuredbefore volume expansion and changes in cardiac index (CCI) following volume expansion. p � .05 wasconsidered significant. �PEPKT, respiratory change in pre-ejection period (with catheter); �PEPPLET,respiratory change in pre-ejection period (with plethysmography); �PP, respiratory variation in pulsepressure.
Figure 2. Distribution of all the individual results (n 22) of the respiratory change in pre-ejectionperiod (with catheter) (�PEPKTd) and respiratory change in pre-ejection period (with plethysmogra-phy) (�PEPPLET, upper panel) and respiratory variation in pulse pressure (�PP, lower panel) inpercentages. Gray boxes, responders (cardiac index increase � 5% after volume challenge). Whiteboxes, nonresponders. Box plots, the middle horizontal bars represent median values, the squaresrepresent the 25th and 75th percentiles, and the lower and upper bars represent the 10th and 90thpercentiles. The upper and lower circles represent the minimum and maximal values.
E2534 Crit Care Med 2005 Vol. 33, No. 11
82
be measured) of the time measurementswas of 10 msecs (sample of 100 Hz) (15).In the present study, the data acquiredonline from the arterial blood pressure-time, pulse plethysmography-time, ECG-time, and airway pressure-time curveswere sampled at 500 Hz, defining a reso-lution of 2 msecs.
In the present study, the threshold�PP value of 17% allowed better discrim-ination between responders and nonre-sponders. When we used the thresholdvalue of 13% reported for �PP in previ-ous studies (9), sensitivity increased(92%) but specificity decreased (78%).Therefore, the present study demon-strates the lack of a universally powerfulthreshold for �PP in mechanically venti-lated septic patients. Indeed, one of themajor drawbacks of this marker is thatthe �PP value is dependent on the capac-itance of the pulmonary vessels (23, 27,33) and the compliance and resistance ofthe arterial vessels (34). We may expectthese three variables to be of differentvalues in a septic vasoplegic patient thanin another patient receiving catechol-amine infusions.
In septic patients, this increased vas-cular capacitance may also affect the pre-dictive value of a marker of fluid respon-siveness. Indeed, the increase in
ventricular preload following volume in-fusion depends on the partitioning of thefluid into different cardiovascular com-partments organized in series (35, 36). Inthis regard, volume infusion may in-crease intravascular blood volume (or in-terstitial volume) but not necessarily car-diac preload, making a patient (whowould otherwise have been a responder tofluid challenge) a nonresponder becausepreload does not increase. In the presentstudy, we have also correlated changes in�PEP following volume infusion withchanges in cardiac index related to fluidexpansion. We found a close correlationbetween the two variables.
In the emergency room, for practicalreasons, reliable fluid responsiveness pre-dictors are needed to monitor and facili-tate treatment of hemodynamically un-stable patients. However, clinicalvariables alone have repeatedly beenproven to be unreliable in the assessmentof cardiopulmonary status of these pa-tients. Furthermore, most fluid respon-siveness indexes are useless before theinsertion of central venous and/or arterialcatheters. In this setting, the noninvasive�PEPPLET index obtained using pulseplethysmography is feasible and safe andresults in a significant time gain in the
accurate assessment of hemodynamicstatus of emergency room patients.
Some limitations of this work shouldbe acknowledged. First, the analysis ofthese dynamic markers is useless in spon-taneously breathing intubated patients(8), a mechanical ventilation mode oftenused in the intensive care unit. Regularcardiac rhythm is also an obligatory con-dition to allow their use (8). Second, asoutlined by Magder (28), dynamic vari-ables predict fluid responsiveness underspecific conditions; Just because a patientmay be predicted to increase cardiac out-put in response to VE does not mean thathe or she needs fluid therapy in absenceof shock (37). Third, the signal quality ofpulse oximetry can be reduced and mea-surement precisions may be affected.However, in patients affected by an earlystate of septic shock, this phenomenon israrely seen, as a pathologic vasodilationoccurs (37). Furthermore, we may expectthat in the future, �PEP will be measuredautomatically (38). Finally, as high-lighted in a preceding paragraph, thesevariables can provide a qualitative guideto fluid responsiveness, but quantitativerecommendations are not generalizableand may be misleading.
CONCLUSIONS
The present study confirms that �PEPis of great value in predicting the hemody-namic effects of volume infusion in deeplysedated, mechanically ventilated, septic pa-tients. Moreover, the noninvasive approachusing pulse oximetry (�PEPPLET) is as ac-curate as �PP to predict fluid responsive-ness. In particular, when one looks at dataon early goal-directed therapy (1), one sees
Figure 3. Receiver operating characteristic (ROC) curves comparing the ability of respiratory variationin pulse pressure (�PP), pre-ejection period (with catheter) (�PEPKT), and respiratory change inpre-ejection period (with plethysmography) (�PEPPLET, upper panel) to discriminate responders(cardiac index increase �15%) and nonresponders to volume expansion. The areas under the ROCcurve were not significantly different (p not significant).
T he present study
confirms that the
respiratory change
in pre-ejection period is of
great value in predicting the
hemodynamic effects of vol-
ume infusion in deeply se-
dated, mechanically venti-
lated, septic patients.
E2534 Crit Care Med 2005 Vol. 33, No. 11
83
that �PEPPLET could be a useful means ofguiding the initial phase of volume therapyin septic mechanically ventilated patientsbefore the insertion of an arterial catheter.
ACKNOWLEDGMENTS
We are grateful for the translationsupport provided by Dr. Alice Wood.
REFERENCES
1. Rivers E, Nguyen B, Havstad S, et al: Earlygoal-directed therapy in the treatment of se-vere sepsis and septic shock. N Engl J Med2001; 345:1368–1377
2. Hollenberg SM, Ahrens TS, Annane D, et al:Practice variables for hemodynamic supportof sepsis in adult patients: 2004 update. CritCare Med 2004; 32:1928–1948
3. Balogh Z, McKinley BA, Cocanour CS, et al:Supranormal trauma resuscitation causesmore cases of abdominal compartment syn-drome. Arch Surg 2003; 138:637–642
4. Boussat S, Jacques T, Levy B, et al: Intravas-cular volume monitoring and extravascularlung water in septic patients with pulmonaryedema. Intensive Care Med 2002; 28:712–718
5. Bunchman TE: Fluid overload in multipleorgan dysfunction syndrome: A prediction ofsurvival. Crit Care Med 2004; 32:1805–1806
6. Foland JA, Fortenberry JD, Warshaw BL, etal: Fluid overload before continuous hemo-filtration and survival in critically ill chil-dren: A retrospective analysis. Crit Care Med2004; 32:1771–1776
7. Michard F, Teboul JL: Predicting fluid re-sponsiveness in ICU patients: A critical anal-ysis of the evidence. Chest 2002; 121:2000–2008
8. Bendjelid K, Romand JA: Fluid responsive-ness in mechanically ventilated patients: Areview of indices used in intensive care. In-tensive Care Med 2003; 29:352–360
9. Michard F, Boussat S, Chemla D, et al: Rela-tion between respiratory changes in arterialpulse pressure and fluid responsiveness inseptic patients with acute circulatory failure.Am J Respir Crit Care Med 2000; 162:134–138
10. Feissel M, Michard F, Mangin I, et al: Respi-ratory changes in aortic blood velocity as anindicator of fluid responsiveness in ventilatedpatients with septic shock. Chest 2001; 119:867–873
11. Barbier C, Loubieres Y, Schmit C, et al: Re-
spiratory changes in inferior vena cava diam-eter are helpful in predicting fluid respon-siveness in ventilated septic patients.Intensive Care Med 2004; 30:1740–1746
12. Vieillard-Baron A, Chergui K, Rabiller A, etal: Superior vena caval collapsibility as agauge of volume status in ventilated septicpatients. Intensive Care Med 2004; 30:1734–1739
13. Weissler AM: Current concepts in cardiology.Systolic-time intervals. N Engl J Med 1977;296:321–324
14. Bendjelid K, Romand JA: Respiratorychanges in preejection period (PEP) predictshemodynamic response to PEEP. Crit CareMed 2001; 29:222
15. Bendjelid K, Suter PM, Romand JA: The re-spiratory change in preejection period: A newmethod to predict fluid responsiveness.J Appl Physiol 2004; 96:337–342
16. Levy MM, Fink MP, Marshall JC, et al: 2001SCCM/ESICM/ACCP/ATS/SIS InternationalSepsis Definitions Conference. Crit Care Med2003; 31:1250–1256
17. Ramsay MA, Savege TM, Simpson BR, et al:Controlled sedation with alphaxalone-alphadolone. BMJ 1974; 2:656–659
18. Michard F, Chemla D, Richard C, et al: Clin-ical use of respiratory changes in arterialpulse pressure to monitor the hemodynamiceffects of PEEP. Am J Respir Crit Care Med1999; 159:935–939
19. Stetz CW, Miller RG, Kelly GE, et al: Reli-ability of the thermodilution method in thedetermination of cardiac output in clinicalpractice. Am Rev Respir Dis 1982; 126:1001–1004
20. Hanley JA: Receiver operating characteristic(ROC) methodology: The state of the art. CritRev Diagn Imaging 1989; 29:307–335
21. Tavernier B, Makhotine O, Lebuffe G, et al:Systolic pressure variation as a guide to fluidtherapy in patients with sepsis-induced hy-potension. Anesthesiology 1998; 89:1313–1321
22. Feissel M, Michard F, Faller JP, et al: Therespiratory variation in inferior vena cavadiameter as a guide to fluid therapy. Inten-sive Care Med 2004; 30:1834–1837
23. Vieillard-Baron A, Chergui K, Augarde R, etal: Cyclic changes in arterial pulse duringrespiratory support revisited by Dopplerechocardiography. Am J Respir Crit CareMed 2003; 168:671–676
24. Vieillard-Baron A, Prin S, Chergui K, et al:Hemodynamic instability in sepsis: Bedsideassessment by Doppler echocardiography.
Am J Respir Crit Care Med 2003; 168:1270–1276
25. Gunn SR, Pinsky MR: Implications of arterialpressure variation in patients in the intensivecare unit. Curr Opin Crit Care 2001;7:212–217
26. Scharf SM, Caldini P, Ingram RH Jr: Cardio-vascular effects of increasing airway pressurein the dog. Am J Physiol 1977; 232:H35–H43
27. Denault AY, Gasior TA, Gorcsan J III, et al:Determinants of aortic pressure variationduring positive-pressure ventilation in man.Chest 1999; 116:176–186
28. Magder S: Clinical usefulness of respiratoryvariations in arterial pressure. Am J RespirCrit Care Med 2004; 169:151–155
29. Jellinek H, Krenn H, Oczenski W, et al: In-fluence of positive airway pressure on thepressure gradient for venous return in hu-mans. J Appl Physiol 2000; 88:926–932
30. Stengele E, Winkler F, Trenk D, et al: Digitalpulse plethysmography as a non-invasivemethod for predicting drug-induced changesin left ventricular preload. Eur J Clin Phar-macol 1996; 50:279–282
31. Starling E: The Linacre Lecture on the Lawof the Heart Given at Cambridge, 1915. Lon-don, Longmans, Green, 1918
32. Frank O: Dynamik des Herzmuskels. Z Biol1895; 32:370–447. Translated in Am Heart J1958; 1958:1282–1317, 1467–1478
33. Brower R, Wise RA, Hassapoyannes C, et al:Effect of lung inflation on lung blood volumeand pulmonary venous flow. J Appl Physiol1985; 58:954–963
34. Pizov R, Ya’ari Y, Perel A: Systolic pressurevariation is greater during hemorrhage thanduring sodium nitroprusside-induced hypo-tension in ventilated dogs. Anesth Analg1988; 67:170–174
35. Magder S, Georgiadis G, Cheong T: Respira-tory variations in right atrial pressure predictthe response to fluid challenge. J Crit Care1992; 7:76–85
36. Michard F, Reuter DA: Assessing cardiac pre-load or fluid responsiveness? It depends onthe question we want to answer. IntensiveCare Med 2003; 29:1396
37. Beale RJ, Hollenberg SM, Vincent JL, et al:Vasopressor and inotropic support in septicshock: An evidence-based review. Crit CareMed 2004; 32:S455–S465
38. Bernardi L, Calciati A, Lumina C, et al: Au-tomatic on-line measurement of systolictime intervals using a personal computer.J Cardiogr 1986; 16:133–139
E2534 Crit Care Med 2005 Vol. 33, No. 11
84
APPENDIX
Preenection Period Variations Predict the Fluid Responsiveness in Septic Ventilated Patients. Online Data Supplement (AcqKnowledge software, BiopacSystems, Santa Barbara, CA). Patient 8: Gray areas, measurement of respiratory change in pre-ejection period; white areas, measurement of respiratorychange in pre-ejection period.
E2534 Crit Care Med 2005 Vol. 33, No. 11
85
Intensive Care Med (2007) 33:993–999DOI 10.1007/s00134-007-0602-6 O R I G I N A L
Marc FeisselJean-Louis TeboulPaolo MerlaniJulio BadieJean-Pierre FallerKarim Bendjelid
Plethysmographic dynamic indices predict fluidresponsiveness in septic ventilated patients
Received: 16 May 2006Accepted: 27 February 2007Published online: 29 March 2007© Springer-Verlag 2007
Electronic supplementary materialThe online version of this article(doi:10.1007/s00134-007-0602-6) containssupplementary material, which is availableto authorized users.
Funding: No external funding
This work was performed in the MedicalIntensive Care Unit, Centre Hospitalier,Belfort, France.
M. Feissel · J. Badie · J.-P. FallerCentre Hospitalier, Intensive Care Unit,Belfort, France
J.-L. TeboulParis Sud Medical School, Réanimationmédicale, Bicêtre Hospital,Le Kremlin Bicêtre, France
P. Merlani · K. Bendjelid (�)Geneva University Hospitals, Intensive CareUnit, Department of Anesthesiology,Pharmacology and Intensive Care,1211 Geneva 14, Switzerlande-mail: [email protected].: +41-22-3827452Fax: +41-22-3827455
Abstract Objectives: In septicpatients, reliable non-invasive pre-dictors of fluid responsiveness areneeded. We hypothesised that therespiratory changes in the amplitudeof the plethysmographic pulse wave(∆PPLET) would allow the predictionof changes in cardiac index followingvolume administration in mechan-ically ventilated septic patients.Design: Prospective clinical investi-gation. Setting: An 11-bed hospitalmedical intensive care unit. Patients:Twenty-three deeply sedated septicpatients mechanically ventilated withtidal volume ≥ 8 ml/kg and equippedwith an arterial catheter and a pulseoximetry plethysmographic sensor.Interventions: Respiratory changesin pulse pressure (∆PP), ∆PPLET andcardiac index (transthoracic Dopplerechocardiography) were determinedbefore and after volume infusion ofcolloids (8 ml/kg). Measurementsand main results: Twenty-eightvolume challenges were performedin 23 patients. Before volume expan-sion, ∆PP correlated with ∆PPLET(r2 = 0.71, p < 0.001). Changes in
cardiac index after volume expansionsignificantly (p < 0.001) correlatedwith baseline ∆PP (r2 = 0.76) and∆PPLET (r2 = 0.50). The patientswere defined as responders to fluidchallenge when cardiac index in-creased by at least 15% after thefluid challenge. Such an event oc-curred 18 times. Before volumechallenge, a ∆PP value of 12% anda ∆PPLET value of 14% alloweddiscrimination between respondersand non-responders with sensitivityof 100% and 94% respectively andspecificity of 70% and 80% respec-tively. Comparison of areas underthe receiver operator characteristiccurves showed that ∆PP and ∆PPLETpredicted similarly fluid responsive-ness. Conclusion: The present studyfound ∆PPLET to be as accurate as∆PP for predicting fluid respon-siveness in mechanically ventilatedseptic patients.
Keywords Fluid resuscitation ·Heart–lung interactions · Volumeresponsiveness · Monitoring
Introduction
There are now a great number of clinical studies support-ing the usefulness of dynamic indices based on heart–lunginteraction for guiding volume resuscitation in patients re-ceiving mechanical ventilation [1, 2]. Accordingly, the res-piratory variations of arterial pulse pressure, of “pulse con-
tour” stroke volume, and of Doppler aortic blood velocityhave been shown to predict volume responsiveness far bet-ter than static markers of preload such as cardiac fillingpressures or dimensions [3–5]. The pulse oximeter couldbe an attractive device for detecting volume responsivenesssince it is non-invasive and easy to use and also since thepulse oximetry plethysmographic signal resembles the pe-
86
994
ripheral arterial pressure waveform [6]. In this regard, res-piratory variation of pulse oximeter waveforms has beencorrelated with that of systolic arterial pressure [6–8] andpulse pressure [9].
The plethysmographic “pulse” wave (nadir–peak)displayed on the monitor is assumed to reflect the pul-satile changes in absorption of the infrared light betweenthe light source and the photo detector of the pulseoximeter [10]. Consequently, the beat-to-beat changes inthe amplitude of the plethysmographic pulse wave areassumed to be the result of the beat-to-beat changes instroke volume transmitted to the arterial blood [11]. In thisrespect, the degree of respiratory changes in the amplitudeof the plethysmographic pulse (∆PPLET) wave should bea potential marker of respiratory stroke volume variationand hence a marker of volume responsiveness [12, 13]. Inthis regard, ∆PPLET was demonstrated to be influencedby changes in preload [14]. In a clinical study, it wasrecently shown that each time ∆PPLET was greater thanthe threshold value of 15%, fluid challenge resulted in anincreased of cardiac output by more than 15% [15]. Onthe other hand, ∆PPLET values lower than 15% poorlypredicted volume responsiveness, maybe because halfof the patients were ventilated with tidal volumes lowerthan 8 ml/kg [15], a condition where dynamic indices likepulse pressure variation fail to predict accurately volumeresponsiveness [16].
The aim of our study was to test the hypothesis that∆PPLET could be as valuable to predict volume respon-siveness as respiratory changes in arterial pulse pressure inseptic patients receiving mechanical ventilation with a tidalvolume > 8 ml/kg and exhibiting neither inspiratory ef-forts nor arrhythmias.
Materials and methods
The institutional review board for human subjects ap-proved the protocol, considering it as a part of routineclinical practice, and patients were informed that theywere participating in this study. We included only me-chanically ventilated patients with septic shock, as definedby the International Sepsis Definitions Conference [17],who were equipped with a systemic arterial catheter andfor whom the decision to give fluid was taken by theirattending physician in the context of standard treatment.We excluded those patients with moderate to severe valvedisease and those who experienced inspiratory efforts orcardiac arrhythmias.
Patient management
Sedation and analgesia were provided by continuous infu-sion of midazolam and remifentanil titrated for a Ramsayscore of 6 [18]. Patients were therapeutically paralysed
(with cisatracurium) if the attending physician deemedthis appropriate. All patients were ventilated with positivepressure ventilation (tidal volume, 8–10 ml/kg of bodyweight). The respiratory rate was set to obtain a PaCO2of 35–45 mmHg. The inspired fraction of oxygen wasadjusted in order to obtain an arterial oxygen saturation> 92%. Inspiratory to expiratory ratio was approximately0.5:1 in all patients.
Haemodynamic monitoring
All pressure transducers were referenced to mid-chest. Allpatients were monitored using a pulse oximetry sensorwith plethysmography (SpO2/Pleth, M3150A technology,Philips Medical Systems, Andover, MA) attached to thepatient’s finger (phalanx) with a clip.
Cardiac output measurements
All patients had a colour-Doppler echocardiography-investigation shortly before and after volume infusion.Complete two-dimensional echocardiography and colour-Doppler ultrasound examinations were performed usinga commercially available echocardiographic system(Sonos 5500, Philips Medical Systems, Eindhoven,Netherlands) in a semi-recumbent position with headat 45 °. All tracings were recorded by one investigator,and each value represented the average of five tracings.Echocardiography–Doppler traces were analysed offline. The cardiac output was measured at the level ofthe aortic annulus. Aortic annulus diameter (DAo) wasmeasured at mid-systole, (T wave on ECG) and during theexpiratory phase of the respiratory cycle, from a zoomedtwo-dimensional image in the parasternal long axis view.From an apical five-chamber view, aortic flow (at theannulus level) was recorded using pulsed Doppler. Veloc-ity–time integral for aortic flow (VTIAo) was measuredat the end of the expiratory period. With the use of thesemeasurements, stroke volume could be calculated usingthe following formula: (DAo)2 × 3.14× VTIAo/4. Toobtain cardiac output, stroke volume was multiplied byheart rate. The cardiac output was divided by the bodysurface area (in m2) to obtain the cardiac index. We didnot recalculate the area of the aortic orifice over timesince it is assumed to be unchanged because of the fibroticnature of the annulus.
Respiratory change in the amplitude of theplethysmographic pulse wave
Arterial blood pressure–time, pulse plethysmography–time, ECG–time and airway pressure–time curves were
87
995
Fig. 1 Simultaneous recording of systemic arterial pressure (Art),plethysmographic “pulse” (Pleth), EKG and airway flow (AWF)curves in one patient with large ∆PP and ∆PPLET. (AcqKnowledgesoftware, Biopac Systems, Santa Barbara, CA, USA)
digitised at 500 Hz and sampled using an analogue/numeric system (Biopac Systems, Goleta, CA, USA).Recording was assessed using an MP100wsw Starter sys-tem for PC/Windows (AcqKnowledge software, BiopacSystems, Santa Barbara, CA, USA). The data acquiredonline were stored on a laptop computer for subsequentanalysis of the respiratory changes in arterial pulsepressure (∆PP) and ∆PPLET (Fig. 1). The inter-observervariability of ∆PP and ∆PPLET measurements was de-termined in a “blinded” fashion, with a second observer(M. F., J. B.). All measurements were made before theanalysis of ∆PP so as not to be influenced by the results.The ∆PP and ∆PPLET were calculated as previouslydescribed [4] and expressed in percentage. Pulse pressurewas calculated on a beat-to-beat basis as the differencebetween systolic and diastolic arterial pressure. Maximalpulse pressure (PPmax) and minimal pulse pressure(PPmin) values were determined over a single respiratorycycle. To assess the respiratory changes in pulse pressure,the percent change in pulse pressure was calculated as:∆PP = 100 × {(PPmax–PPmin)/([PPmax+PPmin]/2)}.
Study protocol
All studies were performed in patients in a semi-recumbentposition with head at 45 ° position. Measurements wereperformed in duplicate, first before volume expansion andthen 30 min after volume expansion using 8 ml/kg 6%hydroxyethyl starch (Voluven; Fresenius Kabi, Sèvres,France). The ventilatory settings and the rate of adminis-tration of vasoactive drugs were not changed throughoutthe study. Regarding the echocardiographic measurementof cardiac output, the area of the aortic orifice has beenmeasured only before fluid infusion as it is assumed to be
unchanged because of the fibrotic nature of the annulus.Therefore, VTIAo was the only variable measured beforeand after fluid challenge.
Statistical analysis
For the statistical analysis, Stata Statistical Software, Re-lease 8.0 ® (Stata Corporation, College Station, TX, USA)was used. Data were compared using paired t-test for con-tinuous variables. Ordinal data or non-normally distributedcontinuous data were compared using the Mann–WhitneyU-test or the non-parametric Wilcoxon rank sum test forpaired observations. Correlations were determined usinglinear regression analysis. We also randomly selected a sin-gle paired observation for each of the n = 23 patients andperformed all analyses that had already been conducted.
For the set of measurements obtained before fluidchallenge, the intraobserver and interobserver variabilityof VTIAo measurements was determined in all patientsand expressed as the mean percent error (i.e. the differencebetween two observations, divided by the mean of the twoobserved values).
Patients were divided into two groups according to thepercent increase in cardiac index in response to volumeexpansion. In accordance with previous studies [1, 2,15, 16, 19], we took the benchmark of 15% for dif-ferentiating responders from non-responders [20]. Wecompared haemodynamic parameters before and after vol-ume expansion in responder and non-responder patientsusing a paired t-test for continuous variables. Receiveroperating characteristic (ROC) curves for responders–non-responders were generated for ∆PP and ∆PPLET, varyingthe discriminating threshold of each parameter. The areasunder the ROC curves (± SE) were calculated for eachparameter and compared [21]. A method of comparingthe areas under ROC curves derived from the same cases.All tests were two-tailed, and a p-value less than 0.05 wasconsidered statistically significant.
ResultsTwenty-three patients (mean age 62 ± 17 years) were in-cluded. Fourteen patients survived. Mean tidal volume was9.0 ± 0.9 ml/kg and plateau pressure less than 30 cmH2Oin all patients. A total of 28 fluid challenges wereanalysed. All patients received catecholamines: dobu-tamine (5 µg/kg/min) in association with norepinephrine(n = 4), norepinephrine alone (n = 15) and dopamine(5 µg/kg/min) alone (n = 4). Mean norepinephrine dosewas 0.42 ± 0.24 µg/kg/min. No patient experiencedhypothermia at the time of the study. Haemodynamicvariables before and after volume infusion are shown inTable 1. Volume infusion produced an increase in cardiacindex from 2.5 ± 0.7 to 3.0 ± 0.9 l/min/m2(p < 0.0001).
88
996
Before VE After VE p a
HR (beats/min) 111 ± 25 101 ± 24 < 0.001MAP (mmHg) 74 ± 16 87 ± 19 < 0.01∆PP (%) 18 ± 11 5 ± 3 < 0.001∆PPLET (%) 23 ± 15 7 ± 5 < 0.001CI (l/min/m2) 2.5 ± 0.7 3.0 ± 0.9 < 0.001
HR, heart rate; MAP, mean arterial pressure; ∆PP, respiratory changes in arterial pulse pressure;∆PPLET, respiratory changes in the amplitude of the plethysmographic pulse wave (with pulse oxime-ter); CI, cardiac index; VE, volume expansion; a Before VE/after VE (paired t-test)
Table 1 Effects of volumeinfusion on patients’haemodynamic parameters (28fluid challenges in 23 patients)
Fig. 2 Linear regression analysis of the relationship between changein ∆PP and change in ∆PPLET following volume infusions (28fluid challenges in 23 patients). ∆PP, respiratory changes in arterialpulse pressure; ∆PPLET, respiratory changes in the amplitude of theplethysmographic pulse wave (with pulse oximeter)
Before volume expansion, ∆PP correlated with∆PPLET (r2 = 0.71, p < 0.001). Changes in ∆PP cor-related with changes in ∆PPLET following volumeexpansion, (p < 0.01; Fig. 2). Changes in cardiac index
Fig. 3 Linear regression analysisof the relationship between ∆PPand ∆PPLET measured beforevolume expansion and changesin cardiac index (CI) followingvolume expansion (28 fluid chal-lenges in 23 patients). p < 0.05was considered significant
after volume expansion significantly (p < 0.001) cor-related with baseline ∆PP (r2 = 0.76) and ∆PPLET(r2 = 0.50) (Fig. 3). The fluid-induced decreases in ∆PPand ∆PPLET were significantly correlated with the fluidinfusion–induced increases in cardiac index (r2 = 0.64 andr2 = 0.38; p < 0.01, respectively). In 18 cases patients wereclassified as responders (cardiac index increase ≥ 15%),and in 10 cases patients were classified as non-responders.Before volume expansion, mean ∆PP and ∆PPLET weresignificantly higher in responders than in non-responders(p < 0.01; Fig. 4). Before volume challenge, a ∆PP valueof 12% and a ∆PPLET value of 14% allowed discrim-ination between responders and non-responders withsensitivity of 100% and 94% respectively and specificityof 70% and 80% respectively. Comparison of areas underthe ROC curves showed that ∆PP and ∆PPLET predictedfluid responsiveness similarly (Fig. 5). The combination ofthe two measurements (∆PP and ∆PPLET) did not improvethe power of prediction.
When a single paired observation for each of the 23patients was selected (after removing five pairs of valuesusing a random selection) the results were statisticallyunchanged (see ESM). For 23 pairs of measurements, theareas under the ROC curves were 0.99 (0.98–1.0) and0.96 (0.85–1.0) for ∆PP (optimal cut-off value of 13%)and ∆PPLET (optimal cut-off value of 12%) respectively.
89
997
Fig. 4 Distribution of all the individual results (28 fluid challengesin 23 patients) of ∆PP and ∆PPLET (measured before volume expan-sion in %). R, Responders (cardiac index increase ≥ 15% after vol-ume challenge); NR, non-responders (cardiac index increase < 15%after volume challenge). Points and arrows indicate mean and SDrespectively
The intraobserver variability of VTIAo measurements was0.5 ± 0.7% and the interobserver (M. F., J. B.) variabilityof VTIAo measurements was 2.2 ± 0.8%.
Discussion
The present study shows that ∆PPLET is as valuable as∆PP for predicting volume responsiveness in mechani-cally ventilated septic patients. Similar threshold valueswere found for ∆PPLET (14%) and for ∆PP (12%).
Previous studies demonstrated that pulse pressure vari-ation was more reliable than static parameters of preload topredict volume responsiveness in critically ill patients re-ceiving mechanical ventilation [1, 4, 16, 19]. The rationalefor guiding fluid therapy on ∆PP or on other heart–lung in-teraction indices [1–3, 21] is that influence of positive pres-sure ventilation on haemodynamics is greater when centralblood volume is low than when it is normal or high.
The finger pulse oximetry plethysmographic signalresembles the peripheral arterial pressure waveform [12].Analysis of the respiratory variation in pulse oxime-ter waveforms has been proposed for a long time asa technique with which to assess blood volume status inmechanically ventilated patients [13]. In a recent study,
Fig. 5 Receiver operating characteristic (ROC) curves comparingthe ability of ∆PP and ∆PPLET to discriminate responders (cardiacindex increase ≥ 15%) and non-responders to volume expansion(n = 28). The areas under the ROC curve were not significantlydifferent (p = NS)
we demonstrated that a derived plethysmographic index—the respiratory change in pre-ejection period—was asaccurate as ∆PP to assess preload responsiveness in septicmechanically ventilated patients [22]. In the present study,we used ∆PPLET since we postulated that this index mightreflect the respiratory changes in left ventricular strokevolume. Indeed, by reflecting the pulsatile changes inabsorption of infrared light between the light source andthe photo detector of the pulse oximeter, the ‘pulse’ waveis assumed to be the result of the beat-to-beat changes instroke volume transmitted to arterial blood, which wasreported to correlate with ∆PP in mechanically ventilatedpatients [9]. In this respect, ∆PPLET is potentially a markerof respiratory stroke volume variation and thus of volumeresponsiveness [14, 15]. Interestingly, we found thresholdvalues of 12% and 14% that allowed discriminationbetween responder and non-responder patients for ∆PPand ∆PPLET respectively. These values were very closeto the threshold values (13%, 11.8%, 17%, 12%) foundin previous studies examining the significance of ∆PPto predict fluid responsiveness in septic patients [4, 16,22, 23]. It has to be noted that the prediction of fluidresponsiveness was not improved by the combinationof the two measurements (∆PP and ∆PPLET). This maysuggest that these indices give similar information interms of prediction of fluid responsiveness. However, asindicated by the data displayed in Fig. 3, the proportion-ality between ∆PP and cardiac index changes followingvolume expansion was closer to the identity line than wasthe proportionality between ∆PPLET and cardiac index
90
998
changes. These results emphasise the clinical usefulnessof ∆PP not only for predicting volume responsivenessbut also for quantifying the haemodynamic response tofluid challenge, thus confirming the findings of a previousstudy [4]. On the other hand, the advantages of ∆PPLETare its acquisition with a non-invasive technique (pulseplethysmography) and its immediate availability, whichallows accurate assessment of volume responsiveness inmechanically ventilated patients before insertion of anyarterial catheter.
Some limitations of this work should be acknow-ledged. First, we studied sedated patients such that ourresults cannot be extrapolated to patients experiencingspontaneous breathing activity, a condition that is fre-quently encountered in the intensive care unit (ICU).Second, our patients had regular cardiac rhythm, a manda-tory condition for the use of heart–lung interactionindices [1]. Third, we used a tidal volume > 8 ml/kg inour patients and thus we cannot extrapolate our resultsto patients ventilated with lower tidal volume. Indeed,in such conditions of low cyclic changes in intrathoracicand transpulmonary pressures, volume responsivenessmay coexist with low values of ∆PP [16] and presumablyin ∆PPLET. In this regard, in a series of 22 hypotensivepatients ventilated with tidal volumes ranging from 6 to10 ml/kg (median value of 8 ml/kg), Natalini et al. showedthat ∆PPLET values lower than the threshold value of15% poorly predicted volume responsiveness, while all∆PPLET values above 15% were associated with a positiveresponse to fluid challenge [15]. Fourth, we defined thepositive response to volume challenge as an increase incardiac index by more than 15% after fluid administration.We chose 15% because this benchmark was employed in
numerous previous studies which addressed the issue offluid responsiveness [4, 15, 16, 22]. Since the diameterof the aortic annulus is assumed to remain constantduring short-term haemodynamic interventions, we onlymeasured the response of VTIAo to volume challenge.In this respect, the benchmark of 15% increase was farabove the low intraobserver variability of the VTIAo(0.5 ± 0.7%) that we calculated. Fifth, we did not measureabdominal pressure since there was no clinical suspicionof increased abdominal pressure in this series of medicalICU patients suffering from septic shock. Our resultscannot be extrapolated to patients with significant increasein abdominal pressure, since an animal study recentlyshowed that increasing intra-abdominal pressure mayresult in increase in ∆PP [24]. Finally, in our study, werecorded correct pulse oximetry signals in all patients whowere not hypothermic and in whom peripheral vasocon-striction was unlikely. Indeed, in this context of septicshock, vasomotor tone was expected to be reduced andcatecholamines were given in the attempt to restore organperfusion pressure. However, the pulse oximetry signalmight be of poor quality in the presence of hypothermiaor arterial vasoconstriction, although the quality of thedisplayed signal has been improved with the currentgeneration of pulse oximetry devices.
In conclusion, the present study shows that ∆PPLETmay be as valuable as ∆PP for predicting volume respon-siveness in septic patients ventilated with a tidal volumegreater than 8 ml/kg. Since ∆PPLET is obtained from pulseoximetry, a totally non-invasive monitoring technique, itmay represent an attractive method to detect fluid respon-siveness in mechanically ventilated patients in whom arte-rial catheters have not yet been inserted.
References
1. Bendjelid K, Romand JA (2003)Fluid responsiveness in mechani-cally ventilated patients: a reviewof indices used in intensive care.Intensive Care Med 29:352–360
2. Michard F, Teboul JL (2002) Predictingfluid responsiveness in ICU patients:a critical analysis of the evidence.Chest 121:2000–2008
3. Feissel M, Michard F, Mangin I,Ruyer O, Faller JP, Teboul JL (2001)Respiratory changes in aortic bloodvelocity as an indicator of fluid re-sponsiveness in ventilated patients withseptic shock. Chest 119:867–873
4. Michard F, Boussat S, Chemla D,Anguel N, Mercat A, Lecarpentier Y,Richard C, Pinsky MR, Teboul JL(2000) Relation between respiratorychanges in arterial pulse pressure andfluid responsiveness in septic patientswith acute circulatory failure. AmJ Respir Crit Care Med 162:134–138
5. Reuter DA, Felbinger TW, Schmidt C,Kilger E, Goedje O, Lamm P, Goetz AE(2002) Stroke volume variations forassessment of cardiac responsivenessto volume loading in mechanicallyventilated patients after cardiac surgery.Intensive Care Med 28:392–398
6. Awad AA, Ghobashy MA, Stout RG,Silverman DG, Shelley KH (2001)How does the plethysmogram de-rived from the pulse oximeter relateto arterial blood pressure in coro-nary artery bypass graft patients?Anesth Analg 93:1466–1471
7. Murray WB, Foster PA (1996) Theperipheral pulse wave: informationoverlooked. J Clin Monit 12:365–377
8. Shelley KH, Murray WB, Chang D(1997) Arterial-pulse oximetry loops:a new method of monitoring vasculartone. J Clin Monit 13:223–228
91
999
9. Cannesson M, Besnard C, Durand PG,Bohe J, Jacques D (2005) Relationbetween respiratory variations inpulse oximetry plethysmographicwaveform amplitude and arterialpulse pressure in ventilated patients.Crit Care 9:R562–568
10. Alexander CM, Teller LE, Gross JB(1989) Principles of pulse oximetry:theoretical and practical considerations.Anesth Analg 68:368–376
11. Awad AA, Stout RG, Ghobashy MA,Rezkanna HA, Silverman DG,Shelley KH (2006) Analysis ofthe ear pulse oximeter waveform.J Clin Monit Comput 20:175–184
12. Monnet X, Lamia B, Teboul JL (2005)Pulse oximeter as a sensor of fluid re-sponsiveness: do we have our finger onthe best solution? Crit Care 9:429–430
13. Shamir M, Eidelman LA, Floman Y,Kaplan L, Pizov R (1999) Pulseoximetry plethysmographic waveformduring changes in blood volume.Br J Anaesth 82:178–181
14. Cannesson M, Desebbe O, Hachemi M,Jacques D, Bastien O, Lehot JJ (2007)Respiratory variations in pulse oximeterwaveform amplitude are influenced byvenous return in mechanically venti-lated patients under general anaesthesia.Eur J Anaesthesiol 24:245–251
15. Natalini G, Rosano A, Taranto M,Faggian B, Vittorielli E, Bernardini A(2006) Arterial versus plethysmo-graphic dynamic indices to test respon-siveness for testing fluid administrationin hypotensive patients: a clinical trial.Anesth Analg 103:1478–1484
16. De Backer D, Heenen S, Piagnerelli M,Koch M, Vincent JL (2005) Pulsepressure variations to predict fluid re-sponsiveness: influence of tidal volume.Intensive Care Med 31:517–523
17. Levy MM, Fink MP, Marshall JC, Abra-ham E, Angus D, Cook D, Cohen J,Opal SM, Vincent JL, Ramsay G (2003)2001 SCCM/ESICM/ACCP/ATS/SISInternational Sepsis Definitions Confer-ence. Intensive Care Med 29:530–538
18. Ramsay MA, Savege TM, Simpson BR,Goodwin R (1974) Controlled seda-tion with alphaxalone–alphadolone.Br Med J 2:656–659
19. Preisman S, Kogan S, Berkenstadt H,Perel A (2005) Predicting fluidresponsiveness in patients under-going cardiac surgery: functionalhaemodynamic parameters includingthe Respiratory Systolic VariationTest and static preload indicators.Br J Anaesth 95:746–755
20. Stetz CW, Miller RG, Kelly GE,Raffin TA (1982) Reliability of the ther-modilution method in the determinationof cardiac output in clinical practice.Am Rev Respir Dis 126:1001–1004
21. Hanley JA, McNeil BJ (1983)A method of comparing the areasunder receiver operating characteristiccurves derived from the same cases.Radiology 148:839–843
22. Feissel M, Badie J, Merlani PG,Faller JP, Bendjelid K (2005) Pre-ejection period variations predict thefluid responsiveness of septic ventilatedpatients. Crit Care Med 33:2534–2539
23. Vieillard-Baron A, Chergui K, Ra-biller A, Peyrouset O, Page B,Beauchet A, Jardin F (2004)Superior vena caval collapsibil-ity as a gauge of volume sta-tus in ventilated septic patients.Intensive Care Med 30:1734–1739
24. Duperret S, Lhuillier F, Piriou V,Vivier E, Metton O, Branche P, An-nat G, Bendjelid K, Viale JP (2007)Increased intra-abdominal pressureaffects respiratory variations in arte-rial pressure in normovolaemic andhypovolaemic mechanically ventilatedhealthy pigs. Intensive Care Med33:163–171
92
B. MONITORAGE NON INVASIF DE LA CIRCULATION CEREBRALE.
L’objectif prioritaire de la réanimation des patients qui souffrent de des lésions cérébrales est
aujourd’hui le maintien de leur débit sanguin cérébral et surtout de son adéquation avec la
demande en oxygène du cerveau. La technique du Doppler trans-crânien (DTC) est
actuellement utilisée pour le traitement et le suivi thérapeutique de ces malades dans les soins
intensifs. Elle pourrait permettre une adaptation thérapeutique plus précoce chez les patients à
risque d’ischémie cérébrale. Cependant, l’utilisation du DTC demande une bonne
connaissance théorique et pratique de cette technique.
L’auteur présente ici une étude clinique prospective, dans laquelle la mesure du débit sanguin
cérébral chez des patients qui ont bénéficié d’une chirurgie cardiaque a été monitoré par DTC
et au moyen d’une méthode au Xénon radioactif (Novocerebrograph®) dans les suites d’une
circulation extracorporelle hypothermique. Ce travail a montré que, dans les suites d’une CEC
hypothermique, le débit sanguin cérébral mesuré par DTC présentait une relation statistique
faible avec la méthode de mesure au Xénon (r2= 0.26). Par ailleurs, le débit sanguin cérébral
post-CEC était très augmenté lorsqu’il avait été comparé au débit préopératoire. Cette
augmentation n’a pu être expliquée par le réchauffement corporel ou par la valeur de la
PaCO2. Par ailleurs, dans un cas clinique, rapporté chez un patient de cette étude, le DTC a
été proposé comme un indice d’hypoperfusion cérébrale précoce dans les suites d’un arrêt
cardio-circulatoire [130].
Les performances du DTC: rapidité, non invasivité, bonne reproductibilité en font une
technique particulièrement attractive pour la prise en charge en neuro-reanimation, secteur
dans lequel le suivi diagnostique et thérapeutique est crucial pour la prévention de la survenue
des lésions ischémiques secondaires, et donc pour le pronostic. C’est une méthode
d'exploration de la vascularisation intra-cérébrale facile à mettre en œuvre et qui permet
d'apprécier le retentissement hémodynamique de lésions artérielles cervicales, de rechercher
des anomalies hémodynamiques sur les artères de la base du crâne et d'évaluer la
fonctionnalité du polygone de Willis. Dans les suites d’une circulation extracorporelle, des
complications neurologiques peuvent survenir surtout chez certains patients qui présentent
des risques emboliques plus importants (chirurgie de l’aorte associé, plaques et débris intra-
aortique à l’écho transoesophagien, ..). Chez ce type de patients ciblés sur la bases d’un fort
93
potentiel aux complications neurologiques, il peut paraître envisageable de réaliser un DTC
en préopératoire et de faire un suivi dans les premières heures post opératoires afin de déceler
toutes chutes d’amplitudes de vélocités qui pourraient être compatibles avec la survenue d’un
accident embolique cérébrale. Reprints of :
• “Bendjelid K, Poblete B, Baenziger O, Romand J-A. Doppler cerebral blood flow variation during the first 24 postoperative hours following hypothermic cardio pulmonary bypass. Interactive Cardiovascular and Thoracic Surgery 2003; 2: 46-52.” [174]
• “Carbutti G, Romand J-A, Carballo J.S, Bendjelid S-M, Suter P.M, Bendjelid K. Transcranial Doppler.
An early predictor of ischemic stroke after cardiac arrest? Anesth Analg 2003; 97: 1262-5.” [130]
94
Work in progress report – Cardiopulmonary bypass
The effects of hypothermic cardiopulmonary bypass on Doppler cerebralblood flow during the first 24 postoperative hoursq
Karim Bendjelida,*, Beate Pobleteb, Oskar Baenzigerc, Jacques-Andre Romanda
aDivision of Surgical Intensive Care, Department APSIC, University Hospitals of Geneva, CH-1211 Geneva 14, SwitzerlandbDivision of Anaesthesiology, Hospital of Lucerne, Lucerne, Switzerland
cDepartment of Paediatrics, University Hospitals of Zurich, Zurich, Switzerland
Received 24 April 2002; received in revised form 5 November 2002; accepted 7 November 2002
Abstract
To provide understanding of influence of cardiopulmonary bypass (CPB) on cerebral blood flow (CBF), we investigated the effect of CPB
on patients’ cerebral haemodynamic parameters. Twenty-three patients were prospectively enrolled. CBF was estimated by transcranial
Doppler (TCD) to measure blood velocity in the middle cerebral artery (MVMCA), preoperatively (T0) and at four postoperative times (T1,
T2, T3, T4). At times T2, T3 and T4, MVMCA remained at higher levels than T0 (P , 0:05). In the multivariate analysis PaCO2 was
independently associated to MVMCA at times T1 and T2 (P ¼ 0:03, P ¼ 0:01, respectively) and temperature was independently associated
with MVMCA at time T1 (P ¼ 0:02). Thus, the present study showed an increase in CBF after CPB, that was correlated with raised
temperature but not with decrease in haematocrit.
q 2002 Elsevier Science B.V. All rights reserved.
Keywords: Coronary artery bypass grafting; Human clinical study; Heart surgery; Rewarming
1. Introduction
After cardiac surgery with cardiopulmonary bypass
(CPB), a high incidence of neuropsychological impairment
has been reported and new neurological deficits are also
common [1]. Different associated factors have been identi-
fied to be responsible for or co-factors in the neurological
deficit: the presence of significant carotid artery stenosis,
cardiac valve surgery, repeated heart surgery and/or a
history of stroke [2]. Other perioperative factors like mate-
rial or air microembolisms [3] and already altered cerebral
perfusion have also been incriminated.
Cerebral blood flow (CBF) auto-regulation is normally
under tight control [4]. However, during CPB this regulation
may be altered. Indeed, Stephan et al., using pH-stat acid-
base management, have demonstrated that during systemic
hypothermic non-pulsatile cardiopulmonary bypass, a
luxury cerebral perfusion is observed with an increase in
CBF contemporary to a decrease of cerebral metabolism
[5]. Furthermore, during the rewarming period following
CPB, an increased CBF has also been reported in studies
using either transcranial Doppler (TCD) sonography [6,7] or
a cerebral blood flow tracer [8,9]. The observed increase in
CBF was not attributed to a luxury cerebral perfusion but to
an increased cerebral oxygen consumption [7,10].
Because modification of cerebral blood flow during the
immediate post CPB has not been systematically examined,
we designed a prospective observational study to evaluate
by repeated TCD, changes in CBF during the first 24 h after
elective coronary artery surgery. We hypothesized that after
CPB, the observed change in CBF could be explained by a
decrease in haematocrit as well as to the rewarming. More-
over, as different degrees of correlation have been reported
between TCD-derived CBF and the 133Xenon-clearance
method, both techniques were compared postoperatively.
2. Materials and methods
Over a period of 6 months, 25 patients (20 men and five
women) scheduled for elective coronary bypass surgery
were prospectively enrolled in this clinical study and inves-
tigated by means of TCD and xenon cerebral blood flow
(XCBF) methods. All patients had preoperative TCD cere-
bral blood flow measurements of the middle cerebral artery
Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–52
1569-9293/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved.
doi:10.1016/S1569-9293(02)00098-1
www.icvts.org
q Preliminary data have been presented as an oral presentation to the 43rd
Congress of SFAR (Societe Francaise d’Anesthesie Reanimation), Paris
(France), September 21–23, 2001.
* Corresponding author. Tel.: 141-22-382-7452; fax: 141-22-382-7455.
E-mail address: [email protected] (K. Bendjelid).
95
(MCA), the high internal carotid, internal and external caro-
tid, vertebral and ophthalmic arteries. Patients with
evidence of stenosis in one or more cerebral arteries deter-
mined by TCD [11–13] were excluded from the study.
Furthermore, patients with associated cardiac valvular or
pre-existing neurological diseases were also excluded. In
all patients TCD measurements were performed in the
supine position. The study protocol was approved by our
Institutional Ethical Committee and all patients gave written
informed consent the day before surgery. Permission to
insert a jugular bulb catheter to measure jugular venous
bulb saturation for this study was not granted by our Institu-
tional Ethical Committee.
2.1. Patients
Twenty-three patients completed the protocol. One
patient presented a cardiac arrest when weaned from CPB,
a successful 13 min CPR procedure took place. Postopera-
tively, a right hemiplegia was observed. Another patient
arrived in the intensive care unit (ICU) with a body tempera-
ture already . 37:0 8C. These two patients were excluded
from the study. Mean age was 64 ^ 10 years, mean aortic
cross-clamping time was 96 ^ 33 min with a mean bypass
time of 144 ^ 46 min. The mean bypass temperature was
27.5 ^ 1.6 8C. Eleven patients (48%) and six patients (26%)
presented an anamnestic hypertension or diabetes mellitus,
respectively. Mean number of coronary grafts bypassed was
2.8 ^ 0.6.
2.2. Pre-, peri- and postoperative management
Prior to transfer to the operating room, all patients
received premedication with diazepam and morphine.
General anaesthesia was induced and maintained with mida-
zolam, fentanyl and pancuronium bromide, adjusted to body
weight and elimination half time. All patients were moni-
tored with a V-lead continuous ECG. A radial arterial and a
central venous catheter, a nasogastric tube and a Foley
catheter were inserted. A flow-directed pulmonary artery
catheter (PAC) was introduced for clinical management.
After anticoagulation with heparin 300–400 IU/kg to obtain
an activated clotting time (ACT) higher than 600 s, systemic
hypothermic (28 8C) nonpulsatile cardiopulmonary bypass
(CBP) with cannulation of the aorta and the right atrium
associated with cold cardioplegic-induced cardiac arrest
were used. The nonpulsatile CBP flow was set at 2 l
min21 m22 and a mean arterial pressure (MAP) was main-
tained at 60 mmHg using vasoactive drugs (sodium nitro-
prusside or phenylephrine accordingly). A standard roller
pump (Stockert Instrumente, Munich, Germany), a
hollow-fibre membrane oxygenator (D703A compact flow,
Dideco, Mirandola, Italy) and a 40 mm arterial filter (Bent-
ley, Irvine, CA) were used. ACT was maintained at a level
. 600 s by additional administration of heparin during the
bypass. The aim levels of arterial pH and PaCO2 was at 7.40
and 5 kPa. respectively during the hypothermic phase of
CPB. Blood gas samples were measured according to the
alpha-stat method at 37 8C with a blood gas analyser (Nova
Biomedical/Waltham, USA). Weaning from CBP was
achieved with inotropic or vasopressor drug support when
required. At the end of the bypass surgery patients were
rewarmed up to 35 8C (rectal temperature). Heparin was
antagonized by protamine 1 mg/100 IU heparin. At the
end of surgery, the patients were transferred sedated to the
surgical ICU.
The postoperative management consisted of rewarming
the patients using a convective device which generates
heated air flow and weaning from mechanical ventilation
followed by orotracheal extubation as early as possible.
For all study period, blood temperature was measured
using PAC. A fixed 2 mg h21 nitrate infusion was adminis-
tered to every patient who had undergone coronary artery
revascularisation with an internal mammary artery. Shiver-
ing was treated with bolus injections of morphine sulfate or
meperidine. Mean arterial blood pressure was aimed to be at
a level between 60 and 90 mmHg with appropriated therapy,
i.e. hypotension was initially treated by volume replacement
with crystalloid infusions and catecholamines according to
measured cardiac output. Hypertension was treated by nitro-
prusside infusions. Pain was controlled by morphine
sulphate bolus injections and if required sedation was
provided with intermittent bolus injections of midazolam
until extubation. Haematocrit was maintained above 25%
after surgery. Patients remained in bed until surgical drains
were removed (usually 48 h postoperatively).
2.3. Study protocol
2.3.1. TCD measurements
An investigator trained in the transcranial Doppler sono-
graphy method (B.P.) performed a TCD measurement of
both MCA (VMCA) flow velocity. Each measuring point
was made after at least 15 min of supine rest, when a steady-
state haemodynamic condition was achieved (change in
blood pressure and/or cardiac output less than 10%). Dura-
tion of TCD measurement was 20 min per patient (10 min
for each MCA side).The examiner looked for the maximal
blood flow velocity with the probe hold by hand:
† before cardiac surgery (T0);
† at ICU arrival (T1);
† 2 h later (T1 1 2 h; T2);
† when body temperature reached 37 8C (T3);
† and at 24 h after surgery (T4).
T1, T2, T3, and T4 were respectively 166 ^ 55, 293 ^ 74
348 ^ 105 and 1501 ^ 81 min after the end of CPB. The
flow velocity measurements were realized in the supine
position. Doppler signals from MCA were measured over
the temporal bone windows at a depth ranging between 40
and 60 mm using a 2-MHz pulsed TCD-2-64 B Doppler
Ultrasound velocimeter (EME TCD-2-64, Ueberlingen,
K. Bendjelid et al. / Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–52 47
96
Germany) with an integrated Fourier real-time frequency
analyser. The ultrasound transducer was placed just above
the zygomatic arch. The orientation and strength of the
Doppler signal was adjusted to obtain the best possible
signal. In each subject, a constant depth-range and angle
of insonation were kept throughout the study. After indivi-
dual adjustment of Doppler parameters, such as gain,
sample volume and power of ultrasound, these were not
changed over the study period. Maximum flow velocity
was followed continuously with minimum angle between
probe and vessel. At each side TCD measurement, the
higher VMCA was recorded every 5 min over a 10-min
duration and then averaged from this period. The Pulsatility
Index (PI) was then calculated. The VMCA or PI values
were obtained only during end-expiration to avoid respira-
tory fluctuations and all patients observed a sinus rhythm
without arrhythmia. To minimize side differences, mean of
the right and the left middle cerebral artery was calculated.
Mean VMCA (MVMCA) was calculated as the average of
the mean velocity of right and left middle cerebral arteries
(see Table 2). Data were printed on a paper chart. Quality
control of the data acquisition was assessed by a blinded
investigator, the MD Neurologist TCD specialist of the
Geneva University Hospital.
Concomitantly, clinical neurological status, haemody-
namic data (systemic arterial pressure, central venous pres-
sure, pulmonary artery pressure, pulmonary wedge pressure,
and cardiac output) using bedside calibrated monitoring
(Hewlett Packard Monitor M1092A; Meyrin, Switzerland),
ventilatory parameters (Erica, Drager, Lubeck, Germany)
and blood gas exchange data (Nova Biomedical, Waltham,
MA, USA) were recorded. Furthermore, demographic data
and drug administration were also noted at each TCD point
measurements. TCD measurements were performed under
steady-state conditions. Specifically, no modification of
drug dosage, depth of anaesthesia, chest therapy or endo-
tracheal suction were allowed in the 30 min preceding or
during TCD measurements. Weaning of the mechanical
ventilation and orotracheal extubation were conducted
according to the clinical progress of each patient and the
partial arterial pressure of carbon dioxide (PaCO2) which
was maintained at similar levels. The first two TCD
measurements were performed under intermittent positive
pressure ventilation (IPPV) without spontaneous breathing.
Thereafter, the ventilatory mode could vary in each patient
according to clinical progress. Finally, seventeen patients
were extubated at T4 TDC measurement.
2.3.2. Xenon cerebral blood flow measurement
XCBF was measured at ICU arrival (T1). Regional CBF
was measured by the non-invasive intravenous 133Xenon
method with ten extra Novo Cerebrograph 10a cranial
cadmium telluride detectors (Novo Diagnostic Systems,
Bagsvaerd, Denmark), five placed over each hemisphere.133Xenon dissolved in normal saline (0.9%) at a dose of
10 mCi was injected into a central vein. Expired air record-
ings were obtained which adequately reflect the arterial
xenon time course. From the clearance curves the Initial
Slope Index was derived [14]. The CBF value is expressed
as ml per 100 g brain tissue per minute.
2.3.3. Recorded data
The following data were recorded simultaneously to TDC
measurement at each time period in the same order by a second
blinded investigator: body temperature from pulmonary artery
catheter (T8), haematocrit (Htc), blood gas data (pH, PaCO2,
PaO2, except for time 0), cardiac output (CO, except for time
0), and ventilatory settings (ventilatory mode, respiratory
frequency, tidal volume and minute volume, except for time
0). Duration of all data recording was 5 min. Haematocrit was
determined from venous samples and analysed using an
haematology analyser (SYSMEX NE 8000, TOA Medical
Electronics, Kobe, Japan).
2.4. Data analysis
The data expressed as mean ^ SD. Parametric tests were
used only when variables passed normality test (* ¼ 0:05).
A one-way repeated-measures analysis of variances
(ANOVA), followed by a Tukey–Kramer multiple compar-
ison procedure when appropriate, was performed to
compare the measured variables for the five, or four respec-
tively, different time periods (T0, T1, T2, T3, T4). Variables
associated with the MVMCA in the univariate analysis
(defined as P , 0:1) were subsequently analysed in a multi-
ple regression analysis independently in every time period.
This multiple regression analysis was used to identify those
variables which statistically correlate with the CBF veloci-
ties values. At T1, CBF values measured by TCD
(MVMCA) was compared to XCBF using linear regression
analysis. For this, GraphPad Prism version 3.00 for
Windows (GraphPad Software, San Diego, CA, USA,
http://www.graphpad.com) and StatView (SAS Institute
5.0.1, Cary, NC, USA) for PC were used. A P , 0:05 was
considered statistically significant.
3. Results
The data were normally distributed. Haemodynamic
parameters, pH, PaCO2, haematocrit and temperature
changes are shown in Table 1.
Doppler measurements were successfully obtained from
both locations in all patients during every study period. PI
remained stable for the four time points of study. Increased
flow velocities were observed, in all patients except patients
2, 9, 13 and 16, at each time point after surgery compared to
T1 (32% at T2 and . 50% at T3, T4, respectively). At times
T2, T3 and T4 MVMCA remained at higher levels than preo-
peratively (the TCD data are summarized in Table 2 and
Fig. 1).
A significant correlation between PaCO2 and MVMCA
was found at T1 and T2 (r ¼ 0:65, P ¼ 0:0007; r ¼ 0:53,
K. Bendjelid et al. / Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–5248
97
P ¼ 0:009 respectively; Fig. 2a,b) but not at T3 and T4. A
correlation between an increase in T8 and mean VMCA was
found at T1, T2 and T3 (r ¼ 0:58, P ¼ 0:003; r ¼ 0:34,
P , 0:05; r ¼ 0:36, P , 0:05, respectively) but not at T4.
In the multivariate analysis PaCO2 was independently asso-
ciated to MVMCA in time T1 and T2 (P ¼ 0:03, P ¼ 0:01,
respectively) and T8 was independently associated to
MVMCA in time T1 (P ¼ 0:02). Other parameters: heart
rate, blood pressure, central venous pressure, pulmonary
artery pressure, pulmonary capillary occlusion pressure,
cardiac output, pH, PaO2, haemoglobin, haematocrit, venti-
latory mode-parameters and Glasgow score scale had no
significant correlation with MVMCA or PI. The mean
XCBF value expressed as ISI was 37.5 ^ 12.8 ml/100 g
brain tissue per minute. The correlation between XCBF
and MVMCA was r ¼ 0:51, P ¼ 0:01 (Fig. 3).
4. Discussion
In the present study, TCD was used to assess changes in
CBF during the first 24 postoperative hours after CPB. In the
absence of significant haemodynamic changes besides a
17% decrease in mean arterial pressure, MVMCA is
increased at T2 and thereafter in comparison to preoperative
values. This increase was correlated with an increase in
temperature but not with a decrease in haematocrit.
The haematocrit and CBF are inversely related [15], and
different studies have demonstrated that a decrease in
haematocrit correlates with an increase in Doppler blood
velocity in the middle cerebral artery (VMCA) [16–17].
Because haemodilution is used during deep hypothermic
cardiopulmonary bypass to reduce red cell rigidity and
vascular resistance, the observed increase in MVMCA in
the present study could have been attributed to the decreased
viscosity. Nevertheless, MVMCA value at T1 did not
change when compared to T0 whereas haematocrit had
K. Bendjelid et al. / Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–52 49
Table 2
Mean MCA velocity and Pulsatility Index (PI) (n ¼ 23)a
T0 T1 T2 T3 T4
RVMCA (cm/s) Mean 51.8 48.4 57.3 67.3 67
SD 14.9 19 23 18.7 17.4
LVMCA (cm/s) Mean 52.9 46.9 63.3 66.5 72
SD 13.8 18 28.3 22.4 24.3
MVMCA (cm/s) Mean 52.3 47.6 60.3* 66.9YY ** 69.6Y**
SD 12.9 17.6 24.8 17.9 16.2
PI Mean 1.09 1.25 1.07 1.08 1.08
SD 0.19 0.31 0.22 0.22 0.21
a YP , 0:05 vs. T0;YYP , 0:001 vs. T0; *P , 0:05 vs. T1; **P , 0:001
vs. T1.
Fig. 1. Changes of mean ^ SD absolute values of arterial pressure (MAP)
and haematocrit (HTc) over time concomitantly with MVMCA. No signif-
icant relation was observed between MAP or HTc and the flow velocities.§P , 0:05 vs. T0;
§§P , 0:001 vs. T0; *P , 0:05 vs. T1; **P , 0:001 vs.
T1.
Table 1
Haemodynamic changes and other variables (n ¼ 23)a
T0 T1 T2 T3 T4
T8 (8C) Mean 37.0 35.6YY 36.5YY** 36.9** 37.2**
SD 0.3 0.8 1.2 0.9 0.5
MAP (mmHg) Mean 82 91 86 79* 75**
SD 13 16 19 17 9
CO (l/min) Mean 5.5 5.9 5.9 5.7
SD 1.9 1.5 1.8 1.3
PaCO2 (kPa) Mean 4.9 5.6 5.5 5.4
SD 0.8 1.1 0.7 0.6
pH Mean 7.39 7.35 7.35 7.38
SD 0.07 0.06 0.05 0.04
Htc Mean 0.41 0.31YY 0.31YY 0.30YY 0.29YY
SD 0.04 0.04 0.06 0.05 0.03
a T8, temperature; MAP, mean arterial pressure; CO, cardiac output; Htc, haematocrit. Coefficient of variation of cardiac output: T1, 0.36; T2, 0.28; T3, 0.29;
T4, 0.27. YP , 0:05 vs. T0;YYP , 0:001 vs. T0; *P , 0:05 vs. T1; **P , 0:001 vs. T1.
98
dropped from 41.4 to 30.9 (P , 0:001). This phenomena
may be explained by a traumatized brain which does not
respond to changes in haematocrit [18].
Different hypotheses can be proposed to explain the
increased of MVMCA.
X First, as suggested by Von Knobelsdorff et al. [7] and
Croughwell et al. [10], during the rewarming period, conco-
mitant to the awakening from anaesthesia, an increased CBF
is observed accompanied by a decrease in cerebral venous
blood O2 saturation. Indeed, during this period, episodes of
decreased O2 saturation of the blood in the bulbus jugularis
are observed [19]. In the absence of jugularis blood O2
saturation measurement, which was refused by the ethical
committee, we are not able to confirm this increased cere-
bral oxygen consumption. However, physiologically, brain
temperature is 0.5–1 8C more than body core temperature,
and this gap increases by several degrees during cooling and
rewarming on CPB [20,21]. Moreover, Bissonette et al.
have recently demonstrated, with a modified retrograde
jugular bulb catheter, an important increase in brain
temperature up to 6 h after termination of CPB; and this
increase was 2–3 8C higher than mean core temperature
[22]. Since the present study observed a positive correlation
between increase in temperature and increase in MVMCA,
we are tempted to discuss the possible link between the
recorded increase in MVMCA and a post CPB observed
cerebral hyperthermia. In the present study, we have used
the pulmonary artery catheter to measure the temperature
during the rewarming even if it may be assumed that it does
not represent the temperature of brain. However, during
both normothermia and hypothermia, venous mixed
temperature is the most closely parameter correlated to
venous jugular bulb temperature [23]. Indeed, compared
to temperatures from different sites (rectal, tympanic, oeso-
phageal, skin surface and axilla temperatures respectively)
the latter closely reflects brain temperature after CPB [22].
X Second, the utilization of vasodilator drugs such as
nitroglycerine during the postoperative period may be
responsible for the observed increase in MVMCA. Never-
theless, the constant increase of MVMCA without change in
the therapeutic dose do not address this hypothesis. More-
over, unchanged PI in our patients rules out nitroglycerine
as drug playing the major role in the increase of MVMCA.
The potential methodological flaw in the current study
could be the reliability of MVMCA to reflects and monitors
CBF. Indeed, even though different studies have demon-
strated a good correlation between CBF and TCD, they
were not conducted in CPB surgery patients [24]. Weyland
et al. demonstrated that the changes in cerebral perfusion
associated with the rewarming period following CPB could
K. Bendjelid et al. / Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–5250
Fig. 3. Linear regression analysis MVMCA to 133XCBF for the time T2.133XCBF, 133Xenon cerebral blood flow; MVMCA, middle cerebral artery
blood flow velocity.
Fig. 2. (a,b) Linear regression analysis comparing change in PaCO2 and
MVMCA during time 1 and time 2. MVMCA, middle cerebral artery blood
flow velocity.
99
not be accurately measured by TCD monitoring [8].
Conversely, Trivedi et al. found a good correlation between133Xenon clearance technique and TCD [9]. The weak corre-
lation (r2 ¼ 0:26) found in the present study at T1 (2 h 46
min ^ 0 h 55 min after the end of CPB and at
T8 ¼ 35:6 ^ 0:8), between CBF values obtained by intrave-
nous 133Xenon method and TCD-derived may plead for the
poor reliability of MVMCA to reflects and monitors CBF.
However, in the present study, the arterial 133Xenon level
was only estimated with end tidal measurements and this
may have disturb the accuracy of CBF measurements as
patients in the early postoperative CPB usually have shunts
due to atelectases. The authors have compared TCD to this
technique, because the 133Xenon clearance technique is the
easiest to accomplish and the most widely used method
available after CPB. Indeed, it also requires the shortest
time, a factor of importance because physiologic factors
(temperature, arterial carbon dioxide tension, haematocrit,
depth of anaesthesia) which affect CBF often are changing
after CPB [25]. And, as no methods measure accurately
CBF and all measurements are relative approximations
with inherent errors and limitations [25], the patient being
its own control allows us to conclude that a significant
change in MVMCA occurs when flow velocities were
measured pre- and postoperatively after cardiac surgery
with cardiopulmonary bypass.
A second limit of the present study is that, for each
measuring time, the steady state (in term of shivering,
pain and sedation, factors which influence cerebral metabo-
lism and MVMCA) has been most likely influenced by
medications which have been administered. However, as
all these medications decrease cerebral metabolism, we
may suspect that the observed increase in MVMCA is not
related to their effects. Furthermore, it may even be
supposed that increase in MVMCA would have been more
important without the administration of these drugs. Finally,
it should be kept in mind that there could be heterogeneity in
this patient population, resulting in increased MVMCA
variability.
In conclusion, the results of the present study demonstrate
that during the postoperative period following CPB in
haemodynamically stable patients with a normal neurologi-
cal outcome, MVMCA value is increased after the fifth post-
operative hour and does not return to preoperative values at
24 h. However, the present study design does not allow us to
draw conclusions on the physiological mechanisms respon-
sible for the observed increase in MVMCA, if this is not
explained by the rewarming or PaCO2 value.
Acknowledgements
We are grateful for the help in TCD quality control
provided by Dr. J. Le Floch-Rohr (Neurologist TCD specia-
list of the Geneva University Hospital). We are grateful for
the translation support provided by Angela Frei and Miriam
Treggiari. Also, we are grateful for the help in Xenon cere-
bral blood flow procedure provided by Professor D. Slos-
man (Nuclear Medecine Division, Geneva University
Hospitals) and the help in statistical analysis provided by
Dr. P. Merlani.
References
[1] Newman MF, Kirchner JL, Phillips-Bute B, Gaver V, Grocott H,
Jones RH, Mark DB, Reves JG, Blumenthal JA. Longitudinal assess-
ment of neurocognitive function after coronary-artery bypass surgery.
N Engl J Med 2001;344:395–402.
[2] Ricotta JJ, Faggioli GL, Castilone A, Hassett JM. Risk factors for
stroke after cardiac surgery: Buffalo Cardiac-Cerebral Study Group.
J Vasc Surg 1995;21:359–363.
[3] Pugsley W, Klinger L, Paschalis C, Treasure T, Harrison M, Newman
S. The impact of microemboli during cardiopulmonary bypass on
neuropsychological functioning. Stroke 1994;25:1393–1399.
[4] Lassen N. Autoregulation of cerebral blood flow. Circ Res
1964;15:201–204.
[5] Stephan H, Weyland A, Kazmaier S, Henze T, Menck S, Sonntag H.
Acid-base management during hypothermic cardiopulmonary bypass
does not affect cerebral metabolism but does affect blood flow and
neurological outcome. Br J Anaesth 1992;69:51–57.
[6] Endoh H, Shimoji K. Changes in blood flow velocity in the middle
cerebral artery during nonpulsatile hypothermic cardiopulmonary
bypass. Stroke 1994;25:403–407.
[7] Von Knobelsdorff G, Hanel F, Werner C, Schulte am Esch J. Jugular
bulb oxygen saturation and middle cerebral blood flow velocity
during cardiopulmonary bypass. J Neurosurg Anesthesiol
1997;9:128–133.
[8] Weyland A, Stephan H, Kazmaier S, Weyland W, Schorn B, Grune F,
Sonntag H. Flow velocity measurements as an index of cerebral blood
flow. Validity of transcranial Doppler sonographic monitoring during
cardiac surgery. Anesthesiology 1994;81:1401–1410.
[9] Trivedi UH, Patel RL, Turtle MR, Venn GE, Chambers DJ. Relative
changes in cerebral blood flow during cardiac operations using xenon-
133 clearance versus transcranial Doppler sonography. Ann Thorac
Surg 1997;63:167–174.
[10] Croughwell ND, Frasco P, Blumenthal JA, Leone BJ, White WD,
Reves JG. Warming during cardiopulmonary bypass is associated
with jugular bulb desaturation. Ann Thorac Surg 1992;53:827–832.
[11] Aaslid R, Markwalder TM, Nornes H. Noninvasive transcranial
Doppler ultrasound recording of flow velocity in basal cerebral
arteries. J Neurosurg 1982;57:769–774.
[12] Grolimund P, Seiler RW, Aaslid R, Huber P, Zurbruegg H. Evaluation
of cerebrovascular disease by combined extracranial and transcranial
Doppler sonography. Experience in 1,039 patients. Stroke
1987;18:1018–1024.
[13] Newell DW, Aaslid R. Transcranial Doppler: clinical and experimen-
tal uses. Cerebrovasc Brain Metab Rev 1992;4:122–143.
[14] Risberg J, Ali Z, Wilson EM, Wills EL, Halsey JH. Regional cerebral
blood flow by 133xenon inhalation. Stroke 1975;6:142–148.
[15] Humphrey PR, Du Boulay GH, Marshall J, Pearson TC, Russell RW,
Symon L, Wetherley-Mein G, Zilkha E. Cerebral blood-flow and
viscosity in relative polycythaemia. Lancet 1979;2:873–877.
[16] Brass LM, Pavlakis SG, DeVivo D, Piomelli S, Mohr JP. Transcranial
Doppler measurements of the middle cerebral artery. Effect of hema-
tocrit. Stroke 1988;19:1466–1469.
[17] Bruder N, Cohen B, Pellissier D, Francois G. The effect of hemodilu-
tion on cerebral blood flow velocity in anesthetized patients. Anesth
Analg 1998;86:320–324.
[18] Todd MM, Wu B, Warner DS. The hemispheric cerebrovascular
response to hemodilution is attenuated by a focal cryogenic brain
injury. J Neurotrauma 1994;11:149–160.
K. Bendjelid et al. / Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–52 51
100
[19] Nakajima T, Kuro M, Hayashi Y, Kitaguchi K, Uchida O, Takaki O.
Clinical evaluation of cerebral oxygen balance during cardiopulmon-
ary bypass: on-line continuous monitoring of jugular venous oxyhe-
moglobin saturation. Anesth Analg 1992;74:630–635.
[20] Stone JG, Young WL, Smith CR, Solomon RA, Wald A, Ostapko-
vitch N, Shrebnick DB. Do standard monitoring sites reflect true brain
temperature when profound hypothermia is rapidly induced and
reversed? Anesthesiology 1995;82:344–351.
[21] Kurth CD, O’Rourke MM, O’Hara IB, Uher B. Brain cooling effi-
ciency with pH-stat and alpha-stat cardiopulmonary bypass in
newborn pigs. Circulation 1997;96:358–363.
[22] Bissonnette B, Holtby HM, Davis AJ, Pua H, Gilder FJ, Black M.
Cerebral hyperthermia in children after cardiopulmonary bypass.
Anesthesiology 2000;93:611–618.
[23] Crowder CM, Tempelhoff R, Theard MA, Cheng MA, Todorov A,
Dacey Jr RG. Jugular bulb temperature: comparison with brain
surface and core temperatures in neurosurgical patients during mild
hypothermia. J Neurosurg 1996;85:98–103.
[24] Halsey JH, McDowell HA, Gelmon S, Morawetz RB. Blood velocity
in the middle cerebral artery and regional cerebral blood flow during
carotid endarterectomy. Stroke 1989;20:53–58.
[25] Young WL, Newman MF, Amory D, Reves JG. Cerebral blood flow
values during cardiopulmonary bypass: relatively absolute or abso-
lutely relative? Ann Thorac Surg 1995;59:558–561.
K. Bendjelid et al. / Interactive Cardiovascular and Thoracic Surgery 2 (2003) 46–5252
101
Transcranial Doppler: An Early Predictor of Ischemic StrokeAfter Cardiac Arrest?Giuseppe Carbutti, MD*, Jacques-Andre Romand, MD, FCCM*, Jean-Sebastien Carballo, MD†,Si-M’hamed Bendjelid, MD‡, Peter M. Suter, MD, FRCA*, and Karim Bendjelid, MD, MS*
*Division of Surgical Intensive Care, Department APSIC, and †Department of Internal Medicine, University Hospitals ofGeneva, Geneva, Switzerland; and ‡Department of Radiology, Hospital of Moulins, Moulins, France
A 69-yr-old woman was admitted to the intensive careunit after cardiac surgery. Immediately after the discon-tinuation of cardiopulmonary bypass, she had a circula-tory arrest. A 13-min open-chest cardiac massage wasfollowed by 70 min of cardiopulmonary bypass. Righthemiplegia and right extensor plantar reflex were notedafter the patient awakened. She had been included in aprospective study protocol measuring, before and after
surgery, cerebral blood flow with transcranial Doppler(TCD). The data were retrospectively analyzed, and it wasestablished that the TCD had recorded cerebral perfusiondefects. This is the first case of acute ischemic stroke aftercardiac arrest with retrospective documentation of asym-metrical cerebral blood flow by a systematic postopera-tive TCD recording.
(Anesth Analg 2003;97:1262–5)
T he diagnosis of ischemic stroke in the acutephase usually relies on a computed tomography(CT) scan, which is, however, often normal at the
start of symptoms or which can show lacunae lesionsduring the first 24–48 h. Early diagnosis can be ob-tained with angio-magnetic resonance imaging (MRI)(1), diffusion-weighted imaging (2), or both. Neverthe-less, the clinical presentation remains the determiningelement, because symptoms are the first to draw at-tention. However, in the intensive care unit (ICU),symptoms of stroke can be masked because of bothsedation and muscle relaxation or because of otherserious alterations of mental state or consciousness.
Here, we report the case of a patient who, aftercardiac surgery, presented a right hemiplegia afterawakening. Interestingly, because the patient hadbeen included in a study protocol measuring pre- andpostoperative cerebral blood flow with a transcranialDoppler (TCD) (3), we had the possibility of retrospec-tively analyzing recordings and noting that thetranscranial examination showed cerebral perfusiondefects.
Case ReportA 69-yr-old woman was admitted to the ICU after a qua-druple coronary artery bypass graft surgery performed foran increasingly symptomatic angina pectoris at rest. Thepatient had a history of systemic hypertension, hypercholes-terolemia, and a cholecystectomy. The preoperative cardiacfunction was moderately diminished (ejection fraction at44%). Neck vessel echo-Doppler examination demonstrateddiscrete atheromatoses of bilateral internal carotid arteries.
Immediately after discontinuation of the cardiopulmonarybypass (CPB) and 10 min after the administration of prota-mine, the patient rapidly developed pulmonary artery hy-pertension (systolic arterial blood pressure �8 kPa meas-ured with a pulmonary artery catheter), with subsequentcirculatory arrest. A 13-min open-chest cardiac massage(mean arterial blood pressure (MAP) between 30 and 40 mmHg) was followed by 70 min of CPB, which was finallyweaned under epinephrine (0.11 �g · kg�1 · min�3). The to-tal CPB time was 285 min, and aortic cross-clamp time was210 min. The thoracic cavity was initially kept open becauseof hemodynamic instability and cardiac dilation and wasclosed after 48 h. Hemodynamic stability was progressivelyobtained, and epinephrine was weaned within 6 h after CPB.A Ramsay score of 3 was targeted until tracheal extubation(7 h after CPB) with intermittent bolus injections of midazo-lam (total of 5 mg). The patient’s renal function was normal.
Awakening with marked agitation was noted after 24 h,with slight motor activity and reflex asymmetry. Righthemiplegia and an extensor plantar reflex were confirmedthe next day. No intracranial hemorrhages or other patho-logic lesions were noted on a noninjected cerebral CT scan(Fig. 1A). The patient was tracheally extubated on the thirdday, after which aphasia, swallowing difficulties, and con-fusion were noted. Clinical improvement enabled the pa-tient to be discharged from the ICU on the sixth day. A right
Accepted for publication June 9, 2003.Address correspondence and reprint requests to Karim Bendjelid,
MD, Chef de Clinique, Surgical Intensive Care Division, UniversityHospital of Geneva, CH-1211 Geneva 14, Switzerland. Addresse-mail to [email protected].
DOI: 10.1213/01.ANE.0000083373.19574.64
©2003 by the International Anesthesia Research Society1262 Anesth Analg 2003;97:1262–5 0003-2999/03
102
amotor and sensory hemisyndrome and swallowing diffi-culties were still present, but with recovery of verbal com-munication. A cerebral CT scan done 12 days after surgeryrevealed a voluminous hypodensity of the left midbrain(protuberance) associated with signs of infarction (Fig. 1B).
This case report is of interest because the patient hadagreed to take part in a prospective study measuring cere-bral blood flow during the pre- and postoperative periodafter CPB (3). Indeed, she had four TCD examinations: pre-operatively, immediately on arrival from the operatingroom, after the warming period, and again 18 h after oper-ation. The results clearly showed a significant reduction inthe flow speed measured in the left hemisphere (middlecerebral artery blood flow mean velocity), the same side onwhich the lesion developed. This was observed on the firstmeasurement on arrival in the ICU. It then indicated dimin-ished left cerebral perfusion during the early stage aftercirculatory arrest (Fig. 2). Thus, the diagnosis of ischemicstroke could have been suspected before the clinical picturebecame apparent on the patient’s awakening.
DiscussionTCD is currently increasingly recognized as a usefultechnique for early detection of cerebral blood flowabnormalities during the acute phase of stroke. The
Figure 2. Changes in values of blood flow velocities in the right and leftcerebral arteries over time. Solid line—R.SVMCA � systolic velocity ofthe right middle cerebral artery; R.MVMCA � mean velocity of theright middle cerebral artery; R.DVMCA � diastolic velocity of the rightmiddle cerebral artery. Dashed line—L.SVMCA � systolic velocity ofthe left middle cerebral artery; L.MVMCA � mean velocity of the leftmiddle cerebral artery; L.DVMCA � diastolic velocity of the left mid-dle cerebral artery. The figure clearly shows a significant decrease inleft velocities after surgery with a compensatory increase in rightvelocities after the warming period in the intensive care unit (ICU2).preop � before surgery; ICU1 � immediately on arrival from theoperating room; ICU3 � 18 h after operation.
Figure 1. A, No intracranial hemorrhageor other pathologic lesion was noted ona first (Day 2) noninjected cerebral com-puted tomography scan. Retrospec-tively, it can show a mild left hypoden-sity of the midbrain (protuberance;arrow). B, A cerebral computed tomog-raphy scan done 12 days after surgeryrevealed a voluminous left hypodensityof the midbrain (protuberance; arrow)associated with signs of infarction.
ANESTH ANALG CASE REPORTS 12632003;97:1262–5
103
procedure is simple to do at the bedside and is non-invasive, quick, reproducible, and sensitive (3–6). In-deed, asymmetrical cerebral blood flow observed byTCD is a sign that can evoke a thromboembolic lesionin the acute phase of ischemic stroke (7), and, there-fore, the examination could provide a basis for earlyreperfusion/thrombolytic therapy (8), which has beenshown to significantly improve outcome after stroke(9).
This case is of particular interest because the patient,whose surgery was complicated by a cardiocirculatoryarrest, had been included in a study protocol thatsystematically measured cerebral blood flow by usingTCD. The patient was intubated and sedated, and,therefore, the detection of clinical stroke symptomswas delayed. However, TCD had detected a clear flowasymmetry two hours after the arrest. It seems clear tous that the approach to and management of stroke inthis case could have been optimized with the intro-duction of a prophylactic therapy applied within thefirst 48 hours. Indeed, even if the clinical examinationshowed the neurological defect at an early postoper-ative stage, the TCD abnormalities were observed be-fore the patient awakened from the anesthesia, whichcould have generated further examinations (MRI, an-giography, and so on) to confirm the diagnosis andtailor an anticipated specific reanimation. AlthoughTCD can record cerebral perfusion defects, it cannotconfirm the etiology of an ischemic stroke. Therefore,in this setting it is difficult to define an appropriatetherapeutic goal. Nevertheless, the primary aim ofmanaging patients with acute brain injury in the ICUis to minimize secondary injury by maintaining opti-mal cerebral perfusion, oxygenation, and normal gly-cemia and natremia (10).
The postoperative neurologic description of our pa-tient disputes the hypothesis of a single lesion ob-served on the CT scan. Nevertheless, a severe periop-erative hemodynamic alteration can be the startingpoint of multiple emboli, even if just a single midbrainlesion is confirmed by the CT scan. In contrast, be-cause the duration of CPB was long, we also cannotexclude that this situation could have been the startingpoint of multiple gaseous embolisms (a rare phenom-enon in non-open-heart surgery). The inconsistent re-lation between TCD recording data and the left mid-brain hypodensity on CT scan can also be attributed toa reverse flow perfusion through the Willis polygon.Finally, a cerebral diaschisis phenomenon cannot beexcluded, because local cerebral dysfunction in re-gions remote from the principal cerebral ischemic le-sions in the acute stage of stroke has been demon-strated (11,12).
Evidently, the cerebral CT scan performed 12 daysafter surgery was not typical of persistent middle ce-rebral artery flow abnormality, because this perfusiondefect is usually responsible for severe hypodensities.However, in the absence of cerebral metabolic assess-ment or MRI during this period, we cannot infer theeffective cerebral blood flow of the middle cerebralartery. Thus, the reduction of the flow recorded on theleft side was possibly not severe enough to causehypodensity in the concerned cerebral region. More-over, the immediate hemodynamic improvement (car-diac output; see Fig. 3) could have participated in thisparadoxical status.
Therefore, in anesthesia and ICUs, given the factthat clinical neurological examinations are often oflimited value, TCD could be used systematically formonitoring deeply sedated patients who have under-gone major hemodynamic instabilities. This new toolfor neurological surveillance of postoperative and ICUpatients could be further explored in careful clinicalinvestigations.
The authors are grateful for the translation provided by SamiaBrunner.
References1. Stadnik TW, Osteaux M. Actual trends in diagnosis/
thrombolytic therapy of acute cerebral stroke. Eur J Radiol2001;38:28–40.
2. Schabitz WR, Fisher M. Diffusion weighted imaging for acutecerebral infarction. Neurol Res 1995;17:270–4.
3. Bendjelid K, Poblete B, Baenziger O, Romand J-A. Dopplercerebral blood flow variation during the first 24 postoperativehours following hypothermic cardiopulmonary bypass. Interac-tive Cardiovascular and Thoracic Surgery 2003;2:46–52.
4. Baracchini C, Manara R, Ermani M, Meneghetti G. The quest forearly predictors of stroke evolution: can TCD be a guiding light?Stroke 2000;31:2942–7.
Figure 3. Changes of cardiac output (CO), Paco2, temperature (T°),and mean arterial blood pressure (MAP) during the rewarmingperiod. preop � before surgery; ICU1 � immediately on arrivalfrom the operating room; ICU2 � after the warming period; ICU3 �18 h after operation.
1264 CASE REPORTS ANESTH ANALG2003;97:1262–5104
5. Akopov S, Whitman GT. Hemodynamic studies in early isch-emic stroke: serial transcranial Doppler and magnetic resonanceangiography evaluation. Stroke 2002;33:1274–9.
6. Leanyvari Z, Vastagh I, Fulesdi B, et al. Computed tomographicand transcranial Doppler sonographic findings in acute andsubacute phases of middle cerebral artery strokes. J Clin Ultra-sound 2002;30:33–7.
7. Toni D, Fiorelli M, Zanette EM, et al. Early spontaneous im-provement and deterioration of ischemic stroke patients: a serialstudy with transcranial Doppler ultrasonography. Stroke 1998;29:1144–8.
8. Alexandrov AV, Bladin CF, Norris JW. Intracranial blood flowvelocities in acute ischemic stroke. Stroke 1994;25:1378–83.
9. Kwiatkowski TG, Libman RB, Frankel M, et al. Effects of tissueplasminogen activator for acute ischemic stroke at one year:National Institute of Neurological Disorders and Stroke Recom-binant Tissue Plasminogen Activator Stroke Study Group.N Engl J Med 1999;340:1781–7.
10. Teener JW, Raps EC, Galetta SL. Intensive care neurology. CurrOpin Neurol 1994;7:525–9.
11. Izumi Y, Haida M, Hata T, et al. Distribution of brain oedema inthe contralateral hemisphere after cerebral infarction: repeatedMRI measurement in the rat. J Clin Neurosci 2002;9:289–93.
12. Takasawa M, Hashikawa K, Ohtsuki T, et al. Transient crossedcerebellar diaschisis following thalamic hemorrhage. J Neuro-imaging 2001;11:438–40.
ANESTH ANALG CASE REPORTS 12652003;97:1262–5
105
AGENDA POUR LA RECHERCHE FUTURE
A. DEFINITION DE LA REPONSE AU REMPLISSAGE VASCULAIRE PAR LA
VARIATION DE LA SATURATION VEINEUSE EN OXYGENE.
L’étude de la réponse au remplissage vasculaire dépend de la mesure directe du débit
cardiaque (DC) avant et après infusion volumique. De ce fait, jusqu’à présent, définir si un
patient est répondeur à l’infusion de volume ou non nécessite qu’il soit équipé d’un cathéter,
invasif, de Swan-Ganz ou qu’un médecin échocardiographeur soit toujours disponible. Le
patient est défini comme répondeur (R) s’il augmente son DC de plus de 15 % (valeur la plus
utilisée dans la littérature) et non répondeur (NR) si cette dernière valeur n’est pas atteinte
[9]. Une autre méthode qui pourrait être utile dans ce contexte serait la mesure de la SvO2
centrale (SCvO2, SvO2 dans la veine cave supérieure) avant et après une expansion
volumique. Cette méthode serait intéressante, puisque le thérapeute pourrait s’assurer de la
réponse au remplissage vasculaire en utilisant seulement une voie veineuse centrale, moins
invasive qu’un cathéter artériel pulmonaire.
En effet, il est connu que la SvO2 est corrélée au DC, en dehors de situations biens définies
[175]. Par ailleurs, de multiples études on montré que la ScvO2 est très bien corrélée à la SvO2
en dehors du choc septique [176]. De ce fait, on peut s’attendre à ce que les changements de
valeurs de ScvO2 (ΔScvO2) dans les suites d’une expansion volumique traduisent les
changements de valeurs du DC (ΔDC) [177] (transitivité mathématique, si a = b et si b = c
alors automatiquement a = c). Dans un modèle animal, ce concept a été récemment confirmé
[178].
Nous proposons d’entreprendre une étude pour vérifier l’hypothèse que la variation de la
ScvO2 (ΔScvO2), après infusion de volume, peut être une méthode utile pour définir la
réponse au remplissage vasculaire et pour distinguer les patients répondeurs des non-
répondeurs, qu’ils soient ou non placés sous ventilation mécanique.
106
B. IMPACT DU POIDS DU PATIENT SUR LES PARAMETRES
HEMODYNAMIQUES MESURES.
L’obésité est un désordre métabolique de plus en plus fréquent dans les sociétés occidentales.
Son incidence est majeure en terme de santé publique. Les conséquences physiopathologiques
de l’obésité sévère, définie par un index de masse corporelle (poids (kg) / hauteur 2 (m2))
supérieur à 31, mettent en cause tous les grands systèmes physiologiques. L’obésité augmente
le risque de cardiopathie et d’accident vasculaire cérébral, et elle s’associe à une morbidité et
à une mortalité anesthésiques accrues. Ce risque est d’autant plus grand lorsqu’il s’agit d’une
répartition androïde des graisses. Plus que le diabète ou la stéatose hépatique, ce sont les
altérations cardio-vasculaires et respiratoires qui doivent préoccuper le médecin des soins
intensifs.
Pour le système cardio-vasculaire, le débit cardiaque est augmenté en proportion de
l’accroissement de la masse grasse (0,1 L.min-1 sont nécessaires pour perfuser chaque kilo de
graisse) [179-181]. La fréquence cardiaque n’étant pas modifiée, cette augmentation porte
d’abord sur le volume d’éjection systolique. Ainsi, le ventricule gauche est souvent dilaté et
ses parois épaissies. La tolérance à l’effort est amoindrie, ce qui fait craindre un stress accru
dans la période opératoire [179, 180]. Ces modifications, observées en dehors des
complications cardio-vasculaires de l’obésité, sont corrigées par la perte de poids.
L’hypertension artérielle est fréquente chez l’obèse et le risque de survenue d’une défaillance
ventriculaire gauche est important. Ainsi, chez le sujet très obèse vieillissant, surtout s’il
existe des signes cliniques évocateurs de cardiopathie, le recours à une évaluation précise de
la fonction ventriculaire avant une intervention s’impose.
L’impact du poids d’un patient sur les paramètres hémodynamiques n’est pas bien connu
[182]. En effet, les valeurs hémodynamiques aux soins intensifs sont indexées à la surface
corporelle selon la formule de Dubois, mais selon le type de poids utilisé (l’actuel, l’idéal
(Lorentz) ou l’ajusté) les valeurs risquent d’être différentes [183]. Cette question est très
importante, car les patients obèses représentent un pourcentage non négligeable des malades
admis dans les soins intensifs, alors que cet état est clairement considéré comme un facteur de
surmortalité [184]. Par ailleurs, chez ces patients, la question de savoir si leur monitorage
hémodynamique nécessite que les valeurs obtenues soient indexées ou non au poids corporel
107
n’est pas résolue [185]. Une première étude rétrospective chez des patients hospitalisés en
cardiologie a montré que le poids du patient ne perturbait pas l’interprétation des valeurs
hémodynamiques [186]. Reste qu’à ce jour, une telle information n’est pas disponible dans le
contexte des soins intensifs, d’autant plus que, dans cette situation particulière, le poids d’un
patient peut varier de plusieurs kilogrammes en 24 heures.
L’auteur de la présente thèse a pour objectif de réaliser une étude, à la fois prospective et
rétrospective, dans les soins intensifs, afin d’estimer l’impact du poids du patient sur
l’interprétation des valeurs d’hémodynamiques (débit cardiaque, volume d’éjection systolique
et résistances vasculaires systémique et pulmonaire).
L’étude se composera de deux parties : la première sera rétrospective, durant laquelle toutes
les mesures hémodynamiques par cathéter artériel pulmonaire réalisées dans le service des
soins intensifs des HUG durant l’année 2005 seront récoltées en association avec les
déterminants de la surface corporelle et du BMI chez chaque patient. Par la suite, l’impact de
la surface corporelle et du BMI sur les valeurs obtenues rétrospectivement sera analysé [186].
Ensuite, nous réaliserons une étude clinique prospective, dans laquelle nous étudierons chez
chaque patient l’impact de la modification du poids au cours de l’hospitalisation sur les
valeurs hémodynamiques.
108
C. MONITORAGE ET QUANTIFICATION NON INVASIVE DE LA
MICROCIRCULATION.
Une atteinte sévère et progressive de la micro-circulation et de la perfusion tissulaire du
patient des soins intensifs est un fait établi lors des dysfonctions d’organes multiples. Cette
altération de la fonction micro-circulatoire se fait principalement au niveau de la sphère
gastro-intestinale avec récemment l’identification du rôle des médiateurs de l'inflammation
dans cette dysfonction progressive. Le status cardio-circulatoire du patient présentant un état
de choc est une situation qui est adaptative à la sédation, l’anxiété, le stress et l’affection
causale. Il est donc fort probable que le monitorage de la macro-circulation du patient choqué
est un objectif insuffisant voir limiter vu que cette mesure renseigne sur les apports mais ne
prend pas en compte les besoins. A l’opposé, le monitorage hémodynamique micro-
circulatoire (et/ou tissulaire) accède à une information qui éclair sur la balance apports
sanguins/besoins cellulaires. De ce fait, depuis de nombreuses années, différents auteurs
essaient de monitorer la micro-circulation de manière non invasive chez les patients de soins
intensifs [187]. Le concept vise à permettre d’éviter ainsi un nombre important de
complications liées aux procédures invasives, techniques qui sont régulièrement mises en
cause comme facteurs de surmortalité [72]. Dans cette optique et suite aux découvertes
récentes d’altérations microcirculatoires chez certains groupes de patients des soins intensifs
(sepsis, choc cardiogène, …) [188, 189], il importe de plus en plus de développer nos
connaissances dans ce domaine et de déterminer quel type de monitorage utiliser dans notre
pratique quotidienne.
La microcirculation comprend l’ensemble des vaisseaux de diamètre inférieur à 200 μm. Elle
représente la surface endothéliale la plus importante du corps (> 0.5 km2). Celle ci participe à
plusieurs tâches, dont la plus importante est l’apport d’oxygène aux tissus. Cet apport en
oxygène est en partie déterminé par les valeurs macro-circulatoires, mais il dépend également
de facteurs spécifiques à ce réseau vasculaire. L’apport d’oxygène aux cellules ne peut être
prédit de manière fiable uniquement par les valeurs macro-circulatoires habituelles, comme le
montrent plusieurs études cliniques dans lesquelles on constate que bien que les valeurs
macro-hémodynamique soient optimalisées, il persiste des anomalies microcirculatoires et
des signes indirects d’apport insuffisant d’oxygène aux cellules [188, 189].
109
Ces altérations de la micro-circulation sont retrouvées dans de nombreuses situations et elles
jouent probablement un rôle important dans la persistance de la souffrance cellulaire et dans
le développement des dysfonctions multiples d’organes [190-193]. Dans les situations de
sepsis sévère, probablement la situation la plus étudiée dans ce domaine, De Backer et al. ont
mis en évidence, de manière semi-quantitative, une diminution significative de la densité
vasculaire sublinguale, ainsi qu’une diminution de la proportion de microvaisseaux (<20 μm)
[194]. Ces altérations étaient plus sévères dans le groupe de patient avec le pronostic le plus
sombre. Cette propriété ‘pronostique’ est retrouvée dans une autre étude prospective
observationnelle où, dans un contexte de choc septique, ces altérations ont été suivies de
manière quotidienne depuis le début du choc jusqu'à sa résolution [190].
En dehors du sepsis, on trouve ces altérations dans d’autres situations, comme l’insuffisance
cardiaque sévère. En effet, dans une étude sur un collectif de 40 patients, dans les 48 h de leur
admission aux soins intensifs, une diminution de la proportion de petits vaisseaux perfusés
avait été objectivée [189]. Ces altérations paraissaient être plus sévères chez les patients non
survivants.
Bien qu’encore peu étudiée dans d’autres situations que le sepsis, la présence d’altérations de
la micro-circulation est suspectée de manière indirecte, notamment par le recours à d’autres
méthodes, comme par capnométrie tissulaire [195] ou par la near-infrared spectroscopy
(NIRS ; diminution de la saturation en oxygène dans les tissus musculaires), par exemple dans
le choc post traumatique. Ainsi, les anomalies détectées sont corrélées à la sévérité de
l’atteinte tissulaire avec la possibilité d’un recrutement des micro-vaisseaux après une
thérapeutique adaptée [196]. Ces constatations ouvrent donc des champs thérapeutiques
nouveaux [187]. En effet, l’optimalisation de la microcirculation permettra probablement
d’améliorer le pronostic des patients, en ciblant de manière plus précise les traitements, et en
évitant la perpétuation de la souffrance cellulaire et le développement de dysfonctions
multiples d’organes [187]. Dans cette idée de recrutement des micro-vaisseaux, plusieurs
agents ont été étudiés, principalement chez l’animal. Chez l’homme de manière non invasive
par une technique de « orthogonal polarization system » sublinguale, Spronk et coll. ont
observé une amélioration significative de la microcirculation chez des patients en choc
septique par l’administration de nitroglycérine [197].
110
L’étude de la micro-circulation permettrait également de mieux cibler les objectifs de la
réanimation, actuellement fondée surtout sur une PAM cible, et de manier avec plus de
finesse les différentes médications habituelles, comme les catécholamines, et le remplissage
vasculaire, en particulier dans les situations d’hypovolémie. Le suivi de l’état de la
microcirculation pourrait permettre d’évaluer et de mieux comprendre les effets de nouvelles
substances vasoactives, dont les effets néfastes ne sont pas encore totalement connu, comme
c’est le cas de la vasopressine. En effet, cette molécule est très efficace en terme de correction
de l’hypotension, mais elle s’associe à une diminution très importante du débit de la
microcirculation.
Plusieurs techniques permettent d’approcher de manière plus précise l’apport en oxygène
,notamment par l’évaluation de la microcirculation. Parmis celles-ci la méthode OPSI pour
« orthogonal polarisation spectral imaging » [194]. Cette technique, utilisée dans la plupart
des études cliniques sur ce sujet, permet une visualisation directe de la microcirculation
sublinguale, et est validée par comparaison avec d’autres méthodes. Basées sur cette
technique, de récentes améliorations ont permis de développer des approches encore plus
fiables telle que le SDF pour « Sidestream Dark Field Imaging » [198]. Dans cette technique,
le tissu est illuminé par une source de lumière (longueur d’onde de 530 nm). La lumière
reflétée est filtrée et « zoomée » avec une absorption de la lumière par les globules rouges qui
permet par réfraction la visualisation directe de la microcirculation. Grâce à ce système, les
anomalies fines de la microcirculation peuvent être évaluées de manière dynamique et, pour
l’instant, semi-quantitative. Ces anomalies sont enregistrées sur une caméra vidéo par
l’intermédiaire d’un appareil de petite taille conçu pour une utilisation au lit du malade. Une
des grandes difficultés de cette technique est de déterminer un index témoin de la perfusion
tissulaire. Il existe actuellement plusieurs index, utilisés dans les différentes études humaines
et validés chez l’homme.
La valeur pronostique de ces anomalies dans le choc septique est de plus en plus étudiée
[199]. Cependant, la présence ou non de ces anomalies dans les différentes pathologies
rencontrées aux SI n’est pas connue. Il parait donc important, dans un domaine en pleine
expansion, de déterminer si ces anomalies peuvent être retrouvées dans certains groupes de
malades ou chez l’ensemble des patients des SI. L’auteur de la présente thèse a l’ambition,
avec des collaborateurs des soins intensifs de Genève, de mettre sur pied une étude clinique
111
prospective qui permettra d’approcher cet objectif. L’autre objectif d’une telle même étude
serait de déterminer si le suivi des altérations de la micro-circulation sublinguale constitue un
marqueur précoce de l’apparition de complications.
D. INFLUENCE DU VOLUME COURANT ET DE LA FREQUENCE
RESPIRATOIRE SUR LES INDICES DYNAMIQUES DE REMPLISSAGE
VASCULAIRE. ETUDE ANIMALE.
Les déterminants de la variation de pression artérielle systolique lors d’une ventilation par
pression positive résidents dans le status volumique et l’amplitude du volume courant utilisé
[200, 201]. En effet, plusieurs études ont démontré qu’un volume courant au dessous de 8
ml/kg pouvait abolir la variation de pression artérielle systolique, présente normalement lors
d’une hypovolémie [200, 201]. De Backer et al, ont démontré dans une investigation portant
sur 60 patients placés en venrtilation mécanique que la variation de la pression pulsée était un
bon indice prédictif de la réponse au remplissage, mais seulement si le volume courant utilisé
était situé au dessus de 8 ml/kg [200]. Cette situation rend le concept d’indices dynamiques
difficilement utilisables chez les patients présentant un syndrome de détresse respiratoire
aiguë et ventilés par une pression positive avec un volume courant de 6-8 ml/kg.
Néanmoins, chez des patients porteurs de cette dernière affection, l’instauration d’une
ventilation mécanique à faible volume courant pour limiter la pression plateau, impose au
clinicien de ventiler les poumons avec une fréquence respiratoire plus élevée, afin de ne pas
limiter la ventilation-minute. De ce fait, la faible utilité des indices dynamiques comme
indicateurs de remplissage vasculaire dans cette situation peut s’expliquer aussi par le fait,
que dans cette dernière situation, un cycle inspiratoire assiste moins de battement cardiaques
(rapport : fréquence cardiaque/fréquence respiratoire).
Nous avons donc proposé une étude animale pour explorer cette question.
Il s’agit d’une étude chez le porc (25-30 kg) profondément sédaté et ventilé en pression
positive, pendant une normovolémie suivie d’une exsanguination de 40 % du volume sanguin
112
circulant puis, finalement, d’une re-transfusion du sang prélevé. Les animaux seront ventilés
avec : (1) un volume courant de 6 ml/kg et avec une fréquence respiratoire de 15/min ou
25/min, ou (2) 10 ml/kg de volume courant, avec une fréquence respiratoire de 15/min. Les
paramètres hémodynamiques seront mesurées au moyen d’un cathéter artériel pulmonaire
(Swan Ganz), et un cathéter artériel (Biopac®, Aknowledge®). La pression pleurale sera
mesurée avec un ballon oesophagien.
Les implications cliniques de ce travail sont évidentes. L’intérêt des indices dynamiques pour
la prédiction de la réponse au remplissage chez les patients placés en ventilation mécanique a
été largement confirmé par plusieurs études. Néanmoins, à ce jour, il semble qu’en plus du
status volumique, le volume courant semble être un déterminant majeur de l’oscillation et de
l’amplitude de ces indices. L’auteur a comme objectif d’investiguer l’impact du
rapport fréquence cardiaque/fréquence respiratoire sur la valeur prédictive des ces indices
comme variables de la réponse au remplissage volémique des patients de réanimation.
113
CONCLUSION
La surveillance d'un certain nombre de paramètres hémodynamiques lors d’un état de choc est
indispensable chez tout patient de soins intensifs présentant ou non des défaillances vitales.
Ce monitorage hémodynamique consiste en des dispositifs dont la fonction est d’assurer une
collecte et une analyse continues et régulières d’un ensemble de paramètres cardiovasculaires.
Ces données permettent de surveiller l’état du patient et d’alerter les soignants (moyennant
des alarmes) en cas de détérioration de l’état clinique. De plus, ces paramètres sont des outils
diagnostiques qui évaluent les réponses thérapeutiques à différentes médications. Ils peuvent
aussi décrire les tendances évolutives de ces paramètres durant les heures qui précèdent l’état
actuel du malade. Le but de cette thèse est d’étudier les intérêts et les limites des diverses
techniques non invasives du monitorage hémodynamique et cardio-vasculaire en réanimation.
Depuis plusieurs années, de nombreuses équipes cherchent à promouvoir l’utilisation de
systèmes de monitorage hémodynamique non invasif, pour éviter ainsi les complications
inhérentes aux techniques sanglantes. L’utilisation de ces technologies en pratique clinique a
été rendue possible grâce au recours à des concepts physiques anciens couplés à la mise au
point de systèmes utilisant des programmes informatiques novateurs. De plus, le
développement de systèmes de réseaux de monitorage à l’intérieur de l’hôpital et entre les
hôpitaux offre un accès facile et rapide à l’information clinique, là où elle est nécessaire (au
chevet du patient, dans les autres services de soins ou à l’extérieur de l’hôpital…).
A ce jour, de nombreuses techniques non invasives sont utilisées aux soins intensifs. Ces
techniques regroupent l’échocardiographie-Doppler, l’impédancemétrie, le Doppler
œsophagien, la réinhalation partielle du CO2 avec l’utilisation de l'équation de Fick, la
thermodilution transpulmonaire, le signal de pléthysmographie, la mesure continue de la
saturation veineuse mêlée en oxygène, la courbe de pression artérielle sanglante avec ses des
dérivées et intégrales, etc. Néanmoins, ces techniques de monitorage se doivent dans l’avenir
d’être aussi précises et plus reproductibles que les anciennes méthodes invasives. En effet,
utilisées pendant une vingtaine d’années, les techniques invasives n’ont jamais fait la preuve
de leur utilité par une diminution de la morbi-mortalité des patients de soins intensifs. Une
amélioration de la prise en charge du patient ne peut donc être attendue avec ces nouvelles
114
techniques qu’au prix d’une exigence en précision et reproductibilité aussi bonne que
celles des techniques invasives, tout en assurant un maximum d’innocuité. En effet, en
dehors de l’échocardiographie-Doppler, le principal obstacle à l’emploi de ces méthodes à ce
jour réside dans une imprécision des différentes mesures et de leur variabilité. C’est pourquoi,
les investigations cliniques futures devront préciser quelle est la place de ces nouvelles
méthodes non invasives chez les patients de soins intensifs.
Reste que si à ce jour, d’énormes efforts ont été fait pour le développement de moyens non
invasifs dans le cadre du monitorage hémodynamique macro-circulatoire, la question reste
posée de savoir si les valeurs mesurées sont toujours interprétables dans la situation clinique
du patient des soins intensifs. Multiplier les informations obtenues sur le monitorage
hémodynamique n’a de sens dans l’optique d’une prise en charge bénéfique pour un patient
que s’il existe un traitement adéquat à l’affection sous jacente. De ce fait, une information
précise doit aussi être collectée au moment approprié, interprétée correctement et assortie au
suivi des décisions thérapeutiques. En effet, il n’est pas rare qu’un monitorage aussi
sophistiqué qu’il soit engendre des résultats erronés et que l’instinct du clinicien et son
scepticisme vis à vis des paramètres obtenus remette en question des informations aberrantes
issues de la technologie de surveillance. C’est pourquoi, finalement, les cliniciens doivent
rester vigilants et continuer à surveiller leurs moniteurs.
115
REFERENCES :
1. Gupta S, Westfall TC, Lechner AJ, Knuepfer MM (2005) Teaching principles of
cardiovascular function in a medical student laboratory. Adv Physiol Educ 29:118-127 2. Pinsky MR (1990) The meaning of cardiac output. Intensive Care Med 16:415-417 3. Gunn SR, Pinsky MR (2001) Implications of arterial pressure variation in patients in
the intensive care unit. Curr Opin Crit Care 7:212-217 4. Teboul JL (2005) SRLF experts recommendations: Indicators of volume resuscitation
during circulatory failure. Ann Fr Anesth Reanim 24:568-576; quiz 577-581 5. Boussat S, Jacques T, Levy B, Laurent E, Gache A, Capellier G, Neidhardt A (2002)
Intravascular volume monitoring and extravascular lung water in septic patients with pulmonary edema. Intensive Care Med 28:712-718
6. Bunchman TE (2004) Fluid overload in multiple organ dysfunction syndrome: a
prediction of survival. Crit Care Med 32:1805-1806 7. Foland JA, Fortenberry JD, Warshaw BL, Pettignano R, Merritt RK, Heard ML,
Rogers K, Reid C, Tanner AJ, Easley KA (2004) Fluid overload before continuous hemofiltration and survival in critically ill children: a retrospective analysis. Crit Care Med 32:1771-1776
8. Bendjelid K, Romand JA (2003) Fluid responsiveness in mechanically ventilated
patients: a review of indices used in intensive care. Intensive Care Med 29:352-360 9. Stetz CW, Miller RG, Kelly GE, Raffin TA (1982) Reliability of the thermodilution
method in the determination of cardiac output in clinical practice. Am Rev Respir Dis 126:1001-1004
10. Michard F, Teboul JL (2002) Predicting fluid responsiveness in ICU patients(*) : a
critical analysis of the evidence. Chest 121:2000-2008 11. Magder S, De Varennes B (1998) Clinical death and the measurement of stressed
vascular volume. Crit Care Med 26:1061-1064 12. Metnitz PG, Moreno RP, Almeida E, Jordan B, Bauer P, Campos RA, Iapichino G,
Edbrooke D, Capuzzo M, Le Gall JR (2005) SAPS 3--From evaluation of the patient to evaluation of the intensive care unit. Part 1: Objectives, methods and cohort description. Intensive Care Med 31:1336-1344
13. Jardin F (2003) Ventricular interdependence: how does it impact on hemodynamic
evaluation in clinical practice? Intensive Care Med 29:361-363
116
14. Robotham JL, Lixfeld W, Holland L, MacGregor D, Bromberger-Barnea B, Permutt S, Rabson JL (1980) The effects of positive end-expiratory pressure on right and left ventricular performance. Am Rev Respir Dis 121:677-683
15. Robotham JL, Scharf SM (1983) Effects of positive and negative pressure ventilation
on cardiac performance. Clin Chest Med 4:161-187 16. Takata M, Robotham JL (1991) Ventricular external constraint by the lung and
pericardium during positive end-expiratory pressure. Am Rev Respir Dis 143:872-875 17. Pinsky MR (1985) The influence of positive-pressure ventilation on cardiovascular
function in the critically ill. Crit Care Clin 1:699-717 18. Pinsky MR (1990) The effects of mechanical ventilation on the cardiovascular system.
Crit Care Clin 6:663-678 19. Pinsky MR (1994) Heart-lung interactions during positive-pressure ventilation. New
Horiz 2:443-456 20. Jardin F, Delorme G, Hardy A, Auvert B, Beauchet A, Bourdarias JP (1990)
Reevaluation of hemodynamic consequences of positive pressure ventilation: emphasis on cyclic right ventricular afterloading by mechanical lung inflation. Anesthesiology 72:966-970
21. Dries DJ, Kumar P, Mathru M, Mayer R, Zecca A, Rao TL, Freeark RJ (1991)
Hemodynamic effects of pressure support ventilation in cardiac surgery patients. Am Surg 57:122-125
22. Scharf SM, Caldini P, Ingram RH, Jr. (1977) Cardiovascular effects of increasing
airway pressure in the dog. Am J Physiol 232:H35-43 23. Vieillard-Baron A, Chergui K, Rabiller A, Peyrouset O, Page B, Beauchet A, Jardin F
(2004) Superior vena caval collapsibility as a gauge of volume status in ventilated septic patients. Intensive Care Med 30:1734-1739
24. Starling E (1918) The linacre lecture on the law of the heart given at cambridge, 1915.
London: Longmans, Green and CO 25. Guyton AC, Lindsey AW, Abernathy B, Richardson T (1957) Venous return at
various right atrial pressures and the normal venous return curve. Am J Physiol 189:609-615
26. Weber KT, Janicki JS, Shroff S, Fishman AP (1981) Contractile mechanics and
interaction of the right and left ventricles. Am J Cardiol 47:686-695 27. Arbab-Zadeh A, Dijk E, Prasad A, Fu Q, Torres P, Zhang R, Thomas JD, Palmer D,
Levine BD (2004) Effect of aging and physical activity on left ventricular compliance. Circulation 110:1799-1805
117
28. Vieillard-Baron A, Chergui K, Augarde R, Prin S, Page B, Beauchet A, Jardin F (2003) Cyclic changes in arterial pulse during respiratory support revisited by Doppler echocardiography. Am J Respir Crit Care Med 168:671-676
29. Denault AY, Gasior TA, Gorcsan J, 3rd, Mandarino WA, Deneault LG, Pinsky MR
(1999) Determinants of aortic pressure variation during positive-pressure ventilation in man. Chest 116:176-186
30. Guyton AC (1955) Determination of cardiac output by equating venous return curves
with cardiac response curves. Physiol Rev 35:123-129 31. Jellinek H, Krenn H, Oczenski W, Veit F, Schwarz S, Fitzgerald RD (2000) Influence
of positive airway pressure on the pressure gradient for venous return in humans. J Appl Physiol 88:926-932
32. Biondi JW, Schulman DS, Matthay RA (1988) Effects of mechanical ventilation on
right and left ventricular function. Clin Chest Med 9:55-71 33. Holt JP (1959) Flow of liquids through collapsible tubes. Circ Res 7:342-353 34. Meltzer RS, McGhie J, Roelandt J (1982) Inferior vena cava echocardiography. JCU J
Clin Ultrasound 10:47-51 35. Magder S (1990) Starling resistor versus compliance. Which explains the zero-flow
pressure of a dynamic arterial pressure-flow relation? Circ Res 67:209-220 36. Magder SA, Lichtenstein S, Adelman AG (1983) Effect of negative pleural pressure
on left ventricular hemodynamics. Am J Cardiol 52:588-593 37. Pinsky MR (1984) Instantaneous venous return curves in an intact canine preparation.
J Appl Physiol 56:765-771 38. Whittenberger JL, Mc GM, Berglund E, Borst HG (1960) Influence of state of
inflation of the lung on pulmonary vascular resistance. J Appl Physiol 15:878-882 39. Permutt S, Howell JB, Proctor DF, Riley RL (1961) Effect of lung inflation on static
pressure-volume characteristics of pulmonary vessels. J Appl Physiol 16:64-70 40. Permutt S (1973) Relation between pulmonary arterial pressure and pleural pressure
during the acute asthmatic attack. Chest 63:Suppl:25S-28S 41. Motley HL, Tomashefski JF (1950) Effect of high and low oxygen levels and
intermittent positive pressure breathing on oxygen transport in the lungs in pulmonary fibrosis and emphysema. J Appl Physiol 3:189-196
42. Brower R, Wise RA, Hassapoyannes C, Bromberger-Barnea B, Permutt S (1985)
Effect of lung inflation on lung blood volume and pulmonary venous flow. J Appl Physiol 58:954-963
118
43. Morgan BC, Dillard DH, Guntheroth WG (1966) Effect of cardiac and respiratory cycle on pulmonary vein flow, pressure, and diameter. J Appl Physiol 21:1276-1280
44. Morgan BC, Abel FL, Mullins GL, Guntheroth WG (1966) Flow patterns in cavae,
pulmonary artery, pulmonary vein, and aorta in intact dogs. Am J Physiol 210:903-909
45. Buda AJ, Pinsky MR, Ingels NB, Jr., Daughters GTd, Stinson EB, Alderman EL
(1979) Effect of intrathoracic pressure on left ventricular performance. N Engl J Med 301:453-459
46. Pinsky MR (1997) Clinical applications of cardiopulmonary interactions. J Physiol
Pharmacol 48:587-603 47. Pinsky MR, Matuschak GM, Klain M (1985) Determinants of cardiac augmentation
by elevations in intrathoracic pressure. J Appl Physiol 58:1189-1198 48. Pinsky MR, Summer WR, Wise RA, Permutt S, Bromberger-Barnea B (1983)
Augmentation of cardiac function by elevation of intrathoracic pressure. J Appl Physiol 54:950-955
49. Perel A, Pizov R, Cotev S (1987) Systolic blood pressure variation is a sensitive
indicator of hypovolemia in ventilated dogs subjected to graded hemorrhage. Anesthesiology 67:498-502
50. Roussos C (1985) Function and fatigue of respiratory muscles. Chest 88:124S-132S 51. Roussos C, Macklem PT (1982) The respiratory muscles. N Engl J Med 307:786-797 52. Pepe PE, Marini JJ (1982) Occult positive end-expiratory pressure in mechanically
ventilated patients with airflow obstruction: the auto-PEEP effect. Am Rev Respir Dis 126:166-170
53. Tuxen DV, Lane S (1987) The effects of ventilatory pattern on hyperinflation, airway
pressures, and circulation in mechanical ventilation of patients with severe air-flow obstruction. Am Rev Respir Dis 136:872-879
54. Tuxen DV, Williams TJ, Scheinkestel CD, Czarny D, Bowes G (1992) Use of a
measurement of pulmonary hyperinflation to control the level of mechanical ventilation in patients with acute severe asthma. Am Rev Respir Dis 146:1136-1142
55. Vieillard-Baron A, Loubieres Y, Schmitt JM, Page B, Dubourg O, Jardin F (1999)
Cyclic changes in right ventricular output impedance during mechanical ventilation. J Appl Physiol 87:1644-1650
56. Jardin F, Farcot JC, Gueret P, Prost JF, Ozier Y, Bourdarias JP (1983) Cyclic changes
in arterial pulse during respiratory support. Circulation 68:266-274
119
57. Pinsky MR, Perlini S, Solda PL, Pantaleo P, Calciati A, Bernardi L (1996) Dynamic right and left ventricular interactions in the rabbit: simultaneous measurement of ventricular pressure-volume loops. J Crit Care 11:65-76
58. Barbier C, Loubieres Y, Schmit C, Hayon J, Ricome JL, Jardin F, Vieillard-Baron A
(2004) Respiratory changes in inferior vena cava diameter are helpful in predicting fluid responsiveness in ventilated septic patients. Intensive Care Med 30:1740-1746
59. Bendjelid K, Suter PM, Romand JA (2004) The respiratory change in preejection
period: a new method to predict fluid responsiveness. J Appl Physiol 96:337-342 60. Feissel M, Michard F, Faller JP, Teboul JL (2004) The respiratory variation in inferior
vena cava diameter as a guide to fluid therapy. Intensive Care Med 30:1834-1837 61. Feissel M, Michard F, Mangin I, Ruyer O, Faller JP, Teboul JL (2001) Respiratory
changes in aortic blood velocity as an indicator of fluid responsiveness in ventilated patients with septic shock. Chest 119:867-873
62. Michard F, Boussat S, Chemla D, Anguel N, Mercat A, Lecarpentier Y, Richard C,
Pinsky MR, Teboul JL (2000) Relation between respiratory changes in arterial pulse pressure and fluid responsiveness in septic patients with acute circulatory failure. Am J Respir Crit Care Med 162:134-138
63. Reuter DA, Felbinger TW, Schmidt C, Kilger E, Goedje O, Lamm P, Goetz AE
(2002) Stroke volume variations for assessment of cardiac responsiveness to volume loading in mechanically ventilated patients after cardiac surgery. Intensive Care Med 28:392-398
64. Tavernier B, Makhotine O, Lebuffe G, Dupont J, Scherpereel P (1998) Systolic
pressure variation as a guide to fluid therapy in patients with sepsis-induced hypotension. Anesthesiology 89:1313-1321
65. Vieillard-Baron A, Prin S, Chergui K, Dubourg O, Jardin F (2003) Hemodynamic
instability in sepsis: bedside assessment by Doppler echocardiography. Am J Respir Crit Care Med 168:1270-1276
66. Magder S (2004) Clinical usefulness of respiratory variations in arterial pressure. Am
J Respir Crit Care Med 169:151-155 67. Robotham JL, Cherry D, Mitzner W, Rabson JL, Lixfeld W, Bromberger-Barnea B
(1983) A re-evaluation of the hemodynamic consequences of intermittent positive pressure ventilation. Crit Care Med 11:783-793
68. Blow O, Magliore L, Claridge JA, Butler K, Young JS (1999) The golden hour and
the silver day: detection and correction of occult hypoperfusion within 24 hours improves outcome from major trauma. J Trauma 47:964-969
120
69. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M (2001) Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 345:1368-1377
70. Bendjelid K (2003) Accurate measurement of blood pressure. Jama 289:2793-2794 71. Swan HJ, Ganz W, Forrester J, Marcus H, Diamond G, Chonette D (1970)
Catheterization of the heart in man with use of a flow-directed balloon-tipped catheter. N Engl J Med 283:447-451
72. Connors AF, Jr., Speroff T, Dawson NV, Thomas C, Harrell FE, Jr., Wagner D,
Desbiens N, Goldman L, Wu AW, Califf RM, Fulkerson WJ, Jr., Vidaillet H, Broste S, Bellamy P, Lynn J, Knaus WA (1996) The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators [see comments]. Jama 276:889-897
73. Binanay C, Califf RM, Hasselblad V, O'Connor CM, Shah MR, Sopko G, Stevenson
LW, Francis GS, Leier CV, Miller LW (2005) Evaluation study of congestive heart failure and pulmonary artery catheterization effectiveness: the ESCAPE trial. Jama 294:1625-1633
74. Richard C, Warszawski J, Anguel N, Deye N, Combes A, Barnoud D, Boulain T,
Lefort Y, Fartoukh M, Baud F, Boyer A, Brochard L, Teboul JL (2003) Early use of the pulmonary artery catheter and outcomes in patients with shock and acute respiratory distress syndrome: a randomized controlled trial. Jama 290:2713-2720
75. Sandham JD, Hull RD, Brant RF, Knox L, Pineo GF, Doig CJ, Laporta DP, Viner S,
Passerini L, Devitt H, Kirby A, Jacka M (2003) A randomized, controlled trial of the use of pulmonary-artery catheters in high-risk surgical patients. N Engl J Med 348:5-14
76. Abraham E, Shapiro M, Podolsky S (1983) Central venous catheterization in the
emergency setting. Crit Care Med 11:515-517 77. Arrighi DA, Farnell MB, Mucha P, Jr., Iistrup DM, Anderson DL (1989) Prospective,
randomized trial of rapid venous access for patients in hypovolemic shock. Ann Emerg Med 18:927-930
78. Conces DJ, Jr., Holden RW (1984) Aberrant locations and complications in initial
placement of subclavian vein catheters. Arch Surg 119:293-295 79. Lefrant JY, Muller L, De La Coussaye JE, Prudhomme M, Ripart J, Gouzes C, Peray
P, Saissi G, Eledjam JJ (2002) Risk factors of failure and immediate complication of subclavian vein catheterization in critically ill patients. Intensive Care Med 28:1036-1041
80. Mansfield PF, Hohn DC, Fornage BD, Gregurich MA, Ota DM (1994) Complications
and failures of subclavian-vein catheterization. N Engl J Med 331:1735-1738
121
81. Polderman KH, Girbes AJ (2002) Central venous catheter use. Part 1: mechanical complications. Intensive Care Med 28:1-17
82. Wiklund CU, Romand JA, Suter PM, Bendjelid K (2005) Misplacement of central
vein catheters in patients with hemothorax: a new approach to resolve the problem. J Trauma 59:1029-1031
83. Baujard C, Ponchon P, Elghozi JL (1996) Effects of graded hemorrhage on short-term
variability of blood pressure in conscious rats. Fundam Clin Pharmacol 10:511-517 84. Jones DW, Appel LJ, Sheps SG, Roccella EJ, Lenfant C (2003) Measuring blood
pressure accurately: new and persistent challenges. Jama 289:1027-1030 85. Wo CC, Shoemaker WC, Appel PL, Bishop MH, Kram HB, Hardin E (1993)
Unreliability of blood pressure and heart rate to evaluate cardiac output in emergency resuscitation and critical illness. Crit Care Med 21:218-223
86. LeDoux D, Astiz ME, Carpati CM, Rackow EC (2000) Effects of perfusion pressure
on tissue perfusion in septic shock. Crit Care Med 28:2729-2732 87. Albanese J, Leone M, Delmas A, Martin C (2005) Terlipressin or norepinephrine in
hyperdynamic septic shock: a prospective, randomized study. Crit Care Med 33:1897-1902
88. Bourgoin A, Leone M, Delmas A, Garnier F, Albanese J, Martin C (2005) Increasing
mean arterial pressure in patients with septic shock: effects on oxygen variables and renal function. Crit Care Med 33:780-786
89. Lima AP, Beelen P, Bakker J (2002) Use of a peripheral perfusion index derived from
the pulse oximetry signal as a noninvasive indicator of perfusion. Crit Care Med 30:1210-1213
90. Cannesson M, Besnard C, Durand PG, Bohe J, Jacques D (2005) Relation between
respiratory variations in pulse oximetry plethysmographic waveform amplitude and arterial pulse pressure in ventilated patients. Crit Care 9:R562-568
91. Monnet X, Lamia B, Teboul JL (2005) Pulse oximeter as a sensor of fluid
responsiveness: do we have our finger on the best solution? Crit Care 9:429-430 92. Berkenbosch JW, Lam J, Burd RS, Tobias JD (2001) Noninvasive monitoring of
carbon dioxide during mechanical ventilation in older children: end-tidal versus transcutaneous techniques. Anesth Analg 92:1427-1431
93. Rohling R, Biro P (1999) Clinical investigation of a new combined pulse oximetry
and carbon dioxide tension sensor in adult anaesthesia. J Clin Monit Comput 15:23-27 94. Sanders MH, Kern NB, Costantino JP, Stiller RA, Strollo PJ, Jr., Studnicki KA,
Coates JA, Richards TJ (1994) Accuracy of end-tidal and transcutaneous PCO2 monitoring during sleep. Chest 106:472-483
122
95. Schallom L, Ahrens T (2001) Hemodynamic applications of capnography. J Cardiovasc Nurs 15:56-70
96. Levine RL, Wayne MA, Miller CC (1997) End-tidal carbon dioxide and outcome of
out-of-hospital cardiac arrest. N Engl J Med 337:301-306 97. White RD, Asplin BR (1994) Out-of-hospital quantitative monitoring of end-tidal
carbon dioxide pressure during CPR. Ann Emerg Med 23:25-30 98. Amathieu R, Minville V, Poloujadoff MP, Adnet F, Smail N (2005) Interest of
prehospital use of cardiac output monitoring with partial CO2 rebreathing technique: a case report. Ann Fr Anesth Reanim 24:1387-1389
99. Eberhard P, Gisiger PA, Gardaz JP, Spahn DR (2002) Combining transcutaneous
blood gas measurement and pulse oximetry. Anesth Analg 94:S76-80 100. Nadkarni UB, Shah AM, Deshmukh CT (2000) Non-invasive respiratory monitoring
in paediatric intensive care unit. J Postgrad Med 46:149-152 101. Bendjelid K, Schutz N, Stotz M, Gerard I, Suter PM, Romand JA (2005)
Transcutaneous PCO2 monitoring in critically ill adults: clinical evaluation of a new sensor. Crit Care Med 33:2203-2206
102. Janssens JP, Howarth-Frey C, Chevrolet JC, Abajo B, Rochat T (1998)
Transcutaneous PCO2 to monitor noninvasive mechanical ventilation in adults: assessment of a new transcutaneous PCO2 device. Chest 113:768-773
103. Vivien B, Marmion F, Roche S, Devilliers C, Langeron O, Coriat P, Riou B (2006) An
Evaluation of Transcutaneous Carbon Dioxide Partial Pressure Monitoring during Apnea Testing in Brain-dead Patients. Anesthesiology 104:701-707
104. Feissel M, Badie J, Merlani PG, Faller JP, Bendjelid K (2005) Pre-ejection period
variations predict the fluid responsiveness of septic ventilated patients. Crit Care Med 33:2534-2539
105. Monnet X, Rienzo M, Osman D, Anguel N, Richard C, Pinsky MR, Teboul JL (2006)
Passive leg raising predicts fluid responsiveness in the critically ill*. Crit Care Med Publish Ahead of Print
106. Boulain T, Achard JM, Teboul JL, Richard C, Perrotin D, Ginies G (2002) Changes in
BP induced by passive leg raising predict response to fluid loading in critically ill patients. Chest 121:1245-1252
107. Monnet X, Rienzo M, Osman D, Anguel N, Richard C, Pinsky MR, Teboul JL (2005)
Esophageal Doppler monitoring predicts fluid responsiveness in critically ill ventilated patients. Intensive Care Med 31:1195-1201
123
108. Botte A, Leclerc F, Riou Y, Sadik A, Neve V, Rakza T, Richard A (2006) Evaluation of a noninvasive cardiac output monitor in mechanically ventilated children*. Pediatr Crit Care Med
109. Bajorat J, Hofmockel R, Vagts DA, Janda M, Pohl B, Beck C, Noeldge-Schomburg G
(2006) Comparison of invasive and less-invasive techniques of cardiac output measurement under different haemodynamic conditions in a pig model. Eur J Anaesthesiol 23:23-30
110. Botero M, Kirby D, Lobato EB, Staples ED, Gravenstein N (2004) Measurement of
cardiac output before and after cardiopulmonary bypass: Comparison among aortic transit-time ultrasound, thermodilution, and noninvasive partial CO2 rebreathing. J Cardiothorac Vasc Anesth 18:563-572
111. Kotake Y, Moriyama K, Innami Y, Shimizu H, Ueda T, Morisaki H, Takeda J (2003)
Performance of noninvasive partial CO2 rebreathing cardiac output and continuous thermodilution cardiac output in patients undergoing aortic reconstruction surgery. Anesthesiology 99:283-288
112. Haryadi DG, Orr JA, Kuck K, McJames S, Westenskow DR (2000) Partial CO2
rebreathing indirect Fick technique for non-invasive measurement of cardiac output. J Clin Monit Comput 16:361-374
113. Jaffe MB (1999) Partial CO2 rebreathing cardiac output--operating principles of the
NICO system. J Clin Monit Comput 15:387-401 114. Maxwell RA, Gibson JB, Slade JB, Fabian TC, Proctor KG (2001) Noninvasive
cardiac output by partial CO2 rebreathing after severe chest trauma. J Trauma 51:849-853
115. Kubicek WG, From AH, Patterson RP, Witsoe DA, Castaneda A, Lillehei RC, Ersek
R (1970) Impedance cardiography as a noninvasive means to monitor cardiac function. J Assoc Adv Med Instrum 4:79-84
116. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA, Mattson RH (1966)
Development and evaluation of an impedance cardiac output system. Aerosp Med 37:1208-1212
117. Schuster CJ, Schuster HP (1984) Application of impedance cardiography in critical
care medicine. Resuscitation 11:255-274 118. Spiering W, van Es PN, de Leeuw PW (1998) Comparison of impedance cardiography
and dye dilution method for measuring cardiac output. Heart 79:437-441 119. Gabriel S, Atterhog JH, Oro L, Ekelund LG (1976) Measurement of cardiac output by
impedance cardiography in patients with myocardial infarction. Comparative evaluation of impedance and dye dilution methods. Scand J Clin Lab Invest 36:29-34
124
120. Tanguy RL, Cadoret Y, Guedes Y, Egreteau JP (1982) Transthoracic electrical impedance plethysmography and thermodilution: comparative evaluation of two methods for cardiac output measurement in man. Ann Fr Anesth Reanim 1:67-75
121. Berton C, Cholley B (2002) Equipment review: new techniques for cardiac output
measurement--oesophageal Doppler, Fick principle using carbon dioxide, and pulse contour analysis. Crit Care 6:216-221
122. Huntsman LL, Stewart DK, Barnes SR, Franklin SB, Colocousis JS, Hessel EA (1983)
Noninvasive Doppler determination of cardiac output in man. Clinical validation. Circulation 67:593-602
123. Colocousis JS, Huntsman LL, Curreri PW (1977) Estimation of stroke volume
changes by ultrasonic doppler. Circulation 56:914-917 124. Cariou A, Monchi M, Joly LM, Bellenfant F, Claessens YE, Thebert D, Brunet F,
Dhainaut JF (1998) Noninvasive cardiac output monitoring by aortic blood flow determination: evaluation of the Sometec Dynemo-3000 system. Crit Care Med 26:2066-2072
125. Singer M, Clarke J, Bennett ED (1989) Continuous hemodynamic monitoring by
esophageal Doppler. Crit Care Med 17:447-452 126. Valtier B, Cholley BP, Belot JP, de la Coussaye JE, Mateo J, Payen DM (1998)
Noninvasive monitoring of cardiac output in critically ill patients using transesophageal Doppler. Am J Respir Crit Care Med 158:77-83
127. Cholley BP, Vieillard-Baron A, Mebazaa A (2006) Echocardiography in the ICU:
time for widespread use! Intensive Care Med 32:9-10 128. Lichtenstein D (1997) Pulmonary echography: a method of the future in emergency
medicine and resuscitation. Rev Pneumol Clin 53:63-68 129. Lichtenstein DA, Meziere G, Lascols N, Biderman P, Courret JP, Gepner A, Goldstein
I, Tenoudji-Cohen M (2005) Ultrasound diagnosis of occult pneumothorax. Crit Care Med 33:1231-1238
130. Carbutti G, Romand JA, Carballo JS, Bendjelid SM, Suter PM, Bendjelid K (2003)
Transcranial Doppler: an early predictor of ischemic stroke after cardiac arrest? Anesth Analg 97:1262-1265
131. Denault A, Ferraro P, Couture P, Boudreault D, Babin D, Poirier C, Buithieu J (2003)
Transesophageal echocardiography monitoring in the intensive care department: the management of hemodynamic instability secondary to thoracic tamponade after single lung transplantation. J Am Soc Echocardiogr 16:688-692
125
132. Mansencal N, Joseph T, Vieillard-Baron A, Langlois S, El Hajjam M, Qanadli SD, Lacombe P, Jardin F, Dubourg O (2005) Diagnosis of right ventricular dysfunction in acute pulmonary embolism using helical computed tomography. Am J Cardiol 95:1260-1263
133. Mansencal N, Redheuil A, Joseph T, Vieillard-Baron A, Jondeau G, Lacombe P,
Jardin F, Dubourg O (2004) Use of transthoracic echocardiography combined with venous ultrasonography in patients with pulmonary embolism. Int J Cardiol 96:59-63
134. Cheitlin MD, Armstrong WF, Aurigemma GP, Beller GA, Bierman FZ, Davis JL,
Douglas PS, Faxon DP, Gillam LD, Kimball TR, Kussmaul WG, Pearlman AS, Philbrick JT, Rakowski H, Thys DM, Antman EM, Smith SC, Jr., Alpert JS, Gregoratos G, Anderson JL, Hiratzka LF, Faxon DP, Hunt SA, Fuster V, Jacobs AK, Gibbons RJ, Russell RO (2003) ACC/AHA/ASE 2003 Guideline Update for the Clinical Application of Echocardiography: summary article. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography). J Am Soc Echocardiogr 16:1091-1110
135. Cheitlin MD, Armstrong WF, Aurigemma GP, Beller GA, Bierman FZ, Davis JL,
Douglas PS, Faxon DP, Gillam LD, Kimball TR, Kussmaul WG, Pearlman AS, Philbrick JT, Rakowski H, Thys DM (2003) ACC/AHA/ASE 2003 guideline update for the clinical application of echocardiography--summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography). J Am Coll Cardiol 42:954-970
136. Cheitlin MD, Armstrong WF, Aurigemma GP, Beller GA, Bierman FZ, Davis JL,
Douglas PS, Faxon DP, Gillam LD, Kimball TR, Kussmaul WG, Pearlman AS, Philbrick JT, Rakowski H, Thys DM, Antman EM, Smith SC, Jr., Alpert JS, Gregoratos G, Anderson JL, Hiratzka LF, Hunt SA, Fuster V, Jacobs AK, Gibbons RJ, Russell RO (2003) ACC/AHA/ASE 2003 guideline update for the clinical application of echocardiography: summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography). Circulation 108:1146-1162
137. Thys DM (2003) Clinical competence in echocardiography. Anesth Analg 97:313-322 138. Bendjelid K, Glas KE, Shanewise JS (2005) ECG monitoring is essential for
echocardiographic analysis. Anesth Analg 100:294--296 139. Jardin F, Vieillard-Baron A (2006) Ultrasonographic examination of the venae cavae.
Intensive Care Med 32:203-206 140. Bendjelid K, Romand JA, Walder B, Suter PM, Fournier G (2002) Correlation
between measured inferior vena cava diameter and right atrial pressure depends on the echocardiographic method used in patients who are mechanically ventilated. J Am Soc Echocardiogr 15:944-949
126
141. Nagueh SF, Kopelen HA, Zoghbi WA (1996) Relation of mean right atrial pressure to
echocardiographic and Doppler parameters of right atrial and right ventricular function. Circulation 93:1160-1169
142. Nageh MF, Kopelen HA, Zoghbi WA, Quinones MA, Nagueh SF (1999) Estimation
of mean right atrial pressure using tissue Doppler imaging. Am J Cardiol 84:1448-1451, A1448
143. Nagueh SF (1999) Noninvasive evaluation of hemodynamics by Doppler
echocardiography. Curr Opin Cardiol 14:217-224 144. Nagueh SF, Lakkis NM, Middleton KJ, Spencer WH, 3rd, Zoghbi WA, Quinones MA
(1999) Doppler estimation of left ventricular filling pressures in patients with hypertrophic cardiomyopathy. Circulation 99:254-261
145. Nagueh SF, Middleton KJ, Kopelen HA, Zoghbi WA, Quinones MA (1997) Doppler
tissue imaging: a noninvasive technique for evaluation of left ventricular relaxation and estimation of filling pressures. J Am Coll Cardiol 30:1527-1533
146. Nagueh SF, Mikati I, Kopelen HA, Middleton KJ, Quinones MA, Zoghbi WA (1998)
Doppler estimation of left ventricular filling pressure in sinus tachycardia. A new application of tissue doppler imaging. Circulation 98:1644-1650
147. Bouchard MJ, Denault A, Couture P, Guertin MC, Babin D, Ouellet P, Carrier M,
Tardif JC (2004) Poor correlation between hemodynamic and echocardiographic indexes of left ventricular performance in the operating room and intensive care unit. Crit Care Med 32:644-648
148. Costachescu T, Denault A, Guimond JG, Couture P, Carignan S, Sheridan P, Hellou
G, Blair L, Normandin L, Babin D, Allard M, Harel F, Buithieu J (2002) The hemodynamically unstable patient in the intensive care unit: hemodynamic vs. transesophageal echocardiographic monitoring. Crit Care Med 30:1214-1223
149. English JB, Hodges MR, Sentker C, Johansen R, Stanley TH (1980) Comparison of
aortic pulse-wave contour analysis and thermodilution methods of measuring cardiac output during anesthesia in the dog. Anesthesiology 52:56-61
150. Huggins RA, Smith EL, Sinclair MA (1949) Comparison of cardiac output by the
direct Fick and pressure pulse contour methods in the open-chest dog. Am J Physiol 159:385-388
151. Williams RR, Wray RB, Tsagaris TJ, Kuida H (1974) Computer estimation of stroke
volume from aortic pulse contour in dogs and humans. Cardiology 59:350-366 152. De Vaal JB, de Wilde RB, van den Berg PC, Schreuder JJ, Jansen JR (2005) Less
invasive determination of cardiac output from the arterial pressure by aortic diameter-calibrated pulse contour. Br J Anaesth 95:326-331
127
153. Della Rocca G, Costa MG, Coccia C, Pompei L, Di Marco P, Vilardi V, Pietropaoli P (2003) Cardiac output monitoring: aortic transpulmonary thermodilution and pulse contour analysis agree with standard thermodilution methods in patients undergoing lung transplantation. Can J Anaesth 50:707-711
154. Donelli A, Jansen JR, Hoeksel B, Pedeferri P, Hanania R, Bovelander J, Maisano F,
Castiglioni A, Ottavio A, Schreuder JJ (2002) Performance of a real-time dicrotic notch detection and prediction algorithm in arrhythmic human aortic pressure signals. J Clin Monit Comput 17:181-185
155. Jansen JR, Schreuder JJ, Mulier JP, Smith NT, Settels JJ, Wesseling KH (2001) A
comparison of cardiac output derived from the arterial pressure wave against thermodilution in cardiac surgery patients. Br J Anaesth 87:212-222
156. Reuter DA, Felbinger TW, Kilger E, Schmidt C, Lamm P, Goetz AE (2002)
Optimizing fluid therapy in mechanically ventilated patients after cardiac surgery by on-line monitoring of left ventricular stroke volume variations. Comparison with aortic systolic pressure variations. Br J Anaesth 88:124-126
157. Pearse RM, Ikram K, Barry J (2004) Equipment review: an appraisal of the LiDCO
plus method of measuring cardiac output. Crit Care 8:190-195 158. Hamilton TT, Huber LM, Jessen ME (2002) PulseCO: a less-invasive method to
monitor cardiac output from arterial pressure after cardiac surgery. Ann Thorac Surg 74:S1408-1412
159. Manecke GR (2005) Edwards FloTrac sensor and Vigileo monitor: easy, accurate,
reliable cardiac output assessment using the arterial pulse wave. Expert Rev Med Devices 2:523-527
160. Schneider AJ, Teule GJ, Groeneveld AB, Nauta J, Heidendal GA, Thijs LG (1988)
Biventricular performance during volume loading in patients with early septic shock, with emphasis on the right ventricle: a combined hemodynamic and radionuclide study. Am Heart J 116:103-112
161. Wagner JG, Leatherman JW (1998) Right ventricular end-diastolic volume as a
predictor of the hemodynamic response to a fluid challenge. Chest 113:1048-1054 162. Diebel LN, Wilson RF, Tagett MG, Kline RA (1992) End-diastolic volume. A better
indicator of preload in the critically ill. Arch Surg 127:817-821 163. Tousignant CP, Walsh F, Mazer CD (2000) The use of transesophageal
echocardiography for preload assessment in critically ill patients. Anesth Analg 90:351-355
164. Magder S, Lagonidis D (1999) Effectiveness of albumin versus normal saline as a test
of volume responsiveness in post-cardiac surgery patients. J Crit Care 14:164-171
128
165. De Backer D (2006) Can passive leg raising be used to guide fluid administration? Crit Care 10:170
166. Heenen S, De Backer D, Vincent JL (2006) How can the response to volume
expansion in patients with spontaneous respiratory movements be predicted? Crit Care 10:R102
167. Brengelmann GL (2003) A critical analysis of the view that right atrial pressure
determines venous return. J Appl Physiol 94:849-859 168. Levy MN (1979) The cardiac and vascular factors that determine systemic blood flow.
Circ Res 44:739-747 169. Bendjelid K (2005) Right atrial pressure: determinant or result of change in venous
return? Chest 128:3639-3640 170. Jue J, Chung W, Schiller NB (1992) Does inferior vena cava size predict right atrial
pressures in patients receiving mechanical ventilation? J Am Soc Echocardiogr 5:613-619
171. Lichtenstein D, Jardin F (1994) Appréciation non invasive de la pression veineuse
centrale par la mesure échographique du calibre de la veine cave inférieure en réanimation. Réan. Urg. 3:79-82
172. Feissel M, Teboul JL, Merlani P, Badie J, Faller JP, Bendjelid K (2007)
Plethysmographic dynamic indices predict fluid responsiveness in septic ventilated patients. In press, Intensive Care Med
173. Coudray A, Romand JA, Treggiari M, Bendjelid K (2005) Fluid responsiveness in
spontaneously breathing patients: a review of indexes used in intensive care. Crit Care Med 33:2757-2762
174. Bendjelid K, Poblete B, Baenziger O, Romand J-A (2003) Doppler cerebral blood
flow variation during the first 24 postoperative hours following hypothermic cardio pulmonary bypass. Interactive Cardiovascular and Thoracic Surgery 2003; 2: 46-52
175. Rivers EP, Ander DS, Powell D (2001) Central venous oxygen saturation monitoring
in the critically ill patient. Curr Opin Crit Care 7:204-211 176. Ladakis C, Myrianthefs P, Karabinis A, Karatzas G, Dosios T, Fildissis G, Gogas J,
Baltopoulos G (2001) Central venous and mixed venous oxygen saturation in critically ill patients. Respiration 68:279-285
177. Dueck MH, Klimek M, Appenrodt S, Weigand C, Boerner U (2005) Trends but not
individual values of central venous oxygen saturation agree with mixed venous oxygen saturation during varying hemodynamic conditions. Anesthesiology 103:249-257
129
178. Krantz T, Warberg J, Secher NH (2005) Venous oxygen saturation during normovolaemic haemodilution in the pig. Acta Anaesthesiol Scand 49:1149-1156
179. Alexander JK (1985) The cardiomyopathy of obesity. Prog Cardiovasc Dis 27:325-334
180. Alexander JK (1993) Obesity and the heart. Heart Dis Stroke 2:317-321 181. Lauer MS, Anderson KM, Kannel WB, Levy D (1991) The impact of obesity on left
ventricular mass and geometry. The Framingham Heart Study. Jama 266:231-236 182. Bendjelid K (2006) Body habitus should not considerably affect the interpretation of
hemodynamic measurements: huge news? Crit Care Med 34:1289-1290 183. Beutler S, Schmidt U, Michard F (2004) Hemodynamic monitoring in obese patients:
a big issue. Crit Care Med 32:1981 184. Bercault N, Boulain T, Kuteifan K, Wolf M, Runge I, Fleury JC (2004) Obesity-
related excess mortality rate in an adult intensive care unit: A risk-adjusted matched cohort study. Crit Care Med 32:998-1003
185. Schultz MJ, Spronk PE (2005) Overweight in ICU patients: is the body mass index
reliable? Chest 127:683-684; author reply 684 186. Stelfox HT, Ahmed SB, Ribeiro RA, Gettings EM, Pomerantsev E, Schmidt U (2006)
Hemodynamic monitoring in obese patients: the impact of body mass index on cardiac output and stroke volume. Crit Care Med 34:1243-1246
187. Verdant C, De Backer D (2005) How monitoring of the microcirculation may help us
at the bedside. Curr Opin Crit Care 11:240-244 188. Sakr Y, Reinhart K, Vincent JL, Sprung CL, Moreno R, Ranieri VM, De Backer D,
Payen D (2006) Does dopamine administration in shock influence outcome? Results of the Sepsis Occurrence in Acutely Ill Patients (SOAP) Study. Crit Care Med 34:589-597
189. De Backer D, Creteur J, Dubois MJ, Sakr Y, Vincent JL (2004) Microvascular
alterations in patients with acute severe heart failure and cardiogenic shock. Am Heart J 147:91-99
190. Sakr Y, Dubois MJ, De Backer D, Creteur J, Vincent JL (2004) Persistent
microcirculatory alterations are associated with organ failure and death in patients with septic shock. Crit Care Med 32:1825-1831
191. Asfar P, De Backer D, Meier-Hellmann A, Radermacher P, Sakka SG (2004) Clinical
review: influence of vasoactive and other therapies on intestinal and hepatic circulations in patients with septic shock. Crit Care 8:170-179
192. Creteur J, De Backer D, Sun Q, Vincent JL (2004) The hepatosplanchnic contribution
to hyperlactatemia in endotoxic shock: effects of tissue ischemia. Shock 21:438-443
130
0
193. Lobo SM, De Backer D, Sun Q, Tu Z, Dimopoulos G, Preiser JC, Nagy N, Vray B,
Vercruy V, Terzi RG, Vincent JL (2003) Gut mucosal damage during endotoxic shock is due to mechanisms other than gut ischemia. J Appl Physiol 95:2047-2054
194. De Backer D, Creteur J, Preiser JC, Dubois MJ, Vincent JL (2002) Microvascular
blood flow is altered in patients with sepsis. Am J Respir Crit Care Med 166:98-104 195. Creteur J, De Backer D, Sakr Y, Koch M, Vincent JL (2006) Sublingual capnometry
tracks microcirculatory changes in septic patients. Intensive Care Med 32:516-523 196. Preiser JC, De Backer D, Vincent JL (2003) Nitroglycerin for septic shock. Lancet
361:880 197. Spronk PE, Ince C, Gardien MJ, Mathura KR, Oudemans-van Straaten HM, Zandstra
DF (2002) Nitroglycerin in septic shock after intravascular volume resuscitation. Lancet 360:1395-1396
198. Boerma EC, Mathura KR, van der Voort PH, Spronk PE, Ince C (2005) Quantifying
bedside-derived imaging of microcirculatory abnormalities in septic patients: a prospective validation study. Crit Care 9:R601-606
199. Spronk PE, Zandstra DF, Ince C (2004) Bench-to-bedside review: sepsis is a disease
of the microcirculation. Crit Care 8:462-468 200. De Backer D, Heenen S, Piagnerelli M, Koch M, Vincent JL (2005) Pulse pressure
variations to predict fluid responsiveness: influence of tidal volume. Intensive Care Med 31:517-523
201. Szold A, Pizov R, Segal E, Perel A (1989) The effect of tidal volume and
intravascular volume state on systolic pressure variation in ventilated dogs. Intensive Care Med 15:368-371
131
