Você está na página 1de 5

Extravascular lung water in sepsis-associated acute respiratory distress syndrome: Indexing with predicted body weight improves correlation

with severity of illness and survival*


Charles R. Phillips, MD; Mark S. Chesnutt, MD; Stephen M. Smith, PhD, FJFICM

Objectives: To determine whether extravascular lung water predicts survival in patients with early acute respiratory distress syndrome, to determine the relationship between extravascular lung water and other markers of lung injury, and to examine if indexing extravascular lung water with predicted body weight (EVLWp) strengthens its discriminative power. Design: Extravascular lung water and other markers of lung injury were measured prospectively in 19 patients with sepsisinduced acute respiratory distress syndrome for 3 days. Setting: The intensive care units of an academic tertiary referral hospital. Measurements and Main Results: Lung injury score, Sequential Organ Failure Assessment score, dead space-tidal volume fraction (VD/VT), and EVLWp were all signicantly higher on day 1 in nonsurvivors compared with survivors (lung injury score, 2.8 0.34 vs. 1.9 0.50; p .004) (Sequential Organ Failure Assessment score, 13 3.4 vs. 7.7 0.8; p .006) (VD/VT, 0.68 0.07 vs. 0.58 0.07; p .009) (EVLWp, 20.6 4.6 vs. 11.6 1.9

mL/kg; p .002). EVLWp correlated with Sequential Organ Failure Assessment score, lung injury score, VD/VT, and PaO2/FIO2. The receiver operator characteristic curve analysis indicated that EVLWp, VD/VT, and extravascular lung water (p .0005, .009, and .013, respectively) but not PaO2/FIO2 (p .311) discriminate between survivors and nonsurvivors. Three-day average EVLWp >16 mL/kg predicted in-hospital mortality with 100% specicity and 86% sensitivity. Conclusions: Increased extravascular lung water is a feature of early acute respiratory distress syndrome and predicts survival. Indexing extravascular lung water to predicted body weight, instead of actual body weight, improves the predictive value of extravascular lung water for survival and correlation with markers of disease severity. (Crit Care Med 2008; 36:6973) KEY WORDS: sepsis; physiologic deadspace fraction; pulmonary edema; acute respiratory distress syndrome; extravascular lung water; transpulmonary thermodilution

cute respiratory distress syndrome (ARDS) is a major cause of respiratory failure and death (1). Many of the current clinical indices used to identify the presence of ARDS and to guide supportive care have been shown to be insensitive markers of disease severity and are poor predictors of outcome (25). There remains

*See also p. 337. From the Division of Pulmonary and Critical Care Medicine and Center for Intensive Care Research, Oregon Health and Science University, Portland, OR (CRP, MSC, SMS); and the Portland Veterans Administration Medical Center, Portland, OR (MSC). Supported, in part, by a grant from the Center of Excellence in Human Research, Oregon Opportunity Funds, Oregon Health and Science University, Portland, OR. Dr. Phillips has been a consultant for Pulsion Medical Systems. Drs. Chesnutt and Smith have not disclosed any potential conicts of interest. For information regarding this article, E-mail: phillipc@ohsu.edu Copyright 2007 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/01.CCM.0000295314.01232.BE

a need to identify sensitive pulmonary specic clinical markers that better reect injury to enhance our understanding of the pathophysiology in ARDS, to aid in prognostication, to determine disease severity, and to guide the development of future therapeutic interventions. The hallmark of ARDS is increased alveolar capillary barrier permeability characterized by accumulation of excess extravascular lung water (EVLW). Studies in animals and humans have shown that manipulation of EVLW with diuresis and -agonists provides a potential avenue for therapeutic intervention in ARDS (6 11). However, there are still a number of questions regarding the utility of bedside determination of EVLW in patients with ARDS. The aims of this study were to determine whether EVLW and other markers of disease severity, in patients with ARDS, differ in survivors and nonsurvivors. Additionally, we proposed that the current method of indexing lung water to actual body weight decreases its sensitivity as a

measure of severity of illness and its predictive value for outcome in an obese population. Because predicted body weight (PBW) reects lung size better than actual body weight, we proposed that an index of EVLW with height- and gender-derived PBW (EVLWp) would improve the utility of lung water measurement. Finally, because the transpulmonary thermodilution technique (TTT) relies on heat exchange across the alveolar endothelial barrier, it has been postulated that EVLW will be underestimated in lung regions with diminished vascular perfusion (high dead space/tidal volume VD/VT ). We tested this proposal by examining the relationship between EVLW and VD/VT in ARDS.

MATERIALS AND METHODS


Patients were studied prospectively at Oregon Health & Science University, a tertiary care university hospital, from January 2004 through February 2006. The Institutional Review Board approved the protocol. Patients who were at least 18 yrs of age, intubated, and

Crit Care Med 2008 Vol. 36, No. 1

69

receiving positive-pressure ventilation were eligible if they met the AmericanEuropean consensus denition of ARDS (12): acuteonset bilateral opacities on the chest radiograph, a PaO2:FIO2 ratio 200 mm Hg, and absence of clinical evidence of left atrial hypertension. Patients who were known to have obstructive lung disease, interstitial lung disease, liver failure, or pulmonary vascular disease were excluded. Enrollment occurred within 24 hrs of meeting the inclusion criteria and measurements were made over the subsequent 3 days. All patients had sepsis and were receiving antibiotics. Subjects were ventilated with low tidal volumes and positive endexpiratory pressure titration in accordance with the ARDSNet-derived protocol (13). Measurement of Extravascular Lung Water. The TTT method for measuring EVLW has been validated in critically ill patients (14 16). The accuracy of this method has been conrmed by comparison with the postmortem gravimetric technique, and with the double dilution (thermo-dye) technique in both animals and humans in a variety of disease states (14 17). Moreover, it has been shown to have a sensitivity that allows detection of clinically relevant changes in EVLW (18). EVLW based on actual body weight was measured using the single-indicator TTT (PiCCO, Pulsion Medical Systems, Munich, Germany). A 15-mL bolus of 0.9% saline at 5C was injected via a central venous catheter. The thermodilution curve was recorded with a femoral artery thermistor and used to calculate the indexed EVLW as previously described (16). The average result from three consecutive 15-mL bolus injections was recorded for each patient. EVLW was calculated by dividing the lung water by actual body weight. EVLWp was calculated by dividing the measured lung water by PBW. PBW was calculated as 50 0.91 (centimeters of height 152.4) for males and 45.5 0.91 (centimeters of height 152.4) for females (13) as discussed at www. halls.md/ideal-weight/devine.htm. Measurement of Pulmonary Dead-Space Fraction. The volumetric capnography and pulmonary mechanics monitor used to measure pulmonary dead space has been validated by comparison with the metabolic monitor method (19 21). VD/VT was determined using the mean expired carbon dioxide fraction, which was determined with a bedside metabolic monitor (NICO, Respironics, Wallington, CT), the PaCO2, and the Bohr-Enghoff equation. Once the PECO2 reading was stable for 5 mins, PaCO2 was determined by arterial blood gas collection then entered into the NICO monitor. Because VD/VT is dependent on tidal volume (21), to allow comparison with earlier work, VD/VT was determined after the patient had been ventilated at 10 mL/kg PBW for 10 mins, with no bias ow. Management of the uid balance

Table 1. Patient characteristics on day 1 All 19) 14 10 13 14 14 0.6 2.2 3.8 7.1 40 Survivors (n 12) 55 10 20 177 74 71 6.7 9.3 26.6 28.3 149 17 7 11 13 11 0.6 2.1 3.8 7.1 35 Nonsurvivors (n 7) 56 3 30 166 85 60 6.9 11 30.9 20.4 135 12 14 14 14 0.4 2.0 1.9 4.0 49 9

Variable, Mean

SD

Except Sex

(n

p Value .89 .16 .03a .08 .12 .10 .93 .13 .005b .007b .642

Age, yrs Male, n APACHE II Height, cm Weight, kg PBW, kg Weight indexed tidal volume, mL/kg PBW PEEP, cm of water Plateau pressure, cm of water Static compliance, mL/cm of water PaO2:FIO2, mm Hg

55 13 23 173 78 67 6.8 9.9 28.2 25.4 144

APACHE, Acute Physiology and Chronic Health Evaluation; PBW, predicted body weight; PEEP, positive end-expiratory pressure. a p .05; bp .01.

and mechanical ventilation was determined by the attending physicians, who were unaware of the study measurements. Lung injury score (LIS) and Sequential Organ Failure Assessment (SOFA) score were calculated using the websites http:// www.medalreg.com/qhc/index.html (22, 23). Statistical Analysis. Data are presented as SD unless otherwise stated. Continmean uous data that were normally distributed (Kolmogorov-Smirnov normality test; GraphPad Prism 4.03, San Diego, CA) were compared using Students t-test with Welchs correction. Non-normal continuous data were compared with the Mann-Whitney U test. Categorical comparisons were made using the chi-square test. The coefcients of variation were compared by repeated measures analysis of variance and groups tested with the Bonferroni method (GraphPad Prism). Correlation (Pearsons product moment correlation) and receiver operator characteristic curves also were constructed using GraphPad Prism. In both of these analyses, data were averaged for the 3 days and the mean values used in graphs and calculations.

RESULTS
From January 2004 through February 2006, 154 patients with sepsis were admitted to the intensive care units and 59 of these patients met criteria for ARDS. Twenty three of these met exclusion criteria, 12 did not have surrogate consent, and ve declined to participate in the study. Nineteen patients were enrolled in this study and the rst measurements were made 14.3 6.4 hrs after fullling criteria for ARDS. Pneumonia was the cause of sepsis in ten patients. There was no difference in mortality in patients with and without pneumonia (40% and 33%, respectively; p .764). In addition, there was no difference in EVLW between patients with and without pneumonia on

day 1 (12.7 3.3 mL/kg vs. 12.1 2.9 mL/kg; p .6325). Table 1 describes patient characteristics on day 1 of the study. The mean age of the patients was 55 14 yrs (range, 31 86 yrs) and 13 of 19 patients (68%) were male. The mean Acute Physiology and Chronic Health Evaluation II score was 23.4 10.1 (range, 10 52), and 12 of the 19 patients survived to hospital discharge. The other seven patients died in the intensive care unit when treatment was withdrawn (37% overall mortality). There were signicant differences between survivors and nonsurvivors in their Acute Physiology and Chronic Health Evaluation II scores (20 7 vs. 30 12; p .028), plateau pressures (26.6 3.7 vs. 30.9 1.9 cm water; p .005), and static compliance (28.3 7.1 vs. 20.4 4.0 mL/cm of water; p .007). Baseline weights were not different (Table 1). Many markers of lung performance were abnormal on day 1: VD/VT was elevated (0.62 0.08), the PaO2/FIO2 ratio was decreased (144 40 mm Hg), and EVLW was increased at 12.4 3.0 mL/kg (normal, 710 mL/kg). LIS, SOFA, VD/VT, and EVLWp were all signicantly higher on day 1 in nonsurvivors compared with survivors (Fig. 1): LIS, 2.8 0.34 vs. 1.9 0.50, p .004; SOFA, 13 3.4 vs. 7.7 0.8, p .006; 0.07 vs. 0.58 0.07, p VD/VT, 0.68 .009; and EVLWp, 20.6 4.6 vs. 11.6 1.9 mL/kg, p .002. EVLW had a higher trend in nonsurvivors (14.3 3.3 vs. 11.3 2.0 mL/kg; p .06) (Fig. 1). PaO2/FIO2 was not different between the two groups (135 49 vs. 149 35 mm Hg; p .64). LIS, SOFA, VD/VT, and EVLWp remained higher in nonsurvivors for all three days; however, the difference
Crit Care Med 2008 Vol. 36, No. 1

70

Figure 1. Lung dysfunction in the rst 3 days following diagnosis of acute respiratory distress syndrome for survivors (unlled circles) and nonsurvivors (lled triangles). Plots show means (black) and individual values (gray) vs. time for the following: A, Sequential Organ Failure Assessment score (SOFA); B, lung injury score (LIS); C, PaO2/FIO2; D, dead space-tidal volume fraction (VD/VT); E, extravascular lung water (EVLW); and F, extravascular lung water indexed to predicted body weight (EVLWp). *p .05. **p .01. #p .001.

coefcient of variation for VD/VT, EVLW, and EVLWp (p .001, Bonferroni method) but VD/VT, EVLW, and EVLWp showed no difference (p .05). Pooled average data from all three days (Fig. 2) demonstrate that EVLW correlated inversely with PaO2/FIO2 (r .773; p .0001) (Fig. 2C, left panel) but did not signicantly correlate with LIS (r .289; p .23) (Fig. 2A, left) or .364; p .13) (Fig. 2D, left). VD/VT (r In addition, SOFA was positively correlated with EVLW (Fig. 2B) (r .516; p .024). In contrast, EVLWp was signicantly correlated with LIS, SOFA, PaO2/ FIO2, and VD/VT (r values of .532, .610, .525, and .621; and p values of .02, .006, .02, and .005, respectively) (Fig. 2, right panels). Receiver operator characteristic curves were generated using the values averaged across all three days to assess the ability of VD/VT, PaO2/FIO2, EVLW, and EVLWp to discriminate between survivors and nonsurvivors (Fig. 3). In descending order, the areas under the curves ( SEM) for EVLWp, VD/VT, EVLW, and PaO2/FIO2 were 0.988 0.019, 0.869 0.112, 0.851 0.113, and 0.643 0.137, respectively. The receiver operator characteristic curve analysis indicated that EVLWp, VD/VT, and EVLW (p .0005, .009, and .013, respectively), but not PaO2/FIO2 (p .311), discriminate between survivors and nonsurvivors (Fig. 3). The area under the curve was highest for the EVLWp receiver operator characteristic curve, although the small number of measurements prevented statistical comparison. A value of the 3-day average EVLWp 16 mL/kg predicted death with 100% specicity and 86% sensitivity.

DISCUSSION
Our ndings support the idea that EVLW can predict outcome and reect severity of ARDS. Additionally, we have found that indexing EVLW to PBW instead of actual body weight may enhance its discriminatory power as a predictor of mortality. Furthermore, our examination of the relationship between VD/VT and EVLWp in patients supports the hypothesis that TTT remains a clinical useful tool even at extremes of high dead space.

decreased to a nonsignicant value for VD/VT by day 3 (p .06) (Fig. 1D). While the average measurements (black symbols) varied only modestly over time, variation of individual measurements of lung dysfunction (gray symbols) varied by greater amounts. For instance, the coefcients of variation for VD/VT, EVLW, EVLWp, and PaO2/FIO2 during the three days were different by repeated measures analysis of variance (0.05 0.04, 0.09 0.03, 0.10 0.04, and 0.18 0.10, respectively; p .0001). PaO2/FIO2 coefcient of variation was greater than the
Crit Care Med 2008 Vol. 36, No. 1

PBW: A Better Index for Lung Water?


EVLW has historically been indexed using lung water and actual body weight.

However, this method is potentially awed because adult lung size does not increase with body weight but is more closely associated with gender and height (24 25). Therefore, indexing EVLW to actual body weight will likely underestimate lung water in obese patients. Such inaccuracy is likely to be increasingly prevalent as the U.S. and European populations become more obese (26). Furthermore, this systematic error will confound the relationship between EVLW and other markers of lung injury, and may decrease the predictive value of lung water for survival. Use of PBW increased the area under the receiver operator characteristic curve from 0.851 0.113 for EVLW to 0.988 0.019 for EVLWp (Fig. 3). Small patient numbers prevents statistical testing and points to the need for a larger investigation of the use of EVLWp. The impressive discriminatory power of the average EVLWp was reected by a cut-off of 16 mL/kg predicting death with 100% specicity and 86% sensitivity. Measures of Lung Dysfunction. Disease severity measures in ARDS patients are potentially useful to guide treatment, identify groups of patients appropriate for new treatments in clinical trials, and to assist prognostication. Previously, Acute Physiology and Chronic Health Evaluation II and SOFA scores have been shown to be good predictors of mortality risk in ARDS of many types (2729). LIS and PaO2/FIO2, despite being more pulmonary-specic markers, have not been shown to consistently discriminate between survivors and nonsurvivors with ARDS (2, 3, 28, 29). In this study, nonsurvivors had signicantly higher scores for Acute Physiology and Chronic Health Evaluation II, SOFA, LIS, EVLWp, and VD/VT (Fig. 1), and showed a trend toward higher scores for EVLW. We conrm the observations of Dr. Kuzkov and colleagues (30) that EVLW indexed to measured body weight is positively correlated with LIS and inversely correlated with PaO2/FIO2. The pulmonary specic markers of lung injury all varied during the 3 days of the study (Fig. 1, CF). Of these, PaO2/FIO2 varied signicantly more than VD/VT, EVLW, and EVLWp (p .001). The large temporal variation of PaO2/FIO2 coupled with its relatively poor discriminative power as a prognostic indicator emphasizes the need for the adoption of additional measures of lung dysfunction in the assessment of patients with ARDS.
71

tion to only sepsis-related ARDS subjects. A larger study is required to determine whether EVLWp is a useful measure of disease severity in patients with all types of ARDS. In addition, a larger study would allow the use of multivariate analysis to determine the relative importance of possible confounders.

CONCLUSIONS
In summary, this study shows that increased EVLW is a feature of early ARDS. Furthermore, when EVLW is indexed to PBW, instead of actual body weight, an improvement is seen in the predictive value of EVLW for survival and improved correlation with markers of disease severity. Finally, TTT appears to reliably measure EVLW in ARDS patients even with extremely elevated VD/VT. Measuring lung water in this manner may help to better characterize ARDS and to guide future therapeutic intervention, and it may be used to identify the most severely ill patients and to determine the benet of treatment in future clinical trials.

ACKNOWLEDGMENTS
Figure 2. Correlation between lung dysfunction and extravascular lung water (EVLW; left column) and extravascular lung water indexed to predicted body weight (EVLWp; right column). The graphs show pooled average data from all 3 days for survivors and nonsurvivors. The squares of correlation coefcients (r2) and p values from Pearsons product moment correlation are given in each graph. LIS, lung injury score; SOFA, Sequential Organ Failure Assessment score; VD/VT, dead space-tidal volume fraction.

We acknowledge the contribution of the patients and their families in allowing us to conduct this research, as well as the physicians, nurses, and support staff in our intensive care units, who made this project possible.

REFERENCES
EVLW and Severe ARDS. It has been postulated that the increased VD/VT seen early in ARDS is a marker of pulmonary capillary endothelial cell injury (1). Vascular injury and hypoxic pulmonary vasoconstriction resulting in regional hypoperfusion will contribute to an increased pulmonary dead space in ARDS. Because TTT relies on heat exchange across the capillaryalveolar barrier, it has been proposed that the technique will not detect lung water in areas of poorly perfused lung with high VD/VT. Whereas regional loss of correlation between VD/VT and EVLWp cannot be excluded, lack of decreases in EVLWp at the most abnormal measures of VD/VT (Fig. 2), and the very high discriminatory value of EVLWp for mortality, all argue against the proposal that TTT lacks utility in severe cases of ARDS. Study Limitations. The observations made in this study are limited because of the small patient number and the restric1. Ware LB, Matthay MA: The acute respiratory distress syndrome. N Engl J Med 2000; 342: 1334 1349 2. Doyle RL, Szaarski N, Modin GW, et al: Identication of patients with acute lung injury. Predictors of mortality. Am J Respir Crit Care Med 1995; 152(6 Pt 1):1818 1824 3. Zilberberg MD, Epstein SK: Acute lung injury in the medical ICU: Comorbid conditions, age, etiology, and hospital outcome. Am J Respir Crit Care Med 1998; 157(4 Pt 1):1159 1164 4. Rubenfeld GD, Caldwell E, Granton J, et al: Interobserver variability in applying a radiographic denition for ARDS. Chest 1999; 116:13471353 5. Meade MO, Cook RJ, Guyatt GH, et al: Interobserver variation in interpreting chest radiographs for the diagnosis of acute respiratory distress syndrome. Am J Respir Crit Care Med 2000; 161:8590 6. Martin GS, Moss M, Wheeler AP, et al: A randomized, controlled trial of furosemide with or without albumin in hypoproteinemic patients with acute lung injury. Crit Care Med 2005; 33:16811687

Figure 3. Receiver operator characteristic curves for extravascular lung water indexed to predicted body weight (EVLWp), dead space-tidal volume fraction (V D /V T ), extravascular lung water (EVLW), and PaO2/FIO2 for mortality with sensitivity vs. 1-specicity for identication of nonsurvivors. The areas under the curves were 0.988 0.019, 0.869 0.112, 0.851 0.113, and 0.643 0.137 for EVLWp, VD/VT, EVLW, and PaO2/FIO2, respectively.

72

Crit Care Med 2008 Vol. 36, No. 1

7. Perkins GD, McAuley DF, Thickett DR, et al: The beta-Agonist Lung Injury Trial (BALTI): A randomized placebo-controlled clinical trial. Am J Respir Crit Care Med 2006; 173:281287 8. Matthay MA, Abraham E: Beta-adrenergic agonist therapy as a potential treatment for acute lung injury. Am J Respir Crit Care Med 2006; 173:254 255 9. Sartori C, Allemann Y, Duplain H, et al: Salmeterol for the prevention of high-altitude pulmonary edema. N Engl J Med 2002; 346: 16311636 10. Mutlu GM, Koch WJ, Factor P: Alveolar epithelial beta 2-adrenergic receptors: Their role in regulation of alveolar active sodium transport. Am J Respir Crit Care Med 2004; 170:1270 1275 11. Eisenberg PR, Hansbrough JR, Anderson D, et al: A prospective study of lung water measurements during patient management in an intensive care unit. Am Rev Respir Dis 1987; 136:662 668 12. Bernard GR, Artigas A, Brigham KL, et al: The American-European Consensus Conference on ARDS. Denitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149(3 Pt 1):818 824 13. The Acute Respiratory Distress Syndrome Network: Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000; 342:13011308 14. Rossi P, Oldner A, Wanecek M, et al: Comparison of gravimetric and a double-indicator dilution technique for assessment of extravascular lung water in endotoxaemia. Intensive Care Med 2003; 29:460 466

15. Roch A, Michelet P, Lambert D, et al: Accuracy of the double indicator method for measurement of extravascular lung water depends on the type of acute lung injury. Crit Care Med 2004; 32:811 817 16. Katzenelson R, Perel A, Berkenstadt H, et al: Accuracy of transpulmonary thermodilution versus gravimetric measurement of extravascular lung water. Crit Care Med 2004; 32: 1550 1554 17. Kirov MY, Kuzkov VV, Kuklin VN, et al: Extravascular lung water assessed by transpulmonary single thermodilution and postmortem gravimetry in sheep. Crit Care 2004; 8:R451R458 18. Fernandez-Mondejar E, Rivera-Fernandez R, Garcia-Delgado M, et al: Small increases in extravascular lung water are accurately detected by transpulmonary thermodilution. J Trauma 2005; 59:1420 1424 19. Nuckton TJ, Alonso JA, Kallet RH, et al: Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome. N Engl J Med 2002; 346:12811286 20. Kallet RH, Alonso JA, Pittet JF, et al: Prognostic value of the pulmonary dead-space fraction during the rst 6 days of acute respiratory distress syndrome. Respir Care 2004; 49:1008 1014 21. Kallet RH, Daniel BM, Garcia O, et al: Accuracy of physiologic dead space measurements in patients with acute respiratory distress syndrome using volumetric capnography: Comparison with the metabolic monitor method. Respir Care 2005; 50:462 467 22. Vincent JL, Moreno R, Takala J, et al: The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/ failure. Intensive Care Med 1996; 22:707710

23. Murray JF, Matthay MA, Luce JM, et al: An expanded denition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988; 138:720 723 24. Crapo RO, Morris AH, Gardner RM: Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981; 123: 659 664 25. Crapo RO, Morris AH, Gardner RM: Reference values for pulmonary tissue volume, membrane diffusing capacity, and pulmonary capillary blood volume. Bull Eur Physiopathol Respir 1982; 18:893 899 26. Ogden CL, Carroll MD, Curtin LR, et al: Prevalence of overweight and obesity in the United States, 1999 2004. JAMA 2006; 295: 1549 1555 27. Luecke T, Muench E, Roth H, et al: Predictors of mortality in ARDS patients referred to a tertiary care centre: A pilot study. Eur J Anaesthesiol 2006; 23:403 410 28. Agarwal R, Aggarwal AN, Gupta D, et al: Etiology and outcomes of pulmonary and extrapulmonary acute lung injury/ARDS in a respiratory ICU in north India. Chest 2006; 130:724 729 29. Monchi M, Bellenfant F, Cariou A, et al: Early predictive factors of survival in the acute respiratory distress syndrome. A multivariate analysis. Am J Respir Crit Care Med 1998; 158:1076 1081 30. Kuzkov VV, Kirov MY, Sovershaev MA, et al: Extravascular lung water determined with single transpulmonary thermodilution correlates with the severity of sepsis-induced acute lung injury. Crit Care Med 2006; 34: 16471653

Crit Care Med 2008 Vol. 36, No. 1

73