Risk factors for the development of acute lung injury in patients with septic shock: An observational cohort study*

Remzi Iscimen, MD; Rodrigo Cartin-Ceba, MD; Murat Yilmaz, MD; Hasrat Khan, MD; Rolf D. Hubmayr, MD; Bekele Afessa, MD; Ognjen Gajic, MD, MSc

Objective: Almost half of the patients with septic shock develop acute lung injury (ALI). The understanding why some patients do and others do not develop ALI is limited. The objective of this study was to test the hypothesis that delayed treatment of septic shock is associated with the development of ALI. Design: Observational cohort study. Setting: Medical intensive care unit in a tertiary medical center. Patients: Prospectively identied patients with septic shock who did not have ALI at the outset, excluding those who denied research authorization. Measurements and Main Results: High frequency cardio-respiratory monitoring, arterial gas analysis, and portable chest radiographs were reviewed to identify the timing of ALI development. Risk factors present before ALI development were identied by review of electronic medical records and analyzed in univariate and multivariate analyses. Seventy-one of 160 patients (44%)

developed ALI at a median of 5 (range 294) hours after the onset of septic shock. Multivariate logistic regression analysis identied the following predictors of ALI development: delayed goal-directed resuscitation (odds ratio [OR] 3.55, 95% condence interval [CI] 1.52 8.63, p .004), delayed antibiotics (OR 2.39, 95% CI 1.06 5.59, p .039), transfusion (OR 2.75, 95% CI 1.22 6.37, p .016), alcohol abuse (OR 2.09, 95% CI .88 5.10, p 0.098), recent chemotherapy (OR 6.47, 95% CI 1.99 24.9, p 0.003), diabetes mellitus (OR .44, 95% CI .17 1.07, p .076), and baseline respiratory rate (OR 2.03 per SD, 95% CI 1.38 3.08, p < .001). Conclusion: When adjusted for known modiers of ALI expression, delayed treatment of shock and infection were associated with development of ALI. (Crit Care Med 2008; 36:15181522) KEY WORDS: shock; metabolic acidosis; hyperventilation; epidemiology; antibiotic; resuscitation

eptic shock is one of the most common, life-threatening medical conditions and is frequently complicated by organ failures, especially acute lung injury (ALI). Development of ALI is associated with short and long term morbidity, mortality, prolonged hospitalization, and high costs (1). While septic shock has long been recognized as a trigger of ALI and of acute respiratory distress syndrome (ARDS) (2, 3), our understanding of why

*See also p. 1666. From the Department of Internal Medicine, Division of Pulmonary and Critical Care Medicine, Mayo Clinic (RI, RC-C, MY, HK, RDH, BA, OG), Rochester, Minnesota; Department of Anesthesiology and Intensive Care, Uludag University (RI), Medical Faculty, Bursa, Turkey; Department of Anesthesiology and Intensive Care, Akdeniz University (MY), Medical Faculty, Antalya, Turkey. Dr. Hubmayr has consulted for Novartis and the DSMB. The remaining authors have not disclosed any potential conicts of interest. Supported by NIH grant: NHLBI 1 K23 HL087843 For information regarding this article, E-mail: gajic.ognjen@mayo.edu Copyright 2008 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e31816fc2c0

some patients with septic shock do and others do not develop ALI has been limited. Previous studies have identied a history of chronic alcohol use (4), hypoalbuminemia (5), transfusions (2, 6), pulmonary source of infection (2), and the absence of diabetes mellitus (7) as signicant modiers of ALI expression. However, the timing of ALI development has generally been poorly dened, making it difcult to separate cause from effect of specic risk factors, particularly those related to treatment. This study aims to assess the effect of delayed treatment of shock and infection on the development of ALI using data from prospectively identied consecutive patients with septic shock. We hypothesized that delayed goal-directed resuscitation and delayed antibiotic administration are risk factors for the development of ALI in patients with septic shock.

MATERIALS AND METHODS

The Mayo Clinic Institutional Review Board approved the study protocol and waived the need for informed consent in this observational study. Daily screening identied con-

secutive adults ( 18 yrs of age) with septic shock who were admitted to the medical intensive care unit (ICU) of a tertiary care center between March 2004 and April 2007. We excluded patients with preexisting ALI or cardiogenic pulmonary edema at the onset of septic shock, those who denied research authorization, patients in whom care was withdrawn within 6 hrs of onset of septic shock, and those in whom the time of onset of septic shock started before hospital admission or could not be accurately determined (i.e., patients transferred from another facility). The characteristics of the ICU have been previously described (8). The onset of septic shock was determined according to the standards of the American College of Chest Physicians/Society of Critical Care Medicine consensus conference criteria (9) when, in a patient with suspected infection, two consecutive measurements revealed the following: 1) Two systemic inammatory response syndrome (SIRS) criteria (temperature 38.3C or 35.6C, heart rate 90 beats/ min, respiratory rate 20/min or white blood cell [WBC] count 12.0 103 or 4.0 3 10 ); and 2) Hypoperfusion as evidenced by a systolic blood pressure 90 mmHg or mean arterial pressure (MAP) 60 mmHg or a fall of 40 mmHg from baseline despite 20 mL/kg

1518

Crit Care Med 2008 Vol. 36, No. 5

uid bolus, and/or serum lactate 4 mmol/L regardless of the blood pressure (10). Primary outcome was the development of ALI, dened according to the standard American-European Consensus Conference (AECC) criteria (11) as acute onset of: a) New or worsening hypoxemia with partial pressure of oxygen in arterial blood divided by inspired oxygen concentration (PaO2/FIO2) 300 mmHg; b) New bilateral pulmonary inltrates on chest radiography consistent with edema; and c) The absence of clinical signs of left atrial hypertension as the principal explanation for pulmonary edema. Secondary outcome measures included hospital mortality and length of stay. To increase interobserver reliability, outcome assessors reviewed a structured ALI tutorial (12) before study onset. Outcome assessors determined the onset of septic shock and ALI using high frequency cardio-respiratory monitoring as previously described (13). Study investigators identied newly admitted patients with septic shock at 8 a.m. every morning including weekends. The investigators reviewed monitoring logs capturing respiratory rate, heart rate, arterial and central venous pressures, oxygen saturation, FIO2, laboratory ndings, ventilator settings, infusions, and other treatments to determine the time when the criteria for septic shock and for ALI (if present) were met. Bedside nurses validated all of the measurements. The diagnosis of ALI was conrmed at the time of the qualifying portable chest radiograph. The absence of ALI at baseline was assured by the absence of hypoxemia (PaO2/FIO2 300 mm Hg) and/or the presence of chest radiograph without bilateral inltrates (at least one lung had to be clear). Risk factors for development of ALI were grouped as follows: 1) Underlying condition before development of septic shock: demographics, source of infection, aspiration, history of chronic alcohol abuse, diabetes mellitus, tobacco use, preexisting chronic obstructive lung disease, and recent ( 6 months) exposure to chemotherapy. Chronic alcohol abuse was dened as a previously established diagnosis of chronic alcoholism, a prior admission for alcohol detoxication, or alcohol withdrawal (14). Tobacco use was dened as history of cigarette smoking based on the information provided in the standardized form on hospital admission. Aspiration was dened as aspiration of gastric contents into the airway witnessed or strongly suspected by the bedside provider. 2) Severity of illness, inammation, and hypoperfusion at the onset of septic shock including white cell count, platelet count, temperature, respiratory rate, lactate, and base excess. The degree of baseline hypoxemia was estimated from the PaO2/FIO2 (if arterial blood gas analysis was obtained at the time of septic shock onset) or O2sat/FIO2 ratio (15).

The baseline Acute Physiologic and Chronic Health Evaluation (APACHE III) scores were calculated from the data available at the onset of septic shock. 3) Treatment during the rst six hours of septic shock: time to the administration of antibiotic therapy (16), source control (17), goal-directed uid resuscitation (10), transfusion of blood products, use of vasopressors, steroid administration, and use of activated protein C. Adequate antibiotic therapy was dened as empirical use of broad spectrum antibiotics that would cover gram positives, gram negatives, and anaerobes according to the suspected site of infection (16). Blood transfusion was dened as infusion of any plasma blood product including red blood cell (RBC), platelets, fresh frozen plasma (FFP), or cryoprecipitate, during the 48 hrs before the development of ALI within 48 hrs of septic shock in patients who did not develop ALI. Adequate early goal-directed resuscitation was dened as central venous oxygen saturation 70 percent (10) and/or a combination of clinical factors (18): central venous pressure 8 mmHg, MAP 65 mmHg, urine output .5 mL/kg/hour, and/or improvement in mental state (Glasgow Coma Scale-GCS), base excess, or lactate. The achievement of resuscitation goals at any point within the rst six hours was considered as adequate goaldirected resuscitation; achievement of resuscitation goals after the rst 6 hrs was considered delayed goal-directed resuscitation. Adequate antibiotic administration 3 hrs after the onset of septic shock was considered as delayed antibiotic therapy. Monitoring logs were reviewed and only the predictors present at baseline (before any worsening in oxygenation decrease in oxygen saturation to 90% requiring an increase in FIO2 by the bedside providers) were considered for the analysis. In patients in whom FIO2 was not directly measured, O2 supplementation was determined according to following approximations: room air .21, 1L/min .24, 2L/min .28, 3L/min .32, 4L/min .35, 4L/min .4. Statistical Analysis. The risk factors were compared between patients who did and did not develop ALI in a univariate followed by multivariate analysis. Continuous and categorical variables were compared using a Wilcoxon rank sum, Fishers exact test, or chisquare test, as appropriate. The variables were considered for multivariable logistic regression models if they occurred before or during the rst six hours of septic shock resuscitation and before the development of ALI, had less than 10% missing data and the following: a) had p values .1; b) had high odds ratios (OR 2 for categorical variables or per SD of a continuous variable); and c) were biologically plausible. In addition to our main predictors of interest, delayed goal-directed resuscitation and delayed antibiotic therapy that were al-

ways in the model, previously identied modiers of ALI (references 2 6, see the Results section) were included in the multivariate analysis as covariates. The nal model was determined based on a combination of forward selection and backward elimination taking into consideration colinearity, interaction, and the number of patients who experienced outcome of interest. In a case of colinearity between some of the variables (aspiration and pulmonary source) the variable with stronger association (based on forward selection process) was used in multivariate analysis. JMP statistical software (version 6.0, SAS institute, Cary, NC) was used for all analyses.

RESULTS
Seventy-one of the 160 patients (44%) developed ALI at a median of 5 (range 294) hrs after the onset of septic shock. Ninety percent of patients developed ALI during the rst 12 hrs of septic shock, 64 of whom (90%) met the criteria for ARDS. Table 1 presents univariate comparisons of chronic conditions, baseline characteristics, resuscitation parameters, co-exposures, and co-interventions between patients with septic shock who did and did not develop ALI. When adjusted for known modiers of ALI expression (transfusion, aspiration, alcohol abuse, chemotherapy, and diabetes mellitus) in a multivariate logistic regression analysis (Table 2), delayed goaldirected resuscitation, delayed antibiotics, and high baseline respiratory rate were signicantly associated with development of ALI in patients with septic shock. When the analysis was restricted to 112 patients who had PaO2/FIO2 measurements at baseline, delayed resuscitation (odds ratio [OR] 5.0, 95% condence interval [CI] 1.8 15) and delayed antibiotics (OR 2.6, 95% CI 1.0 7.8) were associated with development of ALI. In a subgroup analysis of 97 patients with nonpulmonary source of sepsis, delayed resuscitation (OR 4.8, 95% CI 1.5 16.8) but not delayed antibiotics (OR 1.9, 95% CI .625.74) were signicantly associated with development of ALI. To determine whether the high respiratory rate was more likely to represent hyperventilation from shock and metabolic acidosis versus early pulmonary dysfunction in our study, we posthoc correlated baseline respiratory rate with the degree of baseline hypoxemia and pulmonary impairment. Baseline respiratory rate was signicantly associated with the presence of metabolic acidosis, base def1519

Crit Care Med 2008 Vol. 36, No. 5

Table 1. Risk factors for development of ALI in patients with septic shock ALI (n Underlying condition prior to development of septic shock Age, years (IQR) Female gender, n (%) Diabetes mellitus, n (%) Chronic alcohol use, n (%) History of tobacco smoking, n (%) COPD, n (%) Recent chemotherapy, n (%) ACE inhibitors (%) Amiodarone (%) Source of sepsis, n (%) Pulmonary Abdomen Urine Skin and soft tissue Other Positive blood cultures (%) Aspiration (%) Baseline severity of illness, inammation, and hypoperfusion APACHE III (IQR) at the onset of septic shock Temperature, C (IQR) GCS (IQR) WBC count, per mm3 (IQR) Platelet count, per mm3 (IQR) Respiratory rate (IQR) O2 sat/FIO2 (IQR) PaO2/FIO2 (IQR) n 112 pH (IQR) n 153 Base excess (IQR), n 150 Lactate (IQR), n 148 Serum albumin, mg/dL (IQR), n 79 Serum creatinine, mg/dL (IQR) Treatment during the rst 6 hours of septic shock Time to antibiotic administration, minutes (IQR) Delayed antibiotic administration ( 3 hours), n (%) Source control, n (%) Time to goal-directed resuscitation endpoints, minutes (IQR) Delayed goal-directed resuscitation ( 6 hours), n (%) Mean MAP mm Hg (IQR) Mean CVP (IQR), n 121 Mean SvcO2 (IQR), n 83 Fluid intake, mL rst 6 hrs (IQR) Fluid intake, mL 612 hrs (IQR) Colloid (5% albumin) use, n (%) Any transfusion, n (%) RBC only, n (%) FFP or platelets, n (%) Dobutamine, n (%) Vasopressors, n (%) Corticosteroids, n (%) Mechanical ventilationa, n (%) Activated protein C, n (%)b Vt, mL/kg PBW (IQR), n 44 Ppk, cm H2O (IQR), n 40 Outcome ICU mortality, n (%) Hospital mortality, n (%) 71) No ALI (n 89) p Value

68 (56 to 80) 33 (46) 17 (24) 27 (38) 38 (54) 24 (34) 17 (24) 17 (24) 7 (10) 35 (49) 21 (30) 8 (11) 5 (7) 2 (3) 18 (25) 18 (25) 61 (49 to 72) 36.8 (36.1 to 38.2) 13 (9 to 15) 16 (9 to 24) 192 (113 to 293) 27 (23 to 29) 405 (263 to 457) 200 (100 to 260) 7.27 (7.19 to 7.34) 8 ( 12 to 5) 3.2 (1.7 to 5.5) 2.4 (1.9 to 3.2) 2.1 (1.4 to 3.1) 223 (82.5 to 360) 44 (62) 12 (17) 505 (329 to 653) 36 (51) 58 (53 to 62) 6 (4 to 8) 64 (55 to 75) 4518 (2982 to 6723) 2320 (1450 to 3992) 14 (20) 40 (56) 10 (14) 16 (23) 8 (11) 68 (96) 37 (52) 19 (27) 9 (13) 6.9 (6.1 to 7.8) 30 (27 to 35) 27 (38) 36 (51)

73 (61 to 83) 39 (44) 38 (43) 15 (17) 33 (37) 23 (26) 5 (6) 22 (25) 5 (6) 28 (32) 24 (27) 20 (22) 10 (11) 7 (8) 18 (20) 8 (9) 55 (44 to 64) 36.9 (36.4 to 38.2) 14 (13 to 15) 12 (8 to 18) 181 (107 to 283) 20 (18 to 25) 431 (331 to 467) 300 (222 to 400) 7.37 (7.31 to 7.42) 6 ( 8 to 2) 2.4 (1.4 to 4.1) 2.8 (2.4 to 3.3) 1.8 (1.3 to 3.4) 113 (0 to 235) 36 (40) 12 (14) 320 (202 to 480) 20 (22) 58 (55 to 62) 5 (3 to 8) 67 (61 to 72) 5394 (3476 to 6766) 1851 (1072 to 2893) 17 (19) 27 (30) 15 (17) 12 (14) 7 (8) 67 (75) 40 (45) 35 (39) 2 (2) 7.0 (6.5 to 7.9) 26 (23 to 31) 10 (11) 16 (18)

.056 .737 .012 .003 .037 .273 .001 .910 .313 .064

.441 .005 .028 .499 .039 .032 .548 .001 .033 .001 .001 .001 .038 .050 .566 .001 .007 .550 .001 .001 .703 .200 .580 .527 .020 .922 .001 .631 .136 .465 .001 .367 .093 .008 .362 .097 .001 .001

IQR, interquartile range; WBC, white blood cell; ALI, acute lung injury; MAP, mean arterial blood pressure; ACE, angiotensin-converting enzyme; GCS, Glasgow Coma Scale; RBC, red blood cells; FFP, fresh frozen plasma; COPD, chronic obstructive pulmonary disease; Ppk, peak airway pressure; Vt, tidal volume; PBW, predicted body weight; CVP, central venous pressure; APACHE, Acute Physiology and Chronic Health Evaluation; ICU, intensive care unit. a During the rst 6 hours of resuscitation. In patients who developed ALI, only mechanical ventilation started before the development of ALI was included. bActivated protein C was used after the rst 6 hours in all patients and after the development of ALI in patients with ALI.

1520

Crit Care Med 2008 Vol. 36, No. 5

Table 2. Risk factors for development of ALI in patients with septic shock: multiple logistic regression analysis Odds Ratio Delayed goal-directed resuscitation Delayed antibiotics Respiratory rate (per SD) Chemotherapy Chronic alcohol use Transfusion Aspiration Diabetes mellitus 3.55 2.39 2.03 6.47 2.09 2.75 3.48 .44 95% CI 1.528.63 1.065.59 1.383.08 1.9924.9 .885.10 1.226.37 1.2210.78 .171.07 p Value .004 .039 .001 .003 .098 .016 .024 .076

SD, standard deviation; CI, condence interval; ALI, acute lung injury. In addition to delayed goal-directed resuscitation and delayed antibiotic therapy, the following variables were entered into the stepwise logistic regression procedure: O2 sat /FIO2, diabetes, alcohol, aspiration, age, pulmonary source, smoking, transfusion, vasopressors, chemotherapy, baseline respiratory rate, baseline Acute Physiology and Chronic Health Evaluation III score, white cell count, Glasgow Coma Scale, pH, base decit, lactate.

icit (p .004) but not with hypoxemia (p .209), or a pulmonary source of sepsis (p .790). When adjusted for baseline APACHE III scores, age, metastatic cancer, and delayed resuscitation, development of ALI was associated with increased hospital mortality (OR 4.4, 95% CI 2.0 to 11).

DISCUSSION
In this cohort of patients with septic shock who did not have ALI at the outset, nearly half of the patients developed ALI. As determined from high frequency cardio-respiratory monitoring, the majority of patients developed ALI within the rst 12 hrs after the septic shock onset. When adjusted for known modiers of ALI expression (history of recent chemotherapy, transfusion, aspiration, diabetes mellitus, and alcohol abuse), the development of ALI was associated with delayed treatment of shock (goal-directed resuscitation), inadequate control of infection, and high baseline respiratory rate. Patients who developed ALI had higher hospital mortality. Early antibiotic administration and early goal-directed therapy have been identied as the most important treatment options in septic shock and have been strictly recommended by the Surviving Sepsis Campaign and the Institute for Healthcare Improvement (17). Our study adds to the existing knowledge demonstrating that the development of one of the most important complications of septic shock, ALI, is associated with delayed treatment. In the landmark study by Rivers et al., delayed goal-directed resuscitation was associated with sudden cardiovascular collapse and the trend toCrit Care Med 2008 Vol. 36, No. 5

ward development of multiorgan failure (10). However, the authors did not assess the development of ALI in this study. Although short of statistical signicance, our study conrmed known associations among the chronic alcohol abuse and the absence of diabetes mellitus and the development of ALI (4, 7). Previous studies have shown that even submassive transfusions increase the probability of ALI development acting as one of the multiple hits occurring during resuscitation of critically ill patients (2, 6). As previously reported for a general ICU population, transfusion of plasma containing blood products was associated with a higher ALI risk than transfusion of RBCs (6). Multiple studies identied pulmonary aspiration as a signicant risk factor for ALI development (19). In our study, aspiration appeared to be a more important risk factor than having identied the lung as a likely source of sepsis. The strong association between recent cancer chemotherapy and development of ALI comes as no surprise since ALI is a known complication of many cancer chemotherapeutic agents. Indeed, chemotherapy had the highest odds ratio of all risk factors examined in our study. The mechanisms and their possible interactions with septic shock and its management are not known. However, they likely involve the generation of reactive oxygen species and their effects on alveolar barrier properties. The association between high baseline respiratory rate and the development of ALI conrmed the ndings of a recent large prospective cohort of patients at risk for ALI (2). Baseline respiratory rate correlated with metabolic acidosis (but

not with hypoxemia) supporting an interesting experimental observation that (even spontaneous) hyperventilation may be a contributing factor in the development and expression of ALI (20). Since the spontaneous tidal volumes were not measured, we could not exclude the alternative explanation, i.e., patients who had a higher respiratory rate at baseline may have already had a subclinical form of ALI and were going to develop the full expression of the syndrome some hours later. However, the fact that more patients in the control group underwent mechanical ventilation at baseline 39% vs. 27%, p .09) is not consistent with this alternative explanation. In a subgroup of patients who were mechanically ventilated during the rst six hours of resuscitation, tidal volumes were small and similar between the patients who did and did not develop ALI. Peak airway pressures tended to be higher in patients who were to develop ALI, suggesting an already smaller baby lung compartment. Signicant limitations of our study include the inability to assess causality as opposed to association. For example, the inability to achieve resuscitation thresholds within a certain time period may have been due to inadequate and delayed therapy or to more severe original insult. Also, while the patients with septic shock and ALI were prospectively identied, all of the predictor variables were collected by the passive review of cardio-respiratory monitoring logs, nurses and physicians observations, and patient/family provided information in the electronic medical records leading to potential measurement error. For example, rather than on the basis of standard validated questionnaires, the denition of alcohol abuse was based on a known diagnosis of chronic alcoholism or a previous admission for alcohol detoxication or alcohol withdrawal (14). In addition, because a signicant number of patients did not have arterial blood gas performed at the time of septic shock onset, baseline PaO2/ FIO2 was substituted by O2sat/FIO2 (15), which was available in all patients. Nevertheless, our study is one of the rst attempts to assess the exact timing of exposures of interest in relationship to the development of critical care syndromes, such as ALI, using data from continuous monitoring available through high resolution clinical information systems. An additional limitation of our study stems from the complex and some1521

what imprecise denitions of both critical care syndromes (septic shock and ALI) and their supportive interventions (timing of adequate antibiotics and resuscitation). While we paid attention to use precise denitions, formally trained the observers (12), and independently assessed predictors and the outcome, full blinding of the observers was not possible within the integrated electronic information system. Finally, the generalizability of our results is limited as it was done in the medical ICU of a single institution. The fact that patients who developed ALI in our study had higher adjusted hospital mortality is similar to some but not all previous reports. Eggimann et al. recently found that ALI per se was not associated with an increased mortality when adjusted to baseline characteristics and evolving organ failures (21). Since we limited our predictors to only those present before ALI development, we did not adjust for subsequent development of nonpulmonary organ failures. It is not known to what extent nonpulmonary organ failures are caused or contributed by underlying septic shock and its management as opposed to biotrauma related to ALI development and its treatment with mechanical ventilation (22). In conclusion, ALI is a common complication and occurs early in the course of septic shock. In addition to the known modiers of ALI expression such as transfusion, aspiration, chemotherapy, alcohol use, and diabetes, our study identied delayed goal-directed resuscitation, delayed antibiotic administration, and baseline high respiratory rate as important determinants of ALI development. Targeting the modiers of ALI expression may be an attractive strategy for prevention of this common and devastating ICU complication.

REFERENCES
1. Rubenfeld GD, Herridge MS: Epidemiology and outcomes of acute lung injury. Chest 2007; 131:554 562 2. Gong MN, Thompson BT, Williams P, et al: Clinical predictors of and mortality in acute respiratory distress syndrome: potential role of red cell transfusion. Crit Care Med 2005; 33:11911198 3. Hudson LD, Milberg JA, Anardi D, et al: Clinical risks for development of the acute respiratory distress syndrome. Am J Respir Crit Care Med 1995; 151:293301 4. Moss M, Burnham EL: Chronic alcohol abuse, acute respiratory distress syndrome, and multiple organ disfunction. Crit Care Med 2003; 31(4 Suppl):S207S212 5. Mangialardi RJ, Martin GS, Bernard GR, et al: Hypoproteinemia predicts acute respiratory distress syndrome development, weight gain, and death in patients with sepsis. Ibuprofen in Sepsis Study Group. Crit Care Med 2000; 28:31373145 6. Khan H, Belsher J, Yilmaz M, et al: Freshfrozen plasma and platelet transfusions are associated with development of acute lung injury in critically ill medical patients. Chest 2007; 131:1308 1314 7. Moss M, Guidot DM, Steinberg KP, et al: Diabetic patients have a decreased incidence of acute respiratory distress syndrome. Crit Care Med 2000; 28:21872192 8. Afessa B, Keegan MT, Hubmayr RD, et al: Evaluating the performance of an institution using an intensive care unit benchmark. Mayo Clin Proc 2005; 80:174 180 9. Bone RC, Sibbald WJ, Sprung CL: The ACCPSCCM consensus conference on sepsis and organ failure. Chest 1992; 101:14811483 10. Rivers E, Nguyen B, Havstad S, et al: Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001; 345:1368 1377 11. 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; 149:818 824 12. Tutorial: Improving the Radiographic Diagnosis of Acute Lung Injury. Tutorial. June 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

20th 2007; http://depts.washington.edu/ kclip/about.shtml Gajic O, Rana R, Winters JL, et al: Transfusion related acute lung injury in the critically Ill: prospective nested case-control study. Am J Respir Crit Care Med 2007; 176: 839 840 Moss M, Bucher B, Moore FA, et al: The role of chronic alcohol abuse in the development of acute respiratory distress syndrome in adults. JAMA 1996; 275:50 54 Rice TW, Wheeler AP, Bernard GR, et al: Comparison of the SpO2/FiO2 Ratio and the PaO2/FiO2 Ratio in Patients with Acute Lung Injury or Acute Respiratory Distress Syndrome. Chest 2007; 132:410 417 Kumar A, Roberts D, Wood KE, et al: Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006; 34:1589 1596 Dellinger RP, Carlet JM, Masur H, et al: Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med 2004; 32:858 873 Antonelli M, Levy M, Andrews PJ, et al: Hemodynamic monitoring in shock and implications for management. International Consensus Conference, Paris, France, 2728 April 2006. Intensive Care Med 2007; 33: 575590 Garber BG, Hebert PC, Yelle JD, et al: Adult respiratory distress syndrome: a systemic overview of incidence and risk factors. Crit Care Med 1996; 24:687 695 Mascheroni D, Kolobow T, Fumagalli R, et al: Acute respiratory failure following pharmacologically induced hyperventilation: an experimental animal study. Intensive Care Med 1988; 15:8 14 Eggimann P, Harbarth S, Ricou B, et al: Acute respiratory distress syndrome after bacteremic sepsis does not increase mortality. Am J Respir Crit Care Med 2003; 167: 1210 1214 Tremblay L, Valenza F, Ribeiro SP, et al: Injurious ventilatory strategies increase cytokines and c-fos m-RNA expression in an isolated rat lung model. J Clin Invest 1997; 99:944 952

1522

Crit Care Med 2008 Vol. 36, No. 5