V Acute Respiratory Distress Syndrome The Berlin Definition SPECIAL COMMUNICATION
Transcription
V Acute Respiratory Distress Syndrome The Berlin Definition SPECIAL COMMUNICATION
SPECIAL COMMUNICATION ONLINE FIRST Acute Respiratory Distress Syndrome The Berlin Definition The ARDS Definition Task Force* V ALID AND RELIABLE DEFINI - tions are essential to conduct epidemiological studies succ e s s f u lly a n d t o facilitate enrollment of a consistent patient phenotype into clinical trials.1 Clinicians also need such definitions to implement the results of clinical trials, discuss prognosis with families, and plan resource allocation. Following the initial description of acute respiratory distress syndrome (ARDS) by Ashbaugh et al2 in 1967, multiple definitions were proposed and used until the 1994 publication of the American-European Consensus Conference (AECC) definition.3 The AECC defined ARDS as the acute onset of hypoxemia (arterial partial pressure of oxygen to fraction of inspired oxygen [PaO2/FIO2] ⱕ200 mm Hg) with bilateral infiltrates on frontal chest radiograph, with no evidence of left atrial hypertension. A new overarching entity— acute lung injury (ALI)—was also described, using similar criteria but with less severe hypoxemia (PaO2/FIO2 ⱕ300 mm Hg).3 The AECC definition was widely adopted by clinical researchers and clinicians and has advanced the knowledge of ARDS by allowing the acquisition of clinical and epidemiological data, which in turn have led to improvements in the ability to care for patients with ARDS. However, after 18 years of applied research, a number of issues regarding various criteria of the AECC definition have emerged, including a lack of explicit For editorial comment see p 2542. 2526 JAMA, June 20, 2012—Vol 307, No. 23 The acute respiratory distress syndrome (ARDS) was defined in 1994 by the American-European Consensus Conference (AECC); since then, issues regarding the reliability and validity of this definition have emerged. Using a consensus process, a panel of experts convened in 2011 (an initiative of the European Society of Intensive Care Medicine endorsed by the American Thoracic Society and the Society of Critical Care Medicine) developed the Berlin Definition, focusing on feasibility, reliability, validity, and objective evaluation of its performance. A draft definition proposed 3 mutually exclusive categories of ARDS based on degree of hypoxemia: mild (200 mm Hg⬍PaO2/FIO2 ⱕ300 mm Hg), moderate (100 mm Hg⬍PaO2/FIO2 ⱕ200 mm Hg), and severe (PaO2/ FIO2 ⱕ100 mm Hg) and 4 ancillary variables for severe ARDS: radiographic severity, respiratory system compliance (ⱕ40 mL/cm H2O), positive endexpiratory pressure (ⱖ10 cm H2O), and corrected expired volume per minute (ⱖ10 L/min). The draft Berlin Definition was empirically evaluated using patientlevel meta-analysis of 4188 patients with ARDS from 4 multicenter clinical data sets and 269 patients with ARDS from 3 single-center data sets containing physiologic information. The 4 ancillary variables did not contribute to the predictive validity of severe ARDS for mortality and were removed from the definition. Using the Berlin Definition, stages of mild, moderate, and severe ARDS were associated with increased mortality (27%; 95% CI, 24%-30%; 32%; 95% CI, 29%-34%; and 45%; 95% CI, 42%-48%, respectively; P⬍.001) and increased median duration of mechanical ventilation in survivors (5 days; interquartile [IQR], 2-11; 7 days; IQR, 4-14; and 9 days; IQR, 5-17, respectively; P⬍.001). Compared with the AECC definition, the final Berlin Definition had better predictive validity for mortality, with an area under the receiver operating curve of 0.577 (95% CI, 0.561-0.593) vs 0.536 (95% CI, 0.520-0.553; P⬍.001). This updated and revised Berlin Definition for ARDS addresses a number of the limitations of the AECC definition. The approach of combining consensus discussions with empirical evaluation may serve as a model to create more accurate, evidence-based, critical illness syndrome definitions and to better inform clinical care, research, and health services planning. JAMA. 2012;307(23):2526-2533 Published online May 21, 2012. doi:10.1001/jama.2012.5669 criteria for defining acute, sensitivity of PaO2/FIO2 to different ventilator settings, poor reliability of the chest radiograph criterion, and difficulties distinguishing hydrostatic edema (TABLE 1).4 www.jama.com *Authors/Writing Committee and the Members of the ARDS Definition Task Force are listed at the end of this article. Corresponding Author: Gordon D. Rubenfeld, MD, MSc, Program in Trauma, Emergency, and Critical Care, Sunnybrook Health Sciences Center, 2075 Bayview Ave, Toronto, ON M4N 3M5, Canada (gordon [email protected]). ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME For these reasons, and because all disease definitions should be reviewed periodically, the European Society of Intensive Care Medicine convened an international expert panel to revise the ARDS definition, with endorsement from the American Thoracic Society and the Society of Critical Care Medicine. The objectives were to update the definition using new data (epidemiological, physiological, and clinical trials) to address the current limitations of the AECC definition and explore other defining variables. Methods Consensus Process. Three co-chairs were appointed by the European Society of Intensive Care Medicine, who in turn selected panelists based on their work in the area of ARDS and to ensure geographic representation from both Europe and North America. An overview of the consensus process used by the panel is outlined in the FIGURE. In revising the definition of ARDS, the panel emphasized feasibility, reliability, face validity (ie, how clinicians recognize ARDS), and predictive validity (ie, ability to predict response to therapy, outcomes, or both). In addition, the panel determined that any revision of the definition should be compatible with the AECC definition to facilitate interpretation of previous studies. After initial preparations and an in-person consensus discussion, a draft definition was proposed,13 which underwent empirical evaluation. The definition was further refined through consensus discussion informed by these empirical data. Empirical Evaluation of Draft Definition. Cohort Assembly. Through the review of the literature presented at the consensus meeting, discussions with other experts, and review of personal files, the panel identified studies that met the following eligibility criteria: (1) large, multicenter prospective cohorts, including consecutive patients or randomized trials, or smaller, single-center prospective studies with unique radiological or physiological data that enrolled adult patients with ALI as defined by AECC; Table 1. The AECC Definition3—Limitations and Methods to Address These in the Berlin Definition AECC Definition Acute onset AECC Limitations No definition of acute4 ALI category All patients with PaO2/ FIO2 ⬍300 mm Hg Oxygenation PaO2/FIO2 ⱕ300 mm Hg (regardless of PEEP) Misinterpreted as PaO2/FIO2 = 201-300, leading to confusing ALI/ARDS term Inconsistency of PaO2/ FIO2 ratio due to the effect of PEEP and/or FIO25-7 Chest radiograph Bilateral infiltrates observed on frontal chest radiograph PAWP PAWP ⱕ18 mm Hg when measured or no clinical evidence of left atrial hypertension Risk factor None Timing Poor interobserver reliability of chest radiograph interpretation8,9 High PAWP and ARDS may coexist10,11 Poor interobserver reliability of PAWP and clinical assesments of left atrial hypertension12 Not formally included in definition4 Addressed in Berlin Definition Acute time frame specified 3 Mutually exclusive subgroups of ARDS by severity ALI term removed Minimal PEEP level added across subgroups FIO2 effect less relevant in severe ARDS group Chest radiograph criteria clarified Example radiographs created a PAWP requirement removed Hydrostatic edema not the primary cause of respiratory failure Clinical vignettes created a to help exclude hydrostatic edema Included When none identified, need to objectively rule out hydrostatic edema Abbreviations: AECC, American-European Consensus Conference; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, fraction of inspired oxygen; PaO2, arterial partial pressure of oxygen; PAWP, pulmonary artery wedge pressure; PEEP, positive end-expiratory pressure. aAvailable on request. (2) studies collected granular data necessary to apply the individual criteria of both the draft Berlin Definition and the AECC definition; and (3) authors of these original studies were willing to share data and collaborate. The panel identified 7 distinct data sets (4 multicenter clinical studies for the clinical database14-17 and 3 single-center physiological studies for the physiological database18-20) that met these criteria. Further details of these studies are included in the eMethods (http://www.jama .com). Variables. Studies provided data on hospital or 90-day mortality. Ventilatorfree days at 28 days after the diagnosis of ALI were calculated as a composite measure of mortality and duration of mechanical ventilation. Duration of mechanical ventilation in survivors was selected as an indirect marker of severity of lung injury because this outcome is not biased by mortality or decisions ©2012 American Medical Association. All rights reserved. related to the withdrawal of lifesustaining treatments.21 Progression of severity of ARDS within 7 days was assessed using the longitudinal data collected within each cohort. We distinguished patients with more extensive involvement on the frontal chest radiograph (3 or 4 quadrants) from those with the minimal criterion of “bilateral opacities” (2 quadrants). Static compliance of the respiratory system (CRS) was calculated as tidal volume (mL) divided by plateau pressure (cm H2O) minus positive endexpiratory pressure (PEEP) (cm H2O). The corrected expired volume per minute (V̇ECORR) was calculated as the measured minute ventilation multiplied by the arterial partial pressure of carbon dioxide (PaCO2) divided by 40 mm Hg.22 Total lung weight was estimated from quantitative computed tomography (CT) images.23 Shunt was calculated at one site as previously reported.24 JAMA, June 20, 2012—Vol 307, No. 23 2527 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME Figure. Outline of Consensus Process Premeeting preparations (May to September 2011) Selection of panelists by chairs Precirculation of key topics for discussion Preparation of background material by panelists In-person discussions (September 30 to October 2, 2011, Berlin, Germany) Presentations of key background material Development of the conceptual model of ARDS Draft of Berlin Definition based on informal consensus discussions Empirical evaluation of draft definition (October 2011 to January 2012) Assembling clinical and physiologic cohorts Demonstration of patient characteristics and distribution according to definition categories Evaluation of impact of ancillary variables for severe ARDS subgroup Follow-up of consensus discussions and analysis (February 2012 by multiple teleconferences) Presentation of empirical evaluation Final definition created based on further informal consensus discussions Decision to present the results of a post hoc higher-risk subset Testing of predictive validity ARDS indicates acute respiratory distress syndrome. Analytic Framework and Statistical Methods. The analytic framework for evaluating the draft Berlin ARDS Definition was to (1) determine the distribution of patient characteristics across the defined severity categories; (2) evaluate the value of proposed ancillary variables (more severe radiographic criterion, higher PEEP levels, static respiratory compliance, and V̇ECORR) in defining the severe ARDS subgroup in the draft definition; (3) determine the predictive validity for mortality of the final Berlin Definition; and (4) compare the final Berlin Defini2528 JAMA, June 20, 2012—Vol 307, No. 23 tion to the AECC definition. In addition, in a post hoc analysis, we sought thresholds for C RS and V̇ E CORR that would identify a severe group of patients with ARDS who had more than 50% mortality and include more than 10% of the study population. We did not evaluate other PaO2/FIO2 cutoffs or the requirement of a minimum PEEP level (5 cm H2O) as they were selected by the panel using face validity criteria and to ensure compatibility with prior definitions. Similarly, we did not explore other variables that might improve predictive validity, such as age and severity of nonpulmonary organ failure, because they were not specific to the definition of ARDS.25 To compare the predictive validity of the AECC definition and the Berlin Definition, we used the area under the receiver operating curve (AUROC or C statistic) in logistic regression models of mortality with a dummy variable for the ARDS definition categories.26 Because this technique requires independent categories to create the dummy variable and the AECC definition for ARDS is a subset of ALI, we could not compare the AECC definition as specified. Therefore, we modified the AECC definition and divided ALI into the independent categories of ALI nonARDS (200 mm Hg⬍PaO2/FIO2 ⱕ 300 mm Hg) and ARDS alone (Pa O 2 / FIO2 ⱕ200 mm Hg). Although the category of ALI non-ARDS is not explicitly described by the AECC, it has been used by many investigators.27,28 P values for categorical variables were calculated with the 2 test; P values for continuous variables were estimated with the t test, Mann-Whitney, analysis of variance, or Kruskal-Wallis, depending on the distribution and number of variables. The receiver operating curve statistical analyses were performed by using MedCalc for Windows version 12.1.4.0 (MedCalc Software) and other statistical tests were performed with SAS/STAT for Windows version 9.2 (SAS Institute Inc). Statistical significance was assessed at the 2-sided P ⬍.05 level. Results Draft Consensus Definition. The ARDS Conceptual Model. The panel agreed that ARDS is a type of acute diffuse, inflammatory lung injury, leading to increased pulmonary vascular permeability, increased lung weight, and loss of aerated lung tissue. The clinical hallmarks are hypoxemia and bilateral radiographic opacities, associated with increased venous admixture, increased physiological dead space, and decreased lung compliance. The morphological hallmark of the acute phase is diffuse alveolar damage (ie, edema, inflammation, hyaline membrane, or hemorrhage).29 Draft Definition Criteria. Following 2 days of consensus discussions, the panel proposed a draft definition with 3 mutually exclusive severity categories (mild, moderate, and severe) of ARDS. A set of ancillary variables was proposed to further characterize severe ARDS and these were explicitly specified for further empirical evaluation.13 Timing. Most patients with ARDS are identified within 72 hours of recognition of the underlying risk factor, with nearly all patients with ARDS identified within 7 days.30 Accordingly, for a patient to be defined as having ARDS, the onset must be within 1 week of a known clinical insult or new or worsening respiratory symptoms. Chest Imaging. The panel retained bilateral opacities consistent with pulmonary edema on the chest radiograph as defining criteria for ARDS, but also explicitly recognized that these findings could be demonstrated on CT scan instead of chest radiograph. More extensive opacities (ie, 3 or 4 quadrants on chest radiograph) were proposed as part of the severe ARDS category and identified for further evaluation. Origin of Edema. Given the declining use of pulmonary artery catheters and because hydrostatic edema in the form of cardiac failure or fluid overload may coexist with ARDS,10,11 the pulmonary artery wedge pressure criterion was removed from the defini- ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME Table 2. Exploration of Proposed Variables to Define Severe ARDS a Mild Severe ARDS Definition Consensus panel draft PaO2/FIO2 ⱕ100 mm Hg ⫹ chest radiograph of 3 or 4 quadrants ⫹ PEEP ⱖ10 cm H2O ⫹ (CRS ⱕ40 mL/cm H2O or V̇ECORR ⱖ10 L/min) Consensus panel final PaO2/FIO2 ⱕ100 mm Hg Moderate Severe No. (%) of Patients % Mortality (95% CI) No. (%) of Patients % Mortality (95% CI) No. (%) of Patients 220 (22) 27 (24-30) 2344 (64) 35 (33-36) 507 (14) 220 (22) 27 (24-30) 1820 (50) 32 (29-34) 1031 (28) % Mortality (95% CI) 45 (40-49) b 45 (42-48) b,c Abbreviations: ARDS, acute respiratory distress syndrome; CRS, compliance of the respiratory system; FIO2, fraction of inspired oxygen; PaO2, arterial partial pressure of oxygen; PEEP, positive end-expiratory pressure; V̇ECORR, corrected expired volume per minute. a The moderate group includes patients with PaO /FIO ⱕ200 mm Hg and patients with PaO /FIO ⱕ100 mm Hg who do not meet the additional criteria for severe ARDS in the draft 2 2 2 2 definition. All patients are receiving at least 5 cm H2O PEEP and have bilateral infiltrates on chest radiograph. b P⬍.001 comparing mortality across stages of ARDS (mild, moderate, severe) for draft and final definitions. c P=.97 comparing mortality in consensus draft severe ARDS to consensus final severe ARDS definitions. tion. Patients may qualify as having ARDS as long as they have respiratory failure not fully explained by cardiac failure or fluid overload as judged by the treating physician using all available data. If no ARDS risk factor (eTable 1) is apparent, some objective evaluation (eg, with echocardiography) is required to help eliminate the possibility of hydrostatic edema. Oxygenation. The term acute lung injury as defined by the AECC was removed, due to the perception that clinicians were misusing this term to refer to a subset of patients with less severe hypoxemia rather than its intended use as an inclusive term for all patients with the syndrome. Positive end-expiratory pressure can markedly affect PaO2/FIO25,6; therefore, a minimum level of PEEP (5 cm H2O), which can be delivered noninvasively in mild ARDS, was included in the draft definition of ARDS. A minimum PEEP level of 10 cm H2O was proposed and empirically evaluated for the severe ARDS category. Additional Physiologic Measurements. Compliance of the respiratory system largely reflects the degree of lung volume loss.2 Increased dead space is common in patients with ARDS and is associated with increased mortality. 2 4 However, because the measurement of dead space is challenging, the panel chose minute ventilation standardized at a PaCO2 of 40 mm Hg (V̇ECORR = minute ventilation ⫻ Pa CO 2 /40) as a surrogate.22 The draft definition of severe ARDS included the requirement of either a low respiratory system compliance (⬍40 mL/cm H2O), a high V̇ECORR (⬎10 L/min), or both. These variables were identified for further study during the evaluation phase. The panel considered a number of additional measures to improve specificity and face validity for the increased pulmonary vascular permeability and loss of aerated lung tissue that are the hallmarks of ARDS, including CT scanning, and inflammatory or genetic markers (eTable 2). The most common reasons for exclusion of these measures were lack of routine availability, lack of safety of the measure in critically ill patients, or a lack of demonstrated sensitivity, specificity, or both for use as a defining characteristic for ARDS. Empirical Evaluation of the Draft Definition. Patients. A total of 4188 patients in the clinical database had sufficient data to classify as having ARDS by the AECC definition. Of these patients, 518 (12%) could not be classified by the draft Berlin Definition because PEEP was missing or was less than 5 cm H2O. Patients who could not be classified by the draft Berlin Definition had a mortality rate of 35% (95% CI, 31%-39%), a median (interquartile range [IQR]) of 19 (1-25) ventilator-free days, and a median (IQR) duration of mechanical ventilation in survivors of 4 (2-8) days. These patients were excluded from analyses of the draft Berlin Definition and comparisons between the AECC ©2012 American Medical Association. All rights reserved. definition and the draft Berlin Definition. Compared with patients from the population-based cohorts, patients from clinical trials and the academic centers cohorts were younger, had more severe hypoxemia, and had more opacities on chest radiographs. The cohort of patients from the clinical trials had the lowest mortality, likely reflecting the inclusion and exclusion criteria of the trials.31 The cohort of patients from academic centers had the highest mortality and the lowest percentage of trauma patients, reflecting the referral population (eTable 3). There were 269 patients in the physiological database with sufficient data to classify ARDS by the AECC definition, although the numbers of patients in each cohort were small. Patients in the Turin cohort had worse PaO2/FIO2 ratios and had higher mortality than the other studies (eTable 4). Evaluation of Ancillary Variables. The draft Berlin Definition for severe ARDS that included a PaO2/FIO2 of 100 mm Hg or less, chest radiograph with 3 or 4 quadrants with opacities, PEEP of at least 10 cm H2O, and either a CRS of 40 mL/cm H2O or less or a V̇ECORR of at least 10 L/min identified a smaller set of patients with identical mortality to the simpler severe ARDS category of PaO2/FIO2 of 100 mm Hg or less (TABLE 2). To address the possibility that the CRS and V̇ECORR thresholds might be different in patients with higher body weight, we evaluated weight-adjusted cutoffs for JAMA, June 20, 2012—Vol 307, No. 23 2529 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME these variables in one of the cohorts. There was no significant difference in the predictive validity of the weightadjusted criteria. The consensus panel reviewed these results and considered the lack of evidence for predictive validity of these ancillary variables and their potential contribution to face validity and construct validity and decided to use the simpler definition for severe ARDS that relied on oxygenation alone. The Berlin Definition. The final Berlin Definition of ARDS is shown in TABLE 3. Twenty-two percent (95% CI, 21%-24%) of patients met criteria for mild ARDS (which is comparable with the ALI non-ARDS category of the AECC definition; TABLE 4), 50% (95% CI, 48%-51%) of patients met criteria for moderate ARDS, and 28% (95% CI, 27%-30%) of patients met criteria for severe ARDS. Mortality increased with stages of ARDS from mild (27%; 95% CI, 24%-30%) to moderate (32%; 95% CI, 29%-34%) to severe (45%; 95% CI, 42%-48%). Median (IQR) ventilatorfree days declined with stages of ARDS from mild (20 [1-25] days) to moderate (16 [0-23] days) to severe (1 [020] day). Median (IQR) duration of mechanical ventilation in survivors increased with stages of ARDS from mild (5 [2-11] days) to moderate (7 [414] days) to severe (9 [5-17] days). Using the Berlin Definition, 29% (95% CI, 26%-32%) of patients with mild ARDS at baseline progressed to moderate ARDS and 4% (95% CI, 3%-6%) progressed to severe ARDS within 7 days; and 13% (95% CI, 11%-14%) of pa- Table 3. The Berlin Definition of Acute Respiratory Distress Syndrome Timing Chest imaging a Origin of edema Oxygenation b Mild Moderate Severe Acute Respiratory Distress Syndrome Within 1 week of a known clinical insult or new or worsening respiratory symptoms Bilateral opacities—not fully explained by effusions, lobar/lung collapse, or nodules Respiratory failure not fully explained by cardiac failure or fluid overload Need objective assessment (eg, echocardiography) to exclude hydrostatic edema if no risk factor present 200 mm Hg ⬍ PaO2/FIO2 ⱕ 300 mm Hg with PEEP or CPAP ⱖ5 cm H2O c 100 mm Hg ⬍ PaO2/FIO2 ⱕ 200 mm Hg with PEEP ⱖ5 cm H2O PaO2/FIO2 ⱕ 100 mm Hg with PEEP ⱖ5 cm H2O Abbreviations: CPAP, continuous positive airway pressure; FIO2, fraction of inspired oxygen; PaO2, partial pressure of arterial oxygen; PEEP, positive end-expiratory pressure. a Chest radiograph or computed tomography scan. b If altitude is higher than 1000 m, the correction factor should be calculated as follows: [PaO /FIO ⫻(barometric pressure/ 2 2 760)]. c This may be delivered noninvasively in the mild acute respiratory distress syndrome group. tients with moderate ARDS at baseline progressed to severe ARDS within 7 days. All differences between outcome variables across categories of modified AECC (ALI non-ARDS and ARDS alone) and across categories of Berlin Definition (mild, moderate, and severe) were statistically significant (P⬍.001). Compared with the AECC definition, the final Berlin Definition had better predictive validity for mortality with an AUROC of 0.577 (95% CI, 0.5610.593) vs 0.536 (95% CI, 0.5200.553; P⬍.001), with the difference in AUROC of 0.041 (95% CI, 0.0300.050). To ensure that missing PEEP data in one of the cohorts did not bias the results, the regression analysis was repeated without this cohort and yielded similar results. The Berlin Definition performed similarly in the physiological database as in the clinical database (TABLE 5, eFigure 1, and eFigure 2). Twenty-five percent (95% CI, 20%-30%) of patients met criteria for mild ARDS, 59% (95% CI, 54%-66%) of patients met criteria for moderate ARDS, and 16% (95% CI, 11%21%) of patients met criteria for severe ARDS. Mortality increased with stages of ARDS from mild (20%; 95% CI, 11%31%) to moderate (41%; 95% CI, 33%49%) to severe (52%; 95% CI, 36%68%), with P = .001 for differences in mortality across stages of ARDS. Median (IQR) ventilator-free days declined with stages of ARDS from mild Table 4. Predictive Validity of ARDS Definitions in the Clinical Database Modified AECC Definition a No. (%) [95% CI] of patients Progression in 7 d from mild, No. (%) [95% CI] Progression in 7 d from moderate, No. (%) [95% CI] Mortality, No. (%) [95% CI] b Ventilator-free days, median (IQR) b Duration of mechanical ventilation in survivors, median (IQR), d b ALI Non-ARDS 1001 (24) [23-25] ARDS 3187 (76) [75-77] 336 (34) [31-37] Berlin Definition ARDS a Mild 819 (22) [21-24] Moderate 1820 (50) [48-51] 234 (29) [26-32] Severe 1031 (28) [27-30] 33 (4) [3-6] 230 (13) [11-14] 263 (26) [23-29] 20 (2-25) 5 (2-10) 1173 (37) [35-38] 12 (0-22) 7 (4-14) 220 (27) [24-30] 20 (1-25) 5 (2-11) 575 (32) [29-34] 16 (0-23) 7 (4-14) 461 (45) [42-48] 1 (0-20) 9 (5-17) Abbreviations: AECC, American-European Consensus Conference; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, fraction of inspired oxygen; IQR, interquartile range; PaO2, arterial partial pressure of oxygen; PEEP, positive end-expiratory pressure. a The definitions are the following for ALI non-ARDS (200 mm Hg⬍PaO /FIO ⱕ300 mm Hg, regardless of PEEP), ARDS (PaO /FIO ⱕ200 mm Hg, regardless of PEEP), mild Ber2 2 2 2 lin Definition (200 mm Hg⬍PaO2 /FIO2 ⱕ300 mm Hg with PEEP ⱖ5 cm H2O), moderate Berlin Definition (100 mm Hg⬍PaO2 /FIO2 ⱕ200 mm Hg with PEEP ⱖ5 cm H2O), and severe Berlin Definition (PaO2 /FIO2 ⱕ100 mm Hg with PEEP ⱖ5 cm H2O). b Comparisons of mortality, ventilator-free days, and duration of mechanical ventilation in survivors across categories of modified AECC (ALI non-ARDS and ARDS) and across categories of Berlin Definition (mild, moderate, and severe) are all statistically significant (P⬍.001). 2530 JAMA, June 20, 2012—Vol 307, No. 23 ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME Table 5. Predictive Validity of ARDS Definitions in the Physiologic Database Modified AECC Definition a No. (%) [95% CI] of patients Mortality, No. (%) [95% CI] b Ventilator-free days Median (IQR) Missing, No. Duration of mechanical ventilation in survivors, median (IQR), d Lung weight, mg c Mean (SD) Missing, No. Shunt, mean (SD), % c,d ALI Non-ARDS 66 (25) [19-30] 13 (20) [11-31] 8.5 (0-23.5) 10 6.0 (3.3-20.8) 1371 (360.4) 16 21 (21) Berlin Definition ARDS a ARDS 203 (75) [70-80] 84 (43) [36-50] Mild 66 (25) [20-30] 13 (20) [11-31] Moderate 161 (59) [54-66] 62 (41) [33-49] 0 (0-16.0) 26 13.0 (5.0-25.5) 8.5 (0-23.5) 10 6.0 (3.3-20.8) 0 (0-16.5) 25 12.0 (5.0-19.3) 0 (0-6.5) 1 19.0 (9.0-48.0) 1556 (469.7) 32 29 (11) 1828 (630.2) 16 40 (16) 1602 (508.1) 48 32 (13) 1371 (360.4) 16 21 (12) Severe 42 (16) [11-21] 22 (52) [36-68] Abbreviations: AECC, American-European Consensus Conference; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, fraction of inspired oxygen; IQR, interquartile range; PaO2, arterial partial pressure of oxygen; PEEP, positive end-expiratory pressure. a The definitions are the following for ALI non-ARDS (200 mm Hg⬍PaO /FIO ⱕ300 mm Hg, regardless of PEEP), ARDS (PaO /FIO ⱕ200 mm Hg, regardless of PEEP), mild Ber2 2 2 2 lin Definition (200 mm Hg⬍PaO2 /FIO2 ⱕ300 mm Hg with PEEP ⱖ5 cm H2O), moderate Berlin Definition (100 mm Hg⬍PaO2 /FIO2 ⱕ200 mm Hg with PEEP ⱖ5 cm H2O), and severe Berlin Definition (PaO2 /FIO2 ⱕ100 mm Hg with PEEP ⱖ5 cm H2O). b Eight patients are missing in the moderate Berlin Definition ARDS group. P=.001 for difference in mortality across Berlin stages of ARDS. c Comparisons of lung weight and shunt across categories of modified AECC (ALI non-ARDS and ARDS) and across categories of Berlin Definition (mild, moderate, and severe) are statistically significant (P⬍.001). d Only available at 1 site. (8.5 [0-23.5] days) to moderate (0 [016.5] days) to severe (0 [0-6.5] days), with P=.003 for differences in ventilatorfree days across stages of ARDS. Median (IQR) duration of mechanical ventilation in survivors increased with stages of ARDS from mild (6.0 [3.3-20.8] days) to moderate (12.0 [5.0-19.3] days) to severe (19.0 [9.0-48.0] days), with P=.045 for differences in duration of mechanical ventilation in survivors across stages of ARDS. Using the Berlin Definition, stages of mild, moderate, and severe ARDS had increased mean lung weight by CT scan (1371 mg; 95% CI, 1268-1473; 1556 mg; 95% CI, 1474-1638; and 1828 mg; 95% CI, 1573-2082; respectively) and increased mean shunt (21%; 95% CI, 16%26%; 29%; 95% CI, 26%-32%; and 40%; 95% CI, 31%-48%; respectively). Comparisons of lung weight and shunt (from the single site providing these data) across categories of modified AECC (ALI non-ARDS and ARDS alone) and across categories of Berlin Definition (mild, moderate, and severe) were statistically significant (P⬍.001) (Table 5, eFigure 3, and eFigure 4). In a post hoc analysis, combining a PaO2/FIO2 of 100 mm Hg or less with either a Crs of 20 mL/cm H2O or less or a V̇ECORR of at least 13 L/min identified a higher-risk subgroup among pa- tients with severe ARDS that included 15% of the entire ARDS population and had a mortality of 52% (95% CI, 48%56%). Patients with severe ARDS who did not meet the higher-risk subset criteria included 13% of the entire ARDS population and had a mortality rate of 37% (95% CI, 33%-41%). The difference between the mortality of patients with higher-risk severe ARDS and patients with severe ARDS who did not meet these criteria was statistically significant (P ⬍ .001). Comment Developing and disseminating formal definitions for clinical syndromes in critically ill patients are essential for research and clinical practice. Although previous proposals have relied solely on the consensus process, this is to our knowledge the first attempt in critical care to link an international consensus panel endorsed by professional societies with an empirical evaluation. The draft Berlin Definition classified patients with ARDS into 3 independent categories but relied on ancillary variables (severity of chest radiograph, PEEP ⱖ10 cm H2O, CRS ⱕ40 mL/cm H 2 O, and V̇ E CORR ⱖ10 L/min) in addition to oxygenation to define the severe ARDS group. When the ancillary variables selected by the panel ©2012 American Medical Association. All rights reserved. were subjected to evaluation, these parameters did not identify a group of patients with higher mortality and were excluded from the final Berlin Definition after further consensus discussion. Without this evaluation, a needlessly complex ARDS definition would have been proposed. However, static respiratory system compliance and an understanding of minute ventilation are important variables for clinicians to consider in managing patients with ARDS, even though those variables were not included as part of the definition.32 The Berlin Definition addresses some of the limitations of the AECC definition, including clarification of the exclusion of hydrostatic edema and adding minimum ventilator settings, and provides slight improvement in predictive validity. Our study presents data on the outcomes of patients with ARDS defined according to the Berlin Definition in a large heterogeneous cohort of patients including patients managed with modern approaches to lung protective ventilation. Estimates of the prevalence and clinical outcomes of mild, moderate, and severe ARDS can be assessed from this database for research and health services planning. Acute respiratory distress syndrome is a heterogeneous syndrome with comJAMA, June 20, 2012—Vol 307, No. 23 2531 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME plex pathology and mechanisms. The proposed definition does not resolve this problem. Investigators may choose to design future trials using 1 or more of the ARDS subgroups as a base study population, which may be further refined using criteria specific to the putative mechanism of action of the intervention (eg, IL-6 levels for an anti–IL-6 trial or more stringent hypoxemia criteria for a study on extracorporeal membrane oxygenation). Furthermore, some variables that were excluded from the Berlin Definition because of current feasibility and lack of data on operational characteristics may become more useful in the future. We anticipate that clinical research using our model of definition development will be used to revise the definition in the future. There are limitations to our approach. First, although the Berlin Definition had statistically significantly superior predictive validity for mortality compared with the modified AECC definition, the magnitude of this difference and the absolute values of the AUROC are small and would be clinically unimportant if the Berlin Definition was designed as a clinical prediction tool. However, predictive validity for outcome is only one criterion for evaluating a syndrome definition and the purpose of the Berlin Definition is not a prognostication tool.33 Although the Berlin Definition was developed with a framework including these criteria, we did not empirically evaluate face validity, content validity, reliability, feasibility, or success at identifying patients for clinical trial enrollment. Second, it is possible that our results are not generalizable because of the data sets we studied. This seems unlikely because patients from a broad range of populations, including clinical trials, academic centers, and community patients, were included in the analyses. Third, some variables (eg, CRS and PEEP) were missing in some patients in the data sets we used, either due to the mode of mechanical ventilation that precluded their measurement or the practicalities of population-based research. However, bias due to cohort selection or 2532 JAMA, June 20, 2012—Vol 307, No. 23 missing data seem unlikely because our results were robust to sensitivity analyses that excluded individual cohorts. Fourth, it is possible that the ancillary variables did not identify a higher-risk subset because the number of quadrants on the chest radiograph cannot be assessed reliably, PEEP was not used in a predictable fashion, or CRS and V̇ECORR were not accurately measured. However, if this is true, it is likely also to be true in future studies and in clinical practice because the study database was constructed from clinical trial, academic, and community sites reflecting practice in the real world of clinical research. In addition, we evaluated PEEP and CRS as used by clinicians in practice and not as a test of prespecified ventilator settings that may be betterthanthevariablesevaluatedherein, but may not be practical, particularly in observational cohort studies.5,6 Fifth, because our study was not an exercise in developing a prognostic model for ARDS, we only considered the variables and cutoffs proposed by the consensus panel. We could not compare this definition directly to the AECC definition because the categories of that definition overlap. It is possible that the outcomes as well as the relative proportion of patients within each category of ARDS will change if the underlying epidemiology of the syndrome evolves due to changes in clinical practice or risk factors.34 This is particularly true for the post hoc higherrisk subset reported, for which the cut points were derived from the data sets. Conclusion In conclusion, we developed a consensus draft definition for ARDS with an international panel using a framework that focused on feasibility, reliability, and validity. We tested that definition using empirical data on clinical outcome, radiographic findings, and physiological measures from 2 large databases constructed from 7 contributing sources to assess the predictive value of ancillary variables, refine the draft definition, and compare the predictive validity of the definition to the existing AECC definition. This approach for developing the Berlin Definition for ARDS may serve as an example for linking consensus definition activities with empirical research to better inform clinical care, research, and health services planning. Published Online: May 21, 2012. doi:10.1001 /jama.2012.5669 Authors/Writing Committee: V. Marco Ranieri, MD (Department of Anesthesia and Intensive Care Medicine, University of Turin, Turin, Italy); Gordon D. Rubenfeld, MD, MSc (Program in Trauma, Emergency, and Critical Care, Sunnybrook Health Sciences Center, and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); B. Taylor Thompson, MD (Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston); Niall D. Ferguson, MD, MSc (Department of Medicine, University Health Network and Mount Sinai Hospital, and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); Ellen Caldwell, MS (Division of Pulmonary and Critical Care Medicine, University of Washington, Seattle); Eddy Fan, MD (Department of Medicine, University Health Network and Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada); Luigi Camporota, MD (Department of Critical Care, Guy’s and St. Thomas’ NHS Foundation Trust, King’s Health Partners, London, England); and Arthur S. Slutsky, MD (Keenan Research Center of the Li Ka Shing Knowledge Institute of St. Michael’s Hospital; Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada). Author Contributions: Dr Rubenfeld and Ms Caldwell had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ranieri, Rubenfeld, Thompson, Ferguson, Caldwell, Camporota. Acquisition of data: Ranieri, Rubenfeld, Thompson. Analysis and interpretation of data: Rubenfeld, Thompson, Ferguson, Caldwell, Fan, Slutsky. Drafting of the manuscript: Rubenfeld, Ferguson, Caldwell, Slutsky. Critical revision of the manuscript for important intellectual content: Ranieri, Rubenfeld, Thompson, Ferguson, Caldwell, Fan, Camporota, Slutsky. Statistical analysis: Rubenfeld, Caldwell, Slutsky. Obtained funding: Ranieri. Administrative, technical, or material support: Rubenfeld, Thompson, Fan, Camporota. Study supervision: Ranieri, Rubenfeld, Thompson, Slutsky. Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Ranieri reported receiving consulting fees or honoraria from Maquet and Hemodec and board membership from Faron. Dr Rubenfeld reported receiving consulting fees or honoraria from Ikaria, Faron, and Cerus. Dr Thompson reported receiving support for travel from European Society of Intensive Care Medicine; being an advisory board member of Hemodec and AstraZeneca; receiving consultancy fees from US Biotest, Sirius Genetics, sanofi-aventis, Immunetrics, Abbott, and Eli Lilly; and receiving grants from the National Heart, Lung, and Blood Institute. Dr Slutsky reported receiving support for travel expenses from European Society of Intensive Care Medicine; board membership from Ikaria; receiving consultancy fees from GlaxoSmithKline and Tarix; having stock/stock options with Apeiron and Tarix; and sitting on advisory boards for Maquet Medical and NovaLung and steering committees for HemoDec and Eli Lilly. No other authors reported any financial disclosures. Members of the ARDS Definition Task Force: V. Marco ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME Ranieri, MD (Department of Anesthesia and Intensive Care Medicine, University of Turin, Turin, Italy); Gordon D. Rubenfeld, MD, MSc (Program in Trauma, Emergency, and Critical Care, Sunnybrook Health Sciences Center and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); B. Taylor Thompson, MD (Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston); Niall D. Ferguson, MD, MSc (Department of Medicine, University Health Network and Mount Sinai Hospital, and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); Ellen Caldwell, MS (Division of Pulmonary and Critical Care Medicine, University of Washington, Seattle); Eddy Fan, MD (Department of Medicine, University Health Network and Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada); Luigi Camporota, MD (Department of Critical Care, Guy’s and St. Thomas’ NHS Foundation Trust, King’s Health Partners, London, England); and Arthur S. Slutsky, MD (Keenan Research Center of the Li Ka Shing Knowledge Institute of St. Michael’s Hospital; Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); Massimo Antonelli, MD (Dipartimento di Anestesia e Rianimazione, Universita Cattolica del Sacro Cuore, Rome, Italy); Antonio Anzueto, MD (Department of Pulmonary/Critical Care, University of Texas Health Sciences Center, San Antonio); Richard Beale, MBBS (Department of Critical Care, Guy’s and St. Thomas’ NHS Foundation Trust, King’s Health Partners, London, England); Laurent Brochard, MD (Medical-Surgical Intensive Care Unit, Hopitaux Universitaires de Geneve, Geneva, Switzerland); Roy Brower, MD (Division of Pulmonary and Critical Care Medicine, Johns Hopkins University, Baltimore, Maryland); Andrés Esteban, MD, PhD (Servicio de Cuidados Intensivos, Hospital Universitario de Getafe, CIBERES, Madrid, Spain); Luciano Gattinoni, MD (Istituto di Anestesiologia e Rianimazione, Universita degli Studi di Milano, Milan, Italy); Andrew Rhodes, MD (Department of Intensive Care Medicine, St. George’s Healthcare NHS Trust, London, England); Jean-Louis Vincent, MD (Department of Intensive Care, Erasme University, Brussels, Belgium); Provided data for the empiric evaluation of the definition but were not part of the consensus development: Andrew Bersten, MD (Department of Critical Care Medicine, Flinders University, Adelaide, South Australia); Dale Needham, MD, PhD (Outcomes After Critical Illness and Surgery Group [OACIS], Division of Pulmonary and Critical Care Medicine and Department of Physical Medicine and Rehabilitation, Johns Hopkins University, Baltimore, Maryland); and Antonio Pesenti, MD (Department of Anesthesia and Critical Care, Ospedale San Gerardo, Monza, Italy; and Department of Experimental Medicine, University of Milano Bicocca, Milan, Italy). Funding/Support: This work was supported by the Eu- ropean Society of Intensive Care Medicine and grant R01HL067939 from the National Institutes of Health (Dr Rubenfeld). Dr Ferguson is supported by a Canadian Institutes of Health Research New Investigator Award (Ottawa, Canada). Role of the Sponsors: The European Society of Intensive Care Medicine, the National Institutes of Health, the Canadian Institutes of Health Research, and the endorsing professional societies had no role in the design and conduct of the study, in the collection, management, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript. Online-Only Material: The eMethods, eReferences, eTables 1 through 4, and eFigures 1 through 4 are available at http://www.jama.com. Additional Contributions: Salvatore Maggiore, MD, PhD (Department of Anesthesiology and Intensive Care, Agostino Gemelli University Hospital, Università Cattolica del Sacro Cuore, Rome, Italy), and Anders Larsson, MD, PhD (Department of Surgical Sciences, Anesthesiology and Critical Care Medicine, Uppsala University, Uppsala, Sweden), attended the roundtable as representatives of the European Society of Intensive Care Medicine. Drs Maggiore and Larsson received no compensation for their roles. Karen Pickett, MB BCh (Department of Intensive Care, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium), provided technical assistance. Dr Pickett received compensation for her role in the conference. terpretation of pulmonary artery catheter pressure tracings. Chest. 1991;100(6):1647-1654. 13. ESICM Congress Highlights. http://www.esicm .org/07-congresses/0A-annual-congress/webTv .asp. Accessed April 23, 2011. 14. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med. 2005;353(16):1685-1693. 15. Bersten AD, Edibam C, Hunt T, Moran J; Australian and New Zealand Intensive Care Society Clinical Trials Group. Incidence and mortality of acute lung injury and the acute respiratory distress syndrome in three Australian States. Am J Respir Crit Care Med. 2002; 165(4):443-448. 16. Needham DM, Dennison CR, Dowdy DW, et al. Study protocol: The Improving Care of Acute Lung Injury Patients (ICAP) study. Crit Care. 2006;10(1):R9. 17. Britos M, Smoot E, Liu KD, Thompson BT, Checkley W, Brower RG; National Institutes of Health Acute Respiratory Distress Syndrome Network Investigators. The value of positive end-expiratory pressure and Fio criteria in the definition of the acute respiratory distress syndrome. Crit Care Med. 2011;39(9):2025-2030. 18. Bellani G, Guerra L, Musch G, et al. Lung regional metabolic activity and gas volume changes induced by tidal ventilation in patients with acute lung injury. Am J Respir Crit Care Med. 2011;183(9): 1193-1199. 19. Terragni PP, Rosboch G, Tealdi A, et al. Tidal hyperinflation during low tidal volume ventilation in acute respiratory distress syndrome. Am J Respir Crit Care Med. 2007;175(2):160-166. 20. Terragni PP, Del Sorbo L, Mascia L, et al. Tidal volume lower than 6 mL/kg enhances lung protection: role of extracorporeal carbon dioxide removal. Anesthesiology. 2009;111(4):826-835. 21. Spragg RG, Bernard GR, Checkley W, et al. Beyond mortality: future clinical research in acute lung injury. Am J Respir Crit Care Med. 2010;181(10): 1121-1127. 22. Wexler HR, Lok P. A simple formula for adjusting arterial carbon dioxide tension. Can Anaesth Soc J. 1981;28(4):370-372. 23. Gattinoni L, Caironi P, Pelosi P, Goodman LR. What has computed tomography taught us about the acute respiratory distress syndrome? Am J Respir Crit Care Med. 2001;164(9):1701-1711. 24. 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(17):1281-1286. 25. Cooke CR, Kahn JM, Caldwell E, et al. Predictors of hospital mortality in a population-based cohort of patients with acute lung injury. Crit Care Med. 2008; 36(5):1412-1420. 26. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837845. 27. Luhr OR, Antonsen K, Karlsson M, et al; The ARF Study Group. Incidence and mortality after acute respiratory failure and acute respiratory distress syndrome in Sweden, Denmark, and Iceland. Am J Respir Crit Care Med. 1999;159(6):1849-1861. 28. Brun-Buisson C, Minelli C, Bertolini G, et al; ALIVE Study Group. Epidemiology and outcome of acute lung injury in European intensive care units: results from the ALIVE study. Intensive Care Med. 2004;30 (1):51-61. 29. Katzenstein AL, Bloor CM, Leibow AA. Diffuse alveolar damage—the role of oxygen, shock, and related factors: a review. Am J Pathol. 1976;85(1): 209-228. 30. Hudson LD, Milberg JA, Anardi D, Maunder RJ. Clinical risks for development of the acute respiratory distress syndrome. Am J Respir Crit Care Med. 1995;151(2 pt 1):293-301. 31. Phua J, Badia JR, Adhikari NKJ, et al. Has mortality from acute respiratory distress syndrome decreased over time? a systematic review. Am J Respir Crit Care Med. 2009;179(3):220-227. 32. Hager DN, Krishnan JA, Hayden DL, Brower RG; ARDS Clinical Trials Network. Tidal volume reduction in patients with acute lung injury when plateau pressures are not high. Am J Respir Crit Care Med. 2005;172(10):1241-1245. 33. Rubenfeld GD. Epidemiology of acute lung injury. Crit Care Med. 2003;31(4)(suppl):S276-S284. 34. Li G, Malinchoc M, Cartin-Ceba R, et al. Eightyear trend of acute respiratory distress syndrome: a population-based study in Olmsted County, Minnesota. Am J Respir Crit Care Med. 2011; 183(1):59-66. REFERENCES 1. Streiner D, Norman G. Health Measurement Scales. 4th ed. New York, NY: Oxford University Press; 2008. 2. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute respiratory distress in adults. Lancet. 1967; 2(7511):319-323. 3. Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994; 149(3 pt 1):818-824. 4. Phua J, Stewart TE, Ferguson ND. Acute respiratory distress syndrome 40 years later: time to revisit its definition. Crit Care Med. 2008;36(10):29122921. 5. Villar J, Pérez-Méndez L, López J, et al; HELP Network. An early PEEP/FIO2 trial identifies different degrees of lung injury in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med. 2007;176(8):795-804. 6. Ferguson ND, Kacmarek RM, Chiche J-D, et al. Screening of ARDS patients using standardized ventilator settings: influence on enrollment in a clinical trial. Intensive Care Med. 2004;30(6):1111-1116. 7. Gowda MS, Klocke RA. Variability of indices of hypoxemia in adult respiratory distress syndrome. Crit Care Med. 1997;25(1):41-45. 8. Rubenfeld GD, Caldwell E, Granton JT, Hudson LD, Matthay MA. Interobserver variability in applying a radiographic definition for ARDS. Chest. 1999; 116(5):1347-1353. 9. 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(1):85-90. 10. Ferguson ND, Meade MO, Hallett DC, Stewart TE. High values of the pulmonary artery wedge pressure in patients with acute lung injury and acute respiratory distress syndrome. Intensive Care Med. 2002; 28(8):1073-1077. 11. Wheeler AP, Bernard GR, Thompson BT, et al; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network. Pulmonary-artery versus central venous catheter to guide treatment of acute lung injury. N Engl J Med. 2006;354(21):2213-2224. 12. Komadina KH, Schenk DA, LaVeau P, Duncan CA, Chambers SL. Interobserver variability in the in- ©2012 American Medical Association. All rights reserved. JAMA, June 20, 2012—Vol 307, No. 23 2533 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013 ORIGINAL ARTICLE Surgeon-Reported Conflict With Intensivists About Postoperative Goals of Care Terrah J. Paul Olson, MD; Karen J. Brasel, MD, MPH; Andrew J. Redmann, BA, BS; G. Caleb Alexander, MD, MS; Margaret L. Schwarze, MD, MPP Objective: To examine surgeons’ experiences of conflict with intensivists and nurses about goals of care for their postoperative patients. Design: Cross-sectional incentivized US mail-based survey. Setting: Private and academic surgical practices. Participants: A total of 2100 vascular, neurologic, and cardiothoracic surgeons. Main Outcome Measures: Surgeon-reported rates of conflict with intensivists and nurses about goals of care for patients with poor postsurgical outcomes. Results: The adjusted response rate was 55.6%. Forty- three percent of surgeons reported sometimes or always experiencing conflict about postoperative goals of care with intensivists, and 43% reported conflict with nurses. Younger surgeons reported higher rates of conflict than older surgeons with both intensivists (57% vs 32%; P = .001) and nurses (48% vs 33%; P = .001). Surgeons practicing in closed intensive care units reported more frequent conflict than those practicing in open intensive care units (60% vs 41%; P=.005). On multivariate analysis, the odds of reporting conflict with intensivists were 2.5 times higher for surgeons with fewer years of experience compared with their older colleagues (odds ratio, 2.5; 95% CI, 1.6-3.8) and 70% higher for reporting conflict with nurses (odds ratio, 1.7; 95% CI, 1.12.6). The odds of reporting conflict with intensivists about goals of postoperative care were 40% lower for surgeons who primarily managed their intensive care unit patients than for those who worked in a closed unit (odds ratio, 0.60; 95% CI, 0.40-0.96). Conclusions: Surgeons regularly experience conflict with critical care clinicians about goals of care for patients with poor postoperative outcomes. Higher rates of conflict are associated with less experience and working in a closed intensive care unit. JAMA Surg. 2013;148(1):29-35 C Author Affiliations are listed at the end of this article. ONFLICT IN THE INTEN sive care unit (ICU) is a significant public health problem, as more than 70% of ICU clinicians report experiencing conflict weekly.1,2 The combination of caring for acutely ill patients, end-of-life decision making, and coordination of large multidisciplinary teams can lead to frustration, communication breakdown, and discord among members of a health care team. The epidemiology of ICU conflict is well described.1 This conflict has been associated with lower-quality patient care,3,4 higher rates of medical error,5 higher levels of staff burnout,6,7 and greater direct and indirect costs of care.2,8 Intensive care unit conflict can occur between the health care team and patients’ families; among members of the intensive care team (intrateam conflict); and between different JAMA SURG/ VOL 148 (NO. 1), JAN 2013 29 groups of clinicians caring for the same patient (interteam conflict), most notably between surgeons and intensivists.1-4,9 See Invited Critique at end of article Two primary contributors to ICU conflict are particularly relevant to surgeons: patient-doctor relationships formed prior to the ICU admission and discussions of end-of-life care.3,4,10 Others have shown that surgeons have a strong sense of personal responsibility for patient outcomes that may influence surgeons’ interactions with critical care clinicians as well as discussions about end-of-life care.3,10-15 Surgeons are often reluctant to switch goals of care from cure to comfort, particularly in the postoperative period. 3,10,14 Although these sources of conflict have been WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 Author Aff Departmen University Hospital an Olson); Dep Division of (Dr Schwar Wisconsin and Public Redmann), Departmen College of W Milwaukee Wisconsin; Medicine an for Clinical University Departmen Practice, Un Chicago Sc (Dr Alexan Illinois; and Epidemiolo Health and Medicine, J University, (Dr Alexan well described by intensivists, it is unknown whether surgeons appreciate these conflicts. We examined whether surgeons recognized and reported conflict with intensivists and nurses about goals of care for their patients, specifically in the setting of a poor postoperative outcome. In addition, we explored how often surgeons reported conflict with ICU clinicians as well as surgeon factors associated with such conflict. METHODS PARTICIPANTS We selected a random sample of neurologic, vascular, and cardiothoracic surgeons. We chose these specialties because they were likely to perform high-risk operations and have patients who frequently require intensive care postoperatively. We excluded other surgeons who routinely care for patients in the ICU to avoid specific confounding issues. Trauma surgeons were excluded owing to the routine performance of emergency surgery, transplant surgeons were excluded owing to their concern for resource allocation, and surgical oncologists were excluded owing to the heterogeneous nature of surgical oncology whereby surgical oncologists who specialize in breast or endocrine surgery would be unlikely to care for patients in the ICU. We randomly selected participants from the membership lists of the American Association for Neurological Surgery Cerebrovascular Division, regional vascular surgery societies (Midwestern, New England, Eastern, and Western societies), and the Society for Thoracic Surgery. We mailed a total of 2100 surveys, 700 to each specialty, via the US Postal Service. The survey packet also contained a stamped addressed return envelope and a laser-pointer pen valued at $2.85 as an incentive to complete the survey. In March 2010, we mailed the first round of surveys. We sent a second mailing (including a return envelope but no pen incentive) to nonrespondents. Finally, because of a high proportion of nonrespondents from the neurosurgical group owing to incorrect addresses, we mailed a third survey with an additional laserpointer pen as well as a letter of support from a key neurosurgical opinion leader. Prior to this third mailing, we validated these addresses through an Internet search. We could not verify the addresses of 180 members of the original neurosurgical cohort; therefore, we replaced these members with 180 randomly selected new participants. The survey was completed in August 2010. This study was approved as exempt, including a waiver of written consent, by the institutional review boards of the University of Wisconsin and University of Chicago. SURVEY We designed a survey to assess surgeon attitudes and practices regarding advanced directives, withdrawal of lifesupporting therapies, and decision making surrounding highrisk operations. First, we performed a qualitative study using semistructured interviews of surgeons and other physicians involved in perioperative care to examine surgeons’ attitudes and routines surrounding the use of advanced directives and withdrawal of life-supporting therapies postoperatively.15,16 In addition to work by others, this study identified conflict between surgeons and ICU clinicians that stems from decisions about the use or withdrawal of life-supporting therapies in postoperative patients.3,4,10,14,15 Using these results, we designed survey questions to explore the validity and generalizability of our qualitative findings to a larger group of surgeons. To assess the face validity of the survey questions, we performed cognitive interviews with 10 surgeons. To avoid interviewing actual survey recipients, the surgeons who were interviewed performed high-risk operations routinely (transplant, surgical oncology, and trauma) but were not members of the subspecialties included in the study target population. To inform questionnaire design, we asked respondents to think aloud as they read survey items and explain their interpretation of each item to ensure that the intended focus of the question was clearly understood.17,18 We incorporated each respondent’s input in a stepwise fashion. Our final survey included items about the surgeons’ experiences with conflict during the care of patients who had poor postoperative outcomes. Using a 4-point Likert scale of never, rarely, sometimes, or always, respondents were asked to rate how often they experienced conflict about the goals of care for their postoperative patients with poor outcomes with critical care physicians, nursing staff, and other groups. In addition, we gathered data on surgeon sex, specialty, number of years in practice, practice setting (academic vs private practice), number of high-risk procedures performed per month (defined as procedures having ⬎1% operative mortality or significant morbidity), percentage of patients routinely requiring ICU care postoperatively, and the administrative model for the main ICU where the surgeon practices (open, closed, mixed, or other). We defined a closed ICU as a unit where an intensive care physician was primarily responsible for all patients and an open ICU as a unit where the operative surgeon was primarily responsible for his or her patients. A mixed ICU combines elements of both open and closed units. STATISTICAL ANALYSIS After removing all surveys that were returned to sender and surveys completed by ineligible respondents (junior residents and nonphysicians), an Internet search was used to estimate the percentage of nonrespondents owing to inaccurate contact information. We used a sample of 60 respondents (20 from each specialty) and 60 nonrespondents to estimate the percentage of nonrespondents owing to faulty contact information. Using the American Association for Public Opinion Research guidelines, we calculated the adjusted response rate with the following formula: response rate=R/([R]⫹e[T⫺R⫺NE]), where R = eligible respondents, e = proportion of nonrespondents estimated to be ineligible, T=total number of surveys, and NE=ineligible respondents (including return to sender).19 As a surrogate marker for nonresponse bias, we looked for evidence of forward response wave bias. To do this, we calculated response time for each survey and identified clusters of early and late respondents. Rates of surgeon-reported conflict were compared between the early and late respondents. We defined our primary outcome as surgeon-reported conflict with critical care physicians and nursing staff. We collapsed the response frame for conflict into dichotomous variables (never/rarely vs sometimes/always). On sensitivity analysis, we found no difference in our outcomes between the collapsed responses and the 4-point response frame. We performed bivariate analyses using the 2 test. We next developed a multivariate logistic regression model including basic demographic characteristics, such as sex and surgical subspecialty (vascular, cardiothoracic, or neurologic surgery); variables that were significant on bivariate analysis at P = .10 included years in clinical practice, practice type (private, academic, or private with academic affiliation), and ICU administrative model (closed, open, or mixed). Of note, we did not find that the ICU administrative model was a statistically significant variable for conflict with nursing staff on bivariate analysis (P=.85), but we included it in our multivariate model of conflict with JAMA SURG/ VOL 148 (NO. 1), JAN 2013 30 WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 Table 1. Characteristics of 912 Survey Respondents a Characteristic Table 2. Rates of Surgeon-Reported Conflict and Challenges No. (%) Specialty Vascular Neurologic Cardiothoracic Time in practice, y 0-10 11-20 21-30 ⱖ31 Sex Male Female Practice type Private Academic Private with academic affiliation Other ICU administrative model Closed Open Mixed Other No. of high-risk procedures per mo 0 1-5 6-10 ⱖ11 Question 327 (36) 273 (30) 312 (34) 196 (23) 215 (25) 233 (27) 223 (26) 850 (94) 50 (6) 343 (38) 323 (36) 167 (19) 61 (7) 96 (11) 277 (32) 500 (57) 4 (0.5) 34 (4) 311 (37) 256 (30) 239 (28) Abbreviation: ICU, intensive care unit. a Numbers do not sum to 912 because not all respondents answered all questions. nursing staff because this was a variable of considerable interest for our explanatory model of ICU conflict. All statistical analyses were performed using SAS version 9.1 (SAS Institute Inc). RESULTS Answer, % At times, conflicts may arise among different parties involved in the care of a patient who has a poor postsurgical outcome. How frequently, if ever, do you experience conflict with each of these groups about the goals of care for your postoperative patients? Critical care physicians Nursing staff Primary care physicians Ethics consultants Family members of the patient Surgical colleagues How challenging, if at all, do you find each of the following aspects of surgical practice? Communicating with family and/or patient about poor outcomes Addressing patient fears regarding morbidity and mortality Managing personal discomfort about poor outcomes Managing clinical aspects of poor outcomes Emphasis on outcome measures such as physician profiling Sometimes/ always 43 43 23 16 60 18 Somewhat/very challenging 64 42 73 62 50 cians, and a similar percentage (43%) reported sometimes or always experiencing conflict with nursing staff about the goals of postoperative care (Table 2). Although 50% of our respondents felt that the emphasis on outcome measures such as physician profiling was a challenging aspect of surgical practice, a greater number of surgeons identified the difficulty of managing clinical aspects related to poor outcomes or communicating with patients and their families as a somewhat or very challenging aspect of surgical practice. Moreover, 73% of respondents noted that managing their own discomfort about poor outcomes was a significant challenge (Table 2). PARTICIPANTS We received completed surveys from 912 respondents, and 203 were returned to sender. The adjusted response rate as calculated by the American Association for Public Opinion Research guidelines was 55.6%. There was no significant difference in the rates of reported conflict between early and late respondents, suggesting no evidence for forward response wave bias, our surrogate measure for nonresponse bias. Survey respondents were evenly distributed with respect to years in practice (Table 1). There were approximately equal proportions of respondents who practiced in an academic vs a private setting. Most respondents (57%) described the administrative model for their main ICU as mixed. About one-third of respondents practiced primarily with an open ICU model (32%), whereas a smaller proportion worked in hospitals with a closed ICU model (11%). SURGEON-REPORTED CONFLICT Forty-three percent of surgeons reported sometimes or always experiencing conflict with critical care physi- CONFLICT WITH CRITICAL CARE PHYSICIANS On bivariate analysis (Table 3), surgeons with less than 10 years of experience were significantly more likely than surgeons with more than 30 years of experience to report conflict with critical care physicians (57% vs 32%; P ⬍ .001). In addition, surgeons in academic practices reported significantly more conflict compared with those in private practice (52% vs 36%; P ⬍ .001). Forty-one percent of surgeons practicing in an open ICU and 42% practicing in a mixed ICU reported conflict with critical care doctors, whereas 60% of surgeons practicing within a closed ICU model reported conflict with critical care doctors (P = .005). On multivariate logistic regression, a strong and statistically significant association persisted between surgeon experience and surgeon-reported conflict with critical care physicians. The odds of surgeons reporting conflict with critical care physicians were 2.5 times higher for surgeons with less than 10 years of experience than for those with more than 30 years of experience (odds ratio, 2.5; JAMA SURG/ VOL 148 (NO. 1), JAN 2013 31 WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 Table 3. Surgeon Characteristics Associated With Conflict With Intensivists a Characteristic Sex Female Male Specialty Neurologic Vascular Cardiothoracic Time in practice, y ⱖ31 21-30 11-20 0-10 Practice type Private Academic Private with academic affiliation Other ICU administrative model Closed Open Mixed Other Total No. Reported Conflict With Intensivists, % Bivariate P Value Multivariate Odds Ratio (95% CI) 50 838 54 43 .12 1.0 [Reference] 0.8 (0.4-1.6) 269 323 308 38 47 45 .08 1.0 [Reference] 1.6 (1.1-2.3) 1.5 (1.0-2.1) 217 231 213 196 32 40 45 57 ⬍.001 1.0 [Reference] 1.5 (1.0-2.2) 1.8 (1.2-2.8) 2.5 (1.6-3.8) 340 318 163 61 36 52 44 41 ⬍.001 1.0 [Reference] 1.4 (1.0-2.0) 1.3 (0.8-1.9) 1.1 (0.6-2.0) 96 272 496 4 60 41 42 25 .005 1.0 [Reference] 0.6 (0.4-1.0) 0.60 (0.40-0.96) 0.20 (0.02-2.30) Abbreviation: ICU, intensive care unit. a Number does not sum to 912 because not all respondents answered all questions. Table 4. Surgeon Characteristics Associated With Conflict With Nursing Staff a Characteristic Sex Female Male Specialty Neurologic Vascular Cardiothoracic Time in practice, y ⱖ31 21-30 11-20 0-10 Practice type Private Academic Private with academic affiliation Other ICU administrative model Closed Open Mixed Other Total No. Reported Conflict With Nurses, % Bivariate P Value Multivariate Odds Ratio (95% CI) 50 838 54 42 .11 1.0 [Reference] 0.7 (0.4-1.3) 269 323 308 39 47 43 .17 1.0 [Reference] 1.6 (1.1-2.2) 1.3 (0.9-1.9) 217 232 213 195 33 42 50 48 .001 1.0 [Reference] 1.4 (0.9-2.1) 1.8 (1.2-2.8) 1.7 (1.1-2.6) 341 317 163 61 39 50 43 34 .02 1.0 [Reference] 1.4 (1.0-2.0) 1.1 (0.7-1.6) 0.8 (0.5-1.5) 95 273 496 4 45 44 43 25 .85 1.0 [Reference] 1.3 (0.8-2.2) 1.1 (0.7-1.7) 0.40 (0.03-3.90) Abbreviation: ICU, intensive care unit. a Numbers do not sum to 912 because not all respondents answered all questions. 95% CI, 1.6-3.8). In addition, the odds of surgeons reporting conflict with critical care physicians was 40% lower for surgeons practicing within a mixed or open model of ICU administration compared with those practicing in a closed ICU (mixed: odds ratio, 0.60; 95% CI, 0.40-0.96; open: odds ratio, 0.6; 95% CI, 0.4-1.0). CONFLICT WITH NURSING STAFF On bivariate analysis (Table 4), an association between surgeon experience and reported conflict with nurses was demonstrated. Surgeons with less than 10 years of experience were significantly more likely than JAMA SURG/ VOL 148 (NO. 1), JAN 2013 32 WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 surgeons who had been in practice for more than 30 years to report conflict with nursing staff regarding goals of care for postoperative patients who had poor postsurgical outcomes (48% vs 33%; P = .001). Surgeons in academic settings reported more conflict with ICU nurses than their counterparts in private practice (50% vs 39%; P = .02). In contrast to the experience of conflict with critical care physicians, the model of ICU administration was not associated with significantly higher rates of conflict with nurses (open: 44%, mixed: 43%, closed: 45%; P = .85). On multivariate analysis, a statistically significant association was found between surgeons’ experiences and reported rates of conflict with nursing staff. The odds of surgeons reporting conflict with nursing staff was 70% greater in surgeons with less than 10 years of experience compared with surgeons with more than 30 years in practice (odds ratio, 1.7; 95% CI, 1.1-2.6). The administrative model of ICU care was not associated with increased rates of conflict reported by surgeons with nurses. COMMENT More than 40% of surgeons who routinely perform highrisk operations reported conflict with critical care physicians and nurses regarding the goals of care for their patients with poor postoperative outcomes. Surgeons who reported higher rates of conflict had fewer years in practice and worked in an academic setting. Additionally, surgeons who practiced in a closed ICU reported higher rates of conflict about the goals of care with critical care physicians, but not with critical care nurses. These findings have important implications for surgeons, critical care clinicians, and their patients. For surgeons, these findings suggest that learning how to manage personal discomfort in the setting of a poor postoperative outcome may take time and experience. Prior work by our group and others has shown that surgeons have a particularly difficult time switching goals of care from cure to comfort postoperatively, especially given the perception of a direct link between the surgeon’s performance and the patient’s outcome.11,13,14,20,21 When critical care clinicians who have had no direct role in the patient’s operation suggest a change in the goals of care from survival to comfort, conflict may ensue. Surgeons with more experience may be more accepting of the limits of surgical therapy and may have developed more robust coping strategies for the inevitable unwanted outcome. For critical care clinicians, it is important to note that the ICU administrative model can contribute to conflict with surgeons. Surgeons typically have a relationship with their patient that begins preoperatively and extends throughout the patient’s recovery. Because of this longitudinal relationship, surgeons see themselves as the primary decision makers for their postoperative patients. The structure of a closed ICU likely poses a barrier to the continuity of the surgeon-patient relationship and contributes to conflict when the surgeon is replaced as the primary decision maker for his or her patient by an intensivist.3,13-15,21 As Cassell and colleagues14 noted, intensivists working in a closed ICU are charged with allocation of scarce resources for the good of the entire unit, while surgeons focus primarily on their covenantal responsibilities to individual patients. The administrative model of a closed ICU thus promotes conflict with its juxtaposition of clinicians with competing viewpoints. Patients and their families are likely affected by the conflict that surgeons report with critical care clinicians. This can impact the overall experience of patients and their families in the ICU, with perceived conflict leading to decreased satisfaction with care and increased stress for families.1-4,9,22,23 Changing the goals of care postoperatively may trigger conflict when those caring for the patient try to determine what is most in line with the patient’s preferences. Our findings about surgeonreported conflict with intensivists underscore how important it is for patients to discuss their goals and values with their surgeon and surrogate decision makers preoperatively to inform postoperative decisions in the event of an undesired outcome.15,16,24,25 Patients and their surrogates can be referred to other available resources for conflict adjudication to help guide treatment decisions such as a second opinion from a panel of senior clinicians or the hospital ethics committee.3,9,13,21,23,26,27 Our study has several limitations. Survey-based research can be subject to nonresponse bias, where nonrespondents differ systematically from respondents. However, our response rate was robust, and there was no evidence of forward response wave bias, making the possibility of nonresponse bias low. Because conflict is not considered acceptable behavior, questions about conflict are subject to social desirability bias. As such, the true rate of conflict might be higher than our respondents reported. This survey also targeted 3 surgical subspecialties (vascular, cardiothoracic, and neurologic surgery) to ensure the respondents routinely care for critically ill patients. It is unclear whether our results are generalizable to other surgical subspecialties. Our survey did not specifically define the parameters of a poor postoperative outcome, thus we cannot specify a clinical threshold that may prompt such conflicts. Also, we did not distinguish between closed ICUs run by surgeons vs closed ICUs managed by nonsurgeons, although this may prove to be an important determinant in rates of surgeon-reported conflict.21,26,28 Finally, although we noted a high rate of surgeon-reported conflict between surgeons and patients’ family members, we chose not to focus on this finding here. The primary focus of our survey was physician practices and, as such, we did not generate enough information for a complete discussion of this important topic. Surgeons frequently experience conflict with critical care physicians and nurses about the goals of care for their postoperative patients with poor outcomes. Higher rates of conflict are reported by surgeons with fewer years of experience and those working in institutions with a closed model of ICU administration. This conflict is a significant public health problem that diminishes quality of care for critically ill patients and their families. Given the myriad challenges inherent in delivering the highest qual- JAMA SURG/ VOL 148 (NO. 1), JAN 2013 33 WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 ity of care in these settings, clinicians from all backgrounds should focus on eliminating these interteam conflicts to allow energies to be spent more productively on other clinical issues affecting safety and quality. Interventions directed at the individual level as well as the system level will be important to mitigate conflict to provide better care for our critically ill postoperative patients. tistical analyses as well as Ann Chodara at the University of Wisconsin School of Medicine and Public Health for her assistance with data collection and management. Accepted for Publication: April 30, 2012. Author Affiliations: Department of Surgery, University of Wisconsin Hospital and Clinics (Dr Paul Olson); Department of Surgery, Division of Vascular Surgery (Dr Schwarze), University of Wisconsin School of Medicine and Public Health (Mr Redmann), Madison; Department of Surgery, Medical College of Wisconsin, Milwaukee (Dr Brasel), Wisconsin; Department of Medicine and MacLean Center for Clinical Medical Ethics, University of Chicago; Department of Pharmacy Practice, University of Illinois at Chicago School of Pharmacy (Dr Alexander), Chicago, Illinois; and Departments of Epidemiology, School of Public Health and Medicine, School of Medicine, Johns Hopkins University, Baltimore, Maryland (Dr Alexander). Correspondence: Margaret L. Schwarze, MD, MPP, Department of Surgery, Division of Vascular Surgery, University of Wisconsin, G5/315 CSC, 600 Highland Ave, Madison, WI 53792 ([email protected]). Author Contributions: Dr Schwarze had full access to all data in this study and takes responsibility for the integrity of the data and accuracy of the data analysis. Study concept and design: Brasel, Alexander, and Schwarze. Acquisition of data: Redmann and Schwarze. Analysis and interpretation of data: Paul Olson, Brasel, Redmann, Alexander, and Schwarze. Drafting of the manuscript: Paul Olson and Schwarze. Critical revision of the manuscript for important intellectual content: Brasel, Redmann, Alexander, and Schwarze. Statistical analysis: Paul Olson, Brasel, Redmann. Obtained funding: Schwarze. Administrative, technical, and material support: Schwarze. Study supervision: Alexander and Schwarze. Conflict of Interest Disclosures: None reported. Funding/Support: Dr Paul Olson’s work is supported by grant T32 CA090217 from the National Institutes of Health. Mr Redmann’s work was supported by a Shapiro Summer Research Award from the University of Wisconsin School of Medicine and Public Health. Dr Alexander’s work is supported by grants K08 HS15699 and R01 HS0189960 from the Agency for Healthcare Research and Quality. Dr Schwarze’s work is supported by a Greenwall Faculty Scholars Award and the Department of Surgery at the University of Wisconsin. Role of the Sponsors: The funding sources had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; and preparation, review, or approval of the manuscript for publication. Previous Presentation: This work was presented in part at the 7th Annual Academic Surgical Congress; February 16, 2012; Las Vegas, Nevada. Additional Contributions: We thank Glen Leverson, PhD, in the Department of Surgery at the University of Wisconsin, Madison, for his invaluable help with sta- 1. Azoulay E, Timsit JF, Sprung CL, et al; Conflicus Study Investigators and for the Ethics Section of the European Society of Intensive Care Medicine. Prevalence and factors of intensive care unit conflicts: the Conflicus Study. Am J Respir Crit Care Med. 2009;180(9):853-860. 2. Fassier T, Azoulay E. Conflicts and communication gaps in the intensive care unit [published online October 7, 2012]. Curr Opin Crit Care. doi:10.1097/MCC .0b013e32834044f0. 3. Danjoux Meth N, Lawless B, Hawryluck L. Conflicts in the ICU: perspectives of administrators and clinicians. Intensive Care Med. 2009;35(12):2068-2077. 4. Studdert DM, Mello MM, Burns JP, et al. Conflict in the care of patients with prolonged stay in the ICU: types, sources, and predictors. Intensive Care Med. 2003; 29(9):1489-1497. 5. Baldwin DC Jr, Daugherty SR. Interprofessional conflict and medical errors: results of a national multi-specialty survey of hospital residents in the US. J Interprof Care. 2008;22(6):573-586. 6. Embriaco N, Azoulay E, Barrau K, et al. High level of burnout in intensivists: prevalence and associated factors. Am J Respir Crit Care Med. 2007;175(7):686692. 7. Poncet MC, Toullic P, Papazian L, et al. Burnout syndrome in critical care nursing staff. Am J Respir Crit Care Med. 2007;175(7):698-704. 8. Brinkert R. A literature review of conflict communication causes, costs, benefits and interventions in nursing. J Nurs Manag. 2010;18(2):145-156. 9. Breen CM, Abernethy AP, Abbott KH, Tulsky JA. Conflict associated with decisions to limit life-sustaining treatment in intensive care units. J Gen Intern Med. 2001;16(5):283-289. 10. Larochelle MR, Rodriguez KL, Arnold RM, Barnato AE. Hospital staff attributions of the causes of physician variation in end-of-life treatment intensity. Palliat Med. 2009;23(5):460-470. 11. Bosk CL. Forgive and Remember: Managing Medical Failure. 2nd ed. Chicago, IL: The University of Chicago Press; 1979. 12. Katz P. The Scalpel’s Edge: The Culture of Surgeons. Needham Heights, MA: Allyn and Bacon; 1999. 13. Buchman TG, Cassell J, Ray SE, Wax ML. Who should manage the dying patient? rescue, shame, and the surgical ICU dilemma. J Am Coll Surg. 2002; 194(5):665-673. 14. Cassell J, Buchman TG, Streat S, Stewart RM. Surgeons, intensivists, and the covenant of care: administrative models and values affecting care at the end of life: updated. Crit Care Med. 2003;31(5):1551-1557, discussion 1557-1559. 15. Schwarze ML, Bradley CT, Brasel KJ. Surgical “buy-in”: the contractual relationship between surgeons and patients that influences decisions regarding lifesupporting therapy. Crit Care Med. 2010;38(3):843-848. 16. Bradley CT, Brasel KJ, Schwarze ML. Physician attitudes regarding advance directives for high-risk surgical patients: a qualitative analysis. Surgery. 2010; 148(2):209-216. 17. Conrad F, Blair J. From impression to data: increasing the objectivity of cognitive interviews. In: Proceedings of the Survey Research Methods Section of the American Statistical Association. Alexandria, VA: American Statistical Association; 1996. 18. Drennan J. Cognitive interviewing: verbal data in the design and pretesting of questionnaires. J Adv Nurs. 2003;42(1):57-63. 19. American Association for Public Opinion Research. Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys. 7th ed. Deerfield, IL: American Association for Public Opinion; 2011. 20. Schwarze ML, Redmann AJ, Brasel KJ, Alexander GC. The role of surgeon error in withdrawal of postoperative life support. Ann Surg. 2012;256(1):10-15. 21. Penkoske PA, Buchman TG. The relationship between the surgeon and the intensivist in the surgical intensive care unit. Surg Clin North Am. 2006;86(6): 1351-1357. 22. Abbott KH, Sago JG, Breen CM, Abernethy AP, Tulsky JA. Families looking back: one year after discussion of withdrawal or withholding of life-sustaining support. Crit Care Med. 2001;29(1):197-201. 23. Buchman TG, Ray SE, Wax ML, Cassell J, Rich D, Niemczycki MA. Families’ perceptions of surgical intensive care. J Am Coll Surg. 2003;196(6):977-983. 24. Redmann AJ, Brasel KJ, Alexander CG, Schwarze ML. Use of advance directives for high-risk operations: a national survey of surgeons. Ann Surg. 2012;255 (3):418-423. REFERENCES JAMA SURG/ VOL 148 (NO. 1), JAN 2013 34 WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 25. Sudore RL, Fried TR. Redefining the “planning” in advance care planning: preparing for end-of-life decision making. Ann Intern Med. 2010;153(4):256-261. 26. Eachempati SR, Miller FG, Fins JJ. The surgical intensivist as mediator of endof-life issues in the care of critically ill patients. J Am Coll Surg. 2003;197(5): 847-853, discussion 853-854. 27. Schaefer KG, Block SD. Physician communication with families in the ICU: evidencebased strategies for improvement. Curr Opin Crit Care. 2009;15(6):569-577. 28. Whitman GJ, Haddad M, Hirose H, Allen JG, Lusardi M, Murphy MA. Cardiothoracic surgeon management of postoperative cardiac critical care. Arch Surg. 2011; 146(11):1253-1260. INVITED CRITIQUE Patient Ownership and Conflict in the Intensive Care Unit Mine, Yours, or Ours? P aul Olson and colleagues1 attempt to answer the question of patient ownership with a survey of surgeons who frequently have patients in the intensive care unit. In view of the increased prevalence of hospitalists—whether they be intensivists, trauma surgeons, anesthesiologists, or internists—and the current resident work-hour restrictions, the issue of continuity of care is an increasingly important one. In the first few sentences of their article, the authors set the stage, citing recent critical care publications in which 70% of intensive care unit clinicians reported experiencing conflict on a weekly basis! This subject is an important one, particularly as it pertains to surgeons’ difficulty accepting bad outcomes in their patients and de-escalating care. No event in the life of a surgeon surpasses the physical and emotional impact of a bad surgical outcome. We are, by training, taught to help others. When as a consequence of our direct action (an operation or treatment) an adverse outcome ensues, there is often difficulty in giving up—a difficulty in admitting that this situation cannot be salvaged. The article by Paul Olson et al1 provides valuable insight. One of the better messages in this article is the no doubt guilt- accentuated concerns about the patient with a less-thandesired outcome; predictably, the operating surgeon is less prompt to move toward palliative care. Just who is the more consistent and ethical conservator of resources varies widely in the hospitals in our university medical center. I do not believe intensivists are more protective of those resources; more likely, the opposite is true. The reader must also be wondering how intensive care unit nurses view these complex and intense interactions. These kinds of questions are hard to pose meaningfully and even more difficult to answer. Susan Galandiuk, MD Author Affiliation: Department of Surgery, University of Louisville, Kentucky. Correspondence: Dr Galandiuk, Department of Surgery, University of Louisville, Louisville, KY 40292 ([email protected]). Conflict of Interest Disclosures: None reported. 1. Paul Olson TJ, Brasel KJ, Redmann AJ, Alexander GC, Schwarze ML. Surgeonreported conflict with intensivists about postoperative goals of care. JAMA Surg. 2013;148(1):29-35. JAMA SURG/ VOL 148 (NO. 1), JAN 2013 35 WWW.JAMASURG.COM ©2013 American Medical Association. All rights reserved. Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 Factors Associated With Elevated Plateau Pressure in Patients With Acute Lung Injury Receiving Lower Tidal Volume Ventilation Hallie C. Prescott, MD1,2; Roy G. Brower, MD3; Colin R. Cooke, MD, MSc4; Gary Phillips, MAS5; James M. O’Brien, MD, MSc6; for the National Institutes of Health Acute Respiratory Distress Syndrome Investigators Objectives: Lung-protective ventilation with lower tidal volume and lower plateau pressure improves mortality in patients with acute lung injury and acute respiratory distress syndrome. We sought to determine the incidence of elevated plateau pressure in acute lung injury /acute respiratory distress syndrome patients receiving lower tidal volume ventilation and to determine the factors that predict elevated plateau pressure in these patients. Patients: We used data from 1398 participants in Acute Respiratory Distress Syndrome Network trials, who received lower tidal volume ventilation (≤ 6.5 mL/kg predicted body weight). Design: We considered patients with a plateau pressure greater than 30 cm H2O and/or a tidal volume less than 5.5 mL/kg predicted body weight on study day 1 to have “elevated plateau pressure.” We used logistic regression to identify baseline clinical variables a ssociated with elevated plateau pressure and to develop a model to predict elevated plateau pressure using a subset of 1,188 patients. We validated the model in the 210 patients not used for model development. Setting: Medical centers participating in Acute Respiratory Distress Syndrome Network clinical trials. 1Department of Internal Medicine, The Ohio State University Medical Center, Columbus, OH. 2Division of Pulmonary and Critical Care Medicine, University of Michigan Health System, Ann Arbor, MI. 3Pulmonary and Critical Care Medicine, John Hopkins University School of Medicine, Baltimore, MD. 4University of Michigan, Robert Wood Johnson Clinical Scholars Program and Center for Healthcare Outcomes & Policy, Ann Arbor, MI. 5The Ohio State University Center for Biostatistics, Columbus, OH. 6Division of Pulmonary, Allergy, Critical Care & Sleep Medicine, Department of Internal Medicine, The Ohio State University Medical Center, Columbus, OH. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (http://journals.lww.com/ccmjournal). Supported, in part, by National Heart, Lung, and Blood Institute (NHLBI) contract NO1-HR-46054–46064 and NO1-HR 56165–56179.s The authors have not disclosed any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Critical Care Medicine Interventions: None. Measurements and Main Results: Of the 1,398 patients in our study, 288 (20.6%) had elevated plateau pressure on day 1. Severity of illness indices and demographic factors (younger age, greater body mass index, and non-white race) were independently associated with elevated plateau pressure. The multivariable logistic regression model for predicting elevated plateau pressure had an area under the receiving operator characteristic curve of 0.71 for both the developmental and the validation subsets. Conclusions: acute lung injury patients receiving lower tidal volume ventilation often have a plateau pressure that exceeds Acute Respiratory Distress Syndrome Network goals. Race, body mass index, and severity of lung injury are each independently associated with elevated plateau pressure. Selecting a smaller initial tidal volume for non-white patients and patients with higher severity of illness may decrease the incidence of elevated plateau pressure. Prospective studies are needed to evaluate this approach. (Crit Care Med 2013; 41:0–0) Key Words: acute lung injury; acute respiratory distress syndrome; mechanical ventilation; plateau pressure A cute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are common and deadly diagnoses that affect 190,000 patients (1) each year in the United States and are associated with a 36–44% mortality rate (1–3). Mechanical ventilation strategies in ALI patients aim to minimize ventilator-induced lung injury while providing adequate oxygenation and ventilation. Accurate bedside measures of lung stress and lung strain are not readily available, so clinicians rely on surrogate measures of transpulmonary pressure to guide ventilator management. Despite its limitations (4), plateau pressure is used commonly as a surrogate measure of transpulmonary pressure. The association between plateau pressure (Pplat) and mortality, as Copyright © 2013 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI:10.1097/CCM.0b013e3182741790 www.ccmjournal.org 1 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Prescott et al has been demonstrated in retrospective analyses using ARDS Network trial participants, supports this practice (5, 6). In 2000, the ARDS Network published a multicenter trial comparing lower tidal volume (VT) ventilation (6 mL/kg predicted body weight [PBW]) to traditional ventilation (12 mL/kg PBW) in ALI patients. The lower VT arm had an 8.8% lower mortality rate compared with the control arm (7). As a means of scaling VT to lung size, the ARDS Network protocol calculated a PBW for each patient based on sex and height, and then used this PBW to determine a VT (7). VT was then adjusted based on estimates of lung stress (second order scaling). Plateau pressure was monitored every 4 hrs, and VT reduced if Pplat exceeded 30 cm H2O in the lower VT arm or 50 cm H2O in the traditional ventilation arm (7). Since publication of the ARDS Network lower VT trial, physicians have prescribed smaller VTs (5, 8). However, Pplat monitoring is also part of the ARDS Network ventilation protocol, there is less evidence to suggest that Pplat monitoring or reduction of VT in response to elevated Pplat has been adopted on a widespread level. Several studies of physician practice patterns for treating ALI do not comment on Pplat (9, 10) because this variable is frequently unavailable or underreported in datasets (8, 10–13), suggesting that Pplat monitoring does not occur regularly in clinical practice. Moreover, the initial VT often remains in place for over 24 hrs (9, 14) even when the VT exceeds ARDS Network goals, suggesting that physicians do not routinely reassess ventilator settings. Given the association between higher Pplat and mortality, and the concern that Pplat monitoring has not been widely adopted, we sought to determine the incidence of elevated Pplat in ALI patients receiving a VT of 6 mL/kg PBW. Our second objective was to identify factors that predict elevated Pplat in patients receiving lower VT ventilation. MATERIALS AND METHODS Study Subjects We performed a secondary analysis of data collected from ALI/ ARDS patients enrolled in three ARDS Network clinical trials between 1996 and 2005, including a trial of different VTs (7), a trial of different positive end-expiratory pressures (15), and a trial of different hemodynamic management strategies (16, 17). Patients with VT > 6.5 mL/kg PBW and those missing VT or Pplat data for study day 1 were excluded from our study. Definitions The outcome of interest for this analysis was an “elevated Pplat” on study day 1 (postenrollment). We defined “elevated Pplat” as a measured Pplat > 30 cm H2O or a VT < 5.5 mL/kg PBW. The reduction of VT to less than 6.0 mL/kg PBW was assumed to be the result of a prior Pplat measurement greater than 30 cm H2O. Statistical Analyses Descriptive statistics used counts and frequencies for categorical data and means and standard deviations for continuous variables. To determine whether differences existed between the study population and those excluded for missing data, we used Pearson’s 2 www.ccmjournal.org chi-square for categorical variables and Wilcoxon rank-sum test for continuous variables. We used a two-sample t test to evaluate whether differences in Pplat existed across VT and Pearson’s chisquare to test if Pplat > 30 cm H2O was different across VT. Modeling Approach To determine the factors independently associated with elevated Pplat, we used logistic regression to model the effect of baseline (preenrollment) variables on elevated Pplat. We performed two distinct approaches to model development and validation: cross-validation and bootstrap resampling (Online supplement, Supplemental Digital Content 1, http://links.lww.com/CCM/A561). In the cross-validation approach, the study population was randomly divided into development (85%) and validation (15%) groups. The regression model was derived in the development sample, and model performance was assessed in the validation sample. The size of the validation group (210 subjects) was selected in order to provide enough data to perform a goodness-of-fit test while still allowing the majority (1188 subjects) to be used in estimating the model coefficients (18). Model Selection We first conducted a preliminary univariable analysis to determine the associations between baseline variables and the odds of elevated Pplat on study day 1. Candidate variables for the multivariable logistic regression included: age, sex, race, body mass index (BMI; calculated from enrollment height and weight), PaO2/Fio2 ratio, cirrhosis, chronic dialysis, serum albumin, net fluid balance over the 24 hrs preceding study enrollment, use of sedation or neuromuscular blockade (NMB), number of quadrants of infiltrate on enrollment chest radiograph, APACHE III score, and diabetes. We evaluated BMI as both a continuous variable and a categorical variable using the previously defined classifications: underweight < 18.5 kg/m2, normal 18.5–24.9 kg/m2, overweight 25–29.9 kg/m2, obese 30–34.9 kg/m2, and severely obese ≥ 35 kg/ m2. Race was recorded in the ARDS Network trials as white nonHispanic, black non-Hispanic, and other; we reclassified race as white or non-white for our analysis. We included all variables in the multivariable logistic regression if they achieved a p value of less than or equal to 0.2 in a univariable logistic regression with elevated Pplat as the outcome. The form of the continuous variables was assessed using fractional polynomials to ensure that the logit was linear (19). All variables considered for model inclusion were tested for all possible two-way interactions. All of the candidate variables and their interactions were included in the multivariable model. Interactions were dropped from the model one at a time if they were no longer statistically significant (p > 0.050). This step was followed by dropping main effect variables if they were no longer significant. The predictor variables were also assessed for collinearity using variance inflation factor. Model Performance We evaluated the discrimination of the cross-validation model using AUC in both the development and the validation cohorts. We assessed calibration using the Hosmer-Lemeshow goodnessof-fit test (20, 21). Given the low power of the Hosmer-Lemeshow goodness-of-fit test to detect poor performance of models with March 2013 • Volume 41 • Number 3 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Clinical Investigation RESULTS Study Sample The three ARDS Network studies enrolled a total of 2,452 patients. Of these, 785 (32.0%) were excluded from our analysis for VT > 6.5 mL/kg PBW, and 269 (11.0%) were excluded for missing day 1 Pplat and/or day 1 VT data (Fig. 1). Of those with missing data, 185 (68.8%) were missing both Pplat and VT, 60 (22.3%) were missing just Pplat, and 24 (8.9%) were missing just VT. In the VT trial, 429 patients were randomized to a VT of 12 mL/kg (7); these patients account for 54.6% of the patients excluded from this study for VT > 6.5 mL/kg. The 269 patients excluded for missing data had similar baseline characteristics to the 1,398 patients (57.0%) included in the analysis (data not shown), with the exception of a greater mean PaO2/Fio2 ratio (150.4 mm Hg vs. 126.9 mm Hg, p < 0.001). Also, the patients excluded for missing data had a higher proportion enrolled in the VT trial (28.6% vs. 24.5%) and the hemodynamic management trial (49.1% vs. 45.6%) and a smaller proportion enrolled in the PEEP trial (22.3% vs. 30.0%) than those included in the Figure 1. Flow chart of study. All patients enrolled in three Acute Respiratory Distress Syndrome (ARDS) analysis (p = 0.031). Network trials were screened for inclusion in our study. Patients receiving tidal volume (VT) > 6.5 mL/kg Summary statistics for entire predicted body weight (PBW) and patients missing plateau pressure (Pplat) and/or VT data for study day 1 were excluded for our analysis. The remaining 1,398 subjects were categorized as having target vs. elevated Pplat on study population, the developmenstudy day 1. tal cohort, and the validation cohort are provided in Table 1. The develsmall sample sizes (22), we also compared actual and predicted opment and validation cohorts had frequencies of elevated Pplat by decile of predicted probability. similar baseline characteristics. The overall study population of 1,398 patients had a mean Pplat Sensitivity Analyses of 24.3 ± 6.4 cm H2O on day 1. In order to test the assumptions made in building the regression The 183 patients with VT < 5.5 mL/kg PBW had a greater mean Pplat and a higher frequency of Pplat > 30 cm H2O than patients with model, we performed two sensitivity analyses. In the first, we removed patients with VT < 5.5 mL/kg PBW to evaluate whether VT 5.5–6.5 mL/kg PBW (Table 2). Histograms of Pplat by VT demour classification of these subjects as having elevated Pplat was ap- onstrate that the frequency distribution for Pplat is shifted upward propriate. In the second, we removed patients with barotrauma about 5 cm H2O for patients receiving a VT < 5.5 mL/kg PBW in comparison with patients receiving a VT of 5.5–6.5 mL/kg PBW because we were unsure if or how barotrauma may have affected (Fig. 2). The frequency of elevated Pplat did not differ by study Pplat. We refit these revised datasets to the model from the develenrollment: (63/342) 18.4% of patients in the VT trial, (87/419) opmental dataset. 20.8% of patients in the PEEP trial, and (138/637) 21.7% of paThe Institutional Review Board for The Ohio State University tients in the hemodynamic management trial had elevated Pplat approved the study. All statistical analyses were conducted using (p = 0.487). On average, patients gained a net 2240 mL of fluid in Stata 10.1, Stata Corporation, College Station, Texas. the day preceding enrollment. Critical Care Medicine www.ccmjournal.org 3 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Prescott et al Table 1. Baseline Characteristics of Patients by Study Population Developmental Cohort Validation Cohort Characteristic Entire Study Subjects, n (%) 1398 (100) 1188 (85) 210 (15) Pplat (cm H2O), mean ± sd 24.3 ± 6.4 24.3 ± 6.4 24.6 ± 6.4 Elevated Pplat, % 288 (20.6) 240 (20.2) 48 (22.9) Pplat > 30 cm H2O, n (%) 179 (12.8) 147 (12.4) 32 (15.2) VT < 5.5 mL/kg PBW, n (%) 183 (13.1) 153 (13.1) 27 (12.9) 74 (5.3) 63 (5.3) 11 (5.2) 49.9 ± 16.7 49.9 ± 16.6 50.3 ± 17.3 Male, n (%) 810 (57.9) 685 (57.7) 125 (59.5) Non-white, n (%) 420 (30.0) 357 (30.1) 63 (30.0) White, non-Hispanic 978 (70.0) 831 (70.0) 147 (70.0) Black, non-Hispanic 253 (18.1) 219 (18.4) 34 (16.2) Other 167 (11.9) 138 (11.6) 29 (13.8) Pplat > 30 cm H2O and VT < 5.5 mL/kg PBW, n (%) Age (years), mean ± sd Race, n (%) BMI (kg/m ), mean ± sd 27.8 ± 7.1 2 27.8 ± 7.0 27.4 ± 7.3 Underweight BMI (<18.5 kg/m ), n (%) 60 (4.4) 50 (4.3) Normal BMI (18.5–24.9 kg/m ), n (%) 482 (35.7) 403 (35.0) 79 (39.3) Overweight BMI (25.0–29.9 kg/m2), n (%) 394 (29.1) 336 (29.1) 58 (28.9) Obese BMI (30.0–34.9 kg/m2), n (%) 228 (16.8) 197 (17.1) 31 (15.4) Severely obese BMI (≥ 35.0 kg/m2), N (%) 190 (14.0) 167 (14.5) 23 (11.4) 126.9 ± 59.1 127 ± 59.1 126.0 ± 59.2 91.5 ± 30.5 91.3 ± 30.6 92.8 ± 30.3 Four-quadrant infiltrate on chest radiograph, n (%) 1,061 (75.9) 902 (75.9) 159 (75.7) Fluid balance during preceding 24 hrs (mL), mean ± sd 2,240 ± 3,727 2,231 ± 3,750 2,286 ± 3,608 Sedation or NMB, n (%) 1,215 (86.9) 1,031 (86.8) 184 (87.6) 2 2 PaO2/Fio2 ratio, mean ± sd Acute Physiology and Chronic Health Evaluation III, mean ± sd 10 (5.0) Cirrhosis, n (%) 42 (3.0) 34 (2.9) 7 (3.3) Diabetes, n (%) 219 (15.7) 186 (15.7) 33 (15.7) Chronic dialysis, n (%) 20 (1.4) 17 (1.4) 3 (1.4) Serum albumin (g/dL), mean ± sd 2.2 ± 0.7 2.3 ± 0.7 2.1 ± 0.6 ARMA, n (%) 342 (24.5) 289 (24.3) 53 (25.2) ALVEOLI, n (%) 419 (30.0) 252 (21.2) 67 (31.9) FACTT, n (%) 637 (45.6) 547 (46.0) 90 (42.9) BMI = body mass index; NMB = neuromuscular blockade; Pplat = plateau pressure; sd = standard deviation; VT = tidal volume. Plateau Pressure Outcome Groups According to our study definitions, 1110 patients (79.4%) had ontarget Pplat and 288 patients (20.6%) had elevated Pplat on study day 1 (Fig. 1). Of the 288 patients with elevated day 1 Pplat, 105 (36.5%) did not have VT reduced below 5.5 mL/kg in response to their Pplat recording. Three of these 105 patients had an arterial pH < 7.15, which was a protocol exception allowing higher Pplat; for the other 4 www.ccmjournal.org 102 patients, there appears to have been nonadherence with the intended study protocol. Following these patients to study day 3, 86 of the 102 patients were still alive and had Pplat and VT data recorded: 33 of 86 (38.4%) still had Pplat > 30 cm H2O, and 25 (75.8%) of these 33 patients with Pplat > 30 cm H2O still had VT ≥ 5.5 mL/kg PBW. None of these 25 patients had protocol-approved exceptions allowing Pplat > 30 cm H2O, suggesting nonadherence to the study protocol. March 2013 • Volume 41 • Number 3 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Clinical Investigation Table 2. Plateau Pressure Statistics by Tidal Volume Cohort Entire Study VT < 5.5 VT 5.5–6.5 pa 1398 (100) 183 (13.1) 1215 (86.9) — Pplat (cm H2O), mean ± sd 24.3 ± 6.4 30.2 ± 6.5 23.5 ± 5.9 < 0.001 Pplat > 30 cm H2O, n (%) 288 (20.6) 74 (40.4) 105 (8.6) < 0.001 Subjects, n (%) Pplat = plateau pressure; VT = tidal volume. a p values based on comparison of VT < 5.5 vs. VT 5.5–6.5 based on either the two-sample t test or the Pearson’s chi-square test. Univariable Analysis Summary statistics by outcome group are presented in Table 3. Compared with patients with on-target Pplat, those with elevated Pplat were more likely to be non-white, severely obese (BMI ≥ 35 kg/ m2), have four quadrants of infiltrate on chest radiograph, to have received sedation or NMB, and to be receiving chronic dialysis. Patients with elevated Pplat were also younger, had a lower PaO2/ Fio2 ratio, a higher APACHE III score, a higher serum albumin, and received more fluid in the 24 hrs preceding study enrollment (+2093 mL for Pplat < 30 cm H2O vs. +2803 mL for elevated Pplat, p < 0.001). Multivariable Analysis Age, BMI, race, PaO2/Fio2, APACHE III, and four-quadrant chest radiograph infiltrates remained significant in multivariable analysis (Table 4). Each of these variables had a directly or inversely linear relationship with elevated Pplat, with the exception of BMI, which had a J-shaped relationship. Severely obese BMI was associated with increased odds of elevated Pplat (OR = 1.85, p < 0.001), whereas underweight BMI trended towards greater odds of elevated plateau pressures (OR = 1.87, p = 0.080). indicating good discrimination of the model. The HosmerLemeshow goodness-of-fit test showed good calibration with no evidence of inadequate fit for the development or validation models (χ2df=8 = 4.57, p = 0.803; χ2df=10 = 12.97, p = 0.227). Comparing observed vs. predicted rates of elevated Pplat by decile of predicted probability also demonstrated good calibration of the model (Fig. 3). Using conservative estimates, there was no evidence of collinearity in our model. The bootstrap model had the same final variables as the crossvalidation model and similar variable coefficients, discrimination, and calibration (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/CCM/A561). Removing subjects with barotrauma or VT < 5.5 mL/kg PBW and refitting the remaining subjects to the logistic regression model did not significantly alter the model or the variable coefficients (Supplemental Tables 2 and 3, Supplemental Digital Content 1, http://links.lww.com/CCM/A561). DISCUSSION Using data collected from a large sample of participants in ARDS Network trials conducted across multiple centers, we determined that approximately one in five patients with ALI has elevated day 1 Model Performance Pplat despite receiving lower VT ventilation. The rate of elevated Pplat The AUC was 0.71 (95% CI: 0.67–0.74) for the developmental likely varies by patient population, and the incidence observed in dataset and 0.71 (95% CI: 0.62–0.79) for the validation dataset, our study may have been inflated by excluding patients with missing day 1 VT and/or Pplat data who had a greater mean PaO2/Fio2 than the patients included in the study. Still, the 20% incidence seen in our study indicates that a sizeable minority of patients may have elevated Pplat despite receiving lower VT ventilation. This suggests that setting a VT of 6 mL/kg PBW alone does not ensure ALI patients receive the full benefit of lung-protective ventilation. An interesting finding in our study was the incidence of nonadherence to the ventilation protocol. A third of the patients with elevated Pplat did not have VT reduced according to study design. In routine clinical practice, noncompliance with the Figure 2. Histogram of plateau pressure by tidal volume (VT). This histogram shows the distribution of maxiARDS Network ventilation protocol mum plateau pressures recorded on study day 1 for the 1215 patients with VT 5.5–6.5 mL/kg predicted body weight (PBW) and the 183 patients with VT < 5.5 mL/kg PBW. is likely higher than the rate observed Critical Care Medicine www.ccmjournal.org 5 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Prescott et al Table 3. Baseline Characteristics of the Patients by Outcome Group On-target Pplat Characteristic Elevated Pplat pa 1110 (79.4%) 288 (20.6%) Pplat (cm H2O), mean ± sd 22.4 ± 4.7 32.0 ± 6.2 Unacceptable Pplat, n (%) 0 (0.0) 1398 (100) N/Ab Pplat > 30 cm H2O, N (%) 0 (0.0) 179 (62.2) N/Ab VT < 5.5 mL/kg PBW, n (%) 0 (0.0) 183 (63.5) N/Ab Pplat > 30 cm H2O and VT < 5.5 mL/kg PBW, n (%) 0 (0.0) 74 (25.7) N/Ab 51.2 ± 16.7 45.1 ± 15.8 <0.001 Subjects, n (%) Age (years), mean ± sd <0.001 Male, n (%) 654 (58.9) 156 (54.2) 0.145 Non-white, n (%) 297 (26.8) 123 (42.7) <0.001 27.6 ± 6.9 28.4 ± 7.8 Underweight BMI (<18.5 kg/m ), n (%) 43 (3.9) 17 (6.0) Normal BMI (18.5–24.9 kg/m ), n (%) 389 (36.4) 93 (33.0) Overweight BMI (25.0–29.9 kg/m2), n (%) 321 (29.9) 73 (25.9) Obese BMI (30.0–34.9 kg/m2), n (%) 178 (16.6) 50 (17.7) Severely obese BMI (≥ 35.0 kg/m2), n (%) 141 (13.2) 49 (17.4) PaO2/Fio2 ratio, mean ± sd 131.1 ± 60.0 110.6 ± 52.8 <0.001 89.1 ± 29.8 100.8 ± 31.4 <0.001 246 (85.4) <0.001 2802 ± 3398 <0.001 BMI (kg/m ), mean ± sd 2 2 2 Acute Physiology and Chronic Health Evaluation III, mean ± sd 0.178 0.122 Four-quadrant infiltrate on chest radiograph, n (%) 815 (73.4) Fluid balance (mL) during preceding 24 hrs, mean ± sd 2093 ± 3796 Sedation or NMB, n (%) 950 (85.6) 265 (92.0) 0.004 Cirrhosis, n (%) 33 (3.0) 9 (3.1) 0.893 Diabetes, n (%) 166 (15.0) 54 (18.8) 0.115 19 (1.7) 0 (0.0) 0.025 2.2 ± 0.6 0.039 Chronic dialysis, n (%) Albumin (g/dL), mean ± sd 2.3 ± 0.7 ARMA, n (%) 279 (25.1) 63 (21.9) ALVEOLI, n (%) 332 (29.9) 87 (30.2) FACTT, n (%) 499 (45.0) 138 (47.9) 0.208 BMI = body mass index; NMB = neuromuscular blockade; Pplat = plateau pressure; sd = standard deviation; VT = tidal volume. a p values are based on Pearson’s chi-square for categorical variables or Wilcoxon rank-sum test for continuous variables b Different by definition in clinical trial participants, as supported by recent observational study where only 41% of ventilator settings in ALI patients were adherent to ARDS Network goals (13). In this study, each additional ventilator setting adherent to ARDS Network goals was associated with a measureable decrease in 2-year mortality (13). Thus, it is important to identify patients at higher risk of elevated Pplat and to consider additional methods to limit exposure to elevated Pplat. Our regression model identified several factors that predict elevated Pplat. As would be expected, measures of the severity of lung injury and overall illness were associated with elevated Pplat. We also identified that race and BMI are both independently associated with plateau pressure. 6 www.ccmjournal.org In our model, non-white race was associated with an almost two-fold increase in the risk of elevated Pplat. Since the ARDS Network lower VT ventilation protocol was developed, reference equations for normal lung function have been revised to include race (23, 24). For a given height and gender, African Americans have 12% smaller total lung capacity than people of white race (23, 25), and limited available data suggests that other non-white racial groups also have smaller lung volumes (23, 26). Accounting for differences in lung volume by race when selecting VT may limit exposure to elevated Pplat in non-white patients. For example, VT could be set at 5 mL/kg PBW in African Americans to account for the expected 12% reduction in TLC. March 2013 • Volume 41 • Number 3 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Clinical Investigation Logistic Regression to Predict Elevated Plateau Pressure Using the Developmental Dataset lung capacity and functional residual capacity (27–29), likely due to abdominal and intrathoracic fat limiting full diaphragm and lung expansion (30). Thus, calculating VT by PBW results in relatively larger VT (expressed as a proportion of lung capacity) as BMI increases. However, because obesity is also associated with Odds Ratio (95% Confidence decreased respiratory system compliance and elevated pleural Variable Interval) p pressure (31), elevated Pplat may not signify higher transpulmonary pressure. Knowing that pleural pressure is elevated in obese patients Age, 10 yr increase 0.79 (0.72–0.87) <0.001 and patients with elevated intra-abdominal pressure for other Body mass index reasons, physicians may choose to accept a higher Pplat threshold in these patients, use esophageal pressure monitoring for a better 1.87 (0.93–3.87) 0.08 Underweight (<18.5 kg/m2) estimation of transpulmonary pressure, or simply target a VT of 6 mL/kg until further prospective studies provide guidance for 1 — Normal, referent optimizing lung-protective ventilation in this patient population. (18.5–24.9 kg/m2) There are several important limitations to this study. As a retro1.23 (0.83–1.83) 0.31 Overweight spective analysis, this study was limited by available data. Race was 2 (25.0–29.9 kg/m ) reported only as white non-Hispanic, black non-Hispanic, and 1.31 (0.83–2.06) 0.242 Obese (30.0–34.9 kg/m2) other, so we could not evaluate associations between specific races and elevated Pplat. Furthermore, we could not assess whether the 2.11 (1.34–3.33) 0.001 Severely obese (≥35.0 kg/m2) differences in Pplat by race are due to differences in genetic ancestry vs. confounding socioeconomic factors. Similar limitations have Non-white race 1.85 (1.35–2.53) <0.001 been encountered when evaluating the association between race 0.95 (0.92–0.97) <0.001 PaO2/Fio2, 10 unit increase and normal lung function (32, 33). Yet, reference equations for Acute Physiology and 1.12 (1.06–1.18) <0.001 normal lung function are based on self-reported race because this Chronic Health Evaluation information does improve prediction of lung function (32, 33). III, 10 unit increase BMI was calculated from weight measured after the onset of Four-quadrant infiltrate on 1.87 (1.24–2.82) 0.003 critical illness, so fluid balance may have affected these calculations. chest radiograph A previous analysis of subjects from the ARDS Network lower VT trial demonstrated that adjusting for fluid balance changed the BMI The benefit of this approach would need to be confirmed in a category in 12.9% of subjects although the correlation between prospective study. fluid-adjusted and unadjusted BMI was strong (r = 0.978 (34)). Patients with a severely obese BMI (≥35.0 kg/m2) had a more We chose a threshold of 30 cm H2O to define elevated Pplat bethan two-fold increase in risk of elevated Pplat. Obesity, and in cause it prompted VT reduction according to the ARDS Network particular severe obesity, has been associated with decreased total protocol (7) and still remains the point at which VT reduction is recommended (7, 35). However, this is merely a surrogate marker for lung strain. Regional hyperinflation and lung injury are still possible with Pplat < 30 cm H2O (36, 37). Other measurements of lung injury, such as biochemical markers, histopathologic findings, and transpulmonary pressure, determined with esophageal pressure monitoring may be more accurate but are less widely used in clinical practice and were not available in our datasets. Finally, we developed a predictive rather than an explanatory model of factors associated with elevated Pplat. Because of this approach, there may be confounding factors associated with Pplat that are not included in the model. The beneficial or harmful efFigure 3. Predicted vs. observed rates of elevated plateau pressure (Pplat) for the cross-validation logistic fects of hypercarbia as a result of regression model. Patients were stratified according to their probability of elevated Pplat predicted by the decreased VT, which could be more regression model and then divided into 10 groups. The predicted rate of elevated Pplat was compared with the common in certain patient groups observed rate of elevated Pplat for each group and the overall population. Table 4. Critical Care Medicine www.ccmjournal.org 7 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited Prescott et al based on physiologic parameters, was beyond the scope of our analysis. CONCLUSION A ventilation strategy incorporating lower VT with further VT reductions based on Pplat measurements decreases mortality in patients with ALI, yet one in five ALI patients has elevated day 1 Pplat while receiving a VT of 6 mL/kg PBW. Race, BMI, and severity of lung injury are independently associated with Pplat in patients receiving lower VT ventilation. Selecting a smaller initial VT for non-white patients and patients with higher severity of lung injury may decrease the incidence of elevated Pplat. However, we do not suggest this approach in severely obese patients because higher P plats may be necessary to maintain appropriate transpulmonary pressure in this population. Prospective studies are needed to confirm the benefit of lowering initial VTs in non-white patients and patients with higher severity of lung injury. REFERENCES 1.Rubenfeld GD, Caldwell E, Peabody E, et al: Incidence and outcomes of acute lung injury. N Engl J Med 2005; 353:1685–1693 2. Zambon M, Vincent JL: Mortality rates for patients with acute lung injury/ ARDS have decreased over time. Chest 2008; 133:1120–1127 3.Phua J, Badia JR, Adhikari NK, et al: Has mortality from acute respiratory distress syndrome decreased over time?: A systematic review. Am J Respir Crit Care Med 2009; 179:220–227 4.Chiumello D, Carlesso E, Cadringher P, et al: Lung stress and strain during mechanical ventilation for acute respiratory distress syndrome. Am J Respir Crit Care Med 2008; 178:346–355 5.Checkley W, Brower R, Korpak A, et al; Acute Respiratory Distress Syndrome Network Investigators: Effects of a clinical trial on mechanical ventilation practices in patients with acute lung injury. Am J Respir Crit Care Med 2008; 177:1215–1222 6.Brower RG, Matthay M, Schoenfeld D: Meta-analysis of acute lung injury and acute respiratory distress syndrome trials. Am J Respir Crit Care Med 2002; 166:1515–1517 7. 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; 342:1301–1308 8. Esteban A, Ferguson ND, Meade MO, et al; VENTILA Group: Evolution of mechanical ventilation in response to clinical research. Am J Respir Crit Care Med 2008; 177:170–177 9.Umoh NJ, Fan E, Mendez-Tellez PA, et al: Patient and intensive care unit organizational factors associated with low tidal volume ventilation in acute lung injury. Crit Care Med 2008; 36:1463–1468 10.Brun-Buisson C, Doyon F, Sollet JP, et al: Prevention of intravascular catheter-related infection with newer chlorhexidine-silver sulfadiazinecoated catheters: a randomized controlled trial. Intensive Care Med 2004; 30:837–843 11.Akhtar SR, Weaver J, Pierson DJ, et al: Practice variation in respiratory therapy documentation during mechanical ventilation. Chest 2003; 124:2275–2282 12.Cooke CR, Watkins TR, Kahn JM, et al: The effect of an intensive care unit staffing model on tidal volume in patients with acute lung injury. Crit Care 2008; 12:R134 13.Needham DM, Colantuoni E, Mendez-Tellez PA, et al: Lung protective mechanical ventilation and two year survival in patients with acute lung injury: prospective cohort study. BMJ 2012; 344:e2124 14. Gajic O, Dara SI, Mendez JL, et al: Ventilator-associated lung injury in patients without acute lung injury at the onset of mechanical ventilation. Crit Care Med 2004; 32:1817–1824 8 www.ccmjournal.org 15. Brower RG, Lanken PN, MacIntyre N, et al; National Heart, Lung, and Blood Institute ARDS Clinical Trials Network: Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med 2004; 351:327–336 16.Wheeler AP, Bernard GR, Thompson BT, et al; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network: Pulmonary-artery versus central venous catheter to guide treatment of acute lung injury. N Engl J Med 2006; 354:2213–2224 17.Wiedemann HP, Wheeler AP, Bernard GR, et al; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network: Comparison of two fluid-management strategies in acute lung injury. N Engl J Med 2006; 354:2564–2575 18.Hosmer DW, Lemeshow S, Wiley online library: Applied Logistic Regression. Second Edition. New York, Wiley, 2000, pp 373 19.Royston P, Altman DG: Regression using fractional polynomials of continuous covariates: parsimonious parametric modelling. J Roy Stat Soc Ser C (Applied Statistics) 1994; 43:429–467 20.Hanley JA, McNeil BJ: The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29–36 21. Lemeshow S, Hosmer DW Jr: A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982; 115:92–106 22. Zhu BP, Lemeshow S, Hosmer DW, et al: Factors affecting the performance of the models in the Mortality Probability Model II system and strategies of customization: a simulation study. Crit Care Med 1996; 24:57–63 23.Stocks J, Quanjer PH: Reference values for residual volume, functional residual capacity and total lung capacity. ATS Workshop on Lung Volume Measurements. Official Statement of The European Respiratory Society. Eur Respir J 1995; 8:492–506 24. Hankinson JL, Odencrantz JR, Fedan KB: Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med 1999; 159:179–187 25.Lapp NL, Amandus HE, Hall R, et al: Lung volumes and flow rates in black and white subjects. Thorax 1974; 29:185–188 26.Quanjer PH, Tammeling GJ, Cotes JE, et al: Lung volumes and forced ventilatory flows. Report Working Party Standardization of Lung Function Tests, European Community for Steel and Coal. Official Statement of the European Respiratory Society. Eur Respir J Suppl 1993; 16:5–40 27.Jones RL, Nzekwu MM: The effects of body mass index on lung volumes. Chest 2006; 130:827–833 28.Ray CS, Sue DY, Bray G, et al: Effects of obesity on respiratory function. Am Rev Respir Dis 1983; 128:501–506 29. Porhomayon J, Papadakos P, Singh A, et al: Alteration in respiratory physiology in obesity for anesthesia-critical care physician. HSR Proceedings in Intenstive Care and Cardiovascular Anesteria 2011; 3:109–118 30.Watson RA, Pride NB, Thomas EL, et al: Reduction of total lung capacity in obese men: Comparison of total intrathoracic and gas volumes. J Appl Physiol 2010; 108:1605–1612 31.Behazin N, Jones SB, Cohen RI, et al: Respiratory restriction and elevated pleural and esophageal pressures in morbid obesity. J Appl Physiol 2010; 108:212–218 32.Kumar R, Seibold MA, Aldrich MC, et al: Genetic ancestry in lung-function predictions. N Engl J Med 2010; 363:321–330 33.Scanlon PD, Shriver MD: “Race correction” in pulmonary-function testing. N Engl J Med 2010; 363:385–386 34.O’Brien JM Jr, Welsh CH, Fish RH, et al; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome Network: Excess body weight is not independently associated with outcome in mechanically ventilated patients with acute lung injury. Ann Intern Med 2004; 140:338–345 35. Hager DN, Krishnan JA, Hayden DL, et al: Tidal volume reduction in patients with acute lung injury when plateau pressures are not high. Am J Respir Crit Care Med 2005; 172:1241–1245 36. Mattingley JS, Holets SR, Oeckler RA, et al: Sizing the lung of mechanically ventilated patients. Crit Care 2011; 15:R60 37.Terragni PP, Rosboch G, Tealdi A, et al: Tidal hyperinflation during low tidal volume ventilation in acute respiratory distress syndrome. Am J Respir Crit Care Med 2007;175:160–166 March 2013 • Volume 41 • Number 3 Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited The n e w e ng l a n d j o u r na l of m e dic i n e original article High-Frequency Oscillation for Acute Respiratory Distress Syndrome Duncan Young, D.M., Sarah E. Lamb, D.Phil., Sanjoy Shah, M.D., Iain MacKenzie, M.D., William Tunnicliffe, M.Sc., Ranjit Lall, Ph.D., Kathy Rowan, D.Phil., and Brian H. Cuthbertson, M.D., for the OSCAR Study Group* A bs t r ac t Background From John Radcliffe Hospital (D.Y.) and the University of Oxford (D.Y., S.E.L.), Oxford; the Bristol Royal Infirmary, Bristol (S.S.); Queen Elizabeth Hospital, Birmingham (I.M., W.T.); Warwick Clinical Trials Unit, University of Warwick, Warwick (R.L.); and Intensive Care National Audit and Research Centre, London (K.R.) — all in the United Kingdom; and the Department of Critical Care Medicine, Sunnybrook Health Sciences Centre and the Department of Anesthesia, University of Toronto, Toronto (B.H.C.). Address reprint requests to Dr. Young at the Adult Intensive Care Unit, John Radcliffe Hospital, Headley Way, Oxford OX3 9DU, United Kingdom, or at duncan.young@ nda.ox.ac.uk. *Investigators in the Oscillation in ARDS (OSCAR) study group are listed in the Supplementary Appendix, available at NEJM.org. This article was published on January 22, 2013, and updated on January 31, 2013, at NEJM.org. N Engl J Med 2013;368:806-13. DOI: 10.1056/NEJMoa1215716 Copyright © 2013 Massachusetts Medical Society. 806 Patients with the acute respiratory distress syndrome (ARDS) require mechanical ventilation to maintain arterial oxygenation, but this treatment may produce secondary lung injury. High-frequency oscillatory ventilation (HFOV) may reduce this secondary damage. Methods In a multicenter study, we randomly assigned adults requiring mechanical ventilation for ARDS to undergo either HFOV with a Novalung R100 ventilator (Metran) or usual ventilatory care. All the patients had a ratio of the partial pressure of arterial oxygen (Pao2) to the fraction of inspired oxygen (Fio2) of 200 mm Hg (26.7 kPa) or less and an expected duration of ventilation of at least 2 days. The primary outcome was all-cause mortality 30 days after randomization. Results There was no significant between-group difference in the primary outcome, which occurred in 166 of 398 patients (41.7%) in the HFOV group and 163 of 397 patients (41.1%) in the conventional-ventilation group (P = 0.85 by the chi-square test). After adjustment for study center, sex, score on the Acute Physiology and Chronic Health Evaluation (APACHE) II, and the initial Pao2:Fio2 ratio, the odds ratio for survival in the conventional-ventilation group was 1.03 (95% confidence interval, 0.75 to 1.40; P = 0.87 by logistic regression). Conclusions The use of HFOV had no significant effect on 30-day mortality in patients undergoing mechanical ventilation for ARDS. (Funded by the National Institute for Health Research Health Technology Assessment Programme; OSCAR Current Controlled Trials number, ISRCTN10416500.) n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation for Respir atory Distress T he acute respiratory distress syndrome (ARDS) is a severe, diffuse inflammatory lung condition caused by a range of acute illnesses. Mortality in affected patients is high,1 and survivors may have functional limitations for years.2,3 Although mechanical ventilation can initially be lifesaving in patients with ARDS, it can also further injure the patients’ lungs and contribute to death.4 High-frequency oscillatory ventilation (HFOV) was first used experimentally in the 1970s to minimize the hemodynamic effects of mechanical ventilation.5 Patients’ lungs are held inflated to maintain oxygenation, and carbon dioxide is cleared by small volumes of gas moved in and out of the respiratory system at 3 to 15 Hz. This action is thought to minimize the repeated process of opening and collapsing of lung units that causes the secondary lung damage during mechanical ventilation. On the basis of small trials with outdated controls6 and the commercial availability of HFOV equipment, many clinicians use HFOV for patients who have hypoxemia despite the use of standard approaches for improving arterial oxygenation. The increasing use of HFOV in the absence of good evidence of effectiveness led the National Institute for Health Research in the United Kingdom to commission a study to determine the effectiveness of HFOV as a treatment for ARDS. manuscript preparation. The study was approved by national ethics review committees and research governance departments at each center. Patients or their representatives provided written informed consent. Patients Patients who were undergoing mechanical ventilation were eligible for the study if they had a ratio of the partial pressure of arterial oxygen (Pao2) to the fraction of inspired oxygen (Fio2) of 200 mm Hg (26.7 kPa) or less while receiving a positive end-expiratory pressure (PEEP) of 5 cm of water or greater, if bilateral pulmonary infiltrates were visible on chest radiography without evidence of left atrial hypertension, and if they were expected to require at least 2 more days of mechanical ventilation. Patients were excluded if they had undergone mechanical ventilation for 7 or more days, if they were under the age of 16 years, if they weighed less than 35 kg, if they were participating in other interventional studies, if they had lung disease characterized by airway narrowing or air trapping, or if they had undergone recent lung surgery. An independent telephone randomization system assigned patients to either HFOV or conventional mechanical ventilation in a 1:1 ratio. Randomization was by permuted block stratified according to study center, Pao2:Fio2 ratio (≤113 mm Hg [15 kPa] or >113 mm Hg), age Me thods (≤55 years or >55 years), and sex. Each center Study Design had one HFOV ventilator, so recruitment could We conducted a randomized, controlled trial of not take place if the device was in use for anHFOV, as compared with conventional mechani- other study patient. cal ventilation. Patients were recruited from adult general intensive care units (ICUs) in 12 university Study Treatments hospitals, 4 university-affiliated hospitals, and Patients in the HFOV group were treated with the 13 district general hospitals in England, Wales, use of a Novalung R100 ventilator (Metran)7 until and Scotland. Three hospitals had previous expe- the start of weaning. The initial settings were a rience with HFOV with the use of SensorMedics ventilation frequency of 10 Hz, a mean airway 3100B ventilators (CareFusion), and the remain- pressure of 5 cm of water above the plateau airder had limited experience (in 6 hospitals) or no way pressure at enrollment, bias flow rate of experience (in 20 hospitals) with HFOV. Details 20 liters per minute, a cycle volume of 100 ml regarding HFOV training are provided in the Sup- (the volume of gas used to move the oscillating plementary Appendix, available with the full text diaphragm; the tidal volume delivered to the alof this article at NEJM.org. The full protocol is veoli is a fraction of this volume), and an inspired also available at NEJM.org. oxygen fraction of 1. This ventilator has a fixed The ventilators were purchased from Inspira- 1:1 inspiratory:expiratory time ratio. tion Healthcare. The company had no role in the Two algorithms were used to determine changstudy design, data acquisition, data analysis, or es in HFOV settings (for details, see the Supplen engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 807 The n e w e ng l a n d j o u r na l mentary Appendix). The partial pressure of arterial carbon dioxide (Paco2) was controlled to maintain an arterial pH above 7.25 by increasing the cycle volume to the maximum at each frequency. If this was insufficient, the frequency was reduced by 1 Hz. If the minimum frequency (5 Hz) was reached, the on-call study clinician would suggest other measures to control the Paco2 level (see the Supplementary Appendix). The Pao2 level was maintained between 60 mm Hg and 75 mm Hg (8 kPa to 10 kPa). Hypoxemia was treated by increasing the mean airway pressure and then by increasing the Fio2 level. If a patient reached a mean airway pressure of 24 cm of water, at an Fio2 level of 0.4 or less, with a Pao2 level of 60 mm Hg or greater, for 12 hours or more, he or she was switched to pressure-controlled ventilation for weaning from mechanical ventilation, since there was no facility to accommodate patients’ spontaneous respiratory efforts during HFOV. Patients could be restarted on HFOV up to 2 days after the start of weaning. Patients in the conventional-ventilation group were treated according to local practice in the participating ICUs. The participating units were encouraged to use pressure-controlled ventilation at 6 to 8 ml per kilogram of ideal body weight and to use the combinations of PEEP and Fio2 values that were used in the Acute Respiratory Distress Syndrome Network study.4 All other treatment was determined by the patients’ physicians on the basis of assessment of clinical need. Data Collection At the time of enrollment, we recorded data with respect to the patients’ demographic characteristics, ventilation before enrollment, physiology and other data required to calculate the score on the Acute Physiology and Chronic Health Evaluation (APACHE) II, coexisting medical conditions, the use of sedatives and muscle relaxants, and ventilator settings. For each day that a patient was treated in the ICU, we recorded data with respect to the use of antibiotics, sedatives, and muscle relaxants during the previous day or since enrollment on the first day. Data regarding support for respiratory and cardiovascular organ systems were recorded daily during treatment in the ICU with the use of the United Kingdom’s criticalcare minimum data set.8 Vital status at 30 days was known for all patients, but causes of death were not recorded. 808 of m e dic i n e Outcomes The primary outcome, vital status at 30 days, was obtained from hospital records and verified with the use of a national database. Secondary outcomes were all-cause mortality at the time of discharge from the ICU and the hospital, the duration of mechanical ventilation, and the use of antimicrobial, sedative, vasoactive, and neuromuscular-blocking drugs. We recorded the duration of treatment in both the ICU and the hospital. Statistical Analysis Recruitment-rate estimates and sample-size calculations were performed after a systematic review of the incidence and outcome of ARDS, national audits in the United Kingdom, and two randomized, controlled trials of HFOV.9,10 We determined that the enrollment of 503 patients per study group would provide a power of 80% to identify a change of 9 percentage points in an estimated rate of death of 45% in the control group at a P value of 0.05. At a planned interim review, the sample size was revised to 401 patients per group on the basis of accumulated mortality data in the control group and an effect size of 10 percentage points (80% power at P = 0.05). All analyses were conducted on an intentionto-treat basis. Three planned interim analyses were conducted by an independent data and safety monitoring committee after the recruitment of 100, 340, and 640 patients. Formal stopping rules were not specified. Instead, the committee assessed whether the randomized comparisons provided “proof beyond reasonable doubt” that for all or some patients the treatment was clearly indicated or clearly contraindicated and provided evidence that might reasonably be expected to influence future patient treatment. We used chi-square tests to compare betweengroup rates of death at 30 days and among patients in ICU and hospital settings. We performed an analysis of mortality after adjustment for study center, sex, Pao2:Fio2 ratio, and APACHE II score using logistic regression. Continuous variables were compared with the use of Student’s t-tests. Since both the rate and timing of death were similar in the two study groups, data for survivors and those for nonsurvivors were not analyzed separately. All P values are two-sided. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation for Respir atory Distress R e sult s Trial Progression and Recruitment We trained 2306 intensive care nurses, medical staff, physiotherapists, and technicians in 198 face-to-face training sessions. Patients were recruited from December 7, 2007, until the end of July 2012. Of the 2769 patients who were screened, 795 (28.7%) underwent randomization (Fig. 1). The study had 968 ICU-months of recruitment averaging 0.82 patients per ICUmonth. (A graphical summary of recruitment is provided in Figure S3 in the Supplementary Appendix.) The baseline characteristics of the patients at randomization were similar in the two study groups (Table 1). Ventilation HFOV was used for a median of 3 days (interquartile range, 2 to 5) in 388 patients. The longest initial period of receipt of HFOV was 24 days. Figure 2 shows the use of HFOV in the two study groups. Ten patients in the conventional-ventilation group underwent HFOV at some point during the study period, and 10 patients who were assigned to the HFOV group never received this treatment. Table 2 shows ventilatory and other variables for the first 3 days of the study period. Neuromuscular-blocking drugs were used for a mean (±SD) of 2.0±3.4 days in the conventionalventilation group and for 2.5±3.5 days in the HFOV group (P = 0.02). Sedative drugs were used for 8.5±6.9 days in the conventional-ventilation group and for 9.4±7.2 days in the HFOV group (P = 0.07). The patients had 17.6±8.8 ventilator-free days in the conventional-ventilation group and 17.1±8.6 ventilator-free days in the HFOV group (P = 0.42). Mechanical ventilation (including HFOV but excluding noninvasive ventilation) was used for 14.1±13.4 days in the conventional-ventilation group and 14.9±13.3 days in the HFOV group (P = 0.41). the odds ratio for survival in the conventionalventilation group, as compared with the HFOV group, was 1.03 (95% CI, 0.75 to 1.40; P = 0.87 by logistic regression) (Fig. 3). Similar proportions of patients died at each time point in each group. The rates of death at first discharge from the ICU were 42.1% in the conventional-ventilation group and 44.1% in the HFOV group, for an absolute difference of 2.0 percentage points (P = 0.57). At first hospital discharge, 48.4% of patients in the conventional-ventilation group and 50.1% of those in the HFOV group had died, for an absolute difference of 1.7 percentage points (P = 0.62). Data are not provided with respect to the duration of care for survivors and nonsurvivors, since the proportions of patients who died in each study group over time were nearly identical. The total duration of ICU stay was 16.1±15.2 days in the conventional-ventilation group and 17.6±16.6 2769 Patients were screened 1974 Were excluded 913 (33.0%) Did not meet inclusion criteria or met exclusion criteria 134 (4.8%) Did not have access to ventilator 282 (10.2%) Did not provide consent 645 (23.3%) Had other reasons 795 (28.7%) Underwent randomization 397 Were assigned to receive conventional intervention 397 (100%) Received assigned intervention 10 (2.5%) Received HFOV at some point after randomization 398 Were assigned to receive HFOV 388 (97.5%) Received assigned intervention 10 Did not receive assigned intervention 3 Died 4 Had ventilator malfunction 1 Recovered 2 Were treated by clinician who did not comply with assigned therapy 397 (100%) Were included in primary analysis 398 (100%) Were included in primary analysis Outcomes The primary outcome occurred in 166 of 398 patients (41.7%) in the HFOV group and in 163 of 397 patients (41.1%) in the conventional-ventilation group (P = 0.85), for an absolute difference of 0.6 percentage points (95% confidence interval [CI], −6.1 to 7.5). After adjustment for study center, sex, APACHE II score, and Pao2:Fio2 ratio, Figure 1. Enrollment and Outcomes. HFOV denotes high-frequency oscillatory ventilation. n engl j med 368;9 nejm.org february 28, 2013 4 col 22p3 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 809 The n e w e ng l a n d j o u r na l of m e dic i n e Table 1. Baseline Characteristics of the Patients.* Conventional Ventilation (N = 397) HFOV (N = 398) All Patients (N = 795) 55.9±16.2 54.9±18.8 55.4±16.2 Male sex — no. (%) 239 (60.2) 256 (64.3) 495 (62.3) APACHE II score† 21.7±6.1 21.8±6.0 21.8±6.1 Probability of in-hospital death (as calculated from APACHE II score) 0.43±0.19 0.44±0.19 0.43±0.19 PaO2:FiO2 ratio — mm Hg 113±38 113±37 113±38 Exhaled tidal volume — ml 505±173 541±271 523±228 Exhaled tidal volume — ml/kg of ideal body weight‡ 8.3±3.5 8.7±3.5 8.5±3.9 10.17±3.46 10.41±3.25 10.29±3.35 11.3±3.3 11.4±3.5 11.4±3.4 2.1±2.1 2.2±2.3 2.2±2.2 304 (76.6) 302 (75.9) 606 (76.2) Characteristic Age — yr Exhaled minute ventilation — liters/min Positive end-expiratory pressure — cm of water Duration of mechanical ventilation before randomization — days Pulmonary cause of ARDS — no. (%) *Plus–minus values are means ±SD. There was no significant difference between groups except for exhaled tidal volume (P = 0.04). ARDS denotes acute respiratory distress syndrome, FiO2 fraction of inspired oxygen, and PaO2 partial pressure of arterial oxygen. †Scores on the Acute Physiology and Chronic Health Evaluation (APACHE) II scale range from 0 to 71, with higher scores indicating more severe disease.11 ‡Ideal body weight was calculated as 2.3 kg for each inch of height above 60 in. added to 50 kg for men or 45.5 kg for women. days in the HFOV group (P = 0.18); the total durations of hospital stay were 33.1±44.3 days and 33.9±41.6 days, respectively (P = 0.79). As of October 1, 2012, the date that the database was closed, 7 patients remained in acute hospital care. Patients received antimicrobial drugs for 12.4±10.3 days in the conventional-ventilation group and for 12.8±12.0 days in the HFOV group (P = 0.56); 67.5% and 64.4% of these drugs, respectively, were administered to treat pulmonary infections. There was no significant difference in the number of days on which patients received inotropic agents or pressor infusions, with 2.8±5.6 days in the conventional-ventilation group and 2.9±4.5 days in the HFOV group (P = 0.74). Discussion This study, which was designed to help practitioners choose between options for care, met 7 of the 10 criteria of the Pragmatic–Explanatory Continuum Indicator Summary (PRECIS).12 The results were not totally pragmatic because of the tight protocol-specified restrictions on the use of HFOV, protocol-compliance monitoring, and additional follow-up. We found no significant be810 tween-group difference in the primary outcome of mortality up to 30 days after randomization. Our estimate of the 95% confidence interval for the treatment excludes the treatment effect we specified in both the initial and revised samplesize estimates. Since data collection is ongoing, we cannot yet report the longer-term outcomes (including survival and health-related quality of life). We recruited patients with moderate-tosevere ARDS, with an average Pao2:Fio2 ratio of 113 mm Hg (15.1 kPa). The study-entry criterion was a Pao2:Fio2 ratio of less than 200 mm Hg (26.7 kPa), which was in line with the agreed definition of ARDS,13 but the additional requirement of a further 48 hours or more of mechanical ventilation may have excluded milder cases of ARDS. The average Pao2:Fio2 ratio is nearly identical to the mean of 112 mm Hg reported in the recent systematic review of HFOV6 and is similar to the mean values reported in studies of other treatments for ARDS.14-16 The patients had a high severity of illness, as evidenced by the APACHE II scores, which also were nearly identical to those reported in the two other multicenter studies of HFOV in adults.9,10 Thus, we appear to have recruited patients who were similar to those in previous randomized, controlled trials of HFOV. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation for Respir atory Distress 100 Patients Undergoing HFOV (%) HFOV improved oxygenation as expected. The Paco2 value increased as a predicted result of the HFOV treatment algorithms, resulting in a modest respiratory acidosis. A similar effect was seen in the larger of the two reported studies of HFOV in adults10 but not in the smaller study9 or the meta-analysis.6 The conventional-ventilation group was treated with tidal volumes at the upper end of the accepted range of 6 to 8 ml per kilogram of ideal body weight. The use of HFOV was initially associated with an increased use of neuromuscular-blocking drugs, probably because the R100 ventilator has no facility to allow the patient to breathe spontaneously. HFOV has been reported to cause a reduction in cardiac output,17 but as indicated by the use of vasoactive and inotropic drugs, that did not occur in this study. Our results are at variance with the latest metaanalysis of HFOV,6 which showed a reduced risk of death (risk ratio, 0.77; 95% CI, 0.61 to 0.98), as compared with conventional ventilation. This may be simply that our study recruited more than 90 80 70 HFOV 60 50 40 30 20 Conventional ventilation 10 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Day Figure 2. Proportions of Patients Undergoing High-Frequency Oscillatory Ventilation (HFOV) during the First 30 Days, According to Study Group. Shown are the percentages of patients in each study group who underwent HFOV. Ten patients in the conventional-ventilation group underwent HFOV at some point during their treatment, and 10 patients who were assigned to the HFOV group never received this treatment. Day 0 is the day of randomization, and subsequent determinations were made at 8 a.m. each day. Table 2. Ventilatory Variables during the First 3 Study Days.* Variable Day 1 Day 2 Day 3 HFOV Conventional Ventilation HFOV Conventional Ventilation HFOV Conventional Ventilation 370 392 326 374 240 348 Mean airway pressure (HFOV) or plateau pressure (conventional ventilation) — cm of water 26.9±6.2 30.9±11.0 25.3±5.5 29.5±10.7 25.1±5.4 28.5±11.2 Total respiratory frequency — Hz (HFOV) or breaths/min (conventional ventilation) 7.8±1.8 21.7±8.4 7.5±1.8 22.7±9.0 7.2±1.8 23.3±8.2 Cycle volume (HFOV) or tidal volume (conventional ventilation) — ml (HFOV) or ml/kg of ideal body weight (conventional ventilation) 213±72 8.3±2.9 228±75 8.2±2.5 240±75 8.3±3.0 NA 11.4±3.6 NA 11.0±3.6 NA 10.5±3.7 PaO2:FiO2 ratio — mm Hg 192±77 154±61 212±69 163±66 217±69 166±63 PaCO2 — mm Hg 55±17 50±19 56±16 49±13 56±17 48±13 7.30±0.10 7.35±0.10 7.32±0.09 7.37±0.10 7.34±0.10 7.39±0.09 Neuromuscular-blocking agent 209 (52.5) 165 (41.6) 147 (36.9) 115 (29.0) 110 (27.6) 77 (19.4) Vasoactive or inotropic agent 173 (43.5) 177 (44.6) 158 (40.0) 146 (36.8) 126 (31.7) 112 (28.2) Sedative agent 390 (98.0) 388 (97.7) 371 (93.2) 363 (91.4) 341 (85.7) 335 (84.4) No. of patients Positive end-expiratory pressure — cm of water (conventional ventilation only) Arterial pH Medication use — no. (%)† *Measurements were taken at 8 a.m. Day 1 values were recorded the morning after recruitment. The values for high-frequency oscillatory ventilation (HFOV) are only for patients who actually underwent the treatment. The values for conventional ventilation are for all patients assigned to receive conventional ventilation who were receiving any mechanical ventilation. NA denotes not applicable, and PaCO2 partial pressure of arterial carbon dioxide. †Percentages were calculated on the basis of the 398 patients in the HFOV group and the 397 patients in the conventional-ventilation group who underwent randomization. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 811 The n e w e ng l a n d j o u r na l Probability of Survival 1.00 Conventional ventilation 0.75 HFOV 0.50 0.25 0.00 0 5 10 15 20 25 30 Days No. at Risk Conventional ventilation HFOV 397 351 312 281 259 243 236 398 349 311 280 253 241 233 Figure 3. Kaplan–Meier Survival Estimates during the First 30 Study Days. twice the number of patients who were included in the meta-analysis. Adding our results to the meta-analysis changes the estimated risk ratio from the pooled studies to 0.90 (95% CI, 0.76 to 1.07), indicating no significant benefit for HFOV. The use of HFOV is a lung-protection strategy, which may be ineffective if it is used for too brief a period. We used it up to the point at which the HFOV design hindered weaning. In the two other multicenter studies of HFOV in adults,9,10 the duration of ventilation was not reported. In the HFOV group in our study, we used the Novalung R100 ventilator, a device that had not been used before in clinical trials. To date, all studies have used the SensorMedics 3100B ventilator, a device that has an electromechanically driven diaphragm, which normally oscillates with an inspiratory:expiratory time ratio of 1:2. The R100 ventilator uses a pneumatically driven diaphragm with a fixed 1:1 ratio. It seems unlikely that these differences would explain the difference in mortality between our study and the pooled results of studies to date, but it remains a possibility. We recruited patients who met the definition of ARDS13 that was in place at the time the study was planned, and the entry criteria match the 812 of m e dic i n e “moderate” and “severe” categories in the recently revised definition.18 The study has good internal and external validity. Bias was minimized by using centers with equipoise, by concealing treatment assignments before randomization, by concealing interim analyses from all study investigators except for the data and safety monitoring committee, and by using an analysis plan that was agreed on before study closure. There was no loss to follow-up, crossovers were minimal, and the study recruited 99.1% of the planned sample size. External validity was maintained by using a large number of different-sized ICUs spread across the United Kingdom. Most of the centers in this trial were inexperienced with the intervention at the start, but this was unavoidable, since few centers in the United Kingdom have experience with the use of HFOV. We invested heavily in training at each study center. The consent-refusal rate was low. Our report coincides with the publication in the Journal of the results of a large multicenter efficacy study of HFOV, the Oscillation for Acute Respiratory Distress Syndrome Treated Early (OSCILLATE) trial.19 This study showed 47% mortality in the HFOV group and 35% in the control group. The patients who were recruited in both studies were broadly similar. The OSCILLATE trial used the 3100B ventilator, maneuvers to reexpand collapsed areas of lung before HFOV, and a protocol-specified high-PEEP strategy for conventional ventilation. In that study, the patients undergoing HFOV required more inotropic and pressor support than did those in the control group. It is possible that the HFOV strategy used in the OSCILLATE trial was injurious, but the low mortality in their control group also raises the possibility that the control treatment was a very effective ventilation strategy in patients with ARDS. In conclusion, in a large effectiveness study, we were unable to find any benefit or harm from the use of HFOV in adult patients with ARDS. We recommend that this mode of ventilation not be used for routine care. Supported by the National Institute for Health Research Health Technology Assessment Programme (project number 06/04/01). Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation for Respir atory Distress References 1. Erickson SE, Martin GS, Davis JL, Matthay MA, Eisner MD. Recent trends in acute lung injury mortality: 1996-2005. Crit Care Med 2009;37:1574-9. 2. Herridge MS, Tansey CM, Matté A, et al. Functional disability 5 years after acute respiratory distress syndrome. N Engl J Med 2011;364:1293-304. 3. Herridge MS, Cheung AM, Tansey CM, et al. One-year outcomes in survivors of the acute respiratory distress syndrome. N Engl J Med 2003;348:683-93. 4. 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;342:1301-8. 5. Lunkenheimer PP, Rafflenbeul W, Keller H, Frank I, Dickhut HH, Fuhrmann C. Application of transtracheal pressure oscillations as a modification of “diffusing respiration.” Br J Anaesth 1972;44: 627. 6. Sud S, Sud M, Friedrich JO, et al. High frequency oscillation in patients with acute lung injury and acute respiratory distress syndrome (ARDS): systematic review and meta-analysis. BMJ 2010;340: c2327. 7. Metran. Product information: R100. 2012 (http://www.metran.co.jp/en/product/ r100/index.html). 8. Critical care minimum dataset, version 8.0, revised March 2010 (http://www.isb .nhs.uk/documents/isb-0153/amd-812010/0153812010spec.pdf). 9. Bollen CW, van Well GT, Sherry T, et al. High frequency oscillation ventilation compared with conventional mechanical ventilation in adult respiratory distress syndrome: a randomized controlled trial [ISRCTN24242669]. Crit Care 2005;9: R430-R439. 10. Derdak S, Mehta S, Stewart TE, et al. High-frequency oscillatory ventilation for acute respiratory distress syndrome in adults: a randomized, controlled trial. Am J Respir Crit Care Med 2002;166: 801-8. 11. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 12. Thorpe KE, Zwarenstein M, Oxman AD, et al. A Pragmatic-Explanatory Continuum Indicator Summary (PRECIS): a tool to help trial designers. CMAJ 2009; 180:E47-E57. 13. Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818-24. 14. Villar J, Pérez-Méndez L, Basaldúa S, et al. A risk tertiles model for predicting mortality in patients with acute respiratory distress syndrome: age, plateau pressure, and PaO2/FIO2 at ARDS onset can predict mortality. Respir Care 2011;56: 420-8. 15. Villar J, Blanco J, Añón JM, et al. The ALIEN study: incidence and outcome of acute respiratory distress syndrome in the era of lung protective ventilation. Intensive Care Med 2011;37:1932-41. [Erratum, Intensive Care Med 2011;37:1942.] 16. Perkins GD, Gates S, Lamb SE, McCabe C, Young D, Gao F. Beta Agonist Lung Injury TrIal-2 (BALTI-2) trial protocol: a randomised, double-blind, placebocontrolled of intravenous infusion of salbutamol in the acute respiratory distress syndrome. Trials 2011;12:113. 17. Mehta S, Lapinsky SE, Hallett DC, et al. Prospective trial of high-frequency oscillation in adults with acute respiratory distress syndrome. Crit Care Med 2001; 29:1360-9. 18. Ranieri VM, Rubenfeld GD, Thompson BT, et al. Acute respiratory distress syndrome: the Berlin definition. JAMA 2012; 307:2526-33. 19. Ferguson ND, Cook DJ, Guyatt GH, et al. High-frequency oscillation in early acute respiratory distress syndrome. N Engl J Med 2013;368:795-805. Copyright © 2013 Massachusetts Medical Society. an nejm app for iphone The NEJM Image Challenge app brings a popular online feature to the smartphone. Optimized for viewing on the iPhone and iPod Touch, the Image Challenge app lets you test your diagnostic skills anytime, anywhere. The Image Challenge app randomly selects from 300 challenging clinical photos published in NEJM, with a new image added each week. View an image, choose your answer, get immediate feedback, and see how others answered. The Image Challenge app is available at the iTunes App Store. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 813 new england journal of medicine The established in 1812 february 28, 2013 vol. 368 no. 9 High-Frequency Oscillation in Early Acute Respiratory Distress Syndrome Niall D. Ferguson, M.D., Deborah J. Cook, M.D., Gordon H. Guyatt, M.D., Sangeeta Mehta, M.D., Lori Hand, R.R.T., Peggy Austin, C.C.R.A., Qi Zhou, Ph.D., Andrea Matte, R.R.T., Stephen D. Walter, Ph.D., Francois Lamontagne, M.D., John T. Granton, M.D., Yaseen M. Arabi, M.D., Alejandro C. Arroliga, M.D., Thomas E. Stewart, M.D., Arthur S. Slutsky, M.D., and Maureen O. Meade, M.D., for the OSCILLATE Trial Investigators and the Canadian Critical Care Trials Group* A bs t r ac t Background Previous trials suggesting that high-frequency oscillatory ventilation (HFOV) reduced mortality among adults with the acute respiratory distress syndrome (ARDS) were limited by the use of outdated comparator ventilation strategies and small sample sizes. Methods In a multicenter, randomized, controlled trial conducted at 39 intensive care units in five countries, we randomly assigned adults with new-onset, moderate-to-severe ARDS to HFOV targeting lung recruitment or to a control ventilation strategy targeting lung recruitment with the use of low tidal volumes and high positive end-expiratory pressure. The primary outcome was the rate of in-hospital death from any cause. Results On the recommendation of the data monitoring committee, we stopped the trial after 548 of a planned 1200 patients had undergone randomization. The two study groups were well matched at baseline. The HFOV group underwent HFOV for a median of 3 days (interquartile range, 2 to 8); in addition, 34 of 273 patients (12%) in the control group received HFOV for refractory hypoxemia. In-hospital mortality was 47% in the HFOV group, as compared with 35% in the control group (relative risk of death with HFOV, 1.33; 95% confidence interval, 1.09 to 1.64; P = 0.005). This finding was independent of baseline abnormalities in oxygenation or respiratory compliance. Patients in the HFOV group received higher doses of midazolam than did patients in the control group (199 mg per day [interquartile range, 100 to 382] vs. 141 mg per day [interquartile range, 68 to 240], P<0.001), and more patients in the HFOV group than in the control group received neuromuscular blockers (83% vs. 68%, P<0.001). In addition, more patients in the HFOV group received vasoactive drugs (91% vs. 84%, P = 0.01) and received them for a longer period than did patients in the control group (5 days vs. 3 days, P = 0.01). Conclusions In adults with moderate-to-severe ARDS, early application of HFOV, as compared with a ventilation strategy of low tidal volume and high positive end-expiratory pressure, does not reduce, and may increase, in-hospital mortality. (Funded by the Canadian Institutes of Health Research; Current Controlled Trials numbers, ISRCTN42992782 and ISRCTN87124254, and ClinicalTrials.gov numbers, NCT00474656 and NCT01506401.) From the Interdepartmental Division of Critical Care Medicine (N.D.F., S.M., J.T.G., T.E.S., A.S.S.), Departments of Medicine and Physiology (N.D.F.), the Department of Medicine, Division of Respirology (S.M., J.T.G., T.E.S.), and the Departments of Medicine, Surgery, and Biomedical Engineering (A.S.S.), University of Toronto, University Health Network and Mount Sinai Hospital (N.D.F., J.T.G.), Mount Sinai Hospital (S.M., T.E.S.), St. Michael’s Hospital (A.S.S.), and the Critical Care Program, University Health Network (A.M.), Toronto; the Interdepartmental Division of Critical Care, Hamilton Health Sciences (D.J.C., M.O.M.), the Departments of Medicine and Clinical Epidemiology and Biostatistics (D.J.C., G.H.G., S.D.W., M.O.M.), and the CLARITY Research Centre (D.J.C., G.H.G., S.D.W., M.O.M., L.H., P.A., Q.Z.), McMaster University, Hamilton, ON; and the Centre de Recherche Clinique Étienne-Le Bel, Université de Sherbrooke, Sherbrooke, QC (F.L.) — all in Canada; the Intensive Care Department, King Saud bin Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia (Y.M.A.); and the Department of Medicine, Scott and White Healthcare and Texas A&M Health Science Center College of Medicine, Temple (A.C.A.). Address reprint requests to Dr. Meade at 1280 Main St. W., Hamilton, ON L8N 3Z5, Canada, or at [email protected]. *A complete list of the investigators in the Oscillation for Acute Respiratory Distress Syndrome Treated Early (OSCILLATE) trial is provided in the Supplementary Appendix, available at NEJM.org. This article was published on January 22, 2013, at NEJM.org. N Engl J Med 2013;368:795-805. DOI: 10.1056/NEJMoa1215554 Copyright © 2013 Massachusetts Medical Society. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 795 The n e w e ng l a n d j o u r na l T he acute respiratory distress syndrome (ARDS) is a common complication of critical illness.1,2 Mortality is high, and survivors often have long-term complications.3,4 Although mechanical ventilation is life-sustaining for patients with ARDS, it can perpetuate lung injury. Basic research suggests that repetitive overstretching or collapse of lung units with each respiratory cycle can generate local and systemic inflammation, contributing to multiorgan failure and death.5 Consistent with these findings are data from clinical trials that support the use of smaller tidal volumes (6 vs. 12 ml per kilogram of predicted body weight)6 and higher levels of positive end-expiratory pressure (PEEP).7-10 Mortality remains high, however, and additional therapies are needed to protect the lung in cases of severe ARDS.11,12 One such approach is high-frequency oscillatory ventilation (HFOV), which delivers very small tidal volumes (approximately 1 to 2 ml per kilogram13) at very high rates (3 to 15 breaths per second).14-19 Previous randomized trials of the use of HFOV in adults with ARDS have suggested that this strategy results in improvements in oxygenation and survival, but the trials were limited by small sample sizes and outdated ventilation strategies for the control group.20-22 Consequently, despite the frequent use of HFOV in patients who do not have an adequate response to conventional mechanical ventilation and the increased use of HFOV earlier in the course of the disease, this approach remains an unproven therapy for adults with ARDS.23-26 We therefore compared HFOV with a conventional ventilation strategy that used low tidal volumes and high levels of PEEP in patients with new-onset, moderate-to-severe ARDS. Me thods Study Oversight For the pilot phase of the study, we enrolled patients at 11 centers in Canada and 1 in Saudi Arabia from July 2007 through June 2008; for the main trial, we enrolled patients at the same centers and at an additional 27 centers in Canada, the United States, Saudi Arabia, Chile, and India from July 2009 through August 2012 (see the Supplementary Appendix, available with the full text of this article at NEJM.org). The trial protocol, which is available at NEJM.org, was approved by the research ethics board at each participating site. 796 of m e dic i n e The first and last author vouch for the accuracy and completeness of the reported data and for the fidelity of this report to the study protocol. For HFOV, we used the SensorMedics 3100B High-Frequency Oscillatory Ventilator (CareFusion); the manufacturer loaned nine ventilators and provided technical support but had no role in the design of the study, the collection or analysis of the data, or the preparation of the manuscript. Patients Patients were eligible for inclusion if they had had an onset of pulmonary symptoms within the previous 2 weeks, had undergone tracheal intubation, had hypoxemia (defined as a ratio of the partial pressure of arterial oxygen [Pao2] to the fraction of inspired oxygen [Fio2] of ≤200, with an Fio2 of ≥0.5), and had bilateral air-space opacities on chest radiography. Patients were excluded if they had hypoxemia primarily related to left atrial hypertension, suspected vasculitic pulmonary hemorrhage, neuromuscular disorders that are known to prolong the need for mechanical ventilation, severe chronic respiratory disease, or preexisting conditions with an expected 6-month mortality exceeding 50%; if they were at risk for intracranial hypertension; if there was a lack of commitment to life support; if the expected duration of mechanical ventilation was less than 48 hours; if they were younger than 16 years of age or older than 85 years of age; or if their weight was less than 35 kg or more than 1 kg per centimeter of height. We did not enroll patients who had already met the eligibility criteria for more than 72 hours, those who were already receiving HFOV, or those whose physicians declined to enroll them. After enrollment, standardized ventilator settings were used for all the patients: pressurecontrol mode, a tidal volume of 6 ml per kilogram, and an Fio2 of 0.60 with a PEEP level of 10 cm of water or higher if needed for oxygenation. After 30 minutes, if the Pao2:Fio2 ratio remained at 200 or lower, patients underwent randomization; otherwise the standardized ventilator settings were maintained, and the patients were reassessed at least once daily for up to 72 hours. Eligible patients were randomly assigned in a 1:1 ratio to the HFOV group or to the conventional-ventilation group. Randomization was performed in undisclosed block sizes of 2 and 4 with the use of a central Web-based n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation in Early ARDS Table 1. Ventilator Protocols.* Component Variable Ventilator mode HFOV Control Ventilation High-frequency oscillatory ventilation Pressure control Tidal volume target (ml/kg of predicted body weight) NA 6 Tidal volume range (ml/kg of predicted body weight) NA 4–8 Plateau airway pressure (cm of water) NA ≤35 Positive end-expiratory pressure (cm of water) NA Adjusted according to oxygenation† Adjusted according to oxygenation† Measured but not adjusted 3–12 Hz ≤35 breaths/min Mean airway pressure (cm of water) Respiratory frequency Pressure amplitude target (cm of water) 90 NA Partial pressure of arterial oxygen (mm Hg) 55–80 55–80 Oxygen saturation by pulse oximetry (%) 88–93 88–93 7.25–7.35 7.30–7.45 Ratio of inspiratory-to-expiratory time Arterial blood pH 1:2 1:1–1:3 Recruitment maneuvers Yes Yes *The full version of the study protocol is available at NEJM.org. HFOV denotes high-frequency oscillatory ventilation, and NA not applicable. †For more information on the protocol for adjustment, see Table 2. randomization system, stratified according to center. All patients or their legal surrogates provided written informed consent for participation in the study. Table 2. Usual Combinations of the Fraction of Inspired Oxygen (Fio2) and Positive End-Expiratory Pressure (PEEP) or Mean Airway Pressure Used to Adjust Ventilators. HFOV HFOV Protocol The HFOV protocol was designed on the basis of the results of pilot testing and consensus guidelines.24,27 We first conducted a recruitment maneuver, by applying 40 cm of water pressure for 40 seconds to the airway opening in an effort to reopen closed lung units. We then initiated HFOV with a mean airway pressure of 30 cm of water, adjusting the pressure thereafter according to the protocol, targeting a Pao2 of 55 to 80 mm Hg (Tables 1 and 2). We minimized HFOV tidal volumes by using the highest possible frequency that would maintain arterial blood pH above 7.25.13,28 After 24 hours of HFOV, conventional ventilation could be resumed if the mean airway pressure was 24 cm of water or less for 12 hours. This transition was mandatory when airway pressures reached 20 cm of water. Thereafter, mechanical ventilation followed the control protocol. Over the next 48 hours, if an Fio2 of more than 0.4 or a PEEP level of more than 14 cm of water was required for more than 1 hour to achieve oxygenation targets, HFOV was resumed. Fio2 Control Ventilation Mean Airway Pressure Fio2 cm of water PEEP cm of water 0.4 20 0.3 0.4 22 0.3 5 8 0.4 24 0.3 10 0.4 26 0.4 10 0.4 28 0.4 12 0.4 30 0.4 14 0.5 30 0.4 16 0.6 30 0.4 18 0.6 32 0.5 18 0.6 34 0.5 20 0.7 34 0.6 20 0.8 34 0.7 20 0.9 34 0.8 20 1.0 34 0.8 22 1.0 36 0.9 22 1.0 38 1.0 22 1.0 24 n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 797 The n e w e ng l a n d j o u r na l m e dic i n e Control Ventilation Protocol Statistical Analysis The control ventilation protocol, which was adapted from an earlier trial,9 called for a target tidal volume of 6 ml per kilogram, with plateau airway pressure of 35 cm of water or less and high levels of PEEP. After an initial recruitment maneuver (the same as that used for the HFOV group), clinicians applied ventilation using pressure-control mode with a PEEP level of 20 cm of water and then adjusted the PEEP level and the Fio2 according to the protocol (Tables 1 and 2). The protocol permitted the use of volume-assist control mode or pressure-support mode with the same limits for tidal volumes and airway pressures. For patients receiving pressure support with PEEP levels of 10 cm of water or less and an Fio2 of 0.4 or less, there were no limits on tidal volume or airway pressures. The weaning protocol, which has been published previously, included daily trials of spontaneous breathing.9,29 We anticipated that mortality in the control group would be 45%. Assuming a two-sided alpha level of 0.05, we calculated that enrollment of 1200 patients would provide at least 80% power to detect a relative-risk reduction with HFOV of 20%, even if mortality in the control group was as low as 37%. Investigators reviewed feasibility data from the pilot phase, which involved 94 patients, but remained unaware of the clinical outcomes. The independent data monitoring committee reviewed the clinical outcomes from the pilot phase and recommended that the trial continue to the next phase. As originally planned, data from the patients involved in the pilot phase were included in the current analyses. In addition to an interim analysis after 800 patients had undergone randomization, safety analyses of physiological data at the initiation of the study were planned after 300, 500, and 700 patients had undergone randomization. After reviewing these safety data, the data monitoring committee could request analyses of in-hospital mortality, which they did after both the 300-patient and 500-patient safety analyses. With plans to stop the study early only in response to a strong signal of harm in association with the use of HFOV, we used the O’Brien–Fleming method to calculate alpha spending and generated one-sided P values for considering early stopping after random assignment of 300 patients (P≤0.00001), 500 patients (P≤0.0001), and 700 patients (P≤0.0064). We used SAS software, version 9.2, for the statistical analyses. We summarized data using means with standard deviations, medians and interquartile ranges, or proportions. Normally distributed data were compared with the use of Student’s t-test, nonnormally distributed data with the use of the Wilcoxon rank-sum test, and proportions with the use of the Mantel–Haenszel chi-square test, with stratification according to center. We analyzed data from all patients according to their assigned group. The primary outcome was in-hospital mortality, with the outcome compared between the two groups stratified according to center. Other than recording whether death occurred as a result of withdrawal of life support, we did not record specific causes of death. As a sensitivity analysis, we used logistic regression to adjust the treatment effect for prespecified baseline vari- Procedures in Both Groups When hypoxemia persisted despite increases in PEEP or mean airway pressure, or when, on the basis of radiographic or clinical evidence, physicians judged that the lungs were over-distended, they could reduce PEEP or mean airway pressure to a level below that indicated in the assigned protocol (Table 2). For patients with hypoxemia who required an Fio2 of 0.9 or greater, clinicians could institute therapies for hypoxemia (e.g., prone positioning or inhaled nitric oxide) that did not interfere with the assigned ventilator protocols. Physicians could institute any alternative therapy (including HFOV in the control group) for patients who met any one of the following criteria: refractory hypoxemia (Pao2 <60 mm Hg for 1 hour with an Fio2 of 1.0 and neuromuscular blockade), refractory barotrauma (persistent pneumothorax or increasing subcutaneous emphysema despite two thoracostomy tubes on the involved side), or refractory acidosis (pH of ≤7.05 despite neuromuscular blockade). Physicians prescribed fluids, sedatives, and neuromuscular blockers at their discretion. We recorded cardiorespiratory variables daily as well as data on cointerventions applied while patients were undergoing mechanical ventilation for up to 60 days. Intensivists reviewed chest radiographs for evidence of new barotrauma. Patients were followed until their discharge from the hospital. 798 of n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation in Early ARDS ables: age, the Acute Physiology Score component of the Acute Physiology and Chronic Health Evaluation (APACHE) II score,30 the presence or absence of sepsis, and the duration of hospitalization before randomization.9 To compare the two groups with respect to the time to death, we used a survival analysis, in which patients who were discharged alive from the hospital were assumed to be alive at day 60. We conducted prespecified subgroup analyses to determine whether there were interactions of the treatment effect with baseline severity of lung injury (in quartiles of the Pao2:Fio2 ratio) or with center experience with HFOV and study protocols (in thirds of number of patients recruited). In addition, we studied interactions of the treatment effect with baseline dynamic compliance measured from tidal breaths during conventional ventilation (in quartiles), baseline body-mass index (in quartiles), and receipt or no receipt of vasopressors at baseline — all post hoc analyses. R e sult s Early Termination of the Trial After the 500-patient analysis, the steering committee terminated the trial, acting on a unanimous recommendation from the data monitoring committee, although the threshold P value for stopping had not been reached. At the time of termination, 571 patients had been enrolled, of whom 548 had undergone randomization: 275 to the HFOV group and 273 to the control-ventilation group (Fig. 1). Important prognostic factors were similar in the two groups at baseline (Table 3, and Table S1 in the Supplementary Appendix). Mortality A total of 129 patients (47%) in the HFOV group, as compared with 96 patients (35%) in the control group, died in the hospital (relative risk of death with HFOV, 1.33; 95% confidence interval, 1.09 to 1.64; P = 0.005) (Table 4 and Fig. 2). The results were consistent in a multivariable analysis (Table S2 in the Supplementary Appendix), in an analysis of mortality in the intensive care unit (ICU), and in an analysis of 28-day mortality. Subgroup analyses showed no interaction of mortality with baseline severity of hypoxemia, respiratory compliance, body-mass index, or use or nonuse of vasopressors or with center experience in the trial (Fig. S1 in the Supplementary Appendix). Early Physiological Responses to Ventilation Table S3 in the Supplementary Appendix shows early physiological responses to HFOV and to control ventilation. The use of vasopressors was similar in the HFOV and control groups before the initiation of ventilation (66% and 61%, respectively; P = 0.24) but increased in the HFOV group as compared with the control group within 4 hours after initiation (73% vs. 62%, P = 0.01) and increased even more in the HFOV group by the following day (78% vs. 58%, P<0.001). The use of neuromuscular blockers followed a similar pattern: 27% of patients in the HFOV group and 29% of those in the control group received neuromuscular blockers before the initiation of ventilation (P = 0.66), 46% as compared with 31% received them within 4 hours after initiation (P<0.001), and 46% as compared with 26% received them the next day (P<0.001). The mean Fio2 at these time points decreased to a similar extent in both groups: the Fio2 was 0.75 in the HFOV group and 0.73 in the control group before initiation (P = 0.93); 0.62 and 0.64 in the two groups, respectively, 4 hours after initiation (P = 0.94); and 0.51 and 0.50, respectively, the next day (P = 0.97). Cardiorespiratory Results Table S4 in the Supplementary Appendix shows cardiorespiratory data from the first week of the study. On day 1, the mean (±SD) of the mean airway pressure in the HFOV group was 31±2.6 cm of water, with a frequency of 5.5±1.0 Hz; patients in the control group underwent ventilation with a tidal volume of 6.1±1.3 ml per kilogram, PEEP of 18±3.2, and plateau pressure of 32±5.7 cm of water. The mean Fio2 in the control group was similar to or lower than that in the HFOV group, despite lower mean airway pressures. The net fluid balance was higher in the HFOV group than in the control group, but the difference was not significant. In the HFOV group, 270 of the 275 patients (98%) underwent HFOV for a median of 3 days (interquartile range, 2 to 8); a total of 222 patients (81%) survived and were transitioned to conventional ventilation for a further 5 days (interquartile range, 2 to 7). In the control group, 34 patients (12%) crossed over to HFOV (31 according to protocol and 3 in violation of protocol) for 7 days (interquartile range, 5 to 15), beginning 2 days (interquartile range, 1 to 4) after randomization; 24 of those 34 patients (71%) died in the hospital. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 799 The n e w e ng l a n d j o u r na l of m e dic i n e 2720 Patients met inclusion criteria 1571 Were excluded (some patients may have more than 1 exclusion criterion) 411 Had primary cardiac failure 316 Had chronic respiratory disease 259 Had condition with expected >50% 6-mo mortality 193 Were considering palliation rather than aggressive care 118 Were at risk for intracranial hypertension 103 Had vasculitic pulmonary hemorrhage 373 Had other reasons 1149 Were eligible 574 Were not enrolled 254 Did not provide consent 130 Were withdrawn by physician 82 Were in ICU >72 hr 75 Were already undergoing HFOV 24 Were enrolled in related trial 9 Had other reasons 571 Were enrolled 23 Did not undergo randomization 19 Had Pao2:FIo2 ratio >200 mm Hg on standard settings 4 Had other reasons 548 Underwent randomization 273 Were assigned to receive control ventilation 273 Received assigned intervention 275 Were assigned to receive HFOV 270 Received assigned intervention 2 Died before HFOV could be started 2 Withdrew consent after randomization 1 Had approval withdrawn by physician 1 Had premature termination of assigned strategy after withdrawal of consent 3 Had premature termination of assigned strategy after withdrawal of consent 273 Were included in primary analysis 275 Were included in primary analysis Figure 1. Screening, Randomization, and Follow-up. HFOV denotes high-frequency oscillatory ventilation, ICU intensive care unit, and Pao2:Fio2 the ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen. 800 n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation in Early ARDS Table 3. Baseline Characteristics of the Patients.* HFOV Group (N = 275) Characteristic Age — yr Female sex — no. (%) Control Group (N = 273) Patients Eligible but Not Enrolled (N = 472)† P Value‡ 55±16 54±16 53±16 0.18 108 (39) 120 (44) 198 (42) 0.42 26±8 <0.001 APACHE II score§ 29±8 29±7 Duration of hospital stay — days 5.6±8.0 4.9±8.0 Duration of mechanical ventilation — days 2.5±3.3 1.9±2.3 Risk factors for ARDS — no. of patients (%) Sepsis 128 (47) 130 (48) 193 (41) 0.01 Pneumonia 155 (56) 164 (60) 289 (61) 0.37 49 (18) 44 (16) 51 (11) 0.02 Gastric aspiration Trauma 10 (4) 5 (2) 24 (5) 0.07 Other 71 (26) 67 (25) 137 (29) 0.34 Tidal volume — ml/kg of predicted body weight 7.2±1.9 7.1±1.8 Plateau pressure — cm of water 29±6 29±7 27±7 <0.001 11±4 <0.001 13±3 13±4 Minute ventilation — liters/min Set PEEP — cm of water 11.3±3.1 11.2±3.3 Oxygenation index 19.6±11.2 19.9±9.3 17.8±10.2 0.002 Pao2:Fio2 ratio — mm Hg 121±46 114±38 118±47 0.17 PaCo2 — mm Hg Arterial pH Barotrauma — no. of patients (%) 46±13 47±14 45±14 0.01 7.32±0.10 7.31±0.10 7.32±0.12 0.06 19 (7) 14 (5) 184 (67) 171 (63) Cointerventions — no. of patients (%) Inotropes or vasopressors Renal-replacement therapy 29 (11) 28 (10) Glucocorticoids 93 (34) 96 (35) Neuromuscular blockers 84 (31) 94 (34) *Plus–minus values are means ±SD. There were no significant differences between the two study groups in any of the baseline characteristics listed here, with the exception of duration of mechanical ventilation, for which P = 0.003. ARDS denotes acute respiratory distress syndrome, and Pao2 partial pressure of arterial oxygen. †Not all centers had approval from an ethics committee to collect data on patients who were eligible but not enrolled in the study. ‡The P values are for the comparison of patients who were eligible but not enrolled with all patients who underwent randomization, with adjustment for stratification according to center. § Scores on the Acute Physiology and Chronic Health Evaluation II (APACHE II) range from 0 to 71, with higher scores indicating greater severity of illness. Cointerventions During the course of the study, larger proportions of patients in the HFOV group than in the control group received vasoactive drugs (91% vs. 84%, P = 0.01) and neuromuscular blockers (83% vs. 68%, P<0.001); vasoactive drugs were administered for an average of 2 days longer in the HFOV group than in the control group, and neuromus- cular blockers were administered for an average of 1 day longer in the HFOV group (Table S5 in the Supplementary Appendix). Sedatives and opioids (most commonly midazolam and fentanyl) were administered for the same duration in the two groups (median, 10 days [interquartile range, 6 to 18] and 10 days [interquartile range, 6 to 17], respectively; P = 0.99), but during the first week the n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 801 The n e w e ng l a n d j o u r na l of m e dic i n e Table 4. Outcomes. HFOV Group (N = 275) Control Group (N = 273) Relative Risk (95% CI) P Value Death in hospital — no. (%) 129 (47) 96 (35) 1.33 (1.09–1.64) 0.005 Death in intensive care unit — no. (%) 123 (45) 84 (31) 1.45 (1.17–1.81) 0.001 Death before day 28 — no. (%) 111 (40) 78 (29) 1.41 (1.12–1.79) 0.004 New barotrauma — no./total no. (%)* 46/256 (18) 34/259 (13) 1.37 (0.91–2.06) 0.13 New tracheostomy — no./total no. (%)† 59/273 (22) 66/267 (25) 0.87 (0.64–1.19) 0.39 19 (7) 38 (14) 0.50 (0.29–0.84) 0.007 Outcome Refractory hypoxemia — no. (%) Death after refractory hypoxemia — no./total no. (%) 15/19 (79) 25/38 (66) 1.20 (0.87–1.66) 0.31 Refractory acidosis — no. (%) 9 (3) 8 (3) 1.12 (0.44–2.85) 0.82 Refractory barotrauma — no. (%) 2 (<1) 2 (<1) 0.99 (0.14–7.00) 0.99 Use of mechanical ventilation, among survivors — days Median Interquartile range 0.59 11 10 7–19 6–18 15 14 9–25 9–26 Stay in intensive care, among survivors — days Median Interquartile range 0.93 Length of hospitalization, among survivors — days Median Interquartile range 0.74 30 25 16–45 15–41 *Barotrauma was defined as pneumothorax, pneumomediastinum, pneumopericardium, or subcutaneous emphysema occurring spontaneously or after a recruitment maneuver. Excluded from this category were patients who had barotrauma at baseline. †Excluded from this category were patients who had a tracheostomy at baseline. median doses of midazolam were significantly higher in the HFOV group than in the control group (199 mg per day [interquartile range, 100 to 382] vs. 141 mg per day [interquartile range, 68 to 240], P<0.001), and there was a trend toward higher doses of fentanyl equivalents in the HFOV group (2980 μg per day [interquartile range, 1258 to 4800] vs. 2400 μg per day [interquartile range, 1140 to 4430], P = 0.06) (for daily doses of selected sedative and analgesic drugs, see Fig. S2 in the Supplementary Appendix). The rates of use of other cointerventions, including glucocorticoids, renal-replacement therapy, and prone positioning, were similar in the two groups (Table S5 in the Supplementary Appendix). Other Outcomes Refractory hypoxemia developed in significantly more patients in the control group than in the HFOV group; however, the total number of deaths after refractory hypoxemia was similar in the 802 two groups (Table 4). The proportion of deaths after withdrawal of life support was similar in the two groups (55% [71 of 129 patients] in the HFOV group and 49% [47 of 96 patients] in the control group, P = 0.12). The rate of new-onset barotrauma was higher in the HFOV group than in the control group, but the difference was not significant (18% and 13%, respectively; P = 0.13). Among survivors, the duration of ventilation and the length of stay in the ICU were similar in the two groups (Table 4). Discussion The main finding of this multicenter, randomized trial is that among patients with moderateto-severe ARDS, early application of HFOV was associated with higher mortality than was a ventilation strategy that used small tidal volumes and high PEEP levels, with HFOV used only in patients with severe refractory hypoxemia. HFOV n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation in Early ARDS 1.0 0.9 Probability of Survival was associated with higher mean airway pressures and with greater use of sedatives, neuromuscular blockers, and vasoactive drugs. We stopped the trial early on the basis of a strong signal for increased mortality with HFOV, even though the prespecified stopping thresholds had not been reached. Studies that are stopped early on the basis of harm (or benefit) typically overestimate the magnitude of effect.31 We chose to terminate the study for three reasons: there was a consistent finding of increased mortality with HFOV in three consecutive analyses that were conducted after enrollment of 94, 300, and 500 patients; the increased need for vasoactive drugs in the HFOV group suggested a mechanism of harm that was not offset by better oxygenation and lung recruitment; and the effect size was sufficiently large that we concluded that even if early HFOV did not increase mortality, it would be very unlikely to decrease mortality. We believe that continued enrollment would have put patients at risk with little likelihood of benefit. Our results are inconsistent with the physiological rationale for HFOV and with the results of studies in animals. In studies in animals in which benefits of HFOV were observed, lung injury was induced with the use of saline lavage — a highly recruitable model of surfactant deficiency — which our results suggest does not translate directly to human adults with ARDS, in whom recruitability can be heterogeneous.32 Our results also contrast with those of prior randomized trials involving adults.22 A possible explanation, which provided motivation for our trial, is that prior studies used control ventilation strategies that are now known to be potentially harmful.20,21 We found no benefit with HFOV when a current ventilation strategy was used as a control. This finding of no benefit with respect to mortality is consistent with the results of another trial now reported in the Journal; in that trial, conducted in the United Kingdom, current standards for lung protection were suggested but not mandated.33 More surprising was our finding of harm. Several plausible mechanisms may contribute to increased mortality with HFOV. Higher mean airway pressures may result in hemodynamic compromise by decreasing venous return or directly affecting right ventricular function.34 Increased use of vasodilating sedative agents may also contribute to hemodynamic compromise. Moreover, we cannot exclude the 0.8 0.7 Control 0.6 0.5 0.4 HFOV 0.3 0.2 P=0.004 by log-rank test 0.1 0.0 0 15 30 45 60 54 54 26 39 Days since Randomization No. at Risk HFOV Control 275 273 169 181 98 92 Figure 2. Probability of Survival from the Day of Randomization to Day 60 in the HFOV and Control Groups. possibility of increased barotrauma in association with HFOV. The HFOV strategy that we chose, which was supported by preclinical data15,16 and a prospective physiological study,24 aimed to adjust mean airway pressure on the deflation limb of the volume-pressure curve and use the highest frequency possible to limit oscillatory volumes. This approach led to relatively high mean airway pressures, even considering that when mean airway pressures are delivered with a ratio of inspiratory-to-expiratory time of 1:2, as in our study, the pressures measured at the airway opening during HFOV are somewhat higher than those measured in the trachea.35-37 It is possible that an HFOV protocol that uses lower mean airway pressures, a different ratio of inspiratory-to-expiratory time, or a lower oscillatory frequency might have led to different results. The strengths of this trial include its methodologic rigor, the application of protocols designed to open lung units in patients in both groups on the basis of the best available evidence, and enrollment at centers in several countries, which enhances the generalizability of our findings. Because we were cognizant that there is a learning curve associated with the use of HFOV,38,39 we enrolled most patients at centers that were experienced with HFOV, and we did not detect an interaction between treatment effect and the number of enrolled patients per site. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 803 The n e w e ng l a n d j o u r na l Our results raise serious concerns about the early use of HFOV for the management of ARDS in adults. The results of this study increase the uncertainty about possible benefits of HFOV even when applied in patients with life-threatening refractory hypoxemia. In conclusion, in adults with moderate-tosevere ARDS, the early application of HFOV targeting lung recruitment — as compared with a ventilation strategy that uses low tidal volume and high PEEP and that permits HFOV only in cases of refractory hypoxemia — does not reduce mortality and may be harmful. Supported by the Canadian Institutes of Health Research, Randomized Controlled Trial (RCT) Program (Ottawa) and the King Abdullah International Medical Research Center (Riyadh, of m e dic i n e Saudi Arabia). Dr. Ferguson is supported by a Canadian Institutes of Health Research New Investigator Award; Dr. Cook holds a Canada Research Chair; Dr. Lamontagne is supported by a Fonds de Recherche de Québec–Santé Research Career Award; and Drs. Ferguson and Meade and Drs. Lamontagne and Meade were supported by Canadian Institutes of Health Research RCT Mentorship awards. Dr. Ferguson reports receiving grant support through his institution from the Physicians Services Incorporated Foundation; Dr. Granton, receiving consulting fees from Ikaria, lecture fees from Actelion and Eli Lilly, grant support through his institution from Pfizer, Actelion, Eli Lilly, GlaxoSmithKline, and Bayer, payment through his institution from Telus for the sale of a software site license, and support from Actelion for his hospital foundation for pulmonary hypertension research and providing expert testimony for Pfizer regarding patent legal action; and Dr. Slutsky, receiving consulting fees from Maquet Medical, Novalung, Gambro, Ikaria, and Hemodec. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. References 1. ARDS Definition Task Force, Ranieri VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin Definition. JAMA 2012;307:2526-33. 2. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med 2005;353: 1685-93. 3. Phua J, Badia JR, Adhikari NKJ, et al. Has mortality from acute respiratory distress syndrome decreased over time? Am J Respir Crit Care Med 2009;179:220-7. 4. Herridge MS, Tansey CM, Matté A, et al. Functional disability 5 years after acute respiratory distress syndrome. N Engl J Med 2011;364:1293-304. 5. Tremblay LN, Slutsky AS. Ventilatorinduced lung injury: from the bench to the bedside. Intensive Care Med 2006;32: 24-33. 6. 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;342:1301-8. 7. Brower RG, Lanken PN, MacIntyre N, et al. Higher versus lower positive endexpiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med 2004;351:327-36. 8. Mercat A, Richard J-CM, Vielle B, et al. Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008; 299:646-55. 9. Meade MO, Cook DJ, Guyatt GH, et al. Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008;299:637-45. 804 10. Briel M, Meade M, Mercat A, et al. Higher vs lower positive end-expiratory pressure in patients with acute lung injury and acute respiratory distress syndrome: systematic review and meta-analysis. JAMA 2010;303:865-73. 11. Terragni PP, Rosboch G, Tealdi A, et al. Tidal hyperinflation during low tidal volume ventilation in acute respiratory distress syndrome. Am J Respir Crit Care Med 2007;175:160-6. 12. Terragni PP, Del Sorbo L, Mascia L, et al. Tidal volume lower than 6 ml/kg enhances lung protection: role of extracorporeal carbon dioxide removal. Anesthesiology 2009;111:826-35. 13. Hager DN, Fessler HE, Kaczka DW, et al. Tidal volume delivery during high-frequency oscillatory ventilation in adults with acute respiratory distress syndrome. Crit Care Med 2007;35:1522-9. 14. Ferguson ND, Slutsky AS, Kacmarek R. High-frequency ventilation is/is not the optimal physiological approach to ventilate ARDS patients. J Appl Physiol 2008; 104:1230-1. 15. Froese AB. High-frequency oscillatory ventilation for adult respiratory distress syndrome: let’s get it right this time. Crit Care Med 1997;25:906-8. 16. Rimensberger PC, Pache JC, McKerlie C, Frndova H, Cox PN. Lung recruitment and lung volume maintenance: a strategy for improving oxygenation and preventing lung injury during both conventional mechanical ventilation and high-frequency oscillation. Intensive Care Med 2000;26: 745-55. 17. Imai Y, Nakagawa S, Ito Y, Kawano T, Slutsky AS, Miyasaka K. Comparison of lung protection strategies using conventional and high-frequency oscillatory ventilation. J Appl Physiol 2001;91: 1836-44. 18. Muellenbach RM, Kredel M, Said HM, et al. High-frequency oscillatory ventilation reduces lung inflammation: a largeanimal 24-h model of respiratory distress. Intensive Care Med 2007;33:1423-33. 19. Sedeek KA, Takeuchi M, Suchodolski K, et al. Open-lung protective ventilation with pressure control ventilation, highfrequency oscillation, and intratracheal pulmonary ventilation results in similar gas exchange, hemodynamics, and lung mechanics. Anesthesiology 2003;99: 1102-11. 20. Derdak S, Mehta S, Stewart TE, et al. High frequency oscillatory ventilation for acute respiratory distress syndrome: a randomized controlled trial. Am J Respir Crit Care Med 2002;166:801-8. 21. Bollen CW, van Well GT, Sherry T, et al. High frequency oscillatory ventilation compared with conventional mechanical ventilation in adult respiratory distress syndrome: a randomized controlled trial. Crit Care 2005;9:R430-R439. 22. Sud S, Sud M, Friedrich JO, et al. High frequency oscillation in patients with acute lung injury and acute respiratory distress syndrome (ARDS): systematic review and meta-analysis. BMJ 2010;340: c2327. 23. Fort P, Farmer C, Westerman J, et al. High-frequency oscillatory ventilation for adult respiratory distress syndrome — a pilot study. Crit Care Med 1997;25:937-47. 24. Ferguson ND, Chiche J-D, Kacmarek RM, et al. Combining high-frequency oscillatory ventilation and recruitment maneuvers in adults with early acute respiratory distress syndrome: the Treatment with Oscillation and an Open Lung Strategy (TOOLS) Trial pilot study. Crit Care Med 2005;33:479-86. 25. Mehta S, Granton J, MacDonald RJ, et al. High-frequency oscillatory ventilation n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. High-Frequency Oscillation in Early ARDS in adults: the Toronto experience. Chest 2004;126:518-27. 26. Adhikari NKJ, Bashir A, Lamontagne F, et al. High-frequency oscillation in adults: a utilization review. Crit Care Med 2011;39:2631-44. 27. Fessler HE, Derdak S, Ferguson ND, et al. A protocol for high-frequency oscillatory ventilation in adults: results from a roundtable discussion. Crit Care Med 2007;35:1649-54. 28. Fessler HE, Hager DN, Brower RG. Feasibility of very high-frequency ventilation in adults with acute respiratory distress syndrome. Crit Care Med 2008;36: 1043-8. 29. Esteban A, Frutos F, Tobin MJ, et al. A comparison of four methods of weaning patients from mechanical ventilation. N Engl J Med 1995;332:345-50. 30. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 31. Montori VM, Devereaux PJ, Adhikari NK, et al. Randomized trials stopped early for benefit: a systematic review. JAMA 2005;294:2203-9. 32. Gattinoni L, Caironi P, Cressoni M, et al. Lung recruitment in patients with the acute respiratory distress syndrome. N Engl J Med 2006;354:1775-86. 33. Young D, Lamb SE, Shah S, et al. Highfrequency oscillation for acute respiratory distress syndrome. N Engl J Med 2013; 368:806-13. 34. Guervilly C, Forel J-M, Hraiech S, et al. Right ventricular function during high-frequency oscillatory ventilation in adults with acute respiratory distress syndrome. Crit Care Med 2012;40:1539-45. 35. Hatcher D, Watanabe H, Ashbury T, Vincent S, Fisher J, Froese A. Mechanical performance of clinically available, neo- natal, high-frequency, oscillatory-type ventilators. Crit Care Med 1998;26:1081-8. 36. Pillow JJ, Neil H, Wilkinson MH, Ramsden CA. Effect of I/E ratio on mean alveolar pressure during high-frequency oscillatory ventilation. J Appl Physiol 1999;87:407-14. 37. Mentzelopoulos SD, Malachias S, Kokkoris S, Roussos C, Zakynthinos SG. Comparison of high-frequency oscillation and tracheal gas insufflation versus standard high-frequency oscillation at two levels of tracheal pressure. Intensive Care Med 2010;36:810-6. 38. The HiFi Study Group. High-frequency oscillatory ventilation compared with conventional mechanical ventilation in the treatment of respiratory failure in preterm infants. N Engl J Med 1989;320:88-93. 39. Bryan AC, Froese AB. Reflections on the HIFI trial. Pediatrics 1991;87:565-7. Copyright © 2013 Massachusetts Medical Society. n engl j med 368;9 nejm.org february 28, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 805 Clinical Review & Education Special Communication | LESS IS MORE “Less Is More” in Critically Ill Patients Not Too Intensive Matthijs Kox, PhD; Peter Pickkers, MD, PhD The current view in intensive care medicine is that very sick patients need very intensive treatment. However, in this group of highly vulnerable patients, more intensive treatment may promote the chances of unwanted adverse effects and hence, iatrogenic damage. Therefore, we state that critically ill patients probably benefit from a more cautious approach. Using data from large clinical trials of previous years, we exemplify that less intensive treatment is associated with a better outcome in intensive care patients and suggest that we reappraise patient management as well as trial design in intensive care medicine while bearing in mind the “less is more” paradigm. We illustrate our case by describing the intensity of the most relevant treatment options for patients with septic shock, including mechanical ventilation, fluid management, blood pressure–targeted therapy, corticosteroids, patient monitoring, sedation, and nutrition. We conclude that treatment of critically ill patients while keeping in mind the “less is more” paradigm might not only benefit the patient but could also have a notable impact on the ever-increasing intensive care–related health care costs. JAMA Intern Med. 2013;173(14):1369-1372. doi:10.1001/jamainternmed.2013.6702 Published online June 10, 2013. T he phrase “less is more” was first popularized by the German architect Ludwig Mies van der Rohe, one of the pioneers of modern architecture. Because of financial shortages after World War II, the modernist style originally conceived for aesthetic reasons was quickly adopted as a practical program of inexpensive construction. Although not directly related to the current financial crisis, we propose that “less is more” also applies to the treatment of critically ill patients. The current view is that these very sick patients need very intensive treatment. Indeed, the addition of several treatments such as mechanical ventilation and improved hemodynamic monitoring has revolutionized the care of medical and surgical patients who are critically ill. However, because more intensive treatment may promote the chances of unwanted adverse effects, iatrogenic damage is more likely to occur, resulting in unfavorable sequelae and worse outcome in this highly vulnerable group of patients. Therefore, in keeping with the Hippocratic oath (primum non nocere), we state that critically ill patients probably benefit from a more “cautious” approach. In this Special Article, based on data from large clinical trials conducted in the previous years, we show that less intensive treatment is indeed associated with a better outcome in critically ill patients (Table). To improve outcome, we should therefore reappraise patient management as well as trial design in intensive care medicine, bearing in mind with the principles of “less is more.” Respiratory Therapy, Fluids, and Transfusion To illustrate our case, we describe the intensity of the most relevant treatment options for septic shock, the number 1 cause of jamainternalmedicine.com Author Affiliations: Department of Intensive Care Medicine, Radboud University Nijmegen Medical Center, Nijmegen, the Netherlands (Kox, Pickkers); Department of Anesthesiology, Radboud University Nijmegen Medical Center, Nijmegen, the Netherlands (Kox); Nijmegen Institute for Infection, Inflammation, and Immunity (N4i), Nijmegen, the Netherlands (Kox, Pickkers). Corresponding Author: Matthijs Kox, PhD, Department of Intensive Care Medicine, Internal Mail 710, Radboud University Nijmegen Medical Center, Geert Grooteplein 10, 6500 HB Nijmegen, the Netherlands ([email protected]). death in noncardiac intensive care units (ICUs). 1,24 In most patients with septic shock, mechanical ventilation is required. The first Acute Respiratory Distress Syndrome (ARDS) Network trial has clearly shown that ventilation with lower tidal volumes (6 vs 12 mL/kg) reduces lung injury and improves outcome.1 Furthermore, higher levels of positive end-expiratory pressure (13 cm of water) did not improve clinical outcome compared with lower levels (8 cm of water),2 indicating that “more” is not “better.”2 Of interest, recently high-frequency ventilation was found to increase mortality in ARDS patients, illustrating that complicated tertiary approaches to ventilation do not seem to help.3 Related to respiratory therapy, too-liberal supplementation of oxygen resulting in high P O 2 values is associated with, albeit slightly, increased mortality in observational cohorts.4 Randomized trials to investigate the effects of higher and lower oxygen levels are currently not available. In addition, following the initial fluid resuscitation phase, it has been demonstrated that a more conservative fluid management results in improved lung function and shorter stay in the ICU compared with a more liberal regime.5 Recently, an extreme example of the benefits of less fluid administration emerged in a study performed in African children presenting with febrile illness and impaired perfusion.6 Strikingly, a survival benefit was found in children who were withheld from fluid administration compared with those receiving, by US and European standards, conservative fluid boluses. 6 Also, a restrictive strategy of red blood cell transfusion is superior to a liberal one.7 It is clear that for respiratory and hemodynamic management, less is more in the critically ill patient. Remarkably, the same holds true for many adjunctive treatments applied in the ICU. JAMA Internal Medicine July 22, 2013 Volume 173, Number 14 Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013 1369 Clinical Review & Education Special Communication “Less Is More” in Critically Ill Patients Table. Interventions for Which “More” Has Been Shown to Be Associated With Worse Outcome Intervention Respiratory therapy Higher tidal volume (12 vs 6 mL/kg)1; higher PEEP (13 vs 8 cm of water)2; high-frequency ventilation3 High PO24 Effect Type of Study 1,3 1-3 Increased mortality and days on ventilator1; no benefit2 RCT Increased mortality4 Observational study4 Worsened lung function,5 increased days on ventilator, and ICU stay5; increased mortality6 Increased mortality7 RCT5,6 No benefit8; increased mortality9,11; increased mortality (only in less severe septic shock)10 RCT8,10; observational study9; post hoc analysis of RCT11 High-dose corticosteroids12 Cardiac therapy Increased mortality12 RCT12 Prophylactic lidocaine treatment to reduce VF13 Monitoring Increased mortality13 RCT13 Swan-Ganz cathether14 Daily routine x-ray examinations15 Antibiotics No benefit14 RCT14 No benefit15 Observational study15 Longer duration of antibiotic treatment16 Renal therapy No benefit16 RCT16 High vs low volume renal replacement therapy17 Sedation No benefit17 RCT17 Continuous sedation vs daily interruption of sedatives (lower sedative use)18; daily interruption of sedatives vs no sedation19 Nutrition and glucose control High vs moderate caloric intake20; target feeding vs permissive underfeeding21; full enteral nutrition vs trophic nutrition22 Strict vs less strict glucose control in patients that are predominantly fed enterally23 Increased days on ventilator18,19 and ICU stay18 RCT18,19 Increased mortality20,21; no benefit22 Observational study20; RCT21,22 Fluids and transfusion Liberal fluid management vs strict regime5; conservative vs no fluid administration6 Liberal vs restrictive red blood cell transfusion7 Blood pressure–targeted therapy Increasing MAP >65 mm Hg8; higher vasopressor use9; higher vasopressor infusion rates vs lower rates + vasopressin10; higher vs lower dosage of L-NMMA11 Corticosteroids RCT7 Patients with septic shock require vasopressor therapy by definition. Although the use of different vasopressor agents appears to result in similar outcome, increasing mean arterial pressure above 65 mm Hg showed no benefit.8 Furthermore, vasopressor load was shown to be associated with mortality in an observational study.9 While there is no direct evidence for adverse effects of high vasopressor use (there are no randomized clinical trials comparing different dosages), most pharmacological adverse effects are dose dependent. Hence, combinations of (lower dosages) of vasopressors might be beneficial compared with the use of higher dosages of a single vasopressor, and indeed there is some evidence for this. For instance, vasopressin therapy in combination with norepinephrine resulted in a significant survival benefit in patients with mild septic shock compared with patients treated with norepinephrine alone, although this was not the case in patients with severe shock.10 Also, while the nitric oxide synthase inhibitor N G -monomethyl-Larginine (L-NMMA) was associated with worse outcome in the complete group of patients with septic shock, those who received a lower infusion rate (<5 mg/kg/h) had an improved survival rate relative to placebo-treated patients, suggesting that lower dosages are beneficial.11 Corticosteroids, Monitoring, and Antibiotics Patients with refractory shock who require high dosages of vasopressor therapy might benefit from corticosteroids. However, in the early trials, high dosages of corticosteroids resulted in adverse outcome.12 In later trials it was demonstrated that lower corticosteroid dosages may improve outcome, especially when administrated early in those with refractory shock.25,26 In case of impending myocardial infarction, prophylactic treatment with lidocaine, while reducing the incidence of ventricular fibrillation, actually increases mortality.13 Another argument in support of “less is more” in critically ill patients is the fact that studies have failed to demonstrate that more invasive hemodynamic monitoring using SwanGanz catheters is beneficial for the ICU patient.14 Furthermore, daily routine x-ray examinations, another form of monitoring, were not shown to be useful,15 and even a shorter duration of antibiotic treatment (guided by procalcitonin levels) was not associated with worse outcome,16 while liberal use of antibiotics has many disadvantages, among which increased risk of the development of resistance is the most important one. Renal Therapy, Sedation, Nutrition, and Glucose Control Increased mortality23 RCT23 Abbreviations: ICU, intensive care unit; L-NMMA, NG-monomethyl-L-arginine; MAP, mean arterial pressure; PEEP, positive end-expiratory pressure; RCT, randomized clinical trial; VF, ventricular fibrillation. 1370 Blood Pressure–Targeted and Cardiac Therapy Patients with septic shock often develop acute kidney injury, for which renal replacement therapy can be initiated, but increasing ultrafiltrate volume has not showed any benefit.17 Most patients with septic shock will be sedated. The old paradigm was to deeply sedate to facilitate adequate patient management. It is now becoming increasingly clear that less18 or even no19 sedation is superior. Finally, patients need nutrition, and it has emerged that high caloric JAMA Internal Medicine July 22, 2013 Volume 173, Number 14 Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013 jamainternalmedicine.com “Less Is More” in Critically Ill Patients Special Communication Clinical Review & Education intake is associated with increased mortality,20 whereas low caloric intake via the enteral route (permissive underfeeding) reduces hospital mortality in critically ill patients.21 In line, it was recently demonstrated in ARDS patients that full enteral feeding is not superior compared with trophic feeding in terms of time on ventilator or mortality.22 Furthermore, early initiation of parenteral nutrition worsens outcome,27 suggesting it is better to limit caloric intake via the parenteral route until enteral feeding is possible. Related to this issue, beneficial effects of intensive glucose control therapy were found in ICU patients who receive larger amounts of parenteral calories,28 while it was found to increase mortality in patients who are predominantly fed via the enteral route.23 From these studies, it might be concluded that very strict glucose regulation is not beneficial by itself and only nullifies the deleterious effects of early parenteral administration of calories. Exceptions to the Rule Naturally, there are some notable exceptions to the rule; sometimes “more is more.” For instance, early mobilization of critically ill patients,29 infection prevention by means of decontamination of the digestive tract and oropharynx,30 and high-intensity ICU physician31 and nurse32 staffing is associated with better outcome. Furthermore, hemodynamic optimization and/or early goal-directed therapy is associated with better outcome following major surgery33 and in the early phase of sepsis.34 Nevertheless, equipoise still remains, and 3 trials (United States, United Kingdom, and Australia) are under way to evaluate the true effectiveness of early goal-directed therapy. FurARTICLE INFORMATION Accepted for Publication: March 16, 2013. Published Online: June 10, 2013. doi:10.1001/jamainternmed.2013.6702. Author Contributions: Study concept and design: Both authors. Drafting of the manuscript: Kox. Critical revision of the manuscript for important intellectual content: Pickkers. Study supervision: Pickkers. Conflict of Interest Disclosures: None reported. REFERENCES 1. 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(18):1301-1308. 2. Brower RG, Lanken PN, MacIntyre N, et al; National Heart, Lung, and Blood Institute ARDS Clinical Trials Network. Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med. 2004;351(4):327-336. 3. Ferguson ND, Cook DJ, Guyatt GH, et al; OSCILLATE Trial Investigators; Canadian Critical Care Trials Group. High-frequency oscillation in early acute respiratory distress syndrome. N Engl J Med. 2013;368(9):795-805. 4. de Jonge E, Peelen L, Keijzers PJ, et al. Association between administered oxygen, arterial partial oxygen pressure and mortality in jamainternalmedicine.com thermore, a highly intensive and invasive treatment such as extracorporeal circulation membrane oxygenation has proved its value during the 2009 H1N1 influenza pandemic.35 However, a recent observational study indicates that the effectiveness of extracorporeal circulation membrane oxygenation in ARDS patients is highly dependent on chest wall mechanics and that inappropriate use of extracorporeal circulation membrane oxygenation can be avoided in many patients by partitioning respiratory system mechanics between lung and chest wall.36 Conclusions It is becoming clear that following decades of increasingly more intensive treatment of critically ill patients, a more cautious approach for the vulnerable ICU patient is in many cases associated with improved outcome. While it may prove to be a difficult task to convince more and more ICU physicians to do less, implementation of evidence-based reduction of therapy intensities is of vital importance. With regard to ICU trial design, instead of investigating the effects of altering a single therapy, it would be of interest to evaluate the efficacy of combinations of lower dosages or intensities of certain therapies to reduce the chances of adverse effects to occur. Last but not least, keeping the “less is more” paradigm in mind in the treatment of critically ill patients might not only benefit the patient but could also have a notable impact on the ever-increasing ICU-related health care costs and thereby benefit the society as a whole, especially in these times of financial hardship. mechanically ventilated intensive care unit patients. Crit Care. 2008;12(6):R156. 5. Wiedemann HP, Wheeler AP, Bernard GR, et al; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network. Comparison of two fluid-management strategies in acute lung injury. N Engl J Med. 2006;354(24):2564-2575. 6. Maitland K, Kiguli S, Opoka RO, et al; FEAST Trial Group. Mortality after fluid bolus in African children with severe infection. N Engl J Med. 2011;364(26): 2483-2495. 7. Hébert PC, Wells G, Blajchman MA, et al; Transfusion Requirements in Critical Care Investigators for the Canadian Critical Care Trials Group. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med. 1999;340(6):409-417. 8. Bourgoin A, Leone M, Delmas A, Garnier F, Albanèse J, Martin C. Increasing mean arterial pressure in patients with septic shock: effects on oxygen variables and renal function. Crit Care Med. 2005;33(4):780-786. 9. Dünser MW, Ruokonen E, Pettilä V, et al. Association of arterial blood pressure and vasopressor load with septic shock mortality: a post hoc analysis of a multicenter trial. Crit Care. 2009;13(6):R181. 10. Russell JA, Walley KR, Singer J, et al; VASST Investigators. Vasopressin versus norepinephrine infusion in patients with septic shock. N Engl J Med. 2008;358(9):877-887. 11. López A, Lorente JA, Steingrub J, et al. Multiple-center, randomized, placebo-controlled, double-blind study of the nitric oxide synthase inhibitor 546C88: effect on survival in patients with septic shock. Crit Care Med. 2004;32(1):21-30. 12. Bone RC, Fisher CJ Jr, Clemmer TP, Slotman GJ, Metz CA, Balk RA. A controlled clinical trial of high-dose methylprednisolone in the treatment of severe sepsis and septic shock. N Engl J Med. 1987;317(11):653-658. 13. Sadowski ZP, Alexander JH, Skrabucha B, et al. Multicenter randomized trial and a systematic overview of lidocaine in acute myocardial infarction. Am Heart J. 1999;137(5):792-798. 14. Harvey S, Harrison DA, Singer M, et al; PAC-Man study collaboration. Assessment of the clinical effectiveness of pulmonary artery catheters in management of patients in intensive care (PAC-Man): a randomised controlled trial. Lancet. 2005;366(9484):472-477. 15. Graat ME, Choi G, Wolthuis EK, et al. The clinical value of daily routine chest radiographs in a mixed medical-surgical intensive care unit is low. Crit Care. 2006;10(1):R11. 16. Bouadma L, Luyt CE, Tubach F, et al; PRORATA trial group. Use of procalcitonin to reduce patients’ exposure to antibiotics in intensive care units (PRORATA trial): a multicentre randomised controlled trial. Lancet. 2010;375(9713):463-474. 17. Bouman CS, Oudemans-Van Straaten HM, Tijssen JG, Zandstra DF, Kesecioglu J. Effects of early high-volume continuous venovenous hemofiltration on survival and recovery of renal JAMA Internal Medicine July 22, 2013 Volume 173, Number 14 Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013 1371 Clinical Review & Education Special Communication 31. Pronovost PJ, Angus DC, Dorman T, Robinson KA, Dremsizov TT, Young TL. Physician staffing patterns and clinical outcomes in critically ill patients: a systematic review. JAMA. 2002;288(17): 2151-2162. function in intensive care patients with acute renal failure: a prospective, randomized trial. Crit Care Med. 2002;30(10):2205-2211. severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29(7):1303-1310. 18. Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med. 2000;342(20):1471-1477. 25. Annane D, Sébille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA. 2002;288(7):862-871. 19. Strøm T, Martinussen T, Toft P. A protocol of no sedation for critically ill patients receiving mechanical ventilation: a randomised trial. Lancet. 2010;375(9713):475-480. 26. Sprung CL, Annane D, Keh D, et al; CORTICUS Study Group. Hydrocortisone therapy for patients with septic shock. N Engl J Med. 2008;358(2): 111-124. 20. Krishnan JA, Parce PB, Martinez A, Diette GB, Brower RG. Caloric intake in medical ICU patients: consistency of care with guidelines and relationship to clinical outcomes. Chest. 2003;124(1):297-305. 27. Casaer MP, Mesotten D, Hermans G, et al. Early versus late parenteral nutrition in critically ill adults. N Engl J Med. 2011;365(6):506-517. 33. Pearse R, Dawson D, Fawcett J, Rhodes A, Grounds RM, Bennett ED. Early goal-directed therapy after major surgery reduces complications and duration of hospital stay: a randomised, controlled trial [ISRCTN38797445]. Crit Care. 2005;9(6):R687-R693. 21. Arabi YM, Tamim HM, Dhar GS, et al. Permissive underfeeding and intensive insulin therapy in critically ill patients: a randomized controlled trial. Am J Clin Nutr. 2011;93(3):569-577. 28. Vlasselaers D, Milants I, Desmet L, et al. Intensive insulin therapy for patients in paediatric intensive care: a prospective, randomised controlled study. Lancet. 2009;373(9663): 547-556. 34. Rivers E, Nguyen B, Havstad S, et al; Early Goal-Directed Therapy Collaborative Group. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345(19):1368-1377. 22. National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network. Initial trophic vs full enteral feeding in patients with acute lung injury: the EDEN randomized trial. JAMA. 2012;307(8):795-803. 29. Schweickert WD, Pohlman MC, Pohlman AS, et al. Early physical and occupational therapy in mechanically ventilated, critically ill patients: a randomised controlled trial. Lancet. 2009;373(9678):1874-1882. 35. Noah MA, Peek GJ, Finney SJ, et al. Referral to an extracorporeal membrane oxygenation center and mortality among patients with severe 2009 influenza A(H1N1). JAMA. 2011;306(15):1659-1668. 23. NICE-SUGAR Study Investigators. Intensive versus conventional glucose control in critically ill patients. N Engl J Med. 2009;360(13):1283-1297. 30. de Smet AM, Kluytmans JA, Cooper BS, et al. Decontamination of the digestive tract and oropharynx in ICU patients. N Engl J Med. 2009;360(1):20-31. 24. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of 1372 “Less Is More” in Critically Ill Patients 32. Cho SH, Hwang JH, Kim J. Nurse staffing and patient mortality in intensive care units. Nurs Res. 2008;57(5):322-330. 36. Grasso S, Terragni P, Birocco A, et al. ECMO criteria for influenza A (H1N1)-associated ARDS: role of transpulmonary pressure. Intensive Care Med. 2012;38(3):395-403. JAMA Internal Medicine July 22, 2013 Volume 173, Number 14 Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013 jamainternalmedicine.com The n e w e ng l a n d j o u r na l of m e dic i n e original article A Trial of Intraoperative Low-Tidal-Volume Ventilation in Abdominal Surgery Emmanuel Futier, M.D., Jean-Michel Constantin, M.D., Ph.D., Catherine Paugam-Burtz, M.D., Ph.D., Julien Pascal, M.D., Mathilde Eurin, M.D., Arthur Neuschwander, M.D., Emmanuel Marret, M.D., Marc Beaussier, M.D., Ph.D., Christophe Gutton, M.D., Jean-Yves Lefrant, M.D., Ph.D., Bernard Allaouchiche, M.D., Ph.D., Daniel Verzilli, M.D., Marc Leone, M.D., Ph.D., Audrey De Jong, M.D., Jean-Etienne Bazin, M.D., Ph.D., Bruno Pereira, Ph.D., and Samir Jaber, M.D., Ph.D., for the IMPROVE Study Group* A BS T R AC T From the Département d’Anesthésie et Ré animation, Hôpital Estaing (E.F., J.-M.C., J.P., J.-E.B.), Université de Clermont-Ferrand, Retinoids, Reproduction, and Developmental Diseases Unit, Équipe Accueil 7281 (E.F., J.-M.C.), and the Biostatistics Unit, Direction de la Recherche Clinique (B.P.), Centre Hospitalier Universitaire (CHU) de Clermont-Ferrand, ClermontFerrand; Assistance Publique–Hôpitaux de Paris (AP-HP), Département d’Anesthé sie et Réanimation, Hôpital Beaujon, Hôpitaux Universitaires Paris Nord Val de Seine and Université Paris Diderot, Sorbonne Paris Cité (C.P.-B., M.E., A.N.), Départe ment d’Anesthésie et Réanimation, Hôpital Tenon (E.M.), and AP-HP, Départe ment d’Anesthésie et Réanimation, Hôpital Saint-Antoine (M.B., C.G.), Paris; CHU de Nîmes, Section d’Anesthésie and Département d’Anesthésie et Réanimation, Nîmes (J.-Y.L.); CHU de Lyon, Département d’Anesthésie et Réanimation, Hôpital Edouard Herriot, Lyon (B.A.); CHU de Montpellier, Département d’Anes thésie et Réanimation B, Hôpital SaintEloi, and INSERM Unité 1046 and Université Montpellier 1, Montpellier (D.V., A.D.J., S.J.); and Assistance Publique–Hôpital de Marseille, Département d’Anesthésie et Réanimation, Hôpital Nord, Marseille (M.L.) — all in France. Address reprint requests to Dr. Jaber at the Département d’Anesthésie et Réanimation B (DAR B), 80 Ave. Augustin Fliche, 34295 Montpellier, France, or at [email protected]. *Additional investigators in the Intraoperative Protective Ventilation (IMPROVE) Study Group are listed in the Supplementary Appendix, available at NEJM.org. N Engl J Med 2013;369:428-37. DOI: 10.1056/NEJMoa1301082 Copyright © 2013 Massachusetts Medical Society. 428 BACKGROUND Lung-protective ventilation with the use of low tidal volumes and positive endexpiratory pressure is considered best practice in the care of many critically ill patients. However, its role in anesthetized patients undergoing major surgery is not known. METHODS In this multicenter, double-blind, parallel-group trial, we randomly assigned 400 adults at intermediate to high risk of pulmonary complications after major abdominal surgery to either nonprotective mechanical ventilation or a strategy of lung-protective ventilation. The primary outcome was a composite of major pulmonary and extrapulmonary complications occurring within the first 7 days after surgery. RESULTS The two intervention groups had similar characteristics at baseline. In the intention-to-treat analysis, the primary outcome occurred in 21 of 200 patients (10.5%) assigned to lung-protective ventilation, as compared with 55 of 200 (27.5%) assigned to nonprotective ventilation (relative risk, 0.40; 95% confidence interval [CI], 0.24 to 0.68; P = 0.001). Over the 7-day postoperative period, 10 patients (5.0%) assigned to lung-protective ventilation required noninvasive ventilation or intubation for acute respiratory failure, as compared with 34 (17.0%) assigned to nonprotective ventilation (relative risk, 0.29; 95% CI, 0.14 to 0.61; P = 0.001). The length of the hospital stay was shorter among patients receiving lung-protective ventilation than among those receiving nonprotective ventilation (mean difference, −2.45 days; 95% CI, −4.17 to −0.72; P = 0.006). CONCLUSIONS As compared with a practice of nonprotective mechanical ventilation, the use of a lung-protective ventilation strategy in intermediate-risk and high-risk patients undergoing major abdominal surgery was associated with improved clinical outcomes and reduced health care utilization. (IMPROVE ClinicalTrials.gov number, NCT01282996.) n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intr aoper ative Low-Tidal-Volume Ventilation W orldwide, more than 230 million patients undergoing major surgery each year require general anesthesia and mechanical ventilation.1 Postoperative pulmonary complications adversely affect clinical outcomes and health care utilization,2 so prevention of these complications has become a measure of the quality of hospital care.3 Previous, large cohort studies have shown that 20 to 30% of patients undergoing surgery with general anesthesia are at intermediate to high risk for postoperative pulmonary complications.4,5 Mechanical ventilation with the use of high tidal volumes (10 to 15 ml per kilogram of predicted body weight) has traditionally been recommended to prevent hypoxemia and atelectasis in anesthetized patients.6 There is, however, considerable evidence from experimental and observational studies that mechanical ventilation — in particular, high tidal volumes that cause alveolar overstretching — can initiate ventilator-associated lung injury7 and contribute to extrapulmonary organ dysfunction through systemic release of inflammatory mediators.8,9 Lung-protective ventilation, which refers to the use of low tidal volumes and positive endexpiratory pressure (PEEP), and which may also include the use of recruitment maneuvers (periodic hyperinflation of the lungs),10 has been shown to reduce mortality among patients with the acute respiratory distress syndrome11 and is now considered best practice in the care of many critically ill patients.12 Although this approach may be beneficial in a broader population,13,14 some physicians have questioned the benefits of using lung-protective ventilation in the surgical setting,15-18 especially since the use of high tidal volumes and no PEEP is still commonplace and less than 20% of patients receive protective ventilation in routine anesthetic practice.19,20 We conducted the Intraoperative Protective Ventilation (IMPROVE) trial to determine whether a multifaceted strategy of prophylactic lungprotective ventilation that combined low tidal volumes, PEEP, and recruitment maneuvers could improve outcomes after abdominal surgery, as compared with the standard practice of nonprotective mechanical ventilation. ME THODS TRIAL DESIGN AND OVERSIGHT The IMPROVE trial was an investigator-initiated, multicenter, double-blind, stratified, parallel-group, clinical trial. Randomization was performed with the use of a computer-generated assignment sequence and a centralized telephone system. The study protocol and statistical analysis plan were approved for all centers by a central ethics committee (Comité de Protection des Personnes Sud-Est I, Saint-Etienne, France) according to French law. The protocol, including the statistical analysis plan, is available with the full text of this article at NEJM.org. An independent data and safety monitoring committee oversaw the study conduct and reviewed blinded safety data. The members of the steering committee (see the Supplementary Appendix, available at NEJM.org) vouch for the accuracy and completeness of the data and analyses and the fidelity of the study to the protocol. There was no industry support or involvement in the trial. Patients were screened and underwent randomization between January 31, 2011, and August 10, 2012, at seven French university teaching hospitals. Written informed consent was obtained before randomization from each patient, on the day before surgery. Randomization was stratified according to study site and the planned use or nonuse of postoperative epidural analgesia, which is a factor that may influence outcomes.21 Treatment assignments were concealed from patients, research staff, the statistician, and the data and safety monitoring committee. Although the staff members who collected data during surgery were aware of the group assignments, outcome assessors were unaware of these assignments throughout the study. PATIENTS Patients were eligible for participation in the study if they were older than 40 years of age, were scheduled to undergo laparoscopic or nonlaparoscopic elective major abdominal surgery1 with an expected duration of at least 2 hours, and had a preoperative risk index for pulmonary complications5 of more than 2. The risk index uses risk classes that range from 1 to 5, with higher risk classes indicating a higher risk of postoperative pulmonary complications (see the n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 429 The n e w e ng l a n d j o u r na l Supplementary Appendix). Patients were ineligible if they had received mechanical ventilation within the 2 weeks preceding surgery, had a body-mass index (the weight in kilograms divided by the square of the height in meters) of 35 or higher, had a history of respiratory failure or sepsis within the 2 weeks preceding surgery, had a requirement for intrathoracic or emergency surgery, or had a progressive neuromuscular illness. INTERVENTIONS Patients were assigned to receive volume-controlled mechanical ventilation according to one of two strategies: nonprotective ventilation with a tidal volume of 10 to 12 ml per kilogram of predicted body weight, with no PEEP and no recruitment maneuvers, as previously described20 (the nonprotective-ventilation group), or lungprotective ventilation with a tidal volume of 6 to 8 ml per kilogram of predicted body weight, a PEEP of 6 to 8 cm of water, and recruitment maneuvers repeated every 30 minutes after tracheal intubation (the protective-ventilation group). Each recruitment maneuver consisted of applying a continuous positive airway pressure of 30 cm of water for 30 seconds. During anesthesia, a plateau pressure of no more than 30 cm of water was targeted in each group. All other ventilation procedures were identical in the two study groups (see the Supplementary Appendix). The predicted body weight was calculated for each patient with the use of previously defined formulas.11 For episodes of arterial desaturation (defined as a peripheral oxygen saturation of ≤92%), a transient increase in the fraction of inspired oxygen (Fio2) to 100% was permitted, and in patients assigned to nonprotective ventilation, the use of PEEP, recruitment maneuvers, or both was allowed, if required. Decisions about all other aspects of patient care during the intraoperative and postoperative periods, including general anesthesia, administration of fluids, use of prophylactic antibiotic agents, and postoperative pain management, were made by the attending physician according to the expertise of the staff at each center and routine clinical practice. OUTCOMES The primary outcome was a composite of major pulmonary and extrapulmonary complications occurring by day 7 after surgery. Major pulmonary complications were defined as pneumonia 430 of m e dic i n e (defined according to standard criteria; see the Supplementary Appendix) or the need for invasive or noninvasive ventilation for acute respiratory failure. Major extrapulmonary complications were defined as sepsis, severe sepsis and septic shock (defined according to consensus criteria),22 or death. Secondary outcomes within the 30-day follow-up period were the incidence of pulmonary complications due to any cause, graded on a scale from 0 (no pulmonary complications) to 4 (the most severe complications)23 (see the Supplementary Appendix); ventilation-related adverse events during surgery; postoperative gas exchange; unexpected need for admission to the intensive care unit (ICU); extrapulmonary complications; durations of ICU and hospital stays; and the rate of death from any cause 30 days after surgery. Pulmonary complications were analyzed separately; in particular, the need for invasive or noninvasive ventilation because of acute respiratory failure, the development of postoperative atelectasis, pneumonia, acute lung injury, and the acute respiratory distress syndrome, defined according to standard criteria (see the Supplementary Appendix). Extrapulmonary complications included the systemic inflammatory response syndrome (SIRS); sepsis; severe sepsis and septic shock; and surgical complications, including intraabdominal abscess, anastomotic leakage, and unplanned reoperation (all defined according to consensus criteria22,24). STATISTICAL ANALYSIS We calculated that a sample of 400 patients would provide 80% power to detect a relative difference of 50% in the primary outcome, at a twosided alpha level of 0.05, assuming a 20% rate of postoperative complications in the nonprotectiveventilation group.25 For safety reasons, an interim analysis was conducted after the enrollment of the first 200 patients, according to the a priori statistical analysis plan. The data and safety monitoring committee did not recommend discontinuation of the trial on the basis of that analysis, and 400 patients were therefore included. A total of 3 patients were excluded after randomization; surgery was stopped prematurely in 2 of the 3 patients because of extensive illness (duration of surgery, <2 hours), and 1 had undergone randomization in error (violation of exclusion criteria). An additional 3 patients were thus randomly assigned to a study group to obtain the full sample. n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intr aoper ative Low-Tidal-Volume Ventilation All analyses were conducted on data from the modified intention-to-treat population, which included all patients who underwent randomization except the three who were excluded (Fig. 1). An un adjusted chi-square test was used for the primary outcome analysis. Multiple logistic-regression analysis was used to identify relevant baseline covariates associated with the primary outcome, in addition to the stratification variables (use or nonuse of epidural analgesia and study center). Variables tested in the model were selected if the P value was less than 0.10 and if they were clinically relevant. Adjusted analyses were performed with the use of robust Poisson generalized-linear-model re gression26 and are presented as relative risks with 95% confidence intervals. A chi-square test (or Fisher’s exact test, as appropriate) was used for secondary binary outcomes. The Hochberg procedure was used to adjust for multiple testing of components of the composite primary outcome.27 Adjusted analyses were performed with the use of the same adjustment variables that were used in the robust Poisson regression analysis. Continuous variables were compared with the use of an unpaired t-test or the Mann–Whitney U test. Adjusted analyses were performed with the use 1803 Patients were scheduled to undergo elective abdominal surgery 1202 Were ineligible 47 Were ≤40 yr of age 102 Had expected duration of surgery of <2 hr 1053 Had preoperative risk index for pulmonary complications of <2 601 Underwent screening 198 Were excluded 49 Declined to participate 149 Met exclusion criteria 400 Underwent randomization 3 Were excluded after randomization 2 Had extensive illness (surgery <2 hr) 1 Had violation of exclusion criteria (age ≤40 yr) An additional 3 underwent randomization 200 Were assigned to receive nonprotective mechanical ventilation 200 Were assigned to receive lung-protective mechanical ventilation 200 (100%) Were included in 30-day analysis 200 (100%) Were included in 30-day analysis Figure 1. Assessment, Randomization, and Follow-up of Patients. A total of 1803 patients awaiting abdominal surgery were assessed preoperatively for trial eligibility by the research staff. A total of 400 patients were included in the modified intention-to-treat analysis and were followed for 30 days after surgery. After randomization, 3 patients were excluded; 2 patients were excluded because surgery was stopped prematurely (duration, <2 hours), owing to extensive illness, and 1 had undergone randomization in error. An additional 3 patients were then enrolled in the study. n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 431 The n e w e ng l a n d j o u r na l Table 1. Baseline Characteristics of the Patients.* Nonprotective Ventilation (N = 200) Characteristic Lung-Protective Ventilation (N = 200) Age — yr 63.4±10.0 61.6±11.0 Male sex — no. (%) 121 (60.5) 116 (58.0) Height — cm 169.5±9.0 169.1±8.8 Actual 71.3±13.9 71.4±14.2 Predicted† 63.8±9.9 63.3±9.7 24.7±3.8 24.8±3.8 Body weight — kg Body-mass index‡ Mean 25–35 — no. (%) 88 (44.0) 99 (49.5) Risk class 2 100 (50.0) 101 (50.5) Risk class 3 94 (47.0) 93 (46.5) 6 (3.0) 6 (3.0) Current smoking 50 (25.0) 51 (25.5) Any alcohol intake 10 (5.0) 21 (10.5) 8 (4.0) 8 (4.0) Chronic obstructive pulmonary disease 20 (10.0) 20 (10.0) Loss of >10% of body weight in previous 6 mo 44 (22.0) 40 (20.0) 4 (2.0) 7 (3.5) 44 (22.0) 41 (20.5) Pancreaticoduodenectomy 80 (40.0) 84 (42.0) Liver resection 52 (26.0) 44 (22.0) Gastrectomy 17 (8.5) 15 (7.5) Colorectal resection 40 (20.0) 47 (23.5) Other procedure 11 (5.5) 10 (5.0) 164 (82.0) 155 (77.5) 36 (18.0) 45 (22.5) Preoperative risk index — no. (%)§ Risk class 4 or 5 Coexisting condition — no. (%)¶ Not fully independent in activities of daily living Long-term glucocorticoid use Laparoscopic surgery — no. (%) Type of surgery — no. (%) Diagnosis — no. (%) Cancer Diagnosis other than cancer *Plus–minus values are means ±SD. There were no significant differences between the two groups (P>0.05). †The predicted body weight was calculated as follows: for men, 50+0.91(height in centimeters − 152.4); and for women, 45.5 + 0.91(height in centimeters − 152.4).11 ‡The body-mass index is the weight in kilograms divided by the square of the height in meters. § The preoperative risk index for pulmonary complications5 uses risk classes that range from 1 to 5, with higher risk classes indicating a higher risk of postoperative complications. Patients with a risk class of 2 or more were eligible for participation in the study. ¶All factors listed as coexisting conditions were included in the preoperative risk index as predictors of postoperative pulmonary complications. 432 of m e dic i n e of the same adjustment variables that were used in the linear-regression model. The time-to-event curves were calculated with the Kaplan–Meier method. Details regarding the handling of missing data are provided in the Supplementary Appendix. All analyses were conducted with the use of Stata software, version 12 (StataCorp). A two-sided P value of less than 0.05 was considered to indicate statistical significance. R E SULT S STUDY POPULATION From January 2011 through August 2012, a total of 1803 patients awaiting abdominal surgery were assessed for trial eligibility. A total of 400 patients were included in the modified intentionto-treat analysis and were followed for 30 days after surgery (Fig. 1). One patient in the nonprotective-ventilation group received lung-protective ventilation but was included in the analysis for the group to which he was assigned. Data on the primary outcome were available for all patients. Baseline characteristics were similar between the two groups (Table 1). Open laparotomy, mainly for cancer resection, was performed in 156 patients (78.0%) in the nonprotective-ventilation group and in 159 (79.5%) in the protective-ventilation group (P = 0.80). INTRAOPERATIVE PROCEDURES Table 2 shows the distribution of the main intraoperative procedures. Mean (±SD) tidal volumes were 11.1±1.1 ml per kilogram in the nonprotective-ventilation group, as compared with 6.4±0.8 ml per kilogram in the protective-ventilation group (P<0.001), and values remained within target ranges throughout the intraoperative period. In the protective-ventilation group, the median PEEP was 6 cm of water (interquartile range, 6 to 8), and the median number of recruitment maneuvers was 9 (interquartile range, 6 to 12); in the nonprotective-ventilation group, the value for each of these measures was 0 (interquartile range, 0 to 0) (Table 2). There were no significant betweengroup differences in type and duration of surgery, use or nonuse of epidural analgesia, blood loss, volume of fluids administered, and need for vasopressor administration. Five patients in the nonprotective-ventilation group required at least one intraoperative rescue therapy for arterial de- n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intr aoper ative Low-Tidal-Volume Ventilation Table 2. Intraoperative Procedures.* Variable Tidal volume — ml Tidal volume — ml/kg of predicted body weight Nonprotective Ventilation (N = 200) Lung-Protective Ventilation (N = 200) P Value 719.0±127.8 406.7±75.6 <0.001 11.1±1.1 6.4±0.8 <0.001 PEEP — cm of water Baseline <0.001 Median Interquartile range 0 6 0–0 6–8 End of surgery <0.001 Median Interquartile range 0 6 0–0 6–8 No. of recruitment maneuvers <0.001 Median 0 9 0–0 6–12 Baseline 20.1±4.9 18.9±3.6 0.04 End of surgery 20.6±4.4 20.0±4.0 0.15 Baseline 16.1±4.3 15.2±3.0 0.02 End of surgery 16.6±3.5 15.2±2.6 <0.001 Baseline 48.4±17.8 55.2±26.6 0.06 End of surgery 45.1±12.9 55.2±26.7 <0.001 47.2±7.6 46.4±7.3 0.27 Interquartile range Peak pressure — cm of water Plateau pressure — cm of water Respiratory system compliance — ml/cm of water Fio2 — % Volume of fluids administered — liters Crystalloid 0.47 Median Interquartile range 2.0 1.5 1.5–3.5 2.0–3.0 Colloid 0.97 Median Interquartile range 0.5 0.5 0.25–1.0 0.50–1.0 Duration of surgery — no./total no. (%)† 0.95 2–4 hr 76/192 (39.6) 75/195 (38.5) >4–6 hr 75/192 (39.1) 76/195 (39.0) >6 hr 41/192 (21.4) 44/195 (22.6) Duration of mechanical ventilation — min 344±127.9 319±139.4 0.84 Epidural analgesia — no. (%) 77 (38.5) 83 (41.5) 0.61 *Plus–minus values are means ±SD. Detailed data on intraoperative procedures are given in Table S2 in the Supplementary Appendix. Fio2 denotes inspired oxygen fraction, and PEEP positive end-expiratory pressure. †The duration of surgery was calculated as the time between skin incision and closure of the incision. n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 433 The n e w e ng l a n d j o u r na l of m e dic i n e saturation (PEEP in one patient, recruitment ma- ventilation (P<0.001 by the log-rank test) (Fig. S1 neuvers in two, and both in two), as compared in the Supplementary Appendix). with no patients in the protective-ventilation There was no significant difference between group (P = 0.06). the protective-ventilation group and the nonprotective-ventilation group with respect to the proOUTCOMES portion of patients who were unexpectedly adPrimary Outcome mitted to the ICU during the 30-day period after Major pulmonary and extrapulmonary complica- surgery (11.0 and 12.5%, respectively; adjusted relations occurred within the first 7 days after sur- tive risk with protective ventilation, 0.88; 95% CI, gery in 21 patients (10.5%) in the protective-ven- 0.49 to 1.59; P = 0.67), nor was there a significant tilation group, as compared with 55 (27.5%) in the between-group difference in the rate of adverse nonprotective-ventilation group (adjusted relative events (Table S3 in the Supplementary Appendix). risk, 0.40; 95% confidence interval [CI], 0.24 to Mortality at 30 days in the protective-ventilation 0.68; P = 0.001) (Table 3). The results of associated group was similar to that in the nonprotectiveunivariate and multivariate analyses are provided ventilation group (3.0% and 3.5%, respectively; in Table S1 in the Supplementary Appendix. adjusted relative risk with protective ventilation, 1.13; 95% CI, 0.36 to 3.61; P = 0.83). However, the median hospital stay was shorter in the protecSecondary Outcomes One or more pulmonary complications developed tive-ventilation group than in the nonprotectivewithin the first 7 days after surgery in 35 patients ventilation group (Table 3). (17.5%) in the protective-ventilation group, as compared with 72 (36.0%) in the nonprotective-ventilaDISCUSSION tion group (adjusted relative risk, 0.49; 95% CI, 0.32 to 0.74; P<0.001). More patients in the non- In this trial, intraoperative lung-protective meprotective-ventilation group than in the protec- chanical ventilation, as compared with non tive-ventilation group had major (grade ≥3) pul- protective ventilation, led to improved clinical monary complications (Table 3, and Tables S3 and outcomes and reduced health care utilization S4 in the Supplementary Appendix) and major after abdominal surgery. The observed rate of pulmonary and extrapulmonary complications dur- postoperative complications in our study was ing the 30 days after surgery (P<0.001 by the log- slightly higher than predicted.25 This was due, rank test) (Fig. 2). There were no relevant between- in part, to the exclusion of patients with a low group differences in gas exchange after extubation risk of complications, as well as the large proand on day 1 after surgery (Table S5 in the Supple- portion of patients who underwent major abmentary Appendix). dominal procedures, which are associated with The proportion of patients who required post- increased morbidity rates. Of the 400 patients operative ventilatory assistance (noninvasive ven- enrolled, 19 had postoperative pneumonia and tilation or intubation) for acute respiratory failure 47 had respiratory failure requiring intubation was lower in the protective-ventilation group than or noninvasive ventilation. These rates are conin the nonprotective-ventilation group during the sistent with previously reported rates of pulmofirst 7 days after surgery (10 of 200 patients [5.0%], nary complications25,28 and mortality.29 Our lungvs. 34 of 200 [17.0%]; adjusted relative risk, 0.29; protective ventilation strategy resulted in a 69% 95% CI, 0.14 to 0.61; P = 0.001), and the proportion reduction in the number of patients requiring was also lower with protective ventilation during ventilatory support within the first 7 days after the first 13 days after surgery (6.5% vs. 18.5%; surgery. adjusted relative risk, 0.36; 95% CI, 0.19 to 0.70; Several hypotheses could explain some of P = 0.003) (Table 3). In addition, the cumulative the differences between the results of the pres30-day probability of an event requiring intuba- ent study and findings in other trials of lungtion or noninvasive ventilation for postoperative protective ventilation during high-risk surgery. acute respiratory failure was lower among pa- Previous trials have included small numbers of tients who received lung-protective ventilation patients, have focused on different (and not than among those who received nonprotective necessarily clinically relevant) outcomes,17 and 434 n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intr aoper ative Low-Tidal-Volume Ventilation Table 3. Results of Unadjusted and Adjusted Outcome Analyses.* Nonprotective Lung-Protective Ventilation Ventilation (N = 200) (N = 200) Variable Unadjusted Relative Risk or Between-Group Difference (95% CI) P Value† Adjusted Relative Risk or Between-Group Difference (95% CI)‡ P Value Primary composite outcome — no. (%) Within 7 days§ 55 (27.5) 21 (10.5) 0.38 (0.24–0.61) <0.001 0.40 (0.24–0.68) 0.001 Within 30 days 58 (29.0) 25 (12.5) 0.43 (0.28–0.66) <0.001 0.45 (0.28–0.73) <0.001 Grade 1 or 2 30 (15.0) 25 (12.5) 0.69 (0.42–1.13) 0.14 0.67 (0.39–1.16) 0.16 Grade ≥3 42 (21.0) 10 (5.0) 0.24 (0.12–0.46) <0.001 0.23 (0.11–0.49) <0.001 Atelectasis within 7 days‖ 34 (17.0) 13 (6.5) 0.38 (0.21–0.70) 0.001 0.37 (0.19–0.73) 0.004 Pneumonia within 7 days 16 (8.0) 3 (1.5) 0.19 (0.05–0.63) 0.01 0.19 (0.05–0.66) 0.009 6 (3.0) 1 (0.5) 0.17 (0.02–1.37) 0.12 0.21 (0.02–1.71) 0.14 7 (3.5) 2 (1.0) 0.29 (0.06–1.36) 0.51 0.40 (0.08–1.97) 0.26 29 (14.5) 9 (4.5) 0.31 (0.15–0.64) 0.006 0.29 (0.13–0.65) 0.002 100 (50.0) 86 (43.0) 0.86 (0.70–1.06) 0.16 0.87 (0.65–1.17) 0.37 29 (14.5) 13 (6.5) 0.45 (0.24–0.84) 0.04 0.48 (0.25–0.93) 0.03 9 (4.5) 8 (4.0) 0.89 (0.35–2.26) 0.80 1.48 (0.51–4.32) 0.47 7 (3.5) 6 (3.0) 0.86 (0.29–2.51) 0.80 1.13 (0.36–3.61) 0.83 13 11 8–20 8–15 Secondary outcomes — no. (%) Pulmonary complication within 7 days¶ Acute lung injury or ARDS within 7 days Need for ventilation within 7 days Invasive Noninvasive Extrapulmonary complication within 7 days SIRS Sepsis Severe sepsis or septic shock Death within 30 days Duration of stay in hospital and ICU — days Hospital Median Interquartile range 0.02 −2.25 (−4.04 to −0.47) ICU Median Interquartile range 0.006 −2.45 (−4.17 to −0.72) 0.58 7 6 4–9 4–8 −1.48 (−6.87 to 3.91) 0.69 −1.21 (−4.98 to 7.40) *All postoperative complications were defined according to consensus criteria (see the Supplementary Appendix). For additional data on postoperative outcomes, see Tables S3 and S4 in the Supplementary Appendix. ARDS denotes acute respiratory distress syndrome, CI confidence interval, ICU intensive care unit, and SIRS systemic inflammatory response syndrome. Relative risks are shown for outcome variables, and differences between groups are shown for the duration of stays in the hospital and ICU. †Adjustment was performed for stratification variables (use or nonuse of epidural analgesia and study center), preoperative risk index for postoperative pulmonary complications, sex, duration of surgery, and need for blood transfusion (yes or no). ‡The Hochberg procedure was used to adjust for multiple testing of components of the composite primary outcome.27 § The primary outcome was a composite of major pulmonary complications (defined as pneumonia or need for invasive or noninvasive ventilation for acute respiratory failure) and extrapulmonary complications (defined as sepsis, septic shock, or death) within the first 7 days after surgery. ¶Postoperative pulmonary complications were scored with the use of a graded scale23 from 0 (no pulmonary complications) to 4 (the most severe complications) (see the Supplementary Appendix). ‖Atelectasis was defined as opacification of the lung with shift of the mediastinum, hilum, or hemidiaphragm toward the affected area and compensatory overinflation in the adjacent, nonatelectatic lung. have used either very low levels of PEEP or no PEEP.15,16,18 One strength of the present trial is our use of a robust composite outcome that is highly pertinent to this high-risk surgical population.5 Mechanical ventilation itself can induce an inflammatory response30 and can synergize with the response induced by major surgery at both local and systemic levels. This amplification of n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 435 The n e w e ng l a n d j o u r na l 0.50 Probability of Event 0.40 Nonprotective ventilation 0.30 0.20 Lung-protective ventilation 0.10 0.00 1 3 7 15 30 Days since Randomization No. at Risk Nonprotective ventilation Lung-protective ventilation 200 182 163 145 142 142 200 192 184 179 176 175 Figure 2. Kaplan–Meier Estimates of the Probability of the Composite Primary Outcome. Data for the Kaplan–Meier estimates of the probability of the composite primary outcome of major pulmonary or extrapulmonary complications were censored at 30 days after surgery. Major pulmonary complications included pneumonia or the need for invasive or noninvasive ventilation for acute respiratory failure. Major extrapulmonary complications were sepsis, severe sepsis, septic shock, and death. P<0.001 by the log-rank test for the between-group difference in the probability of the primary outcome. the inflammation cascade contributes to the subsequent development of lung injury31 and sys temic organ failure.8,32 The use of very low levels of PEEP in previous trials may have promoted the repeated opening and closing of small airways, leading to atelectasis, which can precipitate the development of pulmonary complications.6,33 We used a multifaceted strategy of lung-protective ventilation that combined low tidal volumes, recruitment maneuvers to open collapsed alveoli, and moderate levels of PEEP to prevent further collapse.34 Other strengths of the present trial include the methods used to minimize bias (blinded and centralized randomization, complete follow-up, and intention-to-treat analyses); the pragmatic nature of the trial protocol, with routine practice being maintained; and the enrollment of patients with characteristics similar to those of patients enrolled in other studies analyzing outcomes after major surgery.29 Our findings are consistent with the observation of transient arterial hypotension during recruitment maneuvers.35 Consequently, recruit436 of m e dic i n e ment maneuvers, in which hemodynamic effects are potentially influenced by the applied level of alveolar pressure,36 should be used with caution in patients with hemodynamic instability. There are several limitations to our study. The trial design did not include standardization of the administration of fluids. However, this limitation is unlikely to have affected our results, since the volume of fluids administered was similar in the two groups. The definition of nonprotective ventilation was arbitrary but is supported in the literature.19,20 The trial protocol did not include standardization of requirements for noninvasive ventilation; however, it was recommended that the study centers follow clinicalpractice guidelines,37,38 and postoperative care was conducted by health care workers who were unaware of the study assignments. The utilization of noninvasive ventilation in our trial is close to that reported in earlier studies.37 We therefore consider it unlikely that any imbalance in interventions affected our results. In conclusion, our study provides evidence that a multifaceted strategy of prophylactic lung-protective ventilation during surgery, as compared with a practice of nonprotective mechanical ventilation, results in fewer postoperative complications and reduced health care utilization. Dr. Futier reports receiving consulting fees from General Electric Medical Systems, lecture fees from Fresenius Kabi, and reimbursement of travel expenses from Fisher and Paykel Healthcare. Dr Constantin reports receiving consulting fees from Baxter, Fresenius Kabi, Dräger, and General Electric Medical Systems, payment for expert testimony from Baxter, Dräger, and Fresenius Kabi, lecture fees from General Electric Medical Systems, Dräger, Fresenius Kabi, Baxter, Hospal, Merck Sharp & Dohme, and LFB Biomedicaments, payment for the development of educational presentations from Dräger, General Electric Medical Systems, Baxter, and Fresenius Kabi, and reimbursement of travel expenses from Bird, Astute Medical, Astellas, Fresenius Kabi, Baxter, and Hospal. Dr. Paugam-Burtz reports receiving consulting fees from Fresenius Kabi, lecture fees and reimbursement of travel expenses from Astellas, and payment for the development of educational presentations from LFB Biomedicaments and Merck Sharp & Dohme. Dr. Allaouchiche reports receiving consulting fees from Fresenius Kabi and lecture fees from Novartis and Astellas. Dr. Leone reports receiving consulting fees from LFB Biomedicaments and lecture fees from Fresenius Kabi and Novartis. Dr. Jaber reports receiving consulting fees from Dräger France and Maquet France, lecture fees from Fisher and Paykel Healthcare, Abbott, and Philips, and reimbursement of travel expenses from Pfizer. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank all the patients who participated in the study; the clinical and research staff at all the trial sites, without whose assistance the study would never have been completed; and Mervyn Singer for valuable advice during the preparation of the manuscript. n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Intr aoper ative Low-Tidal-Volume Ventilation References 1. Weiser TG, Regenbogen SE, Thomp- son KD, et al. An estimation of the global volume of surgery: a modelling strategy based on available data. Lancet 2008;372: 139-44. 2. Khuri SF, Henderson WG, DePalma RG, et al. Determinants of long-term survival after major surgery and the adverse effect of postoperative complications. Ann Surg 2005;242:326-41. 3. Shander A, Fleisher LA, Barie PS, Bigtello LM, Sladen RN, Watson CB. Clinical and economic burden of postoperative pulmonary complications: patient safety summit on definition, risk-reducing interventions, and preventive strategies. Crit Care Med 2011;39:2163-72. 4. Arozullah AM, Daley J, Henderson WG, Khuri SF. Multifactorial risk index for predicting postoperative respiratory failure in men after major noncardiac surgery: the National Veterans Administration Surgical Quality Improvement Program. Ann Surg 2000;232:242-53. 5. Arozullah AM, Khuri SF, Henderson WG, Daley J. Development and validation of a multifactorial risk index for predicting postoperative pneumonia after major noncardiac surgery. Ann Intern Med 2001;135:847-57. 6. Bendixen HH, Hedley-Whyte J, Laver MB. Impaired oxygenation in surgical patients during general anesthesia with con trolled ventilation: a concept of atelectasis. N Engl J Med 1963;269:991-6. 7. Serpa Neto A, Cardoso SO, Manetta JA, et al. Association between use of lungprotective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome: a meta-analysis. JAMA 2012; 308:1651-9. 8. Imai Y, Parodo J, Kajikawa O, et al. Injurious mechanical ventilation and endorgan epithelial cell apoptosis and organ dysfunction in an experimental model of acute respiratory distress syndrome. JAMA 2003;289:2104-12. 9. Lellouche F, Dionne S, Simard S, Bussieres J, Dagenais F. High tidal volumes in mechanically ventilated patients increase organ dysfunction after cardiac surgery. Anesthesiology 2012;116:1072-82. 10. Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998; 338:347-54. 11. 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;342:1301-8. 12. Putensen C, Theuerkauf N, Zinserling J, Wrigge H, Pelosi P. Meta-analysis: ven- tilation strategies and outcomes of the acute respiratory distress syndrome and acute lung injury. Ann Intern Med 2009; 151:566-76. [Erratum, Ann Intern Med 2009;151:897.] 13. Schultz MJ, Haitsma JJ, Slutsky AS, Gajic O. What tidal volumes should be used in patients without acute lung injury? Anesthesiology 2007;106:1226-31. 14. Severgnini P, Selmo G, Lanza C, et al. Protective mechanical ventilation during general anesthesia for open abdominal surgery improves postoperative pulmonary function. Anesthesiology 2013 March 29 (Epub ahead of print). 15. Wrigge H, Uhlig U, Zinserling J, et al. The effects of different ventilatory settings on pulmonary and systemic inflammatory responses during major surgery. Anesth Analg 2004;98:775-81. 16. Wrigge H, Uhlig U, Baumgarten G, et al. Mechanical ventilation strategies and inflammatory responses to cardiac surgery: a prospective randomized clinical trial. Intensive Care Med 2005;31:1379-87. 17. Hong CM, Xu DZ, Lu Q, et al. Low tidal volume and high positive end-expiratory pressure mechanical ventilation results in increased inflammation and ventilator-associated lung injury in normal lungs. Anesth Analg 2010;110:1652-60. 18. Treschan TA, Kaisers W, Schaefer MS, et al. Ventilation with low tidal volumes during upper abdominal surgery does not improve postoperative lung function. Br J Anaesth 2012;109:263-71. 19. Jaber S, Coisel Y, Chanques G, et al. A multicentre observational study of intraoperative ventilatory management during general anaesthesia: tidal volumes and relation to body weight. Anaesthesia 2012; 67:999-1008. 20. Hess DR, Kondili D, Burns E, Bittner EA, Schmidt UH. A 5-year observational study of lung-protective ventilation in the operating room: a single-center experience. J Crit Care 2013 January 29 (Epub ahead of print). 21. Pöpping DM, Elia N, Marret E, Remy C, Tramèr MR. Protective effects of epidural analgesia on pulmonary complications after abdominal and thoracic surgery: a meta-analysis. Arch Surg 2008;143:990-9. 22. Bone RC, Sibbald WJ, Sprung CL. The ACCP-SCCM consensus conference on sepsis and organ failure. Chest 1992;101: 1481-3. 23. Hulzebos EH, Helders PJ, Favié NJ, et al. Preoperative intensive inspiratory muscle training to prevent postoperative pulmonary complications in high-risk patients undergoing CABG surgery: a randomized clinical trial. JAMA 2006;296:1851-7. 24. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. 25. Thompson JS, Baxter BT, Allison JG, Johnson FE, Lee KK, Park WY. Temporal patterns of postoperative complications. Arch Surg 2003;138:596-602. 26. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004;159:702-6. 27. Hochberg Y. A sharper Bonferroni procedure for multiple tests of significance. Biometrika 1988;75:800-2. 28. Lawrence VA, Dhanda R, Hilsenbeck SG, Page CP. Risk of pulmonary complications after elective abdominal surgery. Chest 1996;110:744-50. 29. Pearse RM, Moreno RP, Bauer P, et al. Mortality after surgery in Europe: a 7 day cohort study. Lancet 2012;380:1059-65. 30. Ranieri VM, Suter PM, Tortorella C, et al. Effect of mechanical ventilation on inflammatory mediators in patients with acute respiratory distress syndrome: a randomized controlled trial. JAMA 1999; 282:54-61. 31. Mascia L, Pasero D, Slutsky AS, et al. Effect of a lung protective strategy for organ donors on eligibility and availability of lungs for transplantation: a randomized controlled trial. JAMA 2010;304: 2620-7. 32. Bouadma L, Dreyfuss D, Ricard JD, Martet G, Saumon G. Mechanical ventilation and hemorrhagic shock-resuscitation interact to increase inflammatory cytokine release in rats. Crit Care Med 2007; 35:2601-6. 33. Duggan M, Kavanagh BP. Pulmonary atelectasis: a pathogenic perioperative entity. Anesthesiology 2005;102:838-54. 34. Futier E, Constantin JM, Pelosi P, et al. Noninvasive ventilation and alveolar recruitment maneuver improve respiratory function during and after intubation of morbidly obese patients: a randomized controlled study. Anesthesiology 2011;114: 1354-63. 35. Fan E, Wilcox ME, Brower RG, et al. Recruitment maneuvers for acute lung injury: a systematic review. Am J Respir Crit Care Med 2008;178:1156-63. 36. Lim SC, Adams AB, Simonson DA, et al. Transient hemodynamic effects of recruitment maneuvers in three experimental models of acute lung injury. Crit Care Med 2004;32:2378-84. 37. Jaber S, Chanques G, Jung B. Post operative noninvasive ventilation. Anesthesiology 2010;112:453-61. 38. Keenan SP, Sinuff T, Burns KE, et al. Clinical practice guidelines for the use of noninvasive positive-pressure ventilation and noninvasive continuous positive airway pressure in the acute care setting. CMAJ 2011;183(3):E195-E214. Copyright © 2013 Massachusetts Medical Society. n engl j med 369;5 nejm.org august 1, 2013 The New England Journal of Medicine Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 437 CARING FOR THE CRITICALLY ILL PATIENT Association Between Use of Lung-Protective Ventilation With Lower Tidal Volumes and Clinical Outcomes Among Patients Without Acute Respiratory Distress Syndrome A Meta-analysis Ary Serpa Neto, MD, MSc Sérgio Oliveira Cardoso, MD José Antônio Manetta, MD Victor Galvão Moura Pereira, MD Daniel Crepaldi Espósito, MD Manoela de Oliveira Prado Pasqualucci, MD Maria Cecı́lia Toledo Damasceno, MD, PhD Marcus J. Schultz, MD, PhD M ECHANICAL VENTILATION is a life-saving strategy in patients with acute respiratory failure. However, unequivocal evidence suggests that mechanical ventilation has the potential to aggravate and precipitate lung injury.1 In acute respiratory distress syndrome (ARDS), and in a milder form of ARDS formerly known as acute lung injury (ALI),2 mechanical ventilation can cause ventilatorassociated lung injury. Ventilatorassociated lung injury is a frequent complication in critically ill patients receiving mechanical ventilation, and its development increases morbidity and mortality.1 Higher tidal volume (VT) ventilation causes the alveoli to overstretch in a process called volutrauma, and this overstretching is the main cause of ventilator-associated lung injury.3 The use of a lower VT was shown to reduce morbidity and mortality in For editorial comment see p 1689. Context Lung-protective mechanical ventilation with the use of lower tidal volumes has been found to improve outcomes of patients with acute respiratory distress syndrome (ARDS). It has been suggested that use of lower tidal volumes also benefits patients who do not have ARDS. Objective To determine whether use of lower tidal volumes is associated with improved outcomes of patients receiving ventilation who do not have ARDS. Data Sources MEDLINE, CINAHL, Web of Science, and Cochrane Central Register of Controlled Trials up to August 2012. Study Selection Eligible studies evaluated use of lower vs higher tidal volumes in patients without ARDS at onset of mechanical ventilation and reported lung injury development, overall mortality, pulmonary infection, atelectasis, and biochemical alterations. Data Extraction Three reviewers extracted data on study characteristics, methods, and outcomes. Disagreement was resolved by consensus. Data Synthesis Twenty articles (2822 participants) were included. Meta-analysis using a fixed-effects model showed a decrease in lung injury development (risk ratio [RR], 0.33; 95% CI, 0.23 to 0.47; I2, 0%; number needed to treat [NNT], 11), and mortality (RR, 0.64; 95% CI, 0.46 to 0.89; I2, 0%; NNT, 23) in patients receiving ventilation with lower tidal volumes. The results of lung injury development were similar when stratified by the type of study (randomized vs nonrandomized) and were significant only in randomized trials for pulmonary infection and only in nonrandomized trials for mortality. Metaanalysis using a random-effects model showed, in protective ventilation groups, a lower incidence of pulmonary infection (RR, 0.45; 95% CI, 0.22 to 0.92; I2, 32%; NNT, 26), lower mean (SD) hospital length of stay (6.91 [2.36] vs 8.87 [2.93] days, respectively; standardized mean difference [SMD], 0.51; 95% CI, 0.20 to 0.82; I2, 75%), higher mean (SD) PaCO2 levels (41.05 [3.79] vs 37.90 [4.19] mm Hg, respectively; SMD, −0.51; 95% CI, −0.70 to −0.32; I2, 54%), and lower mean (SD) pH values (7.37 [0.03] vs 7.40 [0.04], respectively; SMD, 1.16; 95% CI, 0.31 to 2.02; I2, 96%) but similar mean (SD) ratios of PaO2 to fraction of inspired oxygen (304.40 [65.7] vs 312.97 [68.13], respectively; SMD, 0.11; 95% CI, −0.06 to 0.27; I2, 60%). Tidal volume gradients between the 2 groups did not influence significantly the final results. Conclusions Among patients without ARDS, protective ventilation with lower tidal volumes was associated with better clinical outcomes. Some of the limitations of the meta-analysis were the mixed setting of mechanical ventilation (intensive care unit or operating room) and the duration of mechanical ventilation. www.jama.com JAMA. 2012;308(16):1651-1659 Author Affiliations: Department of Critical Care Medicine, ABC Medical School, Santo André, São Paulo, Brazil (Drs Serpa Neto, Cardoso, Manetta, Pereira, Espósito, and Damasceno); Department of Internal Medicine, Hospital das Clı́nicas, University of São Paulo, São Paulo (Dr Damasceno); and Department of Intensive Care Medicine and Laboratory of Experimental Intensive Care and Anesthesiology, Academic Medical ©2012 American Medical Association. All rights reserved. Center, University of Amsterdam, Amsterdam, the Netherlands (Dr Schultz). Corresponding Author: Ary Serpa Neto, MD, MSc, Avenue Lauro Gomes, 2000 São Paulo, Brazil (aryserpa @terra.com.br). Caring for the Critically Ill Patient Section Editor: Derek C. Angus, MD, MPH, Contributing Editor, JAMA ([email protected]). JAMA, October 24/31, 2012—Vol 308, No. 16 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 1651 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES patients with ARDS or ALI, thus justifying the progressive decrease in VT used by clinicians over the past decades.4-6 However, in critically ill patients without ALI, there is little evidence regarding the benefits of ventilation with lower V T , partly because of a lack of randomized controlled trials evaluating the best ventilator strategies in these patients.7 Some observational studies have suggested that use of higher V T in patients without ARDS or ALI, at the initiation of mechanical ventilation, increases morbidity and mortality.8-10 As suggested by the “biotrauma hypothesis,” ventilation with higher V T and peak pressures may lead to recruitment of neutrophils and local production and release of inflammatory mediators. 11 We conducted a meta-analysis to determine whether conventional (higher) or protective (lower) tidal volumes would be associated with lung injury, mortality, pulmonary infection, and atelectasis in patients without lung injury at the onset of mechanical ventilation. METHODS Studies were identified by 2 authors through a computerized blinded search of MEDLINE (1966-2012), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science, and Cochrane Central Register of Controlled Trials (CENTRAL) using a sensitive search strategy combining the following Medical Subject Headings and keywords (protective ventilation [text word] OR lower tidal volumes [text word]). All reviewed articles and crossreferenced studies from retrieved articles were screened for pertinent information. Selection of Studies Articles were selected for inclusion in the systematic review if they evaluated 2 types of ventilation in patients without ARDS or ALI at the onset of mechanical ventilation. In 1 group of the study, ventilation was protective (lower VT). Then, this protective ven1652 JAMA, October 24/31, 2012—Vol 308, No. 16 tilation group was compared with another group using conventional methods (higher V T ). A study was deemed eligible if it evaluated patients who did not meet the consensus criteria for ARDS or ALI at baseline.12 We included randomized trials as well as observational studies (cohort, before/ after, and cross-sectional), with no restrictions on language or scenario (intensive care unit or operating room). We excluded revisions and studies that did not report the outcomes of interest. When we found duplicate reports of the same study in preliminary abstracts and articles, we analyzed data from the most complete data set. When necessary, we contacted the authors for additional unpublished data. Data Extraction Data were independently extracted from each report by 3 authors using a data recording form developed for this purpose. After extraction, data were reviewed and compared by the first author. Instances of disagreement between the 2 other extractors were solved by a consensus among the investigators. Whenever needed, we obtained additional information about a specific study by directly questioning the principal investigator. Validity Assessment In randomized trials, we assessed allocation concealment, the baseline similarity of groups (with regard to age, severity of illness, and severity of lung injury), and the early stopping of treatment. We used the GRADE approach to summarize the quality of evidence for each outcome.13 In this approach, randomized trials begin as highquality evidence but can be rated down for apparent risk of bias, imprecision, inconsistency, indirectness, or suspicion of a publication bias. Definition of End Points The primary end point was the development of lung injury in each group of the study. Secondary end points included overall survival, incidence of pulmonary infection and atelectasis, intensive care unit (ICU) and hospital length of stay, time to extubation, change in PaCO2, arterial pH values, and change in the ratio of PaO2 to fraction of inspired oxygen (FIO2). Statistical Analysis We extracted data regarding the study design, patient characteristics, type of ventilation, mean change in arterial blood gases, lung injury development, ICU and hospital length of stay, time to extubation, overall survival, and incidence of atelectasis. For the analysis of lung injury development, mortality, pulmonary infection, and atelectasis, we used the most protracted follow-up in each trial up to hospital discharge. We calculated a pooled estimate of risk ratio (RR) in the individual studies using a fixed-effects model according to Mantel and Haenszel and graphically represented these results using forest plot graphs. We explored the following variables as potential modifiers: incorporation of “open lung” techniques (using the authors’ definitions) into experimental strategies, between-group gradients in tidal volumes and plateau pressures, and case mix effects. We reasoned that each of these might influence the effect of protective ventilation on outcome. To explore whether these variables modified the outcome, we compared pooled effects among studies with and without them. For continuous variables, we used the standardized mean difference (SMD), which is the difference in means divided by a standard deviation. The homogeneity assumption was measured by the I2, which describes the percentage of total variation across studies that is due to heterogeneity rather than chance. I2 was calculated from basic results obtained from a typical metaanalysis as I2 =100%⫻(Q−df)/Q, where Q is the Cochran heterogeneity statistic. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity. When heterogeneity was found ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES (I2 ⬎ 25%) we presented the randomeffects model results as primary analysis. A sensitivity analysis was carried out by recalculating pooled RR estimates for different subgroups of studies based on relevant clinical features. This analysis demonstrates whether the overall results have been affected by a change in the meta-analysis selection criteria. Also, a sensitivity analysis about the treatment effect according to quality components of the studies (concealed treatment allocation, blinding of patients and caregivers, blinded outcome assessment) was conducted. A potential publication bias was assessed graphically with funnel plots, as well as by a Begg and Mazumdar rank correlation and an Egger regression. Interrater reliability was determined by comparing the number of studies included by one author with those of another author in each stage of the search using coefficients. Parametric variables were presented as the mean and standard deviation, and nonparametric variables were presented as the median and interquartile range (IQR). All analyses were conducted with Review Manager version 5.1.1 (The Cochrane Collaboration) and SPSS version 16.0.1 (IBM SPSS). For all analyses, 2-sided P values less than .05 were considered significant. RESULTS Our initial search yielded 2122 studies (458 from MEDLINE, 141 from CENTRAL, 885 from CINAHL, and 638 from Web of Science). After removing 711 duplicate studies, we evaluated the abstracts of 1411 studies. After evaluating the abstract of each study, we excluded 1364 studies because they did not meet inclusion criteria. Subsequently, we carefully read the full text of each of the remaining 47 studies and excluded 27 for the following reasons: no data on outcome of interest in 20 studies and same cohort previously analyzed in 7. Twenty references (2822 participants) were included in the final analysis (FIGURE 1 and TABLE 1). For the comparisons of interrater reliability in each stage of the search, the coefficient was 0.91 in the citation stage (P = .004), 0.86 during the abstract review (P =.03), and 0.90 in the full-text stage (P=.006). Figure 1. Literature Search Strategy 2122 Articles identified 458 From MEDLINE 141 From CENTRAL 885 From CINAHL 638 From Web of Science 711 Excluded (duplicate studies) Study Characteristics Table 1 summarizes the studies’ characteristics. All but 5 studies16,22,23,26,29 were randomized controlled trials, and median follow-up time was 21.0 hours (IQR, 6.28-54.60 hours). The median time of per-protocol mechanical ventilation was 6.90 hours for protective and 6.56 hours for conservative strategy. The development of lung injury was the primary outcome in 4 studies. Eight studies evaluated the levels of inflammatory mediators in bronchoalveolar lavage or blood. Tidal volume was set to 6 mL/kg of ideal body weight (IBW) in the protective group of 13 studies; only in 1 study was the tidal volume in the protective ventilation group above 8 mL/kg IBW. Four studies did not report what weight was used to calculate the tidal volume,14,15,21,25 1 study used the measured weight,19 and 15 studies used the predicted weight.9,16-18,20,22-24,26-32 Of these, 7 used the ARDSnet formula to calculated the predicted body weight.16,18,20,24,28-30 The tidal volume gradient between protective and conventional ventilation ranged from 2 to 6 mL/kg IBW, with a mean (SD) of 4.15 (1.42) mL/kg IBW. The tidal volume gradient was less than 4 mL/kg IBW in 30.0% of the studies, between 4 and 5 mL/kg IBW in 40% of the studies, and above 5 mL/kg IBW in 30% of the studies. In 15 studies, the reason for intubation was scheduled surgery,9,15,17-22,24,25,29-32 and in 5, the reason was mixed (medical or surgery).14,16,23,24,28 Lung injury was diagnosed according to the American-European Consensus Conference definition in 6 of the 8 trials that assessed this outcome.16,23,26,27,31,32 The diagnosis of infection was made by clinical assessment plus laboratory, radiological, and microbiologi- ©2012 American Medical Association. All rights reserved. 1411 Potentially relevant articles screened based on abstracts 1364 Excluded 576 ARDS/ALI at onset of mechanical ventilation 487 Reviews 227 Experimental studies 33 Secondary analysis 21 Older version of a study 20 Other 47 Full-text articles assessed for eligibility 27 Excluded 20 No data on outcome of interest 7 Same cohort previously analyzed 20 Articles included in meta-analysis (2822 study participants) ALI indicates acute lung injury; ARDS, acute respiratory distress syndrome; CENTRAL, Cochrane Central Register of Controlled Trials; CINAHL, Cumulative Index to Nursing and Allied Health Literature. cal evaluation in 2 studies14,26 ; was made by decrease in PaO2/FIO2 plus radiological assessment in 1 study31; and was not specified in the last study.20 eTable 1 (available at http://www.jama .com) summarizes study methods, highlighting features related to the risk of bias. Randomization was concealed in 11 of 15 randomized controlled trials included, and follow-up was excellent with minimal loss. Limitations included a lack of blinding (all trials), a lack of intention-to treat analysis (12 trials), and early stopping for benefit (1 trial). Age, weight, minute-volume (product of respiratory rate and tidal volume), and PaO2/ FIO2 were all similar between the 2 groups analyzed (TABLE 2 and eTable 2). As expected, VT and plateau pressure were lower and positive end-expiratory pressure (PEEP) and respiratory rate were higher in the protective group. PaCO2 was higher in the protective group JAMA, October 24/31, 2012—Vol 308, No. 16 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 1653 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES but remained within normal limits (35-45 mm Hg). Acidosis (pH ⬍7.35) was found in the protective group in 3 studies, and the pH level in the protective group was similar to that of the conventional group. The mechanical ventilation settings for each study are provided in eTable 3. sult of the overall test for heterogeneity was not statistically significant, and the I2 was 0% (no sign of heterogeneity) (FIGURE 2). When stratified by the tidal volume gradient between the 2 groups, the RR for lung injury decreased from 0.35 (95% CI, 0.230.51) in the group with less than 4 Primary Outcome Forty-seven of 1113 patients (4.22%) assigned to protective ventilation and 138 of 1090 patients (12.66%) assigned to conventional ventilation developed lung injury during follow-up (RR, 0.33; 95% CI, 0.23-0.47; number needed to treat [NNT], 11). The re- Table 1. Characteristics of the Included Studies and Summary of Continuous Variables Protective Source a Lee et al,14 1999 Chaney et al,15 2000 Gajic et al,16 2004 Koner et al,17 2004 Wrigge et al,9 2004 Wrigge et al,18 2005 Zupancich et al,19 2005 Michelet et al,20 2006 Cai et al,21 2007 Wolthuis et al,22 2007 Yilmaz et al,23 2007 Determann et al,24 2008 Lin et al,25 2008 No. of Patients 103 VT, mL/kg 6 No. 47 Duration of MV, Mean (SD), h Conservative VT, mL/kg 12 No. 56 Setting SICU Protective 2.30 (0.5) Conservative 3.90 (0.8) Dis ST ⫹ 1 ST ⫹ 1 NS NS 12 9.90 (1.0) 10.0 (1.4) 3 NS NS 16.1 (10.2) 12.9 (4.4) Primary Outcome Duration of MV Jadad Score 3 Pulmonary mechanics LI 2 Cytokines in blood Cytokines in BAL Cytokines in BAL Cytokines in BAL Cytokines in blood CT atelectasis Sedative use 1 25 6 12 12 13 166 9 66 12 100 44 6 15 10 29 CABG 62 6 30 12 32 Surgical 44 6 22 12 22 CABG 40 8 20 10 20 CS 6 NS NS 52 5 26 9 26 OS 18 7.06 (1.81) 7.76 (1.85) 16 36 6 8 8 23 10 10 8 13 Neurosurgery ICU 7.15 6.90 (2.2) NS 7.4 (3.1) NS 375 8 163 11 212 NS NS LI 40 6 21 12 19 Surgical 5 ST ST 40 5 20 9 20 OS 24 4.33 (0.9) 4.23 (0.71) 1091 6 558 9 533 OS 2.93 (1.2) 2.76 (1.0) Cytokines in BAL Cytokines in blood LI 150 6 76 10 74 ICU 672 NS NS 20 6 10 12 10 SICU 672 168.0 72.0 229 8 154 10 75 Surgical NS NS Sundar et al,30 2011 Yang et al,31 2011 149 6 75 10 74 CS 672 7.50 10.71 100 6 50 10 50 OS 168 2.00 (0.68) Weingarten et al,32 2012 Total, Mean (SD) 40 6 20 10 20 Surgical Dis 21.0 (6.28-54.6) b Licker et al,26 2009 Determann et al,27 2010 de Oliveira et al,28 2010 FernandezBustamante et al,29 2011 2822 6.45 (1.09) 1416 10.60 (1.14) 1406 CABG Follow-up, h 168 ICU Dis ICU Cytokines in BAL Cytokines in BAL 3 1 1 3 2 3 1 3 3 Duration of MV; ICULS; mortality Duration of MV 3 2.11 (0.8) LI 3 5.13 (1.86) 5.73 (1.71) Oxygenation 3 6.90 (2.93-9.90) b 6.56 (3.61-10.17) b 2.33 (0.89) Abbreviations: BAL, bronchoalveolar lavage; CABG, coronary artery bypass graft surgery; CS, cardiac surgery; CT, computed tomography; Dis, until patient’s discharge; ICU, intensive care unit; ICULS, ICU length of stay; LI, lung injury; MV, mechanical ventilation; NS, not specified; OS, oncology surgery; SICU, surgical intensive care unit; ST, surgery time; VT, tidal volume. a Most of the studies were randomized controlled trials. The exceptions are as follows: Gajic et al,16 Yilmaz et al,23 and Licker et al,26 were cohort studies; Wolthuis et al22 had a beforeand-after design; and Bustamante et al29 had a cross-sectional design. b Median (interquartile range). 1654 JAMA, October 24/31, 2012—Vol 308, No. 16 ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES mL/kg IBW to 0.26 (95% CI, 0.100.66) in the group with 4 to 5 mL/kg IBW (eFigure 1). The RR for the development of lung injury with conventional ventilation, analyzing only randomized controlled trials, was 0.26 (95% CI, 0.10-0.66; NNT, 10). Secondary Outcomes Overall mortality was lower in patients receiving protective ventilation (RR, 0.64; 95% CI, 0.46 to 0.89; NNT, 23). The incidence of pulmonary infection (using the authors’ definition) and atelectasis were lower in the group receiving ventilation with a lower VT (RR [random-effect], 0.45; 95% CI, 0.22 to 0.92; NNT, 26; and RR, 0.62; 95% CI, 0.41 to 0.95, respectively) (Figure 2). The I2 test indicated moderate heterogeneity only in the analysis of pulmonary infection (32%). Protective ventilation was associated with a shorter mean (SD) hospital stay (6.91 [2.36] vs 8.87 [2.93] days, respectively; SMD, 0.51; 95% CI, 0.20 to 0.82), and showed no difference in ICU stay (3.63 [2.43] vs 4.64 [3.29] days, respectively; SMD, 0.37; 95% CI, −0.53 to 1.27) and time of mechanical ventilation (51.07 [58.08] vs 47.12 [45.00] hours, respectively; SMD, 0.48; 95% CI, −0.27 to 1.23). Mean (SD) levels of Pa CO 2 were higher in the protective ventilation group (41.05 [3.79] vs 37.90 [4.19] mm Hg, respectively; SMD, −0.51; 95% CI, −0.70 to −0.32), and mean (SD) pH levels were lower (7.37 [0.03] vs 7.40 [0.03], respectively; SMD, 1.16; 95% CI, 0.31 to 2.02). The mean (SD) PaO2/ F IO 2 ratio was similar between the groups (304.40 [65.70] vs 312.97 [68.13], respectively; SMD, 0.11; 95% CI, −0.06 to 0.27). All these analyses yield significant heterogeneity and were analyzed by random-effects model (I2 for hospital stay, ICU stay, time of mechanical ventilation, PaCO2, pH, and PaO2/FIO2 of 75%, 95%, 92%, 54%, 96%, and 60%, respectively) (eFigures 2, 3, 4, 5, 6, and 7 and eTable 4). In eTable 5, the GRADE evidence profile is provided. This profile evaluates the effect of protective ventilation Table 2. Demographic, Ventilation, and Laboratory Characteristics of the Patients at the Final Follow-up Visit Mean (SD) Protective Ventilation (n = 1416) Age, y 59.97 (7.92) Conventional Ventilation (n = 1406) 60.22 (7.36) Weight, kg Tidal volume, mL/kg IBW a PEEP, cm H2O a Plateau pressure, cm H2O a 72.71 (12.34) 6.45 (1.09) 6.40 (2.39) 72.13 (12.16) 10.60 (1.14) 3.41 (2.79) 16.63 (2.58) 18.02 (4.14) 21.35 (3.61) 13.20 (4.43) .006 .01 8.46 (2.90) 304.41 (65.74) 41.05 (3.79) 7.37 (0.03) 9.13 (2.70) 312.97 (68.13) 37.90 (4.19) 7.40 (0.03) .72 .51 .003 .11 Respiratory rate, breaths/min a Minute-volume, L/min a,b PaO2/FIO2 a PaCO2, mm Hg a pH a P Value .93 ⬍.001 .01 .93 Abbreviations: FIO2, fraction of inspired oxygen; IBW, ideal body weight; PEEP, positive end-expiratory pressure. a At the final follow-up visit. b Minute-volume is the product of respiratory rate and tidal volume. in patients without ARDS or ALI, only from a systematic review and a metaanalysis of randomized controlled trials. The findings for lung injury, mortality, and pulmonary infection were considered moderate, high, and low quality, respectively, by the GRADE profile. Sensitivity analyses according to quality components of each study are shown in eTable 6. In addition, we excluded each trial one at a time and assessed the results. In lung injury and pulmonary infection analyses, the results were always significant despite the exclusion of any trial. After we excluded the trial by Yilmaz et al,23 the analysis of mortality was no longer significant. Sensitivity Analysis To explore these results, we performed a stratified analysis across a number of key study characteristics and clinical factors, and this analysis is shown in TABLE 3. Protection from lung injury, in the protective group, was more pronounced in studies that were not randomized controlled trials performed in the ICU. These trials did not incorporate recruitment maneuvers, had a higher plateau pressure gradient, and a smaller tidal volume gradient. In the survival analysis, we found significant changes in studies without recruitment maneuvers, in studies that ©2012 American Medical Association. All rights reserved. were not randomized trials, and in studies performed in the ICU with a lower tidal volume gradient. For pulmonary infections, we found no statistically significant association in studies that were not randomized trials, a tidal volume gradient less than 4 mL/kg IBW, and the use of recruitment maneuvers. A tidal volume gradient from 4 to 5 mL/kg IBW and a randomized controlled trial performed in surgical patients were each associated with a significant reduction in pulmonary infections in the protective group. Publication Bias Funnel-plot graphical analysis (eFigure 8), Begg and Mazumdar rank correlation, and Egger regression did not suggest a significant publication bias for the analyses conducted in Figure 2 (Kendall =0.17, P=.63; Egger regression intercept=0.24, P =.68). COMMENT We found evidence that a ventilation strategy using lower tidal volumes is associated with a lower risk for developing ARDS. Furthermore, the strategy was associated with lower mortality, fewer pulmonary infections, and less atelectasis when compared with higher tidal volume ventilation in patients without lung injury at the onset of meJAMA, October 24/31, 2012—Vol 308, No. 16 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 1655 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES ciated with the use of lower tidal volumes. Notably, differences in the levels of PEEP and plateau pressure did not influence the final results of the metaanalysis. Previously, Esteban et al33 showed plateau pressures above 35 cm H2O to be associated with an increased risk of death in ICU patients. Although not definitive, this study at least suggested that chanical ventilation. These benefits were associated with a shorter hospital length of stay. Protective ventilation was associated with higher PaCO2 levels and lower pH values, but no difference in the incidence of acidosis was found. In all studies, although the primary goal of the investigators was to compare 2 different tidal volumes, other ventilator strategy elements were asso- higher VT has the ability to exaggerate lung injury and maybe even cause death in patients who require mechanical ventilation for days. Fernández-Pérez et al34 showed higher VT to be associated with postoperative respiratory failure in patients receiving ventilation for only a few hours in the operating room. In light of this information, over the past decade, V T has progressively de- Figure 2. Effect of Ventilation With Smaller Tidal Volume in Patients With Healthy Lungs at the End of the Follow-up Period for Each Study High VT, No. Events Lung injury Gajic et al,16 2004 Michelet et al,20 2006 Yilmaz et al,23 2007 Licker et al,26 2009 Determann et al,27 2010 Yang et al,31 2011 Fernandez-Bustamante et al,29 2011 Weingarten et al,32 2012 Subtotal (95% CI) Total events 32 6 60 20 10 4 5 1 Total 100 26 212 533 74 50 75 20 Low VT, No. Events 12 3 17 5 2 1 7 0 1090 138 Total Weight, % RR (95% CI) 66 26 163 558 76 50 154 20 18.1 4.6 40.7 17.7 8.6 3.4 5.6 1.3 0.47 (0.22-1.00) 0.43 (0.10-1.97) 0.29 (0.16-0.53) 0.23 (0.09-0.62) 0.17 (0.04-0.82) 0.23 (0.03-2.18) 0.67 (0.20-2.17) 0.32 (0.01-8.26) 1113 100.0 0.33 (0.23-0.47) Favors Low VT Favors High VT 47 Heterogeneity: χ27 = 3.74; P = .81, I 2 = 0% Test for overall effect: z = 6.06; P<.001 0.01 0.1 1.0 10 100 10 100 10 100 10 100 RR (95% CI) Mortality Michelet et al,20 2006 Wolthuis et al,22 2007 Yilmaz et al,23 2007 Licker et al,26 2009 Determann et al,27 2010 Fernandez-Bustamante et al,29 2011 Sundar et al,30 2011 Yang et al,31 2011 Weingarten et al,32 2012 Subtotal (95% CI) Total events 1 2 69 15 23 1 2 1 1 26 13 212 533 74 75 74 50 20 2 3 27 13 24 3 1 0 1 1077 115 26 23 163 558 76 154 75 50 20 1.0 2.5 55.7 16.7 17.7 1.5 2.2 1.7 1.1 2.08 (0.18-24.51) 0.82 (0.12-5.71) 0.41 (0.25-0.68) 0.82 (0.39-1.75) 1.02 (0.51-2.04) 1.47 (0.15-14.38) 0.49 (0.04-5.48) 0.33 (0.01-8.21) 1.00 (0.06-17.18) 1145 100.0 0.64 (0.46-0.86) 74 Heterogeneity: χ28 = 6.94; P = .54, I 2 = 0% Test for overall effect: z = 2.68; P = .007 Pulmonary infection Lee et al,14 1999 Michelet et al,20 2006 Licker et al,26 2009 Yang et al,31 2011 Subtotal (95% CI) Total events 0.01 Subtotal (95% CI) Total events 1.0 RR (95% CI) 10 10 30 7 56 26 533 50 2 6 23 1 665 57 47 26 558 50 16.6 14.6 55.8 13.0 0.20 (0.04-0.99) 0.48 (0.14-1.60) 0.72 (0.41-1.26) 0.13 (0.01-1.06) 681 100.0 0.52 (0.33-0.82) 32 Heterogeneity: χ23 = 4.39; P = .22, I 2 = 32% Test for overall effect: z = 2.79; P = .005 Atelectasis Lin et al,25 2008 Cai et al,21 2007 Licker et al,26 2009 Yang et al,31 2011 Weingarten et al,32 2012 0.1 0.01 0.1 1.0 RR (95% CI) 2 5 47 3 5 20 8 533 50 20 3 7 28 1 4 631 62 Heterogeneity: χ24 = 3.76; P = .44, I 2 = 0% Test for overall effect: z = 2.18; P = .03 20 8 558 50 20 3.1 1.1 83.1 5.4 7.3 1.59 (0.24-10.70) 4.20 (0.33-53.12) 0.55 (0.34-0.89) 0.32 (0.03-3.18) 0.75 (0.17-3.33) 656 100.0 0.62 (0.41-0.95) 43 0.01 0.1 1.0 RR (95% CI) A pooled estimate of risk ratio (RR) was calculated in the individual studies using a fixed-effects model according to Mantel and Haenszel. The size of the data markers indicates the weight of the study in the final analyses. VT indicates tidal volume. 1656 JAMA, October 24/31, 2012—Vol 308, No. 16 ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES creased from greater than 12 to 15 mL/kg IBW to less than 9 mL/kg IBW.6,35 The results of the present metaanalysis support this change in ventilation practice. Our results may even suggest that VT should be further reduced. Protective ventilation in patients with ALI or ARDS is already well established; however, physicians do not always adhere to such guidelines. Mikkelsen et al 36 reported that approximately one-third of the patients were receiving protective ventilation at 48 hours, and the main reason for poor adherence was the uncertainty about the diagnosis of ARDS. Another possible reason is that 82% of the patients who never received protective ventilation had a plateau pressure below 30 cm H2O. However, it is well established that reducing the VT in patients with plateau pressures below 30 cm H2O is associated with a survival benefit.10 In this context, the adoption of protective ventilation in patients without lung injury may be even more difficult. It is possible that the beneficial effects of protective ventilation, regarding the development of lung injury, are even greater than what is suggested by the current analysis. Mechanical ventilation can damage the lung, cause inflammation, and release cytokines into the systemic circulation.20,25 This process may cause fever, leukocytosis, and new pulmonary infiltrates, which could be interpreted as ventilator-associated pneumonia instead of ventilatorassociated lung injury. The absence of strict criteria for the diagnosis of pneumonia, such as microbiological identification in blood and bronchoalveolar lavage, in the studies evaluated may lead to an incorrect diagnosis. Ventilatorassociated lung injury may be incorrectly diagnosed as pneumonia in many cases, underestimating the true incidence of lung injury. It is difficult to diagnose pneumonia in the presence of ARDS or ALI, with a quoted sensitivity using conventional clinical criteria of less than 50%.37 Table 3. Summary of Stratified Analyses of Pooled Relative Risks Stratified Analysis Acute Lung Injury Recruitment maneuvers Yes No Tidal volume gradient, mL/kg IBW ⬍4 4-5 Randomized Yes No Setting Operation room ICU Plateau pressure gradient, cm H2O ⬍4 4-8 Diagnosis AECCD Other Mortality Recruitment maneuvers Yes No Tidal volume gradient, mL/kg IBW ⬍4 4-5 Randomized Yes No Setting Operation room ICU Plateau pressure gradient, cm H2O ⬍4 4-8 Pulmonary Infection Recruitment maneuvers Yes No Tidal volume gradient, mL/kg IBW ⬍4 4-5 ⬎5 Randomized Yes No Setting Operation room ICU Plateau pressure gradient, cm H2O ⬍4 4-8 Infection diagnosis Not specified Specified CLCXR PaO2/FIO2 ⫹ x-ray No. of Trials No. of Patients Risk Ratio (95% CI) P Value Heterogeneity, Q 1 7 1091 1112 0.23 (0.09-0.62) 0.35 (0.24-0.52) .004 ⬍.001 0.80 4 4 1861 342 0.35 (0.23-0.51) 0.26 (0.10-0.66) ⬍.001 .004 0.43 .87 4 4 342 1861 0.26 (0.10-0.66) 0.35 (0.23-0.51) .004 ⬍.001 0.87 0.43 5 3 1512 691 0.34 (0.18-0.63) 0.33 (0.21-0.51) ⬍.001 ⬍.001 0.73 0.43 3 1 368 1091 0.38 (0.21-0.71) 0.23 (0.09-0.62) .002 .004 0.52 6 2 1922 281 0.30 (0.21-0.45) 0.56 (0.22-1.41) ⬍.001 .22 0.83 0.66 1 8 1091 1131 0.82 (0.39-1.75) 0.60 (0.42-0.87) .61 .006 0.49 4 5 1731 491 0.54 (0.36-0.79) 0.97 (0.53-1.78) .002 .92 0.35 0.89 5 4 491 1731 0.97 (0.53-1.78) 0.54 (0.36-0.79) .92 .002 0.89 0.35 6 3 1661 561 0.86 (0.46-1.60) 0.57 (0.38-0.84) .63 .005 0.94 0.10 3 1 351 1091 1.02 (0.54-1.92) 0.82 (0.39-1.75) .95 .61 0.71 1 3 1091 255 0.72 (0.41-1.26) 0.27 (0.12-0.64) .25 .003 1 2 1 1091 152 103 0.72 (0.41-1.26) 0.31 (0.11-0.86) 0.20 (0.04-0.99) .25 .02 .05 3 1 255 1091 0.27 (0.12-0.64) 0.72 (0.41-1.26) .003 .25 0.48 3 1 1243 103 0.59 (0.36-0.95) 0.20 (0.04-0.99) .03 .05 0.27 2 1 155 1091 0.33 (0.13-0.85) 0.72 (0.41-1.26) .02 .25 0.40 1 3 2 1 52 1294 1194 100 0.48 (0.14-1.60) 0.53 (0.32-0.87) 0.60 (0.36-1.01) 0.13 (0.01-1.06) .23 .01 .05 .06 .48 0.28 0.11 0.14 Abbreviations: AECCD, American-European Consensus Conference definition; CLCXR, clinical⫹laboratory⫹culture⫹xray; FIO2, fraction of inspired oxygen; IBW, ideal body weight; ICU, intensive care unit. ©2012 American Medical Association. All rights reserved. JAMA, October 24/31, 2012—Vol 308, No. 16 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 1657 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES Our findings are in line with a recently published retrospective study of cardiac surgery patients.38 Although it should be noted that the lower tidal volumes in that study were much higher than those used in the protective groups of the studies analyzed in this metaanalysis, a tidal volume of more than 10 mL/kg was found as a risk factor for organ failure and prolonged ICU stay after cardiac surgery. The results of this meta-analysis should be interpreted within the context of the included studies. Systematic reviews are subject to publication bias, which may exaggerate the study’s conclusion if publication is related to the strength of the results. Additionally, it may be important to distinguish between mechanical ventilation performed in the operating room and that performed in the ICU. Patients in the operating room receive mechanical ventilation for a much shorter time than those in the ICU. Both surgical patients and critically ill patients are at risk for several causes of lung injury. However, these may not be the same for both patient groups, and mechanical ventilation may have different effects on both groups. In addition, although our meta-analysis found decreased mortality rate with protective ventilation, the interpretation of this finding should be considered cautiously because it was discovered only after the addition of the study by Yilmaz et al. 23 Also, one important limitation is that the patients received ventilation for a relatively short time in most studies, which complicates the extrapolation of the results for patients receiving ventilation for long periods in the ICU. For the lung injury analysis, 4 of 8 studies (accounting for 85.4% and 87.2% of the events in the conservative and protective groups, respectively) were not randomized controlled trials, and the randomized controlled trials were of moderate quality. Furthermore, funnel plots are limited as a test for publication bias for a small number of studies. All the dichotomous analyses yielded significant results, and with the excep1658 JAMA, October 24/31, 2012—Vol 308, No. 16 tion of pulmonary infection, all the results showed no heterogeneity (I2 =0%). Pulmonary infection yielded moderate heterogeneity (I2 = 32%), but the analysis with a random-effects model showed similar results. However, all the continuous analyses showed significant heterogeneity (all I2 ⬎60%) and with the use of a random-effects model only differences in pH level, PaCO2 level, and hospital length of stay showed significant results. Therefore, continuous analyses need to be interpreted with caution because of the heterogeneity. In conclusion, our meta-analysis suggests that among patients without lung injury, protective ventilation with use of lower tidal volumes at onset of mechanical ventilation may be associated with better clinical outcomes. We believe that clinical trials are needed to compare higher vs lower tidal volumes in a heterogeneous group of patients receiving mechanical ventilation for longer periods. Author Contributions: Dr Serpa Neto had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Serpa Neto, Cardoso, Manetta, Pereira, Espósito, Schultz. Acquisition of data: Serpa Neto, Pereira, Espósito, Pasqualucci. Analysis and interpretation of data: Serpa Neto, Cardoso, Manetta, Pereira, Espósito, Damasceno, Schultz. Drafting of the manuscript: Serpa Neto, Pereira, Damasceno, Schultz. Critical revision of the manuscript for important intellectual content: Serpa Neto, Cardoso, Manetta, Pereira, Espósito, Pasqualucci, Damasceno, Schultz. Statistical analysis: Serpa Neto, Pereira. Administrative, technical, or material support: Cardoso, Manetta, Espósito, Pasqualucci, Damasceno, Schultz. Study supervision: Damasceno, Schultz. Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Online-Only Material: The eTables and eFigures are available at http://www.jama.com. REFERENCES 1. Plötz FB, Slutsky AS, van Vught AJ, Heijnen CJ. Ventilator-induced lung injury and multiple system organ failure: a critical review of facts and hypotheses. Intensive Care Med. 2004;30(10):1865-1872. 2. The ARDS Definition Task Force; Ranieri VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012;307 (23):2526-2533. 3. Carney D, DiRocco J, Nieman G. Dynamic alveolar mechanics and ventilator-induced lung injury. Crit Care Med. 2005;33(3)(suppl):S122-S128. 4. Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med. 1998;338(6):347-354. 5. 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(18):1301-1308. 6. Brun-Buisson C, Minelli C, Bertolini G, et al; ALIVE Study Group. Epidemiology and outcome of acute lung injury in European intensive care units: results from the ALIVE study. Intensive Care Med. 2004;30 (1):51-61. 7. Schultz MJ, Haitsma JJ, Slutsky AS, Gajic O. What tidal volumes should be used in patients without acute lung injury? Anesthesiology. 2007;106(6):12261231. 8. Gajic O, Frutos-Vivar F, Esteban A, Hubmayr RD, Anzueto A. Ventilator settings as a risk factor for acute respiratory distress syndrome in mechanically ventilated patients. Intensive Care Med. 2005;31(7): 922-926. 9. Wrigge H, Uhlig U, Zinserling J, et al. The effects of different ventilatory settings on pulmonary and systemic inflammatory responses during major surgery. Anesth Analg. 2004;98(3):775-781. 10. Hager DN, Krishnan JA, Hayden DL, Brower RG; ARDS Clinical Trials Network. Tidal volume reduction in patients with acute lung injury when plateau pressures are not high. Am J Respir Crit Care Med. 2005;172(10):1241-1245. 11. de Prost N, Ricard JD, Saumon G, Dreyfuss D. Ventilator-induced lung injury: historical perspectives and clinical implications. Ann Intensive Care. 2011; 1(1):28-42. 12. Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994; 149(3 pt 1):818-824. 13. Guyatt GH, Oxman AD, Kunz R, Vist GE, Falck-Ytter Y, Schünemann HJ; GRADE Working Group. What is “quality of evidence” and why is it important to clinicians? BMJ. 2008;336(7651):995-998. 14. Lee PC, Helsmoortle CM, Cohn SM, Fink MP. Are low tidal volumes safe? Chest. 1990;97(2):430434. 15. Chaney MA, Nikolov MP, Blakeman BP, Bakhos M. Protective ventilation attenuates postoperative pulmonary dysfunction in patients undergoing cardiopulmonary bypass. J Cardiothorac Vasc Anesth. 2000; 14(5):514-518. 16. Gajic O, Dara SI, Mendez JL, et al. Ventilatorassociated lung injury in patients without acute lung injury at the onset of mechanical ventilation. Crit Care Med. 2004;32(9):1817-1824. 17. Koner O, Celebi S, Balci H, Cetin G, Karaoglu K, Cakar N. Effects of protective and conventional mechanical ventilation on pulmonary function and systemic cytokine release after cardiopulmonary bypass. Intensive Care Med. 2004;30(4):620-626. 18. Wrigge H, Uhlig U, Baumgarten G, et al. Mechanical ventilation strategies and inflammatory responses to cardiac surgery: a prospective randomized clinical trial. Intensive Care Med. 2005;31 (10):1379-1387. 19. Zupancich E, Paparella D, Turani F, et al. Mechanical ventilation affects inflammatory mediators in patients undergoing cardiopulmonary bypass for cardiac surgery: a randomized clinical trial. J Thorac Cardiovasc Surg. 2005;130(2):378-383. 20. Michelet P, D’Journo XB, Roch A, et al. Protective ventilation influences systemic inflammation after esophagectomy: a randomized controlled study. Anesthesiology. 2006;105(5):911-919. 21. Cai H, Gong H, Zhang L, Wang Y, Tian Y. Effect of low tidal volume ventilation on atelectasis in patients during general anesthesia: a computed ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES tomographic scan. J Clin Anesth. 2007;19(2):125129. 22. Wolthuis EK, Veelo DP, Choi G, et al. Mechanical ventilation with lower tidal volumes does not influence the prescription of opioids or sedatives. Crit Care. 2007;11(4):R77-R85. 23. Yilmaz M, Keegan MT, Iscimen R, et al. Toward the prevention of acute lung injury: protocol-guided limitation of large tidal volume ventilation and inappropriate transfusion. Crit Care Med. 2007;35 (7):1660-1666. 24. Determann RM, Wolthuis EK, Choi G, et al. Lung epithelial injury markers are not influenced by use of lower tidal volumes during elective surgery in patients without preexistinglunginjury.AmJPhysiolLungCellMolPhysiol. 2008;294(2):L344-L350. 25. Lin WQ, Lu XY, Cao LH, Wen LL, Bai XH, Zhong ZJ. Effects of the lung protective ventilatory strategy on proinflammatory cytokine release during onelung ventilation [in Chinese]. Ai Zheng. 2008;27 (8):870-873. 26. Licker M, Diaper J, Villiger Y, et al. Impact of intraoperative lung-protective interventions in patients undergoing lung cancer surgery. Crit Care. 2009; 13(2):R41-R50. 27. Determann RM, Royakkers A, Wolthuis EK, et al. Ventilation with lower tidal volumes as compared with conventional tidal volumes for patients without acute lung injury: a preventive randomized controlled trial. Crit Care. 2010;14(1):R1-R17. 28. Pinheiro de Oliveira R, Hetzel MP, dos Anjos Silva M, Dallegrave D, Friedman G. Mechanical ventilation with high tidal volume induces inflammation in patients without lung disease. Crit Care. 2010; 14(2):R39-R47. 29. Fernandez-Bustamante A, Wood CL, Tran ZV, Moine P. Intraoperative ventilation: incidence and risk factors for receiving large tidal volumes during general anesthesia. BMC Anesthesiol. 2011;11(1): 22-29. 30. Sundar S, Novack V, Jervis K, et al. Influence of low tidal volume ventilation on time to extubation in cardiac surgical patients. Anesthesiology. 2011; 114(5):1102-1110. 31. Yang M, Ahn HJ, Kim K, et al. Does a protective ventilation strategy reduce the risk of pulmonary complications after lung cancer surgery? a randomized controlled trial. Chest. 2011;139(3):530-537. 32. Weingarten TN, Whalen FX, Warner DO, et al. Comparison of two ventilatory strategies in elderly patients undergoing major abdominal surgery. Br J Anaesth. 2010;104(1):16-22. 33. Esteban A, Anzueto A, Frutos F, et al; Mechanical Ventilation International Study Group. Character- ©2012 American Medical Association. All rights reserved. istics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study. JAMA. 2002;287(3):345-355. 34. Fernández-Pérez ER, Keegan MT, Brown DR, Hubmayr RD, Gajic O. Intraoperative tidal volume as a risk factor for respiratory failure after pneumonectomy. Anesthesiology. 2006;105(1): 14-18. 35. Checkley W, Brower R, Korpak A, Thompson BT; Acute Respiratory Distress Syndrome Network Investigators. Effects of a clinical trial on mechanical ventilation practices in patients with acute lung injury. Am J Respir Crit Care Med. 2008;177(11):12151222. 36. Mikkelsen ME, Dedhiya PM, Kalhan R, Gallop RJ, Lanken PN, Fuchs BD. Potential reasons why physicians underuse lung-protective ventilation: a retrospective cohort study using physician documentation. Respir Care. 2008;53(4):455-461. 37. Baudouin SV. Ventilator induced lung injury and infection in the critically ill. Thorax. 2001;56(suppl 2): ii50-ii57. 38. Lellouche F, Dionne S, Simard S, Bussières J, Dagenais F. High tidal volumes in mechanically ventilated patients increase organ dysfunction after cardiac surgery. Anesthesiology. 2012;116(5):10721082. JAMA, October 24/31, 2012—Vol 308, No. 16 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013 1659 Evolution of Mortality over Time in Patients Receiving Mechanical Ventilation Andrés Esteban1, Fernando Frutos-Vivar1, Alfonso Muriel2, Niall D. Ferguson3, Oscar Peñuelas1, Victor Abraira2, Konstantinos Raymondos4, Fernando Rios5, Nicolas Nin1, Carlos Apezteguı́a5, Damian A. Violi6, Arnaud W. Thille7, Laurent Brochard8, Marco González9, Asisclo J. Villagomez10, Javier Hurtado11, Andrew R. Davies12, Bin Du13, Salvatore M. Maggiore14, Paolo Pelosi15, Luis Soto16, Vinko Tomicic17, Gabriel D’Empaire18, Dimitrios Matamis19, Fekri Abroug20, Rui P. Moreno21, Marco Antonio Soares22, Yaseen Arabi23, Freddy Sandi24, Manuel Jibaja25, Pravin Amin26, Younsuck Koh27, Michael A. Kuiper28, Hans-Henrik Bülow29, Amine Ali Zeggwagh30, and Antonio Anzueto31 1 Hospital Universitario de Getafe & Centro de Investigación Biomédica en red de Enfermedades Respiratorias, Spain; 2Department of Clinical Biostatistics, Hospital Ramón y Cajal, Instituto Ramón y Cajal de Investigación Sanitaria & Centro de Investigación Biomédica en red de Epidemiologı́a y Salud Pública, Spain; 3Interdepartmental Division of Critical Care Medicine, and Departments of Medicine and Physiology, University of Toronto; 4Medizinische Hochschule Hannover, Germany; 5Hospital Nacional Alejandro Posadas, Buenos Aires, Argentina; 6Hospital Interzonal General de Agudos Dr. Luis Güemes, Haedo, Argentina; 7Cenre Hospitalier Universitaire Henri Mondor, Creteil, France; 8Geneva University Hospital, Geneva, Switzerland; 9Clı́nica Medellı́n & Universidad Pontificia Bolivariana, Medellı́n, Colombia; 10Hospital Regional 1o de Octubre, Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado, México DF, México; 11Hospital de Clı́nicas de Montevideo, Uruguay; 12Alfred Hospital & Monash University, Melbourne, Australia; 13Peking Union Medical College Hospital, Beijing, Popular Republic of China; 14 Policlinico “Agostino Gemelli,” Università Cattolica Del Sacro Cuore, Roma, Italy; 15Dipartimento di Scienze Chirurgiche e Diagnostiche Integrate, Universita’ degli Studi, Genoa, Italy; 16Instituto Nacional del Tórax de Santiago, Chile; 17Clı́nica Las Lilas de Santiago, Chile; 18Hospital de Clı́nicas de Caracas, Venezuela; 19Papageorgiou Hospital, Thessaloniki, Greece; 20Hospital Fattouma Bourguina, Monastir, Tunisia; 21Unidade de Cuidados Intensivos Neurocrı́ticos, Hospital de São José, Centro Hospitalar de Lisboa Central, Lisbon, Portugal; 22Hospital Universitário São José, Belo Horizonte, Brazil; 23King Saud Bin Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia; 24Hospital Obrero No.1, La Paz, Bolivia; 25Hospital Eugenio Espejo, Quito, Ecuador; 26Bombay Hospital Institute of Medical Sciences, Mumbai, India; 27Asan Medical Center, University of Ulsan, Seoul, Republic of Korea; 28Medical Center Leeuwarden, Leeuwarden, The Netherlands; 29Holbak Hospital, Region Zealand University of Copenhagen, Denmark; 30 Hôpital Ibn Sina, Rabat, Morocco; and 31South Texas Veterans Health Care System and University of Texas Health Science Center, San Antonio, Texas Rationale: Baseline characteristics and management have changed over time in patients requiring mechanical ventilation; however, the impact of these changes on patient outcomes is unclear. Objectives: To estimate whether mortality in mechanically ventilated patients has changed over time. Methods: Prospective cohort studies conducted in 1998, 2004, and 2010, including patients receiving mechanical ventilation for more than 12 hours in a 1-month period, from 927 units in 40 countries. To examine effects over time on mortality in intensive care units, we performed generalized estimating equation models. (Received in original form December 5, 2012; accepted in final form May 1, 2013) Supported by Centro de Investigación Biomédica en red Enfermedades Respiratorias, CIBER en Epidemiologı́a y Salud Pública. Instituto de Salud Carlos III, Madrid, Spain, Instituto Ramón y Cajal de Investigación Sanitaria, Madrid, Spain, and by a Canadian Institutes of Health Research (Ottawa, ON, Canada) New Investigator Award (N.D.F.). Author Contributions: A.E., as principal investigator, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: A.E., F.F.-V., O.P., N.D.F., A.A. Coordination of data acquisition: N.D.F., K.R., F.R., N.N., C.A., D.A.V., A.W.T., L.B., A.J.V., J.H., M.G., A.R.D., B.D., S.M.M., P.P., L.S., V.T., G.D., D.M., F.A., R.P.M., M.A.S., Y.A., F.S., M.J., P.A., Y.K., M.A.K., H.-H.B., A.A.Z., A.A. Analysis and interpretation of data: F.F.-V., O.P. Drafting of the manuscript: F.F.-V., O.P., N.D.F. Critical revision of manuscript: A.E., K.R., F.R., N.N., C.A., D.A.V., A.W.T., L.B., A.J.V., J.H., M.G., A.R.D., B.D., S.M.M., P.P., L.S., V.T., G.D., D.M., F.A., R.P.M., M.A.S., Y.A., F.S., M.J., P.A., Y.K., M.A.K., H.-H.B., A.A.Z., A.A. Statistical analysis: A.M., V.A. The funding organizations had no role in the design or conduct of the study, collection, management, analysis, or interpretation of the data, or preparation, review, or approval of the manuscript. Correspondence and requests for reprints should be addressed to Andrés Esteban, M.D., Ph.D., Intensive Care Unit, Hospital Universitario de Getafe, Carretera de Toledo, Km 12,500, 28905 Madrid, Spain. E-mail: [email protected] This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org Am J Respir Crit Care Med Vol 188, Iss. 2, pp 220–230, Jul 15, 2013 Copyright ª 2013 by the American Thoracic Society Originally Published in Press as DOI: 10.1164/rccm.201212-2169OC on April 30, 2013 Internet address: www.atsjournals.org AT A GLANCE COMMENTARY Scientific Knowledge on the Subject Baseline characteristics and management have changed over time in patients requiring mechanical ventilation; however, the impact of these changes on patient outcomes is unclear. What This Study Adds to the Field The implementation of mechanical ventilation has significantly changed in the last decade. These changes have likely resulted in a significant decrease in short-term mortality. Measurements and Main Results: We included 18,302 patients. The reasons for initiating mechanical ventilation varied significantly among cohorts. Ventilatory management changed over time (P , 0.001), with increased use of noninvasive positive-pressure ventilation (5% in 1998 to 14% in 2010), a decrease in tidal volume (mean 8.8 ml/kg actual body weight [SD ¼ 2.1] in 1998 to 6.9 ml/kg [SD ¼ 1.9] in 2010), and an increase in applied positive end-expiratory pressure (mean 4.2 cm H2O [SD ¼ 3.8] in 1998 to 7.0 cm of H2O [SD ¼ 3.0] in 2010). Crude mortality in the intensive care unit decreased in 2010 compared with 1998 (28 versus 31%; odds ratio, 0.87; 95% confidence interval, 0.80–0.94), despite a similar complication rate. Hospital mortality decreased similarly. After adjusting for baseline and management variables, this difference remained significant (odds ratio, 0.78; 95% confidence interval, 0.67–0.92). Conclusions: Patient characteristics and ventilation practices have changed over time, and outcomes of mechanically ventilated patients have improved. Clinical trials registered with www.clinicaltrials.gov (NCT01093482). Keywords: mechanical ventilation; mortality; epidemiology; cohort study Mechanical ventilation remains a cornerstone in the supportive management of patients with acute respiratory failure, and is one Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation 221 TABLE 1. COMPARISON OF GENERAL CHARACTERISTICS OF PARTICIPATING UNITS AND BASELINE CHARACTERISTICS OF PATIENTS Participating units, n Medical, % Medical–Surgical, % Surgical, % Neurological, % Other, % Number of beds, median (IQR) Proportion of mechanically ventilated patients, % Patients included, n Geographical area, n (%) Europe USA/Canada Latin America Africa Asia Oceania Age, mean (SD), yr Female sex, n (%) Weight, mean (SD), kg Body mass index, mean (SD), kg/m2 Simplified Acute Physiology Score II on admission, mean (SD), points Previous tracheotomy, n (%) Noninvasive ventilation at home, n (%) Bilevel positive airway pressure, n (%) Continuous positive airway pressure, n (%) Main reason for mechanical ventilation, n (%)† Chronic obstructive pulmonary disease Asthma Other chronic pulmonary disease Neurologic disease Metabolic, n (% of neurologic) Overdose/intoxication, n (% of neurologic) Hemorrhagic stroke, n (% of neurologic) Ischemic stroke, n (% of neurologic) Brain trauma, n (% of neurologic) Other cause, n (% of neurologic) Neuromuscular disease Acute respiratory failure Postoperative Pneumonia Community acquired, n (% of pneumonia) Hospital acquired, n (% of pneumonia) Sepsis Acute respiratory distress syndrome Congestive heart failure Cardiac arrest Trauma Aspiration Other cause of acute respiratory failure Cohort 1998 Cohort 2004 Cohort 2010 361 19 77 4 — — 10 (8, 14) 33% 5,183 349 16 68 13.5 2 0.4 10 (8, 15) 25% 4,968 494 17 73 5 1.5 3 12 (9, 19) 35% 8,151 2,387 1,455 1,222 119 (46) (28) (24) (2) 0 0 59 (17) 1,985 (39) 72 (17) n.d. 44 (17) 102 73 38 35 (2) (1) (52) (48) 2,133 1,341 1,306 110 78 59 1,967 76 27 42 (43) (27) (26) (2) (2) 0 (17) (40) (20) (7) (18) 123 (2) n.d. P Value* ,0.001 ,0.001 ,0.001 (42) (11) (21) (2) (15) (9) (17) (38) (20) (6.5) (18) ,0.001 180 213 81 132 (2) (3) (38) (62%) 0.22 ,0.001 3,407 929 1,692 167 1,242 714 61 3,105 75 27 45 ,0.001 0.23 ,0.001 0.122 ,0.001 522 (10) 79 (2) 60 (1) 864 (17) n.d. n.d. n.d. n.d. n.d. n.d. 94 (2) 267 63 85 938 129 148 310 119 228 4 58 (5) (1) (2) (19) (14) (16) (33) (13) (24) (0.4) (1) 524 98 142 1,574 265 211 470 214 302 112 74 (6) (1) (2) (19) (17) (13) (30) (13.5) (19) (7) (1) ,0.001 0.26 0.019 ,0.001 1,080 (21) 721 (14) n.d. n.d. 458 (9) 231 (5) 539 (10) 100 (2) 407 (8) 129 (3) 367 (7) 1,053 528 376 152 449 148 285 239 284 139 432 (21) (11) (71) (29) (9) (3) (6) (5) (6) (3) (9) 1,750 819 539 280 726 281 617 473 367 200 506 (21) (10) (66) (34) (9) (3) (8) (6) (4.5) (2.5) (6) 0.68 ,0.001 ,0.001 0.93 ,0.001 ,0.001 ,0.001 ,0.001 0.44 ,0.001 Definition of abbreviations: IQR ¼ interquartile range; n.d. ¼ no data available. * Chi-square tests were used for the comparison of categorical variables between groups; Student’s t test and ANOVA were used for the comparison of continuous variables. y Because of rounding, percentages may not total 100. In 1998, more than one cause of acute respiratory failure per patient was permitted. of the defining interventions of intensive care medicine as a specialty. Basic science, translational, and physiological studies have all informed clinical trials, which, in turn, have shown significant impact in areas such as avoidance of ventilator-induced lung injury (1–5), liberation from mechanical ventilation (6–10), and improved outcomes using noninvasive positive-pressure ventilation (11, 12). Little is known, however, about the impact of these interventions on mortality in a real-world setting. In 1998 (13) and 2004 (14), we performed two observational studies on the use of mechanical ventilation and its associated outcomes. In the first, we focused on describing mortality outcomes and identifying factors that impacted survival (13). In the second, we systematically reviewed clinical trials published in the prior 6 years and examined the implementation of this published evidence (14). The main finding of this study was the increasing use of noninvasive positive-pressure ventilation, a reduction in tidal volumes for patients with acute respiratory distress syndrome (ARDS), and more pressure support and less synchronized intermittent mandatory ventilation being used for liberation from mechanical ventilation. Despite these encouraging findings, our data failed to demonstrate a significant improvement in outcomes. Possible reasons for these results include lack of statistical power, or a true lack of outcome improvements due to differences in intensive care unit (ICU) admission patterns over time, an insufficient magnitude of practice change, or a lack of efficacy of these therapies in unselected populations outside the setting of a clinical trial. 222 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 188 2013 TABLE 2. COMPARISON OF VARIABLES RELATED TO MANAGEMENT Noninvasive positive-pressure ventilation before admission in the ICU, n (%) Noninvasive positive-pressure ventilation at admission in the ICU, n (%) Mode of ventilation, days of use per 1,000 d of invasive mechanical ventilation (excluding days during weaning from mechanical ventilation process) Assist-control SIMV SIMV-PS Pressure support PCV APRV/BIPAP PRVC Other mode Ventilator settings over the course of invasive ventilation Tidal volume, mean (SD) ml/kg actual body weight ml/kg predicted body weight Positive end-expiratory pressure, cm H2O, mean (SD) Total respiratory rate, breaths per minute, mean (SD) Sedation, n (%) Analgesia, n (%) Neuromuscular blocking, n (%) Liberation from mechanical ventilation Met criteria, n/eligible patients (%)† Scheduled extubation, n/patients met criteria (%) Unplanned extubation, n/patients at risk (%)‡ Noninvasive positive-pressure ventilation after extubation, n/eligible patients (%)x Reintubation, n/ patients at risk (%)jj After scheduled extubation After unplanned extubation Tracheotomy, n/patients at risk (%)¶ Percutaneous, n (%) Cohort 1998 (n ¼ 5,183) Cohort 2004 (n ¼ 4,968) n.d. n.d. 256 (5) 479 (10) 627 66 132 65 75 n.d. n.d. 35 8.8 n.d. 4.2 18 3,164 2,241 686 412 21 132 125 125 66 98 21 (2.1) (3.8) (11) (61) (43) (13) 3,640/5,008 (73) 2,858/3,640 (78.5) 179/4,906 (4) n.d. 424/3,037 350/2,858 74/179 546/4,906 n.d. (14) (12) (41) (11) 7.6 9.3 5.4 18 3,486 n.d. 524 (2.1) (2.3) (4.6) (6) (70) 3,005/4,682 2,714/3,005 135/4,559 308/2,849 (64) (90) (3) (11) 319/2,849 278/2,714 41/135 664/4,559 362 (10.5) (11) (10) (30) (15) (54.5) Cohort 2010 (n ¼ 8,151) P Value* 429 (5) — 1,169 (14) ,0.001 330 34 94 237 130 91 61 23 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 6.9 8.2 7.0 19 5,755 5,043 890 (1.9) (2.0) (3.0) (6) (71) (62) (11) ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 5,111/7,323 4,151/5,111 654/7,143 473/4,805 (70) (81) (9) (10) ,0.001 ,0.001 ,0.001 0.17 643/4,805 511/4,151 132/654 993/7,143 526 (13) (12) (20) (14) (53) 0.005 0.001 ,0.001 ,0.001 0.53 Definition of abbreviations: APRV/BIPAP ¼ airway pressure release ventilation/biphasic positive airway pressure; ICU ¼ intensive care unit; PCV ¼ pressure-controlled ventilation; PRVC ¼ pressure regulated volume control; PS ¼ pressure support; n.d. ¼ no data available; SIMV ¼ synchronized intermittent mandatory ventilation. * Chi-square tests were used for the comparison of categorical variables between groups; ANOVA was used for the comparison of continuous variables. y Eligible patients for criteria for liberation from mechanical ventilation: overall cohort excepting patients with successful noninvasive positive-pressure ventilation. z Patients at risk for unplanned extubation: overall cohort excepting patients with a previous tracheotomy and patients with successful noninvasive positive-pressure ventilation. x Eligible patients for noninvasive positive-pressure ventilation after extubation: scheduled and unplanned extubated patients. jj Patients at risk for reintubation: scheduled and unplanned extubated patients. ¶ Patients at risk for tracheotomy: overall cohort excepting patients with a previous tracheotomy and patients with successful noninvasive positive-pressure ventilation. At 6 years after the last study, we have now conducted a new international study with the primary objective of estimating whether mortality in mechanically ventilated patients has changed over time. We reasoned that combining this new study with our two prior databases would give us more statistical power to demonstrate potential improvements in survival. Furthermore, allowing more time for incorporation of previous research findings into clinical practice, and performing a multivariable analysis that accounted for potential changes in case mix over time should improve the study importance. Our secondary aims were to describe the changes in the baseline and management characteristics of ventilated patients over this 12-year period. METHODS Design As in previous studies (13, 14), we conducted a prospective use review of patients receiving invasive mechanical ventilation for at least 12 hours or noninvasive positive-pressure ventilation for at least 1 hour during a 1-month period starting in March 2010. National coordinators recruited local investigators (see full list of Investigators in the Third International Study on Mechanical Ventilation (2010) at the end of this article for 2010 and the online supplement for 2004 and 1998) from eligible ICUs (see Table E1 in the online supplement). Only the investigator and research coordinators at each site were aware of the exact purpose and timing of the study to minimize practice changes in response to observation. The research ethics board of each participating institution approved the protocol and waived the need for informed consent. Protocol We collected baseline characteristics, and daily ventilator, gas exchange, clinical management, and complications data while patients were ventilated or until Day 28. Patients were followed in hospital for mortality and length-of-stay outcomes. For this analysis, we combined these 2010 data with those from our studies in 1998 and 2004 (13, 14). Detailed descriptions of the variables collected in each study, along with their definitions, are shown in Tables E2 and E3. One notable variable not collected in 1998 was height; we therefore express tidal volume as ml/kg actual body weight for comparisons across all three studies. Statistical Analysis Data are expressed as mean (6SD), median (interquartile range), and absolute and relative frequencies, as appropriate. ANOVA or KruskalWallis tests were used to compare continuous variables, and Chi-square tests were used for categorical variables. We rejected the null hypothesis of no difference among cohorts at a nominal significance level of Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation 0.05. These analyses were performed using SPSS 17.0 (SPSS Inc., Chicago, IL). To estimate the effect of the year of study on ICU mortality, we used generalized estimating equation models with a binomial distribution and logit link function to measure the impact of time (by cohort year), accounting for the effects related to the clustering of patients receiving care from the same ICU (15). For the purpose of the analysis, the overall population of patients from the three studies (n ¼ 18,302) was randomly divided into two sets: modeling group (n ¼ 13,644) and validation group (n ¼ 4,658). A first generalized estimating equation estimative model was derived entering the following variables: year of study (reference year, 1998), geographical area (Europe, USA– Canada, Latin American, Africa, Asia, and Oceania), age, sex, ICU characteristics (type of unit [medical, surgical, medical–surgical, neurological, other] and size of the unit [fewer than 10 beds, 10–20 beds, more than 20 beds]), severity at ICU admission (Simplified Acute Physiology Score II), and reason for mechanical ventilation. We built on this model by entering daily variables related to management over the course of ventilation: type of ventilation (invasive or noninvasive), inspired fraction of oxygen, positive end-expiratory pressure (PEEP), tidal volume (ml/kg actual body weight), prone position (yes/no), minute ventilation, use of continuous intravenous sedatives (yes/no), and use of neuromuscular blockers (yes/no). These two models were then evaluated for consistency of effects in the validation set. As a sensitivity analysis, we ran these models in the cohort of patients included from ICUs that participated in all three studies. These analyses were performed using Stata Software 11.0 (StataCorp LP, College Station, Texas). RESULTS Characteristics of Participating ICUs and Included Patients The characteristics of the 927 participating ICUs and 18,302 patients across the studies are shown in Table 1. These include 2,913 patients from 55 ICUs that participated in all three studies. Further details regarding the ICU distribution by country across studies are shown in Table E4. Management during Mechanical Ventilation Variables related to management in the ICU are shown in Table 2. The use of noninvasive positive-pressure ventilation increased significantly over the 12-year period: 5% in 1998, 10% in 2004, and 14% in 2010 (P , 0.001). Meanwhile, the median duration of noninvasive positive-pressure ventilation decreased over time (3 d [interquartile range, 2–6] in 1998 versus 2 d [interquartile range, 2–4] in 2004 and 2 d [interquartile range, 1–3] in 2010; P , 0.001). The need for intubation among these patients was not higher in 2010: 32% in 1998, 40% in 2004, and 29% in 2010 (P , 0.001). During invasive ventilation, volume-cycled assist-control ventilation remained the most common ventilator mode, but its use 223 has decreased over time in favor of other modes, particularly pressure support (Table 2; see also Figure E1). Among all patients, we observed a significant decrease in tidal volume and an increase in PEEP (P , 0.001) (Table 2). In the subgroup of patients with ARDS, tidal volume (expressed as ml/kg actual body weight) was also significantly (P , 0.001) reduced over time (Figure 1). The proportion of patients receiving a ventilation strategy with pressure/volume limitation (arbitrarily defined as tidal volume below 6 ml/kg actual body weight; or tidal volume below 8 ml/kg actual body weight and plateau or peak inspiratory pressure less than 30 cm H2O [13]) increased significantly over time: 29% in 1998, 57% in 2004, and 67.5% in 2010 (P , 0.001). In addition, PEEP levels increased significantly (P , 0.001; Figure 2). The use of adjunctive therapies, such as prone positioning, was infrequent, and remained stable: 9% in 1998, 5% in 2004, and 7% in 2010 (P ¼ 0.07). A process of weaning or liberation from mechanical ventilation began in 73% of patients in 1998, 64% in 2004, and 70% in 2010 (Table 2). The proportion of patients who successfully completed their first attempt of liberation from mechanical ventilation increased over time (49% in 1998, 55% in 2004, and 63.5% in 2010; P , 0.001). Among patients who failed their first attempt at weaning, there was a significant increase in the subsequent use of pressure support as the weaning mode, with a concomitant significant reduction in the use of synchronized intermittent mandatory ventilation with or without pressure support (Table E5). The duration of weaning in these patients, although statistically different (P , 0.001), was not clinically different among the three time periods (median, 3 [interquartile range, 2–5] d in 1998, 3 [interquartile range, 2–4] d in 2004, and 3 [interquartile range, 2–5] d in 2010). The rate of reintubation after scheduled extubation remained similar over time, at around 12% (Table 2). This occurred most commonly within the first 12 hours after extubation (63% of reintubations in 1998, 57% in 2004, and 52% in 2010; P ¼ 0.01). More tracheotomies were performed in 2004 (15%) and 2010 (14%) than in 1998 (11%) (P , 0.001). Median time to tracheotomy from intubation, however, remained stable (12 [interquartile range, 7–17] d in 1998, 9 [interquartile range, 6–14] d in 2004, and 11 [interquartile range, 6–16] d in 2010 (P ¼ 0.73). Outcomes A comparison of patient outcomes across the studies is shown in Table 3. Half of the patients included had at least one complication over the course of mechanical ventilation (50% in 1998, Figure 1. Evolution of the set tidal volume (ml/kg actual body weight) in patients with criteria of acute respiratory distress syndrome (ARDS). Left: mean (6SD); P , 0.001 for all comparisons with ANOVA test. Right: proportion of the time of mechanical ventilation in each category of tidal volume; P , 0.001 for all comparisons with chi-square test. 224 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 188 2013 Figure 2. Evolution of the applied positive end-expiratory pressure (PEEP; cm H2O) in patients with criteria of acute respiratory distress syndrome (ARDS). Left: mean (6SD); P , 0.001 for comparisons, with ANOVA, between 1998 versus 2004 and 1998 versus 2010; P ¼ 0.05 for comparison, with ANOVA, between 2004 and 2010. Right: proportion of the time of mechanical ventilation ventilated in each category of PEEP. P , 0.001 for all comparisons with chi-square test. ZEEP ¼ zero positive end-expiratory pressure. 48% in 2004, and 54% in 2010; P , 0.001). We note higher rates of sepsis and cardiovascular failure in 2010, and a reduction over time in ventilator-associated pneumonia. We observed significant differences in the duration of ventilatory support, ICU, and hospital length of stay. Analysis of Mortality in the ICU There was a decline in the crude ICU mortality, from 31% in 1998 to 28% in 2010 (odds ratio, 0.87; 95% confidence interval, 0.80–0.94; P , 0.001). After adjustment for baseline variables, mortality in 2010 remained lower than in 1998 (odds ratio, 0.74; 95% confidence interval, 0.66–0.84; Table 4). This result was similar after management variables were added to the model (odds ratio, 0.78; 95% confidence interval, 0.67–0.92). These results were quantitatively similar in the validation group (final model odds ratio, 0.86; 95% confidence interval, 0.67–1.09; Table 4). The observed mortality in the 2,913 patients from ICUs that participated in all three studies showed a similar trend: 34% in 1998, 34% in 2004, and 27% in 2010 (P ¼ 0.002). Sensitivity analysis of the final model showed robust results in this subgroup, with a final adjusted odds ratio for 2010 ICU mortality of 0.58 (95% confidence interval, 0.45–0.75) relative to 1998. DISCUSSION Our analysis of these international multicenter studies suggests that, in mechanically ventilated patients, short-term mortality has decreased over time. Although many factors may have contributed to these results, several factors related to mechanical ventilation management may have contributed, including reducing the need for invasive mechanical ventilation, and improving both mechanical ventilation safety and the process of liberation from mechanical ventilation. This study confirms our previous work showing that changes in mechanical ventilation practices continue to follow findings of positive clinical trials, and shows, for the first time, a significant association with reduced mortality. We have observed a significant change of mortality over time after adjustment for case mix and for variables related to ventilatory settings. Despite a similar proportion of patients with complications related to mechanical ventilation and organ dysfunctions, the intensive care mortality in 2010 was lower than in previous studies. The most likely explanation is that the overall management of mechanically ventilated patients has improved over the last 12 years. The implementation of different integrated management strategies for precipitating conditions, such as sepsis (16), protocols related to ventilator management, including sedation (10, 17, 18), and incorporating bundles of care to prevent TABLE 3. COMPARISON OF EVENTS EMERGING OVER THE COURSE OF MECHANICAL VENTILATION AND IN THE OUTCOMES Cohort 1998 (n ¼ 5,183) Events emerging over the course of mechanical ventilation Acute respiratory distress syndrome, n (%) Acquired ICU pneumonia, n (%) Sepsis, n (%) Barotrauma, n (%) Cardiovascular failure, n (%) Renal failure, n (%) Hepatic failure, n (%) Hematological failure, n (%) Outcomes Duration of ventilatory support†, median (IQR), days Length of stay in the ICU, median (IQR), days Length of stay in the hospital‡, median (IQR) days Mortality, n (% [95% confidence interval]) In the ICU At Day 28 after admission in the ICU In the hospital‡ 218 438 457 154 1,145 971 326 552 (4) (8.5) (9) (3) (22) (19) (6) (11) 4 (2, 8) 7 (4, 14) 16 (9,29) 1,590 (31 [29–32]) 1,719 (33 [32–34]) 1,876 (40 [38–41]) Cohort 2004 (n ¼ 4,968) 279 265 400 157 1,193 948 691 795 (6) (5) (8) (3) (24) (19) (14) (16) 6 (3, 11) 8 (4, 15) 17 (9,31) 1,533 (31 [29–32]) 1,605 (32 [31–34]) 1,759 (37 [36–38]) Cohort 2010 (n ¼ 8,151) P Value* (6) (4) (18) (2) (39) (22) (7) (8) ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 5 (3, 10) 7 (4, 14) 17 (9,31) ,0.001 ,0.001 0.002 495 359 1,473 140 3,145 1,775 555 662 2,269 (28 [27–29]) 2,445 (30 [29–31]) 2,733 (35 [34–36]) ,0.001 ,0.001 ,0.001 Definition of abbreviations: ICU ¼ intensive care unit; IQR ¼ interquartile range. * Chi-square tests were used for the comparison of categorical variables between groups; ANOVA or Kruskall-Wallis test was used for the comparison of continuous variables. y Including duration of liberation from mechanical ventilation in the extubated patients. z Patients whose date and status at discharge from hospital was unknown were not included in the calculation (465 patients in 1998, 211 patients in 2004, and 377 patients in 2010). Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation TABLE 4. ESTIMATION OF RELATION OF YEAR OF STUDY TO MORTALITY IN THE INTENSIVE CARE UNIT Study Year Crude odds ratio 1998 2004 2010 After adjustment for ICU and baseline variables 1998 2004 2010 After adjustment for ICU and baseline variables and for daily management variables 1998 2004 2010 Modeling Group (n ¼ 13,644) Validation Group (n ¼ 4,658) 1 1.01 (0.90–1.13) 0.83 (0.75–0.93) 1 1.00 (0.82–1.20) 0.90 (0.75–1.07) 1 1.01 (0.89–1.15) 0.74 (0.66–0.84) 1 1.05 (0.90–1.24) 0.78 (0.67–0.92) 1 1.05 (0.90–1.24) 0.78 (0.67–0.92) 1 1.06 (0.83–1.34) 0.86 (0.67–1.09) Definition of abbreviation: ICU ¼ intensive care unit. Data presented are odds ratios (95% confidence intervals). nosocomial infections (19, 20) could all possibly contribute to these findings. Throughout our studies, we observed a significant decrease in delivered tidal volume, both in the general mechanically ventilated population and particularly among those with ARDS. This lung-protective strategy has been evaluated in several randomized clinical trials (1, 2, 21–23), the largest of which showed reduced mortality when tidal volume and plateau pressure were controlled in patients with acute lung injury (2). These data suggest that routine implementation of a pressure–volume limited ventilation strategy in clinical practice could reduce mortality in these patients. In a recent observational ARDS study (24), adherence to lung-protective ventilation was associated with a 3% decrease in the mortality risk at 2 years (hazard ratio, 0.97; 95% confidence interval, 0.95–0.99; P ¼ 0.002). These results, like ours, suggest that these clinical trial results are robust even applied to a less selective population, as in usual clinical practice. There are also suggestions that lung-protective strategies may be important in patients who do not yet have ARDS (25–28). The fact that the odds ratio for mortality moved closer to 1 in all datasets after accounting for management variables, including tidal volume, is in keeping with the hypothesis that lung-protective ventilation is a key intervention across a broad range of mechanically ventilated patients, not just those with ARDS. PEEP is another essential component of the ventilatory strategy in patients with ARDS. From the first report of this syndrome (29), it has been observed that PEEP can improve oxygenation by keeping recruited alveoli open and decreasing intrapulmonary shunt (30). In addition, PEEP-induced lung recruitment can decrease alveolar overdistention, because the volume of each subsequent tidal breath is shared by more open alveoli. PEEP may also decrease repetitive alveolar opening and closing during the respiratory cycle, thereby preventing lung injury (31). Observational studies have reported an independent association between zero PEEP and mortality (32) in a heterogeneous cohort of patients who were mechanically ventilated and between low values of PEEP and mortality in patients with ARDS (33). An individual patient data metaanalysis (3, 4) from randomized trials (1, 34–38) comparing higher versus lower PEEP levels while controlling tidal volume in patients with acute lung injury overall showed no statistically significant difference in hospital mortality (random effects model relative risk, 0.85; 95% confidence interval, 0.71–1.01). 225 However, in patients with moderate to severe ARDS (39), higher levels of PEEP were associated with a significant 4% absolute mortality reduction (40). In our analysis, we observed a marked reduction in the use of zero end-expiratory pressure and a significant increase in the use of higher PEEP, especially in patients with ARDS. Nevertheless, the mean PEEP level used in the first week of ventilatory support in 2010 (9.3 6 3.3 cm H2O) was still lower than that set in the higher PEEP groups of some randomized clinical trials. The use of noninvasive positive-pressure ventilation at ICU admission has almost doubled every 6 years. This finding is similar to that reported by Chandra and colleagues (41), who observed a 46% increase in noninvasive positive-pressure ventilation use and a 42% decline in invasive mechanical ventilation use between 1998 and 2008 in patients with exacerbations of chronic obstructive pulmonary disease. Encouragingly, despite this increased use, the proportion of patients intubated after failing a trial of noninvasive positive-pressure ventilation has remained stable over time. An important observation is that patients who required invasive ventilation after noninvasive positive-pressure ventilation had higher mortality than patients who were directly intubated. This finding was described previously by our group (13, 14, 42) and others (43), and has several possible explanations. First, failure of noninvasive positive-pressure ventilation could identify patients who are difficult to ventilate or who have a higher severity of illness. In our study, patients who failed noninvasive positive-pressure ventilation had significantly higher Simplified Acute Physiology Score II scores (mean ¼ 42) than successful patients (mean ¼ 37), but they had lower scores than patients invasively ventilated from the beginning (mean ¼ 45). Another explanation is that failing noninvasive ventilation is, in itself, harmful, and that delayed recognition of failure may exacerbate this harm (41). Our study has several limitations. This is an analysis of prospectively collected clinical data from a wide variety of ICUs, patient conditions, and clinical practices, some of which are related to outcome (43). We have performed an analysis accounting for clustering by the ICUs, and we have included ICU type and size in the model to try to minimize the possible bias related to those variables. Although we used multivariable modeling, unmeasured confounders may remain, which could have impacted our results. We constructed an estimative model to evaluate the decreasing ICU mortality in the last 12 years, and we have outlined plausible explanations for this. This model has been validated and adjusted, and the results appear consistent. However, it is not possible to make a causal inference model from our observational data. In addition, we have no information about withdrawal and/or withholding of life support and its possible impact on our results. Another limitation is that we did not have all the same ICUs participating in all three studies. To some extent, this was unavoidable due to closure of some previously participating hospitals, changes in resources available for data collection at participating centers, and our goal to expand the generalizability of results by including ICUs from more countries. In conclusion, our analysis shows that the implementation of mechanical ventilation has significantly changed in the last decade; we speculate that these changes have resulted in the significant decrease in short-term mortality. Author disclosures are available with the text of this article at www.atsjournals.org. Investigators in the Third International Study on Mechanical Ventilation (2010): Argentina: Coordinators: Fernando Rı́os (Hospital Nacional Alejandro Posadas), Damian Violi (Hospital Hospital Interzonal General de Agudo Profesor Dr .Luis Guemes, Haedo). Marisol Rodrı́guez-Goñi, Roger Lamoglie and Fernando Villarejo 226 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE (Hospital Nacional Profesor A. Posadas, Buenos Aires); Norberto Tiribelli and Santiago Ilutovich (Sanatorio de la Trinidad, General Mitre); Matı́as Brizuela and Mariana Monllau (Hospital Tránsito Cáceres de Allende, Córdoba); Fernando Saldarini and Silvina Borello (Hospital General de Agudos donación Francisco Santojanni, Buenos Aires); Alberto Marino and Norberto Tiribelli (Hospital Churruca-Visca, Buenos Aires); Mauricio Vinzio and Karina Bonasegla (Sanatorio de la Trinidad, San Isidro); Julián Hernández and Marı́a Belén Yapur (Sanatorio Nuestra Señora del Rosario, Jujuy); Marı́a Eugenia González (Hospital Privado de Comunidad, Mar del Plata); Sebastián E. Mare (Sanatorio Dr. Julio Méndez, Buenos Aires); Judith Sagardı́a and Marco Bezzi (Hospital General de Agudos P. Piñeiro, Buenos Aires); Cecilia Pereyra and Julian Strati (Hospital Interzonal General de Agudo Profesor Dr. Luis Guemes, Haedo); Daniel Vargas and Claudia Diaz (Hospital Pablo Soria, Jujuy); Pablo Gómez and Marcelo Palavecino (Sanatorio Juncal, Temperley); Graciela Elizabeth and Aguilera Garcı́a, M. Eugenia (Hospital de San Luis); Luis Pablo Cardonnet and Lisandro Betttini (Hospital Provincial del Centenario, Rosario); Hernán Nuñez and Lucas Vallejo (Hospital General de Agudos Juan A. Fernández, Buenos Aires); Fernando Fernández and Jorge Arroyo (Hospital Central, Mendoza); Daniel Duarte and Gerardo Filippa (Hospital Regional Rı́o Grande, Tierra del Fuego); Cayetano Galetti and Hernan Nunia (Sanatorio Allende, Córdoba); Fernando Lambert and Elisa Estenssoro (Hospital Interzonal de Agudos San Martin, La Plata); Marina Busico and Fernando Villarejo (Clı́nica Olivos, Vicente López); Javier Horacio Álvarez (Hospital Universitario Austral, Pilar); Alejandro Raimondi and Gustavo Badariotti (Sanatorio Mater Dei, Buenos Aires); Martı́n Lugaro (Sanatorio Las Lomas, San Isidro); Fernando Lipovestky (Clı́nica Santa Isabel; Buenos Aires); Alan Javier Zazu and Hugo Capponcelli (Clı́nica Privada de Especialidades de Villa Marı́a); Patricia Vogl and Cristina Orlandi (Hospital Zonal Francisco López Lima, General Roca); Alejandro Gómez and Gustavo Jannello (Sanatorio de los Arcos; Buenos Aires); Alejandro Risso (Sanatorio Otamendi y Miroli, Buenos Aires); Leticia Rapetti and Guillermo Chiappero (Hospital Universitario, Universidad Abierta Interamericana, Buenos Aires); Juan Domingo Fernández (Hospital Regional de Comodoro Rivadavia, Chubut); Rodrigo E. Gómez-Paz (Hospital Español, Buenos Aires); Marcos Juan Zec and Pascual Valdez (Hospital General de Agudos Dalmacio Vélez Sársfield, Buenos Aires); Jorgelina Guyon and Ariel Chena (Hospital Lagomaggiore, Mendoza); Sergio Lasdica (Hospital Municipal Coronel Suárez, Buenos Aires); Martin Deheza and Schimdt Alejandra (Hospital General de Agudos Bernardino Rivadavia, Buenos Aires); Francisco Criado (Hospital Naval Puerto Belgrano, Bahı́a Blanca); Norma Beatriz Márquez (Policlı́nico Atlántico del Sur, Rı́os Gallegos); Pablo Desmery and José Luis Scapellato (Sanatorio Anchorena, Buenos Aires); Gonzalo Javier Rı́os and Cristian Casabella (Clı́nica Bazterrica, Buenos Aires) Australia: Coordinators: Jasmin Board and Andrew R. Davies (Alfred Hospital, Melbourne). Andrew Bersten, Elisha Matheson, and Amy Waters (Flinders Medical Center, Adelaide); John Santamaria and Jennifer Holmes (St. Vincent’s Hospital, Melbourne); Cartan Costello, Manoj K Saxena, and John Myburgh (St. George Hospital, Sydney); Ellen Kinkel and Forbes McGain (The Western Hospital, Melbourne); Claire Cattigan and Allison Bone (Barwon Health, Geelong Hospital, Geelong); Ian Seppelt, Leonie Weisbrodt and Cheryl Cuzner (Nepean Hospital, Sydney); Christopher MacIsaac, Deborah Barge and Tania Caf (Royal Melbourne Hospital, Melbourne); Cameron Knott and Graeme Duke (The Northern Hospital, Melbourne); Imogen Mitchell, Helen Rodgers, Rachel Whyte, and Elisha Fulton (Canberra Hospital, Canberra); Jasmin Board, Andrew Davies, and Alistair Nichol (Alfred Hospital, Melbourne); Hergen Buscher, Priya Nair, and Claire Reynolds (St. Vincent’s Hospital, Sydney); Simon J. G. Hockley, Ian Moore, and Katherine Davidson (Calvary Wakefield Hospital, Adelaide); David Milliss, Raju Pusapati, and Helen Wong (Concord Hospital, Sydney); Jason Fletcher and Julie Smith (Bendigo Hospital, Bendigo); Paul Goldrick, Dianne Stephens, and Jane Thomas (Royal Darwin Hospital, Darwin); Anders Aneman, Sutrisno Gunawan, and Tom Cowlam (Liverpool Hospital, Liverpool); George Lukas and Rick McAllister (Royal Hobart Hospital, Hobart); Minka Springham, Joanne Sutton, and Jeff Presneill (Mater Health Services, Brisbane); Tony Sutherland and Dianne Hill (Ballarat Health Services, Ballarat); Howard Connor, Jenny Dennett, and Tim Coles (Central Gippsland Hospital, Sale). Bolivia: Coordinator: Freddy Sandy (Hospital Obrero No. 1, La Paz). Sandro Chavarria and Marcelo Choque (Hospital Obrero No. 1, La Paz); Ronald Pairumani and Juan Guerra (Instituto Gastroenterológico boliviano japonés, Santa Cruz). Brazil: Coordinator: Marco Antônio Soares Reis (Hospital Universitário São José, Belo Horizonte). José Carlos Versiani (Hospital Madre Teresa, Belo Horizonte); Eduardo Fonseca Sad (Hospital Luxemburgo, Belo Horizonte); Maria Aparecida Braga (Hospital Dia e Maternidade Unimed-BH, Belo Horizonte); Dinalva Aparecida Gomes (Hospital Vera Cruz, Belo Horizonte); Fernando Antônio Botoni (Hospital de Pronto Socorro Risoleta Tolentino Neves, Belo Horizonte); Maurı́cio Meireles Góes (Hospital da Baleia, Belo Horizonte); Frederico Costa Val Barros (Hospital da Polı́cia Militar, Belo Horizonte); Rogério de Castro Pereira (Hospital Felı́cio Rocho, Belo Horizonte); Hugo Urbano (Hospital Vila da Serra, Belo Horizonte); Valéria de Carvalho Magela (Hospital Santa Rita, Contagem); Aline Camile Yehia (Hospital Júlia Kubitschek, Belo Horizonte); Bruno Bonaccorsi Fernandino (Hospital São Francisco-Setimig, Belo Horizonte); Marco Antônio Ribeiro Leão (Hospital São João de Deus, Divinópolis); Rovı́lson Lara (Hospital Arnaldo Gavazza Filho, Ponte Nova); Rovı́lson Lara (Hospital São João Batista, Viçosa); Rubens Altair Amaral de Pádua (Hospital Vaz Monteiro, Lavras); Janine Dias Alves (Santa Casa de Misericórdia de Ouro Preto, Ouro Preto); Aloı́sio Marques do Nascimento (Hospital Nossa Senhora das Graças, Sete Lagoas); Bruno do Valle Pinheiro (Hospital Universitário da Universidade Federal de Juiz de Fora, Juiz de Fora); Carlos Alberto Studart Gomes (Hospital de Messejana, Fortaleza); Marcelo Alcântara Holanda (Hospital Universitário Walter Cantı́dio, Fortaleza); Frederico Rodrigues Anselmo (Hospital Nossa Senhora Aparecida, Belo Horizonte). VOL 188 2013 Canada: Coordinator: Niall D. Ferguson (Mount Sinai Hospital and University Health Network, Toronto). Neill Adhikari, Damon Scales, Robert Fowler, Cheromi Sittambalam, Mehar-Un-Nisa Raja, and Nicole Marinoff (Sunnybrook Health Science Centre, Toronto); Lauralyn McIntryre, Shawna Reddie, Laura Jones, and Irene Watpool (Ottawa Hospital, Ottawa); Jeffrey Singh and Madison Dennis (Toronto Western Hospital, Toronto); Andrew Steel, Andrea Matte, Marc Lipkus, Ryan Albert, and Emily Stern (Toronto General Hospital, Toronto); Michael Miletin, Antonio Raso, and Robyn Klages (William Osler Health Science Centre, Brampton); Jan Friedrich, Orla Smith and Laura Wilson (St. Michael’s Hospital, Toronto); Deborah Cook and Mark Bailey (St. Joseph’s Hospital, Hamilton); Sangeeta Mehta, Stephen Lapinsky, Hannah Mathers, Cheryl Ethier, Stephanie Lubchansky, and Samer Haj-Bakri (Mount Sinai Hospital, Toronto); Dietrich Henzler and Lisa Julien (Queen Elizabeth II Health Sciences Centre, Halifax). Chile: Coordinator: Luis Soto-Román (Instituto Nacional del Tórax, Santiago). Juan Carlos Maurelia (Hospital de Copiapo, Copiapo); César Antonio Maquilon (Instituto Nacional del Tórax, Santiago); Luis Soto-Germani (Hospital de Coquimbo, Coquimbo). China: Coordinator: Bin Du (Peking Union Medical College Hospital, Beijing). Yan Kang and Bo Wang (West China Hospital, Chengdu); Fachun Zhou and Fang Xu (Chongqing Medical University 1st Hospital, Chongqing); Haibo Qiu and Yi Yang (Southeast University Zhongda Hospital, Nanjing); Qingyuan Zhan and Bing Sun (Beijing Chaoyang Hospital, Beijing); Zhenjie Hu and Bin Yu (Hebei Medical University 4th Hospital, Shijiazhuang); Xi Zhu and Yu Bai (Peking University 3rd Hospital, Beijing); Gang Li and Yi Li (Sino-Japanese Friendship Hospital, Beijing); Geng Zhang and Jianbiao Meng (Zhejiang Tongde Hospital, Hangzhou); Xiaobo Huang and Hong Pu (Sichuan Provincial Hospital, Chengdu); Bin Du and Daxing Yu (Peking Union Medical College Hospital, Beijing); Chuanyun Qian and Wei Zhang (Kunming Medical College 1st Hospital, Kunming); Yongjie Yin and Debiao Song (Jilin University 2nd Hospital, Changchun); Yunxuan Yue and Zhengxuan Lv (Kunming City 3rd People’s Hospital, Kunming); Chengmin Yu and Qunmei Yao (Yunnan Chuxiong People’s Hospital; Chuxiong); Xue Wang (Xi’an Jiaotong University 1st Hospital, Xi’an); Yuan Xu and Wei He (Beijing Tongren Hospital, Beijing); Mian Chen and Zhihua Hu (Hainan Medical College Hospital, Haikou); Dongpo Jiang and Jian Huang (Daping Hospital, Chongqing); Wei Yu (Yantai Yuhuangding Hospital, Yantai); Juanxian Gu (Zhejiang Haining People’s Hospital, Naining); Yangong Chao (Beijing Huaxin Hospital, Beijing); Zhixiang Li (Fengrun District People’s Hospital, Tangshan); Zhicheng Zhang (PLA Navy General Hospital, Beijing); Wanxia Li (Nanchang University 2nd Hospital, Nanchang); Zhenyang He (Hainan Provincial People’s Hospital, Haikou); Jianguo Li and Chang Liu (Wuhan University Zhongnan Hospital, Wuhan); Tiehe Qin and Shouhong Wang (Guangdong General Hospital, Guangzhou); Feng Li (Nantong 1st People’s Hospital, Nantong); Jun Jin and Jianhong Fu (Suzhou University 1st Hospital, Suzhou); Hongyang Xu (Wuxi Municipal People’s Hospital, Wuxi); Hongyuan Lin and Jianying Guo (PLA 304 Hospital, Beijing); Yalin Liu and Jinghua Wang (Beijing Hospital, Beijing); Maoqin Li and Jiaqiong Li (Xuzhou Central Hospital, Xuzhou); Lei Chen (Sun Yet-Sen University 6th Hospital, Guangzhou); Qing Song and Liang Pan (PLA General Hospital, Beijing); Xianyao Wan and Jiuzhi Zhang (Dalian Medical University 1st Hospital, Dalian); Weihai Yao and Yuhong Guo (Beijing TCM Hospital, Beijing); Pang Wing Yan (Prince Margaret Hospital, Hong Kong); Kelly Choy (Queen Elizabeth Hospital, Hong Kong); Kwan Ming Chit (Pamela Youde Nethersole Eastern Hospital, Hong Kong); Patricia Leung (Prince of Wales Hospital, Hong Kong); Chau Chin Man (North District Hospital, Hong Kong). Colombia: Coordinator: Marco González (Clı́nica Medellı́n and Universidad Pontificia Bolivariana, Medellı́n). Ricardo Buitrago (Clı́nica Shaio, Bogotá); Marcela Granados (Clı́nica Fundación Valle Lili, Cali); Guillermo Ortiz (Hospital Santa Clara, Bogotá); Cesar Enciso (Grupo Cimca Hospital San José, Bogotá); Mario Gómez (Grupo Cimca Hospital San José and Clı́nica Fundadores, Bogotá); Bladimir Alejandro Gil (Clı́nica Las Américas, Medellı́n); Juan Pablo Sedano, Luis Fernando Castro (Centro Medico Imbanaco, Cali); Carlos Alberto Acosta (Hospital Federico Lleras Acosta, Ibague); Marco Gonzalez A.(Hospital San Rafael, Itagui); Francisco Molina (Clı́nica Universitaria Bolivariana, Medellı́n); Camilo Pizarro (Fundación Cardiovascular Colombiana, Bucaramanga); Mario Villabon (Grupo Cimca Hospital de Suba, Bogota); Carmelo Dueñas (Nuevo Hospital Bocagrande, Cartagena de Indias); Carlos Andrés Dı́az (Hospital General de Medellı́n, Medellı́n); Nelson Fonseca (Corbic, Medellı́n); Rubén Camargo (Clı́nica General del Norte, Barranquilla); Juan David Uribe (Clı́nica Cardiovascular, Medellı́n). Denmark: Coordinator: Hans-Henrik Bülow (Holbak). Simona Beniczky and Jens Brushoj (Naestved); Mikkel Præst and Henrik Guldager (Nykobing Falster); Birgitte Viebaek and Sine Wichman (Roskilde); Anette Mortensen (Holbak); Susanne Andi Iversen (Slagelse); Bo Broberg and Lone M. Poulsen (Koge). Dominican Republic: Edgard Luna (Hospital Universitario José Marı́a Cabral y Báez, Santiago). Ecuador: Coordinator: Manuel Jibaja (Hospital Eugenio Espejo, Quito). Leonardo Pazmiño, Katty Trelles, and Fabricio Picoita (Hospital Eugenio Espejo, Quito); Gustavo Paredes and Vanesa Ramı́rez (Hospital Enrique Garcés, Quito); Guillermo Falconı́, Cristian Cevallos and Boris Villamagua (Hospital Carlos Andrade Marı́n, Quito); Marco Escobar and Freddy Sánchez (Hospital de la Policı́a, Quito); Miguel Llano and Miguel Lazcano (Hospital General de las Fuerzas Armadas, Quito); Ramiro Puetate and José Miguel Guerrero (Hospital Pablo Arturo Suárez, Quito); Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation Mijail Játiva and Myriam Montalvo (Hospital de los Valles, Quito); Franklin Villegas (Hospital Metropolitano, Quito); Luis González Zambrano, Ronnie Mantilla, Gina Quinde, Andrea Gimenez, and Luis Gonzalez Mosquera (Hospital Luis Vernaza, Guayaquil); Henry Caballero and Marı́a Fernanda Garcı́a (Hospital de SOLCA, Quito); Marcelo Ochoa, Soraya Puertas, and Jackeline Coello (Hospital José Carrasco Arteaga, Cuenca); Mario Acosta (Hospital San Vicente de Paul, Ibarra). Egypt: Medhat Soliman (Cairo University Hospitals, Cairo). France: Coordinator: Arnaud W. Thille (CHU Henri Mondor, Créteil). Achille Kouatchet and Alain Mercat (CHU d’Angers); Laurent Brochard (CHU Henri Mondor, Créteil); François Collet (Centre Hospitalier De Saint-Malo); Guillaume Marcotte (Hôpital Édouard Herriot, Lyon); Pascal Beuret (Centre Hospitalier De Roanne); Jean-Christophe M. Richard, Gaëtan Bedunaeu, Pierre-Gildas Guitard, Dorothée Carpentier, Benoit Veber, and Fabien Soulis (CHU Charles Nicolle, Rouen); Frédéric Bellec (Centre Hospitalier De Montauban); Philippe Berger (Centre Hospitalier de Châlons en Champagne); Salem Ould Zein, Géraldine Dessertaine (CHU De Grenoble); C. Canevet (Hôpital d’Armentières); Fabien Grelon (Centre Hospitalier Du Mans). Germany: Coordinator: Konstantinos Raymondos (Medizinische Hochschule Hannover, Hannover). Rolf Dembinski and Rolf Rossaint (Universitaetsklinikum Aachen, Aachen); Steffen Weber-Carstens (Charité Universitaetsklinikum, Berlin); Christian Putensen (Universitaetsklinikum Bonn, Bonn); Maximillian Ragaller (Universitaetsklinikum Carl Gustav Carus, Dresden); Michael Quintel (Universitaetsklinikum der Georg-August-Universitaet, Goettingen); Winfried Schubert (Carl-Thiem-Klinikum Cottbus gGmbH, Cottbus); Thomas Bein and Heinrich Paulus (Klinikum der Universitaet Regensburg); Walter Brandt (Universitaetsklinikum, Magdeburg); Lutz Pfeiffer and Silke Frenzel (Hufeland Klinikum GmbH, Mühlhausen); Thoralf Kerner and P. Kruska (Allgemeines Krankenhaus Harburg, Hamburg); Leila Eckholt and Joachim Hartung (Vivantes Krankenhaus am Urban, Berlin-Kreuzberg); Harald Fritz and Monika Holler (Staedtisches Krankenhaus Martha-Maria Halle-Doelau gGmbH, Halle-Doelau); Johannes Busch and Andreas Viehöfer (Evangelisches Waldkrankenhaus Bad Godesberg gGmbH, Bonn); Jens Buettner (Evangelisches Krankenhaus Elisabethenstift gGmbH, Darmstadt); Jörn Schlechtweg and Achim Lunkeit and Roland Schneider (Klinikum Bad Salzungen GmbH, Bad Salzungen); Maria Wussow, Nils Marquardt, and Christian Frenkel (Staedtisches Klinikum Lueneburg, Lueneburg); Falk Hildebrandt (Dietrich-Bonhoeffer Klinikum Neubrandenburg, Neubrandenburg); Tumbass Volker and Thomas Lipp (Ermstalklinik Bad Urach, Bad Urach); Cezar Mihailescu and Thomas Moellhoff (Katholische Stiftung Marienhospital Aachen, Aachen); Thomas Steinke (Universitaetsklinikum der Martin-Luther-Universitaet Halle-Wittenberg, Halle-Wittenberg); Oliver Franke and Marcus Ruecker (Lungenklinik Heckeshorn, Berlin); Markus Schappacher and Steffen Appel (Ermstalklinik Staedtisches Krankenhaus Sindelfingen, Sindelfingen); Heinz Kerger (Evangelisches Diakoniekrankenhaus, Freiburg); Andreas Schwartz (Bundeswehrkrankenhaus Ulm, Ulm); Jan Dittmann and Jörg Haberkorn (Georgius-Agricola-Klinikum Zeitz, Zeitz); Wolfgang Baier (St. Nikolaus-Stiftshospital GmbH, Andernach); Walter Seyde (Staedtisches Klinikum Wolfenbuettel, Wolfenbuettel). Greece: Coordinator: Dimitrios Matamis (Papageorgiou Hospital, Thessaloniki). Eleni Antoniadou (Gennimata Hospital, Thessaloniki); Pertsas Evangelos (Agios Pavlos General Hospital, Thessaloniki); Maria Giannakou (Ahepa Hospital, Thessaloniki). Hungary: Zoltan Szentkereszty (Kenezy Hospital, Debrecen); Zsolt Molnar (University of Szeged, Szeged). India: Coordinator: Pravin Amin (Bombay Hospital Institute of Medical Sciences, Mumbai). Farhad N. Kapadia (Hinduja Hospital, Mumbai); Nagarajan Ramakrishnan (Apollo Hospitals, Chennai); Deepak Govil (Artemis Health Institute, Gurgaon Haryana); Anitha Shenoy and Goneppanavar Umesh (Kasturba Medical College, Manipal); Samir Sahu, (Kalinga Hospital, Odisha); Sheila Nainan Myatra (Tata Memorial Hospital, Mumbai); Subhash Kumar Todi (AMRI Hospitals Kolkata, West Bengal); Sanjay Dhanuka (CHL Apollo Hospital, Indore); Mayur Patel (Saifee Hospital, Mumbai); P. Samaddar (Tata Main Hospital, Jamshedpur); Dhruva Chaudhry (PGIMS, Rohtak); Vivek Joshi and Srinivas Samavedam (CARE Hospitals, Surat); Ankur Devendra Bhavsar (Spandan Multispeciality Hospital, Vadodar); Prachee Sathe (Ruby Hall Clinic, Pune); Sujoy Mukherjee (Calcutta Medical Research Institute, Kolkat). Italy: Coordinator: Salvatore Maurizio Maggiore (Policlinico “Agostino Gemelli,” Università Cattolica Del Sacro Cuore, Rome). Francesco Idone and Federica Antonicelli (Policlinico “Agostino Gemelli,” Università Cattolica Del Sacro Cuore, Rome); Paolo Navalesi, Rosanna Vaschetto, and Arianna Boggero (Ospedale Maggiore Della Carità, Università Del Piemonte Orientale “Amedeo Avogadro,” Novara); Rosalba Tufano, Michele Iannuzzi, and Edoardo De Robertis (Ospedale Policlinico “Federico II,” Università Di Napoli, Naples); Romano Tetamo and Andrea Neville Cracchiolo (Ospedale “Arnas Civico, Di Cristina, Benfratelli,” Palermo), Antonio Braschi, Francesco Mojoli, and Ilaria Currò (ICU 1, Fondazione IRCCS Policlinico “S. Matteo,” Università Di Pavia, Pavia); Mirko Belliato, Chiara Verga, and Marta Ferrari (ICU 2, Fondazione IRCCS Policlinico “S. Matteo,” Università Di Pavia, Pavia); Erika Mannelli, Valerio Mangani, and Giorgio Tulli (Ospedale “San Giovanni Di Dio,” Florence); Francesca Frigieri and Armando Pedullà (Ospedale “Santa Maria Annunziata,” Florence); Monica Rocco, Giorgia Citterio, and A. Di Russo (Policlinico “Umberto I,” Università La Sapienza, Rome); Gaetano Perchiazzi and Loredana Pitagora (Ospedale Policlinico, Università Di Bari, Bari); Antonio Pesenti and Michela 227 Bombino (Ospedale “San Gerardo,” Università Di Milano Bicocca, Monza); Davide Chiumello, Federica Tallarini, and Serena Azzari (Fondazione IRCCS “Cà Grande” Ospedale Maggiore Policlinico, Università Di Milano, Milan); Antonina Pigna, Ivano Aprile, and Marco Adversi (Policlinico Universitario “S. Orsola-Malpighi,” Bologna); Antonio Corcione, Marianna Esposito, and Annunziata Mattei (Ospedale “V. Monaldi,” Naples); Vito Marco Ranieri, Rosario Urbino, and Ilaria Maria Mastromauro (Ospedale “San Giovanni Battista–Molinette,” Università Di Torino, Torino); Antonino Giarratano, Maurizio Raineri Santi and Ambrogio Sansone (Policlinico “P. Giaccone,” Università Di Palermo, Palermo). Japan: Toru Katani (Tokyo Women’s Medical University, Tokyo). Korea: Coordinator: Younsuck Koh (Asan Medical Center, University of Ulsan, Seoul). Moo Suk Park (Hospital Severance, Yonsei University Health System, Seoul); Je Hyeong Kim (Hospital Korea University of Ansan, Ansan); Kyung Chan Kim (Hospital Catholic University of Daegu, Daegu); Hye Sook Choi (Hospital Dongguk University of Gyeongju, Gyeongju); Yun Seong Kim (Hospital Pusan National University of Yangsan, Yangsan); Jin Hwa Lee (Hospital Ewha Womans University Mokdong, Seoul); Myung-Goo Lee (Hospital Chuncheon Sacred Heart Hospital, Hallym University Medical Center, Chuncheon); Won-Yeon Lee (Hospital Yonsei University Wonju Christian, Wonju); Jin Young An (Hospital Chungbuk National University, Chenogiu); Gee Young Suh (Samsung Medical Center, Sungkyunkwan University, Seoul); Ki-Suck Jung (Hallym University Medical Center, Anyang). Mexico: Coordinator: Asisclo J Villagómez Ortiz (Hospital Regional 18 de Octubre, ISSSTE, México DF). César Cruz Lozano (Hospital Regional de Pemex, Ciudad Madero); Zalatiel Maycotte Luna (Hospital Ángeles de las Lomas, Mexico DF); José Francisco López Baca (Hospital Regional de Zona No.1 del IMSS, Mexico DF); José Elizalde (Instituto Nacional De Ciencias Médicas y Nutrición Salvador Zubirán, Mexico DF); Guillermo Cueto Robledo (Hospital General de México, Mexico DF); Mario Alonso Treviño Salinas (Hospital Universitario de Nuevo León ’Dr. Eleuterio González,’ Nuevo León); Ricardo Martinez Zubieta (Hospital Español de México, Miguel Hidalgo); Claudia Olvera-Guzman and Marco Montes De Oca (Centro Médico ABC, Mexico DF), Silvio A. Ñamendys-Silva (Instituto Nacional de Cancerologı́a, Mexico DF); José Salvador Martı́nez Cano (Centenario Hospital Miguel Hidalgo, Aguascalientes); Jose Angel Baltazar Torres (Umae Hospital De Especialidades Dr. Antonio Fraga Mouret, Mexico DF); Gustavo Morales Muñoz (Hospital Regional de Alta Especialidad de la Mujer, Villahermosa); Antonio Villa Delgado (Hospital Mérida Yucatán; Mérida); Javier Ladape Martinez (Hospital Juárez de México, Mexico DF). Morocco: Coordinator: Amine Ali Zeggwagh (Hôpital Ibn Sina, Rabat). Tarek Dendane (Hôpital Ibn Sina, Rabat); Abderrahim Azzouzi (Hôpital Ibn Sina, Rabat); Ahmed Sbihi (Hôpital Ibn Sina, Rabat); Wajdi Maazouzi and Mourad Amor (Hôpital des Specialites, Rabat); Charki Haimeur (Hôpital Militaire D’Instruction Mohamed V, Rabat). Netherlands: Coordinator: Michael A. Kuiper (Medical Center Leeuwarden (MCL), Leeuwarden). Matty Koopmans (MCL, Leeuwarden); Uli Strauch, Dennis Bergmans, and Serge Heines (Universitair Medisch Centrum Maastricht, Maastricht); Sylvia den Boer (Spaarneziekenhuis, Hoofddorp); Bas M. Kors (Kennemer Gasthuis, Haarlem); Peter van der Voort (Onze Lieve Vrouwe Gasthuis, Amsterdam); Paul J. Dennesen (Medisch Centrum Haaglanden, Den Haag); Bert Beishuizen, Ingrid van den Hul, Erna Alberts, Harry PPM Gelissen, and Eduard Bootsma (Vrije Universiteit Medisch Centrum, Amsterdam); Auke Reidinga (Tjongerschans Ziekenhuis, Heerenveen). New Zealand: Coordinators: Jasmin Board and Andrew R. Davies (Alfred Hospital, Melbourne). Kim Heus, Diane Mackle, and Paul Young (Wellington Hospital, Wellington); Rachael Parke, Eileen Gilder, and Jodi Brown (CVICU Auckland City Hospital, Auckland); Lynette Newby and Catherine Simmonds (DCCM Auckland City Hospital, Auckland); Jan Mehrtens and Seton Henderson (Christchurch Hospital, Christchurch); Tony Williams, Judi Tai, and Chantal Hogan (Middlemore, Auckland); Mary La Pine, John Durning, and Sheree Gare (Waikato, Hamilton); Troy Browne, Shirley Nelson, and Jennifer Goodson (Tauranga Hospital, Tauranga). Panama: Julio Osorio (Hospital Rafael Hernández, Chiriquı́). Peru: Coordinator: Isabel Coronado Campos (Hospital Guillermo Almenara Irigoyen, Lima). Rollin Roldán Mori (Hospital Edgardo Rebagliatti Martins, Lima); Rosa Luz López Martı́nez (Hospital Guillermo Almenara Irigoyen, Lima). Poland: Adam Mikstacki and Barbara Tamowicz (Karol Marcinkowski University of Medical Sciences, Poznan). Portugal: Coordinator: Rui Moreno (UCINC, Hospital de São José, Centro Hospitalar de Lisboa Central, E.P.E., Lisbon). Eduardo Almeida (Hospital Garcia de Orta, Almada); Joana Silvestre (Centro Hospitalar de Lisboa Central, Lisbon); Heloisa Castro, Irene Aragão, and Susana Alves Ferreira (Centro Hospitalar Do Porto–Hospital Geral De Santo António, Porto); Nelson Barros (Centro Hospitalar Trás-Os-Montes E Alto Douro, Vila-Real); Filomena Faria (Instituto Portugues de Oncologia, Porto); Carlos André Correia Casado (Hospital Da Luz, Lisboa); Fausto Fialho Moura (Hospital de Cascais, Cascais); Paulo Marcal (Hospital De São Sebastião, São Sebastião); Ricardo Matos (Hospital de Santo Antonio dos Capuchos, Centro Hospitalar de Lisboa Central, E.P.E., Lisbon); António Alvarez (Centro Hospitalar de Lisboa Norte, Lisbon). 228 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE Russian Federation: Coordinator: Edward Nicolayenko (Hospital No.1, Moscow). Mikhail Kirov (Hospital No. 1, Arkhangelsk); Andrey Yaroshetskiy (Hospital No.7, Moscow); Andrei Piontek (Hospital No.14, Ekaterinburg); Valery Subbotin (National Institute of Surgery Named After A. V. Vishnevskij, Moscow). Saudi Arabia: Yaseen Arabi, Olivia Dulfo, Charina Marie Olay, Edgardo E. Tabhan (King Saud Bin Abdulaziz University for Health Sciences, Riyadh). Spain: Coordinator: Nicolas Nin (Hospital Universitario de Getafe, Madrid). Alfonso Muñoz and César Aragón (Hospital Carlos Haya, Málaga); Ana Villagrá (Corporación Sanitaria y Universitaria Parc Taulı́, Sabadell); Ainhoa Rosselló and Joan Maria Raurich (Hospital Universitario Son Espasses, Mallorca); Marı́a Garitacelaya, Miguel Ángel González-Gallego, and Francisco Ortuño (Hospital Clı́nico Universitario San Carlos, Madrid); Miguel Fernández-Vivas (Hospital Virgen de la Arrixaca, Murcia); David Freire (Hospital Juan Canalejo, A Coruña); Francisco Guerrero and Francisco Manzano (Hospital Virgen de las Nieves, Granada); Juan Carlos Sotillo (Hospital Universitario Gregorio Marañón, Madrid); Alejandra Bustos (Hospital de Torrevieja, Torrevieja); Alfredo Padrón, Pedro Rosas, Rafael Morales, and Liliana Caipe (Hospital Doctor Negrı́n, Las Palmas de Gran Canaria); Maurizio Bottiroli (Hospital de la Santa Creu i Sant Pau, Barcelona); José Marı́a Nicolás (Hospital Clinic-IDIBAPS, Barcelona); Marta Ugalde (Hospital de Cruces, Barakaldo); Javier Ruiz (Hospital Sagrado Corazón, Barcelona); Lucia Capilla (Hospital Morales Meseguer, Murcia); Guillermo Muñiz (Hospital Central de Asturias, Oviedo); Jesús Sánchez-Ruiz (Hospital General de Jerez de la Frontera, Jerez de la Frontera); Javier Cebrián, Begoña Balerdi, Elena Parreño, and Alvaro van Bommel (Hospital Universitario La Fe, Valencia); César Pérez-Calvo (Fundación Jiménez Dı́az, Madrid); Irene Dot (Hospital del Mar, Barcelona); Javier Blanco (Complejo Hospitalario de Ciudad Real, Ciudad Real); Raquel Manzanedo and José J. Blanco (Hospital Insular de Gran Canaria, Gran Canaria); Daniel Fontaneda, Raúl González, and Javier Dı́az Domı́nguez (Complejo Hospitalario de León, León); Alfonso Moreno (Hospital San Pedro, Logroño); Antonio Reyes and Ian Carrasco (Hospital de la Princesa, Madrid); Itziar Mintegui, Rosa Sebastián, and Javier Garcı́a-Alonso, (Complejo Hospitalario de Donostia, Donostia); Carolina Lorencio and Josep Maria Sirvent (Hospital Universitario Dr.Josep Trueta, Girona); Patricia Jimeno (Hospital General de Segovia, Segovia); Miguel León (Hospital Arnau de Vilanova, Lleida); Pedro Galdos (Hospital Universitario Puerta de Hierro, Majadahonda); Nuria Alonso (Hospital Universitario Nuestra Santa Marı́a del Rosell, Cartagena); Julia López-Dı́az (Hospital Universitario La Paz, Madrid); Marı́a Victoria de la Torre, Jorge Vidal Hernández, and Nicolás Zamboschi (Hospital Universitario Virgen de la Victoria, Málaga); Francisco Lucena (Hospital Universitario de Valme, Seville); Gemma Rialp (Hospital Son Llatzer, Mallorca); Raquel Montoiro (Hospital Clı́nico Lozano Blesa, Zaragoza); Victoria Goñi, Marı́a Ángeles Pena, and Antonio Maestre (Hospital Virgen del Rocı́o, Seville); Marc Fabra, Jacinto Baena and Eva Benveniste (Hospital Germans Trias I Pujol, Badalona); Susana Temprano (Hospital 12 de Octubre, Madrid); Jesús Sánchez (Hospital de Rio Hortega, Valladolid); Carmen Campos (Hospital Universitario Dr.Peset, Valencia); Sara Cabañes (Hospital Santiago Apóstol, Vitoria); Marı́a Elena Pérez-Losada, Javier González-Robledo, and José Claudio Leo (Complejo Hospitalario de Salamanca, Salamanca); Enrique Piacentini (Hospital Mutua de Terrassa, Terrassa); Marı́a del Carmen de la Torre (Hospital de Mataró, Mataró); Laura Álvarez-Montero and Fernando Sánchez (Hospital Xeral Calde, Lugo); Antonio Viñuales (Hospital Lluis Alcanyis, Xàtiva); Bernabé Álvarez (Hospital General de Alicante, Alicante); Javier Castañeda (Hospital Clı́nico de Valladolid, Valladolid); Ángela Alonso (Hospital de Fuenlabrada, Fuenlabrada); Marı́a Isabel Ruiz (Complejo Hospitalario de Jaén, Jaén); Pedro Jesús Domı́nguez (Hospital Juan Ramón Jiménez, Huelva); Marcos Delgado (Complejo Hospitalari de Manresa, Manresa); Eugenio Palazón (Hospital Universitario Reina Sofı́a, Murcia); Antonio Garcı́a-Jiménez (Hospital Arquitecto Marcide, Ferrol); Rosa Álvaro (Hospital de La Plana, Castellón); Clara Laplaza, Eva Regidor and Enrique Maravı́ (Complejo Hospitalario de Navarra, Pamplona); José Marı́a Quiroga (Hospital de Cabueñes, Gijón); Amalia Martı́nez de la Gandara (Hospital Infanta Leonor, Madrid); Cecilia Carbayo (Hospital Torrecárdenas, Almerı́a); Marı́a Luisa Navarrete (Hospital San Juan, Alicante); Manuel Valledor and Raquel Yano (Hospital San Agustı́n, Avilés); José Marı́a Gutiérrez (Hospital General de Albacete, Albacete); Amparo Ferrandiz, Alberto Belenguer, and Lidón Mateu (Hospital General de Castellón, Castellón); Laura Sayagues (Complejo Hospitalario de Santiago de Compostela, Compostela); Marı́a José Tolón (Hospital Royo Vilanova, Zaragoza); Nieves Franco (Hospital de Móstoles, Móstoles); Elena Gallego (Hospital San Pedro de Alcántara, Cáceres); Félix Lacoma (Hospital Quirón, Madrid); Patricia Albert (Hospital del Sureste, Arganda); Vicente Arraez (Hospital Universitario General, Elche); Mar Gobernado (Hospital General de Soria, Soria); Susana Moradillo (Hospital Rı́o Carrión, Palencia); Carolina Gı́menez-Esparza (Hospital de la Vega Baja, Orihuela); Teresa Sánchez de Dios (Complejo Hospitalario Montecelo, Pontevedra); Carlos Marian Crespo (Hospital General de Guadalajara, Guadalajara); Cecilia Hermosa and Federico Gordo (Hospital del Henares, Coslada); Genis Carrasco (Hospital SCIAS, Barcelona); Marı́a Ángeles Alonso (Trauma ICU, Hospital 12 de Octubre, Madrid); Alejandro Algora (Fundación Hospital Universitario de Alcorcón, Madrid); Raúl de Pablo (Hospital Prı́ncipe de Asturias, Alcalá de Henares); Sofı́a Garcı́a (Hospital del Poniente, El Ejido); Ana Carolina Caballero (Hospital de Zamora, Zamora); José Marı́a Montón (Hospital Obispo Polanco, Teruel); Teresa Mut (Hospital Provincial de Castellón, Castellón); Eva Manteiga (Hospital Infanta Cristina, Parla); Alejandro de la Serna (Hospital de Galdakao, Galdakao); Ana Esther Trujillo (Hospital General de La Palma, La Palma); Rafael Blancas (Hospital del Tajo, Aranjuez); Inmaculada Vallverdú (Hospital Universitario San Juan, Reus); José Manuel Serrano (Hospital Universitario Reina Sofı́a, Córdoba); Miquel Ferrer (Hospital Clinic-IDIBAPS, Barcelona); Juan Diego Jiménez (Hospital de Don Benito, Don Benito); Carlos Gallego (Hospital Infanta Elena, Valdemoro); Luis Marina (Complejo Hospitalario de Toledo, Toledo). VOL 188 2013 Taiwan: Chen Chin-Ming and Ai-Chin Cheng (Chi Mei Medical Center, Tainan City). Tunisia: Coordinator: Fekri Abroug (Hospital Fattouma Bourguina, Monastir). Besbes Mohamed (Hospital Abderrahmane Mami, Ariana); Imed Chouchene (Hospital Universitarie Farhat Hached, Sousse); Mounir Bouaziz (CHU Habib Bourguiba, Sfax); Stambouli Neji and Islem Ouanes (Hospital Fattouma Bourguina, Monastir); Ayed Samia (Hospital Taher Sfar City, Mahdia). Turkey: Coordinator: Nahit Cakar (Istanbul Medical Faculty, Istanbul). Ismail Kati (Medical Faculty of Yuzuncu Yil University, Van) Ali Aydım Altunkan (Faculty of Medicine, Mersin University, Mersin); Remzi Iscimen (Uludag University Faculty of Medicine, Bursa); Zafer Dogan (Sutcu İmam University, Kahramanmaras); Bilge Çetin (Erciye Üniversity, Kayseri); Tayfun Adanir (Atatürk Hospital, Izmir); Sabriye Guvenc (Anadolu Medical Center Istanbul, Istanbul); Unase Büyükkoçak (Kirikkale University, Kirikkale). United States: Coordinator: Antonio Anzueto (University Hospital San Antonio and South Texas Veterans Health Care System, San Antonio, TX). Ashley Ellis and Gary Kinasewitz (Oklahoma University Health Science Center and VA medical Center, Oklahoma City, OK); Allan Walkey and Phil Alkana (Boston Medical Center, Boston, MA); Gregory A. Schmidt, Susan Gillen, Kathleen Lilli, Jennifer Twombley, Denice Wells, and Larry Welder; (University of Iowa Hospitals and Clinics, Iowa City, IA); Alejandro Arroliga, Alfredo Vasquez-Sandoval, Vincent John Scott, Craig Cernosek, and Christopher Spradley (Temple Clinic, Scott and White Healthcare, Temple, TX); Dimple Tejwani, Sindhaghatta Venkatram, and Gilda Diaz-Fuentes (Bronx Lebanon Hospital Center, New York, NY); Amber Monson, Anthony Saleh, Madhav Gudi, and George Liziamma (New York Methodist Hospital, New York, NY); Mohamed A. Saad, Crissie De Spirito, Bryan Beatty, Samir Vermani, and Crissie Despirito (University of Louisville School of Medicine Hospital, Louisville, KY); Zaza Cohen, Amee Patrawalla, Samir Abdelhadi, Rupesh Vakil, and Steven Y. Chang (UMDNJ–New Jersey Medical School, Newark, NJ); Brian Sherman, Rosanna Del Giudice, and John Oropell (Mount Sinai Medical Center, New York, NY); Timothy D Girard, Cayce Strength, Joyce Okahashi, Leanne Boehm, and Matthew Kirchner (Vanderbilt University School of Medicine, Nashville, TN); Ashley Ellis and Gary Kinasewitz (Oklahoma City VA Medical Center, Oklahoma City, OK); Erwin J. Oei, Sebastian Circo, Nelson Medina, and Mohammed Al-Jagbeer (Morristown Medical Center, Morristown, NJ); V. J. Cardenas, Jr. and Smyth Smith (University of Texas Medical Branch at Galveston, Galveston, TX); Shelby Sutton, Marcela Canola-Mazo, Tim Houlihan, Yogeet Kaur, and Travis Parry (University Hospital San Antonio and South Texas Veterans Health Care System, San Antonio, TX); Craig A. Piquette and Kerry Canady (Omaha VA Medical Center, Omaha, NE); Rahul Nanchal and Dana K. Soetaert (Medical College of Wisconsin, Milwaukee, WI); Maria del Mar TorresPerez, Carlos Robles-Arias, and William Rodriguez-Cintron (VA Caribbean Health Care System, San Juan, Puerto Rico); Mark Tidswell, Jennifer Germain, LoriAnn Kozikowski, and Erin Braden (Baystate Medical Center, Springfield, MA); Geneva Tatem (Henry Ford Hospital, Detroit, MI). Uruguay: Coordinator: Javier Hurtado (Hospital Español and CUDAM, Montevideo). Alberto Deicas (CASMU No. 2, Montevideo); Daniel Weiss (Hospital Pasteur, Montevideo); Marta Beron (Hospital Maciel, Montevideo); Román Garrido (Hospital Evangélico, Montevideo); Cristina Santos and Mario Cancela (Hospital de Clı́nicas, Montevideo); Raúl Lombardi (Impasa, Montevideo); Pedro Alzugaray (CAAMOC, Carmelo, Sanatorio Americano, Montevideo and Orameco, Colonia); Jorge Gerez (Hospital Policial, Montevideo); Silvia Mareque (Sanatorio CAMS, Mercedes); Graciela Franca (Circulo Católico, Montevideo); Oscar Cluzet (Sanatorio Americano, Montevideo); Edgardo Nuñez (Sanatorio Mautone and Hospital de Maldonado, Maldonado); Julio Pontet (Hospital de Florida, Florida); Sergio Cáceres (Centro Cardiológico Sanatorio Americano, Montevideo); Elia Caragna (CRAMI, Las Piedras); Alberto Soler (COMEPA and Hospital Escuela del Litoral, Paysandú); Frank Torres (Sanatorio Cantegril, Punta del Este); Gustavo Pittini (CAAMEPA, Pando). Venezuela: Coordinator: Gabriel d‘Empaire (Hospital de Clı́nicas de Caracas, Caracas). Stevens Salva and Fernando Pérez (Hospital de Clı́nicas de Caracas, Caracas); Clara Pacheco and Zoraida Parra (Hospital Clı́nico Universitario de Caracas, Caracas); Ingrid Von der Osten (Hospital Miguel Pérez Carreño, Caracas); Luis Williams and José Salinas (Hospital Centro de Especialidades de Anzoategui, Lecherı́a). Vietnam: Do Danh Quynh, Pham Thi Van Anh, Nguyen Huu Hoang, and Nghuyen ba Tuan (Viet Duc Hospital, Hanoi). References 1. Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schettino GP, Lorenzi-Filho G, Kairalla RA, Deheinzelin D, Munoz C, Oliveira R, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347–354. 2. 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;342:1301–1308. Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation 3. Gordo-Vidal F, Gómez-Tello V, Palencia-Herrejón E, Latour-Pérez J, Sánchez-Artola B, Díaz-Alersi R. PEEP alta frente a PEEP convencional en el síndrome de distrés respiratorio agudo: revisión sistemática y metaanálisis. Med Intensiva 2007;31:491–501. 4. Gordo-Vidal F, Gómez-Tello V, Palencia-Herrejón E, Latour-Pérez J. Impacto de dos nuevos estudios en el resultado de un metaanálisis sobre la aplicación de presión positiva al final de la espiración alta a pacientes con síndrome de distrés respiratorio agudo. Med Intensiva 2008;32:316–317. 5. Protti A, Cressoni M, Santini A, Langer T, Mietto C, Febres D, Chierichetti M, Coppola S, Conte G, Gatti S, et al. Lung stress and strain during mechanical ventilation: any safe threshold? Am J Respir Crit Care Med 2011;183:1354–1362. 6. Brochard L, Rauss A, Benito S, Conti G, Mancebo J, Rekik N, Gasparetto A, Lemaire F. Comparison of three methods of gradual withdrawal from ventilatory support during weaning from mechanical ventilation. Am J Respir Crit Care Med 1994;150:896–903. 7. Esteban A, Frutos F, Tobin MJ, Alía I, Solsona JF, Valverdú I, Fernández R, de la Cal MA, Benito S, Tomás R, et al.; Spanish Lung Failure Collaborative Group. A comparison of four methods of weaning patients from mechanical ventilation. N Engl J Med 1995;332:345–350. 8. Esteban A, Alía I, Gordo F, Fernández R, Solsona JF, Vallverdú I, Macías S, Allegue JM, Blanco J, Carriedo D, et al.; The Spanish Lung Failure Collaborative Group. Extubation outcome after spontaneous breathing trials with T-tube or pressure support ventilation. Am J Respir Crit Care Med 1997;156:459–465. 9. Esteban A, Alía I, Tobin MJ, Gil A, Gordo F, Vallverdú I, Blanch L, Bonet A, Vázquez A, de Pablo R, et al.; Spanish Lung Failure Collaborative Group. Effect of spontaneous breathing trial duration on outcome of attempts to discontinue mechanical ventilation. Am J Respir Crit Care Med 1999;159:512–518. 10. Girard TD, Kress JP, Fuchs BD, Thomason JW, Schweickert WD, Pun BT, Taichman DB, Dunn JG, Pohlman AS, Kinniry PA, et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care (Awakening and Breathing Controlled Trial): a randomised controlled trial. Lancet 2008;371:126–134. 11. Ram FS, Picot J, Lightowler J, Wedzicha JA. Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2004;CD004104. 12. Vital FM, Saconato H, Ladeira MT, Sen A, Hawkes CA, Soares B, Burns KE, Atallah AN. Non-invasive positive pressure ventilation (CPAP or bilevel NPPV) for cardiogenic pulmonary edema. Cochrane Database Syst Rev 2008;CD005351. 13. Esteban A, Anzueto A, Frutos F, Alía I, Brochard L, Stewart TE, Benito S, Epstein SK, Apezteguía C, Nightingale P, et al.; Mechanical Ventilation International Study Group. Characteristics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study. JAMA 2002;287:345–355. 14. Esteban A, Ferguson ND, Meade MO, Frutos-Vivar F, Apezteguia C, Brochard L, Raymondos K, Nin N, Hurtado J, Tomicic V, et al.; VENTILA Group. Evolution of mechanical ventilation in response to clinical research. Am J Respir Crit Care Med 2008;177:170–177. 15. Hanley JA, Negassa A, Edwardes MD, Forrester JE. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003;157:364–375. 16. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, Gea-Banacloche J, Keh D, Marshall JC, Parker MM, et al. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Intensive Care Med 2004;30:536–555. 17. Shehabi Y, Bellomo R, Reade MC, Bailey M, Bass F, Howe B, McArthur C, Seppelt IM, Webb S, Weisbrodt L; Sedation Practice in Intensive Care Evaluation (SPICE) Study Investigators; ANZICS Clinical Trials Group. Early intensive care sedation predicts long-term mortality in ventilated critically ill patients. Am J Respir Crit Care Med 2012;186: 724–731. 18. Mehta S, Burry L, Cook D, Fergusson D, Steinberg M, Granton J, Herridge M, Ferguson N, Devlin J, Tanios M, et al.; SLEAP Investigators; Canadian Critical Care Trials Group. Daily sedation interruption in mechanically ventilated critically ill patients cared for with a sedation protocol: a randomized controlled trial. JAMA 2012;308:1985–1992. 229 19. Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu H, Cosgrove S, Sexton B, Hyzy R, Welsh R, Roth G, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006;355:2725–2732. 20. Resar R, Pronovost P, Haraden C, Simmonds T, Rainey T, Nolan T. Using a bundle approach to improve ventilator care processes and reduce ventilator-associated pneumonia. Jt Comm J Qual Patient Saf 2005;31:243–248. 21. Stewart TE, Meade MO, Cook DJ, Granton JT, Hodder RV, Lapinsky SE, Mazer CD, McLean RF, Rogovein TS, Schouten BD, et al.; Pressure- and Volume-Limited Ventilation Strategy Group. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med 1998;338: 355–361. 22. Brochard L, Roudot-Thoraval F, Roupie E, Delclaux C, Chastre J, Fernandez-Mondéjar E, Clémenti E, Mancebo J, Factor P, Matamis D, et al. Tidal volume reduction for prevention of ventilator-induced lung injury in acute respiratory distress syndrome. The Multicenter Trial Group on Tidal Volume Reduction in ARDS. Am J Respir Crit Care Med 1998;158:1831–1838. 23. Brower RG, Shanholtz CB, Fessler HE, Shade DM, White P Jr, Wiener CM, Teeter JG, Dodd-o JM, Almog Y, Piantadosi S. Prospective, randomized, controlled clinical trial comparing traditional versus reduced tidal volume ventilation in acute respiratory distress syndrome patients. Crit Care Med 1999;27:1492–1498. 24. Needham DM, Colantuoni E, Mendez-Tellez PA, Dinglas VD, Sevransky JE, Dennison Himmelfarb CR, Desai SV, Shanholtz C, Brower RG, Pronovost PJ. Lung protective mechanical ventilation and two year survival in patients with acute lung injury: prospective cohort study. BMJ 2012;344:e2124. 25. Gajic O, Frutos-Vivar F, Esteban A, Hubmayr RD, Anzueto A. Ventilator settings as a risk factor for acute respiratory distress syndrome in mechanically ventilated patients. Intensive Care Med 2005;31:922–926. 26. Gajic O, Dara SI, Mendez JL, Adesanya AO, Festic E, Caples SM, Rana R, St Sauver JL, Lymp JF, Afessa B, et al. Ventilator-associated lung injury in patients without acute lung injury at the onset of mechanical ventilation. Crit Care Med 2004;32:1817–1824. 27. Determann RM, Royakkers A, Wolthuis EK, Vlaar AP, Choi G, Paulus F, Hofstra JJ, de Graaff MJ, Korevaar JC, Schultz MJ. Ventilation with lower tidal volumes as compared with conventional tidal volumes for patients without acute lung injury: a preventive randomized controlled trial. Crit Care 2010;14:R1. 28. Serpa Neto A, Cardoso SO, Manetta JA, Pereira VG, Espósito DC, Pasqualucci MdeO, Damasceno MC, Schultz MJ. Association between use of lung-protective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome: a meta-analysis. JAMA 2012;308:1651–1659. 29. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute respiratory distress in adults. Lancet 1967;2:319–323. 30. Gattinoni L, Pelosi P, Crotti S, Valenza F. Effects of positive endexpiratory pressure on regional distribution of tidal volume and recruitment in adult respiratory distress syndrome. Am J Respir Crit Care Med 1995;151:1807–1814. 31. Caironi P, Cressoni M, Chiumello D, Ranieri M, Quintel M, Russo SG, Cornejo R, Bugedo G, Carlesso E, Russo R, et al. Lung opening and closing during ventilation of acute respiratory distress syndrome. Am J Respir Crit Care Med 2010;181:578–586. 32. Metnitz PG, Metnitz B, Moreno RP, Bauer P, Del Sorbo L, Hoermann C, de Carvalho SA, Ranieri VM; SAPS 3 Investigators. Epidemiology of mechanical ventilation: analysis of the SAPS 3 database. Intensive Care Med 2009;35:816–825. 33. Ferguson ND, Frutos-Vivar F, Esteban A, Anzueto A, Alía I, Brower RG, Stewart TE, Apezteguía C, González M, Soto L, et al.; Mechanical Ventilation International Study Group. Airway pressures, tidal volumes, and mortality in patients with acute respiratory distress syndrome. Crit Care Med 2005;33:21–30. 34. Ranieri VM, Suter PM, Tortorella C, De Tullio R, Dayer JM, Brienza A, Bruno F, Slutsky AS. Effect of mechanical ventilation on inflammatory mediators in patients with acute respiratory distress syndrome: a randomized controlled trial. JAMA 1999;282:54–61. 35. Brower RG, Lanken PN, MacIntyre N, Matthay MA, Morris A, Ancukiewicz M, Schoenfeld D, Thompson BT; National Heart, Lung, and Blood 230 36. 37. 38. 39. AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE Institute ARDS Clinical Trials Network. Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med 2004;351:327–336. Villar J, Kacmarek RM, Pérez-Méndez L, Aguirre-Jaime A. A high positive end-expiratory pressure, low tidal volume ventilatory strategy improves outcome in persistent acute respiratory distress syndrome: a randomized, controlled trial. Crit Care Med 2006;34:1311–1318. Meade MO, Cook DJ, Guyatt GH, Slutsky AS, Arabi YM, Cooper DJ, Davies AR, Hand LE, Zhou Q, Thabane L, et al.; Lung Open Ventilation Study Investigators. Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008;299:637–645. Mercat A, Richard JC, Vielle B, Jaber S, Osman D, Diehl JL, Lefrant JY, Prat G, Richecoeur J, Nieszkowska A, et al.; Expiratory Pressure (Express) Study Group. Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008;299:646–655. Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, Camporota L, Slutsky AS; ARDS Definition Task Force. 40. 41. 42. 43. VOL 188 2013 Acute respiratory distress syndrome: the Berlin definition. JAMA 2012;307:2526–2533. Briel M, Meade M, Mercat A, Brower RG, Talmor D, Walter SD, Slutsky AS, Pullenayegum E, Zhou Q, Cook D, et al. Higher vs lower positive end-expiratory pressure in patients with acute lung injury and acute respiratory distress syndrome: systematic review and meta-analysis. JAMA 2010;303:865–873. Chandra D, Stamm JA, Taylor B, Ramos RM, Satterwhite L, Krishnan JA, Mannino D, Sciurba FC, Holguín F. Outcomes of noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease in the United States, 1998–2008. Am J Respir Crit Care Med 2012;185:152–159. Esteban A, Frutos-Vivar F, Ferguson ND, Arabi Y, Apezteguía C, González M, Epstein SK, Hill NS, Nava S, Soares MA, et al. Noninvasive positivepressure ventilation for respiratory failure after extubation. N Engl J Med 2004;350:2452–2460. Wunsch H, Linde-Zwirble WT, Angus DC, Hartman ME, Milbrandt EB, Kahn JM. The epidemiology of mechanical ventilation use in the United States. Crit Care Med 2010;38:1947–1953. new england journal of medicine The established in 1812 september 16, 2010 vol. 363 no. 12 Neuromuscular Blockers in Early Acute Respiratory Distress Syndrome Laurent Papazian, M.D., Ph.D., Jean-Marie Forel, M.D., Arnaud Gacouin, M.D., Christine Penot-Ragon, Pharm.D., Gilles Perrin, M.D., Anderson Loundou, Ph.D., Samir Jaber, M.D., Ph.D., Jean-Michel Arnal, M.D., Didier Perez, M.D., Jean-Marie Seghboyan, M.D., Jean-Michel Constantin, M.D., Ph.D., Pierre Courant, M.D., Jean-Yves Lefrant, M.D., Ph.D., Claude Guérin, M.D., Ph.D., Gwenaël Prat, M.D., Sophie Morange, M.D., and Antoine Roch, M.D., Ph.D., for the ACURASYS Study Investigators* A bs t r ac t Background In patients undergoing mechanical ventilation for the acute respiratory distress syndrome (ARDS), neuromuscular blocking agents may improve oxygenation and decrease ventilator-induced lung injury but may also cause muscle weakness. We evaluated clinical outcomes after 2 days of therapy with neuromuscular blocking agents in patients with early, severe ARDS. Methods In this multicenter, double-blind trial, 340 patients presenting to the intensive care unit (ICU) with an onset of severe ARDS within the previous 48 hours were randomly assigned to receive, for 48 hours, either cisatracurium besylate (178 patients) or placebo (162 patients). Severe ARDS was defined as a ratio of the partial pressure of arterial oxygen (PaO2) to the fraction of inspired oxygen (FiO2) of less than 150, with a positive end-expiratory pressure of 5 cm or more of water and a tidal volume of 6 to 8 ml per kilogram of predicted body weight. The primary outcome was the proportion of patients who died either before hospital discharge or within 90 days after study enrollment (i.e., the 90-day in-hospital mortality rate), adjusted for predefined covariates and baseline differences between groups with the use of a Cox model. Results The hazard ratio for death at 90 days in the cisatracurium group, as compared with the placebo group, was 0.68 (95% confidence interval [CI], 0.48 to 0.98; P = 0.04), after adjustment for both the baseline PaO2:FIO2 and plateau pressure and the Simplified Acute Physiology II score. The crude 90-day mortality was 31.6% (95% CI, 25.2 to 38.8) in the cisatracurium group and 40.7% (95% CI, 33.5 to 48.4) in the placebo group (P = 0.08). Mortality at 28 days was 23.7% (95% CI, 18.1 to 30.5) with cisatracurium and 33.3% (95% CI, 26.5 to 40.9) with placebo (P = 0.05). The rate of ICU-acquired paresis did not differ significantly between the two groups. Conclusions From Assistance Publique–Hôpitaux de Marseille Unité de Recherche sur les Maladies Infectieuses et Tropicales Émergentes (URMITE), Centre National de la Recherche Scientifique–Unité Mixte de Recherche (CNRS-UMR) 6236 (L.P., J.-M.F., A.R.) and Faculté de Médecine (A.L.), Université de la Méditerranée Aix–Marseille II; Hôpital Sainte-Marguerite (C.P.-R.); Assistance Publique–Hôpitaux de Marseille (G. Perrin); Hôpital Ambroise Paré (J.-M.S.); and Centre d’Investigations Cliniques, Assistance Publique–Hôpitaux de Marseille, INSERM 9502 (S.M.) — all in Marseille; Hôpital Pontchaillou, Rennes (A.G.); Hôpital Saint Eloi, Montpellier (S.J.); Hôpital Font-Pré, Toulon (J.-M.A.); Hôpital Jean Minjoz, Besançon (D.P.); Hôpital Hôtel-Dieu, Clermont-Ferrand (J.-M.C.); Centre Hospitalier, Avignon (P.C.); Hôpital Caremeau, Nîmes (J.-Y.L.); Hôpital de la Croix-Rousse, Lyon (C.G.); and Hôpital de Cavale Blanche, Brest (G. Prat) — all in France. Address reprint requests to Dr. Papazian at Service de Réani mation Médicale, Hôpital Nord, Chemin des Bourrely, 13009 Marseille, France, or at [email protected]. *The ARDS et Curarisation Systematique (ACURASYS) study investigators are listed in the Appendix. N Engl J Med 2010;363:1107-16. Copyright © 2010 Massachusetts Medical Society. In patients with severe ARDS, early administration of a neuromuscular blocking agent improved the adjusted 90-day survival and increased the time off the ventilator without increasing muscle weakness. (Funded by Assistance Publique–Hôpitaux de Marseille and the Programme Hospitalier de Recherche Clinique Régional 2004-26 of the French Ministry of Health; ClinicalTrials.gov number, NCT00299650.) n engl j med 363;12 nejm.org september 16, 2010 1107 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l T he acute respiratory distress syndrome (ARDS) is characterized by hypoxemic respiratory failure; it affects both medical and surgical patients.1 Despite rigorous physiological management,2 in most studies, ARDS has been fatal in 40 to 60% of patients.3-7 Neuromuscular blocking agents are used in a large but highly variable proportion of patients with ARDS.8-12 Current guidelines indicate that neuromuscular blocking agents are appropriate for facilitating mechanical ventilation when sedation alone is inadequate, most notably in patients with severe gas-exchange impairments.10 In a fourcenter randomized, controlled trial of gas exchange in 56 patients with ARDS,13 infusion of a neuromuscular blocking agent for a period of 48 hours was associated with improved oxygenation and a trend toward lower mortality in the intensive care unit (ICU) (46%, vs. 71% among patients who did not receive a blocking agent; P = 0.06). However, this study was not designed or powered to evaluate mortality. Thus, the benefits and risks of adjunctive therapy with neuromuscular blocking agents in patients with ARDS who were receiving lung-protective mechanical ventilation14 require further evaluation. We conducted a multicenter, randomized, placebo-controlled, double-blind trial to determine whether a short period of treatment with the neuromuscular blocking agent cisatracurium besylate early in the course of severe ARDS would improve clinical outcomes. Me thods Patients Patients were enrolled from March 2006 through March 2008 at 20 ICUs in France (see the Appendix). Eligibility criteria were the receipt of endotracheal mechanical ventilation for acute hypoxemic respiratory failure and the presence of all of the following conditions for a period of no longer than 48 hours: ratio of the partial pressure of arterial oxygen (PaO2, measured in millimeters of mercury) to the fraction of inspired oxygen (FiO2, which is unitless) of less than 150 with the ventilator set to deliver a positive end-expiratory pressure of 5 cm of water or higher and a tidal volume of 6 to 8 ml per kilogram of predicted body weight, and bilateral pulmonary infiltrates that were consistent with edema. An additional eligibility criterion was the absence of 1108 of m e dic i n e clinical evidence of left atrial hypertension — that is, a pulmonary-capillary wedge pressure, if available, of less than 18 mm Hg. If the pulmonarycapillary wedge pressure was not available, echocardiography was performed if the patient had a history of, or risk factors for, ischemic heart disease or had crackles on auscultation. Exclusion criteria are listed in Figure 1. The trial was monitored by an independent data and safety monitoring board. Randomization and blinding regarding the study-group assignments were performed according to Consolidated Standards for the Reporting of Trials (CONSORT) guidelines, as indicated in the Supplementary Appendix (available with the full text of this article at NEJM.org). The study protocol and statistical analysis plan (also available at NEJM.org) were approved for all centers by the ethics committee of the Marseille University Hospital (Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale), according to French law. The study was conducted in accordance with the protocol and statistical analysis plan. Written informed consent was obtained from the patients or their proxies. Study Treatment Cisatracurium besylate (150-mg formulation, GlaxoSmithKline) and placebo were prepared in identical separate 30-ml vials for intravenous infusion. Peripheral-nerve stimulators were not permitted. The Ramsay sedation scale was used to adapt sedative requirements. The scale assigns the conscious state a score of 1 (anxious, agitated, or restless) to 6 (no response on glabellar tap). Once the assigned Ramsay sedation score was 6 and the ventilator settings were adjusted (Table 1), a 3-ml rapid intravenous infusion of 15 mg of cis atracurium besylate or placebo was administered, followed by a continuous infusion of 37.5 mg per hour for 48 hours. This regimen was based on the results of two studies of a total of 92 patients monitored for paralysis.13,15 Ventilation and Weaning Protocol The volume assist–control mode of ventilation was used, with a tidal volume of 6 to 8 ml per kilogram of predicted body weight (Table 1). The goal was a saturation of peripheral blood oxygen (SpO2) as measured by means of pulse oximetry of 88 to 95% or a PaO2 of 55 to 80 mm Hg. To achieve this goal, FiO2 and the positive end-expiratory pres- n engl j med 363;12 nejm.org september 16, 2010 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. Neuromuscular Blocking Agents in ARDS 1326 Patients were assessed for eligibility 986 Were excluded 340 Underwent randomization 178 Were assigned to receive cisatracurium 162 Were assigned to receive placebo 1 Withdrew consent 177 Received cisatracurium 162 Received placebo 177 Were included in the analysis 162 Were included in the analysis Figure 1. Randomization and Follow-up the Patients, According to Study Group. 1st RETAKE: AUTHOR:ofPapazian For the 986 patients who were assessed for eligibility but excluded, the reasons for 2nd exclusion were as follows: age FIGURE: 1 of 2 of consent (185 patients, 18.8%), continuous 3rd younger than 18 years (19 patients, 1.9%), lack infusion of a neuroRevised muscular blocking agent at enrollment ARTIST:(42 ts patients, 4.3%), known pregnancy (19 patients, 1.9%), enrollment in SIZE another trial within the previous 30 days, (57 patients, 5.8%), increased intracranial 6 col pressure (18 patients, 1.8%), Line Combo 4-C H/T TYPE: 33p9 severe chronic respiratory disease requiring long-term oxygen therapy or mechanical ventilation at home (95 patients, 9.6%), actual body weight exceeding 1 kgAUTHOR, per centimeter of height, (20 patients, 2.0%), severe chronic liver PLEASE NOTE: Figure8.3%), has been redrawn and type has been reset.or chemotherapy-induced neutro disease (Child–Pugh class C) (82 patients, bone marrow transplantation Please check carefully. penia (97 patients, 9.8%), pneumothorax (18 patients, 1.8%), expected duration of mechanical ventilation of less than 48 hours (15 patients, 1.5%), (168 patients, 17.0%), other reason JOB:decision 36312 to withhold life-sustaining treatment ISSUE: 09-16-10 (103 patients, 10.4%), and time window missed (48 patients, 4.9%). sure were adjusted as in the Prospective, Randomized, Multi-Center Trial of 12 ml/kg Tidal Volume Positive Pressure Ventilation for Treatment of Acute Lung Injury and Acute Respiratory Distress Syndrome (ARMA).14 An open-label, rapid, intravenous injection of 20 mg of cisatracurium was allowed in both groups if the end-inspiratory plateau pressure remained greater than 32 cm of water for at least 10 minutes despite the administration of increasing doses of sedatives and decreasing tidal volume and positive end-expiratory pressure (if tolerated). If this rapid, intravenous injection resulted in a decrease of the end-inspiratory plateau pressure by less than 2 cm of water, a second injection of 20 mg of cisatracurium was allowed. If after the injection, the end-inspiratory plateau pressure did not decrease or decreased by less than 2 cm of water, cisatracurium was not administered again during the following 24-hour period. Organ or System Failure Patients were monitored daily for 28 days for signs of failure of nonpulmonary organs or systems.14 Circulatory failure was defined as systolic blood pressure of 90 mm Hg or less or the need for vasopressor therapy. Coagulation failure was defined as a platelet count of 80,000 or less per cubic millimeter. Hepatic failure was defined as a serum bilirubin level of 2 mg per deciliter (34 μmol per liter) or higher. Renal failure was defined as a serum creatinine level of 2 mg per deciliter (177 μmol n engl j med 363;12 nejm.org september 16, 2010 1109 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. n e w e ng l a n d j o u r na l The of m e dic i n e Weaning attempt: starting on day 3, if FiO2 ≤0.6 findings, and relevant therapeutic interventions were also recorded just before starting the studydrug infusion and again at 24, 48, 72, and 96 hours. Physiological variables were also measured daily between 6 a.m. and 10 a.m. until day 90 or until hospital discharge of a patient who could breath spontaneously. Opioid doses were converted to morphine equivalents. The equivalencies were as follows: 0.01 mg of sufentanil = 10 mg of morphine = 0.1 mg of fentanyl = 0.1 mg of remifentanil.16 Barotrauma was defined as newly developed pneumothorax, pneumomediastinum, subcutaneous emphysema, or pneumatocele larger than 2 cm in diameter. Muscle strength was evaluated with the use of the Medical Research Council (MRC) scale, a previously validated scale that assesses three muscle groups in each arm and leg. The score for each muscle group can range from 0 (paralysis) to 5 (normal strength), with the overall score ranging from 0 to 60.17 The definition of ICU-acquired paresis was an MRC score of less than 48.17 Goals during weaning procedure: SpO2 ≥88% and respiratory rate 26–35 cycles per min Study Outcomes Weaning procedure: decrease PEEP over 20–30 min to 5 cm of water Primary Outcome Table 1. Summary of the Ventilation Procedure.* Variable Ventilator mode: volume assist–control Initial tidal volume: 6–8 ml/kg of predicted body weight Plateau pressure: ≤32 cm of water Oxygenation goal: PaO2 of 55–80 mm Hg or SpO2 of 88–95% Permitted combinations of FiO2 and PEEP, respectively (cm of water): 0.3 and 5, 0.4 and 5, 0.4 and 8, 0.5 and 8, 0.5 and 10, 0.6 and 10, 0.7 and 10, 0.7 and 12, 0.7 and 14, 0.8 and 14, 0.9 and 14, 0.9 and 16, 0.9 and 18, 1.0 and 18, 1.0 and 20, 1.0 and 22, and 1.0 and 24 pH goal: 7.20–7.45 Procedure when oxygenation goal not achieved despite adjustments to FiO2 and PEEP: use inhaled nitric oxide, almitrine mesylate, prone positioning, or any combination thereof Procedure when plateau pressure is >32 cm of water for at least 10 min (in the following order, as needed): increase sedation, reduce tidal volume to 4 ml/kg, decrease PEEP by decrements of 2 cm of water, and perform injection of cisatracurium in a bolus of 20 mg (not to be given again if plateau pressure decreased by <2 cm of water because further doses would probably be futile, but permitted if the drug had its intended effect) Procedure to correct hypercapnia when pH is <7.20 (in the following order, as needed): connect Y-piece directly to endotracheal tube, increase respiratory rate to a maximum of 35 cycles per min, and increase tidal volume to a maximum of 8 ml/kg The primary outcome was the proportion of patients who died before hospital discharge and If weaning procedure fails at a pressure-support ventilation level of 20 cm within 90 days after study enrollment (the 90-day of water, switch to volume assist–control mode of ventilation mortality). Patients who were outside the hospital After at least 2 hr of successful pressure-support ventilation at a level of 5 cm (including those in other types of health care faciliof water, disconnect patient from the ventilator ties) and who were able to breathe spontaneously *FiO2 denotes fraction of inspired oxygen, PaO2 partial pressure of arterial oxy- at day 90 were considered to have been discharged gen, PEEP positive end-expiratory pressure, and SpO2 saturation of peripheral home. Because we anticipated that there would blood oxygen, as measured by means of pulse oximetry. be an imbalance in at least one key risk factor at baseline, the primary outcome was derived from per liter) or higher. The number of days without a Cox regression model in which we adjusted for organ or system failure was calculated by subtract- such imbalance. We also report the crude moring the number of days with organ failure from tality at day 90. 28 days or from the number of days until death, if death occurred before day 28. Organs and systems Secondary Outcomes were considered to be free of failure after hospital Secondary outcomes were the day-28 mortality, the discharge. There was no recommendation regard- numbers of days outside the ICU between day 1 and day 28 and between day 1 and day 90, the number ing volume-resuscitation goals. of days without organ or system failure between Data Collection day 1 and day 28, the rate of barotrauma, the rate of During the 24-hour period before randomization, ICU-acquired paresis, the MRC scores on day 28 we recorded data on demographic characteristics, and at the time of ICU discharge, and the numphysiological variables, relevant interventions per- bers of ventilator-free days (days since successful formed in the ICU, radiographic findings, coexist- weaning from mechanical ventilation) between day ing conditions, and medications. Data on ventila- 1 and day 28 and between day 1 and day 90. It was tor settings, physiological variables, radiographic required that the patient breathe spontaneously, Pressure-support ventilation levels used during weaning procedure: 20, 15, 10, and 5 cm of water 1110 n engl j med 363;12 nejm.org september 16, 2010 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. Neuromuscular Blocking Agents in ARDS without the aid of a ventilator, for a period of at least 48 hours for weaning from the ventilator to be considered successful. The number of ventilatorfree days was considered to be zero for patients who were weaned from mechanical ventilation but who died before day 28 or day 90.18 Statistical Analysis Assumptions for the sample-size calculation were based on our previous studies13,15 that used the same inclusion criteria and on the European epidemiologic study Acute Lung Injury Verification (ALIVE).4 Assuming a 50% mortality at 90 days in the placebo group, we calculated that 340 patients would need to be enrolled to detect a 15% absolute reduction in the 90-day mortality in the cisatra curium group as compared with the placebo group, with 80% statistical power and a two-sided alpha value of 0.05. No interim analysis was performed. We assessed the differences between the groups using Student’s t-test, the Wilcoxon test, the chisquare test, or Fisher’s exact test, as appropriate. All reported P values are two-sided and have not been adjusted for multiple comparisons. Kaplan– Meier curves were plotted to assess the time from enrollment to death and the time to disconnection from the ventilator for a period of at least 48 hours. The primary analysis consisted of evaluating the effect of cisatracurium on the primary outcome (i.e., 90-day mortality), with adjustment by means of a Cox multivariate proportional-hazards model that included two predefined covariates: the baseline Simplified Acute Physiology Score (SAPS) II and the baseline plateau pressure.19 SAPS II is calculated from 12 physiological measurements during a 24-hour period, information about previous health status, and some information obtained at admission. This score ranges from 0 to 163, with higher scores indicating more severe disease. We planned to include all the variables for which there was an imbalance between the two groups at baseline, but the only imbalanced variable was the PaO2:FiO2 ratio. Therefore, we also conducted an analysis based on the baseline PaO2:FiO2 ratio, in which the two thirds of patients with a ratio below 120 (indicating hypoxemia) were compared with the third with a higher ratio. A total of 12 secondary analyses of prespecified outcomes were performed, and results of 9 of these are reported. Only one post hoc analysis was conducted; the results are reported. R e sult s Baseline Characteristics We enrolled 340 patients, of whom 178 were randomly assigned to cisatracurium and 162 to placebo. We excluded 986 patients (Fig. 1). One patient in the cisatracurium group withdrew consent before treatment was started, and data for this patient were therefore not included in the analysis. The median time from the diagnosis of ARDS to study inclusion was 16 hours (interquartile range, 6 to 29) in the study population and did not differ significantly between the cisatracurium group (median, 18 hours; interquartile range, 6 to 31) and the placebo group (median, 15 hours; interquartile range, 7 to 27; P = 0.45). The median time from initiation of mechanical ventilation to study inclusion did not differ significantly between the cis atracurium group (22 hours; interquartile range, 9 to 41) and the placebo group (21 hours; interquartile range, 10 to 42; P = 0.91). The only significant difference between the two groups at baseline was a lower mean PaO2:FiO2 value in the cisatracurium group (P = 0.03) (Table 2, and Table 1 in the Supplementary Appendix). Outcomes Primary Outcome The Cox regression model yielded a hazard ratio for death at 90 days in the cisatracurium group, as compared with the placebo group, of 0.68 (95% confidence interval [CI], 0.48 to 0.98; P = 0.04), after adjustment for the baseline PaO2:FiO2, SAPS II, and plateau pressure (Fig. 2). The crude 90-day mortality was 31.6% (95% CI, 25.2 to 38.8) in the cisatracurium group and 40.7% (95% CI, 33.5 to 48.4) in the placebo group (P = 0.08). Secondary Prespecified Outcomes The beneficial effect of cisatracurium on the 90day survival rate was confined to the two thirds of patients presenting with a PaO2:FiO2 ratio of less than 120. Among these patients, the 90-day mortality was 30.8% in the cisatracurium group and 44.6% in the control group (P = 0.04) (Fig. 2 in the Supplementary Appendix). The absolute difference in 28-day mortality (mortality in the cisatracurium group minus mortality in the placebo group) was −9.6 percentage points (95% CI, −19.2 to −0.2; P = 0.05) (Table 3). The cisatracurium group had significantly more ventilator-free days than the placebo group during n engl j med 363;12 nejm.org september 16, 2010 1111 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e Table 2. Baseline Characteristics of the Patients, According to Study Group.* Cisatracurium (N = 177) Characteristic† Age — yr Placebo N = 162) P Value 58±16 58±15 0.70 Tidal volume — ml/kg of predicted body weight 6.55±1.12 6.48±0.92 0.52 Minute ventilation — liters/min 10.0±2.5 10.1±2.2 0.83 PEEP applied — cm of water 9.2±3.2 9.2±3.5 0.87 Plateau pressure — cm of water 25.0±5.1 24.4±4.7 0.32 Respiratory-system compliance — ml/cm of water 31.5±11.6 31.9±10.7 0.71 FiO2 0.79±0.19 0.77±0.20 0.33 PaO2:FiO2‡ 106±36 115±41 0.03 pH 7.31±0.10 7.32±0.10 0.11 PaO2 — mm Hg 80±24 85±28 0.09 PaCO2 — mm Hg 47±11 47±11 0.62 33 (18.6) 23 (14.2) 0.31 50±16 47±14 0.15 133 (75.1) 125 (77.2) 0.66 129 (72.9) 113 (69.8) 0.52 Surgical, emergency 27 (15.3) 31 (19.1) 0.34 Surgical, scheduled 21 (11.9) 18 (11.1) 0.83 70 (39.5) 73 (45.1) 0.30 142 (80.2) 123 (75.9) 0.34 Prone position or inhaled nitric oxide or almitrine mesylate — no. (%) SAPS II§ Nonfatal condition according to McCabe–Jackson score — no. (%)¶ Main reason for ICU admission — no. (%) Medical Corticosteroids for septic shock — no. (%) Direct lung injury — no. (%) *Plus–minus values are means ±SD. FiO2 denotes fraction of inspired oxygen, ICU intensive care unit, PaCO2 partial pressure of arterial carbon dioxide, PEEP positive end-expiratory pressure, and SpO2 saturation of peripheral oxygen as measured by means of pulse oximetry. †All variables listed except age, nonfatal condition according to McCabe–Jackson score, and main reason for ICU admission were inclusion criteria. ‡Partial pressure of arterial oxygen (PaO2) was measured in millimeters of mercury. § The Simplified Acute Physiology Score (SAPS) II is calculated from 12 physiological measurements during a 24-hour period, information about previous health status, and some information obtained at admission. Scores can range from 0 to 163, with higher scores indicating more severe disease. ¶Possible McCabe–Jackson scores for medical condition are 1 (nonfatal), 2 (ultimately fatal), and 3 (fatal). the first 28 and 90 days (Table 3, and Fig. 3 in the Supplementary Appendix). The Cox regression model yielded an adjusted hazard ratio for weaning from mechanical ventilation by day 90, in the cisatracurium group as compared with the placebo group, of 1.41 (95% CI, 1.08 to 1.83; P = 0.01). The cisatracurium group had more days free of failure of organs, other than the lungs, during the first 28 days (15.8±9.9 days, vs. 12.2±11.1 days in the placebo group; P = 0.01). There were nearly significant between-group differences in the numbers of days without coagulation abnormalities, hepatic failure, and renal failure (Table 3). No patient required dialysis after hospital discharge during the first 28 days. Significantly more days were spent outside the ICU between day 1 and day 90 in the cisatracurium group. 1112 Pneumothorax occurred in a larger proportion of patients in the placebo group (11.7%, vs. 4.0% in the cisatracurium group; P = 0.01) and tended to develop earlier in the placebo group (Fig. 4 in the Supplementary Appendix). During the 48-hour period of study-drug infusion, pneumothorax occurred in one patient (0.6%) in the cisatracurium group as compared with eight patients (4.9%) in the placebo group (P = 0.03). The plateau pressures and minute ventilations for the nine patients are presented in Table 5 in the Supplementary Appendix. Before the development of pneumothorax, none of these patients had an elevated plateau pressure necessitating changes in the mechanical-ventilation settings, changes in the sedation regimen, or open-label administration of cisatracurium. n engl j med 363;12 nejm.org september 16, 2010 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. Neuromuscular Blocking Agents in ARDS The incidence of ICU-acquired paresis, as evaluated on the basis of the MRC score on day 28 or at the time of ICU discharge, did not differ significantly between the two groups (Table 3). Corticosteroids were used during the ICU stay in 189 patients. There was no significant effect of cisatracurium use on the 90-day mortality in the subgroup of patients given corticosteroids (Fig. 6 in the Supplementary Appendix). Ventilator Settings and Lung Function Ventilator settings and lung-function variables during the first week are given in Table 7 in the Supplementary Appendix. On day 7, the PaO2:FiO2 ratio was higher, and the PaCO2 value lower, in the cisatracurium group than in the placebo group. 0.9 0.8 Probability of Survival Secondary Post Hoc Outcome 1.0 Cisatracurium 0.7 0.6 Placebo 0.5 0.4 0.3 0.2 0.1 0.0 0 10 20 30 40 50 60 70 80 90 Days after Enrollment Figure 2. Probability of Survival through Day 90, According to Study Group. Cointerventions During the ICU stay, there were no significant between-group differences in the incidence of cointerventions. A total of 42% of patients in the cisatracurium group and 48% in the placebo group were treated with the use of prone positioning, inhaled nitric oxide, intravenous almitrine mesylate, or a combination of these (Table 8 in the Supplementary Appendix). The criteria for using these interventions were the same in the two groups. Open-label cisatracurium was given more frequently in the placebo group than in the cisatra curium group during the first 48 hours after enrollment. However, the two groups did not differ significantly with respect to the number of patients given at least one open-label cisatracurium bolus during the entire ICU stay after enrollment (Table 8 in the Supplementary Appendix). The required dose of sedatives or analgesics was similar in the two groups during the first week of the study (Table 9 in the Supplementary Appendix). Safety Bradycardia developed during the cisatracurium infusion in one patient. No other side effects were reported. Discussion Treatment with the neuromuscular blocking agent cisatracurium for 48 hours early in the course of severe ARDS improved the adjusted 90-day survival rate, increased the numbers of ventilatorfree days and days outside the ICU, and decreased the incidence of barotrauma during the first 90 days. It did not significantly improve the overall 90-day mortality. Strengths of this trial include the methods used to minimize bias (blinded randomization assignments, a well-defined study protocol, complete follow-up, and intention-to-treat analyses). The recruitment of a large number of patients from 20 multidisciplinary ICUs where international standards of care are followed suggests that our data can be generalized to other ICUs. Limitations of the trial include the fact that our results were obtained for cisatracurium bes ylate and may not apply to other neuromuscular blocking agents. Furthermore, we did not assess the use of a neuromuscular blocking agent late in the course of ARDS or use on the basis of plateaupressure or transpulmonary-pressure measurements.20 Another limitation is the absence of data on conditions known to antagonize or potentiate neuromuscular blockade. However, any condition that increases the duration of neuromuscular blockade would have adversely affected the patients receiving the neuromuscular blocking agent, in particular by increasing the duration of mechanical ventilation. The sample-size calculation was based on our two previous studies performed in four ICUs13,15 that used the same inclusion criteria as were used in the current trial and on the European epidemiologic study ALIVE.4 However, the mortality in the placebo group in this study (40.7%) is lower than n engl j med 363;12 nejm.org september 16, 2010 1113 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e Table 3. Secondary Outcomes, According to Study Group.* Cisatracurium (N = 177) Placebo (N = 162) Relative Risk with Cisatracurium (95% CI) P Value 42 (23.7 [18.1–30.5]) 54 (33.3 [26.5–40.9]) 0.71 (0.51–1.00) 0.05 In the ICU 52 (29.4 [23.2–36.5]) 63 (38.9 [31.7–46.6]) 0.76 (0.56–1.02) 0.06 In the hospital 57 (32.2 [25.8–39.4]) 67 (41.4 [34.1–49.1]) 0.78 (0.59–1.03) 0.08 Outcome Death — no. (% [95% CI]) At 28 days No. of ventilator-free days† From day 1 to day 28 10.6±9.7 8.5±9.4 0.04 From day 1 to day 90 53.1±35.8 44.6±37.5 0.03 No cardiovascular failure 18.3±9.4 16.6±10.4 0.12 No. of days without organ failure, from day 1 to day 28 No coagulation abnormalities 22.6±8.9 20.5±9.9 0.05 No hepatic failure 21.3±9.6 19.1±10.6 0.05 No renal failure 20.5±10.1 18.1±11.6 0.05 None of the four 15.8±9.9 12.2±11.1 0.01 6.9±8.2 5.7±7.8 0.16 39.5±35.6 0.03 No. of days outside the ICU From day 1 to day 28 From day 1 to day 90 Hospital survivors admitted to other health care facilities from day 1 to day 90 — % (95% CI) 47.7±33.5 22.3 (15.8–30.5) 18.8 (12.2–27.8) 0.52 Barotrauma — no. (% [95% CI])‡ 9 (5.1 [2.7–9.4]) 19 (11.7 [7.6–17.6]) 0.43 (0.20–0.93) 0.03 Pneumothorax — no. (% [95% CI]) 7 (4.0 [2.0–8.0]) 19 (11.7 [7.6–17.6]) 0.34 (0.15–0.78) 0.01 At day 28 55 (46–60) 55 (39–60) 1.07 (0.80–1.45) 0.49 At ICU discharge 55 (43–60) 55 (44–60) 0.92 (0.71–1.19) 0.94 MRC score — median (IQR)§ Patients without ICU-acquired paresis¶ By day 28 — no./total no. (% [95% CI]) By ICU discharge — no./total no. (% [95% CI]) 68/96 (70.8 [61.1–79.0]) 52/77 (67.5 [56.5–77.0]) 0.64 72/112 (64.3 [55.1–72.6]) 61/89 (68.5 [58.3–77.3]) 0.51 *Plus–minus values are means ±SD. ICU denotes intensive care unit, and IQR interquartile range. †The number of ventilator-free days was defined as the number of days since successful weaning from mechanical ventilation after a period of spontaneous breathing lasting at least 48 consecutive hours. ‡Barotrauma was defined as any new pneumothorax, pneumomediastinum, subcutaneous emphysema, or pneumatocele larger than 2 cm in diameter. § The Medical Research Council (MRC) scale is a previously validated scale that assesses strength in three muscle groups in each arm and leg. The score for each muscle group can range from 0 (paralysis) to 5 (normal strength), with the overall score ranging from 0 to 60.17 ¶ICU-acquired paresis was defined as an overall MRC score of less than 48. that in the control groups in the earlier studies. Given the observed mortality in our placebo group, the current study was underpowered. Indeed, 885 patients would have been needed to be enrolled to achieve 80% statistical power with a two-sided alpha value of 0.05. Finally, all our patients had severe ARDS. Additional work is needed to determine whether the 1114 use of neuromuscular blocking agents for only 24 hours is beneficial in selected patients. In our general analysis, which was prespecified but with post hoc determination of the threshold value for classifying subgroups, we found that the beneficial effect of the neuromuscular blocking agent on survival was confined to the two thirds of patients with a PaO2:FiO2 ratio below 120. n engl j med 363;12 nejm.org september 16, 2010 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. Neuromuscular Blocking Agents in ARDS The mechanisms underlying the beneficial effect of neuromuscular blocking agents remain speculative. A brief period of paralysis early in the course of ARDS may facilitate lung-protective mechanical ventilation by improving patient– ventilator synchrony and allowing for the accurate adjustment of tidal volume and pressure levels, thereby limiting the risk of both asynchronyrelated alveolar collapse and regional alveolarpressure increases with overdistention. Another possible mechanism of the benefit involves a decrease in lung or systemic inflammation.15 The main safety concern with the use of a neuromuscular blocking agent is muscle weakness; the risk varies among agents.21,22 Steroidal compounds (vecuronium, pancuronium, and ro curonium) may carry the highest risk of myopathy,23 although myopathy has also been reported with benzylisoquinolines, including cisatracurium besylate.24,25 Muscle weakness was not in- creased significantly by the use of the neuromuscular blocking agent in our study. The short duration of use of the neuromuscular blocking agent probably explains this result. In conclusion, this multicenter trial provides evidence that the administration of a neuromuscular blocking agent early in the course of severe ARDS managed with low-tidal-volume ventilation may improve outcomes. Future studies are needed to replicate and expand these findings before they can be widely adopted in clinical practice. Supported by the Assistance Publique–Hôpitaux de Marseille and a grant from the Ministère de la Santé (Programme Hospitalier de Recherche Clinique Régional 2004-26). GlaxoSmithKline France provided the study drugs. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank Kahena Amichi, Daniel Pulina, François Voillet, Patrick Sudour, Teodora Iordanova, Jean-Charles Reynier, Mohamed Fathallah, Karine Barrau, Pascal Auquier, Yann Lebras, Nathalie Boggi, Ammar Zerrar, Cécile Roché-Thiaux, Fabienne Brégeon, Didier Dreyfuss, Marie-Josée Bonavita, Marie-Dominique Chollet, Didier Sanchez, and Antoinette Wolfe. Appendix The ARDS et Curarisation Systematique (ACURASYS) study investigators are as follows (all in France): Marseille: Hôpital Nord — J. Albanese, V. Blaso, M. Leone, F. Antonini, P. Visintini; Hôpital Sainte-Marguerite, Réanimation des Urgences — G. Perrin, D. Blayac, B. Eon, P. Michelet, P. Saux, D. Lambert, V. Fulachier; Hôpital Sainte-Marguerite, Réanimation Médicale — J.M. Forel, A. Roch, C. Guervilly, J. Allardet-Servent, D. Demory, N. Embriaco, M. Gainnier, L. Papazian; Hôpital Ambroise Paré — J.M. Seghboyan, N. Beni Chougrane, R. Soundaravelou, A. Piera, M. Bonnetty. Paris: Hôpital Saint-Louis — E. Azoulay. Besançon: Hôpital Jean Minjoz — D. Perez, G. Capellier, P. Midez, C. Patry, E. Laurent, E. Belle, J.C. Navellou. Clermont-Ferrand: Hôpital Hôtel-Dieu — J.M. Constantin, S. Cayot-Constantin, R. Guerin. Avignon: Centre Hospitalier — P. Courant, T. Signouret, K. Pavaday, P. Garcia, L. Delapierre, K. Debbat. Toulon: Hôpital Font-Pré — J.M. Arnal, J. Durand-Gasselin, G. Corno, A. Orlando, I. Granier, S.Y. Donati. Rennes: Hôpital Pontchaillou — A. Gacouin, Y. Le Tulzo, S. Lavoue, C. Camus. Nîmes: Hôpital Caremeau — C. Gervais, C. Bengler, C. Arich, J.Y. Lefrant, P. Poupard, L. Muller, P. Joubert, R. Cohendy, G. Saissi, P. Barbaste. Lyon: Hôpital de la Croix-Rousse — C. Guérin, F. Bayle, J.C. Richard, J.M. Badet. Bordeaux: Hôpital Pellegrin — G. Hilbert, H. Bruno, E. Rosier, W. Pujol, H.N. Bui. Montpellier: Hôpital Saint-Eloi — S. Jaber, B. Jung, M. Sebbane, G. Chanques. Brest: Hôpital La Cavale Blanche — G. Prat, E. L’Her, J.M. Tonnelier. Aix-en-Provence: Centre Hospitalier du Pays d’Aix — O. Baldesi, C. Leroy, L. Rodriguez, J.L. Le Grand, B. Garrigues. Nice: Hôpital de l’Archet — G. Bernardin, J. Dellamonica. Grenoble: Hôpital Sud — C. Schwebel, J.F. Timsit, R. Hamidfar, L. Hammer, G. Dessertaine, C. De Couchon, A. Bonadona. Saint-Etienne: Hôpital Bellevue — F. Zeni, D. Thevenet, C. Venet, S. Guyomarc’h. References 1. Ware LB, Matthay MA. The acute re- spiratory distress syndrome. N Engl J Med 2000;342:1334-49. 2. Malhotra A. Low-tidal-volume ventilation in the acute respiratory distress syndrome. N Engl J Med 2007;357:1113-20. 3. Bernard GR. Acute respiratory distress syndrome: a historical perspective. Am J Respir Crit Care Med 2005;172:798-806. 4. Brun-Buisson C, Minelli C, Bertolini G, et al. Epidemiology and outcome of acute lung injury in European intensive care units: results from the ALIVE study. Intensive Care Med 2004;30:51-61. 5. Esteban A, Anzueto A, Frutos F, et al. Characteristics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study. JAMA 2002; 287:345-55. 6. Esteban A, Ferguson ND, Meade MO, et al. Evolution of mechanical ventilation in response to clinical research. Am J Respir Crit Care Med 2008;177:170-7. 7. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med 2005;353:1685-93. 8. Hansen-Flaschen JH, Brazinsky S, Basile C, Lanken PN. Use of sedating drugs and neuromuscular blocking agents in patients requiring mechanical ventilation for respiratory failure: a national survey. JAMA 1991;266:2870-5. 9. Mehta S, Burry L, Fischer S, et al. Canadian survey of the use of sedatives, analgesics, and neuromuscular blocking agents in critically ill patients. Crit Care Med 2006;34:374-80. 10. Murray MJ, Cowen J, DeBlock H, et al. Clinical practice guidelines for sustained neuromuscular blockade in the adult critically ill patient. Crit Care Med 2002;30: 142-56. 11. Samuelson KA, Larsson S, Lundberg D, Fridlund B. Intensive care sedation of mechanically ventilated patients: a national Swedish survey. Intensive Crit Care Nurs 2003;19:350-62. 12. Vender JS, Szokol JW, Murphy GS, Nitsun M. Sedation, analgesia, and neuromuscular blockade in sepsis: an evidencebased review. Crit Care Med 2004;32: Suppl:S554-S561. 13. Gainnier M, Roch A, Forel JM, et al. Effect of neuromuscular blocking agents on gas exchange in patients presenting with acute respiratory distress syndrome. Crit Care Med 2004;32:113-9. n engl j med 363;12 nejm.org september 16, 2010 1115 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. Neuromuscular Blocking Agents in ARDS 14. The Acute Respiratory Distress Syn- drome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301-8. 15. Forel JM, Roch A, Marin V, et al. Neuro muscular blocking agents decrease inflammatory response in patients presenting with acute respiratory distress syndrome. Crit Care Med 2006;34:2749-57. 16. Reisine T, Pasternak G. Opioid analgesics and antagonists. In: Hardman J, Limbird L, eds. The pharmacological basis of therapeutics. 9th ed. New York: McGraw-Hill, 1996:521-55. 17. De Jonghe B, Sharshar T, Lefaucheur JP, et al. Paresis acquired in the intensive care unit: a prospective multicenter study. JAMA 2002;288:2859-67. 18. Schoenfeld DA, Bernard GR. Statisti- cal evaluation of ventilator-free days as an efficacy measure in clinical trials of treatments for acute respiratory distress syndrome. Crit Care Med 2002;30:1772-7. 19. Hager DN, Krishnan JA, Hayden DL, Brower RG. Tidal volume reduction in patients with acute lung injury when plateau pressures are not high. Am J Respir Crit Care Med 2005;172:1241-5. 20. Talmor D, Sarge T, Malhotra A, et al. Mechanical ventilation guided by esophageal pressure in acute lung injury. N Engl J Med 2008;359:2095-104. 21. Segredo V, Caldwell JE, Matthay MA, Sharma ML, Gruenke LD, Miller RD. Persistent paralysis in critically ill patients after long-term administration of vecuronium. N Engl J Med 1992;327:524-8. 22. Murray MJ, Coursin DB, Scuderi PE, et al. Double-blind, randomized, multicenter study of doxacurium vs. pancuronium in intensive care unit patients who require neuromuscular-blocking agents. Crit Care Med 1995;23:450-8. 23. Testelmans D, Maes K, Wouters P, et al. Rocuronium exacerbates mechanical ventilation-induced diaphragm dysfunction in rats. Crit Care Med 2006;34:301823. 24. Davis NA, Rodgers JE, Gonzalez ER, Fowler AA III. Prolonged weakness after cisatracurium infusion: a case report. Crit Care Med 1998;26:1290-2. 25. Leatherman JW, Fluegel WL, David WS, Davies SF, Iber C. Muscle weakness in mechanically ventilated patients with severe asthma. Am J Respir Crit Care Med 1996;153:1686-90. Copyright © 2010 Massachusetts Medical Society. posting presentations at medical meetings on the internet Posting an audio recording of an oral presentation at a medical meeting on the Internet, with selected slides from the presentation, will not be considered prior publication. This will allow students and physicians who are unable to attend the meeting to hear the presentation and view the slides. If there are any questions about this policy, authors should feel free to call the Journal’s Editorial Offices. 1116 n engl j med 363;12 nejm.org september 16, 2010 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. new england journal of medicine The established in 1812 november 13, 2008 vol. 359 no. 20 Mechanical Ventilation Guided by Esophageal Pressure in Acute Lung Injury Daniel Talmor, M.D., M.P.H., Todd Sarge, M.D., Atul Malhotra, M.D., Carl R. O’Donnell, Sc.D., M.P.H., Ray Ritz, R.R.T., Alan Lisbon, M.D., Victor Novack, M.D., Ph.D., and Stephen H. Loring, M.D. A bs t r ac t Background Survival of patients with acute lung injury or the acute respiratory distress syndrome (ARDS) has been improved by ventilation with small tidal volumes and the use of positive end-expiratory pressure (PEEP); however, the optimal level of PEEP has been difficult to determine. In this pilot study, we estimated transpulmonary pressure with the use of esophageal balloon catheters. We reasoned that the use of pleuralpressure measurements, despite the technical limitations to the accuracy of such measurements, would enable us to find a PEEP value that could maintain oxygenation while preventing lung injury due to repeated alveolar collapse or overdistention. Methods We randomly assigned patients with acute lung injury or ARDS to undergo mechanical ventilation with PEEP adjusted according to measurements of esophageal pressure (the esophageal-pressure–guided group) or according to the Acute Respiratory Distress Syndrome Network standard-of-care recommendations (the control group). The primary end point was improvement in oxygenation. The secondary end points included respiratory-system compliance and patient outcomes. Results The study reached its stopping criterion and was terminated after 61 patients had been enrolled. The ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen at 72 hours was 88 mm Hg higher in the esophageal-pressure– guided group than in the control group (95% confidence interval, 78.1 to 98.3; P = 0.002). This effect was persistent over the entire follow-up time (at 24, 48, and 72 hours; P = 0.001 by repeated-measures analysis of variance). Respiratory-system compliance was also significantly better at 24, 48, and 72 hours in the esophagealpressure–guided group (P = 0.01 by repeated-measures analysis of variance). From the Department of Anesthesia, Critical Care, and Pain Medicine, Beth Israel Deaconess Medical Center (D.T., T.S., R.R., A.L., S.H.L.); the Division of Pulmonary and Critical Care and the Division of Sleep Medicine, Brigham and Women’s Hospital (A.M.); the Division of Pulmonary, Critical Care, and Sleep Medicine, Beth Israel Deaconess Medical Center (C.R.O.); the Harvard Clinical Research Institute (V.N.); and Harvard Medical School (D.T., T.S., A.M., C.R.O., A.L., S.H.L.) — all in Boston. Address reprint requests to Dr. Talmor at the Department of Anesthesia, Critical Care, and Pain Medicine, Beth Israel Deaconess Medical Center, 1 Deaconess Rd., CC470, Boston, MA 02215, or at dtalmor@ bidmc.harvard.edu. This article (10.1056/NEJMoa0708638) was published at www.nejm.org on November 11, 2008. N Engl J Med 2008;359:2095-104. Copyright © 2008 Massachusetts Medical Society. Conclusions As compared with the current standard of care, a ventilator strategy using esophageal pressures to estimate the transpulmonary pressure significantly improves oxygenation and compliance. Multicenter clinical trials are needed to determine whether this approach should be widely adopted. (ClinicalTrials.gov number, NCT00127491.) n engl j med 359;20 www.nejm.org november 13, 2008 2095 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. The R n e w e ng l a n d j o u r na l ecent changes in the practice of mechanical ventilation have improved survival in patients with the acute respiratory distress syndrome (ARDS), but mortality remains unacceptably high. Whereas low tidal volumes are clearly beneficial in patients with ARDS, how to choose a positive end-expiratory pressure (PEEP) is uncertain.1-4 Ideally, mechanical ventilation should provide sufficient transpulmonary pressure (airway pressure minus pleural pressure) to maintain oxygenation while minimizing repeated alveolar collapse or overdistention leading to lung injury.5 In critical illness, however, there is marked variability among patients in abdominal and pleural pressures6,7; thus, for a given level of PEEP, transpulmonary pressures may vary unpredictably from patient to patient.7 We estimated pleural pressure with the use of an esophageal balloon catheter. Although this technique has been validated in healthy human subjects and animals, it has not been systematically applied in patients in the intensive care setting. We reasoned that we could adjust PEEP according to each patient’s lung and chest-wall mechanics.8-10 We speculated that in patients with high estimated pleural pressure who are undergoing ventilation with conventional ventilator settings, underinflation may cause hypoxemia. In such patients, raising PEEP to maintain a positive transpulmonary pressure might improve aeration and oxygenation without causing overdistention. Conversely, in patients with low pleural pressure, maintaining low PEEP would keep transpulmonary pressure low, preventing overdistention and minimizing the adverse hemodynamic effects of high PEEP.11 We report the results of a randomized, controlled pilot trial involving patients with acute lung injury or ARDS. The trial compared mechan ical ventilation directed by esophageal-pressure measurements with mechanical ventilation managed according to the Acute Respiratory Distress Syndrome Network (ARDSNet) recommendations.12 We tested the hypothesis that oxygenation in patients can be improved by adjusting PEEP to maintain positive transpulmonary pressures. of m e dic i n e was approved by the institutional review board of the center, and written informed consent was obtained from the patients or their nearest relatives. No commercial entities providing equipment or devices had a role in any aspect of this study. Patients were included in the study if they had acute lung injury or ARDS according to the American–European Consensus Conference definitions.13 The exclusion criteria included recent injury or other pathologic condition of the esophagus, major bronchopleural fistula, and solidorgan transplantation. Measurements and Experimental Protocol While undergoing treatment, the subjects were supine, with the head of the bed elevated to 30 degrees. Airway pressure, tidal volume, and air flow were recorded during mechanical ventilation. An esophageal balloon catheter was passed to a depth of 60 cm from the incisors for measurement of gastric pressure and then withdrawn to a depth of 40 cm to record esophageal pressure during mechanical ventilation. Placement of the balloon in the stomach was confirmed by a transient increase in pressure during a gentle compression of the abdomen and by a qualitative change in the pressure tracing (i.e., an increased cardiac artifact) as the balloon was withdrawn into the esophagus. In approximately one third of the patients, the balloon could not be passed into the stomach, and esophageal placement was confirmed by the presence of a cardiac artifact and the changes in transpulmonary pressure during tidal ventilation. The mixed expired partial pressure of carbon dioxide was measured to allow calculation of physiological dead space. After these initial measurements, patients were randomly assigned with the use of a block-randomization scheme to the control or esophagealpressure–guided group. Each patient, while under heavy sedation or paralysis, underwent a recruitment maneuver to standardize the history of lung volume,14 in which airway pressure was increased to 40 cm of water for 30 seconds. If needed, a lower pressure was used to keep the transpulmonary pressure (the difference between the airway pressure and the esophageal pressure) in the physiologic range Me thods (<25 cm of water while the patient is in the supine Patients position).15 After the recruitment maneuver, the We performed the trial in the medical and surgi- patient underwent mechanical ventilation accordcal intensive care units (ICUs) of Beth Israel Dea- ing to the treatment assignment. coness Medical Center in Boston. The protocol The patients in the esophageal-pressure–guided 2096 n engl j med 359;20 www.nejm.org november 13, 2008 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. Mechanical Ventilation in Acute Lung Injury group underwent mechanical ventilation with settings determined by the initial esophageal-pressure measurements. Tidal volume was set at 6 ml per kilogram of predicted body weight. The predicted body weight of male patients was calculated as 50 + 0.91 × (centimeters of height – 152.4) and that of female patients as 45.5 + 0.91 × (centimeters of height − 152.4). PEEP levels were set to achieve a transpulmonary pressure of 0 to 10 cm of water at end expiration, according to a sliding scale based on the partial pressure of arterial oxygen (PaO2) and the fraction of inspired oxygen (FiO2) (Fig. 1). We also limited tidal volume to keep transpulmonary pressure at less than 25 cm of water at end inspiration, although this limit was rarely approached, and tidal volume was never reduced for this purpose. Patients in the control group were treated according to the low-tidal-volume strategy reported by the ARDSNet study of the National Heart, Lung, and Blood Institute.12 This strategy specifies that the tidal volume is set at 6 ml per kilogram of predicted body weight and PEEP is based on the patient’s PaO2 and FiO2 (Fig. 1). In both groups, the goals of mechanical ventilation included a PaO2 of 55 to 120 mm Hg or a pulse-oximeter reading of 88 to 98%, an arterial pH of 7.30 to 7.45, and a partial pressure of arterial carbon dioxide (PaCO2) of 40 to 60 mm Hg, according to the sliding scales in Figure 1. To reduce the need for frequent manipulation of the ventilator settings, the goals for oxygenation in both groups were relaxed from the narrow range of PaO2 values in the ARDSNet study (55 to 80 mm Hg) to a broader range of 55 to 120 mm Hg. All measurements were repeated 5 minutes after the initiation of experimental or control ventilation and again at 24, 48, and 72 hours. Measurements were also performed as needed after changes were made to ventilator settings because of any clinically significant change in the patient’s condition. Therapies other than mechanical ventilation were administered by members of the primary ICU team, who were unaware of the results of the esophageal-pressure measurements. To avert complications, these team members used protocols to guide hemodynamic resuscitation,16 sedation, weaning from ventilation, and other standard interventions related to ventilator care.17 These care standards were aggressively applied in both groups. After the measurements at 72 hours, the results of pressure measurements were made available to the caregivers, who were free to use or not use them for decisions concerning treatment and ventilator management. The primary end point of the study was arterial oxygenation, as measured by the ratio of PaO2 to FiO2 (PaO2:FiO2) 72 hours after randomization. The secondary end points included indexes of lung mechanics and gas exchange (respiratorysystem compliance and the ratio of physiological dead space to tidal volume), as well as outcomes of the patients (the number of ventilator-free days at 28 days, length of stay in the ICU, and death within 28 days and 180 days after treatment). Esophageal-Pressure–Guided Group FIO2 PLexp 0.4 0 0.5 0 0.5 2 0.6 2 0.6 4 0.7 4 0.7 6 0.8 6 0.8 8 0.9 8 0.9 10 1.0 10 Control Group FIO2 PEEP 0.3 5 0.4 5 0.4 8 0.5 8 0.5 10 0.6 10 0.7 10 0.7 12 0.7 14 0.8 14 0.9 14 0.9 16 0.9 18 1.0 20–24 Figure 1. Ventilator Settings According to the Protocol. For the intervention group, keep the partial pressure of arterial oxygen (PaO 55 and 120 mm Hg or keep 2) between 1st RETAKE AUTHOR: Talmor ICM the oxygen saturation, as measured by pulse oximeter, between 88 and 98% by using 2nd the ventilator settings in one FIGURE: 1 of 2 REG F column at a time. Set the positive end-expiratory pressure (PEEP) at such a level that 3rd transpulmonary pressure durRevised ing end-expiratory occlusion (PL exp)CASE stays between 0 and 10 cm of water, and keep transpulmonary pressure during Line 4-C EMail25 cm of water. For the control group, keep SIZE PaO2 between 55 and 120 mm Hg end-inspiratory occlusion at less than ARTIST: ts H/T H/T Enonto pulse oximeter between 88 and 98%)33p9 (or keep oxygen saturation according by using the ventilator settings in one Combo column at a time. Set the PEEP and tidal volume at such levels that the airway pressure during end-inspiratory ocAUTHOR, PLEASE NOTE: clusion stays at less than 30 cm of water. In has both groups, apply ventilation Figure been redrawn and type has beenwith reset.either pressure-control ventilation or Please check carefully. time between 1:1 and 1:3 to minimize dysvolume-control ventilation with a ratio of inspiratory time to expiratory synchrony between the patient and the ventilator while achieving a tidal volume of 6±2 ml per kilogram of predicted JOB:of35915 ISSUE: 10-09-08 maneuvers are permitted to body weight and a respiratory rate 35 breaths per minute or less. Lung-recruitment reverse episodic hypoxemia after suctioning or inadvertent airway disconnection, but not on a routine basis. n engl j med 359;20 www.nejm.org november 13, 2008 2097 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l Statistical Analysis In evaluating the PaO2:FiO2 at 72 hours, we decided a priori that a clinically important change in the PaO2:FiO2 would be approximately 20%, with measurement error taken into account. To determine sample size, we chose a minimal average between-groups difference of 40 in the PaO2:FiO2. We conservatively estimated the standard deviation to be 100 (equivalent to a coefficient of variation of 250%); on the basis of this estimate, a sample of 100 patients per group would be required to detect a difference of 40 in the PaO2:FiO2 with 80% power and a two-tailed alpha value of 0.05. Because of the uncertainty in the estimate of standard deviation, we designed the study with the aid of a data safety and monitoring board, whose members were not involved in patient care or data gathering. The board members were instructed to perform an interim analysis after 60 patients had been enrolled, at which point they could recommend stopping the trial if an overwhelming effect was detected on the basis of the critical significance level (P≤0.02), as adjusted for the Lan–DeMets alpha-spending function with Pocock boundary. The members of this board also participated in the writing of this article. Continuous variables with normal distribution are presented as means (±SD) and compared with the use of Student’s t-test. Continuous variables with non-normal distributions are presented as medians and interquartile ranges and compared with the use of the Mann–Whitney test. Dichotomous or nominal categorical variables are compared with the use of the chi-square test with normal approximation or Fisher’s exact test, as appropriate. We assessed the trend over time in respiratory measurements by comparing the control group and the esophageal-pressure–guided group at 24, 48, and 72 hours with the use of the F test with one degree of freedom for a general linear model with repeated measures. We used sequential hypothesis testing for the assessment of differences between the groups at 72 hours and 24 hours. When a statistically significant difference was found at 72 hours, we performed a repeated-measures analysis and then compared the values at 24 hours. Kaplan–Meier analysis with the log-rank test was applied to compare survival at 180 days between the groups. In a single prespecified analysis, we adjusted the relative risk by using the Acute Physiology and Chronic Health Evaluation (APACHE II) score to 2098 of m e dic i n e estimate the effect of study group on the risk of death within 28 days after treatment. The relative risk was estimated by Poisson regression with conservative robust error variance.18,19 For death within 180 days after treatment, we used a Cox proportional-regression model to compare the control and treatment groups, with adjustment for the APACHE II score at admission. A twotailed P value less than 0.05 was considered to indicate statistical significance. R e sult s The characteristics of the patients in the two groups were well matched at baseline (Table 1). Most patients in both groups were severely ill, with a mean (±SD) APACHE II score of 26.6±6.4 and a median of two failed organs (interquartile range, one to three). We were unable to sedate one patient in the esophageal-pressure–guided group sufficiently to obtain stable esophagealpressure measurements; this patient is included in the analysis on the basis of the intention-totreat principle. There were no adverse events or incidents of barotrauma in either group. We stopped the study after 61 patients had been enrolled, because the planned interim analy sis showed that it had reached the prespecified stopping criterion. The PaO2:FiO2 at 72 hours was 88 mm Hg higher in patients treated with mechanical ventilation with esophageal balloons than in control patients (95% confidence interval [CI], 78.1 to 98.3; P = 0.002) (Table 2). Physiological Measurements The ventilator settings and physiological measurements at baseline were similar in the two groups (Table 2). Forty-nine patients (80%), including the one patient that we were unable to sedate, met the criteria for ARDS (PaO2:FiO2 <200 mm Hg) (see Table 1 in the Supplementary Appendix, available with the full text of this article at www. nejm.org), and there was no significant difference in baseline PaO2:FiO2 between the groups. The average tidal volume was reduced during the first day of therapy by 67 ml in the control group (P<0.001 by paired t-test) and by 44 ml in the esophageal-pressure–guided group (P<0.001 by paired t-test). Oxygenation and respiratory-system compliance improved in the esophageal-pressure–guided group as compared with the control group, whereas the ratio of dead space to tidal volume did not n engl j med 359;20 www.nejm.org november 13, 2008 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. Mechanical Ventilation in Acute Lung Injury Table 1. Baseline Characteristics of the Patients.* Esophageal-Pressure–Guided (N = 30) Characteristic Male sex — no. (%) Conventional Treatment (N = 31) P Value 0.44 19 (63) 17 (55) 54.5±16.1 51.2±23.0 0.52 26 (87) 27 (87) 0.96 Predicted body weight — kg 67.1±8.9 63.2±11.1 0.14 APACHE II score at admission 26.3±6.4 26.8±6.5 0.76 Age — yr White race — no. (%)† Primary physiological injury — no. (%)‡ 0.54 Pulmonary 7 (23) 5 (16) Abdominal 13 (43) 11 (35) Trauma 6 (20) 9 (29) Sepsis 3 (10) 2 (6) Other 1 (3) 4 (13) Organ failure at baseline — no. (%) Cardiac 10 (33) 10 (32) 0.93 Renal 19 (63) 16 (52) 0.36 Neurologic 12 (40) 12 (39) 0.92 Hepatic 11 (37) 10 (32) 0.72 7 (23) 5 (16) 0.48 7.34±0.09 7.32±0.08 0.34 Hematologic Arterial blood gases at baseline pH PaCO2 — mm Hg 42±8 40±8 0.23 PaO2 — mm Hg§ 91±25 107±44 0.09 Bicarbonate — mmol/liter 24±5 22±4 0.05 3.1±3.5 3.4±3.3 0.83 Hemodynamic variables at baseline Lactate — mg/dl Heart rate — beats/min 98±26 100±19 0.71 Systolic blood pressure — mm Hg 108±18 107±18 0.80 Diastolic blood pressure — mm Hg 58±11 54±11 0.20 Central venous pressure — mm Hg 16±5 16±4 0.96 *Plus–minus values are means ±SD. APACHE denotes Acute Physiology and Chronic Health Evaluation, PaCO2 the partial pressure of arterial carbon dioxide, and PaO2 the partial pressure of arterial oxygen. †Race was determined by the investigators. ‡Pulmonary injury included pneumonia (nine patients), aspiration pneumonitis (two), and smoke inhalation (one). Abdom inal injury included bowel obstruction (four patients), abdominal surgery (four), pancreatitis (four), cholangitis (two), small-bowel perforation (three), ruptured aortic aneurysm or surgery for aortic aneurysm (two), gastrointestinal bleeding (one), Crohn’s disease (one), end-stage liver disease (one), ischemic bowel (one), and perforated viscus (one). Trauma included motor vehicle accident (five patients), multiple trauma (eight), abdominal gunshot wound (one), and traumatic brain injury (one). Other injuries included drug overdose (three patients), intraventricular hemorrhage (one), and hypoxic respiratory failure (one). § For the PaO2, the values for the fraction of inspired oxygen were as follows: for the esophageal-pressure–guided group, the median was 0.6 and the interquartile range was 0.5 to 0.8; for the conventional-treatment group, the median was 0.7 and the interquartile range was 0.6 to 1.0. significantly differ between the groups during the first 72 hours (Fig. 2C). The PaO2:FiO2 improved during the first 72 hours by 131 mm Hg (95% CI, 79 to 182) in the esophageal-pressure– guided group and by 49 mm Hg (95% CI, 12 to 86) in the control group (Table 2). The higher value of PaO2:FiO2 in the esophageal-pressure– guided group than in the control group was evident at 24 hours (P = 0.04) (Fig. 2A). Respiratorysystem compliance was significantly improved and was higher in the esophageal-pressure– guided group than in the control group (P = 0.01; n engl j med 359;20 www.nejm.org november 13, 2008 2099 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e Table 2. Measurements of Ventilatory Function at Baseline and 72 Hours.* Measurement Baseline EsophagealConventional Pressure–Guided Treatment (N = 30) (N = 31) PaO2:FiO2 Respiratory-system compliance (ml/cm of water) Ratio of physiological dead space to tidal volume PaO2 (mm Hg) FiO2 PEEP (cm of water) Tidal volume (ml) Tidal volume (ml per kg of predicted body weight) Respiratory rate (breaths/min) Inspiratory time (sec) PEEPtotal (cm of water) Peak inspiratory pressure (cm of water) Mean airway pressure (cm of water) Plateau pressure (cm of water) Transpulmonary end-inspiratory pressure (cm of water) Transpulmonary end-expiratory pressure (cm of water) Esophageal end-inspiratory pressure (cm of water) Esophageal end-expiratory pressure (cm of water) 72 Hr† P Value EsophagealPressure–Guided (N = 29) Conventional Treatment (N = 29) 147±56 36±12 145±57 36±10 0.89 0.94 280±126 45±14 191±71 35±9 P Value 0.002 0.005 0.67±0.11 0.67±0.09 0.95 0.61±0.09 0.64±0.10 0.27 91±25 0.66±0.17 13±5 484±98 7.3±1.3 107±44 0.77±0.18 13±3 491±105 7.9±1.4 0.09 0.02 0.73 0.80 0.12 124±44 0.49±0.17 17±6 472±98 7.1 ±1.3 101±33 0.57±0.18 10±4 418±80 6.8±1 0.03 0.07 <0.001 0.03 0.31 26±6 0.8±0.1 14±5 35±8 20±6 29±7 7.9±6.0 24±6 0.9±0.2 15±4 35±7 20±4 29±5 8.6±5.4 0.32 0.19 0.67 0.85 0.88 0.79 0.61 26±6 0.8±0.1 18±5 32±8 22±6 28±7 7.4±4.4 28±5 0.8±0.1 12±5 28±7 16±5 25±6 6.7±4.9 0.20 0.27 <0.001 0.007 0.001 0.07 0.58 −2.8±5.0 −1.9±4.7 0.49 0.1±2.6 −2.0±4.7 0.06 21.2±4.9 20.7±5.1 0.68 21.7±7.2 17.9±5.2 0.03 17.2±4.4 16.9±5.0 0.79 18.4±5.9 14.3±4.9 0.008 *Plus–minus values are means ±SD. FiO2 denotes the fraction of inspired oxygen, PaO2 the partial pressure of arterial oxygen, PEEP positive end-expiratory pressure applied by the ventilator, and PEEPtotal airway pressure measured during end-expiratory occlusion. †The values are given for the 29 surviving patients in each treatment group. repeated-measures analysis of variance at 24, 48, and 72 hours) (Fig. 2B). On the first therapeutic day, PEEP was changed by less than 5 cm of water in all but one of the control patients, whereas patients in the esophageal-pressure–guided group had variable and often substantial increases in PEEP (Table 3) and significantly higher PEEP at 24, 48, and 72 hours (Fig. 2D, and Fig. 1 in the Supplementary Appendix). At 24 hours, the difference in PEEP between the groups reached 7.7 cm of water (95% CI, 5.5 to 9.9), with a mean PEEP in the esophageal-pressure–guided group of 18.7±5.1 cm of water, although in 3 of the 31 patients in this group, the initial PEEP level was decreased on the basis of initial transpulmonary pressure. At 24, 48, and 72 hours, the mean transpulmonary endexpiratory pressure remained above zero in the esophageal-pressure–guided group, whereas it remained negative in the control group (P<0.001 by 2100 repeated-measures analysis of variance) (Fig. 2E). The plateau airway pressure during end-inspiratory occlusion was higher in the esophagealpressure–guided group than in the control group (P = 0.003 by repeated-measures analysis of variance) (Fig. 2F, and Fig. 1 in the Supplementary Appendix). However, transpulmonary pressures during end-inspiratory occlusion never exceeded 24 cm of water and did not differ significantly between the groups (P = 0.13 by repeated-measures analysis of variance) (Fig. 2G). Clinical Outcomes Table 4 presents the clinical outcomes, all of which were prespecified secondary outcomes. There was no significant difference between the groups in ventilator-free days at day 28 or length of stay in the ICU. The 28-day mortality rate in the entire study cohort was 17 of 61 patients (28%). As would be expected, the APACHE II n engl j med 359;20 www.nejm.org november 13, 2008 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. Mechanical Ventilation in Acute Lung Injury B Respiratory-System Compliance (ml/cm of water) A 350 PaO2:FIO2 300 Esophageal pressure 250 200 150 Conventional treatment P=0.002 100 50 0 Baseline 24 Hr 48 Hr 72 Hr 0.70 20.0 P=0.01 10.0 0.0 Baseline 24 Hr 48 Hr 72 Hr PEEP (cm of water) 0.60 Esophageal pressure 0.55 P=0.29 15.0 10.0 Conventional treatment 5.0 0.50 0 Esophageal pressure 20.0 Conventional treatment 0.65 Baseline 24 Hr 48 Hr 0.0 72 Hr Baseline 24 Hr 48 Hr P<0.001 72 Hr F 3.0 2.0 Esophageal pressure 1.0 0.0 −1.0 −2.0 −3.0 −4.0 −5.0 Conventional treatment Baseline 24 Hr P<0.001 48 Hr Transpulmonary End-Inspiratory Pressure (cm of water) Transpulmonary End-Expiratory Pressure (cm of water) Conventional treatment 30.0 25.0 E G Esophageal pressure 40.0 D 0.75 Plateau Pressure (cm of water) Ratio of Dead Space to Tidal Volume Ratio C 50.0 72 Hr 35.0 Esophageal pressure 30.0 25.0 Conventional treatment 20.0 15.0 P=0.003 10.0 5.0 0.0 Baseline 24 Hr 48 Hr 72 Hr 12.0 Esophageal pressure 10.0 8.0 6.0 4.0 Conventional treatment P=0.13 2.0 0.0 Baseline 24 Hr 48 Hr 72 Hr 1st Esophageal-Pressure–Guided Groups. Figure 2. Respiratory Measurements at Baseline andAUTHOR: at 24, 48,Talmor and 72 Hours in theRETAKE Control and ICM 2nd 2 of 2by repeated-measures analysis Means and standard errors are shown. P values calculated of variance. Panel A shows the ratio of REG were F FIGURE: 3rd the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO :F i O ), Panel B respiratory-system compliance, Panel C the 2 2 CASE Revised Line 4-C ratio of dead space to tidal volume, Panel D positive end-expiratory pressure (PEEP), Panel EMail SIZEE transpulmonary end-expiratory pressure, ARTIST: ts H/T H/T Panel F plateau pressure, and Panel G transpulmonary end-inspiratory pressure. 36p6 Enon Combo AUTHOR, PLEASE NOTE: Figure has been redrawn and type has been reset. patients survivors (153.2±53.7 and Please check carefully. score at admission was higher among 143.8±58.0 mm Hg, who died than among those who survived (31.5±4.5 respectively; P = 0.56). 35915 ISSUE: JOB: baseline vs. 24.7±6.1, P<0.001). However, the The mortality rate at 10-09-08 28 days was lower among PaO2:FiO2 was similar among survivors and non- patients in the esophageal-pressure–guided group n engl j med 359;20 www.nejm.org november 13, 2008 2101 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e Table 3. Changes in PEEP at the Initiation of Ventilation According to the Protocol.* Treatment Group Change in PEEP −1 to −6 cm of Water Esophageal-pressure–guided group Control group 0 to 5 cm of Water 6 to 10 cm of Water 11 to 15 cm of Water 16 to 20 cm of Water 3 9 no. of patients 12 4 2 12 18 1 0 0 *PEEP denotes applied positive end-expiratory pressure. than among control patients, although the difference was not significant (relative risk, 0.43; 95% CI, 0.17 to 1.07; P = 0.06). Multivariable analysis showed that after adjustment for baseline APACHE II score (relative risk per point of score, 1.16; 95% CI, 1.09 to 1.23; P<0.001), the esophageal-pressure protocol was associated with a significant reduction in 28-day mortality as compared with conventional treatment (relative risk, 0.46; 95% CI, 0.19 to 1.0; P = 0.049). The mortality rate at 180 days did not differ significantly between the treatment groups; the point estimate for the relative risk of death in the esophageal-pressure–guided group was 0.59 (95% CI, 0.29 to 1.20) as compared with the control group. However, a Kaplan–Meier survival plot (see Fig. 2 in the Supplementary Appendix) shows separation between the curves that persists at 180 days. Cox regression modeling showed that after adjustment for baseline APACHE II score (hazard ratio per point, 1.12; 95% CI, 1.04 to 1.22), the hazard ratio for 180-day mortality was 0.52 in the esophageal-pressure–guided group (95% CI, 0.22 to 1.25) as compared with the control group. Discussion We found that it is feasible to make repeated measurements of esophageal pressure that are of adequate fidelity and quality to be used to manage the treatment of patients requiring mechanical ventilation. Patients with acute lung injury or ARDS treated in this way had significantly improved oxygenation, as measured by the PaO2:FiO2, and significantly improved respiratory-system compliance. Moreover, these improvements were achieved without elevating transpulmonary pressure at end inspiration above the physiologic range. Finally, these improvements in lung function were associated with a trend toward improved 28-day survival in this group of very sick patients. 2102 Numerous animal models of acute lung injury have shown that reducing end-expiratory lung volume or pressure can be injurious, even when tidal volume or peak pressure is controlled.20-24 In these models, increasing PEEP can be protec tive.25,26 However, in patients with ARDS, effective adjustment of PEEP to the physiological features of the individual patient has been difficult to achieve. For example, in the ARDSNet study of low tidal volume, PEEP and FiO2 were adjusted according to arterial oxygenation without reference to chest-wall or lung mechanics.12 The subsequent Assessment of Low Tidal Volume and Elevated End-Expiratory Volume to Obviate Lung Injury (known as the ALVEOLI trial) (ClinicalTrials.gov number, NCT00000579) compared an increased level of PEEP with standard PEEP, with both levels adjusted according to the patient’s oxygenation, and showed no benefit.1 The recent Lung Open Ventilation Study (NCT00182195) used a similar approach to adjustment of PEEP, without benefit.2 The Expiratory Pressure Study Group trial (NCT00188058) increased PEEP in the intervention group to reach a plateau pressure of 28 to 30 cm of water. This study showed improvements in ventilator-free and organ-failure-free days, oxygenation, and respiratory-system compliance but showed no significant change in survival.3 Other studies, including those using the lower point of maximum curvature on the pressure–volume curve or the stress index, have had mixed results.27-31 The disappointing results of these previous studies may be due in part to their inclusion of patients with elevated pleural or intraabdominal pressure32,33 and elevated esophageal pressure.7 The lungs of such patients may be effectively compressed by high pleural pressures, and their alveoli may collapse at end expiration, despite levels of PEEP that would be adequate in other patients. By using esophageal-pressure measurements to determine PEEP, we may have prevented repeated alveolar collapse or overdistention.5 In the pres- n engl j med 359;20 www.nejm.org november 13, 2008 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. Mechanical Ventilation in Acute Lung Injury Table 4. Clinical Outcomes.* Esophageal-Pressure–Guided (N = 30) Conventional Treatment (N = 31) P Value 28-Day mortality — no. (%) 5 (17) 12 (39) 0.055 180-Day mortality — no. (%) 8 (27) 14 (45) 0.13 15.5 13.0 10.8–28.5 7.0–22.0 Outcome Length of ICU stay — days 0.16 Median Interquartile range No. of ICU-free days at 28 days Median Interquartile range 0.96 5.0 4.0 0.0–14.0 0.0–16.0 11.5 7.0 0.0–20.3 0.0–17.0 12.0 16.0 7.0–27.5 7.0-20.0 No. of ventilator-free days at 28 days Median Interquartile range 0.50 No. of days of ventilation among survivors Median Interquartile range 0.71 *For patients who were deceased at day 28, a value of 0 days was assigned. ICU denotes intensive care unit. ent pilot study, PEEP was lowered in 3 of the 30 patients treated with the use of esophageal pressure to determine PEEP and in 12 of the 31 patients treated according to the ARDSNet protocol. More importantly, PEEP was increased by more than 5 cm of water in 18 patients treated with the use of esophageal pressure to determine PEEP and in only 1 patient treated according to the ARDSnet protocol (Table 3). Thus, the key difference between the two approaches appears to be that measurement of esophageal pressure identifies patients who derive benefit from higher levels of PEEP than would ordinarily be used. Although no adverse events resulting from this strategy were observed, we would have been able to observe only adverse events that occurred at very high frequency, because of the small size of the trial. There is currently mistrust of the use of esophageal-pressure measurements in supine, critically ill patients, largely because of possible artifacts associated with body position and lung pathologic conditions.34 Although transpulmonary pressure–volume curves have been used to characterize lung disease, esophageal-pressure measurements are not usually used to manage mechanical ventilation in patients with acute lung injury or ARDS.35 However, artifacts in esophageal pressure may not be large enough to obscure differences in esophageal and pleural pressures among patients with acute lung injury or ARDS. For example, the average difference in esophageal pressure measured in the upright and the supine position that was imposed by cardiac weight was 2.9±2.1 cm of water,10 and mechanical abnormalities in diseased lungs may reduce tidal excursions in esophageal pressure by a few centimeters of water.34 By contrast, in patients with acute respiratory failure, end-expiratory esophageal pressures ranged from 4 to 32 cm of water.7 Our study has several limitations. It was a single-center study with physiologically expert staff and a small sample, and although we enrolled medical and surgical patients with various diseases, the findings cannot be generalized until they are confirmed in a larger trial powered to detect changes in appropriate clinical end points. Since our primary end point was oxygenation, which is known to be improved by applied PEEP,36 and improvements in patient oxygenation have been associated with unchanged or increased mortality when these were obtained at the cost of higher airway pressures, one cannot be sure of a favorable outcome until a larger trial has been completed.2,3,12 In conclusion, adjustment of the settings of mechanical ventilation for patients with acute lung injury or ARDS on the basis of the patients’ estimated transpulmonary pressure may have clinical benefit. This approach shows promise for improvement in lung function and survival that warrants further investigation. n engl j med 359;20 www.nejm.org november 13, 2008 2103 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. Mechanical Ventilation in Acute Lung Injury Supported in part by a grant from the National Heart, Lung, and Blood Institute (HL-52586). Dr. Malhotra reports receiving consulting fees from Respironics. Mr. Ritz reports receiving consulting fees from INO Therapeutics and lecture fees from Tri-anim and serving on the advisory boards of Cardinal Health and Respironics. No other potential conflict of interest relevant to this article was reported. References 1. Brower RG, Lanken PN, MacIntyre N, et al. Higher versus lower positive endexpiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med 2004;351:327-36. 2. Meade MO, Cook DJ, Guyatt GH, et al. Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008;299:637-45. 3. Mercat A, Richard JC, Vielle B, et al. Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: a randomized controlled trial. JAMA 2008;299:646-55. 4. Grasso S, Fanelli V, Cafarelli A, et al. Effects of high versus low positive endexpiratory pressures in acute respiratory distress syndrome. Am J Respir Crit Care Med 2005;171:1002-8. 5. Slutsky AS. Lung injury caused by mechanical ventilation. Chest 1999;116:Suppl: 9S-15S. 6. Malbrain ML, Chiumello D, Pelosi P, et al. Incidence and prognosis of intra abdominal hypertension in a mixed population of critically ill patients: a multiplecenter epidemiological study. Crit Care Med 2005;33:315-22. 7. Talmor D, Sarge T, O’Donnell CR, et al. Esophageal and transpulmonary pressures in acute respiratory failure. Crit Care Med 2006;34:1389-94. 8. Milic-Emili J, Mead J, Turner JM, Glauser EM. Improved technique for estimating pleural pressure from esophageal baloons. J Appl Physiol 1964;19:207-11. 9. Pelosi P, Goldner M, McKibben A, et al. Recruitment and derecruitment during acute respiratory failure: an experimental study. Am J Respir Crit Care Med 2001; 164:122-30. 10. Washko GR, O’Donnell CR, Loring SH. Volume-related and volume-independent effects of posture on esophageal and transpulmonary pressures in healthy subjects. J Appl Physiol 2006;100:753-8. 11. Beyer J, Beckenlechner P, Messmer K. The influence of PEEP ventilation on organ blood flow and peripheral oxygen delivery. Intensive Care Med 1982;8:75-80. 12. 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;342:1301-8. 13. Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus 2104 Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818-24. 14. Nishida T, Suchodolski K, Schettino GP, Sedeek K, Takeuch M, Kacmarek RM. Peak volume history and peak pressurevolume curve pressures independently affect the shape of the pressure-volume curve of the respiratory system. Crit Care Med 2004;32:1358-64. 15. Colebatch HJ, Greaves IA, Ng CK. Exponential analysis of elastic recoil and aging in healthy males and females. J Appl Physiol 1979;47:683-91. 16. Shapiro NI, Howell MD, Talmor D, et al. Implementation and outcomes of the Multiple Urgent Sepsis Therapies (MUST) protocol. Crit Care Med 2006;34:1025-32. 17. Resar R, Pronovost P, Haraden C, Simmonds T, Rainey T, Nolan T. Using a bundle approach to improve ventilator care processes and reduce ventilator-associated pneumonia. Jt Comm J Qual Patient Saf 2005;31:243-8. 18. McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003;157:940-3. 19. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004;159:702-6. 20. Chiumello D, Pristine G, Slutsky AS. Mechanical ventilation affects local and systemic cytokines in an animal model of acute respiratory distress syndrome. Am J Respir Crit Care Med 1999;160:109-16. 21. Faridy EE, Permutt S, Riley RL. Effect of ventilation on surface forces in excised dogs’ lungs. J Appl Physiol 1966;21:1453-62. 22. Muscedere JG, Mullen JB, Gan K, Slutsky AS. Tidal ventilation at low airway pressures can augment lung injury. Am J Respir Crit Care Med 1994;149:1327-34. 23. Webb HH, Tierney DF. Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures: protection by positive endexpiratory pressure. Am Rev Respir Dis 1974;110:556-65. 24. Wyszogrodski I, Kyei-Aboagye K, Taeusch HW Jr, Avery ME. Surfactant inactivation by hyperventilation: conservation by end-expiratory pressure. J Appl Physiol 1975;38:461-6. 25. Tremblay L, Valenza F, Ribeiro SP, Li J, Slutsky AS. Injurious ventilatory strategies increase cytokines and c-fos m-RNA expression in an isolated rat lung model. J Clin Invest 1997;99:944-52. 26. Imai Y, Parodo J, Kajikawa O, et al. Injurious mechanical ventilation and endorgan epithelial cell apoptosis and organ dysfunction in an experimental model of acute respiratory distress syndrome. JAMA 2003;289:2104-12. 27. Villar J, Kacmarek RM, Pérez-Méndez L, Aguirre-Jaime A. A high positive endexpiratory pressure, low tidal volume ventilatory strategy improves outcome in persistent acute respiratory distress syndrome: a randomized, controlled trial. Crit Care Med 2006;34:1311-8. 28. Maggiore SM, Jonson B, Richard JC, Jaber S, Lemaire F, Brochard L. Alveolar derecruitment at decremental positive endexpiratory pressure levels in acute lung injury: comparison with the lower inflection point, oxygenation, and compliance. Am J Respir Crit Care Med 2001;164:795-801. 29. Jonson B, Richard JC, Straus C, Mancebo J, Lemaire F, Brochard L. Pressurevolume curves and compliance in acute lung injury: evidence of recruitment above the lower inflection point. Am J Respir Crit Care Med 1999;159:1172-8. 30. Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998; 338:347-54. 31. Harris RS, Hess DR, Venegas JG. An objective analysis of the pressure-volume curve in the acute respiratory distress syndrome. Am J Respir Crit Care Med 2000; 161:432-9. 32. Malbrain ML. Abdominal pressure in the critically ill: measurement and clinical relevance. Intensive Care Med 1999;25: 1453-8. 33. Chieveley-Williams S, Dinner L, Puddi combe A, Field D, Lovell AT, Goldstone JC. Central venous and bladder pressure reflect transdiaphragmatic pressure during pressure support ventilation. Chest 2002;121: 533-8. 34. de Chazal I, Hubmayr RD. Novel aspects of pulmonary mechanics in intensive care. Br J Anaesth 2003;91:81-91. 35. Gattinoni L, Carlesso E, Cadringher P, Valenza F, Vagginelli F, Chiumello D. Physical and biological triggers of ventilator-induced lung injury and its prevention. Eur Respir J Suppl 2003;47:15s-25s. 36. Villar J, Pérez-Méndez L, López J, et al. An early PEEP/FIO2 trial identifies different degrees of lung injury in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 2007;176:795804. Copyright © 2008 Massachusetts Medical Society. n engl j med 359;20 www.nejm.org november 13, 2008 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission. Copyright © 2008 Massachusetts Medical Society. All rights reserved. Effects of Prone Positioning on Lung Protection in Patients with Acute Respiratory Distress Syndrome Rodrigo A. Cornejo1, Juan C. Dı́az2, Eduardo A. Tobar1, Alejandro R. Bruhn3, Cristobal A. Ramos2, Roberto A. González1, Claudia A. Repetto1, Carlos M. Romero1, Luis R. Gálvez1, Osvaldo Llanos1, Daniel H. Arellano1, Wilson R. Neira1, Gonzalo A. Dı́az1, Anı́bal J. Zamorano1, and Gonzalo L. Pereira2 1 Unidad de Pacientes Crı́ticos, Departamento de Medicina, Hospital Clı́nico Universidad de Chile; 2Departamento de Radiologı́a, Hospital Clı́nico Universidad de Chile, Santiago, Chile; and 3Departamento de Medicina Intensiva, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile Rationale: Positive end-expiratory pressure (PEEP) and prone positioning may induce lung recruitment and affect alveolar dynamics in acute respiratory distress syndrome (ARDS). Whether there is interdependence between the effects of PEEP and prone positioning on these variables is unknown. Objectives: To determine the effects of high PEEP and prone positioning on lung recruitment, cyclic recruitment/derecruitment, and tidal hyperinflation and how these effects are influenced by lung recruitability. Methods: Mechanically ventilated patients (VT 6 ml/kg ideal body weight) underwent whole-lung computed tomography (CT) during breath-holding sessions at airway pressures of 5, 15, and 45 cm H2O and Cine-CTs on a fixed thoracic transverse slice at PEEP 5 and 15 cm H2O. CT images were repeated in supine and prone positioning. A recruitment maneuver at 45 cm H2O was performed before each PEEP change. Lung recruitability was defined as the difference in percentage of nonaerated tissue between 5 and 45 cm H2O. Cyclic recruitment/derecruitment and tidal hyperinflation were determined as tidal changes in percentage of nonaerated and hyperinflated tissue, respectively Measurements and Main Results: Twenty-four patients with ARDS were included. Increasing PEEP from 5 to 15 cm H2O decreased nonaerated tissue (501 6 201 to 322 6 132 grams; P , 0.001) and increased tidal-hyperinflation (0.41 6 0.26 to 0.57 6 0.30%; P ¼ 0.004) in supine. Prone positioning further decreased nonaerated tissue (322 6 132 to 290 6 141 grams; P ¼ 0.028) and reduced tidal hyperinflation observed at PEEP 15 in supine patients (0.57 6 0.30 to 0.41 6 0.22%). Cyclic recruitment/derecruitment only decreased when high PEEP and prone positioning were applied together (4.1 6 1.9 to 2.9 6 0.9%; P ¼ 0.003), particularly in patients with high lung recruitability. Conclusions: Prone positioning enhances lung recruitment and decreases alveolar instability and hyperinflation observed at high PEEP in patients with ARDS. Ventilator-induced lung injury (VILI) seems to play an important role in patients with acute respiratory distress syndrome (Received in original form July 23, 2012; accepted in final form January 12, 2013) This work was supported by FONDECYT grant 11070156, Chile. Author Contributions: Substantial contributions to conception and design: R.A.C. Acquisition of data: R.A.C., J.C.D., C.M.R., R.A.G., C. A. Repetto, D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P. Analysis and interpretation of data: R.A.C., E.A.T., A.R.B., C.M.R., L.R.G., O.L. Drafting the article or revising it critically for important intellectual content: R.A.C., J.C.D., E.A.T., A.R.B., C.M.R., R.A.G., C. A. Repetto, C. A. Ramos, L.R.G., O.L., D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P. Final approval of the version to be published: R.A.C., J.C.D., E.A.T., A.R.B., C.M.R., R.A.G., C. A. Repetto, C. A. Ramos, L.R.G., O.L., D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P. Correspondence and requests for reprints should be addressed to Rodrigo Cornejo, M.D., Universidad de Chile, Santos Dumont 999, Independencia, Santiago, Chile. E-mail: [email protected] This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org Am J Respir Crit Care Med Vol 188, Iss. 4, pp 440–448, Aug 15, 2013 Copyright ª 2013 by the American Thoracic Society Originally Published in Press as DOI: 10.1164/rccm.201207-1279OC on January 24, 2013 Internet address: www.atsjournals.org AT A GLANCE COMMENTARY Scientific Knowledge on the Subject Experimental and clinical studies suggest that high levels of positive end-expiratory pressure (PEEP) and prone positioning may favor protective mechanical ventilation in patients with acute respiratory distress syndrome. High PEEP may induce lung recruitment and decrease cyclic recruitment/derecruitment; however, increasing PEEP may increase hyperinflation. Prone positioning could have synergistic effects with high PEEP by providing a more uniform recruitment and better distribution of lung stress. What This Study Adds to the Field In ventilated patients with acute respiratory distress syndrome, prone positioning enhances the effects of high PEEP in terms of lung recruitment and reduction of cyclic recruitment/derecruitment and prevents the negative impact of PEEP on tidal hyperinflation. (ARDS) (1). The mechanisms by which mechanical ventilation exerts its detrimental effect are not completely understood, but it appears that hyperinflation of lung units and shear forces generated during cyclic recruitment/derecruitment of unstable alveoli exacerbate, or even initiate, lung injury (1). High levels of positive end-expiratory pressure (PEEP) and prone positioning have been demonstrated to reduce VILI in experimental models of acute lung injury (ALI) (2–5). However, analysis of several large clinical trials in patients with ALI or ARDS suggests that these interventions may be effective only in patients with severe ARDS (6–9). Patients who seem to benefit from prone positioning are frequently subjected to higher levels of PEEP. Thus, there may be a potential interaction between the effects of both interventions on the mechanisms of VILI. Regarding VILI, PEEP may have a protective effect by favoring lung recruitment and by reducing cyclic recruitment/ derecruitment (1, 6), but other mechanisms (e.g., redistribution of extravascular lung water, redistribution of pulmonary blood flow to better aerated units, or preservation of surfactant activity) may be involved. However, PEEP may induce hyperinflation and increase the risk of VILI, especially in patients with low recruitability or lobar ARDS (10–12). In fact, some patients exhibit tidal hyperinflation despite using low VT and moderate PEEP levels according to the ARDS-Net strategy (13, 14). Prone positioning may influence mechanisms of VILI. By recruiting nonaerated tissue and by reducing the vertical pleural pressure gradient, prone positioning may provide a more uniform distribution of transpulmonary pressures during mechanical Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation ventilation (15–21). Therefore, prone positioning may act synergistically with high PEEP to protect the lungs from VILI by enhancing lung recruitment and decreasing the risk of PEEP-induced hyperinflation. The objectives of this study were to determine the combined effects of high PEEP and prone positioning on lung recruitment, cyclic recruitment/derecruitment, and tidal hyperinflation, as assessed by static and dynamic computed tomography (CT) in patients with ARDS. In addition, we sought to determine whether these responses are influenced by lung recruitability (22). Some of the results of this study have been previously reported in the form of abstracts (23, 24). METHODS Study Population Twenty-four patients were studied in a university hospital. Institutional and governmental ethical committees granted their approval. Informed consent was obtained from the patients’ next of kin. Adult patients fulfilling ARDS criteria (25) and on mechanical ventilation for 24 to 72 hours who required lung CT scan for clinical purposes were enrolled. Patients younger than 18 years of age, who were pregnant, or for whom prone positioning was contraindicated (26) were excluded. Study Protocol Patients were evaluated in the ICU and CT room (Figure 1). During the protocol they were kept under deep sedation and neuromuscular paralysis and ventilated in volume-controlled mode with VT of 6 ml/kg ideal body weight. ICU Assessments Patients were ventilated with PEEP 5 and 15 cm H2O, for 20 minutes each, in supine and prone positions (PEEP changes and positioning were applied in random order). Respiratory mechanics, oxygenation, and hemodynamic parameters were assessed at the end of each setting. 441 Before PEEP changes, a recruitment maneuver at 45 cm H2O airway pressure was performed to standardize volume history. Quasistatic compliance of the respiratory system (“compliance”) was calculated by dividing VT by the difference between plateau pressure and total PEEP. CT Assessments Patients underwent whole-lung CT during breath-holding sessions at three different airway pressures: 45 cm H2O end-inspiratory airway pressure and at 5 and 15 cm H2O PEEP. CT scanning (Somaton Sensation, Siemens, Germany) was performed under the following protocol: voltage 120 kVp, current 200 mA, mAs 100, rotation time 0.5 seconds, matrix 512 3 512. Lung compartments were defined according to their CT density in hyperinflated (2901 to 21,000 Hounsfield units [HU]), well aerated (2501 to 2900 HU), poorly aerated (2101 to –500 HU), and nonaerated tissue (2100 to 1100 HU) (22). Lung weight for each compartment was calculated as 1 2 (mean CT number/21,000) 3 volume, where CT number represents lung density (HU). Percentage of potentially recruitable lung was defined as (nonaerated tissue at 5 cm H2O 2 nonaerated tissue at 45 cm H2O)/total weight and high lung recruitability as a percentage of potentially recruitable lung greater than 13.9% in supine, which corresponds to the median value observed in 49 patients with ARDS in a previous study of lung recruitability (22). Contiguous axial sections (5 mm thick) were reconstructed from the volumetric data using a high-definition filter. A 2.4-mm-thick supradiaphragmatic CT slice was selected for dynamic CT. Respiratory rate was transiently decreased to 10 breaths per minutes during dynamic CT capture. Cine-CTs of 12 seconds were performed under the following protocol: voltage 100 kVp, current 80 mA, mAs 40; rotation time 0.5 seconds, 24 images; matrix 512 3 512. Lung compartments were expressed as percentage of tissue weight in the transverse slice. Cyclic recruitment/derecruitment and tidal hyperinflation were determined as tidal changes in percentage of nonaerated and hyperinflated tissue, respectively. CT images were repeated in supine and prone positioning, and the sequence of positions and PEEP levels was applied in random order. Images were analyzed manually by radiologists using Pulmo (Siemens, Germany) and MALUNA (University of Gottingen, Germany) software. Further details are provided in the online supplement. Figure 1. Study protocol. The protocol was performed in the ICU and CT rooms. In the ICU, patients were ventilated with positive end-expiratory pressure (PEEP) 5 and then 15 cm H2O for 20 minutes each, starting with a recruitment maneuver at 45 cm H2O before each period. Respiratory mechanics, oxygenation, and hemodynamic parameters were assessed at the end of each setting. Thereafter, in the CT room, patients underwent whole-lung CT during breath-holding sessions at consecutive airway pressures of 5, 45, and 15 cm H2O; afterward, Cine-CTs were performed on a fixed thoracic transverse slice at PEEP 5 and 15 cm H2O. CT images were repeated in supine and prone positioning, and the sequence of positions and PEEP levels was applied in random order. 442 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 188 2013 TABLE 1. BASELINE CHARACTERISTICS OF THE STUDY POPULATION Patient no. Age (yr) Gender APACHE SOFA 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Mean 6 SD 25 67 47 78 66 80 64 83 30 66 41 63 39 36 52 45 53 74 17 47 65 64 51 37 53 6 18 F F F F M M F M M M F M M M F M M M F M M M F M 18 14 22 26 29 27 12 20 31 23 18 21 16 6 18 10 17 22 22 19 19 24 19 15 20 6 6 5 7 10 12 17 13 9 8 15 8 8 8 8 5 9 9 12 11 12 10 10 11 8 8 10 6 3 Lung Infiltrates* Diffuse Patchy Diffuse Diffuse Patchy Lobar Lobar Patchy Diffuse Diffuse Patchy Patchy Patchy Patchy Diffuse Patchy Patchy Patchy Patchy Lobar Lobar Patchy Patchy Patchy MV (h) 62 28 32 64 68 66 67 26 38 40 69 26 33 29 52 36 34 31 24 38 44 67 69 33 45 6 17 Prestudy position Etiology ARDS PaO2/FIO2 (mm Hg) PEEP (cm H2O) Compliance† (ml/cm H2O) SP SP SP SP SP SP SP SP PP PP SP SP PP PP PP PP PP SP SP SP SP PP SP SP SS1Asp. P SS SS HS SS BP SS BP Asp. P SS SS BP H1N1 H1N1 H1N1 BP H1N11BP BP SS AP AP H1N1 SS H1N1 195 178 154 88 190 170 186 175 101 155 177 183 100 97 78 198 91 185 195 193 105 112 183 86 149 6 44 12 8 10 12 8 8 6 12 14 12 8 8 12 14 13 6 12 7 8 10 13 12 8 11 10 6 3 30 37 39 44 40 43 28 28 32 32 32 35 43 43 29 30 45 38 35 35 29 25 32 43 35 6 6 Definition of abbreviations: AP ¼ acute pancreatitis; APACHE ¼ Acute Physiology and Chronic Health Evaluation II score; ARDS ¼ acute respiratory distress syndrome; Asp. P ¼ aspirative pneumonia; BP ¼ bacterial pneumonia; HS ¼ hemorrhagic shock; MV ¼ mechanical ventilation; PEEP ¼ end-expiratory pressure; PP ¼ prone position; SOFA ¼ Sequential Organ Failure Assessment score; SP ¼ supine position; SS ¼ severe sepsis. * Type of lung infiltrates according to computed tomography. y Quasistatic compliance of respiratory system. Statistical Analysis Results are expressed as mean (6 SD) or median (interquartile range). The effect of PEEP level and position was assessed by applying a twoway ANOVA for repeated measurements with Sidak’s post hoc analysis. Comparisons between patients with low and high lung recruitability were performed with Student’s t test or Wilcoxon Mann-Whitney test. SPSS 20.0 software (Chicago, IL) was used for statistical calculations. Twosided P , 0.05 was considered statistically significant. RESULTS We included 24 patients (15 male; 54 6 18 yr of age) with ARDS. Fourteen patients presented with patchy, six with diffuse, and four with lobar CT attenuations. Baseline characteristics are presented in Table 1. Prone positioning had no significant effect on gas exchange or compliance when compared with supine position at PEEP 5 (Table 2). However, it induced a decrease in nonaerated tissue and an increase in well aerated tissue (Table 3). Increasing PEEP from 5 to 15 cm H2O increased oxygenation and compliance (Table 2), decreased nonaerated tissue, and increased well aerated tissue (Table 3). However, it markedly increased hyperinflation. These effects were observed at supine and prone positioning. Nevertheless, compared with supine, at prone positioning oxygenation and compliance tended to be higher at PEEP 15, whereas nonaerated tissue and hyperinflation were lower (Table 3; Figure 2). The effects of prone positioning and PEEP level on determinants of VILI, namely cyclic recruitment and derecruitment and tidal hyperinflation, are shown in Figure 3. Compared with supine position at PEEP 5, neither prone positioning nor increasing PEEP had a significant effect on cyclic recruitment and derecruitment. However, both strategies applied together (prone positioning at PEEP 15) significantly decreased cyclic recruitment and derecruitment (4.1 6 1.9 to 2.9 6 0.9%; P ¼ 0.003). Tidal hyperinflation was not affected by prone positioning at PEEP 5, but it increased in response to PEEP 15 in supine positioning (0.41 6 0.26 to 0.57 6 0.30%; P ¼ 0.004). The PEEP-induced tidal hyperinflation observed in supine positioning was markedly decreased by prone positioning (0.57 6 0.30 to 0.41 6 0.22%; P ¼ 0.01). Potentially recruitable lung was 18.3 6 11% in supine positioning and 14.9 6 8% in prone positioning (P ¼ 0.036). Figure 4 shows CT images from representative patients obtained at airway pressures of 5 and 45 cm H2O in supine and prone positioning. According to our predefined threshold, 14 patients had high lung recruitability. Increasing PEEP to 15 cm H2O improved oxygenation and compliance only in patients with high lung recruitability. This subgroup exhibited a 48% relative decrease in nonaerated TABLE 2. EFFECTS OF POSITIVE END-EXPIRATORY PRESSURE LEVELS AND PRONE POSITION ON RESPIRATORY VARIABLES Supine PEEP 5 PaO2:FIO2, mm Hg PaO2, mm Hg PaCO2, mm Hg Compliance,‡ ml/cm H2O Plateau pressure, cm H2O Driving pressure,jj cm H2O 143 90 47 34 18 13 6 6 6 6 6 6 58 34 7 6 3 3 Prone PEEP 15 235 155 46 40 27 12 6 6 6 6 6 6 68* 73* 8 9* 3* 3 PEEP 5 170 114 46 35 18 13 6 6 6 6 6 6 71 68 6 8 2 2 PEEP 15 259 175 46 44 26 11 6 6 6 6 6 6 63† 79† 7 10†x 2†x 2†x Definition of abbreviations: PEEP 5 ¼ positive end-expiratory pressure at 5 cm H2O; PEEP 15 ¼ positive end-expiratory pressure at 15 cm H2O. * P , 0.05 between parameters at supine PEEP 5 and 15 cm H2O. y P , 0.05 between parameters at prone PEEP 5 and 15 cm H2O. z Compliance of the respiratory system. x P , 0.05, between parameters at supine PEEP 15 and prone PEEP 15 cm H2O. jj Difference between plateau pressure and positive end-expiratory pressure. Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation 443 TABLE 3. EFFECTS OF POSITIVE END-EXPIRATORY PRESSURE LEVELS AND PRONE POSITION ON LUNG COMPARTMENT DISTRIBUTION Supine PEEP 5 Total lung volume, ml Total lung weight, g Volume of nonaerated compartment, ml Weight of nonaerated tissue, g Volume of poorly aerated compartment, ml Weight of poorly aerated tissue, g Volume of well aerated compartment, ml Weight of well aerated tissue, g Volume of hyperinflated compartment, ml Weight of hyperinflated tissue, g 2,140 1,201 503 501 623 446 872 254 144 5.6 6 6 6 6 6 6 6 6 6 6 643 307 201 201 210 149 406 103 160 6 Prone PEEP 15 3,271 1,282 324 322 753 517 1,767 485 434 19 6 6 6 6 6 6 6 6 6 6 900 316 133* 132* 314 216 661* 148* 318* 15* PEEP 5 2,240 1,216 431 431 657 467 1,027 324 125 4.4 6 6 6 6 6 6 6 6 6 6 614 341 193† 192† 207 157 456† 148† 105 4 PEEP 15 3,268 1,261 291 290 724 497 1,893 535 362 14 6 6 6 6 6 6 6 6 6 6 821 321 143‡x 141‡x 289 203 617‡ 152‡ 220‡x 8‡ Definition of abbreviations: PEEP 5 ¼ positive end-expiratory pressure at 5 cm H2O; PEEP 15 ¼ positive end-expiratory pressure at 15 cm H2O. * P , 0.05 between parameters at supine PEEP 5 and 15 cm H2O. y P , 0.05, between parameters at supine PEEP 5 and prone PEEP 5 cm H2O. z P , 0.05 between parameters at prone PEEP 5 and 15 cm H2O. x P , 0.05 between parameters at supine PEEP 15 and prone PEEP 15 cm H2O. tissue when increasing PEEP; this effect was seen in only 22% of patients with low lung recruitability (Table 4). In contrast to patients with high lung recruitability who showed no additional effect of prone positioning on nonaerated tissue at PEEP 15, patients with low lung recruitability exhibited a significant decrease in nonaerated tissue when subjected to prone positioning (20% additional relative decrease). Cyclic recruitment/derecruitment was significantly lower at baseline in patients with low lung recruitability (2.8 6 1.3% vs. 5.1 6 1.8% in patients with high lung recruitability; P ¼ 0.002), and neither increasing PEEP nor prone positioning had a significant effect on this variable in the former subgroup (Figure 3). In contrast, cyclic recruitment and derecruitment decreased in patients with high lung recruitability when increasing PEEP to 15 cm H2O, and a significant additional effect was obtained after prone positioning at the same PEEP level. Regarding tidal hyperinflation, patients with low and high lung recruitability had similar values at baseline (0.39 6 0.25% vs. 0.45 6 0.24; P ¼ 0.568). The effect of high PEEP level on increasing tidal hyperinflation was prevented by prone positioning in both subgroups of lung recruitability. hyperinflated tissue or plateau pressures. This effect may be explained in part by the suppression of the compressive force of the heart on dorsal lung regions caused by prone positioning (31, 32). Patients with predominantly basal consolidations, such as patients A and B from Figure 4, were those who experienced higher recruitment induced by prone positioning (data not shown), which is in line with the findings of Galiatsu and colleagues (15). These observations may complement the concept by Rouby and colleagues about lung morphology as a predictor of the response to increasing airway pressures (33), although this requires confirmation by further studies. In patients with low potentially recruitable lung, prone positioning was able to recruit a significant amount of nonaerated tissue in addition to that already recruited by high PEEP (Table 4). Therefore, the assessment of potentially recruitable lung between 5 to 45 cm H2O may correctly predict lung recruitability to increasing airway pressures but not the response to prone. Thus, prone positioning may be considered in patients with severe ARDS even if they have poor response to recruitment maneuvers in supine position as assessed by CT. Effects of PEEP DISCUSSION The main finding of the present study is that prone positioning enhances the effects of high PEEP on lung recruitment and cyclic recruitment/derecruitment, whereas it prevents the effects of high PEEP on tidal hyperinflation. In addition, we found that lung recruitability, assessed by CT at high airway pressures (22), does not predict lung recruitment induced by prone positioning. Effects of Prone Positioning Although it has been reported that oxygenation improves after prone positioning (7–9), we only found a trend that did not reach statistical significance. This may be explained by the short sampling period. Data obtained from different series of prolonged prone positioning in patients with ARDS found that oxygenation improves several hours after prone positioning begins (26–30). Another possible explanation is that oxygenation variation and lung recruitment may be dissociated because “anatomical” lung recruitment may differ from “functional” recruitment (22). We observed that prone positioning was an effective recruitment strategy that, in contrast to high PEEP, did not increase Although increasing PEEP from 5 to 15 cm H2O decreased nonaerated tissue and increased oxygenation, it had no consistent effect on cyclic recruitment/derecruitment in the overall population. However, in the subgroup of patients with higher lung recruitability, cyclic recruitment/derecruitment significantly decreased when increasing PEEP to 15 cm H2O (Figure 3). Similar findings were reported in a recent study in which cyclic recruitment/derecruitment was assessed indirectly with static CT images (34). As expected, in supine positioning high PEEP resulted in higher plateau pressures, hyperinflated tissue, and tidal hyperinflation. Hyperinflation is a morphologic description of the lung areas that appear overfilled with gas in CT images, and it has been associated with VILI (14). The low values of tidal hyperinflation may seem irrelevant (,1%). However, fractional analysis of CT data was based on lung weight instead of volume, so the magnitude of real hyperinflated tissue may have been systematically underestimated (35, 36). We used weight for fractional analysis in dynamic CT because total volume is changing throughout the respiratory cycle. If the results were expressed as percentage of volume, tidal hyperinflation would have been 10 times greater (4–6%). PEEP-induced tidal hyperinflation 444 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 188 2013 Figure 2. Distribution of lung compartments expressed in weight and volume. Proportion of total lung weight (A) and volume (B) of each compartment classified as nonaerated tissue (NAT), poorly aerated tissue (PAT), well aerated tissue (WAT), and hyperinflated tissue (HIT), at end expiration with positive endexpiratory pressure (PEEP) 5 and 15 cm H2O, assessed in supine (supine 5 and supine 15) and prone positioning (prone 5 and prone 15). Data are presented for the overall population (left; n ¼ 24), for the subgroup of patients with high lung recruitability (center; n ¼ 14), and for the subgroup of patients with low lung recruitability (right; n ¼ 10). *P , 0.05 between parameters at supine 5 and supine 15 cm H2O. yP , 0.05 between parameters at prone 5 and prone 15 cm H2O. zP , 0.05 between parameters at supine 5 and prone 5 cm H2O. xP , 0.05 between parameters at supine 15 and prone 15 cm H2O. could be an important drawback of high PEEP strategies (14), which may explain the lack of consistency in the results of recent clinical trials (6). The fact that high PEEP levels could favor hyperinflation without a consistent decrease in cyclic recruitment/derecruitment supports the notion that PEEP should be titrated individually based not only on oxygen exchange but also on other variables, such as lung recruitability. Interaction of Prone Positioning with PEEP Few studies have addressed the issue of potential interactions between prone positioning and PEEP (37, 38). In the present study, we found that prone positioning has a synergistic effect with high PEEP in increasing respiratory system compliance, likely because of an increase in well aerated lung tissue. Previous studies have shown contradictory data about the effects of prone positioning on respiratory system compliance (39). These discrepancies may be due to differences in chest wall compliance and lung recruitability of the study groups. Prone positioning and high PEEP induced lung recruitment, resulting in the least amount of nonaerated tissue. Prone positioning also reinforced the effect of high PEEP on cyclic recruitment/ derecruitment, especially in patients with higher lung recruitability. This interaction did not seem to be a simple additive effect because in the whole population neither prone at low PEEP nor high PEEP on supine decreased cyclic recruitment/derecruitment. Thus, prone positioning and high PEEP together showed a synergistic effect on cyclic recruitment/derecruitment. Prone positioning prevented the increase in hyperinflated tissue and in tidal hyperinflation induced by high PEEP levels in supine positioning. Likewise, the increase in plateau pressures Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation 445 Figure 3. Effects of positive end-expiratory pressure (PEEP) and prone positioning on cyclic recruitment/derecruitment (R/D) and tidal hyperinflation (TH). Cyclic R/D and TH with PEEP 5 and 15 cm H2O, assessed in supine (supine 5 and supine 15) and prone positioning (prone 5 and prone 15). Data are presented for the overall population (n ¼ 24) (A), for the subgroup of patients with high lung recruitability (n ¼ 14) (B), and for the subgroup of patients with low lung recruitability (n ¼ 10) (C). *P , 0.05 between parameters at supine 5 and prone 15 cm H2O. yP , 0.05 between parameters at supine 5 and supine 15 cm H2O. zP , 0.05 between parameters at supine 15 and prone 15 cm H2O. induced by high PEEP was lower in prone positioning. These findings may be related to the effects of prone positioning in decreasing pleural pressure gradients and homogenizing transpulmonary pressures in the lungs of patients with ARDS (15, 16, 18, 20). A regional analysis of CT images, as performed by Grasso and colleagues (40), would be a valuable complement to the present study to assess the effects of prone positioning on inhomogeneity and recruitment of individual lung regions. The results of the present study suggest that a high PEEP strategy applied in prone positioning, instead of supine positioning, could have more beneficial and less adverse effects in terms of respiratory mechanics and determinants of VILI. These findings are consistent with the observations of a metaanalysis, which indicate that patients with severe forms of ARDS, who are usually ventilated with high levels of PEEP, may have a survival benefit when treated in prone positioning (7–9). Recruitment and Cyclic Recruitment/Derecruitment: Methodological Issues CT has been the gold standard to assess lung recruitment, although different definitions have been applied (10, 41). We chose the original definition of recruitment based on the decrease of nonaerated tissue expressed in lung weight as we have used in the past (22, 42) because it may be applied to analyze cyclic recruitment/derecruitment in dynamic CT. Other authors have defined recruitment as the reaeration of the nonaerated and poorly aerated compartment (10). Because we applied a definition limited to the nonaerated compartment, we acknowledge that our results for PEEP-induced recruitment may appear as subestimated compared with studies that include the poorly aerated compartment, as shown in a recent study using transthoracic ultrasound (43). The threshold of 13.9% used in the present study to classify patients as having high or low lung recruitability in supine position was predefined arbitrarily based on the median value of the subgroup of 49 patients with ARDS from Gattinoni’s study. The original lung recruitability threshold of Gattinoni’s study was 9%, which corresponded to the median value of the whole ALI/ARDS population (68 patients). By applying a different threshold, our subgroups of higher and lower lung recruitability are not comparable to the subgroups defined in the original study by Gattinoni (22). There is controversy regarding whether cyclic recruitment/ derecruitment, as assessed by CT, corresponds to intratidal opening and closing of lung units or to flooded alveoli that become partially inflated during inspiration (44, 45). Whatever the underlying phenomenon (cyclic mechanical deformation or cyclic recruitment/derecruitment of lung units), a reduction of instability produced by prone positioning at high PEEP, as supported by our study, appears as theoretically positive in terms of lung protection. Several approaches to assess cyclic phenomena have been used (46). A cine-CT analysis of a fixed transverse slice allows dynamic imaging without mechanical ventilation interruption. This method has been recently validated by experimental and clinical studies to determine cyclic recruitment/derecruitment, tidal hyperinflation, and dynamic lung strain (18, 35, 36, 47, 48). 446 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 188 2013 Figure 4. Representative chest CT images obtained during breath-holding sessions. Representative CT slices of the lungs obtained 2 cm above the diaphragm dome at airway pressures of 5 cm H2O (left) and 45 cm H2O (right) from four patients in supine (upper) and prone positioning (lower). The percentage of potentially recruitable lung was defined as the proportion of nonaerated tissue in which aeration is restored when increasing airway pressures from 5 to 45 cm H2O. In patient A, the percentage of potentially recruitable lung was 24% in supine and 18% in prone positioning; in patient B, the percentage of potentially recruitable lung was 24% in supine and 15% in prone; in patient C, the percentage of potentially recruitable lung was 14% in supine and 21% in prone; and in patient D, the percentage of potentially recruitable lung was 27% in supine and 24% in prone. The main limitations of this method are: 1) The more inhomogeneous the lung impairment, the less representative the slice may be (this handicap is particularly true in patients with lobar pattern, but only four of our patients had such pattern); 2) It is impossible for dynamic CT to scan exactly the same anatomical structure in different settings, although the careful definition of anatomical landmarks and fractional analysis used for dynamic CT may avoid artifacts created by the cranio-caudal motion; and 3) the absolute amount of grams of lung tissue undergoing recruitment/derecruitment or tidal hyperinflation cannot be determined. The alternative approach of using static CT images of the whole lung after end-expiratory and end-inspiratory breath holds (15, 16, 18–20, 22, 34) also has several limitations, the most important being the time dependency of the recruitment and derecruitment phenomena (49, 50). Despite the limitations of both methods, a previous study comparing them showed no major differences (48). In conclusion, prone positioning induces lung recruitment even in patients classified as having low potential for lung recruitment. In addition, prone positioning applied together with high TABLE 4. COMPARISONS BETWEEN PATIENTS WITH HIGH AND LOW LUNG RECRUITABILITY High Lung Recruitability PaO2/FIO2 at PEEP 5, mm Hg PaO2/FIO2 at PEEP 15, mm Hg PaCO2 at PEEP 5, mm Hg PaCO2 at PEEP 15, mm Hg Compliance at 5 cmH2O, ml/cm H2O Compliance at 15 cmH2O, ml/cm H2O Total lung weight at 5 cm H2O, ml NAT at 5 cm H2O, g NAT at 15 cm H2O, g PAT at 5 cm H2O, g PAT at 15 cm H2O, g WAT at 5 cm H2O, g WAT at 15 cm H2O, g HIT at 5 cm H2O, g HIT at 15 cm H2O, g Low Lung Recruitability Supine Prone Supine Prone 128 6 61 230 6 69* 49 6 7 47 6 8 34 6 6 40 6 7* 1,307 6 240 576 (410–797) 301 (230–400)* 454 (360–546) 486 (422–786) 208 (163–347) 468 (388–593)* 3.6 (2.1–5.4) 15 (7.3–18)* 156 6 62 272 6 70*† 47 6 7 46 6 8 33 6 8 45 6 11*† 1,312 6 292 448 (354–646)x 296 (231–329)*† 443 (411–561) 459 (410–681) 255 (191–434)x 534 (458–630)*†jj 2.7 (2.0–4.9) 12 (9.0–18)* 164 6 49 244 6 70 45 6 6 46 6 6 35 6 6 40 6 12 1,052 6 334‡ 372 (305–472)‡ 290 (250–334)* 387 (281–431) 378 (241–572) 240 (210–351) 461 (377–581)* 5.8 (4.7–6.9) 17 (14–26)* 191 6 80 241 6 49† 45 6 6 46 6 5 37 6 9 43 6 10† 1,083 6 376 325 (181–390)‡ 215 (163–330)*†jj 457 (275–561) 407 (263–577) 324 (239–343)x 483 (336–593)* 6.0 (4.8–7.9)‡ 15 (10–24)* Definition of abbreviations: HIT ¼ hyperinflated tissue; NAT ¼ nonaerated tissue; PAT ¼ poorly aerated tissue; PEEP 5 ¼ positive end-expiratory pressure 5 cm H2O; PEEP 15 ¼ positive end-expiratory pressure 15 cm H2O; WAT ¼ well aerated tissue. * P , 0.05 between parameters at 5 cm H2O (or PEEP 5) and 15 cm H2O (or PEEP 15). y P , 0.05 between parameters at 5 cm H2O (or PEEP 5) in supine positioning and 15 cm H2O (or PEEP 15) in prone positioning. z P , 0.05 comparing patients with high lung recruitability versus low lung recruitability at the same level of PEEP and position. x P , 0.05 comparing parameters at 5 cm H2O (or PEEP 5) between supine and prone positioning. jj P , 0.05 comparing parameters at 15 cm H2O (or PEEP 15) between supine and prone positioning. Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation PEEP levels in patients with ARDS act synergistically to decrease mechanical determinants of VILI such as cyclic recruitment/derecruitment and tidal hyperinflation. Author disclosures are available with the text of this article at www.atsjournals.org. Acknowledgment: The authors thank Dr. Jerónimo Graf, Dr. Gastón Murias, and Dr. Guillermo Bugedo for critical comments and suggestions for the manuscript; the nurses, respiratory therapists, medical staff, and medical technologists from Hospital Clı́nico Universidad de Chile for support during the execution of the studies; Dr. Hector Gatica for the statistical advice in the present article; Daniel Castro, medical biophysics and radiation protection officer, who assessed and adjusted the level of radiation per CT, allowing that the total dose per patient did not exceed the equivalent of one coronary CT angiography study. References 1. Dos Santos CC, Slutsky AS. Invited review: mechanisms of ventilatorinduced lung injury. A perspective. J Appl Physiol 2000;89:1645–1655. 2. Argiras EP, Blakeley CR, Dunnill MS, Otremski S, Sykes MK. High peep decreases hyaline membrane formation in surfactant deficient lungs. Br J Anaesth 1987;59:1278–1285. 3. Broccard A, Shapiro RS, Schmitz LL, Adams AB, Nahum A, Marini JJ. Prone positioning attenuates and redistributes ventilator-induced lung injury in dogs. Crit Care Med 2000;28:295–303. 4. Dreyfuss D, Soler P, Basset G, Saumon G. High inflation pressure pulmonary edema: respective effects of high airway pressure, high tidal volume, and positive end-expiratory pressure. Am Rev Respir Dis 1988;137:1159–1164. 5. Webb HH, Tierney DF. Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures: protection by positive end-expiratory pressure. Am Rev Respir Dis 1974;110:556–565. 6. Briel M, Meade M, Mercat A, Brower RG, Talmor D, Walter SD, Slutsky AS, Pullenayegum E, Zhou Q, Cook D, et al. Higher vs lower positive end-expiratory pressure in patients with acute lung injury and acute respiratory distress syndrome: systematic review and meta-analysis. JAMA 2010;303:865–873. 7. Sud S, Friedrich JO, Taccone P, Polli F, Adhikari NK, Latini R, Pesenti A, Guerin C, Mancebo J, Curley MA, et al. Prone ventilation reduces mortality in patients with acute respiratory failure and severe hypoxemia: systematic review and meta-analysis. Intensive Care Med 2010;36:585–599. 8. Gattinoni L, Carlesso E, Taccone P, Polli F, Guerin C, Mancebo J. Prone positioning improves survival in severe ARDS: a pathophysiologic review and individual patient meta-analysis. Minerva Anestesiol 2010;76:448–454. 9. Abroug F, Ouanes-Besbes L, Dachraoui F, Ouanes I, Brochard L. An updated study-level meta-analysis of randomised controlled trials on proning in ARDS and acute lung injury. Crit Care 2011;15:R6. 10. Malbouisson LM, Muller JC, Constantin JM, Lu Q, Puybasset L, Rouby JJ. Computed tomography assessment of positive end-expiratory pressure-induced alveolar recruitment in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 2001;163:1444– 1450. 11. Nieszkowska A, Lu Q, Vieira S, Elman M, Fetita C, Rouby JJ. Incidence and regional distribution of lung overinflation during mechanical ventilation with positive end-expiratory pressure. Crit Care Med 2004; 32:1496–1503. 12. Rouby JJ. Lung overinflation: the hidden face of alveolar recruitment. Anesthesiology 2003;99:2–4. 13. Grasso S, Terragni P, Mascia L, Fanelli V, Quintel M, Herrmann P, Hedenstierna G, Slutsky AS, Ranieri VM. Airway pressure-time curve profile (stress index) detects tidal recruitment/hyperinflation in experimental acute lung injury. Crit Care Med 2004;32:1018–1027. 14. Terragni PP, Rosboch G, Tealdi A, Corno E, Menaldo E, Davini O, Gandini G, Herrmann P, Mascia L, Quintel M, et al. Tidal hyperinflation during low tidal volume ventilation in acute respiratory distress syndrome. Am J Respir Crit Care Med 2007;175:160–166. 15. Galiatsou E, Kostanti E, Svarna E, Kitsakos A, Koulouras V, Efremidis SC, Nakos G. Prone position augments recruitment and prevents alveolar overinflation in acute lung injury. Am J Respir Crit Care Med 2006;174:187–197. 447 16. Mentzelopoulos SD, Roussos C, Zakynthinos SG. Prone position reduces lung stress and strain in severe acute respiratory distress syndrome. Eur Respir J 2005;25:534–544. 17. Mure M, Domino KB, Lindahl SG, Hlastala MP, Altemeier WA, Glenny RW. Regional ventilation-perfusion distribution is more uniform in the prone position. J Appl Physiol 2000;88:1076–1083. 18. Perchiazzi G, Rylander C, Vena A, Derosa S, Polieri D, Fiore T, Giuliani R, Hedenstierna G. Lung regional stress and strain as a function of posture and ventilatory mode. J Appl Physiol 2011;110:1374–1383. 19. Richter T, Bellani G, Scott Harris R, Vidal Melo MF, Winkler T, Venegas JG, Musch G. Effect of prone position on regional shunt, aeration, and perfusion in experimental acute lung injury. Am J Respir Crit Care Med 2005;172:480–487. 20. Valenza F, Guglielmi M, Maffioletti M, Tedesco C, Maccagni P, Fossali T, Aletti G, Porro GA, Irace M, Carlesso E, et al. Prone position delays the progression of ventilator-induced lung injury in rats: does lung strain distribution play a role? Crit Care Med 2005;33:361–367. 21. Wiener CM, Kirk W, Albert RK. Prone position reverses gravitational distribution of perfusion in dog lungs with oleic acid-induced injury. J Appl Physiol 1990;68:1386–1392. 22. Gattinoni L, Caironi P, Cressoni M, Chiumello D, Ranieri VM, Quintel M, Russo S, Patroniti N, Cornejo R, Bugedo G. Lung recruitment in patients with the acute respiratory distress syndrome. N Engl J Med 2006;354:1775–1786. 23. Cornejo R, Diaz J, Repetto C, Suarez P, Bozan F, Arellano D, Rouliez K, Diaz G, Romero C, Mezzano E, et al. Comparison of potentially recruitable lung (PRL) in supine versus prone position: preliminary data [abstract]. Intensive Care Med 2009;35:S126. 24. Cornejo R, Diaz J, Repetto C, Suarez P, Bozan F, Neira W, Zamorano A, Tobar E, Castro D, Monsalve J, et al. Effect of prone position and peep on cyclic recruitment-derecruitment (r/d), and tidal hyperinflation (th) using cine-ct: preliminary data [abstract]. Intensive Care Med 2009;35:S103. 25. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M, Legall JR, Morris A, Spragg R. The American-European consensus conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818–824. 26. Romero CM, Cornejo RA, Galvez LR, Llanos OP, Tobar EA, Berasain MA, Arellano DH, Larrondo JF, Castro JS. Extended prone position ventilation in severe acute respiratory distress syndrome: a pilot feasibility study. J Crit Care 2009;24:81–88. 27. Chan MC, Hsu JY, Liu HH, Lee YL, Pong SC, Chang LY, Kuo BI, Wu CL. Effects of prone position on inflammatory markers in patients with ARDS due to community-acquired pneumonia. J Formos Med Assoc 2007;106:708–716. 28. Charron C, Bouferrache K, Caille V, Castro S, Aegerter P, Page B, Jardin F, Vieillard-Baron A. Routine prone positioning in patients with severe ARDS: feasibility and impact on prognosis. Intensive Care Med 2011;37:785–790. 29. Cornejo R, Tobar E, Diaz G, Romero C, Llanos O, Galvez LR, Zamorano A, Fabrega L, Neira W, Arellano D, et al. Systematic approach for severe respiratory failure due to novel a (H1N1) influenza. Minerva Anestesiol 2011;77:510–521. 30. Mancebo J, Fernandez R, Blanch L, Rialp G, Gordo F, Ferrer M, Rodriguez F, Garro P, Ricart P, Vallverdu I, et al. A multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care Med 2006;173:1233–1239. 31. Malbouisson LM, Busch CJ, Puybasset L, Lu Q, Cluzel P, Rouby JJ; CT scan ARDS study group. Role of the heart in the loss of aeration characterizing lower lobes in acute respiratory distress syndrome. Am J Respir Crit Care Med 2000;161:2005–2012. 32. Albert RK, Hubmayr RD. The prone position eliminates compression of the lungs by the heart. Am J Respir Crit Care Med 2000;161:1660–1665. 33. Puybasset L, Cluzel P, Gusman P, Grenier P, Preteux F, Rouby JJ; CT scan ARDS study group. Regional distribution of gas and tissue in acute respiratory distress syndrome: I. Consequences for lung morphology. Intensive Care Med 2000;26:857–869. 34. Caironi P, Cressoni M, Chiumello D, Ranieri M, Quintel M, Russo SG, Cornejo R, Bugedo G, Carlesso E, Russo R, et al. Lung opening and closing during ventilation of acute respiratory distress syndrome. Am J Respir Crit Care Med 2010;181:578–586. 448 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE 35. Bruhn A, Bugedo D, Riquelme F, Varas J, Retamal J, Besa C, Cabrera C, Bugedo G. Tidal volume is a major determinant of cyclic recruitmentderecruitment in acute respiratory distress syndrome. Minerva Anestesiol 2011;77:418–426. 36. David M, Karmrodt J, Bletz C, David S, Herweling A, Kauczor HU, Markstaller K. Analysis of atelectasis, ventilated, and hyperinflated lung during mechanical ventilation by dynamic CT. Chest 2005;128: 3757–3770. 37. Gainnier M, Michelet P, Thirion X, Arnal JM, Sainty JM, Papazian L. Prone position and positive end-expiratory pressure in acute respiratory distress syndrome. Crit Care Med 2003;31:2719–2726. 38. Walther SM, Johansson MJ, Flatebo T, Nicolaysen A, Nicolaysen G. Marked differences between prone and supine sheep in effect of peep on perfusion distribution in zone II lung. J Appl Physiol 2005; 99:909–914. 39. Pelosi P, Brazzi L, Gattinoni L. Prone position in acute respiratory distress syndrome. Eur Respir J 2002;20:1017–1028. 40. Grasso S, Stripoli T, Sacchi M, Trerotoli P, Staffieri F, Franchini D, De Monte V, Valentini V, Pugliese P, Crovace A, et al. Inhomogeneity of lung parenchyma during the open lung strategy: a computed tomography scan study. Am J Respir Crit Care Med 2009;180:415– 423. 41. Gattinoni L, Pelosi P, Crotti S, Valenza F. Effects of positive endexpiratory pressure on regional distribution of tidal volume and recruitment in adult respiratory distress syndrome. Am J Respir Crit Care Med 1995;151:1807–1814. 42. Bugedo G, Bruhn A, Hernandez G, Rojas G, Varela C, Tapia JC, Castillo L. Lung computed tomography during a lung recruitment 43. 44. 45. 46. 47. 48. 49. 50. VOL 188 2013 maneuver in patients with acute lung injury. Intensive Care Med 2003;29:218–225. Bouhemad B, Brisson H, Le-Guen M, Arbelot C, Lu Q, Rouby JJ. Bedside ultrasound assessment of positive end-expiratory pressureinduced lung recruitment. Am J Respir Crit Care Med 2011;183:341– 347. Hubmayr RD. Perspective on lung injury and recruitment: a skeptical look at the opening and collapse story. Am J Respir Crit Care Med 2002;165:1647–1653. Tschumperlin DJ, Oswari J, Margulies AS. Deformation-induced injury of alveolar epithelial cells: effect of frequency, duration, and amplitude. Am J Respir Crit Care Med 2000;162:357–362. Ghadiali S, Huang Y. Role of airway recruitment and derecruitment in lung injury. Crit Rev Biomed Eng 2011;39:297–317. Doebrich M, Markstaller K, Karmrodt J, Kauczor HU, Eberle B, Weiler N, Thelen M, Schreiber WG. Analysis of discrete and continuous distributions of ventilatory time constants from dynamic computed tomography. Phys Med Biol 2005;50:1659–1673. Zinserling J, Wrigge H, Neumann P, Muders T, Magnusson A, Hedenstierna G, Putensen C. Methodologic aspects of attenuation distributions from static and dynamic thoracic ct techniques in experimental acute lung injury. Chest 2005;128:2963–2970. Albert SP, DiRocco J, Allen GB, Bates JH, Lafollette R, Kubiak BD, Fischer J, Maroney S, Nieman GF. The role of time and pressure on alveolar recruitment. J Appl Physiol 2009;106:757–765. Massa CB, Allen GB, Bates JH. Modeling the dynamics of recruitment and derecruitment in mice with acute lung injury. J Appl Physiol 2008; 105:1813–1821. new england journal of medicine The established in 1812 june 6, 2013 vol. 368 no. 23 Prone Positioning in Severe Acute Respiratory Distress Syndrome Claude Guérin, M.D., Ph.D., Jean Reignier, M.D., Ph.D., Jean-Christophe Richard, M.D., Ph.D., Pascal Beuret, M.D., Arnaud Gacouin, M.D., Thierry Boulain, M.D., Emmanuelle Mercier, M.D., Michel Badet, M.D., Alain Mercat, M.D., Ph.D., Olivier Baudin, M.D., Marc Clavel, M.D., Delphine Chatellier, M.D., Samir Jaber, M.D., Ph.D., Sylvène Rosselli, M.D., Jordi Mancebo, M.D., Ph.D., Michel Sirodot, M.D., Gilles Hilbert, M.D., Ph.D., Christian Bengler, M.D., Jack Richecoeur, M.D., Marc Gainnier, M.D., Ph.D., Frédérique Bayle, M.D., Gael Bourdin, M.D., Véronique Leray, M.D., Raphaele Girard, M.D., Loredana Baboi, Ph.D., and Louis Ayzac, M.D., for the PROSEVA Study Group* A bs t r ac t Background Previous trials involving patients with the acute respiratory distress syndrome (ARDS) have failed to show a beneficial effect of prone positioning during mechanical ventilatory support on outcomes. We evaluated the effect of early application of prone positioning on outcomes in patients with severe ARDS. Methods In this multicenter, prospective, randomized, controlled trial, we randomly assigned 466 patients with severe ARDS to undergo prone-positioning sessions of at least 16 hours or to be left in the supine position. Severe ARDS was defined as a ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (Fio2) of less than 150 mm Hg, with an Fio2 of at least 0.6, a positive end-expiratory pressure of at least 5 cm of water, and a tidal volume close to 6 ml per kilogram of predicted body weight. The primary outcome was the proportion of patients who died from any cause within 28 days after inclusion. The authors’ affiliations are listed in the Appendix. Address reprint requests to Dr. Guérin at Service de Réanimation Médicale, Hôpital de la Croix-Rousse, 103 Grande Rue de la Croix-Rousse, 69004 Lyon, France, or at [email protected]. *The Proning Severe ARDS Patients (PROSEVA) study investigators are listed in the Supplementary Appendix, available at NEJM.org. This article was published on May 20, 2013, at NEJM.org. N Engl J Med 2013;368:2159-68. DOI: 10.1056/NEJMoa1214103 Copyright © 2013 Massachusetts Medical Society. Results A total of 237 patients were assigned to the prone group, and 229 patients were assigned to the supine group. The 28-day mortality was 16.0% in the prone group and 32.8% in the supine group (P<0.001). The hazard ratio for death with prone positioning was 0.39 (95% confidence interval [CI], 0.25 to 0.63). Unadjusted 90-day mortality was 23.6% in the prone group versus 41.0% in the supine group (P<0.001), with a hazard ratio of 0.44 (95% CI, 0.29 to 0.67). The incidence of complications did not differ significantly between the groups, except for the incidence of cardiac arrests, which was higher in the supine group. Conclusions In patients with severe ARDS, early application of prolonged prone-positioning sessions significantly decreased 28-day and 90-day mortality. (Funded by the Programme Hospitalier de Recherche Clinique National 2006 and 2010 of the French Ministry of Health; PROSEVA ClinicalTrials.gov number, NCT00527813.) n engl j med 368;23 nejm.org june 6, 2013 2159 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l P rone positioning has been used for many years to improve oxygenation in patients who require mechanical ventilatory support for management of the acute respiratory distress syndrome (ARDS). Randomized, controlled trials have confirmed that oxygenation is significantly better when patients are in the prone position than when they are in the supine position.1,2 Furthermore, several lines of evidence have shown that prone positioning could prevent ventilator-induced lung injury.3-6 In several previous trials, these physiological benefits did not translate into better patient outcomes, since no significant improvement was observed in patient survival with prone positioning.7-10 However, meta-analyses2,11 have suggested that survival is significantly improved with prone positioning as compared with supine positioning among patients with severely hypoxemic ARDS at the time of randomization. We conducted a prospective, multicenter, randomized, controlled trial to explore whether early application of prone positioning would improve survival among patients with ARDS who, at the time of enrollment, were receiving mechanical ventilation with a positive end-expiratory pressure (PEEP) of at least 5 cm of water and in whom the ratio of the partial pressure of arterial oxygen (Pao2) to the fraction of inspired oxygen (Fio2) was less than 150 mm Hg. Me thods Patients of m e dic i n e Randomization was computer-generated and stratified according to ICU. Patients were randomly assigned to the prone group or supine group with the use of a centralized Web-based management system (Clininfo). The protocol, available at NEJM.org, was approved by the ethics committee Comité Consultatif de Protection des Personnes dans la Recherche Biomedicale Sud-Est IV in Lyon, France, and by the Clinical Investigation Ethics Committee at Hospital de Sant Pau in Barcelona. Written informed consent was obtained after the patients’ next of kin read the informational leaflet. If patients were able to read the leaflet at some point after inclusion in the study, they were approached to confirm participation in the trial. An investigator at each center was responsible for enrolling patients in the study, following the protocol, and completing the case-report form. Centers were regularly monitored by research fellows. Data collectors were aware of the study-group assignments, but outcomes assessors were not. The trial was overseen by a steering committee that met monthly. An independent data and safety monitoring board, comprising three experts in the field, was also set up (a list of board members is provided in the Supplementary Appendix). There was no commercial support. No one who is not listed as an author contributed to the writing of this manuscript. All authors vouch for the accuracy of the data and analysis and the fidelity of the study to the protocol. Protocol We included in the study adults who met the following criteria: ARDS, as defined according to the American–European Consensus Conference criteria12; endotracheal intubation and mechanical ventilation for ARDS for less than 36 hours; and severe ARDS (defined as a Pao2:Fio2 ratio of <150 mm Hg, with an Fio2 of ≥0.6, a PEEP of ≥5 cm of water, and a tidal volume of about 6 ml per kilogram of predicted body weight; the criteria were confirmed after 12 to 24 hours of mechanical ventilation in the participating intensive care unit [ICU]). Exclusion criteria are listed in the Supplementary Appendix, available with the full text of this article at NEJM.org. After a patient was determined to be eligible, a stabilization period of 12 to 24 hours was mandated. Inclusion in the study was confirmed only at the end of this period (Fig. S1 in the Supplementary Appendix). Patients assigned to the prone group had to be turned to the prone position within the first hour after randomization. They were placed in a completely prone position for at least 16 consecutive hours. Participating centers were given guidelines (see the Supplementary Appendix) to ensure standardization of prone placement. Standard ICU beds were used for all patients. Patients assigned to the supine group remained in a semirecumbent position. Trial Design Mechanical ventilation13 was delivered in a Patients were recruited from 26 ICUs in France volume-controlled mode with constant inspiraand 1 in Spain, all of which have used prone po- tory flow, with tidal volume targeted at 6 ml per sitioning in daily practice for more than 5 years. kilogram of predicted body weight13 and the PEEP 2160 n engl j med 368;23 nejm.org june 6, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Prone Positioning in Severe ARDS level selected from a PEEP–Fio2 table14 (Table S1 in the Supplementary Appendix). The goal was to maintain an end-inspiratory plateau pressure of the respiratory system (PplatRS), measured after a 1-second period of no air flow, of no more than 30 cm of water and an arterial plasma pH of 7.20 to 7.45. Physiological variables were measured at predetermined times in both groups. In the supine group, measurements were performed every 6 hours; in the prone group, measurements were performed just before the patient was turned to the prone position, after 1 hour of prone positioning, just before the patient was turned back to the supine position, and 4 hours after the patient was returned to the supine position. Adjustments of ventilator settings in specific situations are detailed in the Supplementary Appendix. The criteria for stopping prone treatment were any of the following: improvement in oxygenation (defined as a Pao2:Fio2 ratio of ≥150 mm Hg, with a PEEP of ≤10 cm of water and an Fio2 of ≤0.6; in the prone group, these criteria had to be met in the supine position at least 4 hours after the end of the last prone session); a decrease in the Pao2:Fio2 ratio of more than 20%, relative to the ratio in the supine position, before two consecutive prone sessions; or complications occurring during a prone session and leading to its immediate interruption. Complications leading to the immediate interruption of prone treatment included nonscheduled extubation, main-stem bronchus intubation, endotracheal-tube obstruction, hemoptysis, oxygen saturation of less than 85% on pulse oximetry or a Pao2 of less than 55 mm Hg for more than 5 minutes when the Fio2 was 1.0, cardiac arrest, a heart rate of less than 30 beats per minute for more than 1 minute, a systolic blood pressure of less than 60 mm Hg for more than 5 minutes, and any other life-threatening reason for which the clinician decided to stop the treatment. After patients in the prone group were turned to the supine position, the prone session could be resumed at any time before the planned assessment at 4 hours in the supine position if the criteria for oxygen saturation level, Pao2, or both were met. The prone-positioning strategy was applied every day up to day 28, after which it was used at the clinician’s discretion. Patients in the supine group could not be crossed over to the prone group except as a rescue measure in case of life-threatening hypoxemia when all the following criteria were met simultaneously: a Pao2:Fio2 ratio of less than 55 mm Hg, with an Fio2 of 1.0; maximal PEEP according to the PEEP–Fio2 table; administration of inhaled nitric oxide at a concentration of 10 ppm; infusion of intravenous almitrine bismesylate at a dose of 4 μg per kilogram per minute; and performance of respiratory recruitment maneuvers to increase the amount of aerated lung. Weaning from mechanical ventilation was conducted in the same way for both groups (see the Supplementary Appendix). Details regarding the management of sedation and the use of neuromuscular blocking agents are also provided in the Supplementary Appendix. The investigators assessed patients at least every morning until day 28 or discharge from the ICU. Data Collection At the time of admission, we recorded data on age, sex, the setting from which the patient was admitted to the ICU, the context for admission to the ICU, McCabe score14 (which ranges from A to C, with A indicating no underlying disease that compromises life expectancy, B an estimated life expectancy with the chronic disease of <5 years, and C an estimated life expectancy with the chronic disease of <1 year), ventilator settings, time from intubation to randomization, height, predicted body weight, and the Simplified Acute Physiology Score (SAPS) II15 (which ranges from 0 to 164, with higher scores indicating greater severity of symptoms). We also recorded the number of lung quadrants involved on chest radiography, results of measurements of arterial blood gases, PplatRS, arterial blood lactate levels, the cause of ARDS, the Sepsisrelated Organ Failure Assessment (SOFA) score16 (which ranges from 0 to 24, with higher scores indicating more severe organ failure), the lung injury score (which ranges from 0 to 4, with higher scores indicating more severe lung injury),17 and the time at which the first prone session was started. The following events were recorded daily until day 28: attempts at extubation, administration of inhaled nitric oxide, infusion of almitrine bismesylate, use of extracorporeal membrane oxygenation (ECMO), infusion of sedatives and neuromuscular blockers, complications, and the SOFA score. Ventilator settings, PplatRS, static compliance of the respiratory system, and the results of measurements of arterial blood gases were recorded daily during the first week as indicated above. Data quality was verified by the research n engl j med 368;23 nejm.org june 6, 2013 2161 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e 51,189 Patients were admitted to 27 ICUs in the study period, Jan. 1, 2008–July 25, 2011 47,740 Did not have ARDS 3449 Had ARDS 2015 Were not screened 1434 Were screened 858 Were not eligible 576 Were eligible 102 Were excluded 37 Had improved symptoms after 12–24 hr 55 Had organizational problems in study center 10 Were withdrawn by physician 474 Underwent randomization 234 Were assigned to supine group 240 Were assigned to prone group 5 Were excluded 3 Had PaO2:FIO2 >150 mm Hg 1 Was enrolled before 12-hr stabilization period was over 1 Had guardianship issues 3 Were excluded 2 Were enrolled before 12-hr stabilization period was over 1 Received NIV >24 hr 466 Were included in the intention-to-treat analysis 229 Were in supine group 237 Were in prone group 466 Were included in the 90-day follow-up 229 Were in supine group 237 Were in prone group Figure 1. Enrollment, Randomization, and Follow-up of the Study Participants. ARDS denotes the acute respiratory distress syndrome, ICU intensive care unit, NIV noninvasive ventilation, and Pao2:Fio2 the ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen. 2162 n engl j med 368;23 nejm.org june 6, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Prone Positioning in Severe ARDS fellows, and data were stored in a database (Clin info) that was specifically developed for the study with the use of Epi Info software, version 3.4.3. Table 1. Characteristics of the Participants at Inclusion in the Study.* Supine Group (N = 229) Prone Group (N = 237) 60±16 58±16 152 (66.4) 166 (70.0) Emergency room 98 (42.8) 101 (42.6) Acute care facility 87 (38.0) 86 (36.3) Home 26 (11.4) 31 (13.1) ICU 9 (3.9) 11 (4.6) Other 9 (3.9) 8 (3.4) A 183 (79.9) 197 (83.1) B 45 (19.7) 39 (16.5) C 1 (0.4) 1 (0.4) Diabetes 39 (17.0) 50 (21.1) Renal failure 12 (5.2) 10 (4.2) Hepatic disease 16 (7.0) 15 (6.3) Coronary artery disease 24 (10.5) 24 (10.1) Cancer 30 (13.1) 24 (10.1) COPD 29 (12.7) 23 (9.7) 38 (16.6) 32 (13.5) Characteristic Outcome Measures Age — yr The primary end point was mortality at day 28. Secondary end points were mortality at day 90, the rate of successful extubation, the time to successful extubation, the length of stay in the ICU, complications, the use of noninvasive ventilation, the tracheotomy rate, the number of days free from organ dysfunction, and ventilator settings, measurements of arterial blood gases, and respiratory-system mechanics during the first week after randomization. Successful extubation was defined as no reintubation or use of noninvasive ventilation in the 48 hours after extubation. In patients who had undergone a tracheotomy, successful weaning from the ventilator was defined as the ability to breathe unassisted through the tracheostomy cannula for at least 24 hours. Male sex — no. (%) Statistical Analysis The expected 28-day mortality in the supine group was 60%. We estimated that with a sample of 456 patients, the study would have 90% power to detect an absolute reduction of 15 percentage points (to 45%) with prone positioning, at a onesided type I error rate of 5%. An interim analysis was planned 28 days after half the patients had been enrolled, and two analyses were scheduled, each with a type I error rate set to 2.5% to maintain an overall type I error rate of 5%. The statistician sent the data from the interim analysis to the data and safety monitoring board, which had to decide whether to continue or discontinue the trial. An absolute difference in mortality of 25 percentage points or more between groups at the time of the interim analysis was the only criterion for early trial termination. There was no stopping rule for futility. The analysis was performed on an intentionto-treat basis. Continuous variables were expressed as means with standard deviations. Data were compared between groups with the use of the chi-square test or Fisher’s exact test and analysis of variance as indicated. Patient survival was analyzed with the use of the Kaplan–Meier method and compared between groups with the use of the log-rank test. Cox proportional-hazards regression, with stratification according to center, Setting from which patient was admitted to ICU — no. (%) McCabe score — no. (%)† Coexisting conditions — no. (%) Immunodeficiency — no. (%) SAPS II‡ Sepsis — no./total no. (%)§ SOFA score¶ 47±17 45±15 195/229 (85.2) 194/236 (82.2) 10.4±3.4 9.6±3.2 133 (58.1) 148 (62.4) 29±7 28±6 Vasopressors 190/229 (83.0) 172/237 (72.6) Neuromuscular blockers 186/226 (82.3) 212/233 (91.0) 39/228 (17.1) 27/237 (11.4) 101/225 (44.9) 91/230 (39.6) ARDS due to pneumonia Body-mass index‖ Other interventions — no./total no. (%) Renal-replacement therapy Glucocorticoids *Plus–minus values are means ±SD. There were no significant differences between the groups in any of the characteristics listed, with the exception of the Sepsis-related Organ Failure Assessment (SOFA) score, the use of vasopressors, and the use of neuromuscular blockers. ARDS denotes the acute respiratory distress syndrome, COPD chronic obstructive pulmonary disease, and ICU intensive care unit. A version of this table with additional information is available as Table S2 in the Supplementary Appendix. †A McCabe score of A indicates no underlying disease that compromises life expectancy, B an estimated life expectancy with the chronic disease of less than 5 years, and C an estimated life expectancy with the chronic disease of less than 1 year. ‡The Simplified Acute Physiology Score (SAPS) II ranges from 0 to 164, with higher scores indicating greater severity of symptoms. § Sepsis was defined according to the American–European Consensus Conference criteria. ¶SOFA scores range from 0 to 24, with higher scores indicating more severe organ failure. ‖The body-mass index is the weight in kilograms divided by the square of the height in meters. n engl j med 368;23 nejm.org june 6, 2013 2163 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l Table 2. Ventilator Settings, Respiratory-System Mechanics, and Results of Arterial Blood Gas Measurements at the Time of Inclusion in the Study.* Supine Group (N = 229) Prone Group (N = 237) Tidal volume (ml) 381±66 384±63 Tidal volume (ml per kg of PBW) 6.1±0.6 6.1±0.6 Respiratory frequency (breaths per min) 27±5 27±5 PEEP (cm of water) 10±4 10±3 0.79±0.16 0.79±0.16 23±5 24±5 Variable Fio2 PplatRS (cm of water) CstRS (ml per cm of water) 35±15 36±23 Pao2 (mm Hg) 80±18 80±19 100±20 100±30 52±32 50±14 7.30±0.10 7.30±0.10 25±5 25±5 Pao2:Fio2 (mm Hg) Paco2 (mm Hg) Arterial pH Plasma bicarbonate (mmol per liter)† *Plus–minus values are means ±SD. CstRS denotes static compliance of the respiratory system, Fio2 the fraction of inspired oxygen, Paco2 partial pressure of arterial carbon dioxide, Pao2 partial pressure of arterial oxygen, PBW predicted body weight, PEEP positive end-expiratory pressure, and PplatRS endinspiratory plateau pressure of the respiratory system. †Data are for 227 participants in the supine group and 236 participants in the prone group. was planned to adjust the between-group differences in mortality at day 28 and day 90 for significant baseline covariates. The statistical analysis was performed with the use of SPSS software (SPSS for Windows, version 17.0). The investigators had no access to the database until the study was completed. All reported P values are two-sided, and have not been adjusted for multiple comparisons. A P value of less than 0.05 was considered to indicate statistical significance. R e sult s Participants From January 1, 2008, through July 25, 2011, a total of 3449 patients with ARDS were admitted to the participating ICUs, and 474 underwent randomization (Fig. 1). Eight patients were subsequently excluded (Fig. 1), and 466 patients were included in the analysis: 229 in the supine group and 237 in the prone group. After the interim analysis, the data and safety monitoring board recommended that the trial be continued. of m e dic i n e for the SOFA score and the use of neuromuscular blockers and vasopressors (Table 1). In more than half the cases, the main cause of ARDS was pneumonia (Table 1). Influenza A (H1N1) virus infection was the main cause of ARDS in 28 patients, with no significant difference between the groups in the rate (5.7% in the supine group and 6.3% in the prone group, P = 0.85). The mean (±SD) time from intubation to randomization was 31±26 hours in the supine group and 33±24 hours in the prone group (P = 0.66). The lung injury score was 3.3±0.4 in both groups, and the rate of use of noninvasive ventilation in the 24 hours before inclusion was similar in the two groups (29.3% and 30.8% in the supine and prone groups, respectively). Ventilator settings, respiratory-system mechanics, and results of arterial blood-gas measurements were also similar in the two groups (Table 2). Prone Positioning Patients in the prone group underwent their first prone-positioning session within 55±55 minutes after randomization. The average number of sessions was 4±4 per patient, and the mean duration per session was 17±3 hours. All the patients in this group underwent at least one prone-positioning session. In the prone group, patients were ventilated in the prone position for 73% of the 22,334 patient-hours spent in the ICU from the start of the first session to the end of the last session. Adjunctive Therapies The rates of the use of rescue therapies in the supine and prone groups were 2.6% versus 0.8% for ECMO (P = 0.14), 15.7% versus 9.7% for inhaled nitric oxide (P = 0.05), and 6.6% versus 2.5% for almitrine bismesylate (P = 0.04). Neuromuscular blockers were used for 5.6±5.0 days in the supine group and 5.7±4.7 days in the prone group (P = 0.74), and intravenous sedation was given for 9.5±6.8 and 10.1±7.2 days in the two groups, respectively (P = 0.35). The use of antiviral therapy for H1N1 virus infection was similar in the two groups. Ventilator Settings and Lung Function during the First Week The Pao2:Fio2 ratio recorded in the supine position was significantly higher in the prone group than in the supine group at days 3 and 5, whereCharacteristics at Inclusion as the PEEP and Fio2 were significantly lower The characteristics of the patients at inclusion in (Table S3 in the Supplementary Appendix). The the study were similar in the two groups except PplatRS was 2 cm of water lower by day 3 in the 2164 n engl j med 368;23 nejm.org june 6, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Prone Positioning in Severe ARDS Table 3. Primary and Secondary Outcomes According to Study Group.* Supine Group (N = 229) Prone Group (N = 237) 75 (32.8 [26.4–38.6]) 38 (16.0 [11.3–20.7]) Outcome Hazard Ratio or Odds Ratio with the Prone Position (95% CI) P Value Mortality — no. (% [95% CI]) At day 28 Not adjusted Adjusted for SOFA score† 0.39 (0.25–0.63) <0.001 0.42 (0.26–0.66) <0.001 0.44 (0.29–0.67) <0.001 0.48 (0.32–0.72) <0.001 0.45 (0.29–0.70) <0.001 At day 90 Not adjusted 94 (41.0 [34.6–47.4]) 56 (23.6 [18.2–29.0]) 145/223 (65.0 [58.7–71.3]) 186/231 (80.5 [75.4–85.6]) Survivors 19±21 17±16 Nonsurvivors 16±11 18±14 Survivors 26±27 24±22 Nonsurvivors 18±15 21±20 Adjusted for SOFA score† Successful extubation at day 90 — no./total no. (% [95% CI]) Time to successful extubation, assessed at day 90 — days 0.87 Length of ICU stay, assessed at day 90 — days 0.05 Ventilation-free days At day 28 10±10 14±9 <0.001 At day 90 43±38 57±34 <0.001 13 (5.7 [3.9–7.5]) 15 (6.3 [4.9–7.7]) 0.89 (0.39–2.02) Pneumothorax — no. (% [95% CI]) 0.85 Noninvasive ventilation — no./ total no. (% [95% CI]) At day 28 10/212 (4.7 [1.9–7.5]) 4/228 (1.8 [0.1–3.5]) 0.36 (0.07–3.50) 0.11 At day 90 3/206 (1.5 [0.2–3.2]) 4/225 (1.8 [0.1–3.5]) 1.22 (0.23–6.97) 1.00 Tracheotomy — no./total no. (% [95% CI]) At day 28 12/229 (5.2 [2.3–8.1]) 9/237 (3.8 [1.4–6.0]) 0.71 (0.27–1.86) 0.37 At day 90 18/223 (8.1 [4.5–11.7]) 15/235 (6.4 [3.3–9.5]) 0.78 (0.36–1.67) 0.59 *Plus–minus values are means ±SD. Hazard ratios are shown for mortality and successful extubation; odds ratios are shown for other outcomes. CI denotes confidence interval. †There were no significant differences between the groups in organ dysfunction as assessed from the SOFA score (Table S4 in the Supplementary Appendix). prone group than in the supine group. The partial pressure of arterial carbon dioxide and static compliance of the respiratory system were similar in the two groups. Primary and Secondary Outcomes Mortality at day 28 was significantly lower in the prone group than in the supine group: 16.0% (38 of 237 participants) versus 32.8% (75 of 229) (P<0.001) (Table 3). The significant difference in mortality persisted at day 90 (Table 3). A comparison of the two survival curves showed the same significant difference (Fig. 2). After adjustment for the SOFA score and the use of neuromuscular blockers and vasopressors at the time of inclusion, mortality remained significantly lower in the prone group than in the supine group (Table S5 in the Supplementary Appendix). The rate of successful extubation was significantly higher in the prone group (Table 3). The duration n engl j med 368;23 nejm.org june 6, 2013 2165 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l Cumulative Probability of Survival 1.0 Prone group 0.8 0.6 Supine group 0.4 0.2 0.0 P<0.001 0 10 20 30 40 50 60 70 80 90 Days No. at Risk Prone group Supine group 237 229 202 163 191 150 186 139 182 136 Figure 2. Kaplan–Meier Plot of the Probability of Survival from RandomizaRevised AUTHOR: Guérin tion to Day 90. FIGURE: 2 of 2 ARTIST: ts SIZE 4 col TYPE:of invasive Line Combo 4-C H/T 22p3 length mechanical ventilation, JOB: of stay in the ICU, incidence of pneumothorax, rate of use AUTHOR, PLEASE NOTE: Figure has been redrawnventilation and type has been reset. of noninvasive after extubation, and Please check carefully. tracheotomy rate did not differ significantly be36823 tween the two groups (TableISSUE: 3). 06-06-13 Complications A total of 31 cardiac arrests occurred in the supine group, and 16 in the prone group (P = 0.02). There were no significant differences between the groups with respect to other adverse effects (Table S6 in the Supplementary Appendix). Discussion Videos showing prone positioning of patients with ARDS are available at NEJM.org 2166 Survival after severe ARDS was significantly higher in the prone group than in the supine group. Furthermore, the effect size was large despite the fact that mortality in the supine group was lower than anticipated. Our results are consistent with findings from previous meta-analyses2,11 and an observational study,18 even though prior randomized trials have failed to show a survival benefit with prone positioning. Meta-analyses of ARDS studies have suggested that the outcomes with prone positioning are better in the subgroup of patients with severe hypoxemia.2,11 However, when we stratified our analysis according to quartile of Pao2:Fio2 of m e dic i n e ratio at enrollment, we found no significant differences in outcomes (Table S8 in the Supplementary Appendix). Several factors may explain our results. First, patients with severe ARDS were selected on the basis of oxygenation together with PEEP and Fio2 levels. Second, patients were included after a 12-to-24-hour period during which the ARDS criteria were confirmed. This period may have contributed to the selection of patients with more severe ARDS19 who could benefit from the advantages of the prone positioning, such as relief of severe hypoxemia and prevention of ventilatorinduced lung injury. A previous study has shown that prone positioning, as compared with supine positioning, markedly reduces the overinflated lung areas while promoting alveolar recruitment.5 These effects (reduction in overdistention and recruitment enhancement) may help prevent ventilator-induced lung injury by homogenizing the distribution of stress and strain within the lungs. In our trial, alveolar recruitment was not directly assessed. However, studies have shown that lung recruitability correlates with the extent of hypoxemia20,21 and that the transpulmonary pressure along the ventral-to-dorsal axis is more homogeneously distributed in the prone position than in the supine position.22 We therefore suggest that prone positioning in our patients induced a decrease in lung stress and strain. Third, as in previous investigations,9,10 we used long prone-positioning sessions. Fourth, the prone position was applied for 73% of the time ascribed to the intervention and was concentrated over a period of a few days. Fifth, in our trial, the tidal volume was lower than in previous trials,9,10 and the PplatRS was kept below 30 cm of water. However, because all patients were returned to the supine position at least once a day, the effect of the prone position itself cannot be distinguished from the effects of being moved from the supine to the prone position over the course of a day. We should acknowledge that the technical aspects of prone positioning are not simple and that a coordinated team effort is required (see Videos 1 and 2, available at NEJM.org). All centers participating in this study were skilled in the process of turning patients from the supine to the prone position, as shown by the absence of adverse events directly related to repositioning. Because the experience of the units may explain n engl j med 368;23 nejm.org june 6, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Prone Positioning in Severe ARDS the low rate of complications, our results cannot necessarily be generalized to centers without such experience. We should also emphasize that our results were obtained in the subgroup of severely ill patients with ARDS. It could be argued that our results can be explained by higher mortality in the control group. However, mortality at day 28 in the supine group was similar to that among controls in recent trials.23,24 Furthermore, although the mortality in the control group was lower than that used to compute the power of this study, we calculated that the power of our study was 99%. The study has several limitations. Although we planned to record the data of patients who were eligible but not included, only a few ICUs complied with this request, making it impossible to fully appreciate the physiological condition of the excluded patients. In addition, fluid balance and the cumulative dose of catecholamines were not assessed. The imbalance between the groups in baseline SOFA score, vasopressor use, and the use of neuromuscular blockers could also have influenced the results. However, even after adjustment for these covariates, mortality was significantly lower in the prone group. In conclusion, this trial showed that patients with ARDS and severe hypoxemia (as confirmed by a Pao2:Fio2 ratio of <150 mm Hg, with an Fio2 of ≥0.6 and a PEEP of ≥5 cm of water) can benefit from prone treatment when it is used early and in relatively long sessions. Dr. Guérin reports receiving grant support from Air Liquide; Dr. Mercat, receiving consulting fees from Faron Pharmaceuticals, grant support from Covidien and General Electric, patent royalties on a method for evaluating positive end-expiratory pressure that is licensed to General Electric, and reimbursement for travel expenses from Covidien and Maquet; Dr. Jaber, receiving consulting fees from Maquet and Dräger, lecture fees from Fisher and Paykel, Abbott Laboratories, and Philips Respironics, and reimbursement for travel expenses from Pfizer; and Dr. Mancebo, receiving fees for serving on the data and safety monitoring board of Air Liquide, consulting fees from Faron Pharmaceuticals, ALung, and Philips Respironics, and grant support to his institution from Covidien and General Electric. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank all the physicians, including those on night duty, and nurses in the participating centers for the care provided to patients during the study; the Réseau Européen en Ventilation Artificielle network; and Carolyn Newey for help in editing an earlier version of the manuscript. Appendix The authors’ affiliations are as follows: Réanimation Médicale, Hôpital de la Croix-Rousse, Hospices Civils de Lyon; Université de Lyon; and Creatis INSERM 1044, Lyon (C.G., J.-C.R., F.B., G.B., V.L., L.B.); Réanimation Polyvalente, and Clinical Research in Intensive Care and Sepsis (CRICS) Group, La Roche-Sur-Yon (J. Reignier); Réanimation Polyvalente, Roanne (P.B.); Réanimation Médicale, Hôpital Pontchaillou, Rennes (A.G.); Réanimation Polyvalente, and CRICS Group, Hôpital d’Orléans, Orleans (T.B.); Réanimation Médicale, Hôpital Bretonneau, CRICS Group, and Université de Tours, Tours (E.M.); Réanimation Polyvalente, Hôpital de Chambéry, Chambery (M.B.); L’Université Nantes Angers Le Mans, Université d’Angers, Centre Hospitalier Universitaire Angers, Réanimation Médicale, Angers (A.M.); Réanimation Polyvalente, and CRICS Group, Hôpital d’Angoulême, Angouleme (O.B.); Réanimation Polyvalente Centre d’Investigation Clinique 0801 and CRICS Group, Hôpital de Limoges, Limoges (M.C.); Réanimation Médicale, Hôpital de Poitiers, and CRICS Group, and University of Poitiers, Poitiers (D.C.); Réanimation Chirurgicale, Hôpital Saint Eloi, INSERM Unité 1046, and Université de Montpellier, Montpellier (S.J.); Réanimation Polyvalente, Hôpital Saint Joseph et Saint Luc, Lyon (S.R.); Réanimation Polyvalente, Hôpital d’Annecy, Annecy (M.S.); Réanimation Médicale, Hôpital Pellegrin, and Université de Bordeaux, Bordeaux (G.H.); Réanimation Polyvalente, Hôpital de Nîmes, and Université de Nîmes-Montpellier, Nimes (C.B.); Réanimation Polyvalente, Hôpital de Cergy-Pontoise, Cergy-Pontoise (J. Richecoeur); Réanimation des Urgences, Hôpital de la Timone, and Université de la Méditerranée, Marseille (M.G.); Service d’Hygiène Hospitalière, Groupement Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite (R.G.); and Centre de Coordination et de Lutte contre les Infections Nosocomiales Sud-Est, Hôpital Henri Gabrielle, Saint Genis-Laval (L.A.) — all in France; and Servei de Medicina Intensiva, Hospital de Sant Pau, Barcelona (J.M.). References 1. Abroug F, Ouanes-Besbes L, Elatrous S, Brochard L. The effect of prone positioning in acute respiratory distress syndrome or acute lung injury: a meta-analysis: areas of uncertainty and recommendations for research. Intensive Care Med 2008;34:1002-11. 2. Sud S, Friedrich JO, Taccone P, et al. Prone ventilation reduces mortality in patients with acute respiratory failure and severe hypoxemia: systematic review and meta-analysis. Intensive Care Med 2010; 36:585-99. 3. Broccard A, Shapiro RS, Schmitz LL, Adams AB, Nahum A, Marini JJ. Prone positioning attenuates and redistributes ventilator-induced lung injury in dogs. Crit Care Med 2000;28:295-303. 4. Mentzelopoulos SD, Roussos C, Zakynthinos SG. Prone position reduces lung stress and strain in severe acute respiratory distress syndrome. Eur Respir J 2005; 25:534-44. 5. Galiatsou E, Kostanti E, Svarna E, et al. Prone position augments recruitment and prevents alveolar overinflation in acute lung injury. Am J Respir Crit Care Med 2006;174:187-97. 6. Papazian L, Gainnier M, Marin V, et al. Comparison of prone positioning and high-frequency oscillatory ventilation in patients with acute respiratory distress syndrome. Crit Care Med 2005;33:2162-71. 7. Gattinoni L, Tognoni G, Pesenti A, et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001;345:568-73. 8. Guerin C, Gaillard S, Lemasson S, et al. Effects of systematic prone positioning in hypoxemic acute respiratory failure: a randomized controlled trial. JAMA 2004;292:2379-87. n engl j med 368;23 nejm.org june 6, 2013 2167 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Prone Positioning in Severe ARDS 9. Mancebo J, Fernández R, Blanch L, et al. A multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care Med 2006;173:1233-9. 10. Taccone P, Pesenti A, Latini R, et al. Prone positioning in patients with moderate and severe acute respiratory distress syndrome: a randomized controlled trial. JAMA 2009;302:1977-84. 11. Gattinoni L, Carlesso E, Taccone P, Polli F, Guérin C, Mancebo J. Prone positioning improves survival in severe ARDS: a pathophysiologic review and individual patient meta-analysis. Minerva Anestesiol 2010;76:448-54. 12. Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818-24. 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;342:1301-8. 14. McCabe WR, Treadwell TL, De Maria A Jr. Pathophysiology of bacteremia. Am J Med 1983;75:7-18. 15. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993;270: 2957-63. [Erratum, JAMA 1994;271:1321.] 16. 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; 22:707-10. 17. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138:720-3. [Erratum, Am Rev Respir Dis 1989:139:1065.] 18. Charron C, Bouferrache K, Caille V, et al. Routine prone positioning in patients with severe ARDS: feasibility and impact on prognosis. Intensive Care Med 2011; 37:785-90. 19. Villar J, Pérez-Méndez L, López J, et al. An early PEEP/FiO2 trial identifies different degrees of lung injury in ARDS patients. Am J Respir Crit Care Med 2007; 176:795-804. 20. Gattinoni L, Caironi P, Cressoni M, et al. Lung recruitment in patients with the acute respiratory distress syndrome. N Engl J Med 2006;354:1775-86. 21. Ranieri VM, Rubenfeld GD, Thompson BT, et al. Acute respiratory distress syndrome: the Berlin Definition. JAMA 2012;307:2526-33. 22. Mutoh T, Guest RJ, Lamm WJ, Albert RK. Prone position alters the effect of volume overload on regional pleural pressures and improves hypoxemia in pigs in vivo. Am Rev Respir Dis 1992;146:300-6. 23. Papazian L, Forel J-M, Gacouin A, et al. Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010;363:1107-16. 24. Ferguson ND, Cook DJ, Guyatt GH, et al. High-frequency oscillation in early acute respiratory distress syndrome. N Engl J Med 2013;368:795-805. Copyright © 2013 Massachusetts Medical Society. an nejm app for iphone The NEJM Image Challenge app brings a popular online feature to the smartphone. Optimized for viewing on the iPhone and iPod Touch, the Image Challenge app lets you test your diagnostic skills anytime, anywhere. The Image Challenge app randomly selects from 300 challenging clinical photos published in NEJM, with a new image added each week. View an image, choose your answer, get immediate feedback, and see how others answered. The Image Challenge app is available at the iTunes App Store. 2168 n engl j med 368;23 nejm.org june 6, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 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 identified 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 identified by review of electronic medical records and analyzed in univariate and multivariate analyses. Seventy-one of 160 patients (44%) S 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 conflicts of interest. Supported by NIH grant: NHLBI 1 K23 HL087843 For information regarding this article, E-mail: [email protected] Copyright © 2008 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e31816fc2c0 1518 developed ALI at a median of 5 (range 2–94) hours after the onset of septic shock. Multivariate logistic regression analysis identified the following predictors of ALI development: delayed goal-directed resuscitation (odds ratio [OR] 3.55, 95% confidence interval [CI] 1.52– 8.63, p ⴝ .004), delayed antibiotics (OR 2.39, 95% CI 1.06 ⴚ5.59, p ⴝ .039), transfusion (OR 2.75, 95% CI 1.22– 6.37, p ⴝ .016), alcohol abuse (OR 2.09, 95% CI .88ⴚ5.10, p ⴝ 0.098), recent chemotherapy (OR 6.47, 95% CI 1.99ⴚ24.9, p ⴝ 0.003), diabetes mellitus (OR .44, 95% CI .17ⴚ1.07, p ⴝ .076), and baseline respiratory rate (OR 2.03 per SD, 95% CI 1.38ⴚ3.08, p < .001). Conclusion: When adjusted for known modifiers of ALI expression, delayed treatment of shock and infection were associated with development of ALI. (Crit Care Med 2008; 36:1518–1522) KEY WORDS: shock; metabolic acidosis; hyperventilation; epidemiology; antibiotic; resuscitation some patients with septic shock do and others do not develop ALI has been limited. Previous studies have identified 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 significant modifiers of ALI expression. However, the timing of ALI development has generally been poorly defined, making it difficult to separate cause from effect of specific 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 identified 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 identified 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 inflammatory response syndrome (SIRS) criteria (temperature ⬎38.3°C or ⬍35.6°C, heart rate ⬎90 beats/ min, respiratory rate ⬎20/min or white blood cell [WBC] count ⬎12.0 ⫻ 103 or ⬍4.0 ⫻ 103); 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 Crit Care Med 2008 Vol. 36, No. 5 fluid bolus, and/or serum lactate ⱖ4 mmol/L regardless of the blood pressure (10). Primary outcome was the development of ALI, defined 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 infiltrates 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 identified 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 findings, 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 confirmed 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 infiltrates (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 defined as a previously established diagnosis of chronic alcoholism, a prior admission for alcohol detoxification, or alcohol withdrawal (14). Tobacco use was defined as history of cigarette smoking based on the information provided in the standardized form on hospital admission. Aspiration was defined as aspiration of gastric contents into the airway witnessed or strongly suspected by the bedside provider. 2) Severity of illness, inflammation, 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). Crit Care Med 2008 Vol. 36, No. 5 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 first six hours of septic shock: time to the administration of antibiotic therapy (16), source control (17), goal-directed fluid resuscitation (10), transfusion of blood products, use of vasopressors, steroid administration, and use of activated protein C. Adequate antibiotic therapy was defined 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 defined 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 defined as central venous oxygen saturation ⱖ70 percent (10) and/or a combination of clinical factors (18): central venous pressure ⬎8 mmHg, MAP ⱖ65 mmHg, urine output ⱖ.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 first six hours was considered as adequate goaldirected resuscitation; achievement of resuscitation goals after the first 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, Fisher’s exact test, or chisquare test, as appropriate. The variables were considered for multivariable logistic regression models if they occurred before or during the first 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 ⱖ2 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 identified modifiers of ALI (references 2– 6, see the Results section) were included in the multivariate analysis as covariates. The final 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 2–94) hrs after the onset of septic shock. Ninety percent of patients developed ALI during the first 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 modifiers 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 significantly 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% confidence 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 .62–5.74) were significantly 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 significantly associated with the presence of metabolic acidosis, base def1519 Table 1. Risk factors for development of ALI in patients with septic shock ALI (n ⫽ 71) 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, inflammation, 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 first 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 first 6 hrs (IQR) Fluid intake, mL 6–12 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 (%) No ALI (n ⫽ 89) 68 (56 to 80) 33 (46) 17 (24) 27 (38) 38 (54) 24 (34) 17 (24) 17 (24) 7 (10) 73 (61 to 83) 39 (44) 38 (43) 15 (17) 33 (37) 23 (26) 5 (6) 22 (25) 5 (6) 35 (49) 21 (30) 8 (11) 5 (7) 2 (3) 18 (25) 18 (25) 28 (32) 24 (27) 20 (22) 10 (11) 7 (8) 18 (20) 8 (9) p Value .056 .737 .012 .003 .037 .273 ⬍.001 .910 .313 .064 .441 .005 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) 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) 223 (82.5 to 360) 44 (62) 113 (0 to 235) 36 (40) .001 .007 12 (17) 505 (329 to 653) 12 (14) 320 (202 to 480) .550 .001 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) 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) .028 .499 .039 .032 .548 ⬍.001 .033 ⬍.001 ⬍.001 ⬍.001 .038 .050 .566 ⬍.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 first 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 first 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 Delayed goal-directed resuscitation Delayed antibiotics Respiratory rate (per SD) Chemotherapy Chronic alcohol use Transfusion Aspiration Diabetes mellitus Odds Ratio 95% CI p Value 3.55 2.39 2.03 6.47 2.09 2.75 3.48 .44 1.52–8.63 1.06–5.59 1.38–3.08 1.99–24.9 .88–5.10 1.22–6.37 1.22–10.78 .17–1.07 .004 .039 ⬍.001 .003 .098 .016 .024 .076 SD, standard deviation; CI, confidence 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 deficit, 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 first 12 hrs after the septic shock onset. When adjusted for known modifiers 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 identified 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 significance, our study confirmed 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 identified pulmonary aspiration as a significant risk factor for ALI development (19). In our study, aspiration appeared to be a more important risk factor than having identified 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 confirmed the findings 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 first 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. Significant 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 identified, 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 definition of alcohol abuse was based on “a known diagnosis of chronic alcoholism or a previous admission for alcohol detoxification or alcohol withdrawal” (14). In addition, because a significant 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 first 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 definitions 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 definitions, 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 modifiers of ALI expression such as transfusion, aspiration, chemotherapy, alcohol use, and diabetes, our study identified delayed goal-directed resuscitation, delayed antibiotic administration, and baseline high respiratory rate as important determinants of ALI development. Targeting the modifiers of ALI expression may be an attractive strategy for prevention of this common and devastating ICU complication. 1522 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; 33:1191–1198 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; 151:293–301 4. Moss M, Burnham EL: Chronic alcohol abuse, acute respiratory distress syndrome, and multiple organ disfunction. Crit Care Med 2003; 31(4 Suppl):S207–S212 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; 28:3137–3145 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; 28:2187–2192 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; 101:1481–1483 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. Definitions, 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, 27–28 April 2006. Intensive Care Med 2007; 33: 575–590 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 Crit Care Med 2008 Vol. 36, No. 5 CARING FOR THE CRITICALLY ILL PATIENT Scan for Author Audio Interview ONLINE FIRST Daily Sedation Interruption in Mechanically Ventilated Critically Ill Patients Cared for With a Sedation Protocol A Randomized Controlled Trial Sangeeta Mehta, MD Lisa Burry, PharmD Deborah Cook, MD Dean Fergusson, PhD Marilyn Steinberg, RN John Granton, MD Margaret Herridge, MD Niall Ferguson, MD John Devlin, PharmD Maged Tanios, MD Peter Dodek, MD Robert Fowler, MD Karen Burns, MD Michael Jacka, MD Kendiss Olafson, MD Yoanna Skrobik, MD Paul Hébert, MD Elham Sabri, MSc Maureen Meade, MD for the SLEAP Investigators and the Canadian Critical Care Trials Group C RITICALLY ILL PATIENTS WEAN more quickly from mechanical ventilation, with lower risk of delirium, when clinicians use specific strategies to reduce excessive sedation. 1 - 3 A nursingimplemented sedation titration protocol that specifies clear targets for level of awareness is one approach to miniFor editorial comment see p 2030. Author Audio Interview available at www.jama.com. Context Protocolized sedation and daily sedation interruption are 2 strategies to minimize sedation and reduce the duration of mechanical ventilation and intensive care unit (ICU) stay. We hypothesized that combining these strategies would augment the benefits. Objective To compare protocolized sedation with protocolized sedation plus daily sedation interruption in critically ill patients. Design, Setting, and Patients Randomized controlled trial of 430 critically ill, mechanically ventilated adults conducted in 16 tertiary care medical and surgical ICUs in Canada and the United States between January 2008 and July 2011. Intervention Continuous opioid and/or benzodiazepine infusions and random allocation to protocolized sedation (n=209) (control) or to protocolized sedation plus daily sedation interruption (n=214). Using validated scales, nurses titrated infusions to achieve light sedation. For patients receiving daily interruption, nurses resumed infusions, if indicated, at half of previous doses. Patients were assessed for delirium and for readiness for unassisted breathing. Main Outcome Measure Time to successful extubation. Secondary outcomes included duration of stay, doses of sedatives and opioids, unintentional device removal, delirium, and nurse and respiratory therapist clinical workload (on a 10-point visual analog scale [VAS]). Results Median time to successful extubation was 7 days in both the interruption and control groups (median [IQR], 7 [4-13] vs 7 [3-12]; interruption group hazard ratio, 1.08; 95% CI, 0.86-1.35; P=.52). Duration of ICU stay (median [IQR], 10 [5-17] days vs 10 [6-20] days; P=.36) and hospital stay (median [IQR], 20 [10-36] days vs 20 [10-48] days; P=.42) did not differ between the daily interruption and control groups, respectively. Daily interruption was associated with higher mean daily doses of midazolam (102 mg/d vs 82 mg/d; P=.04) and fentanyl (median [IQR], 550 [50-1850] vs 260 [0-1400]; P⬍.001) and more daily boluses of benzodiazepines (mean, 0.253 vs 0.177; P=.007) and opiates (mean, 2.18 vs 1.79; P⬍.001). Unintentional endotracheal tube removal occurred in 10 of 214 (4.7%) vs 12 of 207 patients (5.8%) in the interruption and control groups, respectively (relative risk, 0.82; 95% CI, 0.36-1.84; P=.64). Rates of delirium were not significantly different between groups (53.3% vs 54.1%; relative risk, 0.98; 95% CI, 0.821.17; P=.83). Nurse workload was greater in the interruption group (VAS score, 4.22 vs 3.80; mean difference, 0.41; 95% CI, 0.17-0.66; P=.001). Conclusion For mechanically ventilated adults managed with protocolized sedation, the addition of daily sedation interruption did not reduce the duration of mechanical ventilation or ICU stay. Trial Registration clinicaltrials.gov Identifier: NCT00675363 JAMA. 2012;308(19):1985-1992 Published online October 17, 2012. doi:10.1001/jama.2012.13872 Author Affiliations are listed at the end of this article. Corresponding Author: Sangeeta Mehta, MD, Mount Sinai Hospital, 600 University Ave, Ste 18-216, Toronto, ©2012 American Medical Association. All rights reserved. www.jama.com ON, M5G 1X5 Canada ([email protected]). Caring for the Critically Ill Patient Section Editor: Derek C. Angus, MD, MPH, Contributing Editor, JAMA ([email protected]). JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 1985 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS mize sedation.4 Daily interruption of sedative infusions may achieve the same goal if infusions are resumed only when necessary and at half the previous dose.5 Early clinical trials evaluating each strategy led to strong recommendations for their use in practice.6 However, results of subsequent clinical trials varied,7-10 and use of these strategies in clinical practice has been inconsistent.11,12 Concerns about daily interruption of sedation include patient discomfort, unintentional device removal, and increased clinician workload.13,14 A systematic review of 5 trials that evaluated daily interruption highlighted the need for further research.15 Avoiding excessive sedation is intuitively appealing. In light of the observed and potential benefits of both protocolized sedation and daily interruption in some settings, we hypothesized that mechanically ventilated adults managed with both strategies would receive less sedation and have a shorter duration of mechanical ventilation than patients managed with protocolized sedation alone. METHODS We conducted this multicenter randomized controlled trial in 16 centers from January 2008 to July 2011, after approval from local institutional review boards. In preparation, we completed a 65-patient, 3-center pilot randomized trial.16 Participants Eligible critically ill adults were those expected by the intensive care unit (ICU) team to require mechanical ventilation for at least 48 hours after enrollment and for whom the ICU team had decided to initiate continuous sedative and/or opioid infusion(s). Patients admitted to the ICU after cardiac arrest or traumatic brain injury were excluded, as were patients receiving neuromuscular blocking agents, those enrolled in another trial, those previously enrolled in the current trial, or those for whom there was a lack of commitment to maximal treatment. Legally authorized surrogates provided written informed consent. 1986 JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 Randomization and Masking Research staff randomized patients to protocolized sedation plus daily interruption (interruption group) or protocolized sedation alone (control group), using an automated telephone system that stratified by center with undisclosed variable block sizes. None of the participants, study personnel, clinicians, or investigators analyzing data was masked to group assignment. Procedures Bedside nurses titrated analgesic and sedative infusions according to a protocol that prioritized pain assessment (eFigures 1 and 2, available at http: //www.jama.com). Morphine, fentanyl, or hydromorphone was administered for analgesia; midazolam or lorazepam, for sedation. Nurses used the Sedation-Agitation Scale17 (8 sites) (eTable 1) or the Richmond Agitation Sedation Scale18 (8 sites) (eTable 2) to assess sedation needs hourly and titrated infusions to maintain, ideally, a comfortable yet rousable state equivalent to a Sedation-Agitation Scale score of 3 or 4 or Richmond Agitation Sedation Scale score of −3 to 0. When the sedation score directed an increase in medication, the bedside nurse judged whether to increase the opioid and/or benzodiazepine infusions. When patients were oversedated, nurses alternately reduced opioid and benzodiazepine infusions. Midazolam and morphine were reduced in 1- to 2-mg decrements, fentanyl in 12.5- to 25-g decrements, and hydromorphone in 0.1- to 0.5-mg decrements at 15- to 30minute intervals. If doses of midazolam, lorazepam, or morphine were less than 3 mg/h, 0.5-mg decrements could be used. If Sedation-Agitation Scale score was 1 to 2 (Richmond Agitation Sedation Scale score −4 or −5), yet the patient showed signs of agitation or distress, bolus doses were administered as needed. When patients were extremely agitated (SedationAgitation Scale score 7; Richmond Agitation Sedation Scale score 3 or 4), nurses could deviate from this protocol. For both groups, infusions were discontinued when a patient was oversedated (Sedation-Agitation Scale score 1 or 2; Richmond Agitation Sedation Scale score −4 or −5) while receiving 0.5 to 1 mg/h of midazolam or morphine (or fentanyl, 12.5-25 g/ h). Intermittent dosing was permitted for procedures. Propofol, ketamine, and dexmedetomidine infusions were not permitted. In the interruption group, bedside nurses interrupted benzodiazepine and opioid infusions daily and assessed hourly for wakefulness, defined as Sedation-Agitation Scale score 4 to 7 (Richmond Agitation Sedation Scale score −1 to 4) and ability to perform at least 3 of the following on request: eye opening, tracking, hand squeezing, and toe moving. If the bedside nurse and a physician agreed that infusions were no longer required (the patient was free of discomfort and agitation and the Sedation-Agitation Scale score was between 2 and 5 or the Richmond Agitation Sedation Scale score was between −4 and 1), oral or bolus intravenous therapy was used at their discretion. Alternatively, if they judged that ongoing benzodiazepine or opioid infusions were required, nurses resumed infusions at half of the previous dose and titrated to achieve the target level of light sedation. If a patient became agitated (Sedation-Agitation Scale score 6 or 7 or Richmond Agitation Sedation Scale score 2 to 4) or exhibited signs of discomfort (respiratory rate ⬎35/ min, oxygen saturation as measured by pulse oximetry ⬍90%, heart rate ⬎140/ min or a change in heart rate of 20% in either direction, systolic blood pressure ⬎180 mm Hg, or increased anxiety and diaphoresis) before the physician’s assessment, nurses promptly resumed infusions at half the previous rate. Daily interruption could be delayed for procedures. When an interruption was not performed or infusions were not restarted at 50% of the previous dosage, the primary reason was documented. We also recorded any interruption of benzodiazepine and opioid infusions among control patients. ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS Patients were weaned from mechanical ventilation at the discretion of the ICU team. To standardize the assessment of a patient’s extubation readiness, respiratory therapists evaluated patients daily at their current ventilator settings for the following criteria: awake, adequate cough with suctioning, PaO2 greater than 60 mm Hg, oxygen saturation greater than or equal to 90%, fraction of inspired oxygen less than or equal to 0.4, positive end expiratory pressure less than or equal to 10 cm H2O, respiratory rate less than or equal to 35/min, ventilation less than or equal to 15 L/minute, no inotrope or vasopressor infusions, mean arterial pressure greater than 60 mm Hg, and no evidence of acute myocardial ischemia (ie, chest pain, consistent electrocardiogram findings, elevated biomarker levels, or new arrhythmia). If all criteria were met, a 1-hour trial of unassisted breathing was initiated, during which ventilatory support was withdrawn and the patient breathed spontaneously at the previous fraction of inspired oxygen through a t-tube circuit, a tracheostomy mask, or the ventilator circuit with continuous positive airway pressure of 5 cm H2O. The breathing trial could be terminated if any of the following signs of failure persisted for more than 5 minutes: respiratory rate greater than 35/min, oxygen saturation less than 90%, heart rate greater than 140/min or a change in heart rate of 20% in either direction, systolic blood pressure less than 90 or greater than 180 mm Hg, or increased anxiety and diaphoresis. A breathing trial was successful if the patient could breathe without mechanical assistance for 1 hour. When patients passed a trial of unassisted breathing, respiratory therapists notified a physician with a view to extubation. Research staff recorded reasons for delayed extubation, and daily screening continued until extubation. If the patient did not pass the unassisted breathing trial, the previous ventilator settings were resumed and the screening and breathing trials were repeated daily until extubation. If reintubation occurred Figure 1. Flow of Patients in the Trial 2091 Patients screened 1661 Excluded 403 No surrogate decision maker 371 Refused consent 179 Physician refused participation 104 Missed/no research staff 86 Enrolled in another trial 13 Not specified 10 Previously in SLEAP 495 Other a 430 Patients randomized 218 Randomized to receive protocolized sedation and daily interruption 212 Randomized to receive only protocolized sedation 4 Consent withdrawn 3 Consent withdrawn 214 Included in analysis 209 Included in analysis a Other includes 362 patients receiving propofol, 39 with open abdomen or chest, 33 needing ongoing deep sedation (because of a plan to return to the operating room, severe agitation, chronic pain, precarious airway, or hemodynamic instability), and 23 receiving high-frequency ventilation. For the remainder, please see the supplemental eAppendix. within 48 hours, study sedation procedures resumed. Bedside nurses also assessed daily for delirium with the Intensive Care Delirium Screening Checklist.19 Patients in both groups were managed “off protocol” during periods of neuromuscular blockade, high-frequency oscillation, or palliative care. Outcomes The primary study outcome was time to successful extubation, defined as time from randomization to extubation (or tracheostomy mask) for 48 hours. Secondary outcomes included unintentional device removal (eg, endotracheal tubes), physical restraint use, delirium, neuroimaging in the ICU, tracheostomy, barotrauma, total doses of sedatives and analgesics during mechanical ventilation, organ dysfunction, ICU and hospital lengths of stay, and death. Twice daily, nurses and respiratory therapists recorded their additional clinical workload attributed to study procedures, using a 10-point visual analog scale (VAS), with 1 corresponding to “very easy” and 10 to “difficult.” For patients assigned to daily interruption, we measured the propor- ©2012 American Medical Association. All rights reserved. tion of days during which sedation was interrupted. Statistical Analysis The sample size estimate assumed a median time to successful extubation of 7 days among controls and a 2-day reduction with the addition of daily interruption (hazard ratio 1.4). We determined that 205 patients per group would provide a power of 90%, with an ␣ level of 5%. Our primary analysis was based on an intention-to-treat principle whereby all patients were analyzed according to their original group allocation, regardless of whether they received the allocated treatment. We used the Kaplan-Meier method to estimate and plot the distributions of time to successful extubation and an unadjusted Cox proportional hazards model to estimate a hazard ratio. For the time-to-extubation analysis, the event occurred when a patient was extubated within 28 days from randomization and remained extubated for more than 48 hours. Patients who died before extubation or who were transferred to another institution before 28 days were censored at death or transfer. Patients undergoing withJAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 1987 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS drawal of life support were censored when that decision was made. We also conducted a per-protocol analysis of patients who had interrup- tions on more than 75% of eligible study days and 1 prespecified subgroup analysis, according to classification of a patient’s ICU admission as medical vs sur- Table 1. Baseline Characteristics No. (%) Characteristics Age, median (IQR), y Women Type of admission a Medical Surgical Trauma BMI, median (IQR) APACHE II score, median (IQR) b SOFA at day 1, median (IQR) c Mechanical ventilation, median (IQR), d Opioid infusions No. (%) Days of infusion, median (IQR) Benzodiazepine infusions No. (%) Days of infusion, median (IQR) ICU admission diagnosis d Bacterial/viral pneumonia Nonurinary sepsis Other respiratory disease Aspiration pneumonia COPD Postoperative respiratory disease Urinary sepsis Gastrointestinal perforation/rupture Hepatic failure Noncardiogenic pulmonary edema Other Pre-ICU conditions e Alcohol use Tobacco use Any psychiatric condition Any neurologic condition Respiratory disease Renal dysfunction Habitual drug use Liver disease Protocolized Sedation and Daily Interruption (n = 214) 57 (46-70) 93 (43.5) 175 (81.8) 30 (14.5) 8 (3.7) 28.2 (23.8-34.2) 24.0 (18-28) 7 (5-10) 2 (1-4) Protocolized Sedation (n = 209) 60 (49-70) 92 (44.0) 179 (86.1) 22 (11.0) 6 (2.9) 28.6 (25.0-33.2) 23.0 (19-29) 6 (4-9) 2 (1-4) 184 (87) 1 (1-3) 186 (89) 1 (1-3) 169 (81) 1 (1-3) 163 (80) 1 (1-3) 39 (18.2) 40 (18.7) 22 (10.3) 11 (5.1) 4 (1.9) 7 (3.3) 3 (1.4) 6 (2.8) 6 (2.8) 5 (2.3) 71 (33.2) 47 (22.5) 36 (17.2) 21 (10.0) 4 (1.9) 10 (4.8) 7 (3.3) 9 (4.3) 5 (2.4) 4 (1.9) 4 (1.9) 62 (29.7) 49 (23.0) 48 (22.5) 42 (19.6) 33 (15.4) 17 (8.0) 20 (9.4) 14 (6.6) 12 (5.6) 44 (21.2) 40 (19.3) 29 (14.4) 36 (17.2) 26 (12.4) 16 (7.7) 10 (4.8) 11 (5.3) Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; IQR, interquartile range; SOFA, Sequential Organ Failure Assessment. a Surgical refers to admission from an operating room or postoperative recovery area. b APACHE II score may range from 0 to 71, with higher scores indicating more severe disease. c SOFA score may range from 0 to 24 points, with higher scores indicating more severe disease. d Diagnoses in this category are mutually exclusive. The 10 most frequent diagnoses are listed, and the remainder are categorized as “other.” e Pre-ICU conditions are listed in descending frequency: neurologic condition defined as stroke, seizure disorder, dementia, neuromuscular disease, Parkinson disease, or other neurologic condition; psychiatric condition includes depression, bipolar disorder, schizophrenia, anxiety disorder, or other psychiatric condition; respiratory disease defined as home oxygen, carbon dioxide retention at baseline, or home ventilation; renal dysfunction defined as chronic renal failure with creatinine level greater than 180 mol/L or chronic dialysis; liver disease defined as Child Pugh Grade C or known esophageal varices; and habitual drug use other than tobacco or alcohol. 1988 JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 gical/trauma. We hypothesized that medical patients would benefit more than surgical patients from daily interruption, given their anticipated longer durations of mechanical ventilation and sedative infusions. Descriptive data are presented as percentages, means with standard deviations for normally distributed variables, and medians with interquartile ranges for nonnormally distributed variables. Sedative and opioid doses are presented as midazolam and fentanyl equivalents, respectively.20 We converted Richmond Agitation Sedation Scale values to SedationAgitation Scale scores for analyses (eTable 3). To examine between-group differences in categorical variables, we used 2 or Fisher exact tests, as appropriate. For dichotomous outcomes, we present relative risks or hazard ratios and their 95% CIs. If all assumptions were met for parametric analyses of the continuous variables, we used a 2-sample t test; otherwise, we used a 2-sample Wilcoxon rank sum test. Mean SedationAgitation Scale and VAS scores per patient and mean differences with 95% CIs were calculated. All statistical tests were 2-sided and considered statistically significant at ␣⬍.05. SAS version 9.2 and S-Plus version 7.0 were used for statistical analysis. An independent data and safety monitoring committee reviewed trial progress and adverse events after randomization of 67, 117, and 292 patients. They also reviewed blinded data for 1 planned interim analysis after enrollment of 211 patients. RESULTS Participants Patients were enrolled in 14 Canadian and 2 US centers. Of 2091 eligible patients, 1661 were not enrolled, primarily because of lack of an authorized decision maker (24.3%), consent refusal (22.3%), or physician refusal (10.8%) (FIGURE 1). Among 430 randomized patients, 7 withdrew consent in the first 3 days of the study and were excluded from the analysis. ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS Patient characteristics were similar in the 2 groups (TABLE 1). Eightyfour percent received medical diagnoses. At enrollment, 359 (84.9%) patients were receiving midazolam infusions; 334 (79.0%), fentanyl; 71 (16.5%), morphine; and 41 (9.5%), propofol. Propofol infusions were discontinued at enrollment according to the study protocol. ence in time to successful extubation between groups. There were no between-group differences in ICU or hospital lengths of stay, hospital mortality, rates of unintentional device removal, delirium, ICU neuroimaging, barotrauma, tracheostomy, or organ dysfunction (TABLE 2). TABLE 3 summarizes data related to sedative and opioid administration. Pa- Figure 2. Kaplan-Meier Curves for Time to Successful Extubation 1.0 Proportion Intubated 0.8 Outcomes The median time to successful extubation was 7 days in both groups (hazard ratio, 1.08; 95% CI, 0.86-1.35; P=.52) (FIGURE 2). Adjustment for age, body mass index, Acute Physiology and Chronic Health Evaluation II score, and admission type gave consistent results (adjusted hazard ratio, 1.04; 95% CI, 0.83-1.31). Similarly, in a perprotocol analysis of patients who had interruptions on more than 75% of eligible study days, there was no differ- 0.6 0.4 P = .52 Protocolized sedation only 0.2 0 Protocolized sedation and daily interruption 0 5 10 15 20 25 34 28 23 16 30 Time, d No. at risk Protocolized sedation only Protocolized sedation and daily interruption 209 214 146 140 72 81 49 42 P value calculated from log-rank statistic. Table 2. Patient Outcomes Days to successful extubation, median (IQR) a Days in ICU, median (IQR) a Days in hospital, median (IQR) a ICU mortality, No. (%) Hospital mortality, No. (%) ICU-acquired organ failure and supportive therapies, No. (%) ARDS Vasopressors/inotropes Renal replacement Neuromuscular blockade Unintentional device removal, No. (%) Gastric tube Endotracheal tube Urinary catheter Central venous or arterial catheter Neuroimaging in ICU, No. (%) Computed tomography Magnetic resonance imaging Physical restraint Patients, No. (%) Study days, mean (SD) Delirium, No (%) b Reintubation within 48 h, No. (%) Tracheostomy, No (%) Protocolized Sedation and Interruption (n = 214) 7 (4 to 13) 10 (5 to 17) 20 (10 to 36) 50 (23.4) 63 (29.6) Protocolized Sedation (n = 209) 7 (3 to 12) 10 (6 to 20) 20 (10 to 48) 52 (24.9) 63 (30.1) Measure of Effect (95% CI) HR, 1.08 (0.86 to 1.35) Mean difference, −3.17 (−6.89 to 0.55) Mean difference, −8.2 (−17.64 to 1.19) RR, 0.94 (0.67 to 1.32) RR, 0.98 (0.73 to 1.31) P Value .52 .36 .42 .72 .89 89 (41.8) 121 (56.8) 50 (23.5) 20 (9.7) 78 (37.3) 130 (62.2) 37 (17.7) 21 (10.2) RR, 1.12 (0.88 to 1.42) RR, 0.91 (0.78 to 1.07) RR, 1.33 (0.91 to 1.94) RR, 0.94 (0.53 to 1.69) .35 .26 .14 .84 18 (8.5) 10 (4.7) 29 (13.9) 12 (5.8) RR, 0.61 (0.35 to 1.07) RR, 0.82 (0.36 to 1.84) .08 .64 6 (2.8) 17 (8.0) 13 (6.2) 10 (4.8) RR, 0.45 (0.17 to 1.17) RR, 1.68 (0.79 to 3.57) .09 .18 29 (13.6) 9 (4.2) 33 (15.9) 7 (3.4) RR, 0.85 (0.54 to 1.35) RR, 1.25 (0.47 to 3.29) .53 .64 166 (76.4) 4.71 (5.67) 113 (53.3) 12 (5.6) 166 (79.4) 5.36 (6.14) 113 (54.1) 16 (7.7) RR, 0.96 (0.87 to 1.07) Mean difference, −0.70 (−1.84 to 0.43) RR, 0.98 (0.82 to 1.17) RR, 0.73 (0.35 to 1.50) .46 .83 .39 49 (23.2) 54 (26.3) RR, 0.88 (0.63 to 1.23) .46 Abbreviations: ARDS, acute respiratory distress syndrome; HR, hazard ratio; ICU, intensive care unit; IQR, interquartile range; RR, relative risk. a Analyses are measured from enrollment. b Patients who ever had a score of 4 or more on the Intensive Care Screening Delirium Checklist.19 ©2012 American Medical Association. All rights reserved. JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 1989 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS tients in the interruption group received higher mean daily benzodiazepine doses (102 vs 82 mg/d midazolam equivalents; median, 8 [IQR, 0-86 vs median, 0 [IQR, 0-50]; P = .04) and a greater number of boluses per day (mean, 0.253 vs 0.177; P = .007). They also received higher daily opioid doses (1780 vs 1070 g/d fentanyl equivalents; P⬍.001), both as infusion and boluses, and a greater number of opioid boluses per day (mean, 2.18 vs 1.79; P⬍ .001). Protocol Adherence and Clinician Workload Adherence with daily interruption was 72.2% of all eligible study days for an average patient and 85.6% for all eligible patient-days. Fifty-three percent of patients missed at least 1 daily interruption, and 6 patients missed every scheduled interruption. The most common reasons for noninterruption were related to mechanical ventilation (38.5%), agitation or pain (16.3%), and first day of study (14.6%) (eTable 4). Infusions were reinitiated at a dose exceeding 50% of the previous dose for 30 patients (14.1%) on a total of 47 days. Propofol infusions were administered to 28 patients, accounting for 3.0% of study days. In the control group, 34 patients (16.4%) had infusions interrupted on 54 occasions, accounting for 2.3% of study days. Forty patients receiving propofol infusions accounted for 2.1% of study days. Overall, mean Sedation-Agitation Scale scores per patient were similar in the 2 groups (3.28 [95% CI, 2.92 to 3.85] in the interruption group vs 3.23 in controls; 95% CI, 3.0 to 3.71, respectively; mean difference, 0.05; 95% CI, −0.10 to 0.19; P = .52). However, nurse workload was significantly higher in the interruption group (mean VAS score, 4.22 vs 3.80; 95% CI, 3.30 to 5.0 vs 2.98 to 4.40; mean difference, 0.41; 95% CI, 0.17 to 0.66; P=.001). Respiratory therapist workload was similar in the 2 groups (mean VAS score, 3.69 in the interruption group vs 3.61 in controls; 95% CI, 2.62 to 4.67 vs 2.70 to 4.33, respectively; mean difference, 0.08; 95% CI, −0.20 to 0.36; P=.57). Adherence with the performance of spontaneous breathing trials and with extubation after a successful spontaneous breathing trial was similar in the 2 groups (eTable 5). Subgroup Analysis Contrary to our hypothesis, surgical and trauma patients randomized to daily interruption had significantly shorter time to successful extubation than those randomized to protocolized sedation alone (6 vs 13 days; hazard ratio 2.55; 95% CI, 1.40 to 4.55), whereas there was no difference among medical patients (9 vs 8 days; hazard ratio, 0.92; 95% CI, 0.72 to 1.18; P value for the interaction=.004). Baseline characteristics and outcomes of the surgical/trauma patients by randomization group are presented in the supplementary appendix (eTables 6 to 8). Table 3. Benzodiazepine and Opioid Administration a Protocolized Sedation and Interruption (n = 214) Midazolam equivalents Total dose/patient, mg Dose/patient/d, mg Dose/patient/d, infusion, mg Dose/patient/d, bolus, mg Infusion, d Boluses/d, No. Fentanyl equivalents Total dose/patient, g Dose/patient/d, g Dose/patient/d, infusion, g Dose/patient/d bolus, g Infusion, d Boluses/d, No. 1087 (4297) 222 (50 to 734) 102 (326) 8 (0 to 86) 101 (325) 6 (0 to 86) 0.99 (5.9) 0 (0 to 0) 5.73 (6.42) 4 (2 to 7) 0.253 (1.145) 0 (0 to 0) 18 997 (59 928) 5286 (1512 to 16 437) 1780 (4135) 550 (50 to 1850) 1664 (4070) 420 (0 to 1725) 116 (215) 0 (0 to 100) 6.44 (6.86) 5 (2 to 9) 2.18 (2.87) 1 (0 to 4) Protocolized Sedation (n = 209) 1038 (4592) 237 (57 to 599) 82 (287) 0 (0 to 50) 82 (287) 0 (0 to 50) 0.49 (2.65) 0 (0 to 0) 5.58 (5.91) 4 (2 to 7) 0.177 (0.808) 0 (0 to 0) 13 532 (23 219) 5936 (2056 to 15 236) 1070 (2066) 260 (0 to 1400) 984 (2002) 80 (0 to 1260) 86 (169) 40 (0 to 150) 6.61 (6.20) 5 (3 to 9) 1.79 (2.67) 0 (0 to 3) Measure of Effect, Mean Difference (95% CI) 48.4 (−804.4 to 901.2) P Value .91 19.23 (2.37 to 37.07) .04 19.22 (1.92 to 36.53) .03 0.50 (0.23 to 0.76) ⬍.001 0.15 (−1.04 to 1.33) .81 0.077 (0.020 to 0.134) .007 5464.6 (−3236.0 to 14 165.2) .22 709.3 (522.0 to 897.7) ⬍.001 679.7 (495.3 to 864.1) ⬍.001 30.13 (19.15 to 41.11) ⬍.001 −0.17 (−1.42 to 1.09) .79 0.395 (0.239 to 0.551) ⬍.001 Conversion factors: For conversion of lorazepam to midazolam, 1 mg midazolam=0.5 mg lorazepam. For conversion of opioids to fentanyl equivalents, 10 mg morphine=2 mg hydromorphone=0.1 mg fentanyl. a Doses are presented as mean (SD) in the first row and median (interquartile range) in the second row. 1990 JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS COMMENT In this multicenter randomized trial, we found that among mechanically ventilated patients receiving continuous sedation, the combined use of protocolized sedation and daily sedative interruption did not improve on the clinical outcomes observed with protocolized sedation alone. Patients in the daily interruption group received more opioids and benzodiazepines, and selfassessed nursing workload was higher for patients in the daily interruption group than the control group; however, these findings are of uncertain clinical importance. Our results contrast with those of 2 earlier trials supporting daily interruption of sedative infusions in mechanically ventilated adults.5,21 In the original single-center trial comparing daily interruption with usual care in 128 mechanically ventilated patients receiving sedative and opioid infusions, daily interruption was associated with shorter durations of mechanical ventilation and ICU stay and less neuroimaging.5 In that trial, research personnel were always present for sedation interruption and had decisional authority regarding resumption of infusions. In a 4-center study, investigators randomized 336 ICU patients receiving ventilation to daily interruption (with up to 4 hours of monitoring by research personnel) or to usual care without a sedation protocol or additional monitoring.21 Patients assigned to daily interruption had shorter durations of mechanical ventilation and ICU and hospital stay; however, unintentional extubation occurred more frequently. Our study is distinct from these earlier trials. First, we compared a sedation strategy adding daily interruption to a control group strategy of protocolized sedation that targeted light sedation, which is likely superior to “usual care” of an earlier era. Second, in this pragmatic trial, sedation was not directed by research staff but was managed by bedside ICU staff with their usual patient assignments, according to well-tested study protocols. Third, the multicenter design reflects actual prac- tice in ICUs with variable workloads and ICU staffing models. Finally, we enrolled surgical patients in addition to medical patients; in this small prespecified subgroup, daily interruption was unexpectedly associated with shorter time to extubation. The potential benefit of nursedirected sedation titration protocols to minimize sedation is recognized,6 although early trials testing this strategy have been conflicting. A nursedirected sedation protocol compared with usual care among 322 medical patients resulted in shorter durations of mechanical ventilation and ICU and hospital stay.4 In contrast, no clinical benefits were found with a different nurse-directed protocol in another center, potentially related to advancedpractice nurses managing ventilators who were already routinely using sedation-minimization strategies.7 The effectiveness of any new intervention to minimize sedation likely depends on the local usual care. In this trial, adherence with sedation interruption of 72% compares favorably with that achieved in previous trials (ranging from 25% to 70%) when research personnel were not managing patient sedation.16,22,23 Reluctance to interrupt sedation infusions is expressed clearly in clinician surveys and practice audits.11,12,24 Common clinical concerns include the potential for patient discomfort, respiratory distress, patient safety, and additional workload.13,14,25,26 These reservations may reflect our unexpected findings of greater opioid and benzodiazepine doses, more bolus doses, and greater nurse workload among patients in the daily interruption group. Strengths of this trial, in addition to the multicenter pragmatic design, include a broad mix of patients and an assessment of perceived additional nursing workload associated with daily sedation interruption. This trial also has limitations. Blinding of caregivers was not feasible, we did not screen for drug withdrawal, and our results may not be applicable to patients receiving shorteracting agents such as propofol or dex- ©2012 American Medical Association. All rights reserved. medetomidine or to patients requiring deeper levels of sedation. In conclusion, for critically ill patients receiving mechanical ventilation, when nurses implemented a sedation protocol that targeted light sedation, daily sedation interruption did not reduce the duration of mechanical ventilation, offered no additional benefits for patients, and may have increased both sedation and analgesic use and nurse workload. Published Online: October 17, 2012. doi:10.1001 /jama.2012.13872 Author Affiliations: Division of Critical Care, Department of Medicine and Interdepartmental Mount Sinai Hospital and University of Toronto, Toronto, Canada (Dr Mehta); Department of Pharmacy and Medicine, Mount Sinai Hospital and University of Toronto, Toronto, Canada (Dr Burry); Departments of Medicine, Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Canada (Drs Cook and Meade); St Joseph’s Healthcare, Hamilton, Canada (Dr Cook); Clinical Epidemiology Program, Ottawa Hospital Research Institute and Faculty of Medicine, University of Ottawa, Ottawa, Canada (Dr Fergusson); Mount Sinai Hospital, Toronto (Ms Steinberg); Toronto General Hospital, Division of Respirology, Interdepartmental Division of Critical Care, Faculty of Medicine, University of Toronto (Dr Granton); Interdepartmental Division of Critical Care and Department of Medicine, University Health Network and University of Toronto (Dr Herridge); Interdepartmental Division of Critical Care Medicine, and Division of Respirology, Department of Medicine, University Health Network and Mount Sinai Hospital, University of Toronto (Dr Ferguson); School of Pharmacy, Northeastern University, Boston, Massachusetts (Dr Devlin); Department of Medicine, Long Beach Memorial Medical Center, Long Beach, California (Dr Tanios); Division of Critical Care Medicine and Center for Health Evaluation and Outcome Sciences, St. Paul’s Hospital and University of British Columbia, Vancouver, Canada (Dr Dodek); Departments of Medicine and Critical Care Medicine, Sunnybrook Hospital, Toronto (Dr Fowler); Interdepartmental Division of Critical Care Medicine (Drs Fowler and Burns) and Institute of Health Policy Management and Evaluation (Dr Burns), University of Toronto; Keenan Research Centre and the Li Ka Shing Knowledge Institute, St Michael’s Hospital, Toronto (Dr Burns); Departments of Anesthesiology and Critical Care, University of Alberta Hospital, Edmonton, Canada (Dr Jacka); Section of Critical Care, Department of Medicine, Faculty of Medicine, University of Manitoba, Winnipeg, Canada (Dr Olafson); Département de Médecine, Soins Intensifs, Hôpital Maisonneuve Rosemont, Université de Montréal, Montréal, Canada (Dr Skrobik); Department of Critical Care, University of Ottawa, Ottawa, Canada (Dr Hébert); Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa (Ms Sabri); and the Department of Critical Care, Hamilton Health Sciences, Hamilton, Canada (Dr Meade). Author Contributions: Dr Mehta had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Mehta, Burry, Cook, Fergusson, Steinberg, Dodek, Fowler, Burns, Skrobik, Hébert, Meade. Acquisition of data: Mehta, Burry, Cook, Steinberg, Granton, Herridge, Ferguson, Devlin, Tanios, Dodek, Fowler, Burns, Jacka, Olafson, Skrobik, Hébert, Sabri, Meade. JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 1991 DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS Analysis and interpretation of data: Mehta, Burry, Cook, Fergusson, Granton, Ferguson, Devlin, Dodek, Fowler, Burns, Hébert, Sabri, Meade. Drafting of the manuscript: Mehta, Burry, Cook, Fergusson, Fowler, Hébert. Critical revision of the manuscript for important intellectual content: Fergusson, Steinberg, Granton, Herridge, Ferguson, Devlin, Tanios, Dodek, Fowler, Burns, Jacka, Olafson, Skrobik, Sabri, Meade. Statistical analysis: Fergusson, Fowler, Hébert, Sabri. Obtained funding: Mehta, Burry, Cook, Fergusson, Fowler, Meade. Administrative, technical, or material support: Mehta, Cook, Steinberg, Herridge, Tanios, Burns, Jacka, Skrobik. Study supervision: Mehta, Steinberg, Granton, Devlin, Tanios, Dodek, Fowler. Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Funding/Support: Funding was provided by the Canadian Institutes of Health Research. Dr Cook is a Canada Research Chair of the Canadian Institutes for Health Research. Dr Burns holds a Clinician Scientist Phase 2 Award of the Canadian Institutes for Health Research. Dr Fowler is a Clinician Scientist of the Heart and Stroke Foundation (Ontario). Role of the Sponsor: The study sponsor had no role in the design of the study; the collection, analysis, or interpretation of the data; or the writing or approval of the manuscript. Online-Only Material: The eAppendix, 2 eFigures, 8 eTables, and Author Audio Interview are available at http://www.jama.com. Additional Contributions: We thank the Canadian Critical Care Trials Group for their collaboration and the ICU nurses for their dedication to patient care and their support of this study. interruption of sedative infusions in an adult medicalsurgical intensive care unit: randomized controlled trial. J Adv Nurs. 2009;65(5):1054-1060. 10. Yiliaz C, Kelebek Girgin N, Ozdemir N, Kutlay O. The effect of nursing-implemented sedation on the duration of mechanical ventilation in the ICU. Ulus Travma Acil Cerrahi Derg. 2010;16(6):521526. 11. Mehta S, Burry L, Fischer S, et al; Canadian Critical Care Trials Group. Canadian survey of the use of sedatives, analgesics, and neuromuscular blocking agents in critically ill patients. Crit Care Med. 2006; 34(2):374-380. 12. Patel RP, Gambrell M, Speroff T, et al. Delirium and sedation in the intensive care unit: survey of behaviors and attitudes of 1384 healthcare professionals. Crit Care Med. 2009;37(3):825-832. 13. Tanios MA, de Wit M, Epstein SK, Devlin JW. Perceived barriers to the use of sedation protocols and daily sedation interruption: a multidisciplinary survey. J Crit Care. 2009;24(1):66-73. 14. Roberts RJ, de Wit M, Epstein SK, Didomenico D, Devlin JW. Predictors for daily interruption of sedation therapy by nurses: a prospective, multicenter study. J Crit Care. 2010;25(4):660.e1-7. 15. Augustes R, Ho KM. Meta-analysis of randomised controlled trials on daily sedation interruption for critically ill adult patients. Anaesth Intensive Care. 2011;39(3):401-409. 16. Mehta S, Burry L, Martinez-Motta JC, et al; Canadian Critical Care Trials Group. A randomized trial of daily awakening in critically ill patients managed with a sedation protocol: a pilot trial. Crit Care Med. 2008; 36(7):2092-2099. 17. Riker RR, Picard JT, Fraser GL. Prospective evaluation of the Sedation-Agitation Scale for adult critically ill patients. Crit Care Med. 1999;27(7):13251329. 18. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond Agitation-Sedation Scale: validity and reliabil- ity in adult intensive care unit patients. Am J Respir Crit Care Med. 2002;166(10):1338-1344. 19. Bergeron N, Dubois M-J, Dumont M, Dial S, Skrobik Y. Intensive Care Delirium Screening Checklist: evaluation of a new screening tool. Intensive Care Med. 2001;27(5):859-864. 20. Patanwala AE, Duby J, Waters D, Erstad BL. Opioid conversions in acute care. Ann Pharmacother. 2007; 41(2):255-266. 21. Girard TD, Kress JP, Fuchs BD, et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care (Awakening and Breathing Controlled trial): a randomised controlled trial. Lancet. 2008;371(9607): 126-134. 22. Weisbrodt L, McKinley S, Marshall AP, Cole L, Seppelt IM, Delaney A. Daily interruption of sedation in patients receiving mechanical ventilation. Am J Crit Care. 2011;20(4):e90-e98. 23. Ruokonen E, Parviainen I, Jakob SM, et al; Dexmedetomidine for Continuous Sedation Investigators. Dexmedetomidine versus propofol/midazolam for long-term sedation during mechanical ventilation. Intensive Care Med. 2009;35(2):282-290. 24. Burry L, Perreault M, Williamson D, et al. A prospective evaluation of sedative, analgesic, antipsychotic and paralytic practices in Canadian mechanically ventilated adults [abstract]. Proc Am Thorac Soc. 2009;179:A5492. 25. O’Connor M, Bucknall T, Manias E. Sedation management in Australian and New Zealand intensive care units: doctors’ and nurses’ practices and opinions. Am J Crit Care. 2010;19(3):285-295. 26. Burry L, Steinberg M, Kim S, et al; SLEAP Investigators and Canadian Critical Care Trials Group. Clinicians’ perspectives on the use of a sedation protocol or daily sedative interruption in mechanically ventilated patients enrolled in a multicenter sedation trial [abstract]. Intensive Care Med. 2011; 37:S83. REFERENCES 1. Arroliga AC, Thompson BT, Ancukiewicz M, et al; Acute Respiratory Distress Syndrome Network. Use of sedatives, opioids, and neuromuscular blocking agents in patients with acute lung injury and acute respiratory distress syndrome. Crit Care Med. 2008;36 (4):1083-1088. 2. Kollef MH, Levy NT, Ahrens TS, Schaiff R, Prentice D, Sherman G. The use of continuous iv sedation is associated with prolongation of mechanical ventilation. Chest. 1998;114(2):541-548. 3. Ouimet S, Kavanagh BP, Gottfried SB, Skrobik Y. Incidence, risk factors and consequences of ICU delirium. Intensive Care Med. 2007;33(1): 66-73. 4. Brook AD, Ahrens TS, Schaiff R, et al. Effect of a nursing-implemented sedation protocol on the duration of mechanical ventilation. Crit Care Med. 1999; 27(12):2609-2615. 5. Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med. 2000;342(20):1471-1477. 6. Jacobi J, Fraser GL, Coursin DB, et al; Task Force of the American College of Critical Care Medicine (ACCM) of the Society of Critical Care Medicine (SCCM), American Society of Health-System Pharmacists (ASHP), American College of Chest Physicians. Clinical practice guidelines for the sustained use of sedatives and analgesics in the critically ill adult. Crit Care Med. 2002;30(1):119-141. 7. Bucknall TK, Manias E, Presneill JJ. A randomized trial of protocol-directed sedation management for mechanical ventilation in an Australian intensive care unit. Crit Care Med. 2008;36(5):1444-1450. 8. de Wit M, Gennings C, Jenvey WI, Epstein SK. Randomized trial comparing daily interruption of sedation and nursing-implemented sedation algorithm in medical intensive care unit patients. Crit Care. 2008; 12(3):R70. 9. Anifantaki S, Prinianakis G, Vitsaksaki E, et al. Daily 1992 JAMA, November 21, 2012—Vol 308, No. 19 Corrected on November 27, 2012 ©2012 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013 new england journal of medicine The established in 1812 january 3, 2013 vol. 368 no. 1 Transfusion Strategies for Acute Upper Gastrointestinal Bleeding Càndid Villanueva, M.D., Alan Colomo, M.D., Alba Bosch, M.D., Mar Concepción, M.D., Virginia Hernandez-Gea, M.D., Carles Aracil, M.D., Isabel Graupera, M.D., María Poca, M.D., Cristina Alvarez-Urturi, M.D., Jordi Gordillo, M.D., Carlos Guarner-Argente, M.D., Miquel Santaló, M.D., Eduardo Muñiz, M.D., and Carlos Guarner, M.D. A BS T R AC T Background The hemoglobin threshold for transfusion of red cells in patients with acute gastrointestinal bleeding is controversial. We compared the efficacy and safety of a restrictive transfusion strategy with those of a liberal transfusion strategy. Methods We enrolled 921 patients with severe acute upper gastrointestinal bleeding and randomly assigned 461 of them to a restrictive strategy (transfusion when the hemoglobin level fell below 7 g per deciliter) and 460 to a liberal strategy (transfusion when the hemoglobin fell below 9 g per deciliter). Randomization was stratified according to the presence or absence of liver cirrhosis. Results A total of 225 patients assigned to the restrictive strategy (51%), as compared with 61 assigned to the liberal strategy (14%), did not receive transfusions (P<0.001). The probability of survival at 6 weeks was higher in the restrictive-strategy group than in the liberal-strategy group (95% vs. 91%; hazard ratio for death with restrictive strategy, 0.55; 95% confidence interval [CI], 0.33 to 0.92; P = 0.02). Further bleeding occurred in 10% of the patients in the restrictive-strategy group as compared with 16% of the patients in the liberal-strategy group (P = 0.01), and adverse events occurred in 40% as compared with 48% (P = 0.02). The probability of survival was slightly higher with the restrictive strategy than with the liberal strategy in the subgroup of patients who had bleeding associated with a peptic ulcer (hazard ratio, 0.70; 95% CI, 0.26 to 1.25) and was significantly higher in the subgroup of patients with cirrhosis and Child–Pugh class A or B disease (hazard ratio, 0.30; 95% CI, 0.11 to 0.85), but not in those with cirrhosis and Child–Pugh class C disease (hazard ratio, 1.04; 95% CI, 0.45 to 2.37). Within the first 5 days, the portal-pressure gradient increased significantly in patients assigned to the liberal strategy (P = 0.03) but not in those assigned to the restrictive strategy. From the Gastrointestinal Bleeding Unit, Department of Gastroenterology (C.V., A.C., M.C., V.H.-G., C.A., I.G., M.P., C.A.-U., J.G., C.G.-A., C.G.), Blood and Tissue Bank (A.B., E.M.), and the SemiCritical Unit (M.S.), Hospital de Sant Pau, Autonomous University, and Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (C.V., A.C., I.G., C.G.) — all in Barcelona. Address reprint requests to Dr. Villa nueva at Servei de Patologia Digestiva, Hospital de la Santa Creu i Sant Pau, Mas Casanovas, 90. 08025 Barcelona, Spain, or at [email protected]. This article was updated on May 16, 2013, at NEJM.org. N Engl J Med 2013;368:11-21. DOI: 10.1056/NEJMoa1211801 Copyright © 2013 Massachusetts Medical Society. Conclusions As compared with a liberal transfusion strategy, a restrictive strategy significantly improved outcomes in patients with acute upper gastrointestinal bleeding. (Funded by Fundació Investigació Sant Pau; ClinicalTrials.gov number, NCT00414713.) n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 11 The n e w e ng l a n d j o u r na l A cute upper gastrointestinal bleeding is a common emergency condition associated with high morbidity and mortality.1 It is a frequent indication for red-cell transfusion, because acute blood loss can decrease tissue perfusion and the delivery of oxygen to tissues. Transfusion may be lifesaving in patients with massive exsanguinating bleeding. However, in most cases hemorrhage is not so severe, and in such circumstances the safest and most effective transfusion strategy is controversial.2,3 Restricted transfusion strategies may be appropriate in some settings. Controlled trials have shown that for critically ill patients, a restrictive transfusion strategy is at least as effective as a liberal strategy, while substantially reducing the use of blood supplies.4,5 However, these studies excluded patients with gastrointestinal bleeding. Observational studies and small controlled trials have suggested that transfusion may be harmful in patients with hypovolemic anemia,6,7 even in those with gastrointestinal bleeding.8-12 Furthermore, studies in animals suggest that transfusion can be particularly harmful in patients with bleeding from portal hypertensive sources, since restitution of blood volume after hemorrhage can lead to a rebound increase in portal pressure, which is associated with a risk of rebleeding.12-14 We performed a randomized, controlled trial in which we assessed whether a restrictive threshold for red-cell transfusion in patients with acute gastrointestinal bleeding was safer and more effective than a liberal transfusion strategy that was based on the threshold recommended in guidelines at the time the study was designed.15,16 Me thods Study Oversight From June 2003 through December 2009, we consecutively enrolled patients with gastrointestinal bleeding who were admitted to Hospital de la Santa Creu i Sant Pau in Barcelona. Written informed consent was obtained from all the patients or their next of kin, and the trial was approved by the institutional ethics committee at the hospital. The protocol, including the statistical analysis plan, is available with the full text of this article at NEJM.org. No commercial support was involved in the study. All the authors vouch for the integrity and the accuracy of the analysis 12 of m e dic i n e and for the fidelity of the study to the protocol. No one who is not an author contributed to the manuscript. Selection of Patients Patients older than 18 years of age who had hematemesis (or bloody nasogastric aspirate), melena, or both, as confirmed by the hospital staff, were considered for inclusion. Patients were excluded if they declined to undergo a blood transfusion. Additional exclusion criteria were massive exsanguinating bleeding; an acute coronary syndrome, symptomatic peripheral vasculopathy, stroke, transient ischemic attack, or transfusion within the previous 90 days; a recent history of trauma or surgery; lower gastrointestinal bleeding; a previous decision on the part of the attending physician that the patient should avoid specific medical therapy; and a clinical Rockall score of 0 with a hemoglobin level higher than 12 g per deciliter. The Rockall score is a system for assessing the risk of further bleeding or death among patients with gastrointestinal bleeding; scores range from 0 to 11, with a score of 2 or lower indicating low risk and scores of 3 to 11 indicating increasingly greater risk. Study Design Immediately after admission, patients were randomly assigned to a restrictive transfusion strategy or a liberal transfusion strategy. Randomization was performed with the use of computer-generated random numbers, with the group assignments placed in sealed, consecutively numbered, opaque envelopes. Randomization was stratified according to the presence or absence of liver cirrhosis and was performed in blocks of four. Cirrhosis was diagnosed according to clinical, biochemical, and ultrasonographic findings. In the restrictive-strategy group, the hemoglobin threshold for transfusion was 7 g per deciliter, with a target range for the post-transfusion hemoglobin level of 7 to 9 g per deciliter. In the liberal-strategy group, the hemoglobin threshold for transfusion was 9 g per deciliter, with a target range for the post-transfusion hemoglobin level of 9 to 11 g per deciliter. In both groups, 1 unit of red cells was transfused initially; the hemoglobin level was assessed after the transfusion, and an additional unit was transfused if the hemoglobin level was below the n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Tr ansfusion Str ategies for Upper GI Bleeding threshold value. The transfusion protocol was applied until the patient’s discharge from the hospital or death. The protocol allowed for a transfusion to be administered any time symptoms or signs related to anemia developed, massive bleeding occurred during follow-up, or surgical intervention was required. Only prestorage leukocytereduced units of packed red cells were used for transfusion. The volume of a unit ranged from 250 to 320 ml, with a hematocrit of approximately 60%. Hemoglobin levels were measured after admission and again every 8 hours during the first 2 days and every day thereafter. Hemoglobin levels were also assessed when further bleeding was suspected. Further bleeding was defined as hematemesis or fresh melena associated with hemodynamic instability (systolic blood pressure of <100 mm Hg; pulse rate of >100 beats per minute, or both) or a fall in hemoglobin level of 2 g per deciliter or more within a 6-hour period. Further bleeding was considered to indicate therapeutic failure; if the bleeding involved nonvariceal lesions, the patient underwent repeat endoscopic therapy or emergency surgery, whereas in the case of further variceal bleeding, transjugular intrahepatic portosystemic shunting (TIPS) was considered. Complications were defined as any untoward events that necessitated active therapy or prolonged hospitalization. Side effects were considered to be severe if the health or safety of the patient was endangered. Treatments and Follow-up All the patients underwent emergency gastroscopy within the first 6 hours. When endoscopic examination disclosed a nonvariceal lesion with active arterial bleeding, a nonbleeding visible vessel, or an adherent clot, patients underwent endoscopic therapy with injection of adrenaline plus multipolar electrocoagulation or application of endoscopic clips. Patients with peptic ulcer received a continuous intravenous infusion of omeprazole (80 mg per 10-hour period after an initial bolus of 80 mg) for the first 72 hours, followed by oral administration of omeprazole. When portal hypertension was suspected, a continuous intravenous infusion of somatostatin (250 μg per hour) and prophylactic antibiotic therapy with norfloxacin or ceftriaxone were administered at the time of admission and continued for 5 days. Bleeding esophageal varices were also treated with band ligation or with sclerotherapy, and gastric varices with injection of cyanoacrylate. In patients with variceal bleeding, portal pressure was measured within the first 48 hours and again 2 to 3 days later to assess the effect of the transfusion strategy on portal hypertension. Portal pressure was estimated with the use of the hepatic venous pressure gradient (HVPG), as described elsewhere.17 Outcome Measures and Definitions The primary outcome measure was the rate of death from any cause within the first 45 days. Secondary outcomes included the rate of further bleeding and the rate of in-hospital complications. Statistical Analysis We estimated that with 430 patients in each group, the study would have the power to detect a between-group difference in mortality of at least 5 percentage points, assuming 10% mortality in the liberal-strategy group (on the basis of results of previous trials with standard care1,3,18), with the use of a two-tailed test and with alpha and beta values of 0.05 and 0.2, respectively. The statistical analysis was performed according to the intention-to-treat principle. Standard tests were used for comparisons of proportions and means. Continuous variables are expressed as means and standard deviations. Actuarial probabilities were calculated with the use of the Kaplan–Meier method and were compared with the use of the log-rank test. A Cox proportional-hazards regression model was used to compare the two transfusion-strategy groups with respect to the primary and secondary end points, with adjustment for baseline risk factors (see the Supplementary Appendix, available at NEJM.org). The hazard ratios and their 95% confidence intervals were calculated. Data were censored at the time an end-point event occurred, at the patient’s last visit, or at the end of the 45-day follow-up period, whichever occurred first. Prespecified subgroup analyses were performed to assess the efficacy of transfusion strategies according to the source of bleeding (lesions related to portal hypertension or peptic ulcer). All P values are two-tailed. Calculations were performed with the use of the SPSS statistical package, version 15.0 (SPSS). n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 13 The n e w e ng l a n d j o u r na l m e dic i n e Mortality R e sult s Study Patients During the study period, 2372 patients were admitted to the hospital for gastrointestinal bleeding and 1610 were screened. Of these, 41 declined to participate and 648 were excluded; among the reasons for exclusion were exsanguinating bleeding requiring transfusion (in 39 patients) and a low risk of rebleeding (329 patients) (Fig. 1). A total of 921 patients underwent randomization and 32 withdrew or were withdrawn by the investigators after randomization (see Fig. 1 for details), leaving 444 patients in the restrictive-strategy group and 445 in the liberal-strategy group for the intention-to-treat analysis. The baseline characteristics were similar in the two groups (Table 1). A total of 277 patients (31%) had cirrhosis, and the baseline characteristics of the patients in this subgroup were similar in the two transfusionstrategy groups (Table 1). Bleeding was due to peptic ulcer in 437 patients (49%) and to esophageal varices in 190 (21%) (Table 1). Hemoglobin Levels and Transfusion The hemoglobin concentration at admission was similar in the two groups (Table 2). The lowest hemoglobin concentration within the first 24 hours was significantly lower in the restrictive-strategy group than in the liberal-strategy group, as was the daily hemoglobin concentration until discharge (P<0.001). The percentage of patients in whom the lowest hemoglobin level was less than 7 g per deciliter was higher in the restrictive-strategy group than in the liberal-strategy group. The hemoglobin concentration at 45 days was similar in the two groups. A total of 225 patients (51%) in the restrictivestrategy group, as compared with 61 patients (14%) in the liberal-strategy group, received no transfusion (P<0.001). The mean (±SD) number of units transfused was significantly lower in the restrictive-strategy group than in the liberalstrategy group (1.5±2.3 vs. 3.7±3.8, P<0.001), and a violation of the transfusion protocol occurred more frequently in the restrictive-strategy group (in 39 patients [9%] vs. 15 patients [3%], P<0.001) (Table 2). The percentage of patients who received a transfusion of fresh-frozen plasma, the percentage of patients who received a transfusion of platelets, and the total amount of fluid administered were similar in the two groups. 14 of Mortality at 45 days was significantly lower in the restrictive-strategy group than in the liberalstrategy group: 5% (23 patients) as compared with 9% (41 patients) (P = 0.02) (Fig. 2). The risk of death was virtually unchanged after adjustment for baseline risk factors for death (hazard ratio with restrictive strategy, 0.55; 95% confidence interval [CI], 0.33 to 0.92) (Table S4 in the Supplementary Appendix). Among all patients with cirrhosis, the risk of death was slightly lower in the restrictive-strategy group than in the liberalstrategy group (Fig. 2). In the subgroup of patients with cirrhosis and Child–Pugh class A or B disease, the risk of death was significantly lower among patients in the restrictive-strategy group than among those in the liberal-strategy group, whereas in the subgroup of patients with cirrhosis and Child–Pugh class C disease, the risk was similar in the two groups. Among patients with bleeding from a peptic ulcer, the risk of death was slightly lower with the restrictive strategy than with the liberal strategy. Death was due to unsuccessfully controlled bleeding in 3 patients (0.7%) in the restrictivestrategy group and in 14 patients (3.1%) in the liberal-strategy group (P = 0.01). Death was caused by complications of treatment in 3 patients (2 in the liberal-strategy group and 1 in the restrictivestrategy group). In the remaining 44 patients (19 in the restrictive-strategy group and 25 in the liberal-strategy group), hemorrhage was controlled and death was due to associated diseases. Further Bleeding The rate of further bleeding was significantly lower in the restrictive-strategy group than in the liberal-strategy group: 10% (45 patients), as compared with 16% (71 patients) (P = 0.01) (Table 3). The risk of further bleeding was significantly lower with the restrictive strategy after adjustment for baseline risk factors for further bleeding (hazard ratio, 0.68; 95% CI, 0.47 to 0.98) (Table S4 in the Supplementary Appendix). In addition, the length of hospital stay was shorter in the restrictive-strategy group than in the liberalstrategy group. In the subgroup of patients with cirrhosis, the risk of further bleeding was lower with the restrictive transfusion strategy than with the liberal transfusion strategy among patients with Child–Pugh class A or B disease and was similar n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Tr ansfusion Str ategies for Upper GI Bleeding in the two groups among patients with Child– Pugh class C disease (Table 3). Among patients with bleeding from esophageal varices, the rate of further bleeding was lower in the restrictivestrategy group than in the liberal-strategy group (11% vs. 22%, P = 0.05). Rescue therapy with balloon tamponade or with transjugular intrahepatic portosystemic shunt was required less frequently in the restrictive-strategy group than in the liberalstrategy group. A baseline hepatic hemodynamic study was performed in 86 patients in the restrictive-strategy group and in 89 in the liberal-strategy group, and it was repeated 2 to 3 days later in 74 and 77 patients, respectively, to assess changes. Patients in the liberal-strategy group had a significant increase in the mean hepatic venous pressure gradient between the first hemodynam ic study and the second (from 20.5±3.1 mm Hg to 21.4±4.3 mm Hg, P = 0.03). There was no significant change in mean hepatic venous pressure gradient in the restrictive-strategy group during that interval. Among patients with bleeding from a peptic ulcer, there was a trend toward a lower risk of further bleeding in the restrictive-strategy group (Table 3). Emergency surgery to control further bleeding was required less frequently in the restrictive-strategy group than in the liberalstrategy group (2% vs. 6%, P = 0.04). Adverse Events 1610 Patients were screened 648 Were excluded 962 Were eligible 41 Declined to participate 921 Underwent randomization 461 Were assigned to restrictive strategy 460 Were assigned to liberal strategy 17 Withdrew 444 Were included in analysis 15 Withdrew 445 Were included in analysis Figure 1. Screening, Randomization, and Follow-up. During the study period, 1610 patients with gastrointestinal bleeding were screened, and 648 patients were excluded. The reasons for exclusion included massive exsanguinating bleeding requiring transfusion before randomization (39 patients) and a low risk of rebleeding (329 patients). A low risk of rebleeding was defined as a clinical Rockall score of 0 and hemoglobin levels higher than 12 g per deciliter. (The Rockall score is a system for assessing the risk of further bleeding or death among patients with gastrointestinal bleeding; scores range from 0 to 11, with higher scores indicating greater risk.) Patients were also excluded if they declined blood transfusion (14 patients); other exclusion criteria were an acute coronary syndrome (58), symptomatic peripheral vasculopathy (12), stroke or transient ischemic attack (7), or transfusion (10) within the previous 90 days; lower gastrointestinal bleeding (51); pregnancy (3); a recent history of trauma or surgery (41); a decision by the attending physician that the patient should avoid medical therapy (9); or inclusion in this study within the previous 90 days or inclusion more than twice (75). A total of 921 patients underwent randomization, of whom 32 were withdrawn: 23 were found to be ineligible, 5 had major protocol violations, and 4 decided to withdraw from the study. The overall rate of complications was significantly lower in the restrictive-strategy group than in the liberal-strategy group (40% [179 patients] vs. 48% [214 patients], P = 0.02), as was the rate of serious adverse events (Table S5 in the Supplementary Appendix). Transfusion reactions and cardiac events, mainly pulmonary edema, occurred more frequently in the liberal-strategy group (Table 3). The rates of other adverse events, such as acute kidney injury or bacterial infections, did not differ significantly between the groups (Table S5 in a liberal transfusion strategy, in which the hemothe Supplementary Appendix). globin threshold was 9 g per deciliter. The most relevant finding was the improvement in survival Discussion rates observed with the restrictive transfusion We found that among patients with severe acute strategy. This advantage was probably related to upper gastrointestinal bleeding, the outcomes a better control of factors contributing to death, were significantly improved with a restrictive such as further bleeding, the need for rescue transfusion strategy, in which the hemoglobin therapy, and serious adverse events. All these facthreshold was 7 g per deciliter, as compared with tors were significantly reduced with the restricn engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 15 The n e w e ng l a n d j o u r na l of m e dic i n e Table 1. Baseline Characteristics of the Patients.* Restrictive Strategy (N = 444) Characteristic Liberal Strategy (N = 445) P Value In-hospital bleeding — no. (%)† 20 (5) 30 (7) 0.19 Rockall score‡ 5.3±2.0 5.4±1.7 0.18 228/444 (51) 209/445 (47) 0.20 Source of bleeding — no./total no. (%) Peptic ulcer Location 0.95 Gastric Duodenal Stomal 76/228 (33) 71/209 (34) 143/228 (63) 131/209 (63) 9/228 (4) 7/209 (3) Stigmata 0.93 Active bleeding Visible vessel Gastroesophageal varices Mallory–Weiss tears 35/228 (15) 33/209 (16) 127/228 (56) 119/209 (57) 101/444 (23) 109/445 (24) 0.58 25/444 (6) 30/445 (7) 0.49 Erosive gastritis or esophagitis 38/444 (9) 29/445 (7) 0.26 Neoplasms 16/444 (4) 20/445 (4) 0.50 Other 36/444 (8) 48/445 (11) Cirrhosis — no. (%) Alcoholic cause — no./total no. (%) 139 (31) 138 (31) 0.94 63/139 (45) 62/138 (45) 0.49 37/139 (27) 30/138 (22) Child–Pugh class — no./total no. (%)§ A 0.57 B 76/139 (55) 79/138 (57) C 26/139 (19) 29/138 (21) 20.1±4.4 20.6±5.2 HVPG — mm Hg¶ 0.61 Causes of bleeding — no./total no. (%) 93/139 (67) 97/138 (70) 0.60 Gastric varices Esophageal varices 8/139 (6) 12/138 (9) 0.36 Peptic lesions 21/139 (15) 18/138 (13) 0.73 *Plus–minus values are means ±SD. †Among patients with in-hospital bleeding, 16 (7 in the restrictive-strategy group and 9 in the liberal-strategy group) were admitted to the intensive care unit with sepsis or for pressure support. ‡The Rockall score is a system for assessing the risk of further bleeding or death among patients with gastrointestinal bleeding; scores range from 0 to 11, with higher scores indicating higher risk. § Child–Pugh class A denotes good hepatic function, class B intermediate function, and class C poor function. The mean Model for End-Stage Liver Disease (MELD) score among patients in all Child–Pugh classes (on a scale from 6 to 40, with higher values indicating more severe liver disease) was 11.9±7 in the restrictive-strategy group and 12.1±6 in the liberal-strategy group (P = 0.95). ¶Portal pressure was measured with the use of the hepatic venous pressure gradient (HVPG), which is the difference between the wedged and free hepatic venous pressures. Measurements were performed within the first 48 hours in 175 patients with variceal bleeding (86 in the restrictive-strategy group and 89 in the liberal-strategy group). tive strategy. Our results are consistent with those from previous observational studies and randomized trials performed in other settings, which have shown that a restrictive transfusion strategy did not increase,5,19 and even de- 16 creased,4,20 the mortality observed with a liberal transfusion strategy. Current international guidelines recommend decreasing the hemoglobin threshold level for transfusion in patients with gastrointestinal n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Tr ansfusion Str ategies for Upper GI Bleeding Table 2. Hemoglobin Levels, Transfusions, and Cointerventions.* Restrictive Strategy (N = 444) Liberal Strategy (N = 445) At admission 9.6.±2.2 9.4±2.4 0.45 Lowest value during hospital stay 7.3±1.4 8.0±1.5 <0.001 9.2±1.2 10.1±1.0 <0.001 11.6±1.7 11.7±1.8 0.67 Variable P Value Hemoglobin level — g/dl At discharge† At day 45 Patients with lowest hemoglobin <7 g/dl — no. (%) 202 (45) 81 (18) <0.001 Patients with lowest hemoglobin >9 g/dl — no. (%) 55 (12) 67 (15) 0.28 219 (49) 384 (86) <0.001 671 1638 <0.001 1.5±2.3 3.7±3.8 <0.001 0 3 <0.001 Red-cell transfusion Any — no. of patients (%) Units transfused — no. Total‡ Mean/patient Median Range 0–19 0–36 1.2±1.8 2.9±2.2 <0.001 Transfusion not adjusted to hemoglobin level — no. of patients (%)¶ 35 (8) 12 (3) 0.001 Major protocol violation — no. of patients (%)‖ ║ 39 (9) 15 (3) <0.001 15 15 During index bleeding§ Duration of storage of red cells — days** Median Range 0.95 1–40 1–42 Fresh-frozen plasma transfusion — no. of patients (%)†† 28 (6) 41 (9) 0.13 Platelet transfusion — no. of patients (%)‡‡ 12 (3) 19 (4) 0.27 5491±3448 5873±4087 0.19 86 (19) 93 (21) 0.62 Crystalloids administered within first 72 hr — ml Receipt of colloids — no. of patients (%) * Plus–minus values are means ±SD. † The average difference in the daily hemoglobin level between the restrictive-strategy group and the liberal-strategy group was 1.0±1.3 g per deciliter, from the time of admission to discharge. ‡ Included are all red-cell transfusions received from the time of admission to discharge. § This category refers to the units of red cells transfused before further bleeding. ¶ Transfusions were administered in 31 patients (26 in the restrictive-strategy group and 5 in the liberal-strategy group) because of symptoms or signs (defined as tachycardia, chest pain, or signs of severe hypoxemia) in 14 patients (8 in the restrictive-strategy group and 6 in the liberal-strategy group) because of massive bleeding, and in 2 patients (1 in each group) because of surgery. ‖ In the restrictive-strategy group, 39 patients without signs or symptoms, massive bleeding, or surgery received a transfusion when the hemoglobin level was higher than 7 g per deciliter. In the liberal-strategy group, 15 patients with a hemoglobin level lower than 9 g per deciliter did not receive a transfusion. ** Red cells were stored for up to 42 days. At least 1 unit stored for more than 14 days was administered in 141 of the 219 patients in the restrictive-strategy group (64%) and 253 of the 384 patients in the liberal-strategy group (66%) who received a transfusion. †† Included are all patients who received a transfusion of fresh-frozen plasma from the time of admission to discharge. ‡‡ Included are all patients who received a transfusion of platelets from the time of admission to discharge. bleeding, from 10 g per deciliter15,16 to 7 g per deciliter.3,21 A reduction in the number of transfusions performed may have accounted for the reduction in mortality from gastrointestinal bleed- ing that has been observed in recent years.22,23 However, current guidelines are based on findings from trials of transfusion triggers involving critically ill patients with normovolemic anemia n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 17 The n e w e ng l a n d j o u r na l of m e dic i n e A Survival, According to Transfusion Strategy 100 90 100 Overall Survival (%) 80 99 98 70 97 60 96 95 50 40 93 P=0.02 by log-rank test 92 30 91 20 Liberal strategy 90 0 10 0 Restrictive strategy 94 0 0 5 5 10 15 10 20 15 25 20 30 35 40 45 25 30 35 40 45 399 386 397 383 395 378 394 375 392 372 Days No. at Risk Restrictive strategy Liberal strategy 444 445 429 428 412 407 404 397 401 393 B Death by 6 Weeks, According to Subgroup Subgroup Restrictive Strategy Liberal Strategy Hazard Ratio (95% CI) P Value no. of patients/total no. (%) Overall Patients with cirrhosis Child–Pugh class A or B Child–Pugh class C Bleeding from varices Bleeding from peptic ulcer 23/444 (5) 15/139 (11) 5/113 (4) 10/26 (38) 10/93 (11) 7/228 (3) 0.55 (0.33–0.92) 0.57 (0.30–1.08) 0.30 (0.11–0.85) 1.04 (0.45–2.37) 0.58 (0.27–1.27) 0.70 (0.26–1.25) 41/445 (9) 25/138 (18) 13/109 (12) 12/29 (41) 17/97 (18) 11/209 (5) 0.1 1.0 0.02 0.08 0.02 0.91 0.18 0.26 10.0 Restrictive Strategy Liberal Strategy Better Better Figure 2. Rate of Survival, According to Subgroup. Panel A shows the Kaplan–Meier estimates of the 6-week survival rate in the two groups. The probability of survival was significantly higher in the restrictive-strategy group than in the liberal-strategy group. The gray arrows indicate the day on which data from a patient were censored. The inset shows the same data on an enlarged y axis. Panel B shows the hazard ratios, with 95% confidence intervals, for death by 6 weeks, according to prespecified subgroups. In the subgroup of patients with Child–Pugh class A or B disease, the Model for End-Stage Liver Disease (MELD) score (on a scale from 6 to 40, with higher values indicating more severe liver disease) was 10.3±5 in the restrictivestrategy group and 10.9±5 in the liberal-strategy group (P = 0.41). In the subgroup of patients with Child–Pugh class C disease, the MELD score was 20.6±6 in the restrictive-strategy group and 18.1±5 in the liberal-strategy group (P = 0.11). — trials from which patients with acute bleeding have been excluded.4,5 Transfusion requirements may be different for patients with acute hemorrhage due to factors such as hemodynamic instability or rapid onset of anemia to extremely low hemoglobin levels. The current study addressed the effects of transfusion in this setting. Previous observational studies and small 18 controlled trials have supported the use of a restrictive transfusion strategy for patients with gastrointestinal bleeding.8-11 Our results, which are consistent with the results from those studies, showed that a restrictive strategy significantly reduced the rates of factors related to therapeutic failure such as further bleeding and the need for rescue therapy, as well as reducing the length of n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Tr ansfusion Str ategies for Upper GI Bleeding Table 3. Study Outcomes.* Restrictive Strategy (N = 444) Liberal Strategy (N = 445) Hazard Ratio with Restrictive Strategy (95% CI) P Value 23 (5) 41 (9) 0.55 (0.33–0.92) 0.02 Overall 45/444 (10) 71/445 (16) 0.62 (0.43–0.91) 0.01 Patients with cirrhosis 16/139 (12) 31/138 (22) 0.49 (0.27–0.90) 0.02 12/113 (11) 23/109 (21) 0.53 (0.27–0.94) 0.04 4/26 (15) 8/29 (28) 0.58 (0.15–1.95) 0.33 10/93 (11) 21/97 (22) 0.50 (0.23–0.99) 0.05 Outcome Death from any cause within 45 days — no. (%) Further bleeding — no. of patients/total no. (%) Child–Pugh class A or B Child–Pugh class C Bleeding from esophageal varices Rescue therapies Balloon tamponade 3/139 (2) 11/138 (8) 0.03 TIPS 6/139 (4) 15/138 (11) 0.04 23/228 (10) 33/209 (16) 20/228 (9) 26/209 (12) 4/228 (2) 12/209 (6) 0.04 9.6±8.7 11.5±12.8 0.01 179 (40) 214 (48) 14 (3) 38 (9) 0.35 (0.19–0.65) 0.001 12 (3) 16 (4) 0.74 (0.35–1.59) 0.56 Patients with bleeding from peptic ulcer 0.63 (0.37–1.07) 0.09 Rescue therapies Second endoscopic therapy Emergency surgery No. of days in hospital 0.21 Adverse events — no. (%)† Any‡ Transfusion reactions Fever 0.73 (0.56–0.95) 0.02 Transfusion-associated circulatory overload 2 (<1) 16 (4) 0.06 (0.01–0.45) 0.001 Allergic reactions 1 (<1) 6 (1) 0.16 (0.02–1.37) 0.12 Cardiac complications§ 49 (11) 70 (16) 0.64 (0.43–0.97) 0.04 8 (2) 13 (3) 0.61 (0.25–0.49) 0.27 12 (3) 21 (5) 0.56 (0.27–1.12) 0.07 Acute coronary syndrome¶ Pulmonary edema Pulmonary complications 48 (11) 53 (12) 0.89 (0.59–1.36) 0.67 Acute kidney injury 78 (18) 97 (22) 0.78 (0.56–1.08) 0.13 3 (1) 6 (1) 0.49 (0.12–2.01) 0.33 119 (27) 135 (30) 0.87 (0.63–1.21) 0.41 Stroke or transient ischemic attack Bacterial infections *Plus–minus values are means ±SD. TIPS denotes transjugular intrahepatic portosystemic shunt. †Patients may have had more than one type of adverse event. ‡Included are all patients who had at least one adverse event during the study period. § This category includes patients with acute coronary syndrome, pulmonary edema, or arrhythmias. ¶Unstable angina developed in 13 patients (8 in the restrictive-strategy group and 5 in the liberal-strategy group), and myocardial infarction occurred in 8 patients (all in the liberal-strategy group). stay in the hospital. These harmful effects of transfusion may be related to an impairment of hemostasis. Transfusion may counteract the splanchnic vasoconstrictive response caused by hypovolemia, inducing an increase in splanchnic blood flow and pressure that may impair the formation of clots.24,25 Transfusion may also induce abnormalities in coagulation properties.8,10 Concerns about transfusion have been raised primarily with respect to patients who have cirrhosis with portal hypertension. Experimental studies have shown that restitution of blood volume can induce rebound increases in portal pressure that may precipitate portal hypertensiverelated bleeding.12-14 Clinical studies have also shown that transfusion increases portal pressure n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 19 The n e w e ng l a n d j o u r na l during acute variceal bleeding, an increase that may be prevented with somatostatin.17 In keeping with these observations, we found that the beneficial effect of a restrictive transfusion strategy with respect to further bleeding was observed mainly in patients with portal hypertension. We also observed that despite treatment with somatostatin, patients in the liberal-strategy group had a significant increase in portal pressure during acute variceal bleeding that was not observed in patients in the restrictive-strategy group. This may have accounted for the higher rate of further bleeding with the liberal strategy. We found a reduction in the rate of complications with the restrictive transfusion strategy. This finding is consistent with results from a previous trial involving critically ill adults.4 However, conflicting results have been shown in other settings.5,19 Several factors, such as coexisting conditions or age, may account for this discrepancy. Cardiac complications, particularly pulmonary edema, occurred more frequently with the liberal transfusion strategy, both in the current study and in the trial that involved critically ill adults.4 The higher level of cardiac complications may indicate a higher risk of circulatory overload associated with a liberal transfusion strategy. Other effects of transfusion, such as transfusion-related immunomodulation,26 may increase the risk of complications or death. These are unlikely to have occurred in the current study given the similar incidence of bacterial infections in the two groups and the universal use of prestorage leukocytereduced red cells. Adverse outcomes have also been associated with long storage time of transfused blood.27 In our study, the storage time was similar in the two groups. However, the median duration of storage was 15 days, and storage lesions become apparent after about 14 days.28 Therefore, the fact that there were more transfusions of blood with these long storage times in the liberal-strategy group may have contributed to the worse outcome. Further research is needed to determine whether the use of newer blood may influence the results with respect to the transfusion strategy. We found that a restrictive transfusion strategy significantly decreased the number of units transfused and the percentage of patients who received no transfusions — findings that were also seen in previous trials.4,5,19 The goal of red-cell transfusions is to improve 20 of m e dic i n e the delivery of oxygen to tissues. The safest and most effective transfusion strategy depends not only on the hemoglobin trigger level but also on factors such as coexisting conditions, age, and hemodynamic status.1,3 Consequently, we allowed transfusions to be performed at the discretion of attending physicians when symptoms related to anemia developed, when massive bleeding occurred, or when surgical intervention was required. Transfusions that were not adjusted to the hemoglobin level and violations of the transfusion protocol occurred more often in the restrictivestrategy group than in the liberal-strategy group. However, both these deviations from the protocol occurred in less than 10% of cases. Our trial has several limitations. First, the results cannot be generalized to all patients with acute gastrointestinal bleeding. Patients with a low risk of rebleeding were not included in this study. However, these patients are less likely to require a transfusion. Patients with massive exsanguinating hemorrhage were also excluded from this trial because red-cell transfusion may be lifesaving for them. However, only a minority of eligible patients were excluded for this reason. Second, because we compared two transfusion strategies, the study was not blinded, and this may have introduced a bias. It is unlikely that bias was introduced, however, owing to the objective definition of the primary outcome and the use of a randomized design with concealed assignments. In summary, we found that a restrictive transfusion strategy, as compared with a liberal transfusion strategy, improved the outcomes among patients with acute upper gastrointestinal bleeding. The risk of further bleeding, the need for rescue therapy, and the rate of complications were all significantly reduced, and the rate of survival was increased, with the restrictive transfusion strategy. Our results suggest that in patients with acute gastrointestinal bleeding, a strategy of not performing transfusion until the hemoglobin concentration falls below 7 g per deciliter is a safe and effective approach. Supported in part by the Fundació Investigació Sant Pau. Dr. Guarner reports receiving consulting fees from Sequana Medical. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank the nursing and medical staffs from the Semi-Critical Unit at the Hospital de la Santa Creu i Sant Pau for their cooperation in this study. n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. Tr ansfusion Str ategies for Upper GI Bleeding References 1. Gralnek IM, Barkun AN, Bardou M. Management of acute bleeding from a peptic ulcer. N Engl J Med 2008;359:928-37. 2. Barkun A, Bardou M, Marshall JK. Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003;139:843-57. 3. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2010;152:101-13. 4. Hébert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340:40917. [Erratum, N Engl J Med 1999;340:1056.] 5. Lacroix J, Hébert PC, Hutchison JS, et al. Transfusion strategies for patients in pediatric intensive care units. N Engl J Med 2007;356:1609-19. 6. Malone DL, Dunne J, Tracy JK, Putnam AT, Scalea TM, Napolitano LM. Blood transfusion, independent of shock severity, is associated with worse outcome in trauma. J Trauma 2003;54:898-905. 7. Robinson WP III, Ahn J, Stiffler A, et al. Blood transfusion is an independent predictor of increased mortality in nonoperatively managed blunt hepatic and splenic injuries. J Trauma 2005;58:437-44. 8. Blair SD, Janvrin SB, McCollum CN, Greenhalgh RM. Effect of early blood transfusion on gastrointestinal haemorrhage. Br J Surg 1986;73:783-5. 9. Villarejo F, Rizzolo M, Lopéz E, Domeniconi G, Arto G, Apezteguia C. Acute anemia in high digestive hemorrhage: margins of security for their handling without transfusion of red globules. Acta Gastroenterol Latinoam 1999;29:261-70. (In Spanish.) 10. Hearnshaw SA, Logan A, Palmer KR, Card TR, Travis SPL, Murphy MF. Out- comes following early red blood cell transfusion in acute upper gastrointestinal bleeding. Aliment Pharmacol Ther 2010; 32:215-24. 11. Halland M, Young M, Fitzgerald MN, Inder K, Duggan JM, Duggan A. Characteristics and outcomes of upper gastrointestinal hemorrhage in a tertiary referral hospital. Dig Dis Sci 2010;55:3430-5. 12. Kravetz D, Sikuler E, Groszmann RJ. Splanchnic and systemic hemodynamics in portal hypertensive rats during hemorrhage and blood volume restitution. Gastroenterology 1986;90:1232-40. 13. McCormick PA, Jenkins SA, McIntyre N, Burroughs AK. Why portal hypertensive varices bleed and bleed: a hypothesis. Gut 1995;36:100-3. 14. Castañeda B, Morales J, Lionetti R, et al. Effects of blood volume restitution following a portal hypertensive–related bleeding in anesthetized cirrhotic rats. Hepatology 2001;33:821-5. 15. British Society of Gastroenterology Endoscopy Committee. Non-variceal upper gastrointestinal haemorrhage: guidelines. Gut 2002;51:Suppl 4:iv1-iv6. 16. de Franchis R. Updating consensus in portal hypertension: report of the Baveno III consensus workshop on definitions, methodology and therapeutic strategies in portal hypertension. J Hepatol 2000;33: 846-52. 17. Villanueva C, Ortiz J, Miñana J, et al. Somatostatin treatment and risk stratification by continuous portal pressure monitoring during acute variceal bleeding. Gastroenterology 2001;121:110-7. 18. Jairath V, Hearnshaw S, Brunskill SJ, et al. Red cell transfusion for the management of upper gastrointestinal haemorrhage. Cochrane Database Syst Rev 2010; 9:CD006613. 19. Carson JL, Terrin ML, Noveck H, et al. Liberal or restrictive transfusion in high- risk patients after hip surgery. N Engl J Med 2011;365:2453-62. 20. Marik PE, Corwin HL. Efficacy of red cell transfusion in the critically ill: a systematic review of the literature. Crit Care Med 2008;36:2667-74. 21. de Franchis R. Revising consensus in portal hypertension: report of the Baveno V consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol 2010;53:762-8. 22. Crooks C, Card TIM, West J. Reductions in 28-day mortality following hospital admission for upper gastrointestinal hemorrhage. Gastroenterology 2011;141: 62-70. 23. British Society of Gastroenterology. UK upper GI bleeding audit (http://www .bsg.org.uk/clinical/general/uk-upper-gibleeding-audit.html). 24. Roberts I, Evans P, Bunn F, Kwan I, Crowhurst E. Is the normalisation of blood pressure in bleeding trauma patients harmful? Lancet 2001;357:385-7. 25. Duggan JM. Transfusion in gastrointestinal haemorrhage — if, when and how much? Aliment Pharmacol Ther 2001;15: 1109-13. 26. Vamvakas EC, Blajchman MA. Transfusion-related immunomodulation (TRIM): an update. Blood Rev 2007;21: 327-48. 27. Koch CG, Li L, Sessler DI, et al. Duration of red-cell storage and complications after cardiac surgery. N Engl J Med 2008; 358:1229-39. 28. Kanias T, Gladwin MT. Nitric oxide, hemolysis, and the red blood cell storage lesion: interactions between transfusion, donor, and recipient. Transfusion 2012; 52:1388-92. Copyright © 2013 Massachusetts Medical Society. nejm clinical practice center Explore a new page designed specifically for practicing clinicians, the NEJM Clinical Practice Center, at www.NEJM.org/clinical-practice-center. Find practice-changing research, reviews from our Clinical Practice series, a curated collection of clinical cases, and interactive features designed to hone your diagnostic skills. n engl j med 368;1 nejm.org january 3, 2013 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission. Copyright © 2013 Massachusetts Medical Society. All rights reserved. 21