Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Feb 1.
Published in final edited form as: J Trauma Acute Care Surg. 2016 Feb;80(2):308–312. doi: 10.1097/TA.0000000000000903

The incidence of ARDS and associated Mortality in severe TBI utilizing the Berlin Definition

Imo P Aisiku 1,✉, Jose-Miguel Yamal 2, Pratik Doshi 3, M Laura Rubin 4, Julia S Benoit 5, H Julia Hannay 6, Barbara C Tilley 7, Shankar Gopinath 8, Claudia S Robertson 9
PMCID: PMC4731296  NIHMSID: NIHMS731008  PMID: 26491799

Abstract

Background

The incidence of ARDS in severe TBI is poorly reported. Recently a new definition for ARDS was proposed: the Berlin Definition. The percentage of patients represented by TBI in the Berlin criteria study is limited. This study describes the incidence and associated mortality of ARDS in TBI patients.

Methods

The study was an analysis of the safety of erythropoietin administration and transfusion threshold on the incidence of ARDS in severe TBI patients. Three reviewers independently assessed all patients enrolled in the study for ALI/ARDS using the Berlin and the AECC definitions. A Cox proportional-hazards model was used to assess the relationship between ARDS and mortality and 6 month GOS.

Results

Two hundred patients were enrolled in the study. Twenty one percent (41/200) and 26% (52/200) of patients developed ARDS using the AECC and Berlin definitions respectively with a median time of 3 (interquartile range=3) days after injury. ARDS by either definition was associated with increased mortality (p=0.04) but not with differences in functional outcome as measured by the GOS at 6 months. Adjusted analysis utilizing the Berlin criteria showed an increased mortality associated with ADS (p=0.01)

Conclusion

Severe TBI is associated with an incidence of ARDS ranging from 20-25%. The incidence is comparable between the Berlin and AECC definitions. ARDS is associated with increased mortality in severe TBI patients but further studies are needed to validate these findings.

Level of Evidence

Level II, Epidemiological

Keywords: Traumatic Brain Injury, Acute Respiratory Distress Syndrome, Acute Lung Injury

Background

Traumatic Brain injury (TBI) is a major cause of disability and mortality among individuals in both developed and developing countries1 The acute respiratory distress syndrome (ARDS) and the less severe pulmonary dysfunction, acute lung injury (ALI), are well-described complications of patients who survive the initial traumatic insult. The worldwide incidence of ALI/ARDS in severe TBI has been reported to range from 20% to 30% and subsequent mortality ranges from 28-38%2-4 However, the majority of previous reports are from retrospective or subgroup analysis3,5-9 which have significant limitations in accurately identifying ARDS and associated mortality in severe TBI patients. Furthermore, the variation in incidence and outcomes is likely multifactorial and likely impacted by the definition of ALI/ARDS. Consequently, the impact of ALI/ARDS on severe TBI mortality still remains unclear.

ARDS was originally described by Ashbaugh, et al. in 196710. The American-European Consensus Conference (AECC)11 provided the first consistent definition for ARDS in 1994. The AECC definition was the first structured definition for ARDS but concerns of its reliability and validity have emerged12, which has led to the recent ARDS Berlin definition (Table 1) in 2012.12 The updated Berlin criteria are meant to address the limitations of the AECC criteria.

Table 1. North American European Consensus Conference and Berlin Definition.

AECC Definition Berlin Definition
1. Acute Onset 1. Acute Onset within 7 days
2. Bilateral pulmonary infiltrates consistent with pulmonary edema 2. Bilateral pulmonary infiltrates consistent with pulmonary edema
3. Impaired oxygenation:

  • PaO2/FiO2 ratio < 300 mm Hg → ALI

  • PaO2/FiO2 ratio < 200 mm Hg → ARDS

 3. Impaired Oxygenation

  • PaO2/FiO2 ratio 200 – 300 mm Hg → mild

  • PaO2/FiO2 ratio 100 – 200 mm Hg → moderate

  • PaO2/FiO2 ratio < 100 mm Hg → severe
4. No evidence of left heart failure PAWP ≤ 18 4. Impaired oxygenation not fully explained by cardiac failure
Open in a new tab

An accurate description of ALI/ARDS incidence, predictors, and impact on neurologic functioning is important for clinicians to implement the results of clinical trials in the most appropriate patient population as well as for prognosis and resource allocation. We report the incidence of ARDS and its association with mortality and long-term outcome from a recently completed randomized controlled trial of erythropoietin in the severe TBI population using the newly developed Berlin criteria.

Methods

Study Population

We examined data from the 200 patients enrolled in the Erythropoietin TBI trial13. Briefly, patients with severe brain injury, the motor component of the Glasgow Coma Score (GCS) <5, at least 15 years of age, and enrolled within 6 hours of injury were recruited at two sites in Houston, Texas. Patients were excluded if their Glasgow Coma Scale score was 3 with fixed and dilated pupils, or if they had penetrating trauma, life-threatening systemic injuries, severe preexisting disease, or were pregnant. After enrollment, patients were randomly assigned to Erythropoietin or placebo and the 7 g/dl or the 10 g/dl transfusion threshold in a 2 × 2 factorial design. Local IRB approval was obtained from the participating sites.

Classification of ARDS

Serial blood gases, ventilatory settings, and serial images of the chest x-rays of the 200 patients were presented to three physicians (a pulmonary critical care physician, a neurocritical care physician, and a trauma surgeon). The physicians, blinded to the trial outcomes and the randomized group assignments, were asked to independently review the cases and determine the pulmonary diagnosis using the AECC and Berlin definitions of ARDS. Using the AECC definition the possible classifications were ‘ARDS’ or ‘no ARDS’ and using the Berlin definition the possible classifications were ‘none’, ‘mild’, ‘moderate’, or ‘severe’ ARDS. At least two physicians were required to agree to obtain a consensus diagnosis for each case.

Outcomes

The primary outcome was the incidence of ARDS by the AECC and the Berlin definitions. The initial study design used the AECC criteria; however, the Berlin criteria was introduced during the study period and therefore the Berlin and AECC criteria were applied to all patients. Secondary outcomes included mortality (within 6 months after injury) and the Glasgow Outcome Scale (GOS) score obtained by a structured interview at 6 months post injury. Although all patients were examined for the presence of ARDS using the both criteria we focused on the Berlin criteria for our secondary outcomes.

Statistical Methods

In the original trial, there were no differences in the incidence of ARDS detected between the erythropoietin and placebo groups and between the transfusion threshold groups; thus, pooled analysis of the groups were performed.13 Baseline characteristics among those who did and did not develop incident ARDS were compared using either a Fisher's exact test for categorical variables or a Wilcoxon rank sum test for continuous variables. Multivariable analysis of risk factors for ARDS was assessed using the Cox proportional hazards model. Lasso-penalized Cox regression, with the penalty parameter selected by 5-fold cross-validation, was used for variable selection.14 The proportional hazards assumption was assessed using Schoenfeld residual plots and testing for interaction terms with time. Fisher's exact test was used to assess the association between ARDS and each of the two secondary outcomes, GOS and mortality. A Cox proportional-hazards model was used to assess the relationship between time-dependent incidence of ARDS and mortality, adjusting for fixed-time covariates transfusion threshold group, EPO treatment group, injury severity score, and IMPACT predicted probability of death. The type I error rate (or alpha level) was set at 0.05.

Results

A total of 200 patients were included in the analysis. There were no instances where consensus was not achieved. Twenty one percent (41/200) and 26% (52/200) of patients developed ARDS using the AECC and Berlin definitions respectively with a median time of 3 (interquartile range=3) days after injury. When the mild ARDS group was removed from the Berlin criteria group, the incidence was almost identical, 20% (38/200) and 21% (41/200) in the Berlin and AECC criteria, respectively. The median tidal volume for all 128 patients with this measure was 8.94mL/kg (25th and 75th percentiles, 8.22mL/kg and 9.76mL/kg respectively) prior to the development of ARDS. Patients clinically were managed using lung protective strategies once ARDS was identified clinically. There were no detectable differences in median tidal volume between the ARDS and non-ARDS groups irrespective of ARDS criteria. Demographic and baseline characteristics stratified by ARDS using the Berlin criteria are shown in Table 2. Acuity of illness, as defined by APACHE II and AIS, was associated with incident ARDS (p≤0.01). Initial ER summary GCS was statistically different between those with and without ARDS (p=0.02). However, this difference (GCS 6 vs. GCS 7) is not clinically significant. There were statistically significant differences between the ARDS and non-ARDS groups with respect to male sex and pre hospital hypoxia. The population was 96.2% male in the ARDS group and 83.1% in the non-ARDS group (p=.03) and instances of pre hospital hypoxia were 30.8% in the ARDS group and 15.5% in the non-ARDS groups (p=.02). In a multivariable model for ARDS development, elevated APACHE II and pre-hospital hypoxia were associated with over a 2-fold increase in the hazard and pre-hospital hypotension was associated with decreased hazard of ARDS (Table 3). The overall mortality in the trial was 15.5% (31/200). Among subjects who developed ARDS, mortality was 25.0% (13/52, 95% confidence interval 14%-39%) (Table 4). ARDS by either definition was associated with increased mortality (p=0.04) but not with differences in functional outcome as measured by the GOS at 6 months (Table 5). In the adjusted Cox proportional-hazards analysis ARDS was found to be associated with an increased risk in mortality (HR: 2.560; p=0.01) (Table 6).

Table 2. Baseline Characteristics of the Study Participants using Berlin criteria.

Characteristic Developed ARDS (N=52) Did not develop ARDS (N=148) p-value
Transfusion threshold <10 g/dL, n (%) 28 (53.8) 73 (49.3) 0.63
Age (yr), median (IQR) 29.0 (19.5) 30.0 (23.0) 0.87
Male sex, n (%) 50 (96.2) 124 (83.8) 0.03
Race, n (%)
 Black 8 (15.4) 35 (23.6) 0.44
 White 16 (30.8) 32 (21.6)
 Hispanic 27 (51.9) 76 (51.4)
 Asian 1 (1.9) 5 (3.4)
ER sum GCS
 >8 7 (13.5) 38 (25.7) 0.04
 6-8 21 (40.4) 68 (45.9)
 3-5 24 (46.2) 42 (28.4)
ER sum GCS, median (IQR) 6.0 (5.0) 7.0 (4.0) 0.02
Marshall Classification on CT
 mild diffuse injury 23 (44.2) 66 (44.6) 0.22
 severe diffuse injury 16 (30.8) 30 (20.3)
 mass lesion 13 (25.0) 52 (35.1)
APACHE II, median (IQR) 24.0 (11.0) 19.0 (7.0) <0.003
Abbreviated Injury Severity Score, median (IQR) 32.0 (14.5) 28.0 (10.0) 0.01
Intracranial Pressure (n=197), n (%) 14.0 (12.0) 14.0 (12.0) 0.84
Mechanism of Injury, n (%)
 Assault 7 (13.5) 15 (10.1) 0.93
 Fall/Jump 7 (13.5) 20 (13.5)
 Motor Vehicle 28 (53.8) 88 (59.5)
 Motorcycle 9 (17.3) 22 (14.9)
 Other 1 (1.9) 3 (2.0)
Surgery on admission, n (%) 13 (25.0) 48 (32.4) 0.38
Pre-hospital hypotension, n (%) 7 (13.5) 18 (12.2) 0.81
Pre-hospital hypoxia 16 (30.8) 23 (15.5) 0.02
Time (days) from injury to ALI, median (IQR) 3.0 (3.0)
Open in a new tab

Table 3. Multivariable Cox Proportional Hazards Model of Risk Factors of ARDS.

HR 95% Confidence Interval P value
Transfusion threshold <10 g/dL 1.14 0.65-2.01 0.648
Male sex 3.69 0.88-15.38 0.073
Apache II ≤16 (reference)
 (16-20] 0.54 0.20-1.41 0.206
 (20-25] 1.47 0.64-3.38 0.364
 >25 2.73 1.21-6.14 0.015
Abbreviated Injury Severity Score 1.02 0.97-1.06 0.474
Surgery on Admission 0.69 0.36-1.33 0.268
Pre-hospital Hypoxia 2.08 1.13-3.83 0.018
Pre-hospital Hypotension 0.42 0.18-0.99 0.048
Abbreviated Injury Severity - Thorax 1.25 0.59-2.67 0.565
Open in a new tab

Table 4. Definition of ARDS/ALI and mortality following an ARDS event.

ALI/ARDS Criteria All-cause Mortality
Berlin Criteria 13/52 (25.0%)
 a. Mild ARDS (PaO2/FIO2 200 - 300 mm Hg), n (%) 2/11 (18.2)
 b. Moderate ARDS (PaO2/FIO2 100 - 200 mm Hg), n (%) 3/27 (11.1)
 c. Severe ARDS (PaO2/FIO2 ≤ 100 mm Hg), n (%) 8/14 (57.1)
Open in a new tab

Table 5. Association of developing incident ARDS (based on the Berlin criteria) with GOSS at 6 months and with mortality.

No ARDS ARDS P value
GOSS at 6 months, n 0.39
 Unfavorable 87 35
 Favorable 53 15
Death, n 0.04
 Did not die 130 39
 Died 18 13
Open in a new tab

Table 6. Cox Regression Analysis –Mortality and time-dependent Incidence of ARDS adjusted for fixed-time Pre-specified Covariates.

Mortality Hazard Ratio 95% Confidence Interval p-value
Incidence of ARDS 2.56 1.21 -5.43 0.01
Epo1 regimen (compared to Placebo) 0.98 0.36-2.67 0.97
Epo2 regimen (compared to Placebo) 0.86 0.34-2.15 0.74
Transfusion Threshold <10 g/dL (compared to TT7) 0.99 0.47-2.06 0.97
Injury Severity Score 1.01 0.97-1.05 0.74
IMPACT probability of Death 1.05 1.03-1.07 <.001
Open in a new tab

Discussion

Severe TBI has been associated with ARDS in previous literature. The incidence has varied significantly in different reports and over time. In one retrospective cohort of patients over a 20-year period, the prevalence of ARDS/ALI increased from 2% to 22% and mortality decreased from 13% to 9%.5 Severe TBI and ARDS carries a significant mortality and valid, reliable definitions of ALI and ARDS are essential to conduct epidemiological studies successfully and for evaluating outcomes related to TBI and ALI/ARDS. This study was initiated using the AECC criteria. We evaluated all patients using both the AECC and the Berlin criteria and reported the first prospective and comparative evaluation of the two criteria in severe TBI patients. The incidence of ARDS was similar but slightly higher using the Berlin definition. It is important to note that exclusion of the mild ARDS criteria from the Berlin definition group, which by PaO2/FiO2 ratio is similar to the ALI group in the AECC definition, yields virtually identical incidences. The incidence of ARDS in our study population is consistent with previous reports with ranges from 20-30% irrespective of which criteria are utilized.

The AECC definition was widely adopted by clinicians (Table 1).11,15 The new criteria are very similar to prior criteria with respect to timing, chest imaging, origin of edema, and degree of hypoxemia.12 The differences are highlighted in Table 1. The term ALI was removed and different levels of ARDS were defined to better categorize patients with respect to severity of disease and lung injury and provide a better predictive model for mortality15. The Berlin criteria were associated with increased mortality and increased median duration of mechanical ventilation when tested in 4188 patients with ARDS from 4 multicenter clinical data sets.12,16-19 The applicability of this cohort to severe TBI is questionable. Although 3 of the 4 studies that comprise the patient population for the Berlin criteria meta-analysis reported trauma as etiology for ARDS, none of the studies differentiate between systemic trauma and TBI. Trauma represented approximately 6% or 250 of the over 4000 patients reported and severe TBI would likely represent an even smaller number. There are conflicting reports in the literature of ARDS as an independent predictor of mortality. One prospective trial of 137 severe TBI patients showed increased mortality and worse long term neurologic outcome in the ALI/ARDS group in contrast to a separate retrospective evaluation of 362 severe TBI patients that showed no difference in mortality or disability at discharge.3,7 ARDS was associated with mortality in our cohort. The incidence of ARDS and associated mortality in our population is more consistent with Holland's prospective trial. These two trials represent the only prospective evaluation of ARDS and mortality in severe TBI patients. Our cohort represents one of the largest prospective studies of severe TBI and ARDS and the first to describe the incidence and mortality using the Berlin criteria.

While severe TBI is known to have an increased risk of ARDS, we did not identify any baseline characteristics or clinical variables that were associated with ARDS except for severity of illness, male sex, and pre-hospital hypoxia. Several studies have demonstrated the importance of high tidal volume ventilation in the etiology and the therapeutic benefits of low tidal volume ventilation as a therapeutic strategy (ARDS Net, plus). The ARDSNet trial 20 demonstrating the impact of lower tidal volume on mortality in ARDS patients was published prior to the initiation of our trial. Data showing the tidal volume per predicted body weight in TBI patients was not presented in that trial. In our trial we found there was no detectable difference in tidal volume between the two groups and all patients were ventilated at less than 10cc/kg IBW. This likely represents a trend towards lower tidal volume ventilation strategies for all patients. The mechanism of injury, the radiographic classification of injury, or the need for surgery was not associated with ARDS either.

Severe TBI is associated with an incidence of ARDS ranging from 20-25% and severity of illness may be the most significant predictor of ARDS. ARDS is associated with increased mortality but further studies are needed to validate these findings. Comparison of the Berlin and the AECC definitions did not yield significant differences in incidence or associated mortality attributable to ARDS in severe TBI patients. The lack of statistically significant differences between ARDS diagnostic criteria affirms previous published data on the incidence of ARDS and more importantly allows for comparison of ARDS in future studies with past studies in severe TBI patients.

We suggest future studies using the Berlin criteria may be compared against previous reports using the AECC definition but include the ALI group to represent the mild ARDS patients.

Acknowledgments

NIH Grant: P01-NS38660

Footnotes

Clinical Trials Number: NCT00313716

No conflict of interests for all authors

Author Contributions: Dr. Aisiku had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Study concept and design: C. Robertson, S. Gopinath, H. J. Hannay

Acquisition of data: C. Robertson, S. Gopinath, I. Aisiku, P. Doshi, H.J. Hannay

Analysis and interpretation of data: I. Aisiku, P. Doshi, C. Robertson, J. Hannay, J-M. Yamal, B.Tilley, M.L. Rubin, J. Benoit

Drafting of manuscript: I. Aisiku, P. Doshi, C. Robertson, H.J. Hannay, J-M. Yamal, B. Tilley, J. Benoit

Critical review of manuscript for important intellectual content: I. Aisiku, C. Robertson, S. Gopinath, H.J. Hannay, J.-M. Yamal, B. Tilley, J. Benoit, M.L. Rubin

Statistical analysis: J-M Yamal, B.Tilley, M.L. Rubin, J. Benoit

Obtained funding: C. Robertson, S. Gopinath, H. J. Hannay

Administrative, technical, or material support: C. Robertson, S. Gopinath, I. Aisiku, H.J. Hannay, J-M. Yamal, B. Tilley

Study supervision: C. Robertson, J-M. Yamal

Contributor Information

Imo P. Aisiku, Email: iaisiku@partners.org, Department of Emergency Medicine, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115.

Jose-Miguel Yamal, Email: Jose-Miguel.Yamal@uth.tmc.edu, Division of Biostatistics, University of Texas School of Public Health.

Pratik Doshi, Email: Pratik.B.Doshi@uth.tmc.edu, Assistant Professor, Department of Emergency Medicine and Internal Medicine, University of Texas Health Science Center at Houston.

M. Laura Rubin, Email: Maria.L.Rubin@uth.tmc.edu, Division of Biostatistics, University of Texas School of Public Health.

Julia S. Benoit, Email: Julia.Benoit@times.uh.edu, Department of Basic Vision Sciences, College of Optometry Texas Institute for Measurement Evaluation and Statistics, University of Houston.

H. Julia Hannay, Email: Julia.Hannay@times.uh.edu, Department of Psychology, University of Houston.

Barbara C. Tilley, Email: Barbara.C.Tilley@uth.tmc.edu, Division of Biostatistics, University of Texas School of Public Health.

Shankar Gopinath, Email: shankarg@bcm.edu, Department of Neurosurgery, Baylor College of Medicine.

Claudia S. Robertson, Email: claudiar@bcm.edu, Department of Neurosugery, Baylor College of Medicine.

References

  • 1.Sosin DM, Sacks JJ, Smith SM. Head injury-associated deaths in the United States from 1979 to 1986. JAMA. 1989;262:2251–5. [PubMed] [Google Scholar]
  • 2.Bratton SL, Davis RL. Acute lung injury in isolated traumatic brain injury. Neurosurgery. 1997;40:707–12. doi: 10.1097/00006123-199704000-00009. discussion 12. [DOI] [PubMed] [Google Scholar]
  • 3.Holland MC, Mackersie RC, Morabito D, Campbell AR, Kivett VA, Patel R, Erickson VR, Pittet JF. The development of acute lung injury is associated with worse neurologic outcome in patients with severe traumatic brain injury. J Trauma. 2003;55:106–11. doi: 10.1097/01.TA.0000071620.27375.BE. [DOI] [PubMed] [Google Scholar]
  • 4.Acker JE, 3rd, Pancioli AM, Crocco TJ, Eckstein MK, Jauch EC, Larrabee H, Meltzer NM, Mergendahl WC, Munn JW, Prentiss SM, et al. Implementation strategies for emergency medical services within stroke systems of care: a policy statement from the American Heart Association/American Stroke Association Expert Panel on Emergency Medical Services Systems and the Stroke Council. Stroke. 2007;38:3097–115. doi: 10.1161/STROKEAHA.107.186094. [DOI] [PubMed] [Google Scholar]
  • 5.Rincon F, Ghosh S, Dey S, Maltenfort M, Vibbert M, Urtecho J, McBride W, Moussouttas M, Bell R, Ratliff JK, Jallo J. Impact of acute lung injury and acute respiratory distress syndrome after traumatic brain injury in the United States. Neurosurgery. 2012;71:795–803. doi: 10.1227/NEU.0b013e3182672ae5. [DOI] [PubMed] [Google Scholar]
  • 6.Mascia L, Sakr Y, Pasero D, Payen D, Reinhart K, Vincent JL. Sepsis Occurrence in Acutely Ill Patients (SOAP) Investigators. Extracranial complications in patients with acute brain injury: a post-hoc analysis of the SOAP study. Intensive Care Med. 2008;34:720–7. doi: 10.1007/s00134-007-0974-7. [DOI] [PubMed] [Google Scholar]
  • 7.Salim A, Martin M, Brown C, Inaba K, Browder T, Rhee P, Teixeira PG, Demetriades D. The presence of the adult respiratory distress syndrome does not worsen mortality or discharge disability in blunt trauma patients with severe traumatic brain injury. Injury. 2008;39:30–5. doi: 10.1016/j.injury.2007.06.015. [DOI] [PubMed] [Google Scholar]
  • 8.Miller PR, Croce MA, Kilgo PD, Scott J, Fabian TC. Acute respiratory distress syndrome in blunt trauma: identification of independent risk factors. The American surgeon. 2002;68:845–50. discussion 50-1. [PubMed] [Google Scholar]
  • 9.Contant CF, Valadka AB, Gopinath SP, Hannay HJ, Robertson CS. Adult respiratory distress syndrome: a complication of induced hypertension after severe head injury. Journal of neurosurgery. 2001;95:560–8. doi: 10.3171/jns.2001.95.4.0560. [DOI] [PubMed] [Google Scholar]
  • 10.Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute respiratory distress in adults. Lancet. 1967;2:319–23. [Google Scholar]
  • 11.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;149:818–24. doi: 10.1164/ajrccm.149.3.7509706. [DOI] [PubMed] [Google Scholar]
  • 12.Force ADT, Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, Camporota L, Slutsky AS. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012;307:2526–33. doi: 10.1001/jama.2012.5669. [DOI] [PubMed] [Google Scholar]
  • 13.Robertson CS, Hannay HJ, Yamal JM, Gopinath S, Goodman JC, Tilley BC, Severe TBI Trial Investigators. Baldwin A, Rivera Lara L, Saucedo-Crespo H, Ahmed O, et al. Effect of erythropoietin and transfusion threshold on neurological recovery after traumatic brain injury: a randomized clinical trial. JAMA. 2014;312:36–47. doi: 10.1001/jama.2014.6490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Tibshirani R. The lasso method for variable selection in the Cox model. Statistics in medicine. 1997;16:385–95. doi: 10.1002/(sici)1097-0258(19970228)16:4<385::aid-sim380>3.0.co;2-3. [DOI] [PubMed] [Google Scholar]
  • 15.Bordes J, Lacroix G, Esnault P, Goutorbe P, Cotte J, Dantzer E, Meaudre E. Comparison of the Berlin definition with the American European Consensus definition for acute respiratory distress syndrome in burn patients. Burns. 2014;40:562–7. doi: 10.1016/j.burns.2014.03.004. [DOI] [PubMed] [Google Scholar]
  • 16.Bersten AD, Edibam C, Hunt T, Moran J New Zealand Intensive Care Society Clinical Trials G. Incidence and mortality of acute lung injury and the acute respiratory distress syndrome in three Australian States. Am J Respir Crit Care Med. 2002;165:443–8. doi: 10.1164/ajrccm.165.4.2101124. [DOI] [PubMed] [Google Scholar]
  • 17.Britos M, Smoot E, Liu KD, Thompson BT, Checkley W, Brower RG Institutes of Health Acute Respiratory Distress Syndrome Network Investigators. The value of positive end-expiratory pressure and Fio(2) criteria in the definition of the acute respiratory distress syndrome. Crit Care Med. 2011;39:2025–30. doi: 10.1097/CCM.0b013e31821cb774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Needham DM, Dennison CR, Dowdy DW, Mendez-Tellez PA, Ciesla N, Desai SV, Sevransky J, Shanholtz C, Scharfstein D, Herridge MS, et al. Study protocol: The Improving Care of Acute Lung Injury Patients (ICAP) study. Crit Care. 2006;10:R9. doi: 10.1186/cc3948. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Rubenfeld GD, Caldwell E, Peabody E, Weaver J, Martin DP, Neff M, Stern EJ, Hudson LD. Incidence and outcomes of acute lung injury. N Engl J Med. 2005;353:1685–93. doi: 10.1056/NEJMoa050333. [DOI] [PubMed] [Google Scholar]
  • 20.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. doi: 10.1056/NEJM200005043421801. [DOI] [PubMed] [Google Scholar]

RESOURCES