Ev olving Epidemio log y

and Definitions of the

A c u t e Re s p i r a t o r y D i s t res s
S y n d ro m e an d E a r l y A c u t e
L u ng In jur y
Andrew J. Sweatt, MD, Joseph E. Levitt, MD, MS*

KEYWORDS
 Acute respiratory distress syndrome  ARDS  Acute lung injury  ALI  Epidemiology  Prevention
 Definition  Criteria

KEY POINTS
 Precise understanding of the epidemiology of the acute respiratory distress syndrome (ARDS) is
limited by evolving clinical criteria and lack of an ideal reference standard.
 The recent Berlin Definition of ARDS addressed some limitations of prior definitions and empirically
validated current criteria.
 As per the Berlin Definition, acute lung injury (ALI) is no longer a classification and ARDS severity is
stratified as mild, moderate, and severe based on the PaO2/FiO2 ratio, and predicts mortality better
than past definitions.
 The Lung Injury Prediction (LIPS) score and the Early Acute Lung Injury (EALI) score are novel
criteria for identifying high-risk patients before progression to ARDS.

INTRODUCTION therapies may have significant implications for the

epidemiology of the syndrome across countries
Precise understanding of the epidemiology of the and time periods because of differences in clinical
acute respiratory distress syndrome (ARDS) has practice patterns. This review discusses the
been limited by evolution of the disease criteria evolving epidemiology and definition of ARDS,
over time and lack of an ideal reference standard. and recent efforts to improve recognition of pa-
Defining ARDS by clinical and physiologic param- tients at high risk of developing ARDS and
eters provides feasibility in clinical practice, and enhance the prevention and early treatment of
has the advantage of conceptualizing the syn- acute lung injury.
drome as a common final pathway of lung injury
in response to a variety of inciting causes. How- THE EVOLUTION OF ACUTE RESPIRATORY
ever, eschewing pathologic correlation or other DISTRESS SYNDROME AND LIMITATIONS OF
reference standards contributes to inclusion of CONSENSUS CRITERIA
heterogeneous patient populations with differing
pathology and potentially very different progno- ARDS was first described in a series of 12 patients
ses. Also, defining a syndrome by criteria that in 1967 by Ashbaugh and colleagues,1 who recog-
chestmed.theclinics.com

are, in part, dependent on the institution of specific nized a common pattern of severe respiratory

Disclosures: The authors have no funding sources of conflicts of interest to disclose.

Division of Pulmonary and Critical Care Medicine, Stanford University, 300 Pasteur Drive, Stanford, CA 94305, USA
* Corresponding author.
E-mail address: jlevitt@stanford.edu

Clin Chest Med 35 (2014) 609–624

http://dx.doi.org/10.1016/j.ccm.2014.08.002
0272-5231/14/$ – see front matter Ó 2014 Elsevier Inc. All rights reserved.
610 Sweatt & Levitt

distress, refractory cyanosis, loss of lung compli- In 2011, the ARDS Definition Task Force of the
ance, and diffuse alveolar infiltrates in a variety of European Society of Intensive Care Medicine
clinical contexts including sepsis, pneumonia, convened in Berlin to address limitations of the
aspiration, and major trauma. Similar syndromes prior AECC definition and provide an empirical
of acute respiratory failure had been previously review of current and novel ancillary criteria.10,11
recognized, but only as distinct conditions named Factors limiting practicality and validity of the defi-
for their specific inciting etiology (eg, Da Nang nition were identified as: lack of clear delineation of
lung, shock lung, posttraumatic lung, respirator “acute”; confusion over inclusion of ARDS within
lung). A better understanding of risk factors for the definition of ALI; failure to account for positive
ARDS emerged in the early 1980s, but proposed end-expiratory pressure (PEEP) in assessment
definitions of the syndrome lacked uniformity.2,3 of the PaO2/FiO2 ratio12–16; poor interobserver reli-
In 1994, the American and European Consensus ability in interpretation of bilateral infiltrates on
Conference (AECC) established specific clinical chest radiograph17,18; inadequate sensitivity of
criteria for ARDS and acute lung injury (ALI), a high left atrial pressure for excluding cases of
novel classification defined by similar criteria but ARDS19,20; and absence of requirement of a
requiring less severe oxygenation impairment.4 known risk factor for ARDS. These issues may
The AECC criteria defined ALI and ARDS as acute contribute to misidentification of ARDS, subopti-
respiratory failure with bilateral pulmonary infil- mal stratification of severity of lung injury, and
trates on chest radiograph; a partial pressure of enrollment of a more heterogeneous population
arterial oxygen (PaO2)/fraction of inspired oxygen into clinical trials.
(FiO2) ratio less than 300 for ALI and less than As proposed solutions, the Berlin Definition
200 for ARDS; and absence of clinical evidence (Table 1) specifies that “acute” respiratory failure
of left atrial hypertension or a pulmonary artery oc- must occur within 1 week of predisposing illness,
clusion pressure less than 18 mm Hg. The AECC as supported by observational data revealing
criteria were subsequently widely adopted, that nearly all patients develop ARDS within
providing uniformity for epidemiologic studies, 7 days of an inciting insult.21 The prior ALI classifi-
multicenter clinical trials, and clinical practice cation was eliminated, and instead ARDS is cate-
guidelines. gorized by severity: mild (200 < PaO2/FiO2 300),
Despite offering feasibility and standardization, moderate (100 < PaO2/FiO2 200), and severe
several limitations of the AECC criteria still exist. (PaO2/FiO2 100). This further stratification of
First, the meaning of respiratory failure was not severity below a PaO2/FiO2 ratio of 200 derives
clearly defined. Most multicenter clinical trials from prior evidence that mortality is highest in
have limited enrollment to patients receiving me- the lowest PaO2/FiO2 quartile independent of venti-
chanical ventilation via an endotracheal tube. lator strategy,22,23 and prior trials indicating differ-
However, in the most rigorous epidemiologic ential success of therapies according to the PaO2/
study to date, respiratory failure was interpreted FiO2 ratio.24–26 The Berlin Definition also requires a
to include mechanical ventilation via a noninvasive minimum PEEP of 5 cm H2O for all severity cate-
face mask or endotracheal tube.5 Other investi- gories in recognition of the influence of PEEP on
gators have since expanded interpretation of the the PaO2/FiO2 ratio. The panel also clarified radio-
consensus criteria to include nonmechanically graphic criteria with supporting teaching exam-
ventilated patients and those outside of the inten- ples, and recognized the potential for ARDS and
sive care unit.6–9 Whether respiratory failure is hydrostatic edema to coexist in the new definition.
interpreted as requiring intubation and/or some Because volume overload is common in patients
level of positive pressure ventilation or purely by with ARDS,20 criteria now exclude clinical evi-
PaO2/FiO2 ratio and radiographic criteria has major dence of isolated left atrial hypertension but with-
implications for anticipated incidence and out- out reference to a specific pulmonary artery
comes. A pediatric study of patients in the emer- occlusion pressure. Finally, the Berlin Definition
gency department with acute hypoxic respiratory specifies that use of noninvasive PEEP is allowed
failure, defined as a PaO2/FiO2 less than 300 (using but limited to the mild ARDS category. This inclu-
a PaO2 derived from recorded saturations and sion is in line with increasing use of noninvasive
charted FiO2), found that only 5% of patients sub- ventilation worldwide27 and will likely facilitate
sequently required intubation.8 Another study of further study of noninvasive ventilation for mild
adults admitted to respiratory isolation rooms ARDS, which continues to be a debated area of
outside the intensive care unit demonstrated that research.28–30
patients with ALI (defined by bilateral infiltrates Importantly the Berlin criteria were empirically
and hypoxemia) had similar mortality to those derived and validated based on a pooled cohort
without one or both ALI criteria (12% vs 10%).9 of patients that comprised separate clinical5,16,31,32
 Acute Respiratory Distress Syndrome 611

Table 1
The Berlin Definition of the acute respiratory distress syndrome

Acute Respiratory Distress Syndrome

Timing Within 1 wk of a known clinical insult or new or worsening respiratory symptoms
Chest imaginga Bilateral opacities not fully explained by effusions, lobar/lung collapse, or nodules
Origin of edema 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
Oxygenation
Mild 200 < PaO2/FiO2 300 with PEEP or CPAP 5 cm H2Ob
Moderate 100 < PaO2/FiO2 200 with PEEP or CPAP 5 cm H2O
Severe PaO2/FiO2 100 with PEEP or CPAP 5 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.
b
May be delivered noninvasively for mild acute respiratory distress.
Adapted from Force AD, Ranieri VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin Definition.
JAMA 2012;307(23):2530. Copyright Ó (2012) American Medical Association. All rights reserved.

and physiologic33–35 databases. Of note, 4 pro- The most rigorous population-based study to date
posed ancillary variables (radiographic severity, from Kings County, Washington identified ALI and
respiratory system compliance, level of PEEP, ARDS in patients receiving mechanical ventilation
and exhaled minute ventilation) were dropped via an endotracheal tube or noninvasive face
from the empirical definition because they did not mask, similarly to the Berlin Definition.5 The inci-
enhance the predictive value of the severe ARDS dence of ALI and ARDS was estimated at 78.9
classification. The distribution of ARDS severity in and 58.7 cases per 100,000 patient-years, res-
the cohort according to the Berlin Definition was pectively, with an age-adjusted extrapolated inci-
22% mild, 50% moderate, and 28% severe, with dence of 190,600 cases per year in the United
associated mortality rates of 27%, 32%, and States. Incidence increased with age to a peak of
45%, respectively (Fig. 1). The Berlin Definition 306 cases per 100,000 patient-years among
was more predictive of mortality (area under the 75- to 84-year-olds. Pneumonia (46%) and nonpul-
curve [AUC] 0.58) than the ALI and ARDS classifi- monary sepsis (33%) were the most common risk
cations by AECC criteria. ARDS severity also corre- factors. Overall hospital mortality was 38.5%. The
lated with lung weight and shunt fraction in the incidence of ARDS and ALI in the Kings County
separate physiologic database. study is significantly higher than that reported by
Whether the Berlin Definition gains the same other less rigorous epidemiologic studies using
widespread adoption as the AECC criteria remains AECC criteria before the era of low tidal volume
to be seen. However, the definition is both prag- ventilation (Table 2). Differences may be explained
matic and empirically derived, addresses many by underrecognition of the syndrome owing to
limitations of the prior definition, and will likely inadequate screening mechanisms or retrospec-
improve standardization for clinical care and tive study design, extrapolation from less than
research. For the purposes of this review, unless 1 year of data, and failure to identify all patients at
otherwise specified, further use of ARDS refers risk within the catchment area in the less rigorous
to the Berlin Definition while use of ALI refers to studies. The incidence also appears to be lower in
the AECC criteria inclusive or ARDS, with acute international studies, which may reflect differing
lung injury (as opposed to ALI) more generally demographics and prevalence of risk factors for
referring to the pathophysiology leading to ARDS. ARDS in addition to differences regarding availabil-
ity or use of mechanical ventilation and noninvasive
ventilation, particularly in the elderly.
EPIDEMIOLOGY OF THE ACUTE RESPIRATORY
Multiple investigators have also reported the
DISTRESS SYNDROME
incidence of ALI and ARDS since widespread
Incidence
adoption of lung-protective mechanical ventilation
The current incidence of ARDS, particularly as and other improvements in supportive care have
defined by the Berlin criteria, is not well established. occurred (Table 3). In the most rigorous of these
612 Sweatt & Levitt

Fig. 1. In the cohort used for Berlin Definition validation and empirical derivation, the distribution of acute
respiratory distress syndrome (ARDS) severity and mortality is shown for: (A) the Berlin criteria and (B) the
American and European Consensus Conference (AECC) Consensus criteria. ALI, acute lung injury. (Data from Force
AD, Ranieri VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin definition. JAMA
2012;307(23):2526–33.)

analyses, Li and colleagues36 demonstrated a are not representative of the United States, and
temporal decline in ARDS incidence from 2001 to the Mayo Clinic (in Olmstead County) has been a
2008 in a population-based study in Olmstead leader in efforts to standardize health care delivery
County, Minnesota. Cases were identified using and limit exposure to potential “second hits” along
a well-validated electronic screening tool followed the pathway to ARDS (see the section Toward Pre-
by investigator confirmation. Despite increasing vention and Early Treatment of Acute Lung Injury).
severity of illness and higher rates of risk factors Therefore, these results may not reliably extra-
for ARDS at presentation, the incidence of ARDS polate to the country as a whole. Moreover, the
decreased from 82.4 to 38.9 cases per 100,000 investigators restricted the definition of ARDS to
person-years. Of importance, the incidence of patients receiving mechanical ventilation via an
patients presenting with ARDS (ie, community- endotracheal tube. Although this decision is rea-
acquired ARDS) did not change and the decline sonable, some of the observed decline in ARDS
in ARDS was due entirely to what the investi- incidence may have been attributed to an
gators deemed hospital-acquired ARDS (Fig. 2). increasing number of patients treated with non-
However, the demographics of Olmstead County invasive ventilation that would otherwise have
 Table 2
Incidence of ALI and ARDS before the era of lung-protective ventilation

Authors,Ref. Year Location Study Period Study Design ICU Admissions ALIa Incidence ARDSa Incidence
5
Rubenfeld et al, 2005 Seattle, USA 1999–2000 Prospective, multicenter Unknown 78.9 58.7
Bersten et al,31 2002 Australia 1999 (8 wk) Prospective, multicenter 1977 34 28
Luhr et al,88 1999 Sweden, Denmark 1997 (8 wk) Prospective, multicenter 13,346 17.9 13.5
Valta et al,89 1999 Finland 1993–1995 Retrospective, single center Unknown N/A 4.9
Goss et al,90 2003 USA (ARDSNet) 1996–1999 Prospective, multicenter 25,392/y 22.4–64.2 N/A

Acute Respiratory Distress Syndrome

Arroliga et al,91 2002 Ohio, USA 1996–1999 Retrospective, single HMO Unknown N/A 15.3
Hudson & Steinberg,92 1999 Seattle, USA 1997 Retrospective, multicenter Unknown 18.9 12.6

Abbreviations: ALI, acute lung injury; ARDS, acute respiratory distress syndrome; HMO, health maintenance organization; ICU, intensive care unit; N/A, not applicable.
a
Incidence per 100,000 patient-years.

613
 614
Sweatt & Levitt
Table 3
Incidence of ALI and ARDS in era of lung protective ventilation

Authors,Ref. Year Location Study Period Study Design ICU Admissions ALIa Incidence ARDSa Incidence
Linko et al,93 2009 Finland 2007 (8 wk) Prospective, multicenter 2670 10.6 5.0
Villar et al,43 2011 Spain 2008–2009 Prospective, multicenter 11,363 N/A 7.2
Sigurdsson et al,94 2013 Iceland 1988–2010 Retrospective, single center 1148/y N/A 3.63–9.63
Li et al,36 2011 Minnesota, USA 2001–2008 Retrospective, 2 centers w1000/y N/A 38.9–82.4
Hernu et al,39 2013 France 2012 (6 mo) Prospective, 10 ICUs 3504 N/A 32
Caser et al,40 2014 Brazil 2006–2007 (15 mo) Retrospective, 14 ICUs 7133 N/A 10.1
a
Incidence per 100,000 patient-years.
 Acute Respiratory Distress Syndrome 615

Fig. 2. (A) Trends of community-acquired ARDS incidence from 2001 to 2008 in Olmsted County, Minnesota;
dotted lines represent 95% confidence intervals. (B) Trends of hospital-acquired acute respiratory distress syn-
drome incidence from 2001 to 2008 in Olmsted County, Minnesota; dotted lines represent 95% confidence inter-
vals. ALI, acute lung injury. (From Li G, Malinchoc M, Cartin-Ceba R, et al. Eight-year 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. Reprinted with permission of the American Thoracic Society. Copyright Ó 2011 American
Thoracic Society. Official Journal of the American Thoracic Society.)

been intubated and met criteria for ARDS.27 patients 18%, 51%, and 31% were classified as
Similar dramatic reductions in the incidence of mild, moderate, and severe ARDS, respectively.
ARDS (from 50% to 90%) have also been demon- Extrapolated to the population of Lyon, the inci-
strated in smaller cohorts of trauma and non- dence was estimated at 32 per 100,000 patient-
trauma surgical patients.37,38 These patient years. However, this estimate does not appear to
populations are likely less susceptible to temporal be adjusted for age or gender, and the observation
changes in the use of noninvasive ventilation, sug- period from March to September may not have
gesting a real effect on improvements in support- captured potential seasonal variations in the inci-
ive care. However, these studies lack the rigor of dence of ARDS.
the study by Rubenfeld and colleagues,5 and a A retrospective review of a prospective cohort of
failure to identify all cases of ARDS may explain 14 adult medical and surgical intensive care units
some of the observed decline in incidence. collected over 15 months in Southeast Brazil from
Two recent international studies have estimated October 2006 to December 2007 identified 130
the incidence of ARDS since the introduction of the cases of ARDS by Berlin criteria among 7133 ad-
Berlin Definition. A 6-month prospective study missions (1.8%).40 The cumulative incidence (10.1
conducted within 10 adult intensive care units affil- cases per 100,000 patient-years) was substantially
iated with the Public University Hospital in Lyon, lower than in Lyon, although the methodology for
France screened 3504 critically ill patients and extrapolating incidence was not described. The
identified 278 who met criteria for ALI and ARDS distribution of ARDS severity was 38% mild, 52%
by AECC criteria.39 Of these 278 patients, 18 did moderate, and 10% severe, but the study only
not meet Berlin criteria because of lack of suffi- included patients receiving invasive mechanical
cient PEEP (a similar 12% of patients did not ventilation, which likely reduced the number of pa-
meet ARDS criteria in the original Berlin cohort tients who would have met criteria for mild ARDS.
because PEEP was <5 cm H2O or missing). An
additional 20 patients could not be classified by
Predisposing Conditions
the Berlin Definition because, despite a PaO2/FiO2
of less than 200, they did not meet criteria for mod- The primary etiology of ARDS has remained consis-
erate or severe ARDS because they were receiving tent across studies of established cases of ARDS,
only noninvasive ventilation. Of the remaining 240 with pneumonia (35%–45%) and nonpulmonary
616 Sweatt & Levitt

sepsis (30%–35%) being the most common modifiers. Consistent with other cohorts, sepsis
followed by aspiration, trauma, pancreatitis, and (33%), pneumonia (27%), and shock (19%) were
multiple transfusions.5,39–43 An in-depth review of the most common predisposing conditions on
risk factors for developing ARDS, including atten- presentation among patients who developed ALI.
tion to important subgroups, is discussed else- However, cardiac surgery (15%), traumatic brain
where.44 In contrast to evaluating cases of ARDS injury (12%), aspiration (9%), and acute abdominal
or populations of critically ill patients, the Lung surgery, lung contusion, and multiple fractures (7%
Injury Prevention Study (LIPS) prospectively evalu- each) were also common. Odds ratios (OR) cal-
ated more than 5000 patients presenting to the culated by multivariable regression for risk factors
emergency department with at least one estab- for developing ALI are shown in Table 4.
lished risk factor for ARDS.45 Risk factors were The presence of multiple comorbidities and
separated into predisposing conditions and risk chronic pulmonary disease are associated with a

Table 4
Multivariable regression of risk factors for the development of ARDS

Odds Ratio 95% CI P value

Predisposing Conditions
Shock 2.2 1.2–3.7 .008
Aspiration 2.2 1.1–4.3 .02
Sepsis 1.4 0.9–2.4 .14
Pneumonia 0.3 0.02–1.7 .3
High-risk surgery
Thoracic (noncardiac) 0.9 0.1–3.2 .9
Orthopedic spine 2.1 0.9–4.6 .07
Acute abdomen 2.5 1.1–5.6 .03
Cardiac 3.7 2.0–7.1 <.001
Aortic vascular 5.9 2.5–13.0 <.001
High-risk trauma
Traumatic brain injury 3.6 2.0–6.8 <.001
Smoke inhalation 2.5 0.8–4.1 .44
Near drowning 5.4 0.06–6.6 .50
Lung contusion 1.5 0.6–3.4 .36
Multiple fractures 1.9 0.8–4.1 .12
Risk Modifiers
Male gender 1.0 0.7–1.5 .91
Alcohol abuse 1.7 0.9–2.9 .08
Obesity (BMI >30) 1.8 1.2–2.5 .004
Chemotherapy 1.6 0.6–3.6 .32
Diabetes mellitus 0.6 0.2–1.2 .14
Smoking 1.1 0.7–1.5 .4
Emergency surgery 3.1 1.6–5.9 <.01
Tachypnea (RR >30/min) 2.0 1.1–3.5 .02
SpO2 <95% 1.4 1.0–2.1 .08
FiO2 >0.35 (>4 L/min) 2.8 1.9–4.1 <.001
Hypoalbuminemia 1.6 1.0–2.4 .03
Acidosis (pH <7.35) 1.7 1.1–2.7 .02
Abbreviations: BMI, body mass index (kg/m2); CI, confidence interval; FiO2, fraction of inspired oxygen; RR, respiratory
rate; SpO2, arterial oxygen saturation.
Adapted from Gajic O, Dabbagh O, Park PK, et al. Early identification of patients at risk of acute lung injury: evaluation
of lung injury prediction score in a multicenter cohort study. Am J Respir Crit Care Med 2011;183(4):466. Reprinted with
permission of the American Thoracic Society. Copyright Ó 2011 American Thoracic Society. Official Journal of the Amer-
ican Thoracic Society.
 Acute Respiratory Distress Syndrome 617

higher incidence of acute lung injury.46 Chronic in crude mortality from 35% to 26% during the
alcohol abuse and active and passive tobacco study period.58 However, clinical trials represent
exposure also seem to be predisposing fac- a more highly selected patient population, and
tors.47,48 By contrast, diabetes mellitus may have results from studies of mortality outside of clinical
a protective effect, a relationship that is not well trials are not consistent. A systematic review of 72
explained but has been found in multiple co- studies and 11,425 patients from 1994 to 2006
horts.45,49,50 A recent analysis of the LIPS cohort found a higher overall mortality rate (44%; 95%
separately analyzed type 1 and type 2 diabetes CI 40%–46%) than in the ARDS Network trials,
and controlled for associated comorbidities such but a similar decline in mortality of 1.1% per year
as obesity, acute hyperglycemia, and diabetes- over the study period.59 By contrast, a subsequent
associated medications.51 Diabetes remained systematic review of the same time period found a
protective (adjusted OR 0.75, 95% confidence static mortality rate of 44.0% for observational
interval [CI] 0.59–0.94); moreover, the associa- studies and 36.2% for clinical trials since the intro-
tion was found for both types of diabetes and, in duction of the AECC criteria in 1994.60 The latter
contrast to prior cohorts, in both septic and non- review postulated that disparate findings may be
septic subgroups.51 explained by the inclusion of non-English studies,
Important gender and racial disparities exist in more accurate assignation of study year for high-
regard to prevalence of risk factors for ARDS, mortality studies that primarily enrolled patients
comorbidities, and severity of illness at presenta- before 1994 but were completed afterward, better
tion in populations of patients at risk. However, exclusion of studies with overlapping patient
race and gender are not clear independent risk populations, and inclusion of 27 additional post-
factors for developing ARDS. In an observational 1994 studies.
study of trauma patients, African Americans were The Berlin Definition did predict mortality better
more likely to present with shock and penetrating than AECC criteria in the Task Force’s validation
trauma and to receive massive transfusions, but cohort; however, prognostic performance was
there was no difference in risk-adjusted rates of limited when evaluated in recent external coho-
ALI among whites, African Americans, Hispanics, rts.39,40,61 Despite the improved empiricism of
and Asian/Pacific Islanders.52 Analysis of a large the Berlin Definition, it remains a purely descriptive
national database of patients with blunt trauma clinical phenotype. Significant sources of hetero-
suggested a protective effect of African American geneity within the syndrome remain, and these
race with respect to ARDS incidence.53 In the will continue to challenge accurate risk stratifi-
5000-patient LIPS cohort, African Americans cation. The prognostic implications of different
were more likely to present with pneumonia, inciting causes, pathophysiologic phases, mecha-
sepsis, shock, and increased severity of illness, nisms of lung injury, and responses to mechanical
but had lower rates of developing ALI either with ventilation are unlikely to be adequately accounted
or without adjustment for baseline characteristics for by any pragmatic clinical definition applied to
(adjusted OR 0.66, 95% CI 0.45–0.96).54 Further- heterogeneous populations.62
more, men had higher rates of developing ALI, For example, the predisposing etiology underly-
but not after controlling for baseline imbalances ing ARDS and patient comorbidities are important
between sexes. independent determinants of outcome. Among
ARDS Network study patients enrolled from 1996
through 2005, the rate of in-hospital mortality
Mortality
was highest for sepsis (37%) and pneumonia
Accurate assessment of mortality attributable to (30%) while exceptionally low in trauma (11%),58
ARDS depends on reliable standard diagnostic although patients with severe trauma were likely
criteria. A systematic review of more than 3000 underrepresented in these trials. A multicenter
patients and 100 studies, between 1967 and clinical trial of activated protein C for treatment
1994 (before AECC criteria), reported a widely of ALI that excluded patients with severe sepsis
ranging but stable mortality rate throughout the and an APACHE II score of at least 25 reported
study period (55% overall mortality, 95% CI an overall mortality rate of only 13%,63 highlighting
33%–77%).55 Although established pharmaco- the importance of underlying conditions for pre-
logic treatments do not currently exist,56,57 studies dicting mortality. Mortality rates with ARDS also
of mortality after adoption of the AECC criteria increase linearly with age.5 Body mass index
suggest improved outcomes, likely attributable to below normal appears be an independent risk
improvements in supportive care. A review of factor for mortality, whereas obesity may portend
2451 patients enrolled in clinical trials within the a lower risk of death.64 Higher acute physiology
ARDS Network from 1995 to 2005 found a decline scores, the presence of shock at admission,
618 Sweatt & Levitt

immune incompetence, a longer hospital stay mechanically ventilated patients,75–77 and trans-
before ALI onset, and a shorter stay in the inten- fusion of blood products.78–81 With respect to
sive care unit preceding ALI onset are also associ- blood products, transfusion of plasma-rich prod-
ated with higher mortatilty.41,65,66 ucts such as platelets and fresh frozen plasma
In contrast to the risk of developing ARDS, race carries a higher risk than transfusing red blood
and ethnicity seem to affect the mortality associ- cells,79 and recent efforts to remove female do-
ated with ARDS. In a case series of greater than nors from the plasma donor pool have reduced
300,000 decedents with ARDS compiled by the the incidence of transfusion-related acute lung
National Center for Health Statistics from 1979 to injury.78,80,81 A recent population-based, nested
1996, annual age-adjusted ARDS mortality rates case-control study from the Mayo Clinic in
were consistently higher for men than for women, Rochester, Minnesota identified multiple potential
and for African Americans in comparison with other hospital exposures related to the development
backgrounds.67 Among patients enrolled in more of ARDS.82 In contrast to other studies, these
recent ARDS Network clinical trials, mortality was patients were carefully matched by baseline risk
greater among blacks and Hispanics.68 After factors for developing ARDS, including their LIPS
adjustment for demographics, clinical covariates, score. Ascertainment of exposures was blinded,
and severity of illness, mortality among blacks and the screening window for risk-factor exposure
was not significantly higher than in whites (adjusted was limited to the time from admission to 6 hours
OR 1.25, 95% CI 0.95–1.66) while the relative before developing ARDS for cases and an equiva-
mortality among Hispanics actually increased lent time period (assigned by unblinded statisti-
(adjusted OR 2.0, 95% CI 1.37–2.90). The reason cians) for controls. In this well-designed study,
for increased mortality among Hispanics is not inadequate empirical antimicrobials, hospital-
known, although it deserves further attention given acquired aspiration, transfusion of blood products,
that this finding occurred in patients enrolled in and higher tidal volumes were highly associated
clinical trials where care is presumably more stan- with developing ARDS. Over the 10-year study
dardized. Similarly, in the aforementioned national period, the investigators found a significant
database of patients with blunt trauma, Hispanics decline in rates of exposures to these risk factors
had higher odds of adjusted ARDS-associated that correlated with a decline in the rate of
mortality (OR 1.76, 95% CI 1.15–2.62).53 hospital-acquired ARDS. Changes implemented
Accurate risk stratification of patients with ARDS to improve standardization of care included:
will be important to the success of future clinical implementation of computerized order entry with
trials. Among ARDS Network study patients, the decision support for pneumonia, sepsis, and
addition of baseline plasma biomarkers (inter- transfusions (with decision support for appropriate
lukin-8, soluble tumor necrosis factor receptor 1, antimicrobial delivery and to limit inappropriate
and surfactant protein D) to the APACHE III score transfusions); respiratory therapist driven lung-
more accurately predicted mortality.69 The predic- protective ventilation protocols for mechanically
tive validity and specificity of future ARDS criteria ventilated patients (including use of a validated
may be further improved through: identification of automated surveillance and notification system
novel genetic polymorphisms and biomarkers70; with documented reduced time of exposure to
more reproducible and practical methods for larger tidal volumes); and increased staffing of
measuring vascular permeability and extravas- the medical intensive care unit including the addi-
cular lung water71,72; incorporation of PaO2/FiO2 tion of a 24-hour on-site intensivist.36 The relative
responses to standard ventilator settings at importance of each of these interventions is un-
24 hours13,14; inclusion of non–plain-film imaging clear; however, there is now abundant evidence
such as fibroproliferation on computed tomogra- that in-hospital exposures to potentially modifiable
phy73 and right ventricular strain on risk factors contribute to the incidence of ARDS.
echocardiogram.74 Efforts to implement commonsense protocols to
limit exposures and improve standardization of
care are likely low-risk, high-reward strategies for
TOWARD PREVENTION AND EARLY
reducing the burden of ARDS.
TREATMENT OF ACUTE LUNG INJURY
Improved Supportive Care
Identification and Treatment of Early Acute
Multiple investigators have identified potentially
Lung Injury
modifiable risk factors that likely contribute to
the incidence of ARDS, including delayed early In the 20 years since publication of the AECC
goal-directed therapy and appropriate antibiotics criteria for ALI and ARDS, numerous pharmaco-
in patients with sepsis,49 larger tidal volumes in logic therapies have been tested and failed to
 Acute Respiratory Distress Syndrome 619

show benefit in large multicenter clinical trials, progressing to ALI and outperformed the LIPS
including several by the National Institutes of score in this cohort. An EALI score greater than
Health (NIH)-funded ARDS Network.83–87 This or equal to 2 identified patients who progressed
recognition has led to increased emphasis on to ALI with 89% sensitivity and 75% specificity.
earlier identification of high-risk patients, and the In this cohort (with a 25% prevalence of ALI), this
restructuring of the ARDS Network into the corresponded to positive and negative predictive
Network for the Prevention and Early Treatment values of 53% and 95%.
of Acute Lung Injury (PETAL) forming in 2014. There are substantial differences in design and
However, established criteria or definitions for intent of the EALI and LIPS scores. For some pul-
early acute lung injury (EALI) to use as inclusion monary specific predisposing conditions and risk
criteria for future trials are lacking. factors included in the LIPS (ie, pneumonia, aspi-
In a large multicenter prospective study of the ration, SpO2 <95% or FiO2 >35%), the distinction
more than 5000 patients, the United States Critical between prevention and early identification may
Illness and Injury Trials (USCIIT) group developed be semantic. However, the distinction is real for
and validated the LIPS score for the early identifi- others (nonpulmonary sepsis, high-risk elective
cation of high-risk patients.45 The LIPS score surgery, comorbidities) and may affect not only
targeted patients presenting to the emergency the incidence but also the time of progression to
department with at least one established risk ALI. The EALI score was derived in patients with
factor for ARDS, and was derived from variables at least some evidence of early bilateral opacities
present within the first 6 hours of presentation. on chest radiography. This requirement increased
Risk factors were divided into predisposing condi- the baseline incidence of developing ALI to 25%
tions (shock, sepsis, aspiration, pneumonia, or compared with 8% in the LIPS cohort, but may
high-risk surgery or trauma) and risk modifiers also limit sensitivity of identifying patients who
(obesity, alcohol abuse, diabetes, hypoalbumine- progress rapidly without an interval qualifying
mia, acidosis, tachypnea, and oxygen supplemen- chest radiograph. Furthermore, the success of
tation). The overall incidence of ARDS in the cohort the EALI score likely derives from the longitudinal
was 7%. The LIPS score showed good discrimina- evaluation of physiologic variables for potentially
tion (AUC 0.80) and calibration, with rates of ARDS up to 6 hours before the onset of ALI. By contrast,
ranging from 1% for a LIPS score of 1 or less to the LIPS score targeted identifying high-risk pa-
36% for a LIPS score of 8 or more. A cutoff of a tients based on variables present within the first
LIPS score greater than 4 had the best overall 6 hours of admission. It is logical that a scoring
discrimination with sensitivity of 69% and speci- system identifying patients at presentation would
ficity of 78%; however, it still identified a relatively require consideration of multiple baseline risk fac-
low-risk population with a positive predictive value tors and risk modifiers, whereas criteria for identi-
of only 18%. fying early but existing lung injury would be more
In a prospective single-center study, the heavily influenced by the acute pulmonary phy-
authors’ research group empirically derived siology predicting impending respiratory failure.
criteria for EALI in patients presenting to the emer- Selection of criteria to identify appropriate target
gency department with bilateral opacities on chest populations will largely depend on the nature of the
radiograph in the absence of isolated left atrial intervention. For identifying patients to target for
hypertension. Clinical variables were collected strategies to prevent exposure to modifiable risk
longitudinally for up to 72 hours or 6 hours before factors, sensitive criteria generalizable to multiple
meeting criteria for ALI (defined by AECC criteria patient populations, such as the LIPS score, may
while receiving positive pressure ventilation via be ideal. However, clinical trials testing novel ther-
face mask or endotracheal tube). Study investi- apies may wish to target higher-risk patients with
gators performed bedside titration of supple- existing physiologic and radiographic surrogate
mental oxygen on a daily basis to record the end points of lung injury. For example, a theoretic
level required to maintain arterial oxygen satura- clinical trial of an intervention with a predicted 50%
tion (SpO2) greater than 90%. Sixty-two of 256 reduction in the rate of progression to ARDS would
patients (25%) developed ALI. Oxygen require- need 400 patients to be appropriately powered if
ment, respiratory rate, and baseline immune sup- the study population has a baseline 20% risk of
pression were the only independent predictors of developing ARDS (with a 2-tailed a of .05). How-
progression to ALI. A pragmatic 3-component ever, if the baseline risk is 40%, only 160 patients
EALI score (1 point for an oxygen requirement would be needed. As with studies of ARDS,
>2–6 L/min or 2 points for >6 L/min; and 1 point markers of lung injury severity to target as surro-
each for a respiratory rate >30 per minute and im- gate end points in clinical trials are not well estab-
mune suppression) accurately identified patients lished, which will challenge the design of early
620 Sweatt & Levitt

exploratory studies and highlights the need for Developing reliable and clinically available biolog-
multicenter coordination for appropriately pow- ical markers of lung injury may help identify more
ered clinical trials in this area. homogenous patient populations, and serve as
At present, the USCITT Group is conducting surrogate markers to facilitate future appropriately
2 multicenter clinical trials for the prevention of powered clinical investigations.
ARDS. The Lung Injury Prevention Study with The ongoing paradigm shift toward prevention
Aspirin (LIPS-A) is a large phase II clinical trial of and early treatment may provide further progress
aspirin versus placebo, targeting 400 patients toward reducing the significant burden of ARDS.
with a LIPS score of at least 4 and a primary Efforts to implement commonsense protocols to
end point of a reduction in the incidence of ALI limit exposures to modifiable risk factors and
while receiving invasive mechanical ventilation improve standardization of care across institutions
(NCT01504867). The Lung Injury Prevention Study are likely to be high-impact strategies to reduce
with Budesonide and Beta Agonists (LIPS-B) is a the incidence of ARDS. The LIPS and EALI scores
smaller phase II study of aerosolized budesonide are novel criteria for identifying high-risk patients
and formoterol versus placebo. This study aims who may particularly benefit from these strategies.
to enroll 40 patients with a LIPS score of at least However, it will be important to continue to refine
4 and a minimal baseline oxygen requirement, and standardize criteria to further understand the
with a primary end point of improvement in the epidemiology of these at-risk populations and to
SpO2/FiO2 ratio (NCT01783821). These trials repre- identify the appropriate target patients for future
sent exciting new strategies to reduce the burden clinical investigations.
of ARDS. This paradigm shift toward prevention
and early treatment is now a major focus of the REFERENCES
NIH through formation of the PETAL Network.
However, with increased emphasis on prevention 1. Ashbaugh DG, Bigelow DB, Petty TL, et al. Acute
and early treatment of ARDS, it will be important respiratory distress in adults. Lancet 1967;
to continue to refine and standardize criteria to 2(7511):319–23.
identify the appropriate target populations and 2. Fowler AA, Hamman RF, Good JT, et al. Adult
inform future clinical trials. respiratory distress syndrome: risk with common
predispositions. Ann Intern Med 1983;98(5 Pt 1):
593–7.
SUMMARY
3. Pepe PE, Potkin RT, Reus DH, et al. Clinical predic-
The AECC criteria for ARDS and ALI have helped tors of the adult respiratory distress syndrome. Am
to standardize patients for epidemiologic studies, J Surg 1982;144(1):124–30.
multicenter clinical trials, and clinical practice 4. Bernard GR, Artigas A, Brigham KL, et al. The
guidelines. However, lack of consistent interpreta- American-European Consensus Conference on
tion of the AECC criteria has limited precise under- ARDS. Definitions, mechanisms, relevant outcomes,
standing of the incidence and epidemiology of and clinical trial coordination. Am J Respir Crit Care
ARDS. Moreover, ARDS remains underrecognized Med 1994;149(3 Pt 1):818–24.
in clinical practice, and accurate epidemiologic 5. Rubenfeld GD, Caldwell E, Peabody E, et al. Inci-
description requires studies with rigorous method- dence and outcomes of acute lung injury. N Engl
ology to reliably identify all cases and relevant J Med 2005;353(16):1685–93.
patients at risk. Given these limitations, improve- 6. Ferguson ND, Frutos-Vivar F, Esteban A, et al.
ments in supportive care apparently have signifi- Clinical risk conditions for acute lung injury in the
cantly reduced the incidence of ARDS in recent intensive care unit and hospital ward: a prospec-
years. tive observational study. Crit Care 2007;11(5):R96.
The recently empirically derived and validated 7. Flori HR, Glidden DV, Rutherford GW, et al. Pediat-
Berlin Definition will help codify the inclusion of ric acute lung injury: prospective evaluation of risk
patients managed with noninvasive positive pres- factors associated with mortality. Am J Respir Crit
sure ventilation and resolve confusion regarding Care Med 2005;171(9):995–1001.
the overlapping criteria of ALI and ARDS. More- 8. Freishtat RJ, Mojgani B, Mathison DJ, et al. Toward
over, establishing mild, moderate, and severe early identification of acute lung injury in the emer-
criteria will help risk-stratify patients for research gency department. J Investig Med 2007;55(8):
purposes. However, ARDS remains a descriptive 423–9.
clinical phenotype representing heterogeneous 9. Quartin AA, Campos MA, Maldonado DA, et al.
inciting causes with likely differing pathology and Acute lung injury outside of the ICU: incidence in
very different prognoses, which will continue to respiratory isolation on a general ward. Chest
challenge clinical trials targeting survival benefit. 2009;135(2):261–8.
 Acute Respiratory Distress Syndrome 621

10. Force AD, Ranieri VM, Rubenfeld GD, et al. Acute 23. Meade MO, Cook DJ, Guyatt GH, et al. Ventilation
respiratory distress syndrome: the Berlin Definition. strategy using low tidal volumes, recruitment ma-
JAMA 2012;307(23):2526–33. neuvers, and high positive end-expiratory pressure
11. Ferguson ND, Fan E, Camporota L, et al. The Berlin for acute lung injury and acute respiratory distress
definition of ARDS: an expanded rationale, justifi- syndrome: a randomized controlled trial. JAMA
cation, and supplementary material. Intensive 2008;299(6):637–45.
Care Med 2012;38(10):1573–82. 24. Briel M, Meade M, Mercat A, et al. Higher vs lower
12. Villar J, Perez-Mendez L, Kacmarek RM. Current positive end-expiratory pressure in patients with
definitions of acute lung injury and the acute respi- acute lung injury and acute respiratory distress
ratory distress syndrome do not reflect their true syndrome: systematic review and meta-analysis.
severity and outcome. Intensive Care Med 1999; JAMA 2010;303(9):865–73.
25(9):930–5. 25. Cesana BM, Antonelli P, Chiumello D, et al. Positive
13. Villar J, Perez-Mendez L, Lopez J, et al. An early end-expiratory pressure, prone positioning, and
PEEP/FIO2 trial identifies different degrees of lung activated protein C: a critical review of meta-ana-
injury in patients with acute respiratory distress lyses. Minerva Anestesiol 2010;76(11):929–36.
syndrome. Am J Respir Crit Care Med 2007; 26. Papazian L, Forel JM, Gacouin A, et al. Neuromus-
176(8):795–804. cular blockers in early acute respiratory distress
14. Villar J, Perez-Mendez L, Blanco J, et al. syndrome. N Engl J Med 2010;363(12):1107–16.
A universal definition of ARDS: the PaO2/FiO2 ratio 27. Esteban A, Frutos-Vivar F, Muriel A, et al. Evolution of
under a standard ventilatory setting–a prospective, mortality over time in patients receiving mechanical
multicenter validation study. Intensive Care Med ventilation. Am J Respir Crit Care Med 2013;188(2):
2013;39(4):583–92. 220–30.
15. Ferguson ND, Kacmarek RM, Chiche JD, et al. 28. Wang S, Singh B, Tian L, et al. Epidemiology of
Screening of ARDS patients using standardized noninvasive mechanical ventilation in acute respira-
ventilator settings: influence on enrollment in a clin- tory failure–a retrospective population-based study.
ical trial. Intensive Care Med 2004;30(6):1111–6. BMC Emerg Med 2013;13:6.
16. Britos M, Smoot E, Liu KD, et al. The value of 29. Agarwal R, Aggarwal AN, Gupta D. Role of noninva-
positive end-expiratory pressure and Fio(2) criteria sive ventilation in acute lung injury/acute respiratory
in the definition of the acute respiratory distress distress syndrome: a proportion meta-analysis.
syndrome. Crit Care Med 2011;39(9):2025–30. Respir Care 2010;55(12):1653–60.
17. Rubenfeld GD, Caldwell E, Granton J, et al. Inter- 30. Zhan Q, Sun B, Liang L, et al. Early use of noninva-
observer variability in applying a radiographic sive positive pressure ventilation for acute lung
definition for ARDS. Chest 1999;116(5):1347–53. injury: a multicenter randomized controlled trial.
18. Meade MO, Cook RJ, Guyatt GH, et al. Interobserver Crit Care Med 2012;40(2):455–60.
variation in interpreting chest radiographs for the 31. Bersten AD, Edibam C, Hunt T, et al, Australian and
diagnosis of acute respiratory distress syndrome. New Zealand Intensive Care Society Clinical Trials
Am J Respir Crit Care Med 2000;161(1):85–90. Group. Incidence and mortality of acute lung injury
19. Ferguson ND, Meade MO, Hallett DC, et al. High and the acute respiratory distress syndrome in
values of the pulmonary artery wedge pressure in three Australian States. Am J Respir Crit Care
patients with acute lung injury and acute respira- Med 2002;165(4):443–8.
tory distress syndrome. Intensive Care Med 2002; 32. Needham DM, Dennison CR, Dowdy DW, et al.
28(8):1073–7. Study protocol: the improving care of acute lung
20. National Heart, Lung, and Blood Institute Acute injury patients (ICAP) study. Crit Care 2006;
Respiratory Distress Syndrome (ARDS) Clinical 10(1):R9.
Trials Network, Wheeler AP, Bernard GR, et al. 33. Bellani G, Guerra L, Musch G, et al. Lung regional
Pulmonary-artery versus central venous catheter metabolic activity and gas volume changes
to guide treatment of acute lung injury. N Engl J induced by tidal ventilation in patients with acute
Med 2006;354(21):2213–24. lung injury. Am J Respir Crit Care Med 2011;
21. Hudson LD, Milberg JA, Anardi D, et al. Clinical 183(9):1193–9.
risks for development of the acute respiratory 34. Terragni PP, Del Sorbo L, Mascia L, et al. Tidal
distress syndrome. Am J Respir Crit Care Med volume lower than 6 ml/kg enhances lung protec-
1995;151(2 Pt 1):293–301. tion: role of extracorporeal carbon dioxide removal.
22. Mercat A, Richard JC, Vielle B, et al. Positive Anesthesiology 2009;111(4):826–35.
end-expiratory pressure setting in adults with acute 35. Terragni PP, Rosboch G, Tealdi A, et al. Tidal hyper-
lung injury and acute respiratory distress syn- inflation during low tidal volume ventilation in acute
drome: a randomized controlled trial. JAMA 2008; respiratory distress syndrome. Am J Respir Crit
299(6):646–55. Care Med 2007;175(2):160–6.
622 Sweatt & Levitt

36. Li G, Malinchoc M, Cartin-Ceba R, et al. Eight-year patients with septic shock: an observational cohort
trend of acute respiratory distress syndrome: a study. Crit Care Med 2008;36(5):1518–22.
population-based study in Olmsted County, Minne- 50. Moss M, Guidot DM, Steinberg KP, et al. Diabetic
sota. Am J Respir Crit Care Med 2011;183(1):59–66. patients have a decreased incidence of acute
37. Ciesla DJ, Moore EE, Johnson JL, et al. Decreased respiratory distress syndrome. Crit Care Med
progression of postinjury lung dysfunction to the 2000;28(7):2187–92.
acute respiratory distress syndrome and multiple 51. Yu S, Christiani DC, Thompson BT, et al. Role of
organ failure. Surgery 2006;140(4):640–7 [discus- diabetes in the development of acute respiratory
sion: 647–8]. distress syndrome. Crit Care Med 2013;41(12):
38. Martin M, Salim A, Murray J, et al. The decreasing 2720–32.
incidence and mortality of acute respiratory 52. Brown LM, Kallet RH, Matthay MA, et al. The influ-
distress syndrome after injury: a 5-year observa- ence of race on the development of acute lung
tional study. J Trauma 2005;59(5):1107–13. injury in trauma patients. Am J Surg 2011;201(4):
39. Hernu R, Wallet F, Thiolliere F, et al. An attempt to vali- 486–91.
date the modification of the American-European 53. Ryb GE, Cooper C. Race/ethnicity and acute respi-
consensus definition of acute lung injury/acute ratory distress syndrome: a National Trauma Data
respiratory distress syndrome by the Berlin defini- Bank study. J Natl Med Assoc 2010;102(10):865–9.
tion in a university hospital. Intensive Care Med 54. Lemos-Filho LB, Mikkelsen ME, Martin GS, et al.
2013;39(12):2161–70. Sex, race, and the development of acute lung
40. Caser EB, Zandonade E, Pereira E, et al. Impact of injury. Chest 2013;143(4):901–9.
distinct definitions of acute lung injury on its inci- 55. Krafft P, Fridrich P, Pernerstorfer T, et al. The acute
dence and outcomes in Brazilian ICUs: prospec- respiratory distress syndrome: definitions, severity
tive evaluation of 7,133 patients*. Crit Care Med and clinical outcome. An analysis of 101 clinical
2014;42(3):574–82. investigations. Intensive Care Med 1996;22(6):
41. Rubenfeld GD, Herridge MS. Epidemiology and 519–29.
outcomes of acute lung injury. Chest 2007;131(2): 56. Calfee CS, Matthay MA. Nonventilatory treatments
554–62. for acute lung injury and ARDS. Chest 2007;
42. Thille AW, Esteban A, Fernandez-Segoviano P, 131(3):913–20.
et al. Comparison of the Berlin definition for acute 57. Levitt JE, Matthay MA. Treatment of acute lung
respiratory distress syndrome with autopsy. Am J injury: historical perspective and potential future
Respir Crit Care Med 2013;187(7):761–7. therapies. Semin Respir Crit Care Med 2006;
43. Villar J, Blanco J, Anon JM, et al. The ALIEN study: 27(4):426–37.
incidence and outcome of acute respiratory distress 58. Erickson SE, Martin GS, Davis JL, et al, NIH NHLBI
syndrome in the era of lung protective ventilation. ARDS Network. Recent trends in acute lung injury
Intensive Care Med 2011;37(12):1932–41. mortality: 1996-2005. Crit Care Med 2009;37(5):
44. Levitt JE, Matthay MA. The utility of clinical predic- 1574–9.
tors of acute lung injury: towards prevention and 59. Zambon M, Vincent JL. Mortality rates for patients
earlier recognition. Expert Rev Respir Med 2010; with acute lung injury/ARDS have decreased over
4(6):785–97. time. Chest 2008;133(5):1120–7.
45. Gajic O, Dabbagh O, Park PK, et al. Early identifi- 60. Phua J, Badia JR, Adhikari NK, et al. Has mortality
cation of patients at risk of acute lung injury: evalua- from acute respiratory distress syndrome
tion of lung injury prediction score in a multicenter decreased over time? A systematic review. Am J
cohort study. Am J Respir Crit Care Med 2011; Respir Crit Care Med 2009;179(3):220–7.
183(4):462–70. 61. Costa EL, Amato MB. The new definition for acute
46. Johnson ER, Matthay MA. Acute lung injury: epide- lung injury and acute respiratory distress syndrome:
miology, pathogenesis, and treatment. J Aerosol is there room for improvement? Curr Opin Crit Care
Med Pulm Drug Deliv 2010;23(4):243–52. 2013;19(1):16–23.
47. Moss M, Bucher B, Moore FA, et al. The role of 62. Phua J, Stewart TE, Ferguson ND. Acute respira-
chronic alcohol abuse in the development of acute tory distress syndrome 40 years later: time to
respiratory distress syndrome in adults. JAMA revisit its definition. Crit Care Med 2008;36(10):
1996;275(1):50–4. 2912–21.
48. Calfee CS, Matthay MA, Eisner MD, et al. Active 63. Liu KD, Levitt J, Zhuo H, et al. Randomized clinical
and passive cigarette smoking and acute lung trial of activated protein C for the treatment of acute
injury after severe blunt trauma. Am J Respir Crit lung injury. Am J Respir Crit Care Med 2008;178(6):
Care Med 2011;183(12):1660–5. 618–23.
49. Iscimen R, Cartin-Ceba R, Yilmaz M, et al. Risk 64. O’Brien JM Jr, Phillips GS, Ali NA, et al. Body mass
factors for the development of acute lung injury in index is independently associated with hospital
 Acute Respiratory Distress Syndrome 623

mortality in mechanically ventilated adults with acute 78. Chapman CE, Stainsby D, Jones H, et al. Ten years
lung injury. Crit Care Med 2006;34(3):738–44. of hemovigilance reports of transfusion-related
65. Brun-Buisson C, Minelli C, Bertolini G, et al. Epi- acute lung injury in the United Kingdom and the
demiology and outcome of acute lung injury in impact of preferential use of male donor plasma.
European intensive care units. Results from the Transfusion 2009;49(3):440–52.
ALIVE study. Intensive Care Med 2004;30(1):51–61. 79. Gajic O, Rana R, Winters JL, et al. Transfusion-
66. Luhr OR, Karlsson M, Thorsteinsson A, et al. The related acute lung injury in the critically ill: prospec-
impact of respiratory variables on mortality in tive nested case-control study. Am J Respir Crit
non-ARDS and ARDS patients requiring mechani- Care Med 2007;176(9):886–91.
cal ventilation. Intensive Care Med 2000;26(5): 80. Eder AF, Herron RM Jr, Strupp A, et al. Effective
508–17. reduction of transfusion-related acute lung injury
67. Moss M, Mannino DM. Race and gender differences risk with male-predominant plasma strategy in the
in acute respiratory distress syndrome deaths in American Red Cross (2006-2008). Transfusion
the United States: an analysis of multiple-cause 2010;50(8):1732–42.
mortality data (1979-1996). Crit Care Med 2002; 81. Toy P, Gajic O, Bacchetti P, et al. Transfusion-
30(8):1679–85. related acute lung injury: incidence and risk
68. Erickson SE, Shlipak MG, Martin GS, et al. Racial factors. Blood 2012;119(7):1757–67.
and ethnic disparities in mortality from acute lung 82. Ahmed AH, Litell JM, Malinchoc M, et al. The role
injury. Crit Care Med 2009;37(1):1–6. of potentially preventable hospital exposures in
69. Calfee CS, Ware LB, Glidden DV, et al. Use of risk the development of acute respiratory distress
reclassification with multiple biomarkers improves syndrome: a population-based study. Crit Care
mortality prediction in acute lung injury. Crit Care Med 2014;42(1):31–9.
Med 2011;39(4):711–7. 83. Ketoconazole for early treatment of acute lung
70. Calfee CS, Eisner MD, Ware LB, et al. Trauma- injury and acute respiratory distress syndrome: a
associated lung injury differs clinically and biologi- randomized controlled trial. The ARDS Network.
cally from acute lung injury due to other clinical JAMA 2000;283(15):1995–2002.
disorders. Crit Care Med 2007;35(10):2243–50. 84. Randomized, placebo-controlled trial of lisofylline
71. Kushimoto S, Endo T, Yamanouchi S, et al. Rela- for early treatment of acute lung injury and acute
tionship between extravascular lung water and respiratory distress syndrome. Crit Care Med
severity categories of acute respiratory distress 2002;30(1):1–6.
syndrome by the Berlin definition. Crit Care 2013; 85. National Heart Lung, Blood Institute Acute Respira-
17(4):R132. tory Distress Syndrome Clinical Trials Network,
72. Jozwiak M, Silva S, Persichini R, et al. Extravas- Matthay MA, Brower RG, et al. Randomized,
cular lung water is an independent prognostic placebo-controlled clinical trial of an aerosolized
factor in patients with acute respiratory distress beta(2)-agonist for treatment of acute lung injury.
syndrome. Crit Care Med 2013;41(2):472–80. Am J Respir Crit Care Med 2011;184(5):561–8.
73. Ichikado K. “The Berlin definition” and clinical 86. Steinberg KP, Hudson LD, Goodman RB, et al.
significance of high-resolution CT (HRCT) imaging Efficacy and safety of corticosteroids for persistent
in acute respiratory distress syndrome. Masui acute respiratory distress syndrome. N Engl J Med
2013;62(5):522–31 [in Japanese]. 2006;354(16):1671–84.
74. Boissier F, Katsahian S, Razazi K, et al. Prevalence 87. Taylor RW, Zimmerman JL, Dellinger RP, et al.
and prognosis of cor pulmonale during protective Low-dose inhaled nitric oxide in patients with acute
ventilation for acute respiratory distress syndrome. lung injury: a randomized controlled trial. JAMA
Intensive Care Med 2013;39(10):1725–33. 2004;291(13):1603–9.
75. Determann RM, Royakkers A, Wolthuis EK, et al. 88. Luhr OR, Antonsen K, Karlsson M, et al. Inci-
Ventilation with lower tidal volumes as compared dence and mortality after acute respiratory failure
with conventional tidal volumes for patients without and acute respiratory distress syndrome in Swe-
acute lung injury: a preventive randomized den, Denmark, and Iceland. The ARF Study
controlled trial. Crit Care 2010;14(1):R1. Group. Am J Respir Crit Care Med 1999;159(6):
76. Gajic O, Dara SI, Mendez JL, et al. Ventilator-asso- 1849–61.
ciated lung injury in patients without acute lung 89. Valta P, Uusaro A, Nunes S, et al. Acute respi-
injury at the onset of mechanical ventilation. Crit ratory distress syndrome: frequency, clinical
Care Med 2004;32(9):1817–24. course, and costs of care. Crit Care Med 1999;
77. Futier E, Constantin JM, Paugam-Burtz C, et al. 27(11):2367–74.
A trial of intraoperative low-tidal-volume ventilation 90. Goss CH, Brower RG, Hudson LD, et al. Incidence
in abdominal surgery. N Engl J Med 2013;369(5): of acute lung injury in the United States. Crit Care
428–37. Med 2003;31(6):1607–11.
624 Sweatt & Levitt

91. Arroliga AC, Ghamra ZW, Perez Trepichio A, et al. a prospective cohort study. Intensive Care Med
Incidence of ARDS in an adult population of north- 2009;35(8):1352–61.
east Ohio. Chest 2002;121(6):1972–6. 94. Sigurdsson MI, Sigvaldason K, Gunnarsson TS,
92. Hudson LD, Steinberg KP. Epidemiology of acute lung et al. Acute respiratory distress syndrome: nation-
injury and ARDS. Chest 1999;116(1 Suppl):74S–82S. wide changes in incidence, treatment and mortality
93. Linko R, Okkonen M, Pettila V, et al. Acute respi- over 23 years. Acta Anaesthesiol Scand 2013;
ratory failure in intensive care units. FINNALI: 57(1):37–45.