Secondary Logo

Journal Logo

RESPIRATORY SYSTEM: Edited by Arthur P. Wheeler

The new definition for acute lung injury and acute respiratory distress syndrome

is there room for improvement?

Costa, Eduardo L.V.a,b; Amato, Marcelo B.P.a

Author Information
Current Opinion in Critical Care: February 2013 - Volume 19 - Issue 1 - p 16-23
doi: 10.1097/MCC.0b013e32835c50b1
  • Free



The acute respiratory distress syndrome (ARDS) was described in 1967 in the seminal paper by Ashbaugh et al.[1]. In their series of 272 patients under mechanical ventilation, they identified 12 patients who did not respond to conventional therapy. The most prominent clinical features described were severe dyspnea, tachypnea, cyanosis refractory to oxygen supplementation, decreased lung compliance, and diffuse chest X-ray infiltrates. The respiratory distress had been preceded by trauma in seven of the patients, by viral infection in four, and by pancreatitis in one. Histopathological changes included interstitial and alveolar hemorrhage and edema, and the presence of alveolar macrophages, although the most striking finding was the presence of hyaline membranes. Ever since its description, there have been several attempts at defining this syndrome, which became known as the ARDS. In this article, we will briefly review the previous definitions of ARDS, present a critical appraisal of the new Berlin definition, and suggest possibilities for future improvement.


In 1988, a definition [2] was proposed that encompassed four different scales to quantify the respiratory impairment based on the level of positive end-expiratory pressure (PEEP), the ratio of the partial pressure of oxygen to the inspired fraction of oxygen (P/F ratio), the static compliance of the respiratory system, and the degree of chest X-ray infiltration. Each scale was scored from 1 to 4 points (Table 1), and the four scales were averaged to provide a final score ranging from 1 to 4, with the term ARDS being recommended for values higher than 2.5. The definition also encompassed the presence or absence of an inciting risk factor and of nonpulmonary organ dysfunction.

Table 1
Table 1:
Lung Injury Score
Box 1
Box 1:
no caption available

In 1994, an American–European Consensus Conference (AECC) of specialists proposed a new definition [3] with the aim to bring clarity and uniformity to the definition of ARDS and acute lung injury (ALI). ALI was defined as a syndrome of acute onset of inflammation and increased permeability of the alveolar–capillary membrane associated with one or more risk factors. Persistent, usually lasting days to weeks, the syndrome was characterized by arterial hypoxemia resistant to oxygen supplementation and by diffuse radiologic infiltrates. Despite internal dissidences, with some members suggesting that there should be different definitions for research, epidemiological, and individual patient care, the panel ultimately decided to propose a single definition for all those areas. Early on, it was recognized that ALI and ARDS represented a clinical spectrum, which included a continuum of P/F ratio and X-ray abnormalities, and that arbitrary cutoff values would have to be established. There was considerable discussion regarding the ideal cutoff value of P/F ratio, specifically whether a cutoff of 150 or 200 mmHg should be used. Finally, the cutoff value of 200 mmHg was chosen with a word of caution to carefully exclude other conditions with altered gas exchange occurring through other pathophysiological pathways, such as postoperative atelectasis. On the same ground, the panel included a criterion to exclude hydrostatic pulmonary edema, which can mimic the gas exchange and radiological deterioration found in ARDS. The panel also acknowledged that PEEP could affect oxygenation. They decided not to include the PEEP values in the oxygenation criterion, however, because they considered that response to PEEP is not consistent and that it is time dependent. The final definition had four components (Table 2) and a broader category of disease called ALI, which encompassed both patients with ARDS (P/F ratio ≤200) and patients with less severe disease (patients with P/F ratios between 200 and 300). In practice, this broader definition brought confusion as most physicians used the term ALI to exclusively indicate patients with ALI without ARDS (P/F ratio from 201 to 300).

Table 2
Table 2:
Criteria for acute lung injury (ALI) and acute respiratory distress syndrome (ARDS)


The AECC definition was widely adopted [4–9], although it suffered several criticisms over the years, especially regarding the P/F ratio cutoff value [10,11], the lack of standard ventilatory settings at the time of arterial blood gases [10–13,14▪▪], and the absence of a clear definition of acute, to mention a few. This year, the definition of ARDS has been revised [15▪▪] by a panel of specialists in a joint effort of the European Society of Intensive Care Medicine, the American Thoracic Society, and the Society of Critical Care Medicine. Their aim was to address the current limitations of the AECC definition and explore other defining variables.

The panel agreed to maintain the previous conceptual model, which defined ARDS as a syndrome of acute, diffuse lung inflammation, with edema because of increased permeability of the alveolar–capillary membrane, and clinically characterized by decreased oxygenation because of increased venous admixture, decreased lung compliance, increased physiological dead space, and bilateral radiographic opacities [better characterized as increased lung weight and decreased aeration on computed tomography (CT) evaluation].

The first step of the revision was to agree on a draft definition that addressed the limitations of the AECC definition, yet maintained compatibility with it. The proposed definition comprised three mutually exclusive categories of hypoxemia: mild (200 mmHg < P/F ≤ 300 mmHg), moderate (100 mmHg < P/F ≤ 200 mmHg), and severe (P/F ≤100 mmHg), and inclusion of four ancillary variables to restrict the definition of severe ARDS: high radiographic score (3–4 quadrants), low respiratory system compliance (≤40 ml/cm H2O), PEEP (≥10 cm H2O), and corrected minute ventilation (≥10 l/min).

This draft definition was subsequently tested on 4188 patients with ARDS from 4 multicenter clinical datasets and 269 patients from 3 single-center datasets containing physiologic information. The use of the four ancillary variables decreased by half (from 28 to 14%) the number of patients classified as severe ARDS, but did not add prediction power for mortality (mortality of the more restricted sample = 45 vs. 45% when considering the simpler definition). Therefore, the panel removed the ancillary variables from the definition.

The final Berlin definition (Table 3) clarified several aspects of the AECC definition, while retaining its ease of use and still improving the predictive ability for death. More specifically, the Berlin definition created a criterion of acuteness of disease onset, reclassified the oxygenation criterion, included a minimum PEEP value for the diagnosis, redefined the exclusion criterion based on the presence of hydrostatic edema, and reformulated the radiologic criterion.

Table 3
Table 3:
The Berlin definition of the acute respiratory distress syndrome (ARDS)

Acuteness of disease

The definition now establishes a maximum period of time of 7 days between the exposure to the risk factor and the development of the syndrome.

This modification emphasizes the search for a known risk factor within a reasonable time frame. The choice of 7 days was based on a previous work [16] that showed that most patients develop ARDS within 72 h and nearly all patients within 7 days of exposure to risk factors.

Categories of oxygenation

The group with ALI without ARDS was replaced by mild ARDS. Additionally, a new category with P/F ratio of 100 mmHg or less was created to represent severe ARDS. The rationale for this choice was likely based on the recent studies that showed worse outcomes in the lowest quartile of oxygenation irrespective of the ventilatory strategy [4,5]. In those studies [4,5], the lowest quartile of P/F ratio roughly matched the P/F cutoff chosen in the Berlin definition [15▪▪].

Minimum positive end-expiratory pressure value

PEEP values are known to influence oxygenation. In fact, this had already been acknowledged in the AECC definition, but the authors refrained from including a PEEP criterion because of the variable and time-dependent effects of PEEP on oxygenation. In the Berlin definition [15▪▪], a lower threshold of 5 cm H2O was chosen and will likely help exclude patients whose hypoxemia is mostly secondary to atelectasis.

Exclusion of hydrostatic pulmonary edema

Ruling out hydrostatic pulmonary edema became simpler with the removal of the pulmonary artery wedge pressure criterion. According to the Berlin definition [15▪▪], clinical judgment suffices in the presence of a known risk factor for ARDS. If no risk factor can be identified, however, a more rigorous evaluation, for example with echocardiography, must be undertaken.

Radiologic criterion

The criterion of bilateral opacities on frontal radiographs was maintained. An important improvement was the recommendation not to consider opacities explained by effusions, lobe or lung collapse, and nodules. To optimize interobserver agreement, a reference set of chest radiographs was also added. Additionally, the Berlin definition recognizes that the imaging findings can be demonstrated on CT instead of chest radiograph.


As detailed below, we identified four key aspects of the definition worth further examination. We tried to explore these aspects with the aim of better defining risk stratification and target populations for clinical trials. To that end, we performed a meta-analysis with data at the patient level using six clinical trials [8,9,17–20] originally designed to test different ventilatory strategies. A total of 1752 patients were included. From those, 15% were excluded because data were missing on at least one of the variables APACHE, age, P/F ratio, or pH. Death at 60 days was used as the outcome variable in logistic regression models. This convenience sample has some overlap with the larger sample of 4188 patients used by the Berlin panel and encompasses only patients included in randomized trials. However, because we could use more detailed information at the patient level, we evaluated some aspects that could not be well explored by the Berlin panel. Our intention was to suggest some promising candidates to improve risk stratification for both clinical and epidemiological purposes, to be checked in a larger, more appropriate sample.

Would the risk stratification improve if the P/F ratio after 24 h of study enrolment (in fact, representing a stable period of a few hours) were used instead of the P/F ratio at baseline?

Yes (Figs. 1 and 2), probably for similar reasons that the APACHE score is calculated over 24 h. Collecting the P/F ratio after a few hours of stabilization, or averaging during the first 24 h, could be more representative of the severity of lung disease, as it would provide enough time for stabilization after intubation. Intubation together with the initiation of sedatives and of positive-pressure ventilation markedly disturbs the ventilation–perfusion matching, thus creating more noise in the P/F ratio signal [21]. Additionally, the P/F ratio at 24 h reflects to some extent the response to the ventilator settings applied.

Odds ratio for death according to the oxygenation at baseline. Unadjusted and adjusted logistic regression analyses of death at 60 days according to the acute respiratory distress syndrome severity using the oxygenation criterion at baseline. Baseline characteristics included were APACHE III, age, and pH.
Odds ratio for death according to the oxygenation at 24 h. Unadjusted and adjusted logistic regression analyses of death at 60 days according to the acute respiratory distress syndrome severity using the oxygenation criterion at 24 h. Baseline characteristics included were APACHE III, age, and pH.

Is it important to have two P/F thresholds?

At baseline, no; after 24 h, yes. At baseline, the moderate category was not statistically different from the mild category in terms of mortality (Fig. 1). A simplified version of the Berlin definition created by merging the categories mild and moderate performed as well as the Berlin definition (P = 0.33). This finding suggests that most of the predictive power of the Berlin definition resides on the knowledge of whether or not the patient belongs to the severe category.

In contrast, after 24 h, both the moderate (P = 0.02) and the severe (P < 0.001) categories were significantly different from the mild category in terms of chance of death (Fig. 2). In this scenario, with both using P/F ratios at 24 h, the Berlin definition performed significantly better than the simplified version (P = 0.02).

What about the ancillary variables?

We reassessed the value of compliance and PEEP from the four originally proposed ancillary variables. We did not have data on the corrected minute ventilation or the radiographic score. In addition to the variables proposed, we tested the addition of FiO2 divided into three categories with cutoff values of 0.5 and 0.7, as recently suggested [14▪▪].

We used the compliance adjusted to ideal body weight (IBW) to take into account the differences in lungs sizes – the same procedure that is now the standard practice for the tidal volumes. Instead of using an arbitrary cutoff value of compliance, we chose the median compliance of the severe oxygenation category to optimize the number of patients in each subcategory. The cutoff obtained was 0.4 ml/cm H2O/kg IBW, corresponding to a compliance of 28 ml/cm H2O for a 70-kg patient. We found that compliance was a strong predictor of mortality (P < 0.0001), along with the P/F ratio (P = 0.0047). Classifying patients according to compliance as shown in Table 4 improved the risk stratification within each stratum of P/F ratio.

Table 4
Table 4:
The Berlin definition stratified according to compliance

Adding PEEP did not improve prediction (P = 0.08), even after taking into account the possible nonlinear relationships between PEEP and mortality. Of note, all but six patients had PEEP values equal or greater than 5 cm H2O. Therefore, we cannot exclude that patients with PEEP values below 5 cm H2O would have a different mortality.

Conversely, the addition of FiO2 improved the prediction confirming the findings of Britos et al.[14▪▪]. One possible reason why FiO2 could add predictive power to a model already containing the P/F categories is that, at higher FiO2, PaO2 correlates better with the mass of collapsed lung tissue [22▪▪].

What happens after multivariate adjustment (inclusion of age, APACHE, and pH)?

In the Berlin definition, the authors 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’. In doing so, however, they failed to adjust for the effect of confounders, an essential step in outcomes research. For example, sicker patients die more often and also tend to have lower P/F ratios. Therefore, it is possible that the mortality differences across the P/F categories be related to nonpulmonary organ failures and not to the severity of the lung disease.

We included the relevant variables, age and APACHE, as suggested by others [19,23▪], as well as baseline pH, a variable strong and independently related to death in our sample. After adjustment for these baseline characteristics, the odds ratio for death of the moderate and severe categories decreased (Figs. 1 and 2), but not enough to change the original message of the Berlin definition.

Additionally, adjustment made it clearer that P/F ratio and compliance were independent risk factors for death and improved the ability to predict death (Fig. 3). Conversely, once P/F ratio and compliance were already considered, FiO2 and PEEP were not informative.

Comparison of the discriminative power between P/F and three ancillary variables. Delta area under the ROC curve for the addition of each of the four variables (P/F ratio, FiO2, PEEP, and compliance) to (a) a null logistic model to predict death at 60 days and (b) to the same model preadjusted to APACHE, age, pH, P/F category, and compliance. The delta area under the ROC curve for P/F and for compliance was calculated by removing those variables from the preadjusted model. Confidence intervals were created with the bootstrap method with 2000 replicates. P/F ratio, the ratio of the partial pressure of oxygen to the inspired fraction of oxygen, was stratified in ≤100, from 101 to 200, and >200 mmHg; FiO2, the inspired fraction of oxygen was stratified as ≤0.5, between 0.5 and 0.7, and >0.7; PEEP or positive end-expiratory pressure was stratified as ≥10 cm H2O or below 10 cm H2O.

Stratification based on compliance was part of the 1988 Murray definition, and indeed, the Berlin panel acknowledged that a threshold of 20 ml/cm H2O might be an interesting cutoff. In this review, we emphasize that compliance is as important as the P/F ratio in risk stratification and should not be disregarded from the definition of ARDS.

There are some limitations associated with the definitional changes we are proposing. First, the use of the P/F ratio after 24 h would postpone the risk stratification and therefore delay, for example, the inclusion in clinical trials. We believe that the improvement in risk stratification justifies such delay. For example, in one study [6], 66% of patients meeting ARDS criteria at baseline did not meet ARDS criteria after 24 h of standard ventilatory settings. For clinical purposes, the delay would be less important, because patients could be diagnosed with ARDS using the baseline P/F ratio, and the risk stratification could be updated 24 h later. Second, measurement of compliance requires absence of muscle effort implying either neuromuscular blockade or heavy sedation. Indeed, such respiratory muscle rest is welcome in the early phase of ARDS [24] and helps ensure the use of protective low tidal volume ventilation.


The Berlin definition brought improvement and simplification over the previous definitions. We believe that the Berlin definition could be further improved by using the data over the first 24 h to reclassify the severity of the disease and by using compliance with optimized cutoffs to stratify each oxygenation category.


The authors thank the National Institutes of Health Acute Respiratory Distress Syndrome Network; the staff working at the respiratory-ICU, Hospital das Clínicas, University of São Paulo; the Canadian Pressure and Volume-Limited Ventilation Group; the centers who participated in the 1994–1996 Multicenter Trial Group on Tidal Volume Reduction in ARDS; and the Baltimore Clinical Research Group.

Conflicts of interest

This study was partially supported by FAPESP (Fundação de Amparo e Pesquisa do Estado de São Paulo), CNPQ (Conselho Nacional de Pesquisa e Desenvolvimento), and FINEP (Fundo de Financiamento de Estudos de Projetos e Programas).


Papers of particular interest, published within the annual period of review, have been highlighted as:

  • ▪ of special interest
  • ▪▪ of outstanding interest

Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 68).


1. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute respiratory distress in adults. Lancet 1967; 2:319–323.
2. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988; 138:720–723.
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. 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; 299:646–655.
5. 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; 299:637–645.
6. Villar J, Kacmarek RM, Perez-Mendez 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.
7. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med 2005; 353:1685–1693.
8. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000; 342:1301–1308.
9. 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; 338:347–354.
10. Villar J, Perez-Mendez L, Lopez 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; 176:795–804.
11. Ferguson ND, Davis AM, Slutsky AS, Stewart TE. Development of a clinical definition for acute respiratory distress syndrome using the Delphi technique. J Crit Care 2005; 20:147–154.
12. Estenssoro E, Dubin A, Laffaire E, et al. Impact of positive end-expiratory pressure on the definition of acute respiratory distress syndrome. Intensive Care Med 2003; 29:1936–1942.
13. Allardet-Servent J, Forel JM, Roch A, et al. FiO2 and acute respiratory distress syndrome definition during lung protective ventilation. Crit Care Med 2009; 37:202–207.e4–e6.
14▪▪. Britos M, Smoot E, Liu KD, et al. The value of positive end-expiratory pressure and FiO(2) criteria in the definition of the acute respiratory distress syndrome. Crit Care Med 2011; 39:2025–2030.

This study evaluated the value of adding PEEP and FiO2 to the AECC definition of ARDS.

15▪▪. Ranieri VM, Rubenfeld GD, Thompson BT, et al. Acute respiratory distress syndrome: the Berlin definition. JAMA 2012; 307:2526–2533.

This is the most important reference of this review. It is the consensus statement on the new (Berlin) definition of ARDS.

16. 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.
17. Stewart TE, Meade MO, Cook DJ, et al. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. Pressure- and Volume-Limited Ventilation Strategy Group. N Engl J Med 1998; 338:355–361.
18. Brower RG, Shanholtz CB, Fessler HE, et al. 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.
19. Brower RG, Lanken PN, MacIntyre N, et al. Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med 2004; 351:327–336.
20. Brochard L, Roudot-Thoraval F, Roupie E, et al. Tidal volume reduction for prevention of ventilator-induced lung injury in acute respiratory distress syndrome. The Multicenter Trail Group on Tidal Volume reduction in ARDS. Am J Respir Crit Care Med 1998; 158:1831–1838.
21. Hedenstierna G. Pulmonary perfusion during anesthesia and mechanical ventilation. Minerva Anestesiol 2005; 71:319–324.
22▪▪. Reske AW, Costa ELV, Reske AP, et al. Bedside estimation of nonaerated lung tissue using blood gas analysis. Crit Care Med (forthcoming).

This is a study correlating the amount of nonaerated lung tissue to the P/F ratio. It highlights the importance of using pure oxygen before drawing arterial blood gases.

23▪. Villar J, Perez-Mendez L, Basaldua 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; 56:420–428.

In this study, a model predictive of mortality was developed based on demographic, pulmonary, and ventilation data collected at ARDS onset.

24. Papazian L, Forel JM, Gacouin A, et al. Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010; 363:1107–1116.

acute respiratory distress syndrome; definition; disease severity

© 2013 Lippincott Williams & Wilkins, Inc.