In August 1967, Ashbaugh et al.9 described for the first time a syndrome that they termed the ARDS. They studied a cohort of 272 patients who were receiving respiratory support, and from this cohort they identified 12 patients with a syndrome that was similar to the Infant Respiratory Distress Syndrome. The respiratory distress was defined as sudden, catastrophic, and often associated with a multiorgan system insult which led to tachypnea, hypoxemia, decreased respiratory system compliance, and bilateral pulmonary infiltrates on chest X-ray due to non-cardiogenic pulmonary edema. The mortality rate was 58% and on pathological examination the lungs of the non-survivors were heavy, atelectatic, with interstitial and alveolar edema, and hyaline membranes. Since that time, the hallmark of this syndrome has included: (i) a risk factor for the development of ARDS, (ii) severe hypoxemia with a relatively high FiO2, (iii) bilateral pulmonary infiltrates, and (iv) no clinical evidence of cardiogenic pulmonary edema, although acute lung injury (ALI) resulting in ARDS can also occur in the setting of left ventricular failure.

ARDS is caused by an inflammatory insult to the alveolar-capillary membrane that results in increased permeability and subsequent interstitial and alveolar edema.10 Unlike other forms of acute respiratory failure and like all forms of tissue inflammation, ALI during ARDS represents a complex process in which multiple cellular signaling pathways propagate or attenuate lung damage. Since it is difficult to measure changes in capillary and alveolar permeability at the bedside, diagnosis of ARDS is based on a combination of clinical, oxygenation, hemodynamic and radiographic criteria. These criteria allow the inclusion of a heterogeneous group of critically ill patients since various types of injury can lead to a similar pulmonary response. The original description of ARDS was incapable of identifying a uniform group of patients. Several patients from the original cohort would not be classified as ARDS today, since fluid overload was an important etiological factor. Thus, a precise definition is important for accurate identification and quantification of various aspects of the underlying pathophysiology and to select the best therapeutic approach in selected subgroups of patients.

Given that severe hypoxemia is the hallmark of ARDS, hypoxemia is crucial to the assessment of the severity of ARDS, for predicting the evolution in any given patient, and for assessing the response to treatment. In order to better characterize the severity of lung damage, the AECC9 defined ALI and ARDS as follows: (i) acute and sudden onset of severe respiratory distress; (ii) bilateral infiltrates on frontal chest radiograph; (iii) the absence of left atrial hypertension (a pulmonary capillary wedge pressure <18mmHg or no clinical signs of left ventricular failure); and (iv) severe hypoxemia (assessed by the PaO2/FiO2 ratio). According to these guidelines, ALI exists when the PaO2/FiO2 ratio is ≤300 and >200mmHg regardless of the PEEP and FiO2, and ARDS when the PaO2/FiO2 ratio is ≤200mmHg again regardless of PEEP and FiO2. Although this definition formalized the criteria for the diagnosis of ARDS and is simple to apply in the clinical setting, it has been challenged over the years in several studies.3,8,11 All patients start off with terrible oxygenation and there is little room for stratifying the patients if there is no further re-evaluation of the hypoxemia. Also, the physiological thresholds of the AECC definition do not require standardized ventilatory support. The use of PEEP can improve oxygenation indices sufficiently to change the physiology of the lung converting patients meeting the definition of ARDS to patients not meeting the ARDS definition. Therefore, a patient could fit the ARDS criteria when the PaO2 is measured with zero PEEP but not when measured at a PEEP of 5 or 10cmH2O, or when measured on FiO2=0.35 but not when measured on FiO2=0.5.3,5,11 These findings illustrate the major problems experienced when trying to compare the results of various clinical trials evaluating ventilatory strategies, since patients with very different levels of lung dysfunction and disease may have been enrolled.

In 1999, Villar et al.3 proposed the need for different guidelines, based on a specific, standard method of evaluating oxygenation status, a proposal that was later advocated by others.11 In order to determine the impact of various PEEP and FiO2 levels on the stratification of patients meeting the AECC ARDS definition, Villar et al.5 evaluated the impact of standard ventilation settings applied on the day patients met the AECC ARDS criteria and 24h later. They studied 170 patients and found that only 58% of them fulfilled ARDS criteria when evaluated on PEEP≥10cmH2O and FiO2≥0.5 at 24h after meeting the AECC ARDS definition. The ICU mortality of those patients was 46%. By contrast, 32% of patients were classified as having ALI (their mortality was 20%), and 10% of patients had a PaO2/FIO2>300mmHg and were simply categorized as having acute respiratory failure (their ICU mortality was 6%). This study demonstrated the large variability in the severity of lung damage in patients who initially meet the AECC definition of ARDS and the strong correlation between oxygenation impairment at 24h after ARDS onset and ICU outcome. The major implication of these findings is that the use of the AECC ARDS definition to enroll patients into clinical trials may result in the inclusion of patients with highly variable severity of lung injury and mortalities. If the subjects in a trial have a very low risk of the condition that the intervention is hypothesized to prevent, the trial – regardless of sample size – will not verify the value of the intervention.12 For example, in the ARDSnet trial,13 17% of enrolled patients did not have ARDS but were included in the overall analysis. Consequently, it can be argued that the ARDSnet trial failed to focus on the highest risk patients.

A PEEP and FiO2 trial 24h after ARDS onset is an easy and simple strategy to find or identify subpopulations of ARDS patients at highest risk.4 It is not difficult to explain why patients initially meeting AECC ARDS criteria would fail to meet the same criteria on standard ventilator setting. It is well established that changes in PEEP and FiO2 alter the PaO2/FiO2 values in lung-injured patients.14,15 PEEP can improve oxygenation sufficiently to change the physiology of the lung.15 If PEEP is inadequate, the lung collapses during expiration allowing alveolar derecruitment and causing ventilator-induced lung injury. The lack of a standard PEEP and FiO2 setting in the AECC definitions may explain the negative results of published trials of various therapeutic interventions on patients with ALI/ARDS.16–18 Few of these negative trials used the AECC definition for ARDS, and none required standard ventilator settings to qualify patients for enrollment. Because many patients without sustained ARDS may have been enrolled, it is conceivable that a disproportionate number of patients meeting ALI or ARF criteria ended up in the control arm, negating the beneficial effect of the treatment because of the lower mortality of these patients. This is contrary to published data from Villar et al.4 From a cohort of 311 patients meeting AECC criteria, only 103 patients (33.1%) qualified as having ARDS on a standard PEEP–FiO2 setting (PEEP≥5cmH2O with FiO2≥0.5) at 24h after screening. This resulted in a homogeneous group of patients in whom the benefit or lack of benefit of a therapy could be appropriately evaluated. Most of the 208 patients (66.9%) who did not qualify as ARDS at 24h had a transient fall of PaO2/FiO2 or the PaO2 was measured while receiving PEEP<5cmH2O and/or FiO2<0.5. Several other patients had a marked improvement of their pulmonary dysfunction within the first 24h.

Initial ventilatory settings in the management of patients with ALI and ARDS may affect the prevalence of ARDS. In a recent epidemiological study, Villar et al.7 found that the current incidence of ARDS in Spain using protective mechanical ventilation (low tidal volume plus moderate to high PEEP levels) before and after the diagnosis of ARDS was very low, and dramatically lower than the incidence values reported in other studies where patients were ventilated with higher tidal volumes and lower levels of PEEP.19–21 In the Villar et al. study,7 most patients (87%) were treated within the first 24h with PEEP≥10cmH2O and 14% received 15–24cmH2O PEEP. Under these circumstances, when patients were categorized by their initial PaO2/FiO2 range at the time of ARDS identification, almost half of patients (46%) had severe ARDS (PaO2/FiO2≤100) and their mortality was 1.5 times higher than the mortality of patients with a PaO2/FiO2 between >100 and 200mmHg (52.1% vs. 34.8%, p<0.01). López-Fernández et al.8 recently performed the largest epidemiological study on pediatric ARDS in the Western world and the only one using the AECC criteria in which most children received protective mechanical ventilation. In addition to reporting that the incidence of pediatric ARDS is lower than those reported in adults, the major contribution of their study was that the comparison of PaO2/FiO2 ratios at baseline (≤100mmHg vs. >100mmHg) and the persistence of PaO2/FiO2≤200 in response to PEEP≥10cmH2O on FiO2≥0.5, was a predictor of outcome since it identified subgroups of patients with different lung severity and mortality. Although the overall mortality of the cohort was low (26%), the stratification based on the range of PaO2/FiO2 ratio measured under standard ventilator settings at 24h after ARDS onset, was able to identify four subgroups of patients with markedly different mortalities: only 44.5% of patients had severe ARDS (PaO2/FiO2≤100), 36.3% had moderate ARDS (PaO2/FiO2 101–200), 11.6% had ALI (PaO2/FiO2 201–300); and 7.5% had ARF (PaO2/FiO2>300), while the mortality rates were 38.5%, 20.7%, 11.8%, and 0%, respectively (p<0.001).

A proposal for an update of the AECC ARDS definition has been published recently6 by a task force panel of experts using a similar terminology as we had previously reported. Using teleconferencing, in-person discussions and retrospective data, they proposed an ARDS classification with three severity categories (mild, moderate, and severe) for empirical evaluation. The term “mild” ARDS was used for defining those patients who are considered as ALI in the AECC definition (300≥PaO2/FiO2>200mmHg). The term “moderate” was used for patients with a PaO2/FiO2>100mmHg but <200, and the term “severe” for those with a PaO2/FiO2≤100mmHg. The panel used 7 datasets: 4 from multicenter studies (enrolling 4188 patients with a PaO2/FiO2≤300mmHg), and 3 from small, single-center studies (enrolling 269 patients). By considering only those patients from the multicenter studies who were managed with PEEP≥5cmH2O at the time of study enrollment, the panel found that hospital mortality increased with every stage of severity (mild 27%, moderate 32%, severe 45%). In the database from the 3 small, single-center studies comprising 269 patients, the hospital mortality increased as well with every stage of ARDS (mild 20%, moderate 41%, severe 52%). Although the authors stated that the purpose of their empirical definition was not to develop a prognostic tool, this exercise should be cautiously generalized for the following methodological reasons.

First, none of the patients included in the empirical analysis were prospectively enrolled for the purpose of revising the ARDS definition and/or evaluating risk stratification. Second, the categorization of patients was done based on the PaO2/FiO2 value at the time of inclusion into their respective observational study or randomized clinical trial. There is no information on whether those baseline values of PaO2/FiO2 were calculated at the time of ARDS onset or during any time of their disease process, or whether the PaO2 was measured under a known FiO2 and PEEP level. In our studies,3,8 PaO2/FiO2 was always calculated from the PaO2 values measured 30min after each standard ventilator setting under a specified FiO2 and PEEP level. Third, 24% of patients included had a PaO2/FiO2>200 at the time of enrollment (classified as non-ARDS by the AECC definition). We did not include those patients in our studies because in many centers they are usually not treated with endotracheal intubation and invasive MV. However, we have reported that in about 40% of ARDS patients, PaO2/FiO2 increased above 200mmHg when they were evaluated on PEEP≥10 and FiO2≥0.5 at 24h after being diagnosed as having ARDS according to the AECC definition.4,5Fourth, the empirical definition did not consider the level of FiO2 for PaO2/FiO2 categorization despite the fact that changes in the applied FiO2 results in changes in PaO2/FiO2.5,22 In addition, since it is likely that a significant proportion of patients included in those multicenter studies were on FiO2≤0.4 at the time of study enrollment, there is no information on how many patients could not meet ARDS criteria if they were evaluated at a minimum level of FiO2=0.5. Fifth, 518 patients were eliminated from the empirical analysis because PEEP was missing or was <5cmH2O. In our prospective studies,3,8 we did not exclude any patient based on the baseline PEEP or FiO2 at the time of ARDS diagnosis. Based on the wide range of FiO2 and PEEP levels at the time of AECC ARDS diagnosis, we found that a standard ventilator setting can improve oxygenation sufficiently to convert patients meeting the AECC ARDS definition on PEEP<5cmH2O and/or on FiO2≤0.4 to non-ARDS patients. Sixth, since there was no standardization of ventilator settings for measuring PaO2, and since more than 50% of patients were on PEEP<10cmH2O at baseline, the experts’ panel only selected a unique level of PEEP (≥5cmH2O) as a requirement for the proposed empirical ARDS definition. We had found that when all patients are evaluated under the same standard ventilator settings, only the evaluation under a PEEP≥10 reached statistical significance for association of PaO2/FiO2 category with ICU mortality.5Seventh, among other considerations (strict entry criteria for the trials), those studies do not include all consecutive ARDS patients admitted into the ICU since patients with pre-existent chronic diseases, patients over a certain age, and patients with established organ dysfunction were excluded. Eighth, the 4 multicenter studies were a case-mix of observational studies and clinical trials performed from 1996 to 2000, when patients were ventilated with tidal volumes ≥10ml/kg predicted body weight (currently considered as an injurious tidal volume) and low levels of PEEP, and studies performed after the year 2000 (when patients were ventilated with lower tidal volumes). It has been postulated that the development of ARDS should have decreased because of advances in supportive care, particularly the application of protective mechanical ventilation.23Ninth, there is a great discrepancy between the number of patients included in the database from the three-single centers and the number of patients reported in the original publications. We have re-examined those publications and have noted that they only studied 75 non-consecutive, selected ALI/ARDS instead of the 269 that are claimed by the experts’ panel. Tenth, the experts’ panel failed to evaluate their patients at 24h after ARDS diagnosis. There is no information about the number of patients who did not meet ARDS criteria at 24h of study entry under the same arbitrary condition (PEEP≥5). Taking into consideration all the above, and based on our previous studies, we postulate that at least 50% of those patients would have a PaO2/FiO2>200mmHg, and at least 25% would have a PaO2/FiO2>300mmHg 24h after ARDS identification on a PEEP≥10 and FiO2≥0.5.

In conclusion, we need more specific guidelines based on a standard method of evaluating oxygenation status (i.e. a specific level of PEEP and FiO2) in order to properly classify the severity of patients with and as ARDS. As we have reported in the several studies discussed in this review, a large variability in the severity of lung damage exists in patients meeting the AECC definition of ARDS and a strong correlation exists between oxygenation impairment at 24h after ARDS onset and ICU outcome. The judicious use of PEEP, FiO2 and tidal volume in critically ill patients in the last decade has made ARDS a rare syndrome in today's modern ICUs. However, no matter how infrequently we observe its presence we need to be able to properly classify its severity. The AECC ARDS definition is dead, long live PEEP!

Supported by Instituto de Salud Carlos III, Spain (PI 10/0393) and the Asociación Científica Pulmón y Ventilación Mecánica.

The authors have no conflict of interest to declare.