An attempt to validate the modification of the American-European consensus definition of acute lung injury/acute respiratory distress syndrome by the Berlin definition in a university hospital

Abstract

The Berlin definition for acute respiratory distress syndrome (ARDS) is a new proposal for changing the American-European consensus definition but has not been assessed prospectively as yet. In the present study, we aimed to determine (1) the prevalence and incidence of ARDS with both definitions, and (2) the initial characteristics of patients with ARDS and 28-day mortality with the Berlin definition. We performed a 6-month prospective observational study in the ten adult ICUs affiliated to the Public University Hospital in Lyon, France, from March to September 2012. Patients under invasive or noninvasive mechanical ventilation, with PaO2/FiO2 <300 mmHg regardless of the positive end-expiratory pressure (PEEP) level, and acute onset of new or increased bilateral infiltrates or opacities on chest X-ray were screened from ICU admission up to discharge. Patients with cardiogenic pulmonary edema were excluded. Patients were further classified into specific categories by using the American-European Consensus Conference and the Berlin definition criteria. The complete data set was measured at the time of inclusion. Patient outcome was measured at day 28 after inclusion. During the study period 3,504 patients were admitted and 278 fulfilled the American-European Consensus Conference criteria. Among them, 18 (6.5 %) did not comply with the Berlin criterion PEEP ≥ 5 cmH2O and 20 (7.2 %) had PaO2/FiO2 ratio ≤200 while on noninvasive ventilation. By using the Berlin definition in the remaining 240 patients (n = 42 mild, n = 123 moderate, n = 75 severe), the overall prevalence was 6.85 % and it was 1.20, 3.51, and 2.14 % for mild, moderate, and severe ARDS, respectively (P > 0.05 between the three groups). The incidence of ARDS amounted to 32 per 100,000 population per year, with values for mild, moderate, and severe ARDS of 5.6, 16.3, and 10 per 100,000 population per year, respectively (P < 0.05 between the three groups). The 28-day mortality was 35.0 %. It amounted to 30.9 % in mild, 27.9 % in moderate, and 49.3 % in severe categories (P < 0.01 between mild or moderate and severe, P = 0.70 between mild and moderate). In the Cox proportional hazard regression analysis ARDS stage was not significantly associated with patient death at day 28. The present study did not validate the Berlin definition of ARDS. Neither the stratification by severity nor the PaO2/FiO2 at study entry was independently associated with mortality.