Impact of positive end-expiratory pressure on the definition of acute respiratory distress syndrome

Objective. We examined whether PEEP during the first hours of ARDS can induce such a change in oxygenation that could mask fulfillment of the AECC criteria of a PaO2/FIO2 less than or equal to 200 essential for ARDS diagnosis. Design and setting. Observational, prospective cohort in two medical-surgical ICU in teaching hospitals. Patients. 48 consecutive patients who met AECC criteria of ARDS on 0 PEEP (ZEEP) at the moment of diagnosis. Measurements and results. PaO2/FIO2 and lung mechanics were recorded on admission (0 h) to the ICU on ZEEP, and after 6, 12, and 24 h on PEEP levels selected by attending physicians. Lung Injury Score (LIS) was calculated at 0 and 24 h. PaO2/FIO2 rose significantly from 121+/-45 on ZEEP at 0 h, to 234+/-85 on PEEP of 12.8+/-3.7 cmH(2)O after 24 h. LIS did not change significantly (2.34+/-0.53 vs. 2.42+/-0.62). These variables behaved similarly in pulmonary and extrapulmonary ARDS, and in survivors and nonsurvivors. After 24 h only 18 patients (38%) still had a PaO2/FIO2 of 200 or lower. Their mortality was similar to that in the remaining patients (61% vs. 53%). Conclusions. The use of PEEP improved oxygenation such that one-half of patients after 6 h, and most after 24 h did not fulfill AECC hypoxemia criteria of ARDS. However, LIS remained stable in the overall series. These results suggest that PEEP level should be taken into consideration for ARDS diagnosis.