Purpose: The PaO2/FiO2 is an integral part of the assessment of patients with acute respiratory distress syndrome (ARDS). The American-European Consensus Conference definition does not mandate any standardization procedure. We hypothesized that the use of PaO2/FiO2 calculated under a standard ventilatory setting within 24 h of ARDS diagnosis allows a more clinically relevant ARDS classification.
Methods: We studied 452 ARDS patients enrolled prospectively in two independent, multicenter cohorts treated with protective mechanical ventilation. At the time of ARDS diagnosis, patients had a PaO2/FiO2 ≤ 200. In the derivation cohort (n = 170), we measured PaO2/FiO2 with two levels of positive end-expiratory pressure (PEEP) (≥ 5 and ≥ 10 cmH2O) and two levels of FiO2 (≥ 0.5 and 1.0) at ARDS onset and 24 h later. Dependent upon PaO2 response, patients were reclassified into three groups: mild (PaO2/FiO2 > 200), moderate (PaO2/FiO2 101-200), and severe (PaO2/FiO2 ≤ 100) ARDS. The primary outcome measure was ICU mortality. The standard ventilatory setting that reached the highest significance difference in mortality among these categories was tested in a separate cohort (n = 282).
Results: The only standard ventilatory setting that identified the three PaO2/FiO2 risk categories in the derivation cohort was PEEP ≥ 10 cmH2O and FiO2 ≥ 0.5 at 24 h after ARDS onset (p = 0.0001). Using this ventilatory setting, patients in the validation cohort were reclassified as having mild ARDS (n = 47, mortality 17 %), moderate ARDS (n = 149, mortality 40.9 %), and severe ARDS (n = 86, mortality 58.1 %) (p = 0.00001).
Conclusions: Our method for assessing PaO2/FiO2 greatly improved risk stratification of ARDS and could be used for enrolling appropriate ARDS patients into therapeutic clinical trials.