Abstract

PURPOSE: To investigate the effect of prolonged glucocorticoid treatment for patients with acute respiratory distress syndrome (ARDS). METHODS: We conducted two sets of intention-to-treat analyses: (1) a primary analysis of individual patients' data (IPD) of four randomized controlled trials (RCTs) which investigated methylprednisolone treatment (n = 322) and (2) a trial-level meta-analysis incorporating four additional RCTs which investigated hydrocortisone treatment in early ARDS (n = 297). We standardized definitions to derive outcomes in all datasets. The primary outcome for the IPD analysis was time to achieving unassisted breathing (UAB) by study day 28. Secondary outcomes included mechanical ventilation (MV) and intensive care unit (ICU)-free days, hospital mortality, and time to hospital mortality by day 28. RESULTS: By study day 28, compared to the placebo group, the methylprednisolone group had fewer patients who died before achieving UAB (12 vs. 29 %; p < 0.001) and more patients who achieved UAB (80 vs. 50 %; p < 0.001). In the methylprednisolone group, time to achieving UAB was shorter [hazard ratio 2.59, 95 % confidence interval (CI) 1.95-3.43; p < 0.001], and hospital mortality was decreased (20 vs. 33 %; p = 0.006), leading to increased MV (13.3 ± 11.8 vs. 7.6 ± 5.7; p < 0.001) and ICU-free days (10.8 ± 0.71 vs. 6.4 ± 0.85; p < 0.001). In those patients randomized before day 14 of ARDS onset, the trial-level meta-analysis indicated decreased hospital mortality (36 vs. 49 %; risk ratio 0.76, 95 % CI 0.59-0.98, I (2) = 17 %, p = 0.035; moderate certainty). Treatment was not associated with increased risk for infections (risk ratio 0.77, 95 % CI 0.56-1.08, I (2) = 26 %; p = 0.13; moderate certainty). CONCLUSIONS: Prolonged methylprednisolone treatment accelerates the resolution of ARDS, improving a broad spectrum of interrelated clinical outcomes and decreasing hospital mortality and healthcare utilization.