Nine studies (n=648 patients) were included in the review: four RCTs (n=341 patients) and five cohort studies (n= 307). The authors stated that study quality was fair in most studies. RCTs provided 75% of the quality assessment items and cohort studies provided 82%.
Mortality: There was no statistically significant difference in mortality between corticosteroids and control for RCTs alone (RR 0.51, 95% CI 0.24 to 1.09; four studies) or for cohort studies alone (RR 0.66, 95% CI, 0.43 to 1.02; five studies). When cohort studies and RCTs were combined, use of corticosteroids was shown to be associated with a statistically significant reduction in mortality (RR 0.62, 95% CI 0.43 to 0.91; nine studies). There was evidence of statistically significant heterogeneity (I2=51%).
Mechanical ventilation: Corticosteroids were associated with a statistically significant reduction in mechanical ventilation duration (MD -4.84, 95% CI -9.28 to -0.39; three RCTs and one cohort study). There was evidence of statistically significant heterogeneity (I2=86%).
Multiple Organ Dysfunction Syndrome (MODS) score: Corticosteroids were associated with a statistically significant reduction in MODS score (MD -0.76, 95% CI -1.10 to -0.42; three RCTs and two cohort studies). No significant heterogeneity was found.
Oxygen saturation (PaO2/FiO2 ratios): Corticosteroids were associated with statistically significant greater oxygen saturation (SMD 0.64, 95% CI 0.15 to 1.13; four RCTs and two cohort studies). There was evidence of statistically significant heterogeneity (I2=78%).
There was no statistically significant difference between corticosteroids and control for length of intensive care stay, lung injury score, infection, neuromyopathy and all major adverse events. Results of additional adverse events were provided in a table. Results of subgroup analyses and meta-regression were reported.