Six trials (389 patients) reported the primary outcome, with a further trial reporting mortality. Sample size ranged from 20 to 87 patients. There was considerable variation in both quality (two trials were quasi-randomised) and the dose and duration of corticosteroid treatment.
Corticosteroids reduced the risk of fat embolism syndrome by 78% (RR 0.22, 95% CI 0.08 to 0.57; NNT 8, 95% CI 5 to 13). The authors stated that funnel plots (not presented) did not suggest publication bias.
Corticosteroids also reduced the risk of hypoxia by 61% (RR 0.39, 95% CI 0.21 to 0.71), but did not have a statistically significant effect on petechiae, mortality or infection. No trials measured avascular necrosis.
There were no significant differences between subgroups based on dose, quality score, degree of blinding or randomisation.
Heterogeneity between trials was not statistically significant in any of the analyses.