Seven RCTs (n=2,632) were included in the review in which 299 patients experienced PEP. Of these, 157 (12 per cent) were randomised to corticosteroids and 142 (11 per cent) received placebo. Two trials were considered low quality and five considered high quality.
There were no significant differences between groups in PEP (seven trials, OR 1.13, 95% CI: 0.89, 1.44, p=0.32), severe PEP (five trials, OR 1.61, 95% CI: 0.74, 3.52, p=0.23) or post-ERCP hyperamylasemia (two trials, OR 0.92, 95% CI: 0.57, 1.48, p=0.73). There was no association between corticosteroid use and case fatality, although adverse effects were reported in three trials. Post-ERCP abdominal pain was not reported in the included trials. No significant heterogeneity was observed for any outcomes.
Subgroup analyses stratifying results by setting and quality did not change the results. Sensitivity analyses showed that outcomes were consistent based on allocation concealment, blinding and trials published as abstracts.
There was no evidence of publication bias.