There were no clear clinically relevant differences between glucocorticosteroid treatment and placebo/no treatment for adults with alcoholic hepatitis for any measured outcome (all‐cause mortality, liver‐related mortality, quality of life, complications, and serious adverse events). However, where quality of the evidence was assessed, this was very low or low. In addition, when scenario analyses were conducted to test for attrition bias (assumptions made about the outcomes for people unavailable for assessment at follow‐up ‐ had an event or not), one scenario led to results in favor of placebo for mortality and serious adverse events, further decreasing confidence in pooled estimates. As a result, no firm conclusions of equivalence can be drawn.