A meta-analysis of all nine available controlled trials of corticosteroids for adjunctive therapy for bacterial meningitis was performed. Risks of various outcomes were assessed for control and treatment groups from each study, and risk differences were determined. For each outcome a weighted average of the individual risk differences was calculated. The results show that corticosteroid administration did not reduce the risk of death or neurologic abnormality at hospital discharge or follow-up examination. Based on statistically combined results of the three most recent trials, there is evidence that dexamethasone reduces the risk of bilateral moderate or more severe hearing loss (risk difference, -9%; 95% confidence limits, -15% and -3%). However, this may be true only for children with meningitis caused by Haemophilus influenzae type b. There are inadequate data in adults or in children with meningitis due to other organisms to demonstrate the benefit of dexamethasone administration. Further study is necessary to fully assess the benefits and risks of corticosteroids for adjunctive therapy for bacterial meningitis.