Objectives: This meta-analysis was performed to determine whether corticosteroid therapy induces the development of peptic ulcer and other putative complications of steroid therapy.
Design: A retrospective investigation in which we analysed all the randomized, double-blind, controlled trials (RDBCT) in which steroids had been administered that we were able to identify. The number of episodes of peptic ulcer, dermatological effects, sepsis, diabetes, hypertension, osteoporosis, psychosis and tuberculosis reported in both the placebo and steroid groups were compared.
Setting: The international medical literature was analysed for any RDBCT in which any steroid or ACTH had been administered in any dosage for any duration, and any putative complication of steroid therapy was reported.
Subjects: Of 1857 articles, 93 satisfied our requirements and were analysed by the meta-analytic techniques of Peto, DerSimonian and Laird. A total of 6602 patients were included.
Main outcome measures: The relative frequencies of each of these eight 'complications' were compared in the placebo and steroid groups using conventional statistics and meta-analysis. The relative frequencies of 'annualized' subgroups of patients who received treatment for 1 to 7 days, 1 week to 1 month, 1 to 3 months and more than 3 months, were similarly analysed.
Results: Nine of 3267 patients in the placebo group (0.3%) and 13 of 3335 patients in the steroid group (0.4%) were reported to develop peptic ulcer (P > 0.05). The dermatological cosmetic effects of steroid therapy were observed more frequently in the steroid group (P < 0.001), as was diabetes (P < 0.001), hypertension (P < 0.01) and psychosis (P < 0.001). Sepsis, osteoporosis and tuberculosis all occurred more frequently in the steroid than in the placebo group, but the differences are not statistically significant.
Conclusions: Peptic ulcer is a rare complication of corticosteroid therapy that should not be considered a contraindication when steroid therapy is indicated.