We performed a meta-analysis of the literature on the treatment of established rheumatic carditis to determine if corticosteroid therapy is superior to salicylates in preventing the sequela of inflammation--valvular damage. We identified 22 reports of comparative trials published since the introduction of corticosteroids in 1949. Five of the 22 studies met the criteria we established for the meta-analysis, which included using randomization and a 1-year follow-up for the presence of a new pathologic apical systolic murmur. Based on the meta-analysis, the advantage of corticosteroid treatment over salicylates in preventing a pathologic murmur at 1 year posttreatment is not statistically significant (estimated odds ratio 0.88; 95% confidence interval: 0.53 to 1.46). However, the meta-analysis is dominated by 1 large negative trial, and there was significant heterogeneity in the results obtained from the studies in the meta-analysis; thus, the question of whether corticosteroid therapy is marginally superior to salicylates for the prevention of valvular heart disease from rheumatic fever remains open.