To assess the efficacy of cimetidine in treating and preventing gastric mucosal lesions associated with nonsteroidal anti-inflammatory drug (NSAID) therapy (NSAID gastropathy), we endoscopically studied 104 patients taking NSAIDs for a variety of rheumatic diseases. Fifty-six percent (22/43) of patients randomized to cimetidine 300 mg four times a day and 52% (22/42) of those randomized to placebo showed progression of endoscopic lesions during the eight-week short-term phase. Thirty-nine patients whose endoscopic lesions improved were then randomized to a ten-month maintenance regimen of either cimetidine 400 mg at bedtime or placebo. Fifty percent (7/14) of placebo-treated and 42% (5/12) of cimetidine-treated patients showed progression of lesions during the maintenance phase. The failure of cimetidine to offer any significant benefit under these protocol conditions reflects the fundamental difference in pathophysiologic features between classic acid-mediated ulcer disease and NSAID gastropathy.