An interim analysis has been performed of a multicenter randomized double-blind parallel-group comparison of misoprostol 50 micrograms (low dose), misoprostol 200 micrograms (high dose), and cimetidine 300 mg in out-patients with endoscopically diagnosed benign gastric ulcers. The drugs were administered q.i.d. for four weeks. Endoscopic assessments were made before treatment and at four weeks. Clinical assessments and laboratory tests were made before treatment and at two and four weeks. At the interim stage 421 subjects of the target 600 patients had been recruited by 22 principal investigators in 13 countries. Demographic data were similar for all groups with the exceptions of a lower proportion of males and smokers in the cimetidine group. Therapeutic success was defined as complete healing of all gastric ulcers. Cure rates at four weeks, for all eligible patients who received at least one dose of study medication (cohort 1), were significantly lower (P = 0.002) for misoprostol 50 micrograms (41.2%) than for misoprostol 200 micrograms (58.2%). There was no significant difference (P = 0.86) in cure rates between the high dose misoprostol and cimetidine (59.6%) groups. Cure rates were approximately 20% higher for non-smokers than smokers in both misoprostol groups, an effect not observed with cimetidine. In all three treatment groups the incidence of complaints, including diarrhea, was similar and therapy was well tolerated.