We performed an open, prospective, randomized, three-cell, 6-month clinical trial on the prevention of duodenal ulcer (DU) relapse, comparing three omeprazole schedules, i.e. 20 mg daily, 20 mg every other day (e.o.d.) and 40 mg on Saturdays and Sundays (S/S). Diagnosis of either healed or relapsed DU was on an endoscopic basis. Follow-up visits were performed at 3-monthly intervals with endoscopy at the baseline, after 6 months and at every symptomatic relapse. METHODS chi 2 test with standardized deviates, Yates' corrected chi 2 test and analysis of variance (one-way). One hundred and fifteen patients were randomized to receive omeprazole 20 mg/day, 123 omeprazole 20 mg e.o.d. and 115 40 mg S/S. Twenty-eight dropped out (11, 8 and 9, respectively). Demonstrated ulcer relapse rates were 5.7% with omeprazole 20 mg/day, 18.1% with 20 mg e.o.d. and 17.6% with 40 mg S/S (p = 0.0124, 'per-protocol' analysis). No clinically significant adverse effects were recorded. In conclusion, of the three schedules studied, omeprazole 20 mg/day proved the most effective maintenance treatment for healed DU.