The aim of this randomized, multicenter, double-masked, parallel-group study was to compare the efficacy of lansoprazole with that of omeprazole monotherapy in duodenal ulcer healing and prevention of relapse. A total of 251 patients with duodenal ulcer were treated with either lansoprazole 30 mg/d (n = 167) or omeprazole 40 mg/d (n = 84). Patients with healed ulcers were then randomly allocated to 12 months of maintenance therapy with lansoprazole 15 mg/d (n = 74), lansoprazole 30 mg/d (n = 71), or omeprazole 20 mg/d (n = 73). Healing rates at 4 weeks (intent-to-treat analysis) were 93.9% (95% confidence interval [CI], 90.2% to 97.6%) with lansoprazole and 97.5% (95% CI, 93.7% to 100%) with omeprazole; there were no significant differences between groups. Endoscopic relapse rates after 6 months were 4.5% (95% CI, 0% to 10.6%) with lansoprazole 15 mg, 0% with lansoprazole 30 mg, and 6.3% (95% CI, 1.5% to 12.5%) with omeprazole 20 mg, compared with 3.3% (95% CI, 0% to 8.2%), 0%, and 3.5% (95% CI, 0% to 8.8%), respectively, at 12 months. Again, there were no significant differences between groups. The incidence of adverse events during acute treatment was 6.0% and 7.1% in the lansoprazole and omeprazole groups, respectively; during maintenance therapy, the incidences were 12.2% (lansoprazole 15 mg), 5.6% (lansoprazole 30 mg), and 11.0% (omeprazole 20 mg). Within treatment groups, pain was significantly ameliorated after the acute phase but not after maintenance therapy (P < 0.05); no differences were observed between groups. Gastrin values increased significantly after acute therapy (P < 0.05), persisted at these increased levels during maintenance therapy, and returned to normal after 6-month follow-up. Both lansoprazole and omeprazole were highly effective and well tolerated in the treatment of duodenal ulcer; relapse rates were similar for all doses studied. Thus no additional benefit is to be gained from using a proton-pump inhibitor at a dose > 15 mg lansoprazole to prevent relapse.