A meta-analysis was performed on pooled data from five large double-blind studies (a total of 1057 patients), which were conducted in Asia to compare the effects of omeprazole with H2-receptor antagonists (H2-RA) in duodenal ulcer. As each study followed the same protocol and data evaluation procedures, a detailed analysis of ulcer healing, symptom relief and influence of prognostic factors across the data was possible. Patients received omeprazole 20 mg om or standard doses of H2-RA for a minimum of 2 and a maximum of 4 weeks, depending on healing (as verified by endoscopy). All efficacy analyses were based on per protocol data. The mean healing rates at 2 weeks were 72% for omeprazole and 42% for H2-RA (difference 30%; 95% CI: 24-36%; P < 0.0001) and at 4 weeks were 96% for omeprazole and 83% for H2-RA (difference 13%; 95% CI: 10-17%; P < 0.0001). In addition to treatment, ulcer size had a significant influence on healing, with large ulcers (diameter > 10 mm) taking longer to heal than small ulcers. There was no significant influence of smoking and alcohol drinking on ulcer healing. Patients on omeprazole experienced significantly less epigastric pain after 2 weeks than those on H2-RA, 79% of them being completely symptom-free on omeprazole compared with 65% on H2-RA. The incidence of adverse events was approximately 5% on each treatment and profiles were similar for each drug. It is concluded that omeprazole, even in the presence of adverse prognostic influences, results in significantly better healing of duodenal ulcer and relief from symptoms than H2-RA.