The study objective was to study the ulcer healing effects and safety of the proton pump inhibitor, omeprazole, given in a dose of 20 mg once daily before breakfast. The study design was a randomized, double-blind, multicenter comparison of omeprazole and placebo using endoscopy to assess ulcer healing after two or four weeks of therapy. One hundred fifty-three patients with endoscopically documented active duodenal ulcer were studied. One hundred two patients received omeprazole and 51 received placebo. Patients in both groups were similar with regard to age, sex, duration of disease, initial ulcer size, smoking history, and alcohol use. A "per protocol" analysis of healing rates showed a significant advantage for omeprazole (P less than 0.01) at both week 2 (41% vs 13%) and week 4 (75% vs 27%). Concomitant factors (including smoking and ulcer size) did not alter the significance of the differences in healing rates between omeprazole and placebo. Complete relief of day and night pain was more often achieved (P less than 0.01) in the omeprazole group. "All-patients treated" analyses for healing and pain relief gave results similar to the respective "per protocol" analyses. Omeprazole was well tolerated; fewer patients had clinical and laboratory adverse experiences in the omeprazole group than in the placebo group. Fasting serum gastrin levels increased with omeprazole therapy (mean 34.9 to 73.5 pg/ml) but exceeded the normal range (greater than 150 pg/ml) in only 12.3% of patients. Two weeks after therapy was stopped, serum gastrin levels showed a decrease toward baseline but had not yet completely returned to pretreatment levels (mean 49.7 pg/ml).(ABSTRACT TRUNCATED AT 250 WORDS)