To elucidate whether pretreatment with omeprazole decreases the cure rate of Helicobacter pylori infection with a new quadruple therapy, and thus, whether this pretreatment should not be used in clinical practice, we conducted a randomized trial. Ninety patients with chronic peptic ulcer disease and nonulcer dyspepsia, with biopsy-proven H. pylori infection were randomly assigned to the two following regimens: Group 1 (n = 45) received omeprazole 20 mg once daily for 2 weeks (days 1-14), and 500 mg amoxicillin granules and 250 mg metronidazole thrice daily, and roxithromycin 150 mg twice daily for 1 week (days 8-14), Group 2 (n = 45) received the same antibiotic treatment as group 1 for 1 week (days 1-7), in addition to omeprazole treatment for 2 weeks (days 1-14). Four weeks after the treatment ended, endoscopy was repeated, with two biopsy specimens each taken from the antrum and the corpus (total of four specimens) for a urease test, histological analysis, and culture to establish cure of infection. A patient was regarded as cured only if all three methods gave negative results for H. pylori. In the intention-to-treat analysis, 42 of 45 patients (93.3%; 95% confidence intervals [CI], 81.7%-98.6%) in group 1 were cured compared with 43 of 45 patients (95.6%; 95% CI, 84.9%-99.5%) in group 2. In the per-protocol analysis, the corresponding figures were 42/44 (95.5%; 95% CI 84.5%-99.4%) and 43/44 (97.7%; 95% CI, 88.0%-99.9%). There were no significant differences in the cure rate between the two groups on either analysis. All patients, except for one who had an allergic reaction, completed the treatment regimens. Fifty to sixty percent of the patients had no side effects while the rest had mild to moderate side effects. The new quadruple therapy consisting of omeprazole, amoxicillin, metronidazole, and roxithromycin appears suitable for use in clinical practice, as the cure rate was 95% and no severe side effects were observed. Pretreatment with omeprazole did not reduce the cure rate for this new quadruple therapy.