A randomized trial of lansoprazole, amoxycillin, and clarithromycin versus lansoprazole, bismuth, metronidazole and tetracycline in the retreatment of patients failing initial Helicobacter pylori therapy

Abstract

Background/aim: 10-30% of the patients treated for Helicobacter pylori fail to clear the infection after initial therapy. Little is known as to the efficacy of retreatment regimens in these patients. Proton pump inhibitor (PPI) -based triple and quadruple therapies demonstrate efficacies of 80-90% as initial therapy for H. pylori infection, but whether these regimens are as effective when used for retreatment is unknown. The efficacy of a metronidazole-containing regimen in this situation is also unknown. Our aim was to compare the efficacy of a nonmetronidazole-containing PPI-based triple versus a PPI-based quadruple therapy containing metronidazole in patients failing previous H. pylori therapy.

Methods: 48 patients were enrolled in this study at two sites after failure of previous H. pylori therapy as determined by a positive (14)C-urea breath test. Patients were stratified by prior treatment with a metronidazole-containing regimen and were then randomized to either lansoprazole (L) 30 mg twice daily, amoxycillin (A) 1,000 mg twice daily, and clarithromycin (C) 500 mg twice daily for 14 days (LAC) or L 30 mg four times daily, bismuth subsalicylate (B) 2 tablets four times daily, metronidazole (M) 250 mg four times daily and tetracycline (T) 250 mg four times daily for 14 days (LBMT). Side effects and compliance (pill count) were assessed at the completion of therapy. A repeat (14)C-urea breath test was performed 4 or more weeks after completion of therapy, and cure was defined as a negative test result.

Results: 48 patients (16 males and 32 females) were enrolled in this study. 20 patients received LAC (18 prior M), and 28 received LBMT (23 prior M). Per protocol and intention-to-treat efficacies were 82% (95% CI 64-100%) and 75% (95% CI 56-94%) for LAC and 80% (96% CI 64-96%) and 71% (95% CI 54-88%) for LBMT (p = 0.85 per protocol and p = 0.78 intention to treat between LAC and LBMT), respectively. The compliance (> or =80% of pills taken) was found to be 89% in both treatment groups. Side effects were noted in 84% for LAC and in 82% for LBMT, but were mild and did not cause discontinuation of therapy.

Conclusions: PPI-based triple and quadruple therapy with both LAC and LBMT are effective in retreating patients failing initial metronidazole-based H. pylori therapies. LAC was not statistically superior to LBMT as a 'retreatment' regimen in this clinical situation, but the small sample size and wide confidence limits do not preclude the possibility of a smaller but significant difference in efficacy between the regimens. To determine whether LAC or LBMT is as effective for retreating patients failing non-metronidazole-containing regimens requires further study.