Worldwide and Regional Efficacy Estimates of First-line Helicobacter pylori Treatments: A Systematic Review and Network Meta-Analysis

Abstract

Background: Eradication of Helicobacter pylori infection is challenging. We aimed to determine the optimal first-line H. pylori treatments at global and regional levels.

Methods: We searched Embase, PubMed, Cochrane CENTRAL, Web of Science, Scopus, WHO ICTRP, ClinicalTrials.gov, and ISRCTN registry, for randomized controlled trials published during 2011-2020. Utilizing a network meta-analysis in a Bayesian framework, success rates of 23 regimens were compared. The effect size was standardized risk ratio (RR) with 95% credible interval (CrI). Pooled eradication rate (ER) with 95% CrI was also reported for top combinations. The reference regimen was 7-day clarithromycin-based triple therapy.

Results: This review identified 121 trials comprising 34,759 participants. Globally, 14-day levofloxacin-based sequential therapy was the most efficient (RR: 1.43; 95% CrI, 1.26-1.59) with low certainty of evidence, followed by modified bismuth-containing quadruple therapy (proton pump inhibitor+bismuth compounds+clarithromycin+amoxicillin) for 10 days (RR: 1.35; 95% CrI, 1.22-1.48) and 14 days (RR: 1.27; 95% CrI, 1.12-1.42), and 14-day hybrid therapy (RR: 1.27; 95% CrI, 1.19-1.36). The corresponding ERs were 98.7% (95% CrI, 86.9-100.0), 93.2% (95% CrI, 84.2-100.0), 87.6% (95% CrI, 82.1-93.8), and 87.6% (95% CrI, 77.3-98.0), respectively. Continentally, the most effective combinations were: 10-day clarithromycin-based sequential therapy [(RR: 1.21; 95% CrI, 1.02-1.42), (ER: 89.5%, 95% CrI, 75.5-100.0)] for Africa, 14-day levofloxacin-based sequential therapy [(RR: 1.41; 95%CrI, 1.27-1.58), (ER: 98.7%, 95% CrI, 88.9-100.0)] for Asia, and 14-day clarithromycin-based triple therapy [(RR: 1.58; 95% CrI, 1.25-2.04), (ER: 94.8%; 95% CrI, 75.0-100.0)] for Europe. For Northern America, no sufficient data were found for network meta-analysis. In South America, none of the combinations were superior to the reference regimen.

Conclusion: Although results of this network meta-analysis revealed optimal combinations for empiric therapy, the treatment preference would be based on the local pattern of antibacterial resistance, when the necessary information exists.