LAUSR

BackgroundHelicobacter pylori (H. pylori) is considered to be the most important risk factor for gastric cancer (GC). The International Agency for Research on Cancer reported that H. pylori eradication could reduce the risk of developing GC. Several clinical studies have investigated this relationship as well; however, their results are inconsistent owing to the varied inclusion criteria. To address the effect of H. pylori eradication on GC incidence, we conducted a comprehensive meta-analysis with several subgroup analyses to resolve these inconsistencies.MethodsWe searched MEDLINE and Ichushi-Web to identify randomized control trial and cohort study articles (English or Japanese) through December 2016. Manual searches were also conducted to identify unlisted references in these databases. Eligible studies reported GC incidence as an outcome, with comparisons between H. pylori eradication and control groups. Subgroup analyses were conducted by country, conditions at baseline, and follow-up periods.ResultsWe selected 28 studies among 1583 references in the databases and 4 studies by manual searches. The H. pylori eradication group showed significantly lower risk of GC [odds ratio (OR) 0.46; 95% confidence interval 0.39–0.55]. The subgroup analyses indicated that the beneficial effect of eradication was greater in Japan (OR 0.39; 95% CI 0.31–0.49), particularly among those with benign conditions (OR 0.32; 95% CI 0.19–0.54), although none of them was statistically significant. However, reduction of gastric cancer after eradication was significantly greater (p = 0.01) in the groups with long-term (5 years or longer) follow-up (OR 0.32; 95% CI 0.24–0.43) as compared to those with shorter follow-up (less than 5 years) (OR 0.55; 95% CI 0.41–0.72).ConclusionReal world data showed that large-scale eradication therapy has been performed mostly for benign conditions in Japan. Since eradication effects in preventing gastric cancer are conceivably greater there, GC incidence may decline faster in Japan than expected from the previous meta-analyses data which were based on multi-national, mixed populations with differing screening quality and disease progression.