Meta‐analysis: Use of proton pump inhibitors and risk of gastric cancer in patients requiring gastric acid suppression

BACKGROUND: Proton-pump inhibitors (PPIs) are suspected to increase the risk of gastric cancer. AIM: To assess the risk of gastric cancer associated with the use of PPIs. METHODS: We systematically searched Medline/PubMed, Embase and Scopus databases (until June 1, 2022) for randomised and non-randomised studies (NRS) of the association between PPIs and gastric cancer having considered Histamine-2 receptor antagonists (H2RAs) users as controls. We chose this comparison to minimise confounding by indication, and focus on patients requiring gastric acid suppression. Two authors independently extracted study data and assessed each study's risk of bias. Maximally-adjusted relative risk (RR) estimates were extracted. Heterogeneity and small-study effect were examined, and summary estimates were calculated using random- and fixed-effect models. Stratified analyses and meta-regression were employed to explore heterogeneity. We used GRADE to evaluate the certainty of evidence. RESULTS: Of 8375 records, 12 NRS (>6 million patients; 11,554 gastric cancers) and two randomised clinical trials (498 patients; 1 gastric cancer) fulfilled the eligibility criteria. Randomised evidence was very imprecise and provided very-low certainty evidence. Meta-analysis of six NRS providing a comprehensive adjustment for confounding (2.5 million patients; 7372 gastric cancers) did not show any association between PPIs and gastric cancer (RR_random  = 1.07, 0.97-1.19; RR_fixed  = 1.05, 0.98-1.12). The certainty of the evidence was low. No convincing evidence of dose-response, or increased risk with long-term use, was found. Lack of or minimal adjustment for confounding was associated with larger effect sizes. CONCLUSIONS: We found no association between PPIs and gastric cancer in NRS having adequately controlled for confounding. Published studies may suffer residual confounding. SYSTEMATIC REVIEW REGISTRATION: CRD42022335971.