Objective: Gastric cancer is a major global health challenge and a leading cause of cancer mortality. Despite therapeutic advances, most patients do not survive beyond one year. In 2024, the U.S. National Cancer Institute (NCI) Think Tank on Advancing Gastric Cancer Prevention developed a gastric cancer prevention roadmap focussed on mitigating Helicobacter pylori infection and improving detection of precursor lesions. For the vast majority, H. pylori infection is a benign condition which usually remains undiagnosed, and its prevalence has been declining globally; gastric intestinal metaplasia is also a benign condition and only a small fraction of premalignant intestinal metaplasia lesions progress to adenocarcinoma. We aimed to establish a global gastric cancer consortium within the Pooling Project of Prospective Studies of Diet and Cancer, a working group of the NCI Cohort Consortium, to identify additional modifiable risk factors to inform more effective disease prevention.
Methods: We harmonised and pooled individual-level data from over 2 million participants across 21 prospective cohorts, among whom 8,803 developed incident invasive gastric adenocarcinoma (22% cardia, 53% non-cardia; 25% unclassified/overlapping) during median follow-up of 9–28 years across cohorts. Baseline diet was assessed using food frequency questionnaires or diet histories with standardised food group definitions applied across studies. A range of body size measures, as well as demographic, medical, and other lifestyle factors, was also assessed at baseline. Cox proportional hazards regression estimated study- and sex-specific associations between potential risk factors (‘exposures’) and gastric cancer risk (‘endpoint’), adjusting for confounders including smoking habits and H. pylori (in a subset). Study- and sex-specific estimates were combined using random-effects meta-analysis. Whether associations varied by tumour subtype or by population subgroup was assessed using an aggregated dataset where individual level data from each study were combined into a single dataset.
Results: In analyses completed thus far, obesity and higher processed meat intake were associated with increased gastric cancer risk (multivariable-adjusted hazard ratio, HR, 1.18 for BMI ≥30 vs 18.5-<25 kg/m2; 95% confidence interval, CI, 1.07-1.30 and HR, 1.15 for ≥45 vs <3 g/day; 95% CI, 1.00-1.32) while higher dietary fibre intake, specifically from grains, was associated with lower risk (HR 0.87 for ≥25 vs <10 g/day; 95% CI, 0.79–0.96), all independent of the effects of H. pylori infection. Higher alcohol consumption was associated with increased risk only for Asian cohorts (HR, 1.21 for ≥30 vs 0.1-<5 g/day; 95% CI, 1.02-1.42). Higher egg intake was also positively associated with gastric cancer risk, while no associations were observed for unprocessed red meat, poultry, or seafood. These associations were largely consistent for cardia and non-cardia as well as diffuse- and intestinal-type cancers (P[heterogeneity between subtypes]>0.05). Ongoing analyses will evaluate associations for additional exposures.
Conclusions: We provide new and robust evidence from by far the largest prospective dataset assembled to date targeting gastric cancer prevention focused on modifiable risks. Our findings support translation of evidence into population-level prevention strategies that complement H. pylori infection control and early detection, highlighting opportunities to reduce gastric cancer burden globally through integrated lifestyle and public health interventions.
 

Schematic illustration of modifiable risk factors and prevention of gastric cancer