Background. Despite steady declines in incidence, gastric cancer (GC) remains the fifth leading cause of cancer-related mortality. Infection with Helicobacter pylori (H.pylori) is the primary causal factor; however, other modifiable (diet, lifestyle) and metabolic risk factors also contribute to GC development. Gastric carcinogenesis is a multifactorial process that follows a long precancerous cascade, progressing through atrophic gastritis, which offers opportunities for prevention and early detection.

Objectives. In this study, we aimed to identify novel risk factors, exposures and blood biomarkers that may promote GC development.

Methods. A case-control study nested within six countries from EPIC was conducted, including 437 GC-cases and their 437 matched controls. H.pylori infection status, and pepsinogen (Pg) I and II levels were determined using ELISA. Severe atrophic gastritis (SAG) was serologically defined as PgI < 30 µg/L or PgI-to-PgII ratio <3. Untargeted plasma metabolomic profiles were obtained using Liquid-Chromatography Mass-Spectrometry. We investigated GC risk association with H.pylori infection, Pg levels, dietary intakes (fruits and vegetables, red-and-processed meat, dairy), lifestyle factors (smoking, alcohol consumption, waist-to-hip ratio, educational level) and metabolic features. Conditional logistic regression was used to compute odds ratios (OR) and 95% confidence intervals (95% CI), and false discovery rate correction for multiple comparisons was applied to identify metabolic features associated with GC. In planned secondary analyses, we will conduct mediation analyses to explore how decreases in Pg levels impact the association between H.pylori and GC risk, while considering the follow-up duration. We will also examine associations between the identified metabolic features and SAG.

Results. Among the study participants, 67% were infected with H. pylori, and 20% presented a SAG. Infection prevalence varied from 81% in Spain to 50% in France and was higher in participants with no education (84%) than in the highest education group (60%). Infection was more prevalent in non-cardia GC cases (86%) than in cardia cases (42%). SAG was most frequent in Germany (26%), Italy (19%), and the Netherlands (18%), and 72% of SAG cases developed GC. Infected participants and those with SAG had an increased risk of GC (respectively, OR= 2.60 (95%IC: 1.83-3.68) & 3.30 (2.18-4.99)). A decrease in the PgI-to-PgII ratio increased GC risk 1.24-fold (1.17-1.32). Additionally, smoking, red-and-processed meat intake, and an elevated waist-to-hip ratio were significantly associated with a greater GC risk. Moreover, 16 molecular features were identified, and elevated levels of these features were associated with increased GC risk; ORs ranged from 1.12 to 1.71. The secondary analyses described above are currently underway.

Conclusion. Our study, conducted in a large prospective European cohort with pre-diagnostic biospecimen collections, detailed questionnaire data and a long follow-up, allowed deeper evaluation of GC risk. H.pylori infection and serologically defined SAG were strongly associated with higher GC risk, as well as various dietary and lifestyle factors. Additionally, we identified 16 plasma metabolic features associated with increased GC risk. Ongoing efforts aim to clarify the association of H.pylori with GC risk, considering the effect of precancerous mucosal changes, and to elucidate the role of these factors and metabolic features in gastric carcinogenesis.