Abstract Previous studies have shown inconsistent associations of folate intake and polymorphisms of the methylenetetrahydrofolate reductase (MTHFR) gene with gastric cancer risk. Our nested case-control study within the European Prospective Investigation into Cancer and Nutrition cohort is the first prospective study of blood folate levels and gastric cancer. Gastric cancer cases (n = 247) and controls (n = 631) were matched for study center, age, sex, and time of blood donation. Two common single nucleotide polymorphisms of the MTHFR gene were determined, as were plasma concentrations of folate, cobalamin (vitamin B12), total homocysteine, and methylmalonic acid (cobalamin deficiency marker) in prediagnostic plasma. Risk measures were calculated with conditional logistic regression. Although no relations were observed between plasma folate or total homocysteine concentrations and gastric cancer, we observed a trend toward lower risk of gastric cancer with increasing cobalamin concentrations (odds ratio, 0.79 per SD increase in cobalamin; P = 0.01). Further analyses showed that the inverse association between cobalamin and gastric cancer was confined to cancer cases with low pepsinogen A levels (marker of severe chronic atrophic gastritis) at the time of blood sampling. The 677 C→T MTHFR polymorphism was not associated with gastric cancer, but we observed an increased risk with the variant genotype of the 1298 A→C polymorphism (odds ratio, 1.47 for CC versus AA; P = 0.04). In conclusion, we found no evidence of a role of folate in gastric cancer etiology. However, we observed increased gastric cancer risk at low cobalamin levels that was most likely due to compromised cobalamin status in atrophic gastritis preceding gastric cancer. (Cancer Epidemiol Biomarkers Prev 2007;16(11):2416–24)