Abstract A133: Associations between Dietary and Circulating B-Vitamins and Incident Stroke: Evidence from Two Large Prospective United States Cohorts

Introduction: While B-vitamins are promising agents for primary stroke prevention, evidence is largely confined to the homocysteine-lowering effects of folate, pyridoxine, and cobalamin. The roles of other B-vitamins, their interactions, and the relevance of different biomarkers in fortified populations remain unclear. Methods: In two large prospective cohorts of women and men in the United States (US), the Women's Health Initiative (WHI, 154,718 postmenopausal women) and the All of Us Research Program (AoU, 102,680 adults), we investigated the roles of B-vitamins (dietary exposures and plasma biomarkers) in the development of incident stroke using multivariable Cox models. Mendelian randomization (MR), interaction among B-vitamins, mediation by homocysteine and subgroup analyses by known risk factors were also evaluated. Results: In the WHI (7,577 incident strokes in a median follow-up of 19.2 years), higher baseline intakes of thiamin, riboflavin, niacin, pyridoxine, and folate were prospectively and significantly associated with lower stroke risk (hazard ratios HRs for highest vs. lowest quintile: 0.86-0.91). In AoU (1,745 incident strokes in a median follow-up of 5.9 years), higher plasma pyridoxine (HR: 0.29; 95% confidence interval CI: 0.09, 0.92 for Q5 vs Q1) and folate (HR: 0.79; 95% CI: 0.64, 0.99 for Q4 vs Q1) were associated with lower stroke risk, consistent with MR analysis for small vessel stroke. Antagonistic interactions were observed for some dietary pairs, while the combination of high plasma pyridoxine with folate or cobalamin showed the strongest association with lower risk. Homocysteine lowering pathway mediated only 6% (95% CI: -19%, 57%) of the association between plasma folate and stroke. No significant differences across strata of known risk factors were observed. Conclusions: In two large national cohorts of US adults, we observed consistent and significant associations of plasma levels of folate and pyridoxine with stroke risk, independent of known risk factors as well as the homocysteine-dependent pathway. Thiamin, riboflavin, and niacin were also linked to lower stroke risk, with evidence of interactions among them.