Abstract 143: Stroke-Associated Metabolites And Dietary Patterns In The Reasons For Geographic And Racial Differences In Stroke (REGARDS) Cohort

Abstract

Introduction: Several diet patterns have previously been associated with incident stroke risk in the REasons for Geographic And Racial Differences in Stroke (REGARDS) study. We used metabolomics to assess biological state changes that could link diet exposures to stroke risk. Hypothesis: Metabolites linked to stroke risk are correlated with adherence to unhealthy and protective dietary patterns. Methods: Using liquid chromatography-tandem mass spectrometry, we analyzed 162 metabolites in baseline plasma samples from REGARDS subjects with or without incident stroke (n = 2,249). Survey-weighted Cox models adjusted for age, race, and sex identified metabolites associated with incident ischemic stroke. Candidate metabolites were correlated with diet patterns using linear regression. Results: Correlations between stroke-associated metabolites and diet patterns linked to higher (Southern diet) or lower (Mediterranean and plant-based diets) stroke risk were consistent with known effects on stroke risk conferred by each metabolite and diet pattern. Notably, the “harmful” nucleoside guanosine was positively correlated with Southern diet (β = 0.11, FDR = 4.7x10 -3 ) and negatively with Mediterranean (β = -0.11, FDR = 3.2x10 -2 ) and plant-based diets (β = -0.13, FDR = 1.6x10 -4 ). In contrast, the “protective” indole-3-propanoic acid, derived from tryptophan metabolism by gut microbiota, was negatively correlated with Southern diet (β = -0.23, FDR = 3.0x10 -6 ) but positively with Mediterranean (β = 0.15, FDR = 6.1x10 -3 ) and plant-based diets (β = 0.23, FDR = 1.5x10 -5 ). Inverse odds ratio weighting mediation analyses showed that guanosine levels partly mediated the protective effect of a plant-based diet (indirect effect hazard ratio [HR] = 0.94, 95% CI = 0.86-0.99) and Mediterranean diet (indirect effect HR = 0.94, 95% CI = 0.85-0.99) on stroke risk. We also examined associations between stroke-linked metabolites and 265 specific food items, finding that unhealthy intakes were positively correlated with harmful metabolites and vice versa. Conclusion: Metabolites relevant to ischemic stroke risk correlate with broader diet patterns and specific food exposures. Metabolome perturbations might link dietary intake and the gut microbiota to stroke risk.