Abstract P317: Dietary Carbohydrate Quantity and Quality Measures and Risk of Stroke

Abstract

Background: Dietary carbohydrates with high glycemic index and low in cereal fiber have been associated with increased risk of diabetes, a risk factor for stroke. Few prospective studies have examined the relation between measures of carbohydrate quality and risk of incident stroke. The relation between carbohydrate to fiber ratio and stroke risk remains unknown. Methods: We prospectively examined the associations of dietary carbohydrate quantity and quality measures (total carbohydrate, fiber [total, cereal, fruit and vegetable], glycemic index and load, and carbohydrate/fiber ratio) and risk of incident stroke among 40,689 men in the Health Professionals Follow-up Study, aged 40 to 75 years and free of diagnosed cardiovascular disease and cancer at baseline in 1986. Diet was collected using a validated questionnaire every 4 years. We calculated the hazard ratio of total, ischemic, and hemorrhagic strokes by quintiles of energy-adjusted intake of dietary variables using multivariate Cox proportional hazard models. Multivariate models were adjusted for age, lifestyle (BMI, smoking, alcohol, physical activity, multivitamins and aspirin use), history of diseases (family history of CHD, baseline history of hypertension, diabetes and hypercholesterolemia) and diet (total energy, red meat, magnesium, polyunsaturated fat to saturated fat ratio and trans fat). Results: During 24 years of follow-up, 1,530 total stroke events (including 847 ischemic, 178 hemorrhagic and 505 unspecified types) were documented. In multivariate analyses adjusted for age, lifestyle risk factors, history of diseases and total energy intake, carbohydrate (RR=0.74, 95% CI: 0.61 - 0.91, Ptrend 0.008), cereal fiber (RR=0.77, 95% CI: 0.64 - 0.93, Ptrend 0.003) and glycemic load (RR=0.78, 95% CI: 0.64 - 0.95, Ptrend 0.03) were associated with reduced risk of total stroke, comparing the extreme quintiles. With further adjustment for dietary variables, these associations were attenuated to non-significant levels, carbohydrate (RR=0.84, 95% CI: 0.67 - 1.05, Ptrend 0.20), cereal fiber (RR=0.91, 95% CI: 0.73 - 1.14, Ptrend 0.40) and glycemic load (RR=0.84, 95% CI: 0.68 - 1.05, Ptrend 0.26). Vegetable, fruit and total fiber, glycemic index and carbohydrate to fiber ratio were also not significantly associated with incident stroke risk. The results for total stroke were relatively similar to those for ischemic and hemorrhagic stroke types. Conclusions: Consumption of high quality carbohydrate foods did not appear to confer protection against incident stroke risk among this population of US men.