Abstract 13226: Low-Carbohydrate Diets and Risk of Type 2 Diabetes in U.S. Men and Women

Introduction: Evidence regarding associations between overall low-carbohydrate diets (LCDs) and the risk of developing type 2 diabetes (T2D) remains to be mixed. In the current analysis, we aimed to prospectively assess associations of various LCDs that emphasize different quality of protein, fat, and carbohydrate with T2D risk in U.S. men and women. Methods: A prospective cohort study was conducted in 203,541 men and women participating in the Nurses’ Health Study (NHS), NHSII, and Health Professionals Follow-Up Study who were free of T2D, cardiovascular disease, and cancer at baseline and were followed for up to >30 years. Five LCD scores based on sources of protein, fat, and carbohydrate differentiated by their qualities were derived every four years based on dietary assessments by a validated food frequency questionnaire. Results: During 5,095,048 person-years of follow-up, 19,675 T2D cases were documented. In the multivariable model, a higher overall LCD score was associated with higher T2D risk in a dose-response manner: the hazard ratio (HR) comparing highest vs. lowest quintile was 1.28 (95% confidence interval: 1.22-1.34; P -trend <0.001). A vegetable-based LCD score that emphasizes vegetable protein and fat intake was associated with 6% lower T2D risk (0.94 [0.90-0.98; P -trend =0.003]), and a healthy LCD score that further de-emphasizes refined carbohydrate intake was associated with 15% lower T2D risk (0.85 [0.81-0.89; P -trend <0.001]). In contrast, LCD scores emphasizing high animal protein and fat intake were associated with higher T2D risk. The HRs (95% CIs) comparing extreme quintiles were 1.35 (1.29-1.42; P -trend <0.001) for an overall animal-based LCD score and 1.39 (1.32-1.46; P -trend <0.001) when such a score that further de-emphasizes whole grain intake. Conclusions: LCDs were not beneficial for the primary prevention of T2D unless they emphasize the quality of the macronutrient compositions.