Blood lipids-related dietary patterns derived from reduced rank regression are associated with incident type 2 diabetes

Abstract

Blood lipids play a critical role in the pathogenesis of type 2 diabetes, and they are closely related to dietary factors. However, the associations between blood lipids-related dietary patterns and risk of type 2 diabetes are controversial and not fully clear. In this study, we aimed to derive dietary patterns that explained variation in blood lipids and to investigate their associations with incident type 2 diabetes. The analysis was based on 39,000 women and 25,777 men participating in the Lifelines cohort study (aged 18-65 years, mean 43.2 years for women and 43.5 years for men). Dietary intake was measured using a 110-item semi-quantitative food frequency questionnaire. Reduced rank regression was used to derive dietary patterns with blood lipids (HDL-cholesterol, LDL-cholesterol, triglycerides, total cholesterol, and total cholesterol:HDL-cholesterol ratio) as response variables for women and men separately. The first dietary pattern identified for each sex was selected because they explained the largest variance in blood lipids. The associations between the identified dietary patterns and incident type 2 diabetes were subsequently investigated using multivariate logistic regression models. All analyses were performed separately for women and men. During an average follow-up of 43 months, 479 new cases (incidence 0.74%) of type 2 diabetes were identified. Using reduced rank regression, we identified two sex-specific blood lipids-associated dietary patterns characterized by high intake of sugary beverages, added sugar, and low intake of vegetables, fruits, tea, and nuts/seeds. These two sex-specific dietary patterns were similar in food groups but differed in factor loadings. High dietary pattern scores were associated with increased risk of type 2 diabetes after adjustment for age, total energy intake, body mass index, waist-hip ratio, and blood pressure (ORs for the fifth quintile [Q5] using the first quintile [Q1] as reference, 1.87 [95% CI 1.23, 2.83] for women [P-trend<0.001], and 1.72 [95% CI 1.11, 2.66] for men [P-trend=0.018]). The associations were attenuated but remained significant after further adjustment for lifestyle and socio-economic factors. Dietary patterns associated with adverse blood lipids are associated with incidence of type 2 diabetes. The present study provides new insights in optimizing blood lipids for the prevention of type 2 diabetes through dietary approaches.