Diabetes Mellitus and Risk of Future Stroke: Evidence From CHARLS and Mendelian Randomization Analyses

Abstract

Objectives: This study leveraged the China Health and Retirement Longitudinal Study (CHARLS) to explore the association between diabetes and stroke in middle-aged and older adults in East Asia and assess the causality of this relationship using Mendelian randomization.

Methods: Data from the 2011–2020 CHARLS cohort identified individuals with diabetes at baseline. Stroke incidence was self-reported through standardized questionnaires. Logistic regression and restricted cubic spline analysis examined the relationship between diabetes and stroke risk alongside nonlinear correlations between glucose levels and stroke. Mendelian randomization clarified the causal link and analyzed the mediating effect between diabetes and stroke using genetic methods.

Results: In the study population aged 45 and above, stroke incidence was 5.99% in normoglycemic, 6.82% in prediabetic, and 9.93% in diabetic individuals. Over 7 years, 473 strokes occurred. Diabetes was associated with a 1.35-fold increased stroke risk compared to normoglycemia (OR = 1.35; 95% CI: 1.03–1.79). Subgroup analyses highlighted higher stroke risks in middle-aged women, nonsmokers, and nondrinkers. Mendelian randomization supports a genetic causal relationship between diabetes and stroke. Diabetes may indirectly lead to stroke through the mediating effects of hypertension and high cholesterol.

Conclusion: The findings confirm a significant association and causal link between diabetes and stroke risk in an East Asian population. In addition, the results indicate that controlling blood glucose in prediabetic individuals reduces stroke risk, with no similar benefits in diabetes.