Statins, Type 2 Diabetes, and Body Mass Index: A Univariable and Multivariable Mendelian Randomization Study
Guoyi Yang, C. Mary Schooling
Abstract
CONTEXT: Statins and possibly other lipid modifiers increase type 2 diabetes risk and body mass index (BMI). However, to what extent BMI mediates the diabetogenic effects of lipid modifiers remains unclear. OBJECTIVE: We used Mendelian randomization (MR) to investigate the effects of commonly used lipid modifiers on type 2 diabetes risk and glycemic traits, and any mediation by BMI. METHODS: Using established genetic variants to mimic commonly used lipid modifiers (ie, statins, PCSK9 inhibitors, and ezetimibe), we assessed their associations with type 2 diabetes risk, glycated hemoglobin (HbA1c), fasting insulin, fasting glucose, and BMI in the largest relevant genome-wide association studies (GWAS) in people of European ancestry, and where possible, in East Asians. We used multivariable MR to examine the role of lipid modifiers independent of BMI. RESULTS: Genetically mimicked effects of statins and ezetimibe, but not PCSK9 inhibitors were associated with higher risk of type 2 diabetes (odds ratio [OR] 1.74 [95% CI, 1.49 to 2.03]; 1.92 [1.22 to 3.02]; 1.06 [0.87 to 1.29] per SD reduction in low-density lipoprotein (LDL)-cholesterol). Of these lipid modifiers, only genetic mimics of statins were associated with higher BMI (0.33 SD [0.29 to 0.38] per SD reduction in LDL-cholesterol), which explained 54% of the total effect of statins on type 2 diabetes risk. CONCLUSION: Higher BMI mediated more than half of the diabetogenic effects of statins, which did not extend to other commonly used lipid modifiers. Further investigations are needed to clarify drug-specific mechanisms underlying the effects of lipid modifiers on type 2 diabetes.