Mendelian randomization study of diabetes and dementia in the Million Veteran Program

Abstract Introduction Diabetes and dementia are diseases of high health‐care burden worldwide. Individuals with diabetes have 1.4 to 2.2 times higher risk of dementia. Our objective was to evaluate evidence of causality between these two common diseases. Methods We conducted a one‐sample Mendelian randomization (MR) analysis in the US Department of Veterans Affairs Million Veteran program. The study included 334,672 participants ≥65 years of age with type 2 diabetes and dementia case‐control status and genotype data. Results For each standard deviation increase in genetically predicted diabetes, we found increased odds of three dementia diagnoses in non‐Hispanic White participants (all‐cause: odds ratio [OR] = 1.07 [1.05–1.08], P = 3.40E‐18; vascular: OR = 1.11 [1.07–1.15], P = 3.63E‐09, Alzheimer's disease [AD]: OR = 1.06 [1.02–1.09], P = 6.84E‐04) and non‐Hispanic Black participants (all‐cause: OR = 1.06 [1.02–1.10], P = 3.66E‐03, vascular: OR = 1.11 [1.04–1.19], P = 2.20E‐03, AD: OR = 1.12 [1.02–1.23], P = 1.60E‐02) but not in Hispanic participants (all P > 0.05). Discussion We found evidence of causality between diabetes and dementia using a one‐sample MR study, with access to individual level data, overcoming limitations of prior studies using two‐sample MR techniques.