Type 2 diabetes in midlife and risk of cerebrovascular disease in late life: a prospective nested case−control study in a nationwide Swedish twin cohort

AIMS/HYPOTHESIS: We aimed to examine the association between midlife type 2 diabetes mellitus and cerebrovascular disease (CBD) in late life, and further to explore whether genetic and early-life familial environmental factors (such as shared childhood socioeconomic status and adolescent environment) play a role in this association. METHODS: In this prospective nested case−control study based on the Swedish Twin Registry, 33,086 twin individuals who were born in 1958 or earlier and were CBD-free before the age of 60 were included. Midlife (40–59 years) type 2 diabetes was ascertained from self-report, the National Patient Registry (NPR) and glucose-lowering medication use. CBD diagnosis (cerebral infarction, occlusion of cerebral arteries, subarachnoid haemorrhage, intracerebral haemorrhage and unspecified CBD) and onset age were identified from the NPR. Late-life CBD was defined as CBD onset age ≥60 years. Generalised estimating equation (GEE) models were used to analyse unmatched case−control data (adjusted for the clustering of twins within a pair). Conditional logistic regression was used in co-twin matched case−control analyses in CBD-discordant twin pairs. RESULTS: Of all the participants, 1248 (3.8%) had midlife type 2 diabetes and 3121 (9.4%) had CBD in late life. In GEE models adjusted for age, sex, education, BMI, smoking, alcohol consumption, marital status, hypertension and heart disease, the ORs (95% CIs) of type 2 diabetes were 1.29 (1.03, 1.61) for cerebral infarction, 2.03 (1.20, 3.44) for occlusion of cerebral arteries, 0.52 (0.12, 2.21) for subarachnoid haemorrhage and 0.78 (0.45, 1.36) for intracerebral haemorrhage. In multi-adjusted conditional logistic regression, the OR of the type 2 diabetes–cerebral infarction association was 0.96 (0.51, 1.80). The differences in ORs from the GEE and co-twin control analyses were not statistically significant (p = 0.780). CONCLUSIONS/INTERPRETATION: Midlife type 2 diabetes is significantly associated with increased risk of cerebral infarction and occlusion of cerebral arteries, but not intracerebral haemorrhage or subarachnoid haemorrhage in late life. Genetic and early-life familial environmental factors do not appear to account for the type 2 diabetes–cerebral infarction association, but further clarification is needed. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s00125-019-4892-3) contains peer-reviewed but unedited supplementary material, which is available to authorised users.