The relationship of family history and risk of type 2 diabetes differs by ancestry

Aim. - Type 2 diabetes (T2DM) in a first-degree relative is a risk factor for incident diabetes. Americans of African ancestry (AA) have higher rates of T2DM than Americans of European ancestry (EA). Thus, we aimed to determine whether the presence, number and kinship of affected relatives are associated with race-specific T2DM incidence in a prospective study of participants from the Genetic Study of Atherosclerosis Risk (GeneSTAR), who underwent baseline screening including a detailed family history. Methods. - Nondiabetic healthy siblings (n = 1405) of patients with early-onset coronary artery disease (18-59 years) were enrolled (861 EA and 544 AA) and followed for incident T2DM (mean 14 +/- 6 years). Results. - Baseline age was 46.2 +/- 7.3 years and 56% were female. T2DM occurred in 12.3% of EA and 19.1% of AA. Among EA, 32.6% had >= 1 affected first-degree relatives versus 53.1% in AA, P < 0.0001. In fully adjusted Cox proportional hazard analyses, any family history was related to incident T2DM in EA (HR = 2.53, 95% CI: 1.58-4.06) but not in AA (HR = 1.01, 0.67-1.53). The number of affected relatives conferred incremental risk of T2DM in EA with HR = 1.82 (1.08-3.06), 4.83 (2.15-10.85) and 8.46 (3.09-23.91) for 1, 2, and >= 3 affected, respectively. In AA only >= 3 affected increased risk (HR = 2.45, 1.44-4.19). Specific kinship patterns were associated with incident T2DM in EA but not in AA. Conclusions. - The presence of any first-degree relative with T2DM does not discriminate risk in AA given the high race-specific prevalence of diabetes. Accounting for the number of affected relatives may more appropriately estimate risk for incident diabetes in both races. (C) 2018 Elsevier Masson SAS. All rights reserved.