To examine the joint effect of family history and BMI on diabetes.

Cross-sectional study.

A rural community in Saskatchewan, Canada.

The analysis was based on data from 2081 adults, 18–79 years of age, who participated in the Humboldt Study conducted in 2003. Doctor-diagnosed diabetes and family history of diabetes of biological parents and siblings were self-reported. Body weight and height were objectively measured. The interaction of family history and BMI on diabetes was assessed on an additive scale.

The prevalence of diabetes was 7·9 %, and BMI and history of diabetes were two important predictors. The adjusted prevalence ratios were 1·76 (95 % CI 1·37, 2·27) and 2·59 (95 % CI 2·05, 3·31) for those with a BMI of 25·0–29·9 kg/m2 and of at least 30 kg/m2, respectively, compared with a BMI of less than 25 kg/m2, and was 2·41 (95 % CI 2·08, 2·80) for those with a family history of diabetes v. those without. The data indicated an additive interaction of family history and BMI on diabetes.

When exposed to both family history and overweight/obesity, individuals would have an increased risk that was greater than the sum of their single effects. Reduction of BMI would also reduce the risk of diabetes associated family history.