Purpose: We sought to determine whether the association between family history, a surrogate for genetic predisposition, and diabetes was modified by any known diabetes risk factors and if these relationships were constant across different ethnic groups. Methods: We examined 10,899 adults from the National Health and Nutrition Examination Survey (1999Survey ( -2004 to identify interactions between family history and clinical, demographic, and lifestyle variables for the outcome of diabetes using logistic regression analysis in racial/ethnic subgroups. Results: There was significant heterogeneity by race/ethnicity in the interaction between covariates and family history in relation to diabetes. In black (P ϭ 0.0001) and Hispanic (P ϭ 0.013), but not white (P ϭ 0.75) subgroups, high-familial risk was a strong risk factor for diabetes among lean individuals but less so among overweight or obese subjects. Among blacks, high-familial risk conferred a 20-fold increased odds of diabetes among lean subjects and only a sixfold increased odds among obese individuals. Conclusions: These findings suggest possible race/ ethnic-specific differences in gene by environment interaction and identify body mass index as an important effect modifier of familial risk in diabetes in non-white populations. These findings may help guide future genetic studies and improve the utility of family history as a public health screening tool. Genet Med 2009:11(7):542-547.