Impact of body mass index on the predictive ability of body fat distribution for Type 2 diabetes risk in Koreans

AIMS: The optimal anthropometric measure of obesity or body fat distribution that best predicts the risk of Type 2 diabetes in Asians is unclear. Moreover, it has not been determined whether BMI modifies the effect of body fat distribution on diabetes risk in Asians. METHODS: We analysed the anthropometric and laboratory data of 7658 non-diabetic Korean adults (5061 men and 2597 women, aged 20-79 years) who underwent routine medical check-ups at 5-year intervals. BMI, waist circumference, waist-to-height ratio, and bioelectrical impedance (to calculate fat mass and per cent body fat) were measured at baseline. RESULTS: Of the 7658 participants, 278 subjects (3.6%) developed diabetes over 5 years. Each of the anthropometric measures of general obesity (BMI, fat mass, per cent body fat) and central body fat distribution (waist circumference and waist-to-height ratio) was a good predictor of Type 2 diabetes. However, when the areas under the receiver-operating characteristic curves were compared, BMI (0.697; 95% CI, 0.669-0.725), waist circumference (0.709, 0.682-0.736) and waist-to-height ratio (0.718, 0.692-0.743) were better predictors of diabetes risk than fat mass (0.672, 0.643-0.700) or per cent body fat (0.657, 0.628-0.686). In the low- (< 23 kg/m(2)) and mid- (23-27 kg/m(2)) BMI groups, the addition of waist-to-height ratio or waist circumference to BMI could improve the prediction of diabetes risk. CONCLUSIONS: BMI, waist circumference and waist-to-height ratio were good predictors of Type 2 diabetes risk in Koreans. In non-obese or less obese subjects, measures of central body fat distribution can help improve the prediction of Type 2 diabetes risk when added to measures of general obesity.