In order to analyse the relationship of genetic factors and the role of obesity in the pathogenesis of Type 2 (noninsulin-dependent) diabetes mellitus, we compared body mass index (BMI) and other parameters in 71 monozygotic and 17 dizygotic twins (MZT and DZT) with Type 2 diabetes, and analysed the relationship of maximal BMI in the past (max BMI) with the concordance and discordance for diabetes. Data were collected by mail from the members of the Japan Diabetes Society. Concordance rate for diabetes was higher in MZT than in DZT (72% vs. 41%, P = 0.017). Concordance rate in MZT varied in relation to the max BMI of the index twins who developed diabetes earlier. It was higher in non-obese and slightly obese groups (82% and 80% for those with max BMI < 24 and 24-27.9, respectively) than in obese group (36%, max BMI greater than or equal to 28) (P = 0.003). In MZT, diabetic twins of discordant pairs were significantly more obese at the survey and in the past than index twins of concordant pairs (mean BMI: 23.9 vs. 21.7, P = 0.007; mean max BMI, 27.7 vs. 24.3, P = 0.001). A similar tendency of differences in BMI was also observed in DZT. Intrapair comparison in discordant MZT and DZT pairs showed that diabetic twins were more obese than nondiabetic cotwins (mean max BMI, 27.7 vs. 24.7, P = 0.004 for MZT; and 27.3 vs. 23.9, P = 0.002 for DZT). In concor dant DZT pairs, index twins with earlier onset of diabetes had been less obese than cotwins (mean max BMI, 23.6 vs. 26.0, P = 0.004). In conclusion, the comparison of the max BMI in twins suggests that the degree of diabetogenic role of obesity varies in heterogeneous genetic background. The index twins of discordant pairs had been more obese than those of concordant pairs, suggesting that obesity plays a more important role in those with weaker genetic susceptibility for diabetes.
