The energy intake modulates the association of the -55CT polymorphism of UCP3 with body weight in type 2 diabetic patients.

ABSTRACT

BACKGROUND: Previous association studies of the -55CT polymorphism of the uncoupling protein 3 (UCP3) gene with body mass index (BMI) have provided inconsistent results. The study aim is twofold (1) to evaluate the association of the -55CT polymorphism of UCP3 with BMI in two independent populations to verify the reproducibility of the finding; (2) to evaluate whether this association is modulated by energy intake. METHODS: Study participants are 736 males and females with type 2 diabetes belonging to independent populations (N=394 population 1; N=342 population 2). Anthropometry and laboratory parameters were measured; in population 2, energy intake and physical exercise were also assessed. RESULTS: The -55CT polymorphism was associated with a significantly lower BMI in population 1 (27.8±3.9 vs 28.9±4.6 kg m(-2); P<0.02), the finding was confirmed in population 2 (that is, 30.3±6.0 vs 32.1±5.9 kg m(-2); P<0.01) independent of gender, age, HbA1c, use of drugs and energy intake. To evaluate the role of diet in population 2, the study participants were stratified by genotype and tertiles of energy intake. In both genotype groups, BMI increased with increasing caloric intake with a significant trend (P<0.001), the BMI difference between the two genotype groups was large and statistically significant in the lower tertile (27.6 vs 31.2 kg m(-2); P<0.001), intermediate in the second tertile and negligible in the upper tertile (32.8 vs 32.9; kg m(-2); nonsignificant). The multivariate regression analysis confirmed a significant interaction between genotype and energy intake as correlates of BMI independent of age, gender, glucose control, physical activity and medications for diabetes (P=0.004). CONCLUSIONS: The study replicates in two independent populations the association between the -55CT polymorphism of UCP3 and a lower BMI. This association was modulated by energy intake, thus suggesting that the unmeasured effect of diet may partly account for inconsistencies of prior association studies.