Obesity Metrics And Their Associations With Type 2 Diabetes And Glucose Levels In A Large International Multi-Ethnic Cohort

Background: Obesity is a major risk factor for diabetes, and its prevalence is increasing worldwide. We examined the associations of several obesity metrics, including body mass index (BMI), waist circumference (WC), hip circumference (HC) and waist-to-hip ratio (WHR) with type 2 diabetes (T2D) and glucose levels in a large multi-ethnic cohort. Methods: Participants and those screened for entry into the DREAM clinical trial from 21 different countries make up the EpiDREAM cohort (n= 24 595). Participants had obesity metrics taken and an oral glucose tolerance test at baseline. Multiple logistic and linear regression were used to determine the associations of obesity metrics with odds of T2D and area-under-the-glucose-curve (AUC; mmol*h/L). Results: An increase in one standard deviation of BMI (OR T2D = 1.32, 95% C.I. 1.28–1.37, β AUC = 0.68, 95% C.I. 0.63– 0.74), WC (OR T2D = 1.39, 95% C.I. 1.34–1.45, β AUC = 0.76, 95% C.I. 0.70 – 0.81), HC (OR T2D = 1.20, 95% C.I. 1.15–1.24, β AUC = 0.44, 95% C.I. 0.39 – 0.50) or WHR (OR T2D = 1.42, 95% C.I. 1.36–1.48, β AUC = 0.79, 95% C.I. 0.72– 0.85) was associated with an increased odds of T2D or AUC after adjusting for age, sex, smoking and ethnicity. After adjusting for HC, the standardized associations of WC with outcomes (OR T2D = 1.76, 95% C.I. 1.65–1.88, β AUC = 1.21, 95% C.I. 1.11–1.31) were strongest. HC was inversely associated with outcomes after adjusting for WC (OR T2D = 0.75, 95% C.I. 0.71– 0.81, β AUC = −0.55, 95% C.I. −0.64- −0.45). Women had a significantly higher odds of T2D or AUC than men for an increase in WC, WC adjusted for HC, and HC adjusted for WC. Conclusions: WC adjusted for HC is the obesity metric most strongly associated with odds of T2D and AUC. A significant sex interaction was observed. The inverse association of HC with odds of T2D and AUC in this study suggests that anthropometric measures of obesity should include both WC and HC. The role of these metrics in predicting incident dysglycemia is being assessed in the prospective phase of the study.