Objective: To investigate whether central obesity variables are more important than general obesity variables in predicting the incidence of type 2 diabetes in Iranians.
Methods: In this population-based longitudinal study, a representative sample of 1852 males aged > or = 20 years, participants of the Tehran Lipid and Glucose Study, were followed for a mean duration of 3.6 years. Demographic data were collected; blood pressure and anthropometric variables such as body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR) were measured according to a standard protocol. Biochemical analyses including measurements of plasma glucose in the fasting state and 2 hours after ingestion of 75 g glucose as well as fasting serum lipids were done. Diabetes and abnormal glucose tolerance (impaired fasting glucose or impaired glucose tolerance) were defined according to ADA criteria. Logistic regression models with the stepwise conditional method were used to estimate the relative risk (RR) of developing diabetes associated with each quartile of each anthropometric variable in a multivariate model adjusted for age, hypertension, smoking, family history of diabetes (model 1), and a full model adjusted for all the variables in model 1 plus abnormal glucose tolerance (model 2). Receiver operating characteristic (ROC) curves were used to determine the predictive power of each variable for development of type 2 diabetes, after adjustment for age.
Results: A total of 69 new cases of type 2 diabetes (3.7%) were diagnosed during 3.6 years of follow-up, with an incidence rate of approximately one percent per year. The RR of developing diabetes associated with the highest quartile of anthropometric variables, was highest for WHtR in model 1. After further adjustment for abnormal glucose tolerance (model 2) only WHtR and WC were significantly associated with type 2 diabetes. According to the ROC curve analysis, among central obesity variables only WHtR had a higher ROC curve than BMI. WHtR also continued to be the best predictive central obesity variable compared to BMI, even when the analysis was restricted to subjects with BMI < 27 kg/m2 (0.678 vs. 0.631, p < 0.05). In those with BMI > or = 27 kg/m2, none of the central obesity variables proved to be superior to BMI.
Conclusion: Among abdominal obesity variables, WHtR appears to be stronger than BMI in identifying men with type 2 diabetes in the future.