Comparison of six anthropometric measures in discriminating diabetes: A cross‐sectional study from the National Health and Nutrition Examination Survey

BACKGROUND: Traditional anthropometric measures, including body mass index (BMI), are insufficient for evaluating the risk of diabetes. This study aimed to evaluate the performance of new anthropometric measures and a combination of anthropometric measures for identifying diabetes. METHODS: A total of 46 979 participants in the National Health and Nutrition Examination Survey program were included in this study. Anthropometric measures, including weight, BMI, waist circumference (WC), waist‐to‐height ratio (WtHR), conicity index (CI), and A Body Shape Index (ABSI), were calculated. Logistic regression analysis and restricted cubic splines were used to evaluate the association between the anthropometric indices and diabetes. The receiver operating characteristic (ROC) curve analysis was performed to compare the discrimination of different anthropometric measures. RESULTS: All anthropometric measures were positively and independently associated with the risk of diabetes. After adjusting for covariates, the per SD increment in WC, WtHR, and CI increased the risk of diabetes by 81%, 83%, and 81%, respectively. In the ROC analysis, CI showed superior discriminative ability for diabetes (area under the curve 0.714), and its optimum cutoff value was 1.31. Results of the combined use of BMI and other anthropometric measures showed that among participants with BMI <30 kg/m(2), an elevated level of another metric increased the risk of having diabetes (P < .001). Similarly, at low levels of weight, CI, and ABSI, an elevated BMI increased diabetes risk (P < .001). CONCLUSIONS: WtHR and CI had the best ability to identify diabetes when applied to the US noninstitutionalized population. Anthropometric measures containing WC information could improve the discrimination ability.