Anthropometric indices of obesity and their relationships with diabetes risk by race and ethnicity among postmenopausal women.

BACKGROUND AND AIMS: Traditional measures of obesity such as body mass index (BMI) and waist circumference (WC) have shown inconsistent predictive utility for diabetes across diverse populations. Novel anthropometric measures that focus on abdominal adiposity and body shape may offer better risk assessment for diabetes. Yet, the utility of traditional and novel anthropometric measures in postmenopausal women and in different racial and ethnic groups remains unclear. The predictive utilities of traditional and novel anthropometric measures for diabetes risk were comprehensively assessed, among postmenopausal women, overall, and across racial and ethnic groups.

METHODS AND

RESULTS

Using data from 91,392 diabetes-free Women's Health Initiative participants, predictive values of anthropometric measures were examined using Receiver Operating Characteristic (ROC) and Cox regression with Harrell's c-statistics. ROC analyses suggested that novel anthropometric measures with best predictive abilities were 'a Body Shape Index' (ABSI), 'Abdominal Volume Index', and 'Body Roundness Index', although WC and novel anthropometric measures had similarly modest predictive abilities to BMI (Areas Under the Curve <0.6). Cut-off points for anthropometric measures varied by race/ethnicity. Multivariable Cox regression modeling suggested that 'Clinica Universidad de Navarra-Body Adiposity Estimator', WC, ABSI, and BMI had the strongest associations with diabetes risk (adjusted hazard ratios [HRs]: 1.27, 1.26, 1.23, and 1.22 per 1-SD increase, respectively), although predictive accuracies involving any measure were modest (Harrell's c-statistics∼0.58-0.59).

CONCLUSIONS

In this comprehensive evaluation, anthropometric measures were marginally predictive of diabetes risk, and novel measures did not outperform traditional measures among postmenopausal women, irrespective of race/ethnicity.