RESUMEN
OBJECTIVES: Excessive adiposity is associated with cardiovascular disease (CVD). Anthropometric indices are useful in screening individuals at higher risk for these diseases. However, there are no studies that show which of these indices has the best discriminatory power among Afro-descendant Brazilian women. The objective of this study was to assess the accuracy of anthropometric indices in identifying risk factors for CVD in Afro-descendant Brazilian women and define the one most suitable for use under the operating conditions prevailing in Quilombola communities. METHODS: A household random sample of 1661 women descendants of African slaves were analyzed. The anthropometric predictors analyzed were waist circumference (WC), body mass index, waist-to-height ratio, conicity index (C-index), body shape index, and percentage of body fat (%BF; estimated by bioimpedance). The assessed risk factors for CVD were arterial hypertension, diabetes mellitus, and dyslipidemias (hypertriglyceridemia; hypercholesterolemia; low high-density lipoprotein). To identify the statistical significance between the differences in the areas under the ROC curves (AUC) obtained with the different predictors and outcomes, was used the Bonferroni test adjusted for multiple analyses by the Sidak method. RESULTS: The AUC obtained with WC was higher (p < .05) or similar (p > .05) to those obtained with the other predictors 29 times out of 30 possibilities (six predictors x five outcomes). Only the AUC obtained with C-index in identifying hypercholesterolemia was significantly higher than that with WC. CONCLUSION: Due to its accuracy and greater operational simplicity, WC was the most adequate predictor for identifying Afro-descendant women at greatest risk for CVD.