Diagnostic performance of body mass index to identify adiposity in women.

Abstract

The aim of this study was to determine the diagnostic performance of body mass index (BMI) and to detect the optimal BMI cutoff points to define adiposity in women of various ages. A total of 2409 women participated. Fat mass was measured using a multifrequency bioelectrical impedance analysis. The diagnostic performance of BMI to identify adiposity was evaluated using a fat mass percentage cutoff point of ⩾35%. Although 40% of women were overfat, the BMI-based obesity prevalence was 21%. In the total sample, BMI had low overall performance, which resulted in a sensitivity of 51.9% (95% confidence interval (CI): 48.7-55.2%) and a specificity of 99.2% (95% CI: 98.7-99.6%). BMI failed to identify overfat women with intermediate BMI ranges. An analysis of the receiver operating characteristic curves of all of the subjects demonstrated that optimal cutoff point corresponded to a BMI value of 26.4 kg/m(2). The diagnostic performance of BMI did not differ as age increased. BMI has a high specificity but a low sensitivity to detect adiposity, and it fails to identify nearly half of women with excess fat mass. We provide evidence that a commonly used BMI cutoff value to diagnose obesity is too high among women.