Abstract
OBJECTIVES: Body mass index (BMI) is commonly used in epidemiological study or clinical center. However, it is not exactly correlated with body fat composition and does not reflect sex, age, or race. The aim of this article is to evaluate the validity of BMI standards relative to total body fat (TBF) and to estimate new BMI criteria that correspond to TBF for obesity, especially for Asian postmenopausal women. METHODS: A total 3,936 patients were included in this cross-sectional study, including 1,565 premenopausal and 2,371 postmenopausal women. At the time of visit, demographic data were collected. We demonstrated the validity of BMI cut-point of 25 kg/m(2) by using area under the curve (AUC), and presented the empirical optimal BMI cut-point by using Youden's index and overall accuracy in both premenopausal and postmenopausal women. RESULTS: BMI-defined obesity (≥ 25 kg/m(2)) represents high AUC values (> 0.9) for each TBF. In premenopausal women, TBF ≥ 38% and corresponding BMI value was 29.45 kg/m(2) indicated the highest both Youden's index and overall accuracy. In comparison, postmenopausal women who were TBF ≥ 38% showed the highest Youden's index and overall accuracy, and corresponding BMI value was 26.45 kg/m(2). CONCLUSIONS: We proposed new BMI criteria for obesity by using TBF reference. With application of bioelectrical impedance analysis, the diagnosis of obesity using BMI criteria may differ between premenopausal and postmenopausal women.