Predictive accuracy of body mass index in estimating body fatness measured by bioelectrical impedance

The aim of this study was to determinate the performance of body mass index (BMI) for the diagnosis of obesity and as a predictor of body fatness in adult Chilean subjects. We conducted a study in 433 women (18-73 years old; BMI: 19.7 to 69.7 kg/m2) and 264 men (18-83 y.; BMI: 19.1 to 54.8 kg/m2). Bioelectrical resistance was measured by impedance method and fat mass percent (FM%) was calculated by fatness-specific equations developed by Segal et al. Obesity was defined as a BMI of at least 30 kg/m2. Increased fatness was defined by the FM% cut-off points of at least 25% for men and at least 30% for women. Sixty-four percent of women and 23.6% of men with BMI below 30 kg/m2 had FM% higher than 30% and 25%, respectively. A value of BMI of 26 kg/m2 in women and 30 kg/m2 in men had the best agreement to the cut-off points of fatness according to sensitivity vs. specificity analysis The following equations were developed to predict individual fatness: women FM% = 0.96 x BMI + 0.154 x age + 1.44 (r2 = 0.75; standard error 3.8%); men FM% = 0.99 x BMI + 0.141 x age - 9.914 (r2 = 0.66; standard error 4.4%). Differences between measured and predicted FM% presented a wide variation, with a range of +/- 2 sd of 7.5% in women and 8.8% in men. The commonly used value of BMI 330 kg/m2 as a cut-off point for obesity does not have adequate sensitivity and specificity for the screening of increased fatness subjects, specially in women. In this study BMI shows a low reliability as a predictor of individual body fatness, particularly in men and in subjects with a BMI below 30 kg/m2.