Development and cross-validation of predictive equations for fat-free mass estimation by bioelectrical impedance analysis in Brazilian subjects with overweight and obesity

Abstract

Introduction: Obesity is a public health problem worldwide, and body composition assessment is a very important diagnostic tool. Bioelectrical Impedance Analysis (BIA) is a fast, non-invasive, relatively low-cost, and user-friendly technique; however, to obtain greater validity of the estimates, the predictive equations used must be population specific. Thus, the objectives of this study were: (1) to test the validity of four BIA equations used for fat-free mass (FFM) estimation and one model for fat mass (FM) estimation in adults with overweight or obesity; (2) develop and cross-validate new equations to estimate FFM to adults with overweight or obesity, and specific for those with obesity. Methods: The non-probabilistic sample included 269 individuals, 53.2% with overweight and 46.8% with obesity, aged 18–79 years, randomly divided into two groups: development (n = 178) and cross-validation (n = 91), stratified by sex and classification as overweight or obese. The criterion technique was dual-energy-x-ray absorptiometry (DXA), whereas a tetrapolar single-frequency BIA equipment was used as the alternative method. Paired t-test, multiple regression, concordance correlation coefficient, and Bland–Altman analysis were used. Results: Most existing equations were not valid and new equations were derived: (1) for individuals with overweight or obesity: CCC = 0.982; r2 = 0.95; standard error of estimate (SEE) = 2.50 kg; limits of agreement (LOA) = -5.0 to 4.8; and (2) specific for individuals with obesity: CCC = 0.968; r2 = 0.94; SEE = 2.53 kg; LOA = -5.3 to 5.2. No FFM differences were observed between the new models and the reference method (p > 0.05). Conclusion: The new proposed models provide valid options to estimate FFM in an adult population with overweight/obesity.