Main content area

Body composition using bioelectrical impedance: Development and validation of a predictive equation for fat-free mass in a middle-income country

Gonzalez, Maria Cristina, Orlandi, Silvana Paiva, Santos, Leonardo Pozza, Barros, Aluísio J.D.
Clinical nutrition 2019 v.38 no.5 pp. 2175-2179
adults, bioelectrical impedance, dual-energy X-ray absorptiometry, equations, lean body mass, men, models, regression analysis, women, Brazil, United States
We developed and described a new equation to estimate fat-free mass derived from BIA using a sample population from Brazil and having Dual-energy X-ray absorptiometry (DXA) as the reference method. We also compared this new equation with two published and widely used equations developed in high-income countries.Cross-sectional study with 294 healthy adults from Pelotas, Brazil. DXA was used to assess total fat mass and fat-free mass aiming to obtain reference measures for the development of the new BIA equation. Multivariable linear regression models including fractional polynomials were used to find the best predictive model for FFM, using resistance, reactance, age, weight and height as the independent variables. Models were developed separately for men and women. The bootstrapping method was used to test the validity of the new equation. Finally, the Bland-Altman approach was used to assess the agreement of our equations and the two others widely used equations with the FFM measured by DXA.The new equations explained more than 80% of the variation in fat-free mass percentage from DXA. In the bootstrapping analysis, the new equations presented good validity, as the corrected RMSE was similar to those found in regression analysis. Finally, the new equations presented a better concordance when compared to two validated equations from US and Switzerland.The new developed equations appear to be the best options to predict fat-free mass percentage in Brazilian adults by bioelectrical impedance and appear to fit well in all Brazilian population due to the good validity presented.