Comparison between Anthropometry and Bioelectrical Impedance in Estimating Fat and Muscle Masses in Soccer Players

Abstract

Anthropometry-based models are broadly used to indirectly estimate body fat and muscle masses. Bioelectrical impedance analysis is an advanced method for indirect measurement of the main components of body composition. PURPOSE: This study intended to evaluate agreement in assessing body fat and muscle masses in soccer players between an anthropometry-based model and the bioelectrical impedance analysis. METHODS: Data from 46 male competitive soccer players were analyzed (Age = 25.6 ± 3.5 yr, Body mass = 75.8 ± 6.7 kg, BMI = 24.9 ± 1.6 kg·m−2; mean ± SD). The anthropometry-based estimations of fat and muscle masses were obtained by means of the four-compartment model of De Rose and Guimaraes (1980) (DRG). This model was modified using the regression equation of Withers et al. (1987, cited by Norton, 1996) to estimate body density, and Siri formula (1961) to compute the percentage of body fat. Fat and muscle masses were also assessed by means of bioelectrical impedance analysis (BIA). The paired t-test was used to evaluate the differences between the two measurement techniques. The Bland-Altman approach was applied to estimate 95% limits of agreement. The statistical significance level was set at p<0.05. RESULTS: Significant differences were observed in fat and muscle masses between DRG and BIA (respectively: 8.71 kg vs. 10.13 kg, p<0.001; and 36.80 kg vs. 37.85 kg, p<0.001). The 95% limits of agreement were from −5.13 kg to 2.29 kg for fat mass, and from −4.67 kg to 2.58 kg for muscle mass. On a percentage scale, the results were as follows: 11.36 % vs. 13.16 % (p<0.001) for fat mass, and 48.63 % vs. 50.04 % (p<0.001) for muscle mass; the 95% limits of agreement were, respectively, from −6.80 % to 3.20 %, and from −5.97 % to 3.16 %. The difference between methods revealed negligible to low sample correlations with the range of measurement for the two variables, both in absolute and percentage scales. CONCLUSIONS: The anthropometry-based model on average underestimated the fat and muscle masses compared to BIA, in both cases in a quantity less than 1.5 kg (and less than 2 %). The bias between the methods had a harmful effect on the limits of agreement. Additional testing would be necessary to confirm these results.