The main advantage of the bioelectric impedance method (BIA) in the determination of body fat (%BF) is the simplicity of the procedure. However, its accuracy and reliability have been criticized. The purpose of this study was to compare the %BF obtained by BIA (RJL-101; Biodynamics A-310, Maltron BF-900 e BF-906), by skinfold thickness (ST), and by underwater weighing (UW). Twenty-five subjects, divided in homogenous groups according to age (18 to 36 years), sex (men), and race (white) participated in the study. BIA measures were taken using the Lukaski et al. standardization (1985,1986). ST was taken by using the equation of 3 and 7 skinfolds (Jackson, Pollock, 1978). The values of %BF and residual volume for the UW were estimated by the Siri (1961) and Goldman and Becklake (1959) equations. Statistical analysis was calculated by: a) repeated measures ANOVA followed by the Tukey post-hoc test; b) Pearson (r) correlation; c) standard error of estimate (SEE) of the BIA and ST compared to UW. The results indicated that: a) there were no significant differences for %BF measures obtained by the BIA devices; b) the %BF obtained by the A-310 and BF-906 devices did not match with the UW measures (p < 0.01); c) SEE was high for all devices, except for the RJL-101; d) the correlation coefficients were low to moderate for all devices, the BF-906 showing the higher values; e) The ST showed greater values of r and fewer SEE than the BIA. In conclusion, the findings do not allow to state that a BIA device is better than the others to assess %BF, but the ST seems to be more powerful and reliable than the BIA for %BF estimation. Notwithstanding, the results should be confirmed by more sophisticated experimental designs, with a closer control of sampling bias for type I and II errors.
Body composition; Bioelectric impedance; Underwater weighing; Skinfold thickness; Testing
