Body fat gain and loss differentially influence validity of dual-energy x-ray absorptiometry and multifrequency bioelectrical impedance analysis during simultaneous fat-free mass accretion

The validity of dual-energy x-ray absorptiometry (DXA) and multifrequency bioelectrical impedance analysis (MFBIA) for detecting changes in fat mass (FM), fat-free mass (FFM), and body fat percentage (BF%) was evaluated, as compared to a rapid 4-component (4C) model, in 31 females completing 8 weeks of resistance training. Analyses were performed in all participants (ALL) and in subgroups that gained FFM but lost FM (R subgroup) or gained both FFM and FM (G subgroup). It was hypothesized that methods would comparably detect changes in ALL, but discrepancies would occur in subgroup analysis. Changes in body composition did not significantly differ between 4C, DXA, and MFBIA. Equivalence testing indicated that similar changes were detected by DXA and MFBIA, compared to 4C, for ΔFFM in all analyses and ΔBF% in ALL and R subgroup. ΔFM was equivalent to 4C only in R subgroup for DXA and G subgroup for MFBIA. For ΔFM and ΔBF%, DXA and MFBIA produced similar magnitude errors in ALL. However, DXA exhibited lower error in R subgroup, whereas MFBIA exhibited lower error in G subgroup. For ΔFFM, DXA and MFBIA exhibited relatively similar errors in ALL and R subgroup, although MFBIA displayed proportional bias and weaker correlations with 4C than DXA. In G subgroup, MFBIA exhibited lower errors and a higher correlation with 4C ΔFFM than DXA. Although both DXA and MFBIA may have utility for estimating body composition changes during FFM accretion, DXA may be superior during simultaneous FM loss, whereas MFBIA may produce lower error during simultaneous FM gain.