Body weight difference between dual-energy X-ray absorptiometry and multi-frequency bioelectrical impedance analysis attenuates the equivalence of body-composition assessment.

Abstract

Low agreement of body-composition analysis (BCA) using dual-energy X-ray absorptiometry (DXA) and multi-frequency bioelectrical impedance analysis (MF-BIA) has been reported. We examined whether this discrepancy is influenced by the precision of body weight (BW) measurement using DXA. This cross-sectional study enrolled 1353 participants aged 53-83 years. A whole-body DXA scan and an eight-polar tactile-electrode impedance-meter using four electronic frequencies of 5, 50, 250, and 500 kHz were employed for BCA. The level of agreement between BW estimated using DXA and actual BW (WgtA) was calculated. The agreement of BCA parameters using DXA and MF-BIA across WgtA groups was also assessed. DXA incorrectly estimated BW, especially in men. In total, 13.5%, 5.1%, and 5.6% of the participants had BW bias levels of 2%, 3%, and ≥4%, respectively. Correlations of BCA parameters measured using DXA and MF-BIA, including body fat mass, percent body fat, and lean body mass (LBM), were gradually reduced, whereas the root mean square error was increased in accordance with the reduction in WgtA. DXA provided a lower LBM amount compared to MF-BIA and this difference increased significantly across groups with poor WgtA. Lower WgtA greatly contributed to the difference in BCA measured using DXA and MF-BIA.