Evaluation Of Bioelectrical Impedance Analysis For Estimating Percent Body Fat In Children

PURPOSE: To compare percent body fat (%BF) measurements estimated by bioelectrical impedance analysis (BIA) with dual-energy x-ray absorptiometry (DXA) in children. METHODS: Forty-nine children (28 girls; 21 boys) volunteered to participate in this study (age = 14.5 ± 2.9 years; body mass index = 22.4 ± 5.6 kg/m2). During the assessment, %BF was measured using hand-to-hand bioelectrical impedance analysis (HBIA; Omron HBF-306), leg-to-leg bioelectrical impedance analysis (LBIA; Tanita 300-A), segmental bioelectrical impedance analysis (SBIA; Tanita BC-418), and DXA, which served as the criterion. %BF data from the four devices was assessed using one-way repeated measures ANOVA with Bonferroni-adjusted post-hoc tests. Alpha level was set a priori to p < 0.05. RESULTS: LBIA (24.2 ± 9.7%, p < 0.001) and SBIA (26.4 ± 8.0%, p = 0.028) significantly underestimated mean %BF when compared to DXA (28.7 ± 11.5%) while no differences were observed between DXA and HBIA (28.4 ± 7.3%, p = 1.0). Linear regression revealed significant (p < 0.001) correlations between %BF determined by DXA and the BIA methods: LBIA (r = 0.87), SBIA (r = 0.91), and HBIA (r = 0.65). For a large percentage of the children (LBIA = 37%, SBIA = 53%, HBIA = 67%), %BF values were outside the ±3.5% minimally acceptable standard for accurate estimation. CONCLUSIONS: The HBIA analyzer produced mean %BF similar to DXA supporting the potential use of this technology when group assessments in this population are of interest. However, due to the large intra-individual variability observed in this study, we do not recommend using the HBIA analyzer when precision of a specific child’s %BF is critical.