Concordance between Body Composition Indices Measured with Dual-Energy X-Ray Absorptiometry and Bioelectrical Impedance Analysis in Obese Children in Sri Lanka

Abstract

Dual-energy X-ray absorptiometry (DXA) is the reference standard in the measurement of body composition indices. But, its utility is limited due to the high cost, expertise required, lack of portability, and restricted availability. Therefore, bioelectrical impedance analysis (BIA) has gained recognition in resource-limited settings for the measurement of body composition indices in the screening of children for childhood obesity. To determine whether BIA represents a viable alternative to DXA in the assessment of body composition in obese children in the community setting in Sri Lanka, the concordance between BIA and DXA was determined. Fat mass (FM), percentage fat mass (%FM), and fat-free mass (FFM) were measured in 97 obese children using DXA and BIA, and the concordance between the methods was analyzed using independent sample $t$ -test, regression analysis, and Bland-Altman plots. Significant mean differences were observed between DXA and BIA in measuring FM and FFM. However, high correlations were seen in DXA- and BMI-derived FM and FFM measurements (FM $r=0.92$ and FFM 0.83, $P<0.001$ for both). Compared to DXA, BIA overestimated FM and %FM and underestimated FFM. When compared with DXA-derived measurements, the accuracy errors (SEE) of BIA for FM, FFM, and %FM were relatively higher in boys (3.56 kg, 4.49 kg, and 5.46%, respectively) than in girls (2.44 kg, 3.72 kg, and 3.5%), respectively. BA plots showed a systematic error in the measurements of FM, FFM, and %FM in both sexes. Despite the limitations inherited, BIA is a viable alternative to DXA for the measurement of body composition in obese children of 5-15 yrs. The accuracy errors observed, however, need to be taken into consideration when interpreting results at the individual level.