Segmental fat-free mass and lean soft mass: a comparative study with dual X-ray absorptiometry (DXA), bioelectrical impedance analysis (BIA) and anthropometry and development of anthropometric prediction models.

Abstract

Background: Although dual X-ray absorptiometry (DXA), bioelectrical impedance ana-lysis (BIA), and anthropometry (ANT) are commonly used to evaluate body composition,evidence on their agreement at the segmental level remains limited. This study aimed tocompare intra-subject differences in DXA, BIA, and ANT for estimating segmental weight(SW), fat-free mass (FFM), and lean soft mass (LSM) and to examines sex-related influ-ences, and to develop anthropometric prediction equations using DXA as the reference.Methods: A cross-sectional study was conducted on 258 young adults (157 males, 101females). Participants were assessed using DXA, BIA, and ANT. SW, FFM, and LSM wereestimated in kilograms and percentages for the upper limbs, trunk, and lower limbsusing DXA and BIA. FFM was also estimated using anthropometry (ANT).Results: Significant intra-subject differences were observed between methods acrossmost body segments (p ≤ 0.049), except for SW in the right upper limb (p = 0.328) andLSM in the trunk (p = 0.186) for females. Sex covariable showed a significant influence onthe differences found in these comparisons (p ≤ 0.032). Females exhibited lower valuesof SW, FFM, and LSM in the upper limbs, while males showed higher values of FFM andLSM specifically in the trunk and lower limbs. Bland – Altman analysis revealed a generallack of agreement between DXA and both BIA and ANT, except for BIA when estimatingSW in the right upper limb in females (p = 0.167). The new anthropometric equationsdemonstrated high predictive accuracy (R2 > 0.750), with slightly lower values in theupper limbs of females for FFM and LSM (R2 = 0.688–0.723). Key predictors includedbody mass, corrected girths, and segmental lengths.Conclusions: DXA, BIA, and ANT showed significant differences in estimating SW, FFM,and LSM, highlighting their lack of interchangeability. Sex was a significant factor,indicating its importance in body composition estimation. The new anthropometricequations showed good comparability with DXA, although their predictive performancewas slightly lower in the upper limbs of females for FFM and LSM