Utilization of BIA-Derived Bone Mineral Estimates Exerts Minimal Impact on Body Fat Estimates via Multicompartment Models in Physically Active Adults

Abstract

The purpose of this study was to compare body fat estimates and fat-free mass (FFM) characteristics produced by multicompartment models when utilizing either dual energy X-ray absorptiometry (DXA) or single-frequency bioelectrical impedance analysis (SF-BIA) for bone mineral content (BMC) in a sample of physically active adults. Body fat percentage (BF%) was estimated with 5-compartment (5C), 4-compartment (4C), 3-compartment (3C), and 2-compartment (2C) models, and DXA. The 5C-Wang with DXA for BMC (i.e., 5C-Wang_DXA) was the criterion. 5C-Wang using SF-BIA for BMC (i.e., 5C-Wang_BIA), 4C-Wang_DXA (DXA for BMC), 4C-Wang_BIA (BIA for BMC), and 3C-Siri all produced values similar to 5C-Wang_DXA (r > 0.99; total error [TE] < 0.83%; standard error of estimate < 0.67%; 95% limits of agreement [LOAs] < ±1.35%). The 2C models (2C-Pace, 2C-Siri, and 2C-Brozek) and DXA each produced similar standard error of estimate and 95% LOAs (2.13%-3.12% and ±4.15%-6.14%, respectively). Furthermore, 3C-Lohman_DXA (underwater weighing for body volume and DXA for BMC) and 3C-Lohman_BIA (underwater weighing for body volume and SF-BIA for BMC) produced the largest 95% LOAs (±5.94%-8.63%). The FFM characteristics (i.e., FFM density, water/FFM, mineral/FFM, and protein/FFM) for 5C-Wang_DXA and 5C-Wang_BIA were each compared with the "reference body" cadavers of Brozek et al. 5C-Wang_BIA FFM density differed significantly from the "reference body" in women (1.103 ± 0.007 g/cm³; p < 0.001), but no differences were observed for 5C-Wang_DXA or either 5C model in men. Moreover, water/FFM and mineral/FFM were significantly lower in men and women when comparing 5C-Wang_DXA and 5C-Wang_BIA with the "reference body," whereas protein/FFM was significantly higher (all p ≤ 0.001). 3C-Lohman_BIA and 3C-Lohman_DXA produced error similar to 2C models and DXA and are therefore not recommended multicompartment models. Although more advanced multicompartment models (e.g., 4C-Wang and 5C-Wang) can utilize BIA-derived BMC with minimal impact on body fat estimates, the increased accuracy of these models over 3C-Siri is minimal.

Keywords: Bioelectrical impedance analysis; body fat percentage; fat mass; fat-free mass; multicompartment models.