Aim: The aim of this study is to compare total weight, % body fat (% BF), fat mass (FM) and fat-free mass (FFM) measured by bioelectrical impedance analysis (BIA) and dual-energy X-ray absorptiometry (DXA).
Methods: This cross-sectional study included 159 women (mean age: 49.1 +/- 10.0 years) and 124 men (mean age: 51.4 +/- 8.0 years) subdivided according to sex and body mass index (BMI): BMI < 30 kg/m(2) (66 women and 50 men); BMI 30-35 kg/m(2) (53 women and 44 men) and BMI > or = 35 kg/m(2) (40 women and 30 men). Bioelectrical impedance was performed in the fasting state on a Tanita TBF-215 leg-to-leg analyser (Tanita, Tokyo, Japan). Whole-body DXA scans were performed on a Hologic QDR 4500 A bone densitometer (Hologic, Bedford, MA, USA). Total weight, % BF, FM and FFM were tested for intermethod differences. Linear regression and correlation analysis was performed. Limits of agreement and Bland-Altman plots were built.
Results: DXA-derived body composition parameters were not significantly different from BIA estimates and were highly correlated (e.g. for FFM, r = 0.82-0.95). In lean individuals, BIA tended to produce lower values for FM and % BF and higher ones for FFM in comparison with DXA. This trend was reversed at BMI > 35 kg/m(2). The correlations decreased with increasing BMI. The limits of agreement were much better in men than in women and increased with increasing BMI in both sexes.
Conclusions: Compared with DXA, the leg-to-leg Tanita TBF-215 analyser accurately assessed body composition in a heterogeneous group of both sexes. In the very obese women (BMI > 35 kg/m(2)), BIA measurements should be viewed with caution.