Comparison of methods for assessing body-composition changes over 1 y in postmenopausal women

Background: Advances in dual-energy X-ray absorptiometry (DXA) software algorithms have improved the accuracy of this method for body-composition measurement. Objective: Our objective was to compare the utility of DXA, underwater weighing (UWW), and a multicomponent model (MC) for assessing changes in body composition. Design: Previously sedentary women aged 40-66 y were randomly assigned to exercise training (ET; n = 36) and no exercise training (NT; n = 40). ET subjects exercised 3 d/wk; NT subjects remained sedentary. Changes in body mass, fat mass, and fat-free mass over 1 y were assessed by the 3 methods. Results: Correlations among methods were significant and large (0.73-0.97). Body weight did not change significantly in either group. In the ET group, fat-free mass increased significantly as assessed by DXA (0.7 ± 1.0 kg) but changes assessed by MC and UWW were not significant. Changes in fat mass and percentage body fat in the ET group were not significant. SDs for changes in fat mass and percentage body fat, respectively, from DXA were 2.5 kg and 2.7%; for MC, 5.5 kg and 7.1%; and for UWW, 4.4 kg and 5.8%. In the NT group, changes in fat-free mass, fat mass, and percentage body fat were significant (P ≤ 0.02) as assessed by MC (fat-free mass, - 1.5 ± 3.7 kg; fat mass, 2.3 ± 4.1 kg; percentage body fat, 2.8 ± 4.7%) and UWW (fat-free mass, -1.1 ± 2.5 kg; fat mass, 2.1 ± 3.6 kg; percentage body fat, 2.5 ± 3.5%), but changes by DXA were not significant (fat-free mass, 0.2 ± 1.2 kg; fat mass, 1.0 ± 3.9 kg; percentage body fat, 0.6 ± 3.2%). Conclusion: DXA was the most sensitive method for assessing small changes in body composition of postmenopausal women.