Measurement of body composition changes with weight loss in postmenopausal women: Comparison of methods

Background: The accurate measurement of body composition changes is important when evaluating the efficacy of medical nutrition therapy and weight management programs, yet is not well documented in older women. Objective: We compared methods of estimating energy-restriction-induced body composition changes in postmenopausal women. Design: 27 women (59 +/- 8 y; BMI 29.0 +/- 2.9 kg/m(2); mean SD) completed a 9-wk energy restriction period (5233 kJ/d, (1250 kcal/d)). Changes in % body fat (Delta%BF), fat mass (Delta FM), and fat-free mass (Delta FFM) were measured by hydrostatic weighing (HW), air-displacement plethysmography (ADP), dual-energy x-ray absorptiometry (DXA), and deuterium oxide dilution (D2O). The Baumgartner et al. (Am J Clin Nutr 53:1345-1353, 1991) four-compartment (4C) model with body volume from HW was the criterion method. The 4C model with body volume from ADP was also compared. Regression equations were developed based on 4C(HW) (dependent variable) utilizing results of change (POST-PRE) for each method. Results: The women lost 6.8 +/- 3.2 kg; 9% of baseline weight. Based on 4C(HW), the body composition changes were -2.4 +/- 4.5 Delta%BF, -4.7 +/- 3.3 kg Delta FM, and -2.6 +/- 4.4 kg Delta FFM. No differences were detected by ANOVA for Delta%BF, Delta FM, and Delta FFM among 4C(HW), HW, ADP, DXA, D2O, and 4C(ADP). Bland-Altman limits of agreement showed differences between methods that ranged from 14.5 to -14.1 Delta%BF, 7.8 to -8.1 kg Delta FM, and 7.5 to -8.4 kg Delta FFM for individuals. A bias was shown with 4C(ADP) overestimating Delta%BF (1.4 %) and FM (0.6 kg) and underestimating Delta FFM (-1.2 kg) compared to 4C(HW). The regression model was acceptable for %BF (4C(ADP), 2C(HW), and 2C(D2O)); FM and FFM (4C(ADP), 3C(DXA), 2C(HW), and 2C(D2O)), but not for other estimates of %BF, FM, FFM. Conclusions: These body composition assessment methods may be used interchangeably to quantify changes in % body fat, fat mass, and fat-free mass with weight loss in groups of postmenopausal women. 4CADP overestimates Delta%BF and underestimates Delta FFM. When utilizing one of these comparison methods (4CADP, 3C(DXA), 2C(HW), 2C(D2O)) to quantify changes in fat mass and fat-free mass for an individual postmenopausal woman, regression equations may be used to relate the data to 4CHW.