Validity Of Single-frequency Bioelectrical Impedance Method To Assess Changes In Body Composition During Exercise Training

Abstract

PURPOSE: To investigate the validity of single-frequency bioelectrical impedance method (BI; Muscle-alpha, Physion) to assess body composition changes during a 12-week aerobic exercise program in comparison with dual-energy X-ray absorptiometry (DXA; QDR 4500, Hologic). METHODS: Eighteen obese men (age: 49.1 ± 11.3 years, body mass index: 28.7 ± 1.8 kg/m2) participated in this study. They enrolled in a 90-min supervised exercise program mainly consisting of aerobic exercise (60-80% of heart rate reserve) on a regular basis of 3 d/wk for 12 weeks. Percentage of fat mass (%fat), fat mass (FM) and fat-free mass (FFM) were measured before and after the program using BI and DXA. RESULTS: With -2.7 ± 2.4 kg of weight reduction during the program, %fat and FM significantly decreased (P < 0.05), but FFM remained unchanged when assessed by BI and DXA. Significant differences (P < 0.05) were observed between BI and DXA at baseline for %fat (28.2% vs. 25.4%), FM (23.5 kg vs. 21.3 kg) and FFM (59.7 kg vs. 62.6 kg). Further, significant differences (P < 0.05) were observed after the program for %fat (25.2% vs. 23.3%), FM (20.4 kg vs. 19.0 kg) and FFM (60.2 kg vs. 62.3 kg) and change in FM (-3.1 kg vs. -2.3 kg). At each time point, %fat, FM and FFM measured by BI were significantly lower (P < 0.05) than those by DXA. The correlation coefficients between BI and DXA were similar at baseline and after the program, while rather low for the change in %fat (r = 0.63, 0.69, and 0.33, respectively), FM (r = 0.81, 0.86, and 0.59, respectively) and FFM (r = 0.98, 0.98, and 0.15, respectively). A Bland-Altman analysis revealed no significant bias (change%fat: r = -0.21, standard error of estimate: SEE = 1.36, P = 0.40; changeFM: r = -0.09, SEE = 1.54, P = 0.72; changeFFM: r = -0.35, SEE = 0.96, P = 0.16). CONCLUSIONS: BI and DXA are able to track changes in %fat and FM similarly during an aerobic exercise program in obese men. Additionally, there are no significant differences in %fat and FM between the two methods, suggesting that BI can be used in a clinical setting.