Comparison whole-body muscle mass assessed by dual-energy X-ray absorptiometry and multi-frequency bioelectrical impedance in overweight and obese adults

Abstract

Objective: To investigate consistency of whole-body muscle mass (WBMM) assessed using multi-frequency bioelectrical impedance analysis (MF-BIA) and dual energy X-ray Absorptiometry (DXA) measurement, and to delvelop the correction regression formulas for the method of MF-BIA. Methods: 1 488 participants who resided in Beijing longer than one year, under 22 to 55 years of age, with self-report BMI ≥24 kg/m(2) were voluntarily recruited based on convenience sampling in Beijing, from April 1(st) 2014 to May 2(nd) 2014. After excluded 171 participants who BMI ≤24 kg/m(2) or with organic diseases when physical examination, Finally 1 317 overweight/obese participants were enrolled into the current study. All the subjects received the measurement of WBMM using both MF-BIA and DXA methods. The paired-samples T test was used to compare the difference of measurement values between MF-BIA and DXA methods. We evaluated the agreement of WBMM measured by MF-BIA and DXA using the interclass correlation coefficients (ICC) and the Bland-Altman plots, and using linear regression method established correction regression formulas. Results: The differences of overweight male, obese male, overweight female, and obese female were (3.29 ± 1.94), (2.05 ± 2.65), (0.17 ± 1.85), and (-2.22 ± 2.89) kg between methods of MF-BIA and DXA, respectively; Except the overweight female, the others groups had the statistically significant difference (P<0.05). The ICC for overweight male, obese male, overweight female, and obese female were 0.958, 0.956, 0.911, and 0.851, respectively (P<0.001). Bland-Altman plots showed that the limit of agreement for four groups were -0.51-7.09, -3.14-7.24, -3.46-3.80, and -7.88-3.44 kg, respectively. Correction regression formulas: overweight male population: y(DXA)=-0.648+ 0.952 x(MF-BIA); obese male population: y(DXA)=-3.646 + 1.026x(MF-BIA); overweight female groups: y(DXA)=-4.800 + 1.117 x(MF-BIA); obese female group: y(DXA)=-9.884+ 1.287x(MF-BIA). Conclusion: The correlation and agreement of WBMM measured by the methods of BIA and DXA were relative weak in Chinese excess weight male and obese female adults. Therefore, measuring WBMM by MF-BIA method in these adults should be corrected to reduce bias compared with the DXA method.