Comparative analysis of bone mineral content measured by bioelectrical impedance analysis and dual energy X ray absorption among children and adolescents / 中国学校卫生

ABSTRACT

Objective@#To compare bioelectrical impedance analysis (BIA) and dual energy X ray absorptiometry (DXA) for measuring body mineral content (BMC) of children and adolescents, and to provide a basis for BIA to accurately measure BMC in children and adolescents.@*Methods@#By using the convenience sampling method, among 1 469 children and adolescents aged 7-17 were recruited in Guangzhou from April to May 2019, the BMC was measured by DXA and BIA. The intraclass correlation coefficient ( ICC ) and Bland Altman analysis were used to evaluate the agreement between BIA and DXA. Bland Altman analysis was performed on log transformed data. The BMC was categorized into age and specific tertiles, and the agreement between methods was evaluated based on the kappa coefficients. Treating the BMC with DXA as the dependent variable, a prediction model was constructed for correcting the BIA measure.@*Results@#The ICC s were 0.93 and 0.94 for boys and girls, respectively. In Bland Altman analysis, the limits of agreements for the BIA to DXA ratio were wide in boys and girls, ranging from 0.27-0.76 and 0.17-0.72, respectively. The kappa coefficients for categorized BMC levels were 0.57 and 0.45 for boys and girls, respectively, showing a fair to good degree of agreement. When sub grouped by BMI, the kappa coefficients for all BMI groups of boys and overweight girls were all >0.75 , with an excellent agreement. The prediction models for boys and girls were as follows BMC DXA =-0.51+0.44× BMC BIA + 0.06× Age +0.02× BMI ; and BMC DXA =-0.55+0.43× BMC BIA +0.06× Age +0.02× BMI , respectively. The R 2 for models of boys and girls were 0.87 and 0.87, respectively.@*Conclusion@#The agreement between BIA and DXA was poor for measuring BMC, but acceptable when evaluating the categorized BMC levels, suggesting the BIA may be applied in assessment of the BMC levels when compared to the age and gender specific population. Additionally, the prediction model for correcting BMC by BIA fis well to the measurement by DXA.