Estimation of appendicular skeletal muscle: Development and validation of anthropometric prediction equations in Chinese patients with knee osteoarthritis

Abstract

To develop anthropometric prediction equations for estimating appendicular skeletal muscle (ASM) in Chinese knee osteoarthritis patients. Subjects were divided into the model development group (MD group: 104 cases, 47 men and 57 women) and cross-validation group (CV group: 69 cases, 38 men and 31 women). Stepwise multiple linear regression analyses were undertaken in the MD group to identify the best equations. Agreement between the estimated ASM and ASM measured by dual-energy X-ray absorptiometry (DXA) was tested in the CV group. Two models were developed in the MD group. Validation in the CV group showed that our models (R2 =0.83 and R2 =0.90) had a high coefficient of determination. The mean bias of ASM estimated by the two models from the ASM measured by DXA in the CV group showed no significant difference (P>0.05). These models could be useful for older Chinese patients with knee osteoarthritis to estimate ASM.