Evaluation of the Agreement and Clinical Utility of Anthropometric Equation-Estimated Appendicular Skeletal Muscle Mass Compared to Dual-Energy X-ray Absorptiometry in Patients with Type 2 Diabetes Mellitus.

Abstract

Objective: Sarcopenia is increasingly recognized as a critical complication of type 2 diabetes mellitus (T2DM), representing the convergence of global population ageing and escalating metabolic disease prevalence. Reliable assessment of appendicular skeletal muscle mass (ASM) is essential for timely detection and effective intervention. This study evaluated the validity of a widely adopted anthropometric equation for estimating ASM in healthy adults and T2DM patients, using dual-energy X-ray absorptiometry (DXA) as the reference method.

Methods: We retrospectively analyzed 402 adults who underwent DXA at Linyi People's Hospital from January 2016 to December 2022, including 175 patients with T2DM and 227 healthy controls. DXA-derived ASM and skeletal muscle index (SMI) were obtained from whole-body scans. Predicted ASM (ASM2) was calculated using the Wen equation, and predicted SMI (SMI2) was derived accordingly. Low muscle mass was defined by Asian Working Group for Sarcopenia 2019 criteria. Correlation, Bland-Altman analysis, Lin's concordance correlation coefficient, calibration, receiver operating characteristic (ROC) analysis, multivariable regression, and age-stratified analyses were performed. Shannon entropy was used exploratorily to describe subgroup distributional heterogeneity.

Results: ASM2 correlated strongly with DXA-measured ASM in both healthy controls (r = 0.882) and T2DM patients (r = 0.871; both P < 0.001), with no significant difference between groups. However, Bland-Altman analysis showed systematic overestimation of ASM by the Wen equation, with only moderate agreement overall and wider limits of agreement in T2DM, especially in women. Concordance was better in men than women. For low muscle mass detection, SMI2 performed well in healthy men (area under the curve (AUC) = 0.851) and men with T2DM (AUC = 0.858), acceptably in healthy women (AUC = 0.793), but poorly in women with T2DM (AUC = 0.596), who also had the highest misclassification rate (40.0%). Higher body mass index independently predicted greater estimation error; age did not.

Conclusions: The Wen equation may be useful for preliminary population-level screening, but its individual diagnostic utility is limited by systematic overestimation and moderate agreement with DXA. Caution is especially needed in women with T2DM, for whom confirmatory DXA should be considered when feasible.