Agreement between bioelectrical impedance analysis and equation-derived estimates of appendicular skeletal muscle in acutely hospitalized multiethnic older patients

BACKGROUND: Accurate assessment of appendicular skeletal muscle (ASM) is essential for diagnosing sarcopenia. While bioelectrical impedance analysis (BIA) is widely used, equation-derived estimates may be useful when device-based methods are unavailable. We evaluated agreement between BIA-measured ASM and two equations: the COCONUT Study Group formula (NHANES-based) and the Japanese equation by Furushima et al. METHODS: We conducted a cross-sectional study of hospitalized, multiethnic older adults in Singapore (n = 295). ASM was measured using multi-frequency BIA S10. Equation-derived ASM was calculated using calf circumference, age, and sex (COCONUT) or age, anthropometry, and handgrip strength (HGS; Furushima). Agreement with BIA was assessed by Lin's concordance correlation coefficient (CCC) and Bland-Altman analysis. Diagnostic agreement for low appendicular skeletal muscle mass index (ASMI; AWGS 2019 cut-offs) was evaluated with Cohen's κ and area under the receiver operating characteristic curve (AUC). Analyses were stratified by sex and obesity. RESULTS: Furushima's equation showed stronger concordance with BIA than COCONUT (CCC 0.696 vs. 0.538) and smaller underestimate bias (1.56 vs. 3.03 kg). For low ASMI, Furushima achieved moderate agreement (κ = 0.48; AUC = 0.78) compared with poor agreement for COCONUT (κ = 0.17; AUC = 0.62). COCONUT performed poorly among women (CCC = 0.192; bias 4.40 kg), whereas Furushima improved performance but remained low overall (CCC = 0.371). Among obese patients, Furushima demonstrated the highest concordance (CCC = 0.862; bias 0.42 kg). CONCLUSIONS: Furushima's equation aligns better with BIA-measured ASM than COCONUT and may complement HGS in sarcopenia screening. Equation-based estimation could aid risk stratification when BIA is unavailable, though subgroup-specific biases require caution.