Performance of Four Sarcopenia Screening Tools for Possible Sarcopenic Obesity in Hospitalised Patients

Abstract Background Sarcopenic obesity (SO) is increasingly recognised as a high-risk geriatric condition, particularly among hospitalised older adults. In routine inpatient care, however, its identification is often limited by the availability of body-composition measurements and physical performance testing. Simple screening approaches that can be applied at the bedside are therefore of practical importance. This study compared the performance of four commonly used sarcopenia screening tools for identifying possible sarcopenic obesity in hospitalised older adults. Methods Inpatients from multiple departments of a tertiary hospital were enrolled. Owing to the lack of uniform ASMI and physical performance measurements, possible sarcopenic obesity was operationally defined as BMI ≥ 24 kg/m² combined with low handgrip strength (men < 28 kg; women < 18 kg) following AWGS guidelines. ROC curves and AUC values were computed. A cross-sectional study was conducted among 1,233 inpatients aged 60 years or older from multiple clinical departments of a tertiary hospital in China. Under real-world clinical constraints, possible sarcopenic obesity was defined as a body mass index (BMI) ≥ 24 kg/m² in combination with low handgrip strength (< 28 kg in men and < 18 kg in women), based on Asian Working Group for Sarcopenia recommendations. The screening performance of SARC-F, SARC-CalF, Ishii score and SARC-F + EBM was evaluated using receiver operating characteristic curves and area under the curve values. Analyses were repeated using a higher BMI threshold (≥ 28 kg/m²). Results When BMI ≥ 24 kg/m² plus low grip strength was used as the reference, SARC-F showed the highest overall discriminatory ability (AUC 0.56, 95% CI 0.53–0.60), with a cut-off score of 2 providing a reasonable balance between sensitivity and specificity. In contrast, SARC-CalF demonstrated poor performance, largely driven by very low sensitivity despite high specificity. The Ishii score and SARC-F + EBM showed limited discriminatory capacity. Increasing the BMI threshold to ≥ 28 kg/m² had little effect on the performance of SARC-F, whereas tools incorporating calf circumference showed further reductions in accuracy. Conclusion Among the screening tools examined, SARC-F appeared to be the most stable and feasible option for identifying possible sarcopenic obesity in hospitalised older adults. Functional screening tools may be more suitable than anthropometric-based measures for bedside risk stratification in inpatient populations with higher levels of adiposity.