RESUMO
BACKGROUND AND AIMS: Heart failure with concomitant sarcopenia has a poor prognosis; therefore, simple methods for evaluating the appendicular skeletal muscle mass index (ASMI) are required. Recently, a model incorporating anthropometric data and the sarcopenia index (i.e., serum creatinine-to-cystatin C ratio [Cre/CysC]), was developed to estimate the ASMI. We hypothesized that this model was superior to the traditional model, which uses only anthropometric data to predict prognosis. This retrospective cohort study compared the prognostic value of low ASMI as defined by the biomarker and anthropometric models in patients with heart failure. METHODS AND RESULTS: Among 847 patients, we estimated ASMI using an anthropometric model (incorporating age, body weight, and height) in 791 patients and a biomarker model (incorporating age, body weight, hemoglobin, and Cre/CysC) in 562 patients. The primary outcome was all-cause mortality. Overall, 53.4% and 39.1% of patients were diagnosed with low ASMI (using the Asian Working Group for Sarcopenia cut-off) by the anthropometric and biomarker models, respectively. The two models showed a poor agreement in the diagnosis of low ASMI (kappa: 0.57, 95% confidence interval: 0.50-0.63). Kaplan-Meier curves showed that a low ASMI was significantly associated with all-cause death in both models. However, this association was retained after adjustment for other covariates in the biomarker model (hazard ratio: 2.32, p = 0.001) but not in the anthropometric model (hazard ratio: 0.79, p = 0.360). CONCLUSION: Among patients hospitalized with heart failure, a low ASMI estimated using the biomarker model, and not the anthropometric model, was significantly associated with all-cause mortality.