External validation of the prognostic performance of a heart failure-specific skeletal muscle mass equation: insights from the J-Proof HF registry.

Abstract

Background and aims: Skeletal muscle depletion in heart failure (HF) represents a key metabolic abnormality associated with adverse outcomes. However, most anthropometric equations for estimating appendicular skeletal muscle mass index (eASMI) were developed in non-HF populations, limiting their clinical applicability in HF. This study aimed to externally validate the prognostic performance of an HF-specific eASMI equation and compare it with that of general-population equations.

Methods and results: This secondary analysis included 9693 patients aged ≥65 years hospitalized for HF from a nationwide multicenter registry. The HF-specific eASMI equation was compared with five established equations using Cox regression, Akaike Information Criterion (AIC), C-statistics, and decision curve analysis (DCA). Lower eASMI was independently associated with increased 1-year mortality (adjusted hazard ratio per 1-SD decrease, 1.50; 95% confidence interval, 1.40-1.60), with a nonlinear dose-response relationship (p for nonlinearity, 0.016 in males and 0.009 in females). In males, the HF-specific equation showed better model fit than the general-population equations (ΔAIC: 4 to 60), whereas in females, it did not (ΔAIC: -12 to 43); discrimination was similar across equations in both sexes. Despite these differences, DCA demonstrated comparable clinical utility across all equations.

Conclusions: In older patients with HF, the HF-specific eASMI equation provides robust prognostic information reflecting disease-specific metabolic alterations. While statistical advantages were observed in males, all equations demonstrated comparable clinical utility. These findings support the use of anthropometric muscle mass estimation as a practical tool for metabolic risk stratification in HF.