Abstract
BACKGROUNDS: The optimal multidimensional frailty assessment approach in heart failure (HF) remains uncertain. This study aimed to compare the prognostic value of a multi-domain approach versus the Tilburg Frailty Indicator (TFI) in elderly HF patients. METHODS: This prospective single-center cohort study consecutively enrolled 461 elderly HF patients at Fuzhou University Affiliated Provincial Hospital, China. Frailty was assessed using both the TFI and a multi-domain approach. The primary outcome was the composite of heart failure hospitalization or all-cause death within six months. Clinical outcomes were analyzed using Kaplan-Meier and Cox proportional hazards regression models, with predictive performance compared by receiver operating characteristic curves. RESULTS: Among 461 participants, physical, psychological, and social frailty prevalence rates were 41.0%, 59.9%, and 56.8%, respectively. Multidimensional frailty, assessed by the TFI, was present in 77.4% of the cohort. The risk of composite outcome increased significantly with the number of frailty domains [2 domains: adjusted HR = 1.74 (95% CI: 1.08-2.79); 3 domains: HR = 2.07 (1.28-3.35)]. TFI-based analysis yielded consistent results. Notably, the predictive accuracy of frailty domains (area under the curve (AUC) = 0.63, 95% CI: 0.58-0.69) did not significantly differ from that of the TFI (AUC = 0.64, 95% CI: 0.59-0.70; P = 0.713), indicating comparable discriminative ability. CONCLUSIONS: In elderly HF patients, both the multi-domain frailty assessment and the TFI demonstrated comparable predictive value for heart failure hospitalization or all-cause death within six months.