Multisystem risk profiling at discharge identifies a high-risk phenotype in acute decompensated heart failure

Introduction Acute decompensated heart failure (ADHF) remains associated with substantial post-discharge mortality despite optimized guideline-directed medical therapy (GDMT). Traditional markers such as left ventricular ejection fraction (LVEF) and natriuretic peptides incompletely capture residual risk at discharge, and determinants of vulnerability across heart failure phenotypes remain insufficiently defined. Aim To identify independent predictors of 24-month all-cause mortality among clinical, biochemical, and echocardiographic parameters assessed prior to discharge in patients hospitalized with ADHF. Methods We prospectively enrolled 556 consecutive patients hospitalized with ADHF in intensive care unit between January 2023 and January 2024. All patients were discharged hemodynamically stable on GDMT. Pre-discharge assessment included clinical characteristics; renal function; natriuretic and myocardial injury biomarkers; inflammatory and fibrosis-related indices (including CRP/albumin ratio, platelet-to-leukocyte ratio, neutrophil-to-lymphocyte ratio, and FIB-4); and comprehensive echocardiography. Patients were classified according to LVEF as HFrEF (n=194), HFmrEF (n=110), or HFpEF (n=252). Cox regression and Kaplan–Meier analyses assessed associations with 24-month mortality. Results The cohort was elderly (74.3 ± 10.6 years), with atrial fibrillation (AF) present in 64% of patients; 59.2% were discharged in AF. Mean LVEF was 43.8 ± 13.6%, mean eGFR 61.8 ± 23.3 ml/min/1.73 m², and mean hemoglobin 131.8 ± 21.1 g/L. During 24-month follow-up, all-cause mortality was 36.3% (202/556). In multivariable Cox regression analysis (Table 1), AF emerged as the strongest independent predictor of mortality. Lower eGFR, lower hemoglobin, and higher FIB-4 index were also independently associated with increased risk. In contrast, traditional cardiac markers (including LVEF, NT-proBNP, and high-sensitivity troponin I) were not independently predictive after adjustment. Kaplan–Meier analysis confirmed significantly reduced survival among patients with AF (mean survival 17.6 vs 20.4 months; log-rank p=0.001)- figure 1. The AF × LVEF interaction was not significant (p=0.107), indicating a consistent prognostic effect of AF across HF phenotypes. Conclusion In this contemporary ADHF cohort discharged euvolemic on optimized GDMT, AF emerged as the strongest phenotype-independent predictor of 24-month mortality, surpassing conventional systolic markers and natriuretic peptides. Renal dysfunction, anemia, and elevated FIB-4 further defined a high-risk phenotype. These findings underscore the need for integrated, discharge-based risk stratification and suggest that rhythm status at discharge may represent a modifiable therapeutic target in ADHF.