Risk stratification at discharge identifies predictors of recurrent hospitalization in acute decompensated heart failure

Background Rehospitalization after acute decompensated heart failure (ADHF) remains frequent despite optimized guideline-directed medical therapy (GDMT). Residual congestion, renal dysfunction, anemia, reduced ejection fraction, and recurrent atrial fibrillation contribute to recurrent decompensation, yet reliable discharge markers of vulnerability remain insufficiently defined. Aim To identify independent predictors of 24-month rehospitalization 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 the intensive care unit between January 2023 and January 2024. Pre-discharge assessment included clinical characteristics; renal function; natriuretic peptides; high-sensitivity troponin; inflammatory indices, including neutrophil-to-lymphocyte ratio and platelet-to-leukocyte ratio; the C-reactive protein-to-albumin ratio (CAR) as a composite marker of inflammation and nutritional status; the FIB-4 index as a fibrosis marker; and comprehensive echocardiography. Patients were classified according to left ventricular ejection fraction as HFrEF (n=194), HFmrEF (n=110), or HFpEF (n=252), and were discharged on optimized GDMT. Cox regression and Kaplan–Meier analyses evaluated time to first rehospitalization. Results The cohort was elderly (74.3 ± 10.6 years), with mean LVEF 43.8 ± 13.6%, mean eGFR 61.8 ± 23.3 ml/min/1.73 m², and permanent atrial fibrillation present in 59% of patients. During 24-month follow-up, rehospitalization occurred in 42.1% (234/556). In multivariable Cox regression analysis (Table 1), lower eGFR, higher E/e′ ratio (reflecting elevated filling pressures), and elevated log-transformed CAR were independently associated with shorter time to recurrent decompensation. Other evaluated biomarkers and echocardiographic parameters were not independently predictive after adjustment. Kaplan–Meier analysis demonstrated progressively reduced event-free survival across CAR tertiles (low, n=186; intermediate, n=184; high, n=186), with event rates of 32%, 45%, and 49%, respectively (log-rank p<0.001; Figure 1). Conclusion In this ADHF cohort discharged on optimized GDMT, rehospitalization was primarily associated with cardiorenal dysfunction and inflammation–nutritional burden. These findings highlight the importance of comprehensive patient assessment at discharge and support a multidimensional approach to post-discharge risk stratification.