Abstract
Background Heart failure imposes a significant healthcare burden, with early unplanned readmissions post-discharge linked to poor outcomes. Identifying risk factors and their predictive value is crucial for targeted interventions.
Objective To explore factors influencing 30-day unplanned readmission in heart failure and their predictive value.
Methods A prospective study of heart failure patients hospitalized in Beijing Hospital from October 2023 to March 2024. Patients were assessed for nutritional status using Mini-Nutritional Assessment Scale Short Version (MNA-SF), frailty using Groningen Frailty Index (GFI), and Appendicular Skeletal Muscle Mass Index (ASMI) calculation. Multifactorial COX regression analyse was conducted, and ROC curves plotted for predictive modeling.
Results A total of 121 heart failure patients (60.3% males), aged (69.87±11.9) years were included. Within 30 days-median follow-up duration, 25 (20.7%) patients with readmission. COX regression analysis stratified by gender showed that age, regular smoking, nutritional status, left ventricular ejection fraction(LVEF), brain natriuretic peptide(BNP), GFI, and ASMI were independent influences on readmission within 30 days in patients with heart failure (P<0.050). ROC curve analysis showed that age, BNP, ASMI, smoking status, LVEF, nutritional status, and GFI individually as well as in combination predicted readmission within 30 days in patients with heart failure; the joint model performed optimally, with an AUC value reaching 0.877 (95%CI 0.801~0.952, P<0.001), corresponded to a sensitivity of 0.920 and a specificity of 0.729.
Conclusion A multifactorial approach including age, BNP, ASMI, smoking status, LVEF, nutritional status, and GFI predicts 30-day readmission risk, offering a basis for clinical intervention strategies to improve patient outcomes.