RESUMO
BACKGROUND: Identifying readmission predictors in heart failure (HF) patients may help guide preventative efforts and save costs. We aimed to identify 15- and 30-day readmission predictors due to cardiovascular reasons. METHODS AND RESULTS: A total of 1831 patients with acute HF admission were prospectively followed during a year. Patient-associated variables were gathered at admission/discharge and events during follow-up. A multivariate Fine and Gray competing risk regression model and a cumulative incidence function were used to identify predictors and build a risk score model for 15- and 30-day readmission. The 15- and 30-day readmission rates due to cardiovascular reasons were 7.1% and 13.9%. Previous acute myocardial infarction, congestive signs at discharge, and length of stay > 9 days were predictors of 15- and 30-day readmission, while much weight loss and large NT-ProBNP reduction were protective factors. The NT-ProBNP reduction was larger at 30 days (> 55%) vs 15 days (> 40%) to protect from readmission. Glomerular filtration rate at discharge < 60 mL/min/1.73m2 and > 1 previous admissions due to HF were predictors of 30-day readmission, while first post-discharge control at an HF unit was a protective factor. CONCLUSIONS: Previous identified factors for early readmission were confirmed. The NT-ProBNP reduction should be increased (> 55%) to protect from 30-day readmission.