Recurrent hospitalizations are associated with increased mortality across the ejection fraction range in heart failure

AIMS: The proportion of patients hospitalized for heart failure (HF) with preserved left ventricular ejection fraction (LVEF) is rising, but no approved treatment exists, in part owing to incomplete characterization of this particular HF phenotype. In order to better define the characteristics of HF phenotypes in Finland, a large cohort with 12 years' follow‐up time was analysed. METHODS AND RESULTS: Patients diagnosed between 2005 and 2017 at the Hospital District of Southwest Finland were stratified according to LVEF measure and N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP) levels. For this retrospective registry study, previously diagnosed HF patients were defined as follows: patients with reduced ejection fraction (HFrEF; LVEF ≤ 40%; n = 4042), mid‐range ejection fraction (HFmrEF; LVEF > 40–50% and NT‐proBNP ≥ 125 pg/mL; n = 1468), and preserved ejection fraction (HFpEF; LVEF > 50% and NT‐proBNP ≥ 125 pg/mL; n = 3122) and followed up for 15 022, 4962, and 10 097 patient‐years, respectively. Cardiovascular (CV) hospitalization and mortality, influence of pre‐selected covariates on hospitalization and mortality, and the proportion of HFpEF and HFmrEF patients with a drop in LVEF to HFrEF phenotype were analysed. All data were extracted from the electronic patient register. HFrEF patients were rehospitalized slightly earlier than HFpEF/HFmrEF patients, but the second, third, and fourth rehospitalization rates did not differ between the subgroups. Female gender and better kidney function were associated with reduced rehospitalizations in HFmrEF and HFrEF, with a non‐significant trend in HFpEF. Each additional hospitalization was associated with a two‐fold increased risk of death and 2.2‐ to 2.3‐fold increased risk of CV death. All‐cause mortality was higher in patients with HFpEF. Although CV mortality was less frequent in HFpEF patients, it was associated with increased NT‐proBNP concentrations at index in all patient groups. During the 10 years following the index date, 26% of HFmrEF patients and 10% of HFpEF patients progressed to an HFrEF phenotype. CONCLUSIONS: These findings suggest that disease progression, in terms of increased frequency of hospitalizations, and the relationship between increased number of hospitalizations and mortality are similar by LVEF phenotypes. These data highlight the importance of effective treatments that can reduce hospitalizations and suggest a role for monitoring NT‐proBNP levels in the management of HFpEF patients in particular.