Abstract Body (Do not enter title and authors here): Background: There are currently no robust clinical markers for assessing prognosis in patients with heart failure (HF) with recovered left ventricular ejection fraction (LVEF).
Hypothesis: We hypothesized that N-terminal pro-B-type natriuretic peptide (NT-proBNP) measured at the time of LVEF recovery is an independent predictor of LVEF relapse and adverse clinical outcomes in patients with HF with recovered LVEF.
Methods: This retrospective cohort study (2009–2024) included 3935 patients with HF with recovered LVEF (≥50%) and available NT-proBNP data at the time of LVEF recovery. Patients were categorized into seven NT-proBNP groups, which were compared using Kaplan-Meier analysis and multivariable Cox regression to evaluate the outcome of LVEF relapse (decrease by ≥10% to <50%) and the composite outcome of HF hospitalization or all-cause death. NT-proBNP was also modeled as a continuous variable using natural cubic splines. Utilization of medications constituting guideline-directed medical therapy (GDMT) was compared across NT-proBNP groups at 6 months of follow-up.
Results: The median value of baseline NT-proBNP was 1341 pg/mL (IQR, 400–4207). In general, the probability of remaining free from LVEF relapse and the composite outcome decreased across NT-proBNP groups. In the fully adjusted multivariable Cox regression model, NT-proBNP was an independent predictor of both LVEF relapse and the composite outcome, with higher NT-proBNP levels associated with higher risk of both outcomes in a dose-response manner. Notably, even near-normal NT-proBNP levels (125–299 pg/mL) were associated with poorer prognosis relative to normal levels (<125 pg/mL), with a 46% higher risk of LVEF relapse and 82% higher risk of the composite outcome. This relationship was consistent across age, sex, atrial fibrillation status, and renal function, but was modified by body mass index (BMI), with higher BMI associated with higher risk. NT-proBNP was even predictive of the composite outcome when patients sustained LVEF recovery and did not experience LVEF relapse. Patients with higher baseline NT-proBNP levels did not have higher utilization of GDMT medications at 6 months of follow-up, and in some cases even had lower utilization.
Conclusions: NT-proBNP is an independent and robust predictor of prognosis in patients with HF with recovered LVEF, warranting routine measurement of NT-proBNP in these patients for further GDMT optimization.