Long-term outcomes in heart failure with preserved ejection fraction: Predictors of cardiac and non-cardiac mortality

Aims: Heart failure (HF) with preserved ejection fraction (HFpEF) is associated with cardiovascular (CV) and non-CV events, but long-term risk is poorly studied. We assessed incidence and predictors of the long-term CV and non-CV events.

Methods and results: Patients presenting with acute HF, EF ≥ 45%, and N-terminal pro-brain natriuretic peptide > 300 ng/L were enrolled in the Karolinska-Rennes study in 2007-11 and were reassessed after 4-8 weeks in a stable state. Long-term follow-up was conducted in 2018. The Fine-Gray sub-distribution hazard regression was used to detect predictors of CV and non-CV deaths, investigated separately from baseline acute presentation (demographic data only) and from the 4-8 week outpatient visit (including echocardiographic data). Of 539 patients enrolled [median age 78 (interquartile range: 72-84) years; 52% female], 397 patients were available for the long-term follow-up. Over a median follow-up time from acute presentation of 5.4 (2.1-7.9) years, 269 (68%) patients died, 128 (47%) from CV and 120 (45%) from non-CV causes. Incidence rates per 1000 patient-years were 62 [95% confidence interval (CI) 52-74] for CV and 58 (95% CI 48-69) for non-CV death. Higher age and coronary artery disease (CAD) were independent predictors of CV death, and anaemia, stroke, kidney disease, and lower body mass index (BMI) and sodium concentrations of non-CV death. From the stable 4-8 week visit, anaemia, CAD, and tricuspid regurgitation (>3.1 m/s) were independent predictors of CV death, and higher age of non-CV death.

Conclusions: In patients with acute decompensated HFpEF, over 5 years of follow-up, nearly two-thirds of patients died, half from CV and the other half from non-CV causes. CAD and tricuspid regurgitation were associated with CV death. Stroke, kidney disease, lower BMI, and lower sodium were associated with non-CV death. Anaemia and higher age were associated with both outcomes. [Correction added on 24 March 2023, after first online publication: In the first sentence of the Conclusions, 'two-thirds' has been inserted before 'of patients died...' in this version.].