Mode of Death in Patients With Heart Failure and Preserved Ejection Fraction: Insights From PARAGON-HF Trial

Abstract

Background: Patients with heart failure (HF) and preserved left ventricular ejection fraction comprise a heterogeneous group including some with mildly reduced EF. We hypothesized that mode of death differs by EF in ambulatory patients with HF and preserved left ventricular ejection fraction. Methods: PARAGON-HF trial (Prospective Comparison of Angiotensin Receptor–Neprilysin Inhibitor With Angiotensin-Receptor Blocker Global Outcomes in Heart Failure With Preserved Ejection Fraction) compared clinical outcomes in 4796 patients with chronic HF and EF ≥45% randomly assigned to sacubitril/valsartan or valsartan. We examined the mode of death in relation to baseline EF in logistic regression models and the effect of randomized treatment on cause-specific death in Cox regression models. Nonlinear relationships with continuous EF were modelled using quadratic and cubic terms. Results: Of 691 deaths during the trial, 416 (60%) were ascribed to cardiovascular, 220 (32%) to noncardiovascular, and 55 (8%) to unknown causes. Of cardiovascular deaths, 154 (37%) were due to sudden death, 118 (28%) were due to HF, 35 (8%) to stroke, 27 (6%) to myocardial infarction, and 82 (20%) to other cardiovascular causes. Rates of all-cause, cardiovascular, and sudden death were higher in those with lower left ventricular ejection fraction (all P <0.001), while rates of non-cardiovascular death were greater in patients with higher EF. Sacubitril/valsartan did not reduce overall death, cardiovascular death, or sudden death compared with valsartan, irrespective of baseline EF (all P for interaction >0.30). Conclusions: Among patients with HF and preserved left ventricular ejection fraction enrolled in PARAGON-HF, the proportion of cardiovascular and sudden death were higher in those with lower left ventricular EF, and the proportion of noncardiovascular death rose with EF. Regardless of EF, sacubitril/valsartan did not reduce death from any cause compared with valsartan. Registration: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT01920711.