Sudden Death in HFpEF: TOPCAT Analysis

Study Questions:

What are the rates and predictors of sudden death (SD) or aborted cardiac arrest (ACA) in heart failure with preserved ejection fraction (HFpEF)?


They study cohort was comprised of 1,767 patients with HFpEF (EF ≥45%) enrolled in the Americas region of the TOPCAT trial. In this study, SD was defined as an unexpected death in an otherwise stable patient and was classified as either witnessed (if death was observed or if last seen within 24 hours) or presumed (if last seen ≥24 hours with the clinical context suggestive of SD). Aborted cardiac arrest (ACA), a prespecified component of the primary composite endpoint for the overall TOPCAT trial, was defined as successful resuscitation after cardiac arrest (with or without antecedent myocardial infarction [MI] or HF) with meaningful recovery. For the purposes of this analysis, all deaths not adjudicated as SD or ACA were considered to be non-SD/ACA and treated as a competing risk. The study investigators identified independent predictors of composite SD/ACA with step-wise backward selection using competing risks regression analysis accounting for nonsudden causes of death.


During a median of 3.0 (1.9-4.4) years, 385 patients died: 72 (19% of total deaths) of SD (63 witnessed, 9 presumed) and 313 (81% of total deaths) of non-SD. There were 6 adjudicated ACA events with variable ultimate outcomes after resuscitation (4 alive in follow-up, 1 SD, 1 infectious death). The outcome of interest of composite SD or ACA occurred in 77 patients and 312 experienced non-SD/ACA. Corresponding incidence rates were 1.4 (1.1-1.8) and 5.8 (5.1-6.4) events/100 patient-years. SD/ACA was numerically lower, but not statistically reduced in those randomized to spironolactone: 1.2 (0.9-1.7) vs. 1.6 (1.2-2.2) events/100 patient-years; sub-distributional hazard ratio, 0.74; 95% confidence interval, 0.47-1.16; p = 0.19. Compared with patients who experienced non-SD/ACA, patients who experienced SD/ACA tended to be younger (70 ± 9 vs. 75 ± 9 years), men (68% vs. 54%), with higher body mass indexes (35 ± 9 vs. 32 ± 9 kg/m2) and rates of diabetes mellitus (58% vs. 45%); p ≤ 0.05 for all comparisons. Left ventricular EF (LVEF) was similar in patients who experienced SD/ACA (56 ± 8%) and non-SD/ACA (57 ± 8%; p = 0.20). Rates of prior MI (29% vs. 24%) and coronary artery disease (57% vs. 49%) did not differ in those who experienced of SD/ACA and non-SD/ACA (p > 0.05 for both comparisons). After accounting for competing risks of non-SD, male gender and insulin-treated diabetes mellitus were independently predictive of composite SD/ACA with modest discrimination (C-statistic 0.65). Covariates—including eligibility criteria, age, LVEF, coronary artery disease, left bundle branch block, and baseline therapies—were not independently associated with SD/ACA. Gender and diabetes mellitus status remained independent predictors in sensitivity analyses excluding patients with implantable cardioverter-defibrillators and when predicting SD alone. Although spironolactone appeared to reduce clinical events, including cardiovascular death, in the Americas region of TOPCAT, spironolactone did not significantly reduce risk of SD/ACA as compared with placebo.


The authors concluded that SD accounted for approximately 20% of deaths in HFpEF. Male gender and insulin-treated diabetes mellitus identify patients at higher risk for SD/ACA with modest discrimination.


Although the overall number of SD/ACA events was limited, this is an important paper because it suggests that the ‘burden’ of SD in HFpEF is significant. As the authors point out, better predictors of SD is the next step to better understanding the natural history of this condition. Clinical trials targeting HFpEF patients should consider SD as an important therapeutic goal.

Clinical Topics: Arrhythmias and Clinical EP, Geriatric Cardiology, Heart Failure and Cardiomyopathies, Atherosclerotic Disease (CAD/PAD), Implantable Devices, EP Basic Science, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Acute Heart Failure, Chronic Heart Failure

Keywords: ACC18, ACC Annual Scientific Session, Arrhythmias, Cardiac, Bundle-Branch Block, Coronary Artery Disease, Death, Sudden, Defibrillators, Implantable, Diabetes Mellitus, Geriatrics, Heart Arrest, Heart Failure, Heart Failure, Diastolic, Insulins, Myocardial Infarction, Risk, Spironolactone, Stroke Volume

< Back to Listings