By Jamie L. W. Kennedy, MD, FACC

Associate Professor, Division of Cardiology, Advanced Heart Failure & Transplant Cardiology, University of California, San Francisco

Dr. Kennedy reports no financial relationships relevant to this field of study.

SYNOPSIS: Long-term follow-up of SCD-HeFT did not show any benefit in installing implantable cardioverter-defibrillator devices in patients with New York Heart Association class III symptoms or nonischemic cardiomyopathy.

SOURCE: JE Poole, B Olshansky, DB Mark, et al. Long-term outcomes of implantable cardioverter-defibrillator therapy in the SCD-HeFT. J Am Coll Cardiol 2020;76:405-415.

Most heart failure deaths fall into two categories: sudden death caused by ventricular arrhythmias and pump failure. The landmark Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT), in combination with the Multicenter Automatic Defibrillator Implantation Trial II (MADIT-II), established the role of implantable cardioverter-defibrillators (ICDs) for primary prevention of sudden cardiac death.

The authors of SCD-HeFT enrolled 2,521 patients with New York Heart Association (NYHA) class II-III symptoms and left ventricular ejection fraction ≤ 35% between 1997 and 2001. Patients were randomized to placebo, amiodarone, or single-lead ICD. Medical management in all arms was similar. On enrollment, nearly 100% of subjects received angiotensin-modulating drugs, although only 69% were prescribed beta-blockers. There were slightly more ischemic than nonischemic patients (52% vs. 48%). Over a median follow-up of 45.5 months, the mortality in the ICD group was 22%, significantly lower than both amiodarone (29%) and placebo (28%) groups (P = 0.007). There were two prespecified subgroup analyses: functional class and cardiomyopathy etiology. NYHA class II patients clearly exhibited a mortality benefit (hazard ratio [HR], 0.54; P < 0.001) while NYHA class III patients did not (HR, 1.16; P = 0.30). The mortality benefit of ICD was similar in ischemic (HR, 0.79; P = 0.05) and nonischemic (HR, 0.73; P = 0.06) patients. Eleven percent of patients in the placebo and amiodarone arms underwent ICD implant during the study, one-third as part of a cardiac resynchronization device. In the ICD arm, 31% of patients received an ICD shock, only 68% of which were appropriate, and 2.5% underwent upgrade to a biventricular pacemaker (CRT-D).

In a long-term follow-up study, Poole et al collected data between 2010 and 2011 for a median follow-up of 11 years. Vital status was known for 91% of the 2,521 originally enrolled patients. Medical management patients were offered ICDs at the end of the original trial (59% of amiodarone and 55% of placebo patients are known to have undergone implantation). About 40% of implanted devices were CRT-Ds. In the ICD arm, 27% are known to have undergone CRT-D upgrades.

Overall, mortality in the ICD arm was 52.5%, amiodarone 52.7%, and placebo 57.2%. The difference between ICD and placebo was statistically significant (P = 0.028). Ischemic patients demonstrated improved survival in the ICD arm compared to placebo (mortality, 59.4% vs. 68%; P = 0.009). However, in the nonischemic patients, mortality was similar between the ICD and placebo groups (45.1% vs. 44.0%; P = 0.802). NYHA class II patients in the ICD group demonstrated improved survival compared to placebo (44.6% vs. 52.1%; P = 0.001). However, NYHA class III patients did not derive benefit from ICD implant (mortality, 69.7% vs. 68.9% in the placebo arm; P = 0.575). In an as-treated analysis, ICD implantation resulted in reduced mortality (HR, 0.82; 95% confidence interval [CI], 0.72-0.96; P = 0.008). In a time-based analysis, the benefit of ICD implantation was unclear more than six years after enrollment. The authors concluded the long-term follow-up of SCD-HeFT-enrolled patients failed to show ICD implantation was beneficial in patients with NYHA class III symptoms or nonischemic cardiomyopathy.


The results of this long-term analysis are limited by incomplete follow-up and a high crossover rate. Medical management has continued to improve. In the period of this analysis, beta-blocker use increased, and the use of aldosterone antagonists and cardiac resynchronization therapy have become widespread. The observed loss of benefit of ICD implantation six years after enrollment is not unexpected considering these factors. The lack of benefit in NYHA class III patients was seen at the conclusion of the initial trial and persisted through the long-term follow-up phase. Similarly, a meta-analysis of SCD-HeFT, the Multicenter Automatic Defibrillator Trial I (MADIT-I), MADIT-II, and the Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation (DEFINITE) revealed statistically significant reduction in mortality in NYHA class II patients, but not in NYHA class III patients. Attempts at risk stratification in heart failure have been limited, but some common themes have emerged. Older age, hyponatremia, renal dysfunction, hypotension, frequent hospitalizations, poor functional status, and inability to tolerate guideline-directed medical therapy all convey poor prognosis. In clinical practice, patients with one or more of these features warrant a thorough assessment. If patients do not improve with medical optimization (e.g., adequate diuresis or transition from angiotensin-converting enzyme inhibitors to sacubitril/valsartan), then ICD implantation seems reasonable. If not, then it seems unlikely that a primary prevention ICD will be beneficial. The benefit of ICD implantation in patients with nonischemic cardiomyopathy has been unclear despite several clinical trials. The DEFINITE trial concerned nonischemic cardiomyopathy patients with non-sustained ventricular tachycardia or frequent premature ventricular contractions, the reduction in mortality with ICD implantation was not considered statistically significant at two years of follow-up (14.1% control vs. 7.9% ICD; P = 0.08). The DANISH trial authors enrolled patients with nonischemic cardiomyopathy and similarly found no statistically significant reduction in mortality at a mean 67.6 months of follow-up (HR, 0.87; 95% CI, 0.68-1.12; P = 0.28).

Part of this uncertainty may be the heterogeneity of the nonischemic population. Some genetic cardiomyopathies, like arrhythmogenic right ventricular cardiomyopathy, lamin A/C, and sarcoidosis, carry substantial arrhythmic risks. More detailed characterization of nonischemic patients enrolled in clinical trials, and in practice, is long overdue to facilitate the next generation of tailored therapy.