Author + information
- Rachel Lampert, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Rachel Lampert, Yale University School of Medicine, Section of Cardiovascular Medicine, 789 Howard Avenue, Dana 319, New Haven, Connecticut 06520
The benefit of prophylactic implantation of an implantable cardioverter-defibrillator (ICD) or cardiac resynchronization therapy (CRT) ICD (CRT-D) in improving mortality in patients with heart failure has been firmly established by multiple randomized controlled trials (RCT) (1–4). However, it has also been well documented that patients receiving ICDs in the “real world” are older and sicker, with higher prevalences of diabetes, hypertension, and atrial fibrillation than are those enrolled in the landmark RCTs (1,3,5). It is an inherent limitation of RCTs that strict entry criteria limit “real-world” generalizability (6), and attempting to understand whether the findings of the RCTs can be extrapolated to these sicker patients is an important endeavor. Several previous studies of varying design have suggested that overall the benefit of the ICD for primary prevention of sudden cardiac death in patients in the real world is similar to that in the trials. Using administrative data from the Medicare database, Groeneveld et al. (7) compared 7,125 ICD recipients to a propensity-score matched group, finding a hazard ratio (HR) for mortality of 0.62 with the ICD. In a population-based study, Parkash et al. (8) compared outcomes between all patients receiving an ICD in Nova Scotia and comparable patients in a province-wide database. They found a dramatic survival benefit from the ICD, with an unadjusted HR for mortality of 0.46, going to 0.55 to 0.59 after multivariable and propensity analyses. In the most rigorous study to date of real-world benefit, Al-Khatib et al. (9) used patient-level data to compare mortality between ICD recipients in the NCDR (National Cardiovascular Data Registry) meeting criteria for MADIT-II (Multicenter Automatic Defibrillator Implantation Trial II) (3) or SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) (1) and the participants in those trials, looking at both ICD recipients and those randomized to standard-of-care medical therapy. Despite the fact that the NCDR patients were older and had greater comorbidities, the mortality in the NCDR patients was similar to ICD recipients in the trials, and significantly better than that in the trial control groups (HR: 0.73; 95% confidence interval [CI]: 0.59 to 0.92; p = 0.007 for the MADIT-II comparison, and HR: 0.82; 95% CI, 0.7 to 0.96; p = 0.01 for the SCD-HeFT comparison).
The study by Chen et al. (10) in this issue of the Journal, analyzing data from the Centers for Medicare and Medicaid Services subset of the NCDR ICD registry, adds to the characterization of real-world ICD recipients as sicker overall than were participants in MADIT-II and SCD-HeFT, with a higher burden of chronic heart failure, with 15% having had 1 heart failure admission within the previous year and 5% having had 2 or more. They also describe a high burden of acute heart failure, with 31% in the hospital 2 or more days before the implantation, suggesting a nonelective implantation. The higher the burden of both chronic and acute heart failure, the higher the mortality in the 1, 2, and 3 years after ICD implantation. They also contrast the mortality of this Medicare cohort with the published data on mortality in the ICD groups of MADIT-II and SCD-HeFT, noting higher mortality in the NCDR subgroups of patients with higher heart failure burden compared with ICD recipients in these trials.
It is important to note, however, that 48% of the patients analyzed were recipients of CRT-D devices. Whereas patients in SCD-HeFT and MADIT-II had stable heart failure, those in COMPANION (Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure), the pivotal study demonstrating CRT-D benefit, had class III or IV congestive heart failure (CHF), and at least 1 previous heart failure hospitalization in the past year was required for inclusion in that trial (4). Only 5% of this Medicare population had a greater burden of CHF, as defined here based on previous hospitalizations, than was required for the COMPANION participants. Similarly, with the exception of the very highest heart failure burden groups, the mortalities are more similar to those of COMPANION CRT-D recipients. The combining of ICD and CRT-D patients in the current analysis makes contrast with previous study populations difficult.
Despite this limitation, the very high mortality at 1 year in the highest-heart-failure burden groups raises the provocative question: Does the benefit of the ICD seen in the RCTs extend to those with significantly higher heart failure burden than trial participants, such as the 5% with >2 CHF admissions in the preceding year or 6% with a prolonged preceding hospital stay? Without comparison against a group of comparable patients who did not receive an ICD, this study cannot in itself answer this important question. Although those with a high burden of heart failure are more likely to die after an ICD than are those with a low burden, the question most relevant to patient care is whether or not patients with a high burden of heart failure are less likely to die with an ICD than they are without one, and the question cannot be determined by the data here. As discussed by Chen et al. (10), there are theoretic reasons why patients with severe CHF might not benefit from an ICD, as the percentage of death that is sudden decreases as heart failure severity increases. In a subanalysis of MADIT-II (11), among those receiving dual-chamber (but not single-chamber) devices, the ICD no longer conferred benefit after a CHF hospitalization. To what extent those findings could be extrapolated to all-comers regarding pre-implant hospitalization, and to CRT candidates, is unknown. As the investigators note, given the high competing mortality demonstrated in the current study in patients with high burden of heart failure, further research is needed to determine whether ICD implantation during a prolonged CHF hospitalization or following multiple hospital admissions will confer survival benefit.
Identification of subgroups in whom the ICD will not confer benefit, particularly due to competing early mortality, is an important endeavor. Whereas ICD patients for the most part enjoy quality of life similar to that of other cardiac patients (12), shocks (13), particularly at the end of life (14), do affect quality of life for patients and families. Whereas ICD deactivation can avert shocks at the end of life (15), implantation with deactivation within a year is not of benefit either to patients or a cost-conscious society. It is often difficult to predict impending death in CHF, which waxes and wanes, underscored by the fact that many patients in this cohort had had 3 or more CHF hospitalizations in the preceding year, which is aggressive care in itself. The very high mortality documented here emphasizes the importance of integration of palliative care, underused in advanced heart failure (16), and of careful consideration of aggressive therapies for those with highest burden of CHF.
Dr. Lampert has received significant research support from Medtronic, Boston Scientific, and St. Jude Medical; and modest honoraria from Medtronic.
↵⁎ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
- American College of Cardiology Foundation
- Parkash R.,
- Sapp J.L.,
- Basta M.,
- et al.
- Chen C.-Y.,
- Stevenson L.W.,
- Stewart G.C.,
- et al.
- Goldenberg I.,
- Moss A.J.,
- Hall W.J.,
- et al.,
- MADIT-II Investigators
- Leosdottir M.,
- Sigurdsson E.,
- Reimarsdottir G.,
- et al.
- Lampert R.,
- Hayes D.L.,
- Annas G.J.,
- et al.,
- American College of Cardiology,
- American Geriatrics Society,
- American Academy of Hospice and Palliative Medicine,
- American Heart Association,
- European Heart Rhythm Association,
- Hospice and Palliative Nurses Association