Author + information
- Received April 12, 2016
- Revision received June 27, 2016
- Accepted July 28, 2016
- Published online January 16, 2017.
- Kairav Vakil, MDa,b,∗ (, )
- Sue Duval, PhDb,
- Rebecca Cogswell, MDb,
- Peter Eckman, MDc,
- Wayne C. Levy, MDd,
- Inderjit Anand, MD, DPhil (Oxon), PhDa,b,
- Todd Dardas, MD, MSd and
- Selcuk Adabag, MD, MSa,b
- aCardiovascular Division, Veterans Affairs Medical Center, Minneapolis, Minnesota
- bCardiovascular Division, University of Minnesota Medical School, Minneapolis, Minnesota
- cDivision of Cardiology, Minneapolis Heart Institute, Minneapolis, Minnesota
- dDivision of Cardiology, University of Washington, Seattle, Washington
- ↵∗Reprint requests and correspondence:
Dr. Kairav Vakil, Division of Cardiology, Department of Medicine, Veterans Affairs Medical Center & University of Minnesota, 111 C, One Veterans Drive, Minneapolis, Minnesota 55417.
Objectives This study sought to assess the impact of implantable cardioverter-defibrillators (ICDs) on waitlist mortality in patients listed for heart transplantation (HT).
Background The impact of ICDs on preventing sudden cardiac death in patients awaiting HT has not been studied in large multicenter cohorts. Furthermore, whether ICDs benefit patients with a left ventricular assist device (LVAD) is unknown.
Methods Adults (age ≥18 years) listed for first-time HT in the United States between January 1, 1999, and September 30, 2014, were retrospectively identified from the United Network for Organ Sharing registry. The primary predictor variable was the presence of an ICD at the time of listing. Primary outcome variable was all-cause waitlist mortality.
Results Data on 32,599 patients (mean age 53 ± 12 years, 77% male, 70% Caucasian) were analyzed. During median follow-up of 154 days, 3,638 patients (11%) died on the waitlist (9% in ICD group vs. 15% in no-ICD group; p < 0.0001), whereas 63% underwent HT. Having an ICD at listing was associated with an adjusted 13% relative reduction in mortality (hazard ratio: 0.87; 95% confidence interval: 0.80 to 0.94). In the subgroup of patients with LVAD (n = 9,478), having an ICD was associated with an adjusted 19% relative reduction in mortality (hazard ratio: 0.81; 95% confidence interval: 0.70 to 0.94).
Conclusions ICD use was associated with improved survival on the HT waitlist in patients with or without LVADs. These findings strengthen the current guideline recommendations of using ICDs in nonhospitalized patients awaiting HT and provide new insight into the effectiveness of ICDs on survival in LVAD-supported patients.
- heart transplantation
- implantable cardioverter-defibrillator
- left ventricular assist device
Over the last decade, the number of adult patients active on the heart transplantation (HT) waitlist has increased by ∼25%. Waiting times for all HT listing categories have also steadily increased during this period (1). In addition to a shortage of donor organs, another reason for the increase in census and waiting times is the reduction in waitlist mortality (1–3), perhaps because of improvements in heart failure care including use of left ventricular assist devices (LVADs) (4,5). Despite this improvement in mortality over the years, the number of deaths on the waitlist remains high. In the United States, of the 2,784 patients who were removed from the waitlist in 2012, 73% underwent HT while 18% either died or became ineligible for HT because of clinical deterioration (1). As such, continued efforts to improve outcomes on the HT waitlist are needed.
Implantable cardioverter-defibrillators (ICDs), used for primary prevention of sudden cardiac death (SCD), reduce mortality in patients with left ventricular ejection fraction ≤35% and New York Heart Association (NYHA) functional class I to III heart failure (6,7). Patients with end-stage or NYHA functional class IV heart failure, on the other hand, have increased competing risk of nonsudden death. Thus, ICDs are not routinely recommended in these patients unless they are awaiting advanced treatments such as HT or LVAD implantation. Given that such patients were not included in primary prevention ICD trials, the data on the utility of ICDs in such patients are limited to a few small studies (8–11). Based on these limited data backed by expert consensus, the current practice guidelines recommend ICD use in nonhospitalized patients awaiting HT (12).
Furthermore, recent publications have raised the question of whether ICDs prolong survival in patients with an LVAD. Although ventricular arrhythmias are associated with poor outcomes in patients with an LVAD, the data on ICDs benefiting these patients have been conflicting (13–18). This study was designed to assess the impact of ICDs on HT waitlist mortality using data from the largest HT cohort in the United States. We hypothesized that ICDs would improve waitlist mortality, but this effect would be limited to patients without an LVAD.
Adults (age ≥18 years) who were listed for initial HT in the United States and reported in the Organ Procurement and Transplantation Network (OPTN)/United Network for Organ Sharing (UNOS) registry were retrospectively identified. This registry consists of data on pre- and post-transplant variables captured and entered by each transplant center into an online database at the time of listing, at the time of HT, and during post-transplant follow-up. For this study, “Standard Transplant Analysis and Research” files with de-identified patient data for all HT listings were obtained from the OPTN/UNOS as previously described (19,20). Because no patient or center identifiers were used and the dataset is publicly available from OPTN/UNOS, this study qualified for exempt status from our institutional review board.
A total of 81,139 adult patients who were listed for HT in the United States since 1987 were identified. Because waitlist acuity classification scheme changed significantly in 1999 and listing status was an important covariate for this analysis, subjects listed before 1999 were excluded (n = 36,813). Of the remaining 44,326 patients, those with total artificial heart, prior HT, congenital heart disease, right ventricular or biventricular assist device, and/or restrictive cardiomyopathy were excluded (n = 7,998). Furthermore, because patients without an LVAD who were listed as status 1A and remained in 1A status until HT, delisting, or death (i.e., never changed their 1A listing status) by definition would be in the hospital and thus may not be impacted by the presence or absence of an ICD, this group also was excluded (n = 3,140). Patients who were listed as 1A initially and changed listing status to 1B or II even once during follow-up were included in the analyses. Lastly, 589 patients with missing records on ICD status also were excluded. The final study cohort consisted of 32,599 patients listed for heart-only transplantation between January 1, 1999, and September 30, 2014 (Figure 1).
The primary outcome was time to all-cause mortality while listed for HT. Patients were censored at the time of HT or at the time of delisting because of clinical reasons or loss to follow-up.
The date of the last follow-up was December 5, 2014. Patients were divided into LVAD and no-LVAD subgroups. Patients who did not have an LVAD at listing but subsequently had an LVAD implanted while on the waitlist were included in the LVAD group. Follow-up for patients in the no-LVAD group started on the date of HT listing. However, follow-up for patients in the LVAD group started either on the date of listing or on the date of LVAD implantation while on the waitlist, whichever was later.
Continuous variables are summarized as mean ± SD or median (25th, 75th percentile), and categorical variables as frequency and percentage. Mann-Whitney U tests or Student t tests were used to compare continuous variables. Chi-square tests or Fisher exact tests were used for categorical variables. Effect of ICDs on survival was assessed by Kaplan-Meier analysis and log-rank test. All Kaplan-Meier survival curves were truncated to 4 years of follow-up. For patients with an LVAD already implanted before listing, follow-up began at the date of listing. Follow-up time in patients who received an LVAD after being listed began at the date of LVAD implantation. For cases in which the LVAD implant date was missing (n = 604), the date was imputed to be the median time to implant for those with available dates of implant. Adjusted analyses were performed using Cox proportional hazards regression with the following covariates: age, sex, ethnicity (Caucasian/non-Caucasian), serum creatinine, ischemic etiology, initial UNOS listing status, LVAD, and listing era (1999 to 2005 or 2006 to 2014). The specified categories of listing era were selected because in 2005 the Centers for Medicare and Medicaid Services approved ICD implantation for primary prevention of SCD. All comparisons were 2-sided, and p < 0.05 was considered statistically significant. To test for multiplicative interaction between subgroups and ICD in their association with the primary endpoint (all-cause mortality), separate Cox regression models with fixed effects for subgroup, ICD, and subgroup by ICD interaction were fitted. The models were adjusted for potential confounders, and tests for interaction were deemed statistically significant when p < 0.05. Analyses were performed using Stata statistical software version 13 (Stata Corp., College Station, Texas).
The mean age of the 32,599 patients included in our analyses was 53 ± 12 years; 77% were male and 70% were Caucasian (Table 1). The indication for HT was ischemic cardiomyopathy in 43% of the patients. A total of 9,478 patients (29%) had an LVAD, either at listing (14%) or after being listed (15%). Furthermore, 13%, 40%, and 47% of patients were listed as UNOS status 1A, 1B, and 2, respectively. Median follow-up was 154 days (interquartile range [IQR]: 50 to 416 days) for the entire cohort and median time on the list for patients who underwent HT was 153 days (IQR: 49 to 415 days).
At the time of HT listing, 21,498 patients (66%) had an ICD. Patients with an ICD were more likely to be older, male, and diabetic compared to those without an ICD (Table 1). On the other hand, ICD patients were less likely to be on life support (intra-aortic balloon pump/extracorporeal membrane oxygenation) or listed as status 1A (Table 1). Median follow-up time was 153 days (IQR: 50 to 397 days) for the ICD group and 156 days (IQR: 49 to 468 days) for the no-ICD group.
Trends in ICD and LVAD implantation between 1999 and 2014
The prevalence of ICD use in patients listed for HT increased steeply from 1999 to 2006 before reaching a plateau at ∼80% in the subsequent years. Similarly, the prevalence of LVAD use among patients listed for HT nearly doubled between the years of 2003 and 2009 before reaching a plateau at ∼45% in the subsequent years (Figure 2).
Mortality in relation to ICD
A total of 3,638 patients (11%) died while awaiting HT (9% in the ICD group vs. 15% in the no-ICD group; p < 0.0001), 63% underwent HT, and 18% were delisted. Approximately 17 patients required treatment with an ICD to prevent 1 death on the waitlist. Events on the waitlist in relation to the presence of an ICD are given in Table 2. Cumulative survival to HT was significantly better in patients with an ICD compared to those without an ICD (Figure 3). On multivariable Cox regression analyses, having an ICD was associated with a 13% relative reduction (hazard ratio [HR]: 0.87; 95% confidence interval [CI]: 0.80 to 0.94) in all-cause mortality on the HT waitlist (Table 3).
Subgroup analyses for patients with and those without LVADs
Among the 9,478 patients with an LVAD, 6,529 (69%) had an ICD. A total of 908 patients (10%) with an LVAD died while awaiting HT (8% in the ICD group vs. 14% in the no-ICD group; p < 0.0001), and 65% underwent HT. Cumulative survival to HT was significantly better in LVAD patients with an ICD compared to those without an ICD (Figure 4). On multivariable Cox regression analyses, having an ICD was associated with a 19% relative reduction (HR: 0.81; 95% CI: 0.70 to 0.94) in all-cause mortality on the HT waitlist in patients with an LVAD (Table 3, model 2). Adjusting for timing of LVAD implantation (before or after listing for HT) did not significantly alter these results.
Among the 23,121 patients without an LVAD, 14,969 (65%) had an ICD. A total of 2,730 patients (12%) without an LVAD died while awaiting HT. Even in this group, cumulative survival to HT was significantly better in patients with an ICD compared to those without an ICD (Figure 5). On multivariable Cox regression analyses, having an ICD was associated with a 13% relative reduction (HR: 0.87; 95% CI: 0.80 to 0.96) in all-cause mortality on the HT waitlist in patients without an LVAD (Table 3).
The results from this large nationwide registry for HT listings between 1999 and 2014 in the United States showed that having an ICD at listing was associated with a 13% reduction in waitlist mortality. This effect was even larger at 19% in patients with an LVAD. Cumulatively, only two-thirds of the patients listed for HT since 1999 had an ICD. Although the ICD utilization rate has significantly increased after 2005, nearly 20% of the non-hospitalized patients (at least at some point while on the waitlist as defined by our inclusion criteria) did not have an ICD even in the most recent era.
The decision to pursue HT for heart failure is typically the culmination of a long journey of conventional medical therapy, usually involving dozens of health care providers. It is surprising that many “ICD-eligible” patients have achieved HT listing without receiving an ICD somewhere along that journey. Several potential explanations merit discussion. First is the patient’s preference to forgo implant. Based on the American Heart Association’s Get with the Guidelines database, <40% of the 13,034 patients meeting ICD eligibility received an ICD (21). Common reasons for patient election to forgo an ICD include lack of insight, perception of strength of physician recommendation, concerns with device malfunction/recall, and implant risk (22). Of note, the percentage of patients with an ICD was much higher in this UNOS cohort compared to the Get with the Guidelines report. Second is the physician’s perception that time to HT will be so short that there may not be enough time to realize the mortality benefit associated with ICD implantation. The median duration to HT in this cohort was ∼150 days. However, in large clinical trials of ICD for primary prevention, survival benefit from ICD became evident ∼12 to 18 months after enrollment, which may have been a reason for HT physicians to defer implant. Third, a patient’s critically ill status may have tilted the decision away from ICD implantation. Indeed, in the present study, twice as many patients without an ICD were supported with an intra-aortic balloon pump. Finally, ICD implantation in NYHA functional class IV patients is discouraged because: 1) this population has been excluded from major ICD trials; and 2) pump failure is the predominant mode of death in these patients, which could limit the benefit from ICD use. Although we do not have information on NYHA functional class for the cohort studied, it is reasonable to assume that this cohort of HT candidates would be highly enriched with NYHA functional class IV patients.
The increase in ICD utilization over the years as noted followed the publication of several landmark clinical trials such as MADIT-II (Multicenter Automatic Defibrillator Implantation Trial II) and SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) and the approval of ICD implantation for primary prevention of SCD by the Centers for Medicare and Medicaid Services in 2005 (6,7). Similarly, the increase in LVAD utilization paralleled 2 eras (REMATCH [Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure] published in 2001, followed by the HeartMate II bridge-to-transplant trial published in 2007) (4,23). Reasons for the plateau in ICD utilization at ∼80% after 2006 are not immediately clear, although one could surmise that the combination of the factors discussed posed an “upper boundary” on the population of patients listed for HT in whom ICD implantation is thought to be suitable. In light of the significant mortality benefit suggested in this study, it will be important for future efforts to determine the reasons why ICD penetration in the population of patients listed for HT remains at ∼80%.
In this analysis, ICD use was associated with a 19% relative reduction and a 6% absolute risk reduction in mortality in patients with an LVAD. Earlier reports have shown that ventricular arrhythmias are associated with poor outcomes in LVAD patients (24). Given the hemodynamic support provided by LVADs, SCD is an uncommon mode of death in these patients (25–28). As such, the utility of an ICD in this population has been a matter of controversy. In a recent meta-analysis of 6 observational studies, our group showed that ICD use was associated with improved survival in patients with an LVAD (29). The current findings expand our understanding of utility of ICDs in this population and validate our preliminary findings in a much larger sample size. Ventricular arrhythmias can significantly worsen right ventricular function and cause renal dysfunction, both of which are associated with poor outcomes in LVAD patients. Although the mechanism of survival benefit from an ICD in LVAD patients remains unclear, one could speculate that the beneficial effects may have been related to reduction in deleterious effects of ventricular arrhythmias on right ventricular function, thereby leading to a reduction in heart failure deaths as opposed to SCD. Further, one could also not exclude the possibility of a selection bias of not implanting ICDs in relatively sicker LVAD patients. Of note, in all patients with an LVAD in this analysis the LVAD was used as bridge to transplantation. As such, the effect of ICD use from this analysis cannot be generalized to patients with an LVAD as destination therapy.
The data were observational and retrospective and may have been subject to selection bias and institutional practice patterns. Information on left ventricular ejection fraction was unavailable and could not be adjusted for. However, patients with restrictive cardiomyopathies, who account for ∼5% of the patients listed for HT, were excluded. As such it would be reasonable to assume that the patients included in this analysis would meet traditional guideline-based indications for ICD implantation. Data on primary or secondary indication for ICD use were not available for all patients in the UNOS registry. Lastly, information on ICD implantation after listing was unknown. However, given the survival benefit with ICD implantation as noted in our analyses, this misclassification would have only reduced the differences in mortality between the ICD and no-ICD groups.
Among patients listed for HT in the United States, having an ICD was associated with improved survival to HT. This included the subgroup of patients with LVADs. These findings support and strengthen the current guideline recommendations of using ICDs in non-hospitalized patients awaiting HT and provide new insight into the effectiveness of ICDs on survival in LVAD-supported patients.
COMPETENCY IN MEDICAL KNOWLEDGE: Results from this study are the largest to date highlighting the importance of ICD use in patients listed for and awaiting HT. Most patients listed for HT have advanced heart failure and are not deemed to be ideal candidates for ICD implantation given their proportionately higher risk for nonsudden death. However, the findings from this study suggest that the use of ICDs in such end-stage heart failure patients may be beneficial if they are awaiting definitive therapeutic treatment options such as HT. Furthermore, these findings support the use of ICDs in patients with an LVAD. Having an LVAD provides significant hemodynamic support against ventricular arrhythmias, and as such one would expect ICDs to be ineffective in such a population. The findings from this study significantly advance our understanding of the utility of ICDs in the LVAD population, although these results would be limited to patients receiving LVADs as a bridge to transplantation rather than destination therapy.
TRANSLATIONAL OUTLOOK: There is currently a lack of randomized controlled clinical trial data addressing the effectiveness of ICD use in patients with and those without an LVAD who are awaiting HT. In the absence of strong data supporting such a practice, most guideline recommendations are based on either small observational studies or expert consensus. As such, the findings from this large registry-based analysis strongly calls for a randomized controlled clinical trial to assess the effectiveness of ICD use in patients with the “newer” current-generation LVADs who are awaiting HT so that survival on the HT waitlist can be further improved.
This work was supported in part by Health Resources and Services Administration contract 234-2005-37011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. Dr. Vakil has received an investigator-initiated research grant from Medtronic for this study. Dr. Eckman is a consultant/advisor for Medtronic and St. Jude Medical. Dr. Levy has served on the steering committee/clinical endpoint committee for GE Healthcare Steering Committee, ADMIRE ICD, Novartis CEC-RELAX-ASIA, St. Jude Medical CEC, CardioMems, and CHAMPION Post Approval Study; has served as a consultant for Novartis and Abbott; and participated in research-clinical trials for Resmed, Amgen, Novartis, Respircardia, Bayer, and Merck. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- confidence interval
- hazard ratio
- heart transplantation
- implantable cardioverter-defibrillator
- interquartile range
- left ventricular assist device
- New York Heart Association
- Organ Procurement and Transplantation Network
- sudden cardiac death
- United Network for Organ Sharing
- Received April 12, 2016.
- Revision received June 27, 2016.
- Accepted July 28, 2016.
- 2017 American College of Cardiology Foundation
- ↵UNOS/OPTN Annual Heart Transplantation Report 2012. Available at: http://srtr.transplant.hrsa.gov/annual_reports/2012/pdf/05_heart_13.pdf. Accessed December 7, 2015.
- Lietz K.,
- Miller L.W.
- Singh T.P.,
- Almond C.S.,
- Taylor D.O.,
- Graham D.A.
- Tracy C.M.,
- Epstein A.E.,
- Darbar D.,
- et al.
- Enriquez A.D.,
- Calenda B.,
- Miller M.A.,
- Anyanwu A.C.,
- Pinney S.P.
- Garan A.R.,
- Yuzefpolskaya M.,
- Colombo P.C.,
- et al.
- Lee W.,
- Tay A.,
- Subbiah R.N.,
- et al.
- Vakil K.,
- Duval S.,
- Sharma A.,
- et al.
- Oz M.C.,
- Rose E.A.,
- Slater J.,
- Kuiper J.J.,
- Catanese K.A.,
- Levin H.R.
- Vakil K.,
- Kazmirczak F.,
- Sathnur N.,
- et al.