Author + information
- Received September 13, 2015
- Revision received November 2, 2015
- Accepted November 11, 2015
- Published online February 2, 2016.
- aCardiology Division, Section of Cardiac Electrophysiology, Department of Medicine, Emory University School of Medicine, Atlanta, Georgia
- bAtlanta Veterans Administration Medical Center, Department of Veterans Affairs, Atlanta, Georgia
- cEmory Cardiovascular Clinical Research Institute, Emory University, Atlanta, Georgia
- ↵∗Reprint requests and correspondence:
Dr. Faisal M. Merchant, Emory University Hospital Midtown, 550 Peachtree Street, MOT 6th floor, Atlanta, Georgia 30308.
Although implantable cardioverter-defibrillators (ICDs) are frequently viewed as a lifelong commitment in that patients are routinely scheduled for generator exchange (GE) at end of battery life, several considerations should prompt a reevaluation of risks and benefits before GE. Compared with initial ICD implant, patients receiving replacement devices are older, and have more comorbidities and shorter life expectancy, all of which may limit the benefit of ICD therapy following GE. Additionally, GE is associated with significant complications, including infection, which may increase the risk of mortality. In this paper, we review recent data regarding opportunities for risk stratification before GE, with a particular focus on those with improved left ventricular function and those who have not experienced ICD therapies during the first battery life. We also provide a broader perspective on ICD therapy, focusing on how decisions regarding GE may affect goals of care at the end of life.
- cardiac arrhythmias
- generator exchange
- heart failure
- risk assessment
- sudden cardiac death
- ventricular function
Implantable cardioverter-defibrillator (ICD) therapy is associated with significant reductions in all-cause mortality among appropriately selected patients at heightened risk of sudden cardiac death (SCD) resulting from ventricular arrhythmia (VA). The decision to implant an ICD is complex, taking into account the risk of SCD/VA, along with noncardiac comorbidities and overall life expectancy. However, several large, randomized clinical trials have been performed to assess the efficacy of ICD implantation in both primary and secondary prevention of SCD/VA. These studies serve as the foundation for the American College of Cardiology Foundation/American Heart Association/Heart Rhythm Society guidelines on ICD implantation (1) and, in conjunction with large registry studies assessing the safety and risks associated with ICD implantation (2,3), provide the basis for patients and providers to have an informed discussion about the benefits and risks of ICD implantation.
More than 100,000 ICDs are implanted annually in the United States, of which approximately three-quarters are new device implants and about one-quarter are generator exchanges (GEs) for end of battery life (4). Whereas a robust body of literature exists to support informed decision making at the time of initial ICD implant, there is a relative paucity of data to support decision making at the end of battery life. ICD therapy is frequently viewed as a lifelong commitment in that patients are scheduled for GE as a matter of course at the end of battery life. However, several important considerations should prompt a reevaluation of risks and benefits to ongoing ICD therapy before GE. First, compared with patients undergoing initial ICD implantation, those receiving replacement devices are older, have more comorbidities, and have shorter life expectancy (5,6). This raises the possibility that, as competing risks of nonarrhythmic death accrue, the potential benefit of ICD therapy may be diminished among those undergoing GE compared with those undergoing initial implant. Additionally, although GE is generally considered a relatively straightforward procedure, elective ICD GE is associated with a major complication rate of approximately 4% (7), and the occurrence of major complications in this setting may be associated with an increased risk for mortality (8). In light of these considerations, the risk/benefit ratio of elective GE may be very different than that at the time of initial ICD implant. However, there is a paucity of data on outcomes and benefits to ongoing ICD therapy after GE, which significantly limits the ability of patients and providers to have an informed discussion.
Opportunities for Risk (Re)-Stratification at the Time of GE
Improved versus persistently impaired left ventricular systolic function
In this country, most ICDs are implanted for primary prevention, that is, in patients with impaired left ventricular ejection fraction (LVEF), but without a history of documented SCD/VA. Therefore, the mean LVEF at the time of initial ICD implant (n = 359,993) for all devices implanted between 2005 and 2010 in the National Cardiovascular Data Registry was 27.7 ± 10.8% (5). In contrast, among patients in the same registry undergoing ICD GE (n = 103,985), mean LVEF was significantly higher, at 32.6 ± 13.7%. This finding highlights that some patients who undergo initial ICD implant for impaired LVEF (i.e., ≤35%) (1) may have improvement in ventricular function between the time of initial implant and GE. There is a well-established relationship between lower LVEF and higher risk of SCD/VA, and the seminal trials establishing the efficacy of primary prevention ICD therapy (9,10) were designed, in part, on the basis of this relationship. Therefore, it is conceivable that an improvement in LVEF between initial implant and GE may alter the risk of SCD/VA, such that the risk of arrhythmic death is no longer sufficiently high to warrant ongoing ICD therapy.
Several studies have looked at outcomes after GE as a function of improvement in LVEF. In a recent study in a Veterans Affairs cohort of 231 patients undergoing GE who were initially implanted for primary prevention, ongoing ICD therapy was considered no longer indicated at the time of GE in 59 patients (26%) on the basis of LVEF improvement to ≥40% and never having received appropriate ICD therapy during the first battery life (i.e., an “uneventful” first battery life) (11). Mean LVEF at the time of GE among those in whom ICD therapy was considered no longer indicated was 49 ± 9% versus 25 ± 11% among those with persistently impaired LV function. Importantly, all patients in this cohort underwent GE. During a mean follow-up of 3.5 ± 2.0 years after GE, the incidence of appropriate ICD therapy among those in whom ICD therapy was considered no longer indicated was 2.8%/person-year, compared with 10.7%/person-year in those in whom ongoing ICD therapy was considered indicated (p < 0.001). These findings highlight the significantly lower risk of SCD/VA among those with improved LVEF and uneventful first battery life.
Similar findings were recently demonstrated in a follow-up study from the MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial With Cardiac Resynchronization Therapy) trial, in which patients who were randomized in the initial study to CRT-defibrillator (CRT-D) and had paired echocardiograms at baseline and at 12 months (n = 752) were evaluated to assess the impact of improvement in LV function on subsequent ICD therapies (12). All patients had LVEF ≤30% at the time of initial CRT-D implant and ICDs were implanted for primary prevention. At the 12-month follow-up, patients were categorized into 3 groups: LVEF ≤35%; LVEF 36% to 50%; and LVEF >50% (“normalized” LVEF group). During a mean follow-up of 2.2 ± 0.8 years after the initial CRT-D implant, the primary endpoint of appropriate ICD therapy for VA ≥200 beats/min occurred in only 1 of 55 patients (2%) with normalization of LVEF; this event was treated without need for ICD shock. There were no appropriate ICD shocks among patients with normalized LVEF. In contrast, the incidence of VA ≥200 beats/min was 7% among those with LVEF 36% to 50% (n = 594) and 18% among those with LVEF ≤35% (n = 103), supporting the notion of an inverse relationship between LV function and risk of SCD/VA. Two important aspects of the MADIT-CRT substudy should be noted. First, this cohort addressed improvement in LVEF and reduction in ICD therapies during the first battery life and did not specifically address prognosis after GE. Second, this study looked only at patients who had recovery of LV function with CRT, which may be mechanistically different than spontaneous improvement in LVEF with medical therapy, as in the study by Kini et al. (11).
In a recently published multicenter, prospective cohort of 1,189 patients implanted with primary prevention ICDs, Zhang et al. (13) reported on the outcomes of 538 patients (45%) who underwent paired assessments of LVEF at baseline and during follow-up (mean time between first and last LVEF measurement of 4.9 years). Among the 538 patients, 215 (40%) had improvement in LVEF (defined as an absolute increase >5%). Compared with those with unchanged LVEF, the hazard ratio for all-cause mortality among those with improved LVEF was 0.33 (95% CI: 0.18 to 0.59) and for appropriate ICD shocks was 0.29 (95% CI: 0.11 to 0.78). There was a consistent relationship noted between improving LVEF and decreased risk of all-cause mortality and appropriate ICD shocks. Only 4 of 215 patients (1.9%) with improved LVEF experienced an appropriate ICD shock. In aggregate, all 3 of these studies are consistent in demonstrating a significantly lower risk of ICD therapies among those with improvement in LVEF after initial implant and reinforce the importance of LVEF in assessing future risk of ICD therapies in those undergoing GE.
A few smaller studies have demonstrated contradictory findings with regard to prognosis after GE among those with recovered LVEF. In another Veterans Affairs cohort, Naksuk et al. analyzed 97 consecutive patients undergoing GE following initial implantation for primary prevention (14). From this group, 25 patients (27%) had improved LVEF to >35% and at least a 10% increase since the initial implant. Mean LVEF among those in the improved LVEF group was 49 ± 8% versus 36 ± 13% among those deemed to have unchanged LVEF (p < 0.001). During a mean follow-up of 6.2 ± 2.2 years after GE, 9 patients (36%) with improved LVEF had appropriate ICD therapies, compared with 19 (29%) among those with unchanged LVEF (p = 0.51). On the basis of these findings, the authors conclude that patients with improved LVEF continue to exhibit a high rate of appropriate ICD therapies after GE and that GE is, therefore, likely indicated, even if LVEF has improved. However, it is important to note that of the 25 patients in the cohort with improved LVEF, 6 had appropriate ICD therapies during the first battery life and therefore would likely be considered to have an ongoing ICD indication. Only 3 patients in the improved LVEF group had appropriate ICD therapies after GE without having experienced appropriate ICD therapy during the first battery life. Additionally, most patients in this study had multizone ICD programming with a lowest rate cutoff for which a shock could be delivered of 171 ± 14 beats/min in the improved LVEF group and 165 ± 16 beats/min in the unchanged LVEF cohort. Although these programming practices may have been broadly consistent with the time period during which these patients underwent GE (2006 to 2010), more contemporary data suggest that programming primary prevention ICDs as a single zone with high-rate cutoffs both reduces unnecessary ICD shocks and improves clinical outcomes (15,16). It is well-acknowledged that “appropriate” ICD shocks overestimate the true incidence of aborted SCD in defibrillator recipients (17), and the use of multizone programming with relatively low rate cutoffs in the study by Naksuk et al. may have led to inclusion of ICD shocks that were “appropriate,” but potentially unnecessary and not life-saving. Last, in this study’s unchanged LVEF cohort, the mean LVEF at the time of GE was 36 ± 13%, suggesting that a significant number of patients who were considered to have unchanged LVEF had some improvement from the time of initial implant, even if LVEF had not increased by >10% compared with baseline (14). Therefore, the lack of difference in ICD therapy event rates between the 2 groups in this study may have been a function of both groups exhibiting some improvement in LVEF, albeit to a lesser extent among those considered to have unchanged ejection fraction. A more clear control group might have included those with LVEF persistently ≤35% at the time of GE.
In a follow-up analysis from DEFINITE (Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation), Schliamser et al. (18) identified 187 patients from the overall study cohort of 449 (42%) who had a follow-up assessment of LVEF between 90 and 730 days after enrollment (median 458 days). Ninety-six patients had LVEF improvement >5% from baseline, 79 had stable LVEF (≤5% change from baseline), and 12 had a decrease in LVEF >5% from baseline. Those with improved LVEF had significantly better overall survival than those with stable or decreased LV function, and the combined endpoint of appropriate ICD shocks, SCD, and resuscitated cardiac arrests also occurred significantly less frequently in the group with improved LVEF. However, in the subgroup randomized to ICD implant, although the incidence of appropriate ICD shocks was lower among those with improved LVEF, the difference was not significant compared with the LVEF stable or decreased groups. Although the lack of a significant difference in incidence of ICD shocks among those with improved LVEF may have been due to limited power in the subgroup with ICDs, the investigators concluded that the presence of an ongoing risk for ICD shocks should prompt routine GE, even among those with improved LVEF.
Two additional factors should be noted with regard to the relevance of changes in LVEF and subsequent risk of VAs. First, it is well-acknowledged that fibrosis, electrical uncoupling, and other ultrastructural changes play a role in the substrate that gives rise to VAs in cardiomyopathy (19), and that this substrate may persist, even among those with improvement in LVEF, giving rise to a persistent arrhythmic risk. However, the extent of this ongoing risk and the way it should be incorporated into decisions about GE among those with improved LVEF remain unclear. Second, there is significant interobserver variability with regard to LVEF assessment, particularly when echocardiography serves as the diagnostic modality. Among patients enrolled in PROSPECT (Predictors of Response to CRT Trial) (20), more than 20% were found to have an LVEF >35% when studies were rereviewed in a core laboratory, suggesting that real-world variability in LVEF assessment may confound the assessment of LV function improvement in gauging arrhythmic risk before GE.
Presence of appropriate ICD therapies during first battery life
In addition to the status of LV function at the time of GE (persistently impaired vs. improved), another marker that has been evaluated for risk-stratifying patients at the end of ICD battery life is the presence or absence of appropriate ICD therapies during the first battery life. In a national cohort of ICD patients enrolled in a remote monitoring system (Latitude, Boston Scientific Corp., Boston, Massachusetts), 24,203 patients underwent GE for end of first battery life, of whom two-thirds (n = 16,230) did not receive ICD shocks before GE (uneventful first battery life), whereas 7,973 experienced at least 1 shock with the first generator (21). At 3 years following GE, the cumulative incidence of ICD shocks for VA with detection heart rate >200 beats/min was significantly lower among those with an uneventful first battery life (9.2% vs. 24.3%, p < 0.001) (Figure 1). Of note, given the large number of patients included in this analysis, ICD shocks were not adjudicated for being appropriate versus inappropriate. The multicenter European INSURE (Incidence Free Survival after ICD Replacement) registry followed 510 patients after first ICD GE, of whom 265 (52%) had not received appropriate ICD therapy with the first battery life. At 3 years following GE, the incidence of appropriate ICD therapies was 21.4% among those with uneventful first battery life compared with an incidence of 48.1% among those who did receive appropriate therapy before GE. Notably, more than 85% of patients in the INSURE registry were initially implanted with ICDs for secondary prevention (22), which may, in part, account for the higher 3-year event rate in this study compared with the analysis by Merchant et al. (21).
In a single-center study of 154 primary prevention ICD recipients undergoing GE for end of battery life, Van Welsenes et al. (23) reported that 114 (74%) had not received appropriate ICD therapy during the first generator. In this group, a first appropriate ICD therapy occurred in 14% of patients after GE following an uneventful first battery life. In a similar 2-center study of 403 primary prevention ICD recipients undergoing GE, 68% did not receive appropriate ICD therapy during the first generator (24); among these, the 3-year incidence of appropriate therapy (shocks and antitachycardia pacing) after GE was 13.7%.
These studies have all been generally consistent in demonstrating that approximately two-thirds to three-quarters of primary prevention ICD recipients can be expected to reach end of first battery life without receiving appropriate ICD therapies. And although the likelihood of ICD therapies after GE is lower among those with an uneventful first battery life, the incidence of appropriate therapy after GE in this cohort still appears to be on the order of 10% to 15% at 3 years (∼5%/year). These findings are consistent with the time dependence of appropriate ICD therapies in primary prevention recipients reported by Alsheikh-Ali et al. (25), demonstrating the highest rate of therapy within the first 2 years after ICD implant (12% to 20%/year), followed by a lower, but persistent incidence of ICD therapies from years 3 to 7 after implant (6% to 11%/year).
A significant percentage of ICD recipients do not survive to reach the end of first battery life. Attrition of the sickest patients with ICDs may conceivably lead to lower than expected shock rates with second devices among those who survive to GE, and may also account for the apparently higher LVEF among those receiving replacement devices compared with initial implants (5). The extent to which attrition accounts for these observations is unclear, but this remains an important consideration when assessing outcomes after GE.
Competing risks of arrhythmic versus nonarrhythmic death following GE
The previously mentioned studies demonstrate that the likelihood of receiving appropriate ICD therapy after GE appears to be lower among those with improvement in LV function and in those without appropriate ICD therapies during the first generator. However, even among these patients, there continues to be a measureable rate of appropriate ICD therapies after GE. This rate is higher than the SCD/VA risk among the general population, which has been estimated at approximately 0.1%/year for men 50 years of age, increasing to ∼0.8%/year for men 75 years of age (26). The incidence of SCD among women in the general population is generally lower than among men (26). That the incidence of appropriate ICD therapies after GE among those with improved LVEF and an uneventful first battery life still appears to be higher than the risk of SCD in the general population has been used by some to suggest that the residual arrhythmic risk in this population may justify routine GE among all patients implanted with ICDs (14,18,27).
However, the notion that SCD/VA risk among those with improved LVEF and uneventful first battery life is not zero or does not reach the low levels observed in the overall population does not necessarily imply that ongoing primary prevention ICD therapy is justified or that GE should be performed routinely in these patients. The benefit of ICD therapy likely does not depend specifically on the absolute risk of SCD/VA, but, more importantly, on the relative balance between competing risks of arrhythmic and nonarrhythmic death. Although age should not be the sole criterion for decisions about performing a GE, it remains an important and clinically useful marker of nonarrhythmic mortality risk. An analysis of 6,252 patients pooled from 13 clinical trials demonstrated that, although the absolute risk of SCD increased per decade of life from <50 to >80 years of age, the risk of nonarrhythmic death increased to an even greater degree, such that the ratio of SCD to all-cause mortality decreased with increasing age (Figure 2) (28). Given that patients undergoing GE are significantly older than those receiving new ICDs and have a significantly higher prevalence of comorbidities (5), it is plausible that the risk of nonarrhythmic death may be higher after GE than after initial ICD implant. Additionally, among those undergoing GE who have improved LVEF and have had an uneventful first battery life, the risk of SCD or appropriate ICD therapies is likely lower than after the initial ICD implant (25). Therefore, the relative balance between the risks of arrhythmic and nonarrhythmic death may be sufficiently altered after GE such that ongoing ICD therapy may no longer be beneficial. An analysis from the National Cardiovascular Data Registry provides some evidence to support this notion by demonstrating that all-cause mortality at 1 and 3 years after GE is significantly higher than at the same time points following initial ICD implant: 9.9% versus 9.4% at 1 year and 27.4% versus 23.5% at 3 years (6). This difference in survival remained significant, even after propensity-score matching of patients after GE and after initial ICD implant.
Although there is a paucity of data on predictors of mode of death after GE, several attempts have been made to predict the risk of specific modes of death after initial ICD implant. The MADIT-II investigators developed a risk model to predict all-cause mortality at 2 years after ICD implant and identified 5 significant markers of total mortality: New York Heart Association functional class >II; age >70 years; blood urea nitrogen >26 mg/dl; QRS duration >120 ms; and history of atrial fibrillation (AF) (29). In this analysis, compared with patients in the control arm of the MADIT-II study, ICD therapy was associated with a significant reduction in all-cause mortality among those with either 1 or 2 of the 5 risk markers, whereas no significant benefit of ICD therapy was seen among those with ≥3 risk markers, presumably because the competing risk of nonarrhythmic death was too high in this subgroup, thus negating any potential benefit. In a similar substudy from the MUSTT (Multicenter Unsustained Tachycardia Trial), the investigators identified predictors of total and arrhythmic mortality and found that both increasing age and history of AF were significant predictors of all-cause mortality, but not of arrhythmic death (30).
In aggregate, these data suggest that due in large part to older age and a higher prevalence of comorbidities, including AF, the risk of nonarrhythmic death likely increases to a greater extent than the risk of arrhythmic death following GE. It remains to be determined whether the ratio of arrhythmic to nonarrhythmic death after GE among those with improved LVEF and uneventful first battery life is sufficiently low that ongoing primary prevention ICD therapy is no longer beneficial. However, this type of analysis, assessing both arrhythmic and nonarrhythmic risks of death, is likely to provide a more clinically relevant determination of whether GE is beneficial rather than focusing solely on the absolute risk of SCD/VA.
More broadly, for many patients with ICDs, both cardiac (i.e., heart failure burden) and noncardiac comorbidities may progress during the time between initial ICD implant and GE. In a similar vein, with advancing age, the development of frailty, defined as loss of physiological reserve and vulnerability to stressors that may span multiple physical, cognitive, and functional domains, may have important implications for the decision to perform a GE (31). Although the guidelines suggest that candidates for ICD implantation and, by extension, for GE, should have a “reasonable expectation of survival with an acceptable functional status for at least 1 year” (1), making such determinations for individuals with heart failure and other significant comorbidities can be challenging. Numerous risk models exist for predicting mortality in patients with advanced heart failure; however, most have modest accuracy, and the specific trajectory for an individual patient is often unclear (32). Even in the absence of high-quality tools to assess prognosis and functional status, at least a qualitative assessment of progression of cardiac and noncardiac comorbidities and a discussion of how they may have affected patient preferences regarding ongoing ICD therapy, should be undertaken before GE.
Risks Associated With GE
ICD GE is often viewed as a straightforward procedure, yet published reports suggest that it is associated with important complications, most notably infection. The multicenter REPLACE (Implantable Cardiac Pulse Generator Replacement) registry (7) had a 4.9% major complication rate associated with ICD GE, with an even higher event rate noted among those with CRT-defibrillators. The rate of device infection associated with GE in the REPLACE registry was 1.4% for ICDs and 2.3% for CRT-D generators (33). The risk of device infection associated with GE is roughly 2-fold higher than after initial ICD implant (34), and the occurrence of major complications associated with ICD generator replacement has been associated with a nearly 10-fold increased risk of all-cause mortality at 45 days post-procedure compared with those without major complications (8).
These data reinforce the notion that ICD GE is associated with a significant rate of major complications that may increase mortality post-procedure. Therefore, to minimize exposure to GE-related complications, the impetus to identify patients who are the most or least likely to benefit from GE should be even stronger.
Other Considerations on GE
Although prevention of SCD/VA is usually the predominant consideration in making a decision to initially implant an ICD and to perform a GE, other factors may also be taken into consideration when making decisions at the end of battery life. For instance, among patients with atrial leads, device-detected subclinical AF is associated with an increased risk of stroke or systemic embolism (35) and may prompt anticoagulation therapy among those without clinically evident AF. The ability to monitor for incidence and burden of atrial arrhythmias may represent a benefit to GE unrelated to SCD/VA. Similarly, some devices are also capable of providing adjunctive data for management of heart failure, such as thoracic impedance monitoring, heart rate derivatives, and hemodynamic pressure sensors (36), which may be beneficial in certain patient subgroups. Last, an important consideration among CRT-D recipients who have improvement in LVEF is the potential for replacement of the generator with a CRT-pacemaker (CRT-P) (12). Replacement with a CRT-P generator may also be a good option for patients in whom ICD therapy is no longer consistent with goals of care, but who may still receive symptomatic benefit from CRT. In the COMPANION (Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure) trial, compared with pharmacological therapy alone, CRT-P reduced the incidence of the primary endpoint (death or all-cause hospitalization) by 34%, and CRT-D reduced the incidence by 40%, suggesting that CRT provides robust benefit, even in the absence of defibrillator therapy (37).
Perspectives on ICD Therapy at End of Life
In many cases, ICD therapy is viewed as a lifelong commitment. Guidelines exist on deactivation of ICD therapy and clearly state that discussions about the potential for withdrawal of ICD therapy should begin before device implant and continue to occur at other sentinel events during a patient’s course (e.g., time of ICD shocks, progression of cardiac and noncardiac disease, initiation of an advanced directive) (38). However, in practice, discussions regarding withdrawal of ICD therapy are often not held until patients reach a more critical state in disease progression, often when palliative or hospice-related care is being considered. For many community-dwelling patients with ICDs, the failure to proactively address goals of care with regard to ICD therapy may result in the delivery of painful ICD shocks immediately before death. In a Swedish cohort of explanted ICDs after death, 35% of patients received ICD shocks within the last hour of life, but in only 13% of cases was death felt to be primarily from an arrhythmic cause, whereas death was due to progressive heart failure or noncardiac causes in the vast majority (39), suggesting that the delivery of shocks in this setting was unlikely to change the ultimate outcome. Additionally, more than one-half of patients in this cohort had an advanced directive including a do-not-resuscitate order, but nearly two-thirds of these patients still had tachytherapies enabled within 24 h of death, and nearly one-half received ICD shocks within the final 24 h of life. These data suggest that a significant opportunity exists to more closely align patients’ desires regarding care at the end of life with ICD programming and with the decision to perform GE, with the recognition that, for many patients, the end-of-life period is foreseeable months in advance because of advancing noncardiac conditions, such as such as dementia, progressive lung disease, cancer, or frailty/debility syndromes.
Several studies have suggested that only one-half to two-thirds of ICD recipients are aware that deactivation is an option, but the overwhelming majority believe that patients should be informed of the option to deactivate ICD therapy, ideally before initial implantation or at the time of end of battery life (40,41). Many patients are also unclear about the implications of deactivation; specifically, that this does not necessarily mean shorter lifespan or decreased quality of life. In a survey of 294 ICD recipients, the only multivariate predictor of a favorable attitude toward ICD deactivation was the desire to die “a worthy death,” defined, in part, by the “avoidance of shocks during dying” (40). However, data on the preferences of ICD recipients regarding device deactivation are still relatively limited, and better tools are needed to help providers gauge patient preferences with regard to ICD management. Educational initiatives would also be useful to help providers traverse the conversation about end-of-life goals and to support patients and families, particularly when concerns exist regarding quantity and quality of life. The end of ICD battery life represents an opportune time for a different form of risk stratification, focusing not on reductions in mortality or complications associated with generator exchange, but more broadly on patient goals of care and the potential to avoid painful ICD shocks at the end of life.
Opportunities for Improved Decision Making at End of ICD Battery Life
Whereas a relatively robust body of literature exists to identify which patients are most likely to benefit from primary prevention ICD implantation, there is a notable paucity of data to identify which patients are most likely to benefit from ICD generator exchange. Available data suggest that the profiles of patients undergoing GE are significantly different from those undergoing initial ICD implant and, as such, the relative risk of arrhythmic versus nonarrhythmic death following GE may differ from that following initial ICD implant. There is a critical need for better evidence to support informed decision-making at the time of ICD battery depletion:
• Multicenter, prospective registries are necessary to provide a better assessment of presumed causes of death after GE (arrhythmic vs. nonarrhythmic), along with better estimates of the incidence of ICD shocks immediately before death.
• Ultimately, randomized trials are likely needed to clearly define which patients are most or least likely to benefit from ICD GE. The most obvious groups to begin with would include those with improved LVEF since the time of initial ICD implant and those with uneventful first ICD battery life because both of these markers have consistently been associated with a lower risk of appropriate ICD therapies after GE. Because the potential benefit to ICD therapy depends on a balance between arrhythmic and nonarrhythmic risks of death, it seems likely that only randomized trials will be capable of defining whether that balance continues to favor ICD therapy after GE. However, given the lower rates of appropriate ICD therapies after GE among those with improved LVEF and uneventful first battery life (11), designing adequately powered trials would likely require large numbers of patients and extended duration of follow-up to accrue enough events. Performing such studies is costly and logistically challenging. The use of registry-based clinical trials, leveraging electronic health records and other novel approaches to trial design, may provide alternatives to gathering high-quality clinical data (42).
• Additional data are needed on patient perspectives regarding continued ICD therapy and, in particular, how those perspectives may change over time after initial ICD implant.
• Better tools are needed to help providers communicate the risks and benefits of ongoing ICD therapy to patients and how ICD therapy may affect goals of care at the end of life.
Given the large number of ICDs implanted annually and the high likelihood of patients surviving through their initial ICD generator, these areas of investigation would allow for a more systematic approach and facilitate more robust decision-making by patients and providers at the time of ICD GE (Central Illustration). Until more data are available, it is also important for providers to communicate the uncertainty about decision making at the time of GE to patients and to engage patients in the decision-making process.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- cardiac resynchronization therapy-defibrillator
- cardiac resynchronization therapy-pacemaker
- generator exchange
- implantable cardioverter-defibrillator
- left ventricular
- left ventricular ejection fraction
- sudden cardiac death
- ventricular arrhythmia
- Received September 13, 2015.
- Revision received November 2, 2015.
- Accepted November 11, 2015.
- 2016 American College of Cardiology Foundation
- Tracy C.M.,
- Epstein A.E.,
- Darbar D.,
- et al.
- Dodson J.A.,
- Reynolds M.R.,
- Bao H.,
- et al.
- Haines D.E.,
- Wang Y.,
- Curtis J.
- Kramer D.B.,
- Kennedy K.F.,
- Noseworthy P.A.,
- et al.
- Poole J.E.,
- Gleva M.J.,
- Mela T.,
- et al.,
- REPLACE Registry Investigators
- Krahn A.D.,
- Lee D.S.,
- Birnie D.,
- et al.,
- Ontario ICD Database Investigators
- Kini V.,
- Soufi M.K.,
- Deo R.,
- et al.
- Ruwald M.H.,
- Solomon S.D.,
- Foster E.,
- et al.
- Zhang Y.,
- Guallar E.,
- Blasco-Colmenares E.,
- et al.
- Ellenbogen K.A.,
- Levine J.H.,
- Berger R.D.,
- et al.,
- Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation (DEFINITE) Investigators
- Chung E.S.,
- Leon A.R.,
- Tavazzi L.,
- et al.
- Merchant F.M.,
- Jones P.,
- Wehrenberg S.,
- et al.
- Erkapic D.,
- Sperzel J.,
- Stiller S.,
- et al.,
- INSURE Investigators
- Yap S.C.,
- Schaer B.A.,
- Bhagwandien R.E.,
- et al.
- Deo R.,
- Albert C.M.
- Patton K.K.
- Goldenberg I.,
- Vyas A.K.,
- Hall W.J.,
- et al.,
- MADIT-II Investigators
- Buxton A.E.,
- Lee K.L.,
- Hafley G.E.,
- et al.,
- MUSTT Investigators
- Bell S.P.,
- Orr N.M.,
- Dodson J.A.,
- et al.
- Allen L.A.,
- Stevenson L.W.,
- Grady K.L.,
- et al.,
- American Heart Association Council on Quality of Care and Outcomes Research,
- Council on Cardiovascular Nursing; Council on Clinical Cardiology,
- Council on Cardiovascular Radiology and Intervention,
- Council on Cardiovascular Surgery and Anesthesia
- Klug D.,
- Balde M.,
- Pavin D.,
- et al.,
- PEOPLE Study Group
- Merchant F.M.,
- Dec G.W.,
- Singh J.P.
- Bristow M.R.,
- Saxon L.A.,
- Boehmer J.,
- et al.,
- Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) Investigators
- Kinch Westerdahl A.,
- Sjöblom J.,
- et al.