Battery longevity of implantable cardioverter-defibrillators and cardiac resynchronization therapy defibrillators: technical, clinical and economic aspects. An expert review paper from EHRA

Abstract

In recent years an extension of devices longevity has been obtained for implantable cardioverter-defibrillators (ICDs), including ICDs for cardiac resynchronization therapy (CRT-D) through improved battery chemistry and device technology and this implies important clinical benefits (reduced need for device replacements and associated complications, particularly infections), as well as economic benefits, in line with patient preferences and needs. From a clinical point of view, the availability of this improvement in technology allows to better tune the choice of the device to be implanted, taking into account that the reasons supporting the value of an extended device longevity as a clinical priority may differ according to the clinical setting (purely electrical diseases or left ventricular dysfunction/heart failure, respectively). From an economic point of view, extension of device longevity may have an important impact in reducing long-term costs of device therapy, with substantial daily savings in favour of devices with extended longevity, up to 30%, depending on clinical scenarios. In studies based on projections, an extension of device longevity allowed to calculate that the cost per day of ICDs may be substantially reduced, and this allows to overcome the frequent perception of ICD and CRT-D devices as treatments with unaffordable costs and to overturn the misconception that up-front costs are the only metric with which to value device treatments. In view of its clinical and economic value, device longevity should be a determining factor in device choice by physicians and healthcare commissioners and should be appropriately considered and valued in comparative tenders.

Introduction

Implantable cardioverter-defibrillator (ICD) therapy is effective for both primary and secondary prevention of sudden cardiac death due to ventricular tachycardia or fibrillation while cardiac resynchronization therapy with defibrillation (CRT-D) back-up is effective in improving patient symptoms and outcome in the setting of severe, moderate as well as mild heart failure. The longevity of ICD and CRT-D devices has rarely been discussed as a topic where technology may meet clinical needs in order to improve patient management, reduce device therapy complications, and reduce costs for the healthcare system. The aim of this review is to analyse the clinical and economic value of extending device longevity in the perspective of the patients and the health care system.

The outcome in patients with and without left ventricular dysfunction implanted with an implantable cardioverter-defibrillator and the need for devices with an extended longevity

Since the first relevant cohort studies in ICD patients dating from the mid-eighties, changes in ICD technology and in expanding opportunities for medical treatment, previously considered improbable, have been implemented in daily practice. In parallel to this evolution, indications were expanded from secondary to primary prevention of sudden cardiac death, so that nowadays the majority of ICDs are implanted for primary prevention of sudden cardiac death. The expected survival of patients currently implanted with ICD or CRT-D is of major relevance for assessing the impact of these device therapies at long term.

Whether today’s patients are in a similar overall health status is difficult to determine. In the early phase of ICD therapy, younger patients usually with coronary artery disease (CAD) and more preserved left ventricular ejection fraction (LVEF) were implanted for secondary prevention. The study from Echt et al.,¹ for instance, enrolled 77 patients with mean age of 54 ± 13 years, 62% with CAD, and ejection fraction (EF) above 40% in 38%. In this population, 1-year survival was 90%. Similar patient characteristics were reported in a study from Winkle et al.² published in 1989 (58 ± 12 years, 80% CAD, and EF >40% in 30%). In retrospect, it is noteworthy that the first randomized trials, both on primary prevention with LVEF <35% (MADIT)³ and on secondary prevention independently on LVEF (AVID)⁴ were published almost 10 years later. When compared with the situation in contemporary patients, heart failure medical treatment was far from being optimal (only 40% of patients were on beta-blocker, and 68% on angiotensin-converting enzyme inhibitors in MADIT). In this early phase comorbidities were not properly addressed. Over time, further randomized trials^5–7 and registries^8,9 evaluated the benefit of ICDs in primary prevention; however, mortality in the ICD group was never a key issue. Table 1 summarizes survival according to different points in time. As can be appreciated, there have been no major changes over the decades, in spite of all improvements in medical therapy. We can only speculate that this improvement is counterbalanced by ICD implantation in weaker patients with more severe comorbidities. Unfortunately, as will be outlined further below, reported ICD longevity is lower than patient survival at 5 years.⁹

Few data are available on survival in patients with a CRT-D. However, survival is better when compared with the past, in part due to the inclusion of lower risk patients, in part due to the positive effect of cardiac resynchronization therapy (CRT) on the functional status and on LVEF in the majority of patients. There is, however, a discrepancy between survival rates in randomized studies and in registry data (Table 1).

Data on patients with hypertrophic cardiomyopathy (HCM) are scarce and the number of patients studied rather small. In a nationwide Swedish study¹³ of 342 patients implanted with an ICD, 45 (13.2%) died during a mean follow-up period of 5.4 years. All-cause mortality was higher in HCM patients compared with the age and sex-matched Swedish general population [standardized mortality ratio 3.4, 95% confidence interval (CI) 2.4–4.5; P < 0.001]. Hypertrophic cardiomyopathy was the main cause of death in 76% of the cases, mainly because of progressive heart failure. Similar results were obtained in a study analysing 134 patients with an ICD followed in two tertiary referral clinics.¹⁴ Annualized cardiac mortality rate was 3.4% per year and was associated with New York Heart Association (NYHA) Class III or IV [hazard ratio (HR) 5.2, 95% CI 2.0–14; P = 0.002] and cardiac resynchronization therapy (HR 6.3, 95% CI 2.1–20; P = 0.02). Inappropriate ICD intervention occurred in 21 patients (3.7%/year) and in 20 patients device related complications were documented (3.6%/year). Thus, for HCM patients, ICDs almost eliminate premature arrhythmic death and result in a shift to heart failure as the cause of death in the majority of cases. As a consequence, in order to further reduce mortality in HCM patients improvement in prevention and management of heart failure and comorbidities appears as key factors.

Patients implanted with an ICD for inherited channelopathies or cardiomyopathies characterized by an arrhythmic risk, such as arrhythmogenic right ventricular cardiomyopathy (ARVC), had a better survival compared with those with structural heart disease and left ventricular dysfunction, and this finding is associated with a much younger age at implant.^15,16 Of 176 Brugada patients followed for mean 84 ± 57 months, only 8 (4.5%) patients died.¹⁵ Of 312 patients with ARVC followed for mean 8.8 ± 7.3 years, only 5 (2%) patients died.¹⁷ Finally, a meta-analysis published in 2016, which include more than 4400 patients that were followed for mean 51 ± 38 months, reported survival rates of 96.5% for HCM patients, 97.7% for ARVC patients, 99.4% in long QT patients, and 99.9% in Brugada patients.¹⁶ It is obvious that especially these young patients with an extremely good survival are the first advocates for maximizing the longevity of ICDs, bearing in mind the risk of infection with every single device replacement (corresponding to 2.6% in 12 800 ICD and CRT-D procedures).¹⁸

It is clear from the expected survival of patients implanted with an ICD or a CRT-D device, according to currently accepted indications, that avoidance of any major mismatch between patient survival and device longevity is critical.

Recent evolution of device technology and improvement in projected and actual longevity

Replacement of cardiac implantable devices is associated with a notable risk of complications.¹⁹ Moreover, replacements involve incremental costs, and extending longevity has been shown to significantly improve cost-effectiveness of ICD and CRT-D therapy.²⁰

Industry-independent studies on defibrillator longevity revealed large discrepancies in device lifespan depending on device type, generation, and manufacturer. Overall, dual-chamber ICDs exhibited shorter longevity than single-chamber ICDs, but lasted longer than CRT-Ds. Cardiac resynchronization therapy defibrillators are the most demanding antiarrhythmic devices in terms of battery consumption because of the need for continuous biventricular pacing and because pacing thresholds tend to be higher in the left ventricle.²¹

The results of previous studies comparing longevity of ICDs across manufacturers are frequently contradictory. This can be ascribed to the single-centre nature of the studies, the variety of population, the low number of devices in analysis and mainly to the heterogeneity of device technologies and generations. The results of the studies published before 2012^22–25 on devices manufactured before 2007 (most of which were no longer available at the time of publication) have been contradicted by more recent observations.^9,26–28

Indeed, device longevity is based on battery technology, the efficiency of the electronic circuitry, the availability of specific algorithms for reducing battery consumption, and varies not only among manufacturers but also among generations of devices from the same manufacturer. Figure 1 reports the evolution of battery technology used by manufacturers over time.

Alam et al.²⁶ determined longevity of CRT-Ds of three major manufacturers in 646 patients implanted between 2008 and 2010. They found important differences among manufacturers, with a 4-year survival rate ranging from 67% to 94%. These results have been confirmed by Landolina et al.²⁷ The authors stratified 1726 consecutive CRT-D devices by manufacturers and time of market release (before or since 2007) and demonstrated an overall 5-year longevity of 54%, with significant differences among currently available CRT-D systems from different manufacturers. Moreover, they also demonstrated significant longevity improvements in recent devices from all manufacturers. In their study, battery longevity was independent of the burden of defibrillator therapy, confirming previous reports,^22–24,26 but was significantly associated with the left ventricular output, confirming previous findings by Thijssen et al.²⁴ and Alam et al.²⁶ and with the unipolar left ventricular pacing configuration. Among the device characteristics, Lithium Manganese Dioxide cell chemistry and batteries from the manufacturer who adopted a 2.0 Ah cell capacity were protective factors against early device replacement.

More recently, Von Gunten et al.⁹ performed a similar analysis in patients who received devices at two institutions from 1994 to 2014. They included not only patients who received CRT-D, but also single-chamber and dual-chamber ICDs. They confirmed that recently all manufacturers have improved longevity of their devices, and that large differences in longevity still exist. More importantly, their data also confirmed that in spite of these improvements, survival of patients continued to be better than ICD longevity and not vice versa as should be the case.

In order to overcome recurring problems with transvenous leads, a subcutaneous ICD (S-ICD) has recently been developed. A Class IIa recommendation for S-ICD has been added to the most recent European Society of Cardiology (ESC) Guidelines for patients with ventricular arrhythmias.³⁰ At present, no data are available with regard to the longevity of these new devices, except for the analysis of the patients enrolled in the European Regulatory Trial recently published by Theuns et al.³¹ They reported that compared with contemporary transvenous ICD systems, longevity of the first generation S-ICD was shorter, with a median longevity of 5.0 years, and not affected by the number of delivered shocks. Nonetheless, also for S-ICD a new generation of devices is currently on the market, designed to provide greater longevity. Moreover, the severity of possible complications associated with replacement of an S-ICD are expected to be lower than with a transvenous ICD.

Further confirmation of the importance of ICD longevity comes from the medical technology guidance published in March 2017 by the National Institute of Health and Care Excellence (NICE) in the framework of the Medical Technologies Evaluation Programme.³² The NICE committee evaluated the literature about currently available ICD battery technology, and concluded that there is good evidence to support the clinical benefit of longer battery life and the associated reduction in device replacements. According to the document, the available evidence suggests that the advantages of longer battery life have not been surpassed by other types of technical advances. The committee therefore concluded that adopting devices powered by batteries with 2 Ah capacity and manganese dioxide chemistry is likely to reduce the number of avoidable replacement procedures a patient may have to undergo, thereby offering improved outcomes for patients and potential savings to the National Health System. Moreover, the committee encouraged further studies that provide data on the battery life of cardiac devices in clinical practice.

Data and projections from the industry on devices longevities

Although real-world industry-independent longevity data are crucial for the assessment of the performance of implantable cardiac devices, the main limitation of all published studies on ICD and CRT-D longevity in clinical practice is the applicability of results to currently available ICD models, due to the high rate of release onto the market of newer device families.

Other information on the performance of devices is provided by the industry and comes from product performance reports. These reports are continuously updated with data from the analysis of returned devices and are periodically published.^33–37 These data suffer from the same limitation of industry-independent longevity publications, i.e. they may no longer apply to newer device families. Moreover, cases of premature battery depletion may be underreported, and clinical events such as removal for infection, system upgrade, heart transplantation, or patient death are not considered and then censored and are estimated in analyses only by applying statistical adjustments and correction factors. Indeed, Alam et al.³⁸ recently demonstrated significant discrepancies between actual battery longevity of CRT-Ds and data reported by manufacturers, with product performance reports highly overestimating longevity.

At the time of the release onto the market of a new ICD, the manufacturer provides projections on battery longevity, based on laboratory testing under controlled conditions. Apart from the possible mismatch between real-life conditions and accelerated battery life tests, data presented in the device manuals are difficult to interpret, due to the lack of homogeneity in the assumptions for longevity projections and the paucity of information about the impact of specific features on battery drain. As example, Table 2 summarizes the information reported in the device manual of some currently available devices from different manufacturers.

In conclusion, an effort should be made to allow for technology comparison, and public health authorities should promote standardization in the assessment and reporting of device performance. For this purpose, since 2015 the French Health Technology Assessment body (Haute Autorité de Santé—HAS) has been requiring manufacturers to predict longevity with exactly the same settings.³⁹ Moreover, the HAS has defined the minimal longevity requirements based on the specified settings: 7 years for single-chamber ICD, 6.5 years for dual-chamber ICD, 5.5 years for CRT-D.

At the same time, an effort should be made by the manufacturers in providing data as transparent and complete as possible, and the availability of tools for the automatic calculation of longevity projections under specific working conditions may be useful.⁴⁰

Battery technology

Battery depletion is the dominant cause for ICD replacements.^27,41 The high power demands of ICDs necessitate a battery, which is capable of delivering high current pulses of 2–3 A in order to rapidly charge the capacitors of the device.⁴² Furthermore, the battery must supply a constant low current to power the other functions of the ICD, i.e. data logging, telemetric communication, standard pacing, biventricular pacing for CRT, anti-tachycardia pacing, and the integration of innovative physiologic sensors. The capacity of a battery is measured in ampere-hour (A⋅h) that is the amount of electric charge that will support a current of 1 ampere for 1 h (1 A⋅h = 3, 600 coulombs). In theory, the capacity can be made as high as required but at the expense of a larger can size. In the last decade, ICD development has focused on enlarging the battery capacity, which has almost doubled without major increase in the can size. Currently available ICDs have capacities ranging from 1 to 2.0 A⋅h, and among them a larger battery capacity was shown to be associated with prolonged device longevity in clinical use.²⁷

Electrochemically, a battery consists of an anode and a cathode physically separated. All batteries used today to power ICDs are lithium anode based, while the cathode is made of silver vanadium oxide (SVO), carbon monofluoride (CFx), SVO-CFx hybrid, or manganese dioxide (MnO₂), with temporal differences in adoption of these types of batteries by each specific manufacturer, as shown in Figure 1.

Since the late 1980s, the power source chemistry used in ICDs has been the Li/SVO. It provides greater energy density than previous chemistries and over the years has undergone numerous improvements, becoming the most commonly used battery for ICDs. However, the discharge curve of Li/SVO is non-linear and makes it difficult to determine elective replacement. A similar behaviour is shown by the more recent SVO-CFx hybrid cells, whose cathode consists of CFx sandwiched between two layers of SVO material.²⁹ Moreover, the internal cell impedance changes and this leads to variable capacitor charge times over the life of the ICD. By contrast, modern MnO₂ batteries are able to maintain a high battery voltage and a relatively stable internal resistance during most of their discharge. This allowed to increase the level of usability from 70% of the total capacity obtained with standard SVO cells to 90% with MnO₂ batteries. According to data collected in clinical practice,²⁷ MnO₂ batteries were shown to be independent protective factors against early depletion in clinical practice.

Factors affecting device longevity

In addition to the cell capacity, the current drain of the system represents a major factor affecting device longevity. Modern ICD and CRT-D devices are endowed with several diagnostic and therapeutic algorithms and features that may have a relevant impact on the device longevity. Table 2 summarizes the information reported in the device manual of currently available devices from different manufacturers. Table 3 summarizes the variables identified as predictors of early battery depletion in the literature.

Background current

The background current needed to run the device’s electronic circuitry is typically in the μA range.²⁹ Over the last years, more efficient circuitry has allowed to decrease energy consumption regardless of the extra functions incorporated into devices.²⁹

Collection and storage of electrograms

The collection and storage of electrograms before tachycardia detection is intended to facilitate the subsequent clinical diagnosis of the arrhythmia. However, continuous storage of pre-episode electrograms, in addition to the episode summary data and the post-episode electrogram, may significantly reduce device longevity. According to the information reported in the device manual of currently available ICD and CRT-D from different manufacturers, electrograms collection and storage may reduce longevity as much as 16%.⁴⁴

Capacitor reformation

In the high-voltage capacitors for ICDs, an oxide layer is formed on the anode surface to provide the dielectric for the capacitor. However, this oxide layer can degrade through chemical and physical processes over time, causing possible capacitor deformations, heating and even explosions.

To avoid degradation, an electric current is passed through the capacitor periodically (capacitor reformation process). In modern ICDs two reformations per year are required in absence of any shock therapy. However, this process requires energy and each reformation may consume up to 24 days of the total device longevity.⁴⁴

Previous studies showed an association between the frequency of capacitor reformations and the time to battery depletion,²³ and trials demonstrated that reducing capacitor charges prolonged battery longevity.⁴⁵ Similarly, a recent analysis showed an association between battery drain and the number of shocks delivered and diverted.⁴³ Therefore, increasing the detection rate or lengthening the monitoring delay before shock delivery could be effective programming strategies not only to reduce inappropriate therapies and improve outcome,^46,47 but also to save energy.

Pacing

Among the factors associated with battery drain, pacing parameters play an important role but differences exist among device types. The impact of pacing on battery drain is negligible for single-chamber ICD, while it may be relevant for dual-chamber ICD and in particular for CRT-D, because of the need for continuous biventricular pacing and because pacing thresholds tend to be higher in the left ventricle.^27,43 Indeed, when the required output for a pacing channel exceeds 2.5 V, voltage amplification of the battery voltage is needed and this is not always an energy-efficient process.²⁹ Careful programming of pacing output is therefore crucial, and algorithms for automatic verification of myocardial capture and continuous optimization of pacing outputs have been developed. These are primarily intended to protect the patient against loss of pacing capture in case of rise in the chronic threshold, but also to prolong device life.⁴⁸

For CRT devices, alternative pacing options have also been proposed that may have an impact on device longevity. The Medtronic Adaptive CRT algorithm periodically measures intrinsic conduction and delivers single-chamber left ventricular pacing in case of preserved atrioventricular conduction. According to the estimations of the manufacturer,⁴⁴ the activation of this feature that essentially reduces the percentage of right ventricular pacing may potentially increase the device longevity by 0.2–0.8 years.

Quadripolar left ventricular leads and multisite pacing

Quadripolar left ventricular leads have become the standard of care in CRT. The availability of multiple pacing options allows the implanter to manage issues represented by phrenic nerve stimulation and high pacing capture thresholds. A quadripolar system allows to choose among up to 17 pacing vectors, increasing the chance (at the time of implantation and during follow-up) of finding a low left ventricular capture threshold, which may ensure effective CRT delivery and possibly prolong device longevity.

Recently, CRT devices with quadripolar left ventricular leads have been endowed with the possibility to deliver simultaneous or timed pacing from two electrodes of the left ventricular lead, in order to improve CRT response by simultaneously recruiting a larger volume of myocardium. Preliminary results seem positive⁴⁹ and additional large studies are currently ongoing. However, the addition of a second left ventricular pacing channel may have a relevant impact on the longevity of the device and the expected clinical benefits from this pacing strategy must justify the premature ICD depletion and the consequent anticipated device replacement. According to the information reported in the device manual of CRT-D systems from different manufacturers, multi-site left ventricular pacing may reduce longevity by more than 1 year.^50–52 This has been recently confirmed by Akerstrom et al.⁵³ who estimated the impact of multi-site pacing activation on battery longevity. Based on a simulation using real-life device interrogation lead and programming data, they found that multi-site pacing activation significantly reduces battery longevity. Moreover, maximal interelectrode cathode spacing for the excitation of larger myocardial area could only be obtained by accepting configurations with high pacing thresholds. Therefore, they concluded that balancing the shorter time to generator replacement against the potential benefits of multi-site pacing is a rather difficult task.

Device interrogations and remote transmissions

The ICD exchanges data with a programmer at the time of implantation and in-office visits, and with a base station in the patient’s home if remote monitoring is adopted and enabled. The short-range (around 5 cm) communication performed by applying the programmer wand on the patient chest is based on inductive telemetry. This communication involves not only the exchange of data between ICD and programmer, but also the transmission of energy to the ICD for the communication. Therefore, this type of interrogation does not consume ICD battery energy. By contrast, longer-range communications with the programmer and with the remote monitoring base stations are based on radio frequency telemetry, in accordance with international standards [Industrial, Scientific, and Medical (ISM) or the Medical Implant Communication Service (MICS) standards]. Radio frequency transmission between the external device and the antenna housed in the plastic header of the ICD is extremely energy-consuming.²⁹ Manufacturers adopted different schemes to initiate and complete the communication, in order to save as much as possible battery energy by minimizing the time the ICD must spend in an active state. The remote station can ‘wake up’ the ICD by sending signals, and in this case the ICD will spend energy to periodically detect possible calls. Alternatively, the ICD powers up periodically to transmit a notification and, if the remote station is in proximity and the communication is established, the transmission occurs. This second scheme may have a higher impact on device longevity and this is the reason why ICDs adopting it are typically programmed to attempt transmissions only for few hours per day, when the probability to be near the remote station is higher (e.g. during the night when the patient is in his bed).

In addition to periodical transmissions, the majority of remote monitoring systems allow to perform daily ICD checks to verify possible alert conditions without performing a complete data transmission. In general, this has a minor impact on device longevity.

Estimations of the possible impact of remote data transmissions on device longevity are reported in Table 2.

However, although remote transmission is one of the possible factors associated with battery drain, it has been shown that the appropriate use of remote monitoring may have a favourable impact on device longevity. The Effectiveness and Cost of ICDs Follow-up Schedule with Telecardiology (ECOST) trial showed that remote monitoring was associated with a significant saving in longevity, as it facilitated optimal device programming and consequent reduction in inappropriate shocks.⁴⁵ More recently, Campana et al.⁵⁴ showed that remote monitoring facilitated the identification of the proper time of ICD replacement, reducing at the same time the frequency of scheduled visits. Therefore, in clinical practice remote monitoring enables patients to be closely monitored, ensuring that replacement occurs before an unsafe status ensues, but also possibly decreasing the high proportion of devices that are replaced before elective replacement indicator notification.⁵⁵

Clinical benefit of extending implantable cardioverter-defibrillator/cardiac resynchronization therapy longevity from a patient perspective

Extending device longevity is important for three principle reasons; (i) to reduce complications, (ii) to comply with patient preference, and (iii) to improve cost efficacy. Complications may be prevented by delaying or avoiding altogether a generator replacement. Whilst the surgical procedure itself imparts risk, the patients at the time of replacement are older and have more comorbidities than at initial device implantation. Actually, they have an increased risk of complications and an increased likelihood of dying of non-arrhythmic causes. The risk/benefit ratio of device replacement is unclear which is compounded by the lack of data supporting decision making at time of ICD/CRT-D replacement compared with the evidence for indications at implant. Thus, it is difficult to provide patients with accurate information at the time of consent under which circumstances their preference cannot be ignored and becomes a key issue. Patients must be supported if they decide not to proceed with an ICD/CRT-D replacement but must also understand the potential consequences.

Complications

The overall complications that characterize cardiac electronic implantable device (CIED) generator replacements are different compared with primary implantations because most complications at the time of the implant are related with the leads (dislodgement or perforation) while replacements are associated with an increased risk of infections.

Data from the literature on the occurrence of major complications in case of device replacement are shown in Table 4. The REPLACE registry prospectively studied patients undergoing pacemaker and ICD replacements in 72 US academic and private practice centres.¹⁹ Major complications were reported in 4.0% (95% CI 2.9–5.4) of 1031 patients without lead insertion or replacement including a 6 month infection rate of 1.4%. Overall complications were higher with ICD compared with pacemaker generator replacements.

In a further prospective, multicentre, population-based registry of 1081 lone ICD replacement patients at 18 centres in Canada, a 4.3% complication rate was found at 45 days.⁵⁶ The most common complications were infection (2.1%), lead revision (3.2%), electrical storm (1.3%), and pulmonary oedema (1.2%). Risk factors associated with them were the use of antiarrhythmic therapy (adjusted HR 6.29, 95% CI 2.07–19.09), implanter volume (adjusted HR 10.4, CI 1.32–82.14 for <60/year vs. <120/year; P = 0.026), and Canadian Cardiovascular Society angina class (adjusted HR 3.00, CI 1.11–8.15 for class 2–4 vs. 0–1). A population-based cohort study⁵⁷ in all Danish patients, who underwent a CIED procedure from May 2010 to April 2011, reported lower risk of any major complication after generator replacement (3.5%) compared with new implants (5.8%). Generator replacement procedures were associated with increased infection risk compared with first implants (1.5%, 95% CI 0.8–2.2; P = 0.001), confirming previous findings. A 4.05% median rate of major complications, with a range between 0.55% and 7.37%, was found in a recent systematic review of all available literature.⁵⁸ Major complications included infections, haematoma, reoperation for other reasons, and stroke. Infections requiring antibiotic therapy and/or lead extraction had an incidence of 2.2%, whilst those for haematoma requiring evacuation and for reoperation due to any other reason (lead damage, repositioning of the device due to intolerable pain, and pocket erosion) were 0.7% and 2.0% respectively. Mortality ranged between 0% and 0.44%. Minor complications, such as incisional infections, haematoma or discomfort were reported in a further 3.5% of cases. Some of the studies found more complications when the replacement was done by a single operator or by operators with non-electrophysiology training.

Other studies have found that the number of previous procedures at the same pocket site was associated with an increase in complications.⁵⁹ This emphasises the importance of extending device longevity. The same study also found that combined consultant and fellow operators were associated with a decrease in complications compared with a single operator.

Infections

Infections are the most common complication after ICD and CRT replacement with an incidence of approximately 1.5–2%. The 6-month infection rates found in the REPLACE registry were 1.4% (95% CI 0.7–2.3) and 1.1% (95% CI 0.5–2.2), respectively.⁶⁰ All patients received preoperative intravenous antibiotics and 68.7% received post-operative systemic antibiotic therapy. Patients with infections were more likely to have had post-operative haematomas (22.7% vs. 0.98%, P = 0.002). Centres with infection rates >5% were more likely to use povidone-iodine for topical antisepsis and had lower implantation volume.

In the population-based cohort study on Danish patients where detailed information on complications after CIED procedures was collected by systematic review of all patient charts and was therefore particularly accurate, the highest infections rates were found in case of upgrade/lead revisions (1.9%, 95% CI 0.6–3.2 when compared with 0.6%, 95% CI 0.3–0.8 for normal implants).⁵⁷

Evaluation of more than 200 000 ICD procedures in the US National Cardiovascular Data Registry has shown that ICD generator replacement results in a higher rate of infection compared with initial implant (1.9% vs. 1.6%, P < 0.001).⁶¹ This has been supported by the work of Borleffs et al.,⁶² who showed that occurrence of surgical re-intervention due to infections in ICD replacements was overall 2.5 (95% CI 1.6–3.7) times higher compared with first implanted ICDs. This study also demonstrated the increasing risk of sequential replacements, ranging from 1.5% for the first to 8.1% for the fourth implanted ICD. A one-third of ICD patients undergoing generator replacement or lead revision have an asymptomatic bacterial colonization of the generator pocket.⁶³ After the revision, 7.5% of these patients develop a device infection with the same species of microorganism. The most common bacteria isolated was coagulase negative staphylococci (68%). Povidone-iodine irrigation of the subcutaneous pocket did not alter the rates of pocket infection after pacemaker/defibrillator implantation.⁶⁴

Additional clinical considerations

There are no randomized controlled trials testing the benefits and true risk reduction of ICD/CRT-D replacements. There is however, observational data about all-cause mortality in a large registry of 111 826 patients, which showed 9.8%, 27.0% and 41.2% mortality rate at 1, 3, and 5 year follow-up, respectively.⁶⁵

Other investigators have studied the incidence of appropriate therapies following ICD/CRT-D replacement. In these reports, patients were divided in two groups: those with an ongoing indication for ICD due to LVEF <35% and/or history or appropriate therapies by the device, and those not fulfilling these criteria. A retrospective review of 231 primary prevention patients found that 26% no longer had indications for an ICD at the time of replacement.⁶⁶ These patients received appropriate ICD therapies at a significantly lower rate than patients with ongoing ICD indications following generator replacement (2.8% vs. 10.7% annually, P < 0.001). Similar findings were observed in another study of primary prevention ICD patients 3.3 years (median) after generator replacement.⁶⁷ Patients with EF ≤35% were more likely to experience ICD therapy compared with those with EF >35% (12% vs. 5% per year; HR 3.57; P = 0.001).

Patients with CRT-D often demonstrate an improvement in left ventricular EF by the time of device generator replacement, and this makes unclear whether to replace the device with a CRT-D or a CRT-P. In the Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) 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.¹⁰ Sebag et al.⁶⁸ studied 107 patients with an initial primary prevention CRT-D and found that 37% had left ventricular EF ≥40%, and no history of appropriate ICD therapy at the time of device generator replacement. After 26.4 ± 14.4 months follow-up post-device replacement, 95% patients of these latter patients were free of appropriate therapy vs. 49 of 68 (72%) patients with ongoing ICD indication (P = 0.007). The only independent predictor of appropriate therapy after generator replacement was an ongoing ICD indication (HR 6.43; P = 0.01).

Patients preference

The option of no device replacement or replacement by a pacing only device should be considered in patients with significant comorbidity or with a short life expectancy. A recent study of explanted ICDs after death found that 35% of patients received ICD shocks within the last hour of life, but in only 13% of cases was death considered of an arrhythmic cause.⁶⁹ This suggests that the delivery of shocks in this setting is unlikely to change the ultimate outcome. This has to be considered whenever the end-of-life of the patient period is foreseeable months in advance because of advancing non-cardiac conditions, such as such as dementia, progressive lung disease, cancer, or frailty/debility syndromes.

Device replacement is however, not only associated with complications and cost but with surgical stress and discomfort to the patient. It has thus been proposed that patient preference should be considered and discussed at the time of all ICD/CRT-D replacements.⁷⁰

A detailed survey found that the vast majority (90.1% vs. 9.9%, P < 0.0001) of patients would prefer a long lasting device with a larger size rather than a smaller device with a shorter battery life.⁴¹ This preference was consistent even in thin (79.5%), female (89.2%), and elderly (>72 years) (90.5%) patients.

Pedersen et al.,⁷¹ found that most elderly patients were unaware that it was possible to decline device replacement. Furthermore, they had confused ideas regarding the risk/benefit balance with replacement and about the consequences of ending life with an ICD. However, the majority of ICD patients seemed to favour device deactivation at the end of life, primarily due to the wish for a dignified, pain free death. This indicates that patients think about deactivation but not avoiding ICD replacement near the end of life and might welcome discussing it with their physician.⁴¹ Healthcare professionals have an ethical and moral responsibility to discuss this with them.

Economic benefit of extending the longevity of implantable cardioverter-defibrillator/cardiac resynchronization therapy defibrillator devices

Healthcare systems need to face an increasing demand for costly medical interventions and this includes also medical devices. Clinicians, public health physicians, regulatory institutions, and policy makers need to find a reasonable balance between what is best for the individual patient and what society can realistically afford. Health technology assessment (HTA) has been proposed as a bridge between evidence and policy-making, with the aim to offer guidance to decision-making in a multidimensional approach, and with a circular pathway as depicted in Figure 2.^72–74

The approach of HTA appears particularly appropriate in the case of CIEDs, which typically imply an asymmetrical, discontinuous distribution of costs, quite different from the pattern of costs due to drug therapies, usually associated with a progressive cumulative increase of resources consumption.⁷² In an economic approach, the budget impact of a given therapy must consider the upfront costs as well as the overall costs over the duration of therapy.

In the case of ICD and CRT-D devices the costs for the device and leads, the cost of hospitalization for the implant and potential complications, and the cost for device replacement (device purchase and cost of procedure) are all important determinants of the costs at long term, as shown in Figure 3. In consideration of the characteristics of current technology, of patient age at implant and of expected patient survival, an extension of current device longevity appears an important priority in the case of ICD and CRT-D devices, while a less pronounced impact is expected in the case of pacemakers implanted for bradycardia.^75,76

The value of extending the longevity of ICDs and CIEDs devices can be appreciated by using economic models. A specific model has been proposed, the Longevity Model⁷⁷ for calculating the costs associated with ICD and CRT-D device therapy, in a hospital perspective, according to varying device longevities (ranging from 4 to 15 years) in a ‘real-world’ clinical practice setting. The model included inputs corresponding to the costs of complications related to initial implantation and replacement of devices, as well as inputs related to life expectancy of different categories of potential device recipients (patients with or without left ventricular dysfunction and heart failure). The economic model assumed that, other patients’ characteristics being equal, the longer the device longevity, the lower the number of device replacements that patients will face during lifetime. The modelling study showed that, in a 15-year time window, the extension of device longevity (e.g. from 5 to 9 years for single-chamber ICDs) can have an important impact in reducing long-term costs of device therapy, with substantial daily savings in favour of devices with extended longevity, in the range of 29–34%, depending on clinical scenarios. The specific economic advantage of any increase in device longevity for specific patient characteristics, in the setting of single-chamber ICDs and CRT-D, respectively, can be detected in the Supplementary material online, Figures S1 and S2. The model allowed also to calculate that the cost per day of a single-chamber ICD with a longevity of 9 years is in the range of €3.8, while for a CRT-D device with a longevity of 7 years it may correspond to €5.5. These data clearly allow to have a different approach to device treatment and to overcome the frequent perception of ICD and CRT-D devices as treatments with unaffordable costs, higher than costs of other commonly accepted treatments.^22,78,79 Conversely, these data could help to overturn the misconception that up-front costs are the only metric with which to value device treatments.

All the projections on longevity of devices and related costs are based on modelling and obviously have some limitations, but data from large datasets appear to confirm the reliability of these estimates based on device battery consumption.⁴³ National registries play a major role in providing appropriate feedback on actual longevity of implanted devices, with appropriate comparisons between predicted longevities and actual performances.^21,80,81

In a real-world analysis of around 1800 CRT-D implants published in 2016⁸¹ the need for device replacements at 6 years was reduced by use of most recent generation devices from 83% to 68%, and this had a major impact on costs since the need for replacement increased total therapy costs by more than 50% over the initial implantation cost for hospitals and by more than 30% for the healthcare system. In consideration of the differences in battery chemistry adopted by different devices, the maximum difference in therapy costs between manufacturers was 40% for hospitals and 19% for the healthcare system.

Furthermore, a national study based on real world data from Sweden and data from the Swedish Pacemaker and ICD registry, based on data collected between 2008 and 2010 showed at base-case analysis that up to 603 replacements and up to SEK 60.4 million cumulative-associated costs (around €6.04 million) could be avoided over 6 years by using devices with extended battery life.⁸² The assumption at the basis of this analysis and the methods for cost calculation are shown in Supplementary material online, Table S1 and Figure S3. The pattern of savings over time suggests that savings are modest initially but they increase rapidly beginning in the third year of Follow-up with each year’s cumulative savings 2–3 times the previous year (Figure 4 for CRT-D). Evaluating CRT-D, ICD-VR, and ICD-DR devices together over a longer 10-year period, the sensitivity analysis showed 2820 fewer replacement procedures and associated cost savings of SEK 249.3 million (around €24.93 million) for all defibrillators with extended battery life.

The economic impact of an improved longevity of CIEDs has also been assessed in a macro-economic perspective. NICE considered a cost modelling based on published data of CIEDs, and found that the price and lifespan of the CRT-D have the greatest effect on overall treatment costs.³² According to these projections, adoption of the most advanced technology in batteries of CRT-D devices (batteries with Li/MnO₂ chemistry) would result in the perspective of the UK National Health Care system in a saving of around £6 million in the first 5 years.³²

The marked reduction in the daily costs of device therapy related to extended device longevity should lead to consider device longevity as a determining factor in device choice by physicians and healthcare commissioners and should be appropriately considered and valued in comparative tenders.

Open issues and areas of research

An important open issue that limits any comparative analysis of data on projected devices longevities provided from the different manufacturers derives by the lack of a standard in the assumptions used for making projections. It is time to establish a standard in order to allow proper comparisons. This is an important limitation for comparative analysis in tenders, where device longevity should be appropriately evaluated and valued.

Moreover, it will be important to collect independent data from ‘real world’ registries in order to validate the data obtained by projections on ICD and CRT-D devices longevities.

By lengthening device longevity, it could be hypothesized that the evolution of ICDs and CRT-Ds in terms of therapeutic algorithms will not become available to patients for an extended time period, until the next device replacement. This is an issue that could be solved, in the future, by the possibility to update devices’ software thus making available the most advanced device technology in terms of software.

The use of devices with extended longevity is currently not valued by reimbursement practices,⁸³ but a revision should be considered in view of available data.

We mainly discussed technical data related to transvenous ICD, with limited focus on S-ICDs, for whom, anyway, an additional improvement in device longevity is desirable.

Finally as an important weak point of ICD/CRT-D systems is represented by the leads, a parallel improvement of the long term performance of ICD and CRT-D leads when compared with devices has to be pursued in order to guarantee an extended reliability of the entire device system.

Conclusions

The importance of extending the longevity of ICDs is quite obvious in the setting of purely electrical diseases (i.e. without left ventricular dysfunction), a setting where patient survival is expected to be quite similar to that of the normal population, provided that sudden arrhythmic death is prevented by the ability of the ICD to terminate life-threatening ventricular tachyarrhythmias. However, the relevance of extending the longevity of ICDs and CRT-D devices is currently actual also in the setting of patients with left ventricular dysfunction and heart failure. Indeed, long-term outcome of heart failure patients has been recently improved when compared with previously established drug treatments⁸⁴ or will even more improve with new agents under appropriate evaluation,⁸⁵ even if the contribution of comorbidities has to be considered for all the unselected patients treated in ‘real world’ practice.^86–88

Extension of device longevity may reduce the mismatch that existed in the past between patient survival and device service life. This lead to an increased need for device replacements, with increased risk of complications, particularly increased risk of infections. An extension of device longevity has been obtained by improved battery chemistry and device technology and allows to obtain both clinical and economic benefits, in line with patient preferences and needs. From a clinical point of view, the availability of this improvement in technology allows to better tune the choice of the device to be implanted, taking into account that the reasons supporting the value of an extended device longevity as a clinical priority may differ according to the clinical setting, as shown in Figure 5. From an economic point of view, extension of device longevity may have an important impact in reducing long-term costs of device therapy, with substantial daily savings in favour of devices with extended longevity, up to 30%, depending on clinical scenarios. In studies based on projections, an extension of device longevity allowed to calculate that the cost per day of a single-chamber ICD with a longevity of at least 9 years is in the range of €3.8, while for a CRT-D device with a longevity of at least 7 years it may correspond to €5.5. These values imply that we could consider a different approach to device treatment and to overcome the frequent perception of ICD and CRT-D devices as treatments with unaffordable costs, higher than the costs of other commonly accepted treatments and to overturn the misconception that up-front costs are the only metric with which to value device treatments. In view of its clinical and economic value, device longevity should be a determining factor in device choice by physicians and healthcare commissioners and should be appropriately considered and valued in comparative tenders.

Supplementary material

Supplementary material is available at Europace online.

Acknowledgements

The authors thank EHRA Scientific Documents Committee: Gregory Y.H. Lip, Laurent Fauchier, David Arnar, Carina Blomstrom-Lundqvist, Zbigniew Kalarus, Gulmira Kudaiberdieva, Georges H. Mairesse, Tatjana Potpara, Irina Savelieva, Jesper Hastrup Svendsen, Vassil B. Traykov.

Funding

This review was prepared by a writing group that was invited on behalf of the European Heart Rhythm Association (EHRA) Education Committee. The document was supported with an educational grant from Boston Scientific. However, the content has not been influenced in any way by its sponsor.

Conflict of interest: G.B. reported speaker's fees from Biotronik, Boston and Medtronic; J.L.M. reported speaker's fees and honoraria from Abbott, Bayer, Biotronik, Biosense Webster, Boston, Cardiome, Daiichi Sankyo, Medtronic, Sorin and research grants from Boehringer-Ingelheim, Boston, Daiichi Sankyo, Medtronic, Pfizer; D.J.W. has received research grants and consultancy fees from Medtronic and Boston Scientific and educational fees from St Jude Medical; F.G. reported honoraria from Abbott/St.Jude Medical and Medtronic; B.S. reported to be part of Boston and Medtronic speaker’s bureau; M.L. reported to be part of speakers’ bureau appointment with Medtronic, LivaNova and Boston Scientific and an advisory board relationship with Medtronic.

References