Implantable Cardioverter-Defibrillator Use in Patients with Left Ventricular Assist Devices

Abstract

This chapter is developed with the intention of discussing the use of implantable defibrillator cardioverters (ICDs) in patients with left ventricular assist devices (LVADs). LVADs have become the standard treatment for patients with advanced heart failure who require prolonged mechanical circulatory support as a bridge to transplantation or as destination therapy. Patients with advanced heart failure have a major risk of sudden death due to ventricular dysrhythmias (VD) so an ICD could be indicated, but it remains unclear within the LVAD population due to several factors including sustained VD good tolerance and inappropriate therapies (due to supraventricular tachycardias or electromechanical interferences) as well as the risk of infections with complex antibiotic therapy or device replacements. Previous VD before LVAD placement, concomitant atrial fibrillation, type of LVAD device, and chronic ischemic heart disease can predict future episodes of VD. The evidence that supports ICD use in patients with LVAD is very limited, and current guidelines are based primarily on the consensus of experts and observational studies. Nowadays, an ICD implant is only recommended for LVAD patients who develop postoperative VD associated with hemodynamic collapse, and it should be programmed in a very conservative mode (higher rate and larger intervals to detection) to avoid undesirable electric shocks.

1. Introduction

Left ventricular assist devices (LVADs) have become the standard treatment for patients with advanced heart failure who require prolonged mechanical circulatory support as a bridge to transplantation or as destination therapy [1, 2, 3, 4, 5]. The first generation of LVAD consisted of pulsatile pumps that were successful in unloading the heart, but they were limited by size and poor long-term durability. The second and third generations use continuous flow (i.e., Heartware or Heartmate II devices). Heartmate III is nowadays the only third-generation device available to implant due to its magnetic levitation rotor related to lower rates of thrombus and hemolysis. One-year survival rates are approximately 80% and 70% at 2 years [4].

Patients with advanced HF have a major risk of sudden death due to ventricular dysrhythmias (VD). Therefore, an implantable cardioverter defibrillator (ICD) could be indicated. Primary prevention needs the device prescription before a fatal episode and is indicated if severe LVEF dysfunction (less than 35%) is diagnosed using any cardiac imaging tool such as transthoracic echocardiography or cardiac magnetic resonance, 40 days after myocardial infarction, despite full-medical protocol, the patient is not stable and if the estimation of the patient survival is more than 1 year. On the other hand, the secondary is indicated after ventricular arrhythmia [3, 6, 7, 8, 9].

Patients in stage D often die of pump failure rather than sudden death, so ICD therapies have not been recommended in this population. Cardiac output impairment is less common [10] in LVAD subjects, and VD is better tolerated [3] than in the non-LVAD population. Moreover, inappropriate therapies and complex infections may increase the risk of ICD therapy. Therefore, ICD therapies remain controversial in the LVAD population, and further research is needed.

2. Epidemiology of ventricular arrhythmia and cardiac sudden death

The highest rates of VD occur in the first 30 days after LVAD placement; however, late ventricular dysrhythmias have also been described. VD incidence ranges from 22% to 59% after LVAD implant [10, 11, 12] and 35% of post-LVAD recipients within 30 days [13, 14].

Several factors can predict VD events after LVAD implantation, i.e., previous ventricular dysrhythmias before LVAD implant (hazard ratio 3.28) [15], previous history of atrial fibrillation [14, 15, 16], the type of LVAD implanted, or the presence concomitant chronic ischemic heart disease [17]. In total, 42.4% of patients with a history of VD pre-LVAD experienced VD events post-LVAD in comparison with 16.7% without previous VD [12]. Ischemic heart disease was the major cause of cardiomyopathy in 71% of patients in the VD group and 45% of patients in the group without VD where dilated cardiomyopathy was more frequent (no difference was made between inheriting cardiomyopathy, enolic, cardiotoxicity, or idiopathic cause). Chronic ischemic heart disease could trigger VD due to persistent subendocardial ischemia, arrhythmogenic substrate associated with myocyte remodeling, and fibrosis [17].

VD often leads to sudden cardiac death in the non-LVAD population. However, long-term survival has been reported in those patients with LVADs despite VD [11]. Interestingly, neither the presence of VD nor ICD therapies (appropriate in 19.1% or inappropriate in 3.1%) were associated with higher mortality rates after 10 months of follow-up [15].

Controversially, other studies have described higher mortality in patients with VD events. Bedi et al. showed an absolute 15% or higher risk of death in the first week after LVAD implantation [17]. Brenyo et al. describe an increase in mortality of up to 10 times higher in patients with LVADs if there is concomitant VD, although 1 year of mean follow-up makes it difficult to correlate VD as a cause of death or as a progression marker of disease [18].

3. Mechanisms and management of ventricular dysrhythmias

The mechanisms of VD in LVAD include:

1. The perioperative adrenergic stimulation and adrenergic agonists may promote early VD [14] because of myocardial or systemic inflammation after LVAD implant, which can elevate catecholamine blood level and promote changes in electrical properties making LV more sensitive to VD. Indeed, isoproterenol, norepinephrine, and dobutamine are treatments often used to avoid RV failure during the perioperative time; however, they can also help to achieve an electrical imbalance and generate a VD substrate.

2. The cardiomyopathy severity, above all, chronic ischemic disease is related to the presence of intrinsic scars that perpetuate reentrant circuits and VD [14].

3. The apical insertion site of cannula inflow has been correlated to the morphological origin of ventricular tachycardia [14]. Although 75% of ventricular tachycardia mapped during electrophysiological studies correlated to an intrinsic scar [19] unrelated to the device.

4. The suction events, an excessive left ventricular discharge due to a mismatch between LVEF filling and LVAD output. When LVEF is excessively unloaded, it can collapse and lead to VD event, monomorphic or polymorphic. But also, an inadequate left ventricular discharge can promote VD [14].

5. The type of LVAD can predict the risk of VD. There was a significant twofold increase in risk for LVAD versus biventricular-VAD [20]. Continuous flow (CF) pumps compared with pulsatile pumps could increase the incidence of VD [21], but it could be related to suction events, more frequent in CF-LVAD due to continuous ventricular unloading.

6. Acute mechanical unloading of the LVEF during recent postoperative time can change electrophysiological properties such as an increase in QT segment within 1 week after LVAD and a decrease after 1 week of mechanical support. A longer QT segment after LVAD placement has been associated with a threefold higher risk for postoperative VD [14].

In a multivariate analysis, VD after LVAD placement in the recent perioperative time was associated with a higher risk of all-cause mortality compared with the population without VD (hazard ratio 7.28). This report suggests that aggressive treatment must be considered [16].

Three main treatment options exist, including adjusting LVAD settings, medical treatment, and VT ablation.

Suction events are a common trigger for ventricular arrhythmias in patients with LVAD; therefore, reduction in LVAD speed or an increase in intravascular volume could solve VD. Once preload and unload are assessed, the next line of treatment should be medical therapy.

Limited literature suggests any potential benefit from the use of β-blockers, amiodarone, sotalol (take care of changes in QT), and sodium channel blockers (lidocaine/mexiletine); but these potential treatments need further studying [14]. A correct repositioning of potassium and magnesium ions is also required.

Radiofrequency ablation therapy has been described as an alternative option in some reports with low complication rates when VD persists, and there is a hemodynamic compromise or a worsening of RV function [13].

If there is prior history of ventricular dysrhythmias in LVAD candidates, surgical ablation at the time of LVAD placement should be considered, as it offers direct visualization of the myocardium and epicardial ablation without defects of epicardial mapping and ablation via subxiphoid pericardial access. Endocardial electroanatomic mapping before the LVAD implantation procedure may let localize VD circuits and guide surgical planning. Once LVAD has been placed, options for ablation of VD are more limited due to pericardial adhesions from the device, and LVAD inflow cannula limits endocardial access to the LV apex.

If catheter ablation is planned, several anatomic and physiologic challenges must be considered. Retrograde access can be limited by insufficient native flow to open the aortic valve, so a transseptal approach to the left ventricle (through the left atrium) is the preferred option. The LVAD often causes magnetic interference that may affect mapping systems.

Nonetheless, a recent systematic review of 18 studies showed that catheter ablation was associated with a decrease in rates of ICD therapies (57 vs. 24%), but VD recurred in 44% at a mean follow-up of 264 days [22].

4. Ventricular arrhythmia tolerance and Fontan-type circulation

LVAD population has a better tolerance for sustained VD, including ventricular fibrillation (VF) than the non-LVAD population, most likely due to the “Fontan-type circulation” phenomena. This population has the same ability to withstand insults as patients with congenital heart diseases and Fontan-type circulation [3] described by Fontan and Baudet in 1971 to palliate tricuspid atresia relying on central venous circulation enabling blood to be directed toward the RV in the absence of pulmonary hypertension.

In patients with LVAD, the right ventricle behaves as a passive corridor driving blood from right to left if adequate preload condition is met. High pulmonary pressure or low central venous pressure is associated with more difficulties to remain stable once VD appears [23, 24].

Sims et al. [25] described how a continuous flow-LVAD could avoid collapse in a patient with sustained ventricular fibrillation over 12 hours when an implantable defibrillator was not able to terminate arrhythmia and external defibrillation was required, due to a correct preload to the LV and normal or low pulmonary resistances. The same findings were described by Busch et al. and Smith et al. [26, 27].

5. Evidence for implantable cardioverter defibrillator device

The evidence that supports ICD indication in patients with LVAD is very limited, and current guidelines are based primarily on the consensus of experts and observational studies [13].

A meta-analysis by Vakil et al. showed that ICD therapy was associated with an absolute risk reduction of 16% and a relative risk reduction of 39% in all-cause mortality at a 7-month follow-up. The number of patients who needed to treat to avoid an outcome was six subjects [11]. Nevertheless, there is not any statistical significance in the continuous-flow LVAD subgroup (34% of the total population). The ICD cohort showed a mortality up to 14% vs. 25% in the non-ICD group (and an absolute risk reduction of 11% and a 24% relative risk (p = 0.17). The lack of significant results was due to the small sample size due to the better long-term results and hemodynamic stability with new continuous-flow assistance [13] with less dependence on native cardiac activity than pulsatile pumps [28].

Rorris et al. did not find any difference in all-cause mortality events, cardiovascular mortality, or right heart failure between LVAD patients with and without an ICD. Although, the study finds important differences in baseline patient characteristics between European and United States populations [29].

A subanalysis of the INTERMACS Registry that included of 2209 patients with an ICD and 2209 patients without one concluded that the presence of ICD was not related to reduced mortality, among patients with a continuous-flow LVAD. The presence of an ICD was associated with increased mortality risk and unexpected death during device support because of the result of propensity score matching, which resulted in an ICD group that had decreased prevalence of certain VD risk factors (as beta-blocker or amiodarone use, prior to VD, or chronic HF) [28].

Kutyfa et al. demonstrated that neither implant ICD before nor after LVAD reached significant survival benefit in a 191-patient study after 23 months of follow-up [30].

Younes et al. designed a 1444 patients study included in a waitlist HF bridged to transplant with LVAD. These authors suggested that the presence of ICD was not associated with lower mortality, cardiovascular or global mortality, or delisting although arrhythmic death was more common in the non-ICD group compared with ICD [12, 31].

A 94-patient study about the use of ICD after continuous-flow LVAD implant showed that only patients with VD before LVAD placement had any benefit [19]. Therefore, it is reasonable to not implant ICD in LVAD patients without previous VD [20].

Alvarez et al. [3] studied 487 patients, mostly with ICDS, 79.6%. The presence of ICD before LVAD was not related to a significant benefit in terms of mortality. However, ICD patients before LVAD presented complications such as episodes of shocks (31%), ventricular lead dysfunction (4.6%), and 2% of infections associated with ICD [3, 32].

The high rates of inappropriate shocks and infection events may decrease the benefits of the ICD indication [5]. Indeed, ICD shocks are associated with worse long-term outcomes and poorer quality of life and can negatively affect mental status [33]. Patients experience inappropriate shocks due to supraventricular tachycardias or electromagnetic interference; for that reason; benefit of appropriate shocks for prolonged but hemodynamically critical VD remains unclear [12].

Infection is a frequent complication in the LVAD population (i.e., driveline infection or device-related sepsis) that can occur in up to 20% of patients. Malnutrition and high comorbidity including diabetes, surgical technique, and quality of percutaneous lead care are important risk factors for infection. Aggressive management often is needed including long-term antibiotics, device exchange, or even urgent transplants [4]. Therefore, infection treatment could be more complex in the presence of both devices (LVAD and ICD).

Riaz et al. described a cohort of 215 ICD-LVAD patients and six (2.8%) developed ICD infections. Three patients had pocket infections related to device generator exchange, and three patients had ICD lead-related endocarditis due to prior LVAD-related infections. In all cases, ICD was removed along with antibiotics. Three patients with a history of previous LVAD infections received longer antimicrobial therapy, and one patient had their LVAD exchanged [32].

Finally, at the last 2019 EACTS Expert Consensus on long-term mechanical circulatory support [34] indications are: (i) patients with LVAD who develop postoperative ventricular dysrhythmias associated with hemodynamic collapse, ICD implantation is recommended (IC); (ii) patients who have an ICD implanted before LVAD, it should be kept activated for prevention of adverse effects due to right ventricular dysfunction, but its programming should be very conservative (IIaC); (iii) primary prevention is not recommended in patients without VD before LVAD (IIIC).

On the other hand, the 2022 ESC Guidelines for the management of patients with ventricular dysrhythmias and the prevention of sudden cardiac death [35] recommends: ICD implantation should be considered in LVAD recipients with symptomatic sustained VD (IIaB) Table 1.

2019 EACTS Expert Consensus on long-term mechanical circulatory support
In patients with long-term mechanical circulatory support who develop postoperative ventricular arrhythmia with hemodynamic compromise, ICD implantation is recommended.	I	C
To prevent adverse sequelae of right ventricular dysfunction, a continuation of ICD therapy should be considered.	IIa	C
Prophylactic ICD implantation in patients without arrhythmias at the time of long-term mechanical circulatory support implantation is not recommended.	III	C
2022 ESC Guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death
ICD implantation should be considered in LVAD recipients with symptomatic sustained VAs.	IIa	B

Table 1.

Summary of current guidelines about the use of implantable cardioverter defibrillator devices in the LVAD population.

6. How to program implantable cardioverter defibrillator devices

Current guidelines recommend conservative programming for ICD [34, 35, 36, 37] based on good tolerance of VD and high rates of inappropriate shocks due to atrial tachycardia (AT), including atrial fibrillation (high rates in HF patients).

Richardson et al. [10] carry out an 83-patient randomized study (well-balanced between ischemic and non-ischemic cardiomyopathy) divided into a first standard programming arm according to the treating physician (Table 2) and a second ultra-conservative ICD mode (Table 3).

VF zone active	38 patients (100%)
Detection to rate	214 bpm (200–228)
Intervals to detection	16 (12–24)
VT zone active	26 patients (68%)
Detection rate	176 (167–181)
Intervals to detection	19 (16–27)
FV/VT-2 zone active	9 patients (24%)
Detection to rate	187 (182–188)
Intervals to detection	24 (18–30)

Table 2.

Adapted from Richardson et al. shows the programming values of ICD in the standard mode.

VF: ventricular fibrillation, VT: ventricular tachycardia, bpm: beats per minute.

Manufacturer	VT zone detection	VI zone therapy	VF zone detection	VF zone therapy
Medtronic Inc.	Rate: 180 bpm 100 intervals (33 s) to detection	ATP × 5; 25 J × 2	Rate: 222 bpm 120/160 intervals to detection (32,4s)	25 J, 35 J × 5
Boston Scientific Inc.	Rate 180 bpm 30 s to detection	ATP × 8, 21 J, 41 J × 6	Rate: 220 bpm 15 s to detection	29 J, 41 J × 7
St. Jude Medical	Rate 180 bpm 100 intervals (33 s) to detection	ATP × 3, 36 J, 40 J × 2	Rate: 240 bpm 100 intervals to detection (25 s)	36 J, 40 J × 5

Table 3.

Adapted from Richardson et al. shows the programming values of the ICD in the ultraconservative mode.

ATP: anti-tachycardia pacing, J: Joules.

As it is known, detection to rate and intervals to detection are important parameters to program an ICD to achieve the highest benefit and to avoid inappropriate shocks. The first one is related to the range of heart rate where ICD acts, and intervals are the numbers of cycles or time to detect before applying therapies. As Table 1 shows, different therapy zones can be chosen, e.g., VT zone or VT1, VT2/VF zone, and VF zone based on these parameters to reach the optimal status for the patient. These zones must be correlated to patient profile, for instance, ischemic cardiomyopathy vs. myocardiopathy and primary vs. secondary prevention.

As indicated in Table 3, the ultra-conservative mode uses larger intervals of detection in the VF and VT zones and numerous anti-tachycardia pacing programming (ATP) therapies than the standard mode.

No difference related to time to the first ICD shock or the total number of shocks was found between the two groups. No statistical difference was observed in mortality terms, arrhythmic events, or heart failure hospitalization events. Inappropriate shocks resulted in 6% of the total population, although no significant differences between both groups were found [10].

However, the MADIT-RIT study showed fewer inappropriate shocks and lower all-cause mortality in those patients scheduled for ICD therapies greater than 200 bpm [38]. This study compared conventional programming versus another more conservative population with prior AT and patients without prior AT and described a statistical reduction of inappropriate shocks with conservative programming able.

Moreover, a randomized ADVANCE III trial (1902 patients) also demonstrated that larger detection intervals (30 of 40 intervals) decreased therapies delivered and inappropriate shocks without difference in mortality or arrhythmic syncope events compared with standard detection (18 of 24 intervals) in not LVAD carriers [39].

7. Resynchronization therapy on ventricular dysrhythmias after LVAD devices

Cardiac resynchronization therapy (CRT) is recommended for symptomatic patients with HF and a QRS duration ≥150 ms and left branch bundle block QRS morphology with LVEF ≤35% despite optical treatment can improve ventricular remodel to improve symptoms and reduce morbidity and mortality [1]. Nevertheless, its potential benefit or antiarrhythmic effect in LVAD patients remains unclear.

A prospective single-center study indicates a lower incidence of implantable cardioverter defibrillator device (ICD) therapies in the LVAD population if CRT mode was activated [40]. Richardson et al. [10] and Gopinathannair et al. [41] failed to find any differences in terms of mortality and hospitalization, but the first one corroborates the lower ICD discharges in the CRT-activated group, which could benefit from the antiarrhythmic effect of CRT.

The lack of additive effect from CRT in the LVAD population could be explained for several reasons including (i) LV unloading by LVAD surpasses the electrical correction by CRT; (ii) CRT population had a more advanced myocardiopathy to begin limiting any additive effect; (iii) CRT could improve clinical outcomes in younger patients with nonischemic dilated cardiomyopathy who receive LVAD as a bridge to recovery, but this population was not included. Muratsu et al. describe a case report of a 15-year-old male with acute decompensated heart failure and LVAD implantation as a bridge to recovery. Despite optimal treatment, a CRT was required to improve cardiac function and finally perform LVAD removal [41, 42].

More randomized studies are necessary to better know the mechanism and the benefit of maintaining CRT turned on in the LVAD population despite its use being associated with higher generator replacement rates [3].

8. Electromagnetic interactions between implantable cardioverter defibrillator and left ventricular assist devices

Both Heartware and HeartMate III (HM3) ventricular assist devices developed an electromagnetic interference (EMI) that can it make impossible to interrogate ICD with external programmers.

EMI can be explained by a pulse-width modulator (PWM), which can emit frequencies of 8 kHz, the same frequency that Abbot and Biotronik ICD/CRT programmers emit to start reading ICD. Distance between the LVAD and the ICD programmer and the speed of the LVAD rotor can also interfere [43]. HM3 manufacturing website itself provides information on possible incompatibilities (Table 4) [44].

Manufacturer	Family / model
Biotronik	Acticor/ Rivacor
Biotronik	Ilivia Neo/Intica Neo
Biotronik	Ilivia/Intica/Inlexa
Biotronik	Itrevia/Iperia/Inventra
Biotronik	Iforia/Ilesto/Idova
Biotronik	Lumax
ELA Medical (Sorin)	Paradyme CRT
Medtronic	Cardia CRT
Boston Scientific	ORIGEN CRT

Table 4.

Adapted from Cardiovascular Abbott. The difficulty or inability to communicate with the external programmer has been reported for or may occur with the following manufacturer families by using HM3.

Schnegg et al. [43] analyzed in vitro the programming of 24 explanted ICDs from several manufacturers in the presence of a running LVAD (Heartware and HM3 LVAD). Heartware LVAD interaction was only observed in the case of Biotronik and Microport devices if the distance was shorter than 6 cm, while HM3 LVAD only Medtronic ICD devices showed no interaction (needing up to 18 cm with some Biotronik devices).

Three different strategies were tested to improve connectivity between the ICD programmer and ICD: (1) Pseudo-faraday cage to achieve electromagnetic isolation of the ICD. The isolation of the ICD device with a metal pan (pseudo-faraday cage) was not effective in those devices that had previously failed (Boston and Biotronik devices); (2) LVAD parameters modification (above all speed of rotor). Once the speed of HM3 was significantly changed (−2000 rpm, +1000 rpm), communication improved up to 45% of total devices, but 55% remained unchanged; and (3) to increase the distance between LVAD and ICD by separating the arm from the thorax. This latter option, separating the arm from the thorax by bringing the hand over the head to increase the distance between ICD and LVAD, was the most effective measure. An increase of 3 cm was achieved. This posture should be maintained for the first 2–10 seconds, and then the arm can be repositioned. Alternatively, the ICD may be removed to the contralateral side as the last solution, but given the doubtful benefit of therapies in LVAD patients, it is not usually carried out [43].

A retrospective single-center study carried out by Yalcin et al. confirmed that the EMI from the Heartmate II LVAD was only present in patients with a St Jude/Abbott device (6 of the 23 St Jude/Abbott devices. In HM3 patients, EMI was mainly present in patients with Biotronik devices: four out of the 18 and only one patient with a Medtronic device. While initial interrogation of these devices was not successful, none of the 11 cases experienced pacing inhibition or inappropriate shocks [45].

Although, two studies reported a decrease in ventricular sensing with a smaller R-wave amplitude, an impedance decrease, and a capture threshold increase after LVAD placement that could lead to failure of ventricular dysrhythmias sensing, capturing failure, and inappropriate pacing [46, 47, 48]. It should be noted that the proximity of the ICD to the LVAD controller did not affect the programming values of the ICD or the shock therapies [43, 46, 49].

Black-Maier et al. describe several major findings in a systematic review of subcutaneous-ICD after LVAD implantation: ventricular sensing amplitudes reduction, EMI appears in primary and secondary vectors that lead to inappropriate shocks (above all in the postoperative period) but rarely in alternate vector, and parameters improved spontaneously during follow-up without need for device revision or extraction [50]. However, Lopez Gil et al. [51] describe a case of a 24-year-old man implanted with an S-ICD because of idiopathy dilated cardiomyopathy and self-limiting sustained VT. After the placement of an HM3, the S-ICD became useless because of inadequate sensing due to EMI and reduced QRS voltage.

9. Conclusion

There is not enough evidence to recommend the use of ICD in patients with LVAD because cardiac output impairment is less common, and VD is better tolerated than in the non-LVAD population.

The presence of VD and ischemic cardiomyopathy is a major risk for suffering VD events after an LVAD implant. VD usually occurs in the early period, the first month after placement, and they usually have good tolerance due to Fontan-like circulation physiology. Although, up to 45% of patients experience symptoms, and 24% required cardioversion or defibrillation. But only up to 4% suffer syncope or 2% required support with RVAD.

Current guidelines are based on expert consensus and observational studies recommend ICD in the LVAD population if concomitant postoperative ventricular dysrhythmias associated with hemodynamic collapse are present, and its programming should be very conservative to avoid inappropriate shocks due to AT.

Until ICD therapies have been more thoroughly investigated and have shown significant evidence to benefit LVAD patients, there will be resistance to deactivating ICD, particularly in patient’s bridge to transplant.

Conflict of interest

The authors have no conflicts of interest to declare.