Characterization of ventricular tachycardia ablation in end-stage heart failure patients with left ventricular assist device (CHANNELED registry)

Abstract

Aims

Patients with left ventricular assist devices (LVADs) are at high risk for ventricular tachycardia (VT), and data on VT ablation in patients with LVAD are scarce. This multicentre registry assessed the mechanism of VT, procedural parameters, and outcome of VT ablation in patients with LVAD (NCT06063811).

Methods and results

Data of patients with LVAD referred for VT ablation at nine tertiary care centres were collected retrospectively. Parameters included VT mechanisms, procedural data, VT recurrence, and mortality. Overall, 69 patients (90% male, mean age 60.7 ± 8.4 years) undergoing 72 ablation procedures were included. Most procedures were conducted after intensification of antiarrhythmic drug (AAD) treatment (18/72; 25%) or a prior combination of ≥2 AADs (31/72; 43%). Endocardial low-voltage areas were detected in all patients. The predominant VT mechanism was scar-related re-entry (76/96 VTs; 79%), and 19/96 VTs (20%) were related to the LVAD cannula. Non-inducibility of any VT was achieved in 28/72 procedures (39%). No LVAD-related complication was observed. The extent of endocardial scar was associated with VT recurrence. The median follow-up was 283 days (interquartile range 70–587 days). A total of 3/69 patients were lost to follow-up, 10/69 (14%) were transplanted, 26/69 (38%) died, and 16/69 (23%) patients were free from VT.

Conclusion

Although often a last resort, VT ablation in patients with LVAD is feasible and safe when performed in experienced centres. These patients suffer from a high scar burden, and cardiomyopathy-associated rather than cannula-related scar seems to be the dominant substrate. Ventricular tachycardia recurrence is high despite extensive treatment, and the overall prognosis is limited.

Graphical Abstract

Left ventricular assist device (LVAD), Ventricular tachycardia, Substrate characterization, End-stage heart failure, Multicentre registry

What’s new?

• This study provides new insights into ventricular tachycardia (VT) ablation in patients with left ventricular assist devices (LVADs). Despite technical challenges, acute procedural success rates were satisfactory, with no observed association between electromagnetic interference and LVAD model.

• Extensive low-voltage areas were detected in most patients, and unlike in patients without LVAD, no difference in endocardial scar burden was found between ischaemic and non-ischaemic cardiomyopathy. This suggests that disease progression and the end-stage condition may have a greater impact on myocardial substrate than cardiomyopathy type in these patients.

• Importantly, VT recurrence was primarily predicted by the extent of myocardial scar rather than cardiomyopathy type or the procedural endpoint, highlighting the critical role of scar burden in arrhythmogenesis and treatment outcomes.

Introduction

Left ventricular assist device (LVAD) implantation has emerged as a cornerstone therapy for selected end-stage heart failure patients, either as a bridge to transplantation or as destination therapy. Expanding the use of LVADs as destination therapy has led to a growing LVAD patient population.¹ Ventricular arrhythmias (VAs) are common in these end-stage heart failure patients both before and after LVAD implantation. Despite the haemodynamic support, VAs are associated with increased mortality, substantial morbidity from recurrent implantable cardioverter defibrillator (ICD) shocks, and rapid heart failure progression, especially of the unsupported right ventricle.^2,3 Regarding VA management in patients with LVAD, antiarrhythmic drugs (AADs) are mostly used as first-line therapy, but as VAs become AAD refractory, there are few options other than catheter ablation (CA).⁴ Consequently, ventricular tachycardia (VT) ablation plays an important role in the management of VAs in patients with LVAD. Alongside common challenges of VT ablation in advanced heart failure patients, technical and procedural challenges unique to VT ablation in LVAD carriers exist. Technical challenges, such as the risk of catheter entrapment, electromagnetic interference (EMI) with the surface ECG, or electroanatomic mapping (EAM) systems, add to the complexity of VT ablation in these patients.^5,6

Existing data on CA of VT in patients with LVAD are either outdated or derived from a limited number of small case series. Hence, the extent to which these challenges impact on procedural and patients’ outcomes is not well quantified. In an era of modern LVAD devices, novel high-density EAM, and image integration availability, new data are needed to better characterize the role of VT ablation in patients with LVAD. Therefore, the present multicentre registry systematically assessed substrate burden, VT mechanisms and analysed procedural parameters and outcomes of VT ablation in patients with LVAD.

Methods

This multicentre registry collected patient data from a collaboration of nine tertiary care centres in Germany and Switzerland. All patients undergoing VT ablation after LVAD implantation in the respective study centre were included retrospectively in this registry. Data were gathered from a variety of sources: electronic medical records, detailed electrophysiological procedure reports, three-dimensional (3D) mapping systems, and ICD interrogation. Institutional review boards approved the study, with patient consent obtained as per individual centre protocols. The study complies with the Declaration of Helsinki and this registry is registered on ClinicalTrials.gov (NCT06063811).

Left ventricular assist devices

In contemporary clinical practice, there are two different LVAD models available. The HeartWare ventricular assist device (HVAD™; Medtronic Inc., Dublin, Ireland) and the HeartMate 3 (Abbott, Abbott Park, IL, USA). The HVAD™ has an impeller suspended in both magnetic and hydrodynamic forces; although the distribution and sale of the HVAD™ System was discontinued in 2021, this third-generation LVAD system is still widely present in current clinical practice.⁷

The HeartMate 3 (Abbott, Abbott Park, IL, USA) device received CE mark in 2015, and Food and Drug Administration approval in 2017, and is a fully magnetically levitated centrifugal-flow third-generation LVAD. It has shown significant survival benefit over previous axial-flow pump LVADs, e.g. the HeartMate 2 (Abbott, Abbott Park, IL, USA), and is currently the only device available for implantation.⁸

Clinical ventricular tachycardia

Clinical VT was defined according to the HRS/EHRA/APHRS/LAHRS expert consensus statement on CA of ventricular arrhythmias as a spontaneously occurring VT episode leading to the patient’s symptoms or ICD therapy, previously recorded via ECG or ICD Holter.⁹

Substrate and ventricular tachycardia mapping

CARTO3® (Biosense Webster, Inc., Diamond Bar, CA, USA), EnSite™Precision or EnSite™X (Abbott, Chicago, IL, USA) was used to create substrate and/or activation maps. Regarding substrate mapping, a peak-to-peak bipolar electrogram amplitude <0.5 mV was defined as dense scar, voltage ≥0.5 and <1.5 mV as border zone. For the assessment of scar distribution and burden, the LV was subclassified in six different segments (basal, anterior, inferior, septal, lateral, and peri-cannular), and an extensive burden was defined as the presence of low voltage in ≥3 segments. A multipolar mapping catheter for high-density mapping (PentaRay®; Biosense Webster, Inc. or Advisor™HD-Grid, Abbott, Chicago, IL, USA) was applied at the operators’ discretion.^5,10 As an image integration technology, either the CARTOUNIVU™ Module (Biosense Webster, Inc.) was used to merge a fluoroscopy image with the electroanatomic map, or a pre-procedural perfusion computed tomography (CT) scan was performed, which was subsequently segmented and analysed with the inHEART system (IHU Liryc, Pessac, France). The resulting 3D model containing information about the underlying substrate was subsequently merged with the endocardial electroanatomic map (Figure 1).^10,11 Ventricular tachycardia mapping was defined as the determination of sites for ablation if diastolic potentials were noted, a post-pacing interval within 30 ms of the VT cycle length after tachycardia entrainment manoeuvre was observed and/or by identification of the critical isthmus after activation mapping.⁹ Substrate modification was defined as ablation of sites with low amplitude fractionated electrograms, long stimulus to QRS, late potentials, or the best pace map sites.^5,6,10 Primary goal of all procedures was the endpoint of non-inducibility of any VT after programmed right ventricular stimulation per individual centre protocol at the end of the procedure.

Figure 1

Combination of a CT-based 3D model created with InHEART and a propagation map of the clinical VT in a patient with NICM and LVAD. (A) The 3D model shows an extensive scar burden with substrate detected at basal, septal, apical, and inferior scar with a distinct basal and apical wall thinning. The model is derived from the arterial phase of the pre-procedural CT scan. (B) The propagation map shows the clinical VT with a basal scar-related re-entry. CT, computed tomography; 3D, three-dimensional; LVAD, left ventricular assist device; NICM, non-ischaemic cardiomyopathy; VT, ventricular tachycardia.

Statistical analysis

Data analysis was performed using SPSS Statistics (Version 27; IBM, Chicago, IL, USA). Categorical variables were presented as numbers and percentages, and continuous variables as mean with standard deviation or median (25th to 75th percentile) were appropriate. Categorical variables were compared between groups with χ² tests or Fisher’s exact tests. Student’s t-tests or Mann–Whitney U tests were used to compare continuous measures between groups. Analysis of variance (ANOVA) was used for multivariate analysis.

Results

Patient characteristics

A total of 69 patients (mean age 60.7 ± 8.4 years, 90% male) undergoing 72 VT ablation procedures between January 2018 and May 2024 were included in the study.

All patients had third-generation LVAD devices implanted, of which 42/69 (61%) HM3 and 27/69 (39%) HVAD™ devices. The underlying cardiomyopathy was ischaemic in 41/69 (59%) patients and non-ischaemic in 28/69 patients (41%) with a mean left ventricular ejection fraction of 18.4±7.1%. Additionally, at least moderate right ventricular dysfunction was present in 33% of patients. Detailed patient characteristics are present in Table 1. Echocardiographic parameters and medications are provided in Supplementary material online.

Table 1

Patient characteristics

Patient characteristics	n = 69
Male	62 (90%)
Age (years)	60.7 ± 8.4
BMI (kg/m2)	26.9 ± 84.9
Ischaemic cardiomyopathy	41 (59%)
LVEF (%)	18.4 ± 7.1
LVAD model
HeartMate 3	42 (61%)
HeartWare	27 (39%)
LVAD indication
Bridge to transplantation	32 (46%)
Bridge to candidacy	4 (6%)
Bridge to decision	2 (3%)
Destination therapy	30 (44%)
Cardiac implantable electronic devices
Patients without defibrillator	6 (9%)
Single-chamber ICD	21 (30%)
Dual-chamber ICD	12 (17%)
CRT-D	30 (44%)
Comorbidities
Atrial fibrillation	41 (59%)
Coronary artery disease	46 (67%)
Diabetes mellitus	17 (25%)
Renal failure (GFR ≤50 mL/h)	23 (33%)

BMI, body mass index; CRT-D, cardiac resynchronization therapy defibrillator; GFR, glomerular filtration rate; ICD, implantable cardioverter defibrillator; LVAD, left ventricular assist device; LVEF, left ventricular ejection fraction.

Timing and indication for ventricular tachycardia ablation after left ventricular assist device implantation

Prior to LVAD implantation, 64% of patients had experienced sustained VT and ICD shocks. After LVAD implantation, VT occurred after a median of 279 days [interquartile range (IQR): 42–862 days], with 20% of patients requiring urgent CA within <30 days post-implantation. Electrical storm (≥3 sustained VTs or ICD shocks in 24 h) was common (68%). Ventricular tachycardia ablation typically followed AAD escalation, either due to Amiodarone intensification (18/72, 25%, additional loading dose 400–600 mg for 2 weeks) or VT recurrence despite ≥2 AADs (31/72, 43%).

Cardinal symptoms of VT at emergency room admission are detailed in Table 2.

Table 2

Characteristics and symptoms of ventricular arrhythmias in patients with LVAD

Characteristics of ventricular arrhythmias	n = 69
History of ventricular arrhythmia
VT prior to LVAD implantation	44 (64%)
Time from LVAD implantation to VA occurrence (days)	279 IQR (42–862)
VA occurrence <30 days post-LVAD implantation	14 (20%)
Electrical storm post-LVAD implantation	47 (68%)
Mean number of ICD shocks post-LVAD implantation	8 (min. 0; max. 110)
Cardinal symptom at emergency room admission
No symptoms besides ICD shock	10 (15%)
LVAD alarm	12 (17%)
Syncope	6 (9%)
Light headedness	6 (9%)
Palpitations	11 (16%)
Heart failure worsening	24 (35%)

ICD, implantable cardioverter defibrillator; VA, ventricular arrhythmia; VT, ventricular tachycardia.

Procedural characteristics

All procedures were performed at tertiary care centres with a dedicated VAD team present on-site. An LVAD technician was present in 57% of procedures. The median procedure duration was 198 min (IQR 149–250 min). Most procedures were performed under deep sedation (68%), general anaesthesia was used in 32% of cases. Intraprocedural vasopressor therapy was required in 30% of cases, and ultrasound guidance was used for vascular access in 38% of procedures. Transseptal access was the preferred approach for LV entry (74%). Pericardial access was not attempted. An EAM system was used in all procedures, and high-density mapping with a multipolar catheter was performed in 74% of patients. Additionally, an inHEART model with CT-scan-derived substrate information was integrated in 5/72 procedures and in 21/72 procedures, the CARTOUNIVU™ Module was used.

The predominant ablation strategy was either a combined approach of critical isthmus ablation and substrate homogenization (56%) or, in cases with multiple VT inducible, a primarily substrate-based approach with scar homogenization and late potential ablation (40%). Irrigated radiofrequency energy was delivered at a power of 30–65 W with the non-inducibility of any VT as the primary goal. Detailed procedural characteristics are presented in Table 3.

Table 3

Procedural data

Procedural data	n = 72
LVAD technician present during procedure	41 (57%)
Anaesthesia
Sedation	49 (68%)
General anaesthesia	23 (32%)
Mapping system
Ensite Precision	5 (7%)
Ensite X	6 (8%)
CARTO	61 (85%)
Access
Ultrasound-guided vascular access	27 (38%)
LV access: retrograde	9 (13%)
LV access: transseptal	53 (74%)
LV access: combined retrograde and transseptal	10 (14%)
Predominant mapping strategy
HD-mapping with multipolar catheter	53 (74%)
Primarily activation map	34 (47%)
Primarily voltage/substrate map	38 (53%)
Entrainment based only	0
Additional entrainment mapping applied	24 (33%)
Additional pace mapping applied	39 (54%)
Predominant ablation strategy
Primarily substrate-based (scar homogenization, LAVA)	29 (40%)
Ablation exclusively during VT	3 (4%)
Combination of ablation during VT and substrate homogenization	40 (56%)
Total number of VTs targeted	96
VTs targeted per patient	1.9 ± 1.8
Procedure characteristics
Total RF time (s)	1899 IQR (1093–2798)
Fluoroscopy duration (min)	15.8 IQR (9.2–25.9)
Procedure duration (skin-to-skin, min)	197.5 (IQR 148.8–250.3)
Acute outcomes
Non-inducibility of any VT	28 (39%)
Non-inducibility of clinical VT	31 (42%)
Substrate-based due to prior non-inducibility	8 (11%)
Clinical VT still inducible	5 (7%)
Post-procedural monitoring
Intensive care	35 (49%)
Intermediate care	25 (35%)
Normal ward	12 (17%)
Days until discharge	5.5 (IQR 3–14)

LAVA, late abnormal ventricular activation; IQR, interquartile range; LV, left ventricle; LVAD, left ventricular assist device; RF, radiofrequency current; VT, ventricular tachycardia.

Electromagnetic interference

High-frequency noise on the surface ECG was observed in all patients, regardless of the LVAD model, and a pre-procedural low-pass filter adjustment was applied in all cases. Severe EMI with the 3D mapping system occurred in 41% of procedures, but only 11% of these cases resulted in incomplete map acquisition. Transient catheter visualization loss was also observed in 41%. Electromagnetic interference susceptibility was comparable between LVAD models (HM3: 12/33 vs. HVAD™: 12/25; P = 0.37). In a single case, severe EMI led to procedure abortion. Left ventricular assist device flow adaptation to mitigate EMI was not attempted in any case.

Scar distribution and ventricular tachycardia characteristics

In all patients, low-voltage areas were detected in the endocardial voltage map, and in most procedures (78%), low-voltage areas were identified in more than one segment. The distribution of endocardial scar is depicted in Figure 2. An extensive scar burden was observed in 57% of patients. Of note, in most patients with non-ischaemic cardiomyopathy (NICM), a high burden of endocardial scar was seen (Figure 1). There was no difference in endocardial scar burden between ischaemic cardiomyopathy (ICM) and NICM patients (P = 0.08; Table 4). The predominant VT morphology was a superior axis (41%) and right bundle branch block-like morphology (69%; Table 5).

Figure 2

A schematic illustration of an apical two chamber view (A) and apical parasternal short axis (B) of the left ventricle. Percentages show the detection of low voltage/scar in the respective LV segment over the study cohort. At an overall high burden of endocardial scar in end-stage heart failure patients with LVAD the illustration shows a homogenous distribution of scar with all LV segments being affected comparably. LVAD, left ventricular assist device.

Table 4

Comparison of scar burden and outcome between ICM and NICM

Comparison of ICM and NICM	ICM	NICM	P-value
Scar detected in endocardial voltage map	41/41	30/30
Number of affected LV segments (mean)	2.6 ± 1.4	3.3 ± 1.5	0.08
Extensive endocardial scar (≥3 LV segments affected)	20/42	21/30	0.08
Endpoint of non-inducibility of any VT	18/42	10/30	0.46
VT recurrence	23/38	13/28	0.26

ICM, ischaemic cardiomyopathy; IQR, interquartile range; LV, left ventricle; NICM, non-ischaemic cardiomyopathy; VT, ventricular tachycardia.

Table 5

ECG morphologies of VTs targeted during VT ablation in patients with LVAD

VT characteristics and morphology	n = 96
Inferior axis (II, III, aVF positive)	35 (37%)
Superior axis (II, III, aVF negative)	39 (41%)
Left axis deviation (I positive, II, III negative)	6 (6%)
Right axis deviation (III positive, I, II negative)	16 (17%)
Left bundle branch block morphology	15 (16%)
Right bundle branch block morphology	66 (69%)
Negative concordance	9 (9%)
Positive concordance	6 (6%)

aVF, augmented voltage foot; ECG, electrocardiogram; LVAD, left ventricular assist device; VT, ventricular tachycardia.

Overall, 96 VTs (1.9 ± 1.8 VTs per patient) were specifically mapped and targeted for ablation. Table 6 shows the reported sites of successful ablation. The underlying mechanism was determined to be a scar-related re-entry in the majority of VTs (79%). Less commonly, the VT mechanism was attributed to be a re-entry utilizing per-cannula scar around the LVAD’s inflow cannula (20%), and in one case, a Purkinje-related mechanism was identified.

Table 6

Sites of VT termination/successful ablation

Sites of successful ablation
Apical/peri-cannula	10 (19%)
Apico-septal	2 (4%)
Anterior	6 (12%)
Antero-lateral	2 (4%)
Antero-septal	3 (6%)
Inferior	8 (15%)
Infero-basal	2 (4%)
Infero-septal	3 (6%)
Infero-lateral	3 (6%)
Septal	3 (6%)
Baso-septal	2 (4%)
Lateral	4 (8%)
Baso-lateral	1 (2%)
Midcardiac vein	2 (4%)
LVOT	1 (2%)

LVOT, left ventricular outflow tract; VT, ventricular tachycardia.

Acute outcome and discharge

The endpoint of non-inducibility of any VT was achieved in 28/72 (39%) of procedures. Acute outcomes were similar in patients with ICM and NICM (Table 4). Most patients were subsequently transferred to the intensive care unit (49%) or intermediate care unit (35%) for post-procedural monitoring and were discharged from the hospital after a median of 5.5 days (IQR 3–14 days). The majority was discharged with continued oral amiodarone therapy (64%) or a combined AAD therapy with amiodarone and mexiletine (21%).

Procedural complications

Procedure-related complications were reported for 9/72 procedures (13%). Of those, the majority were access related (6/9, 67%). All but one access site–related complication occurred after conventional vascular access. One pericardial effusion occurred, and one iatrogenic aortic dissection between the right iliac artery and the renal branches occurred. One procedure-related death (1.4%) was reported: the patient suffered from a post-procedural oxygen desaturation; despite immediate intubation, the patient died later from hypoxic brain injury. No post-interventional pump thrombosis or other LVAD-associated complications were reported.

Long-term outcome and mortality

Follow-up information was available for 66/69 patients, with a median follow-up of 283 days (IQR 70–587 days). Recurrence of VT occurred in 36/66 (55%) of patients. There was no difference in VT recurrence rates (Figure 3) between NICM and ICM patients (NICM 13/28, 46% vs. ICM 23/38, 61%; logrank P = 0.19). The number of low-voltage areas and consequently the extension of left ventricular scar was associated with VT recurrence (ANOVA, P = 0.03; r² = 0.1), whereas the procedural endpoint of non-inducibility of any VT did not predict VT-free survival (ANOVA P = 0.22). A total of nine patients (14%) experienced an electrical storm during follow-up. A repeat ablation procedure was performed in three patients. During follow-up, 10/66 patients (15%) received a cardiac transplant. Overall, 26/66 patients (39%) were deceased. Of these, 9 patients (14%) died during the first 30 days after VT ablation, and a further 17/26 patients deceased during follow-up primarily from heart failure, LVAD failure, sepsis, or intracranial bleeding. None of the deaths was directly attributed to VA, but one patient died from procedure-associated hypoxic brain damage as a result of post-procedural hypoxaemia.

Figure 3

Kaplan–Meier analysis of VT-free survival after VT ablation. ICM, ischaemic cardiomyopathy; NICM, non-ischaemic cardiomyopathy; VT, ventricular tachycardia.

Discussion

Ventricular tachycardia ablation in patients with LVAD is among the most complex and challenging procedures in interventional electrophysiology. Existing data are limited, outdated, or derived from small case series.^5,6 This multicentre analysis represents the largest cohort of patients with third-generation LVADs undergoing VT ablation, and provides several important findings:

1. Ventricular tachycardia ablation is feasible and safe in experienced centres, though 30 day mortality was 14% in this critically ill patient cohort. Despite frequent LVAD-related challenges, acute procedural success was satisfactory, with no association between EMI and LVAD model.

2. Most patients with LVAD requiring VA treatment had prior VA, and scar-related re-entry was the predominant mechanism (79% of VTs).

3. A high burden of endocardial low voltage was observed. Unlike in non-LVAD patients, no difference in endocardial scar between ICM and NICM was found in patients with LVAD, possibly due to disease progression and the end-stage condition.

4. Ventricular tachycardia recurred in over 50% of patients; 14% experienced electrical storm.

5. Myocardial scar extent, rather than procedural endpoints or cardiomyopathy type, was associated with VT recurrence.

Safety

The rate of major complications was low, with most procedure-related complications attributed to vascular access, potentially reducible with wider adoption of ultrasound-guided access.¹² Although VT ablation has been identified as a risk factor for pump thrombosis, no pump thrombosis or other LVAD-related complications were observed. This may be due to strict anticoagulation management, including uninterrupted oral anticoagulation (mean INR: 2.3 ± 0.6 on the procedure day) and a target ACT ≥ 300, supporting recent EHRA consensus recommendations.¹² The overall procedural complication rate (13%) was comparable with non-LVAD VT ablation cohorts in structural heart disease,^13,14 suggesting a reasonable safety profile for VT ablation in patients with LVAD.

Thirty day mortality was 14% and most deaths occurred during the initial hospitalization after VT ablation. Of those, one procedure-related death was reported, but most patients died from terminal heart failure (7/9 patients, 78%) or non-procedure-related septic shock (1/9 patients) and LVAD failure (1/9 patients) underlining the critical condition of these patients.

Technical aspects

Technical challenges in LVAD VT ablation include EMI affecting the surface ECG and mapping system, transient catheter visualization loss, difficulty assessing inflow cannula proximity, and incomplete map acquisition.^6,10,15 Concerns about the HM3 system’s magnetically levitated impeller increasing EMI risk were not confirmed, as EMI severity was similar between HVAD™ and HM3.¹⁰ Despite observed EMI, procedural efficacy was minimally affected. While ECG filter settings were adapted, LVAD flow or turbine speed was never adjusted, and not all procedures had an LVAD technician present.

Ventricular tachycardia mechanisms and scar burden

Existing literature^6,10 suggests that, despite new potential arrhythmogenic factors, most VTs are scar-related re-entries, a finding strongly supported by this study. Among the VTs observed, 79% were scar-related re-entries, and most patients with LVAD requiring VA treatment had a prior history of VA and ICD shocks, indicating that pre-existing arrhythmic substrate was the primary contributor.

We observed a high burden of endocardial scar in EAM and CT-derived models (Figure 1), irrespective of the underlying cardiomyopathy, which contrasts with known substrate distribution in NICM patients. While extensive data exist on NICM patients without LVADs, no study has systematically analysed substrate burden and cardiomyopathy-specific differences in patients with LVAD. We hypothesize that these differences result from disease progression and the end-stage condition, potentially aggravated by implantation surgery. This aligns with Strecker et al.¹⁶ who found severely altered LV apex myocardium during LVAD implantation, regardless of the initial aetiology.

Outcome

Given the patients’ advanced disease stage and procedural complexity, VT ablation in patients with LVAD achieved a satisfactory acute outcome, with 82% non-inducible for any VT or the clinical VT, comparable with previous studies.^6,17,18 However, VT recurrence (Figure 3) was common (>50%), and 14% of patients experienced electrical storm as a manifestation of VT recurrence. All patients with electrical storm required repeat ablation, urgent transplantation, or died, highlighting its poor prognosis.^19,20

As in prior studies, non-inducibility did not predict VT-free survival.^10,21 Instead, LV scar extent correlated with recurrence, suggesting a progression driven mechanism: heart failure progression leads to further myocardial remodelling, increasing the likelihood of VT recurrence despite acute procedural success.

Despite maximal therapy, long-term prognosis remains poor, with 39% mortality during follow-up, consistent with LVAD cohorts without VT ablation.²²

Future considerations for ventricular tachycardia management in left ventricular assist device patients

Ventricular tachycardia ablation in patients with LVAD is highly complex, requiring specialized centres with multidisciplinary teams, including heart failure specialists, LVAD technicians, and electrophysiologists.^12,23 Based on present findings and current literature, the following considerations may further improve the workflow and outcomes of VT ablation in patients with LVAD:

As VTs are often haemodynamically tolerated in patients with LVAD, ICD reprogramming with a long detection time and favouring ATP over shocks may be useful.¹²

Ventricular tachycardia ablation is often a last resort after multiple ICD shocks, electrical storm, or recurrent VT despite ≥2 AADs.^12,24 Given the high mortality, impact of ICD shocks, reasonable safety profile, technological advances,²⁵ and recently demonstrated survival benefit in patients with LVAD,²⁶ earlier VT ablation may be considered.

Electromagnetic interference is an inherent challenge in VT ablation for patients with LVAD. While no differences were observed between LVAD models, reducing LVAD rotor speed, as recently suggested in the EHRA consensus statement, may further help mitigate EMI, particularly when mapping and ablating near the inflow cannula. Advanced imaging techniques, such as CT-based 3D models (Figure 1), may further enhance catheter visualization and manoeuvrability in these challenging areas.^12,23

In patients with LVAD with severely affected myocardium, distinguishing between cardiomyopathy types may be less relevant. Neither cardiomyopathy type nor non-inducibility appears to consistently predict VT recurrence, while myocardial scar extent may play a more significant role. High-density mapping, pre-procedural imaging, and advanced substrate modification, including bipolar ablation for septal scars, may help improve outcomes and should be further evaluated. To enhance safety, vascular access should be performed under ultrasound guidance.¹² Whether VT ablation before or during LVAD implantation improves efficacy and reduces VT occurrence remains unclear. The ongoing PIVATAL trial (NCT05034432) is investigating prophylactic intraoperative ablation. However, pre-LVAD VT ablation requires careful consideration, as epicardial ablation is typically not feasible post-implantation.

Limitations

The present study is a multicentre, non-randomized retrospective registry with the inherent limitations. As a multicentre registry, it includes different procedural approaches, mapping, and ablation strategies. However, it reflects the current practice in nine tertiary care centres. Although the present registry provides the largest cohort of patients with LVAD undergoing VT ablation, given this specific subset of patients studied, the overall sample size remains small. This precludes advanced statistical analysis. Furthermore, all considerations regarding future improvements of VA treatment in patients with LVAD should be viewed as explorative and hypothesis generating. Prospective randomized trials are needed to determine the optimal timing of VT ablation, optimal ablation strategies, endpoints, and impact of VT ablation on quality of life and mortality in end-stage heart failure patients with LVAD.

Conclusions

Catheter ablation of VT in patients with LVAD is feasible but should be performed only in experienced centres due to its complexity. Irrespective of the underlying cardiomyopathy, patients with LVAD have a high burden of endocardial scar. Intrinsic myocardial scar rather than LVAD-associated factors seems to be the dominant arrhythmogenic mechanism. The extent of endocardial scar was associated with VT recurrence. The recurrence of VT after ablation remains high despite maximal therapeutic efforts, and the prognosis of these patients is poor.

Supplementary material

Supplementary material is available at Europace online.

Funding

None declared.

Data availability

Data are available upon reasonable request made with the corresponding author.

Author notes

Conflict of interest: J.-H.v.d.B. reports having received lecture fees from Johnson & Johnson and Abbott outside the submitted work. E.H. received travel grants from Bayer, Edwards LifeSciences, Medtronic, and Pfizer outside the submitted work. D.D. is an editor of EP Europace and was not involved in the peer review process or publication decision. D.D. received lecture honorary, travel grants, and/or a fellowship grant from Abbott, Astra Zeneca, Bayer, Biotronik, Boehringer Ingelheim, Boston Scientific, Bristol Myers Squibb, Medtronic, Microport, Pfizer, Sanofi, and Zoll, outside the submitted work. K.S. discloses lecture fees from Bristol Myers Squibb, Bayer, Astra Zeneca and consultant fees from Bristol Myers Squibb, and Bayer. J.W. reports having received lecture fees from Abbott and Boston Scientific and educational fees from Boston Scientific and Johnson & Johnson. A.S. reports having received lecture fees from Medtronic, Boston Scientific, Abbott, and Johnson & Johnson. J.-H.S. reports having received educational fees from Boston Scientific and Johnson & Johnson and lecture fees from Abbott. L.E. is an editorial consultant of EP Europace and was not involved in the peer review process or publication decision: L.E. discloses consultant fees, speaking honoraria, and travel expenses from Abbott, Bayer Healthcare, Johnson & Johnson, Biotronik, Boehringer, Boston Scientific, Bristol Myers Squibb, Daiichi Sankyo, Medtronic, Pfizer, and Sanofi Aventis. Research has been supported by German Research Foundation (DFG) and German Heart Foundation outside the submitted work. T.R. reports research grants from the Swiss National Science Foundation, the Swiss Heart Foundation, the sitem-insel support funds, Biotronik, Boston Scientific, and Medtronic, all for work outside the submitted study. Speaker/consulting honoraria or travel support from Abbott/SJM, Bayer, Johnson & Johnson, Biotronik, Boston Scientific, Farapulse, Medtronic, Pfizer-BMS, all for work outside the submitted study. Support for his institution’s fellowship program from Abbott, Johnson & Johnson, Biotronik, Boston Scientific, and Medtronic for work outside the submitted study. P.S. is an Editor of EP Europace and was not involved in the peer review process or publication decision. D.S. discloses lecture fees from Abbott, Boston Scientific, Johnson & Johnson, Pfizer, research grants from Abbott, Johnson & Johnson and consultant fees from Edwards and Abbott. J.L. discloses lecture fees from Abbott, Boston Scientific, Johnson & Johnson and consulting fees from Boston Scientific. All other authors report nothing to declare.