An unmet need: arrhythmia detection by implantable loop recorder in the systemic right ventricle

Abstract

Aims

Patients with systemic right ventricles are at high risk of sudden cardiac death. Arrhythmia is a significant risk factor. Routine Holter monitoring is opportunistic with poor adherence. The aim of this study was to determine if continuous rhythm monitoring with an implantable loop recorder (ILR) could allow early detection of clinically important arrhythmias.

Methods and results

Implantable loop recorder implantation was offered to patients with atrial switch repair for transposition of the great arteries. Recordings were made with symptoms or, automatically for pauses, significant bradycardia or tachycardia and reviewed by the multi-disciplinary team. Twenty-four out of 36 eligible patients underwent ILR implantation with no complication. Forty-two per cent had preserved ventricular function, 75% were NYHA functional class I, 88% had low sudden cardiac death risk, 33% had previous intra-atrial re-entrant tachycardia (IART), and none had known conduction disease. Eighteen out of 24 (75%) patients made 52 recordings (52% automated) over 39.5 months (1.6–72.5). Thirty-two out of 52 (62%) recordings in 15/24 (63%) of the cohort were clinically significant and included sinus node disease (two patients), atrioventricular block (two patients), IART (seven patients), and IART with sinus node disease or atrioventricular block (four patients). Implantable loop recorder recordings prompted medication change in 11 patients [beta-blockers (n = 9), anti-coagulation (n = 5), and stopping anti-coagulation (n = 1)] and device therapy recommendation in seven patients [five pacemakers (three: atrioventricular block) and two defibrillators]. Two patients declined intervention; one suffered an arrhythmic death. Intra-atrial re-entrant tachycardia and clinically relevant conduction disease were detected in patients irrespective of sudden cardiac death risk.

Conclusion

Continuous monitoring with an ILR in patients with systemic right ventricle following atrial switch detects clinically relevant arrhythmias that impact decision-making. In this cohort, clinically relevant arrhythmias did not correlate with sudden cardiac death risk.

Graphical Abstract

Graphical Abstract.

What’s new?

Adults with atrial switch repair for transposition of the great arteries are 50 times more likely to experience sudden cardiac death than the general population. Arrhythmia contributes significantly to sudden death and incipient heart failure although mechanisms are unclear.

• Implantable loop recorders (ILRs) offer a convenient, minimally invasive strategy for continuous rhythm monitoring to enable earlier detection and intervention than conventional electrocardiogram and Holter screening.

• In this low-risk cohort of adults with systemic right ventricle after atrial switch, 66% had clinically significant arrhythmia detected by ILR (atrial tachyarrhythmia and/or conduction disease), which changed management in 63% (medication, pacemakers, and defibrillators). Conventional surveillance and risk scoring would not have identified these issues in the same timeframe.

• The prognostic impact of early intervention is unclear; however, continuous monitoring with ILR in this population is feasible and safe and impacts clinical decision-making. Further studies should determine prognostic benefit and cost-effectiveness.

Introduction

Sudden cardiac death (SCD) is a serious concern for adults following atrial switch for patients with transposition of the great arteries (TGA), with an incidence between 1.7 and 9.5 per 1000 patient-years.¹ This represents a 50 times increased risk compared with the age-matched general population² and 10 times increased risk compared with young people with hypertrophic cardiomyopathy.³ After heart failure, sudden death is the second most common cause of mortality associated with this physiology⁴ and an unresolved issue despite advances in surgical and heart failure management. A better strategy to identify patients at risk of SCD is of paramount importance particularly with the predicted growth in heart failure and arrhythmia within this ageing population. The precise mechanism of SCD is often unclear¹; however, arrhythmias, particularly intra-atrial re-entrant tachycardia (IART), are amongst the strongest identified risk factor.^5–7 Early recognition and treatment of IART may preserve cardiac output, avoid myocardial ischaemia, reduce the risk of ventricular arrhythmias, and prompt anti-coagulation to reduce the considerable risk of thromboembolism.^1,6,8,9 It remains unclear whether conduction disease contributes to SCD within this physiology although it is an established risk factor within the structurally normal heart.^10,11 A recently developed risk score is promising in identifying those at greatest risk who may benefit from a primary prevention implantable cardiac defibrillator (ICD) but did not find conduction disease to be an independent predictor.⁴ However, arrhythmias including IART and conduction disease may carry a significant burden of morbidity and potential mortality irrespective of risk scores.^5–7

The current strategy of rhythm surveillance with electrocardiograms (ECGs) and intermittent ambulatory monitoring has significant limitations related to its short temporal coverage. It frequently fails to detect paroxysmal IART,¹² offers a narrow window for symptom–rhythm correlation, and has poor adherence within this population (see Supplementary material online, Table S1). Furthermore, there is a high burden of care to the patient and healthcare provider in delivering this sub-optimal monitoring strategy that is expected to grow with ageing of the adult congenital heart disease (ACHD) population. The current generation of the injectable implantable loop recorder (ILR) presents a convenient alternative for continuous rhythm monitoring. Limited data on symptom-driven application in unselected congenital heart disease have shown good tolerance and a high detection of atrial and ventricular arrhythmias with clinical implications.^13,14 Implantable loop recorders have also been shown to have benefit in empirical arrhythmia surveillance in other high-risk populations such as Fabry’s disease and hypertrophic cardiomyopathy.^15,16

The aim of this study was to prospectively assess the feasibility and utility of arrhythmia detection with ILR in a single-centre population of adults with systemic right ventricle following atrial switch.

Methods

Study population

Since 2016, all adults (>18 years) with a bi-ventricular circulation and systemic right ventricle following atrial switch repair (Mustard or Senning procedure) for TGA attending the ACHD clinic at Newcastle upon Tyne NHS Foundation Trust were considered for an empirical ILR implant. Those who had subsequent heart transplant, ventricular assist device (VAD) implantation, or surgical restoration of a systemic left ventricle (i.e. Rastelli procedure and arterial switch) were excluded. Final screening also excluded those with a pre-existing intra-cardiac monitoring device or who were deemed excessively frail by their responsible clinician (Figure 1). Thirty-six patients were identified, and their responsible clinician discussed the rationale for ILR implantation at their routine clinical review. Twenty-four patients agreed to ILR implantation (Figure 1). All patients subsequently agreed to automatic, elective replacement of their ILR at the end of battery life.

Figure 1.

Patient selection.

Data collection

Demographic, anatomical, and arrhythmia history, arrhythmia (symptomatic or automatic) during follow-up, clinical action following recordings, and clinical outcomes following intervention were recorded. Survival outcome at the end of the study was recorded as alive, transplant, transplant assessment, VAD, or death.

Implantable loop recorder implantation, arrhythmia detection, and management

Implantation was conducted under local anaesthetic in the usual fashion.¹⁷ One patient underwent Medtronic Reveal XT® implantation with all subsequent patients having an injectable Medtronic Reveal Linq® in the left parasternal position. Patients were instructed to activate a recording if they experienced palpitations, dizziness, or syncope. Automated recordings were made for brady-arrhythmias, tachy-arrhythmias, or atrial fibrillation. Brady-arrhythmia detection was set for sinus pauses of greater than 3 s or bradycardia less of than 30 bpm for 12 consecutive beats. Tachy-arrhythmias were set to detect a suspected supra-ventricular tachycardia of greater than 194 bpm for 16 beats.

All recordings were transmitted remotely and reviewed by a cardiac physiologist within 1 day. The patient’s responsible clinician was informed in the event of suspected arrhythmia and management agreed in the ACHD and electrophysiology multi-disciplinary team meeting (ACHD-EP MDT). Recommendations for intervention included review of consensus guidelines,¹⁸ expert opinion, and, following publication, calculation of the risk of SCD using the published scoring system utilizing age, heart failure, syncope, right ventricular function, left ventricular outflow tract obstruction, and QRS duration.⁴

Ethical considerations and data protection

The decision to offer ILR implantation was a clinical consensus reached in the ACHD-EP MDT after an audit of the existing practice of Holter monitoring confirmed poor adherence and thus limited arrhythmia detection. Written informed consent was taken from patients prior to ILR implantation. This study represents a re-audit following a change in clinical practice and as such did not require separate ethical approval. The audit was registered with the clinical effectiveness database of The Newcastle upon Tyne NHS Foundation Trust. Data were managed in accordance with General Data Protection Regulations.

Patient and public involvement

This study reports the results of a change in clinical practice that was instigated after an audit of the recommended arrhythmia monitoring strategy for this patient group identified poor adherence. Whilst patients and the public were not directly involved in the design of this study, the results have led to subsequent engagement with the Somerville Foundation and patient focus groups funded by the British Heart Foundation and led by the senior author to understand what issues matter to patients with respect to monitoring and how this can help inform future study design.

Statistical methods

Continuous data were summarized as median and range (following confirmation of non-normality with the Shapiro–Wilk test) and categorical data as count and percentage of the whole cohort. Freedom from arrhythmia was described with the Kaplan–Meier method. Comparison of different arrhythmia events was compared with a two-tailed logrank test. Group analysis was conducted for five variables (right ventricular function, left ventricular function, tricuspid regurgitation severity, type of surgical repair, and previous arrhythmia) for dependence on survival outcome (death, transplant, or VAD), detected arrhythmia and future intervention using Fisher’s exact test. Ventricular function and tricuspid regurgitation were grouped as normal or mild and moderate or severe as assessed by echocardiography. A P value of <0.05 was considered significant (Figure 2).

Figure 2.

Freedom from ILR detected arrhythmia from time of implant. Patients censored at time of device battery depletion or date of study analysis (whichever was latest).

Analysis was conducted in GraphPad Prism® 9.5 (GraphPad Software LLC, 2022).

Results

Study population

Twenty-four patients (29–49 years) underwent ILR implantation. Half had residual cardiac lesions and 46% had additional non-cardiovascular co-morbidities. Seventy-five per cent were NYHA functional class 1. Forty-two per cent had moderate or severe systemic right ventricular impairment, and 30% had moderate or severe tricuspid regurgitation. All but one patient had baseline sinus rhythm with a third having previously documented IART and a third taking regular beta-blocker therapy. None were taking a class I or III anti-arrhythmic agent. Demographics of the study population are shown in Table 1. Most patients, 88%, had a low predicted risk of SCD (less than 5%).

Table 1.

Study population

Demographic	Median (range)	Number (%)
Age (years)	35 (29–49)
Male		18 (75)
Mustard repair		13 (54)
Senning repair		11 (46)
Left ventricular outflow tract obstruction		4 (17)
Ventricular septal defect		5 (21)
Baffle stenosis/occlusion		5 (21)
Single coronary		2 (8)
Pulmonary hypertension (known)		2 (8)
Straddling tricuspid valve		1 (4)
Epilepsy/prior stroke		5 (20)
Thyroid disease		2 (8)
Type 2 diabetes mellitus		1 (4)
Depression		1 (4)
Moderately/severely impaired right ventricular function		10 (42)
Moderately/severely impaired left ventricular function		2 (8)
Moderate/severe tricuspid regurgitation		7 (29)
NYHA 1		18 (75)
NYHA 2/3		6 (25)
Sinus rhythm		23 (96)
Junctional rhythm		1 (4)
QRS duration (>120 ms)		4 (17)
Previously documented IART [ablated]		8 (33) [2(8)]
Previously documented NSVT		1 (4)
Previously documented sinus pause (<6 s)		1 (4)
No medication		16 (66)
Beta-blocker		8 (33)

IART, intra-atrial re-entrant tachycardia; NSVT, non-sustained ventricular tachycardia.

Implantable loop recorder implantation, arrhythmia detection, and management

All ILR implantations were conducted without procedural complication, and all had interpretable electrograms. Two patients had a second ILR implant following the battery depletion of their first device. There were no complications related to the devices during the surveillance period. The median duration of surveillance was 39.5 months (1.6–72.5) during which a total of 52 recordings (52% automated and 48% activated) were made in 18 patients (Figure 2). This equated to a median of 2.8 recordings per patient (range 0–8) or 0.5 recordings per patient-year (range 0–14.5).

Twenty-five per cent (6/24) of the cohort had no arrhythmia detected during the monitoring period. Anti-coagulation was stopped in one individual based on multiple sinus rhythm recordings that correlated with symptoms and patient discussion of the risk (Table 2).

Table 2.

Characteristics and clinical outcomes of ILR recordings

Recording	Number (%)	Clinical management (number)	Current clinical status
No recording	6 (25)	—	All alive and well
Insignificant recording	3 (13)
Sinus rhythm	2 (8)	Reassurance (1)	All alive and well
		Anti-coagulation stopped (1)
Short sinus pause (<6 s)a	1 (4)	Observation (1)	Alive and well
Conduction disease	4 (17)
Long sinus pause (>6 s)	2 (8)	Exercise test and observation (1)	Alive and well
		PPM (1)	Alive and well
Atrioventricular block	2 (8)	PPM (1)	Alive. transplant assessment
		ICD (1)	Alive and well
Atrial arrhythmia	11 (46)
Non-sustained IART b	2 (8)	Observation (1)	Alive and well
		Beta-blocker (1)	Further runs of non-sustained IART
Sustained IARTc	5 (21)	Beta-blocker (1)	Further runs of non-sustained IART
		Beta-blocker, anti-coagulation (1)
		Beta-blocker, declined primary prevention ICD (1)	Further runs of non-sustained IART
		Beta-blocker, anti-coagulation, ablation (1)	Heart failure, IART, cardiac arrest, death
		Beta-blocker, anti-coagulation, awaiting ablation (1)
			Heart failure, ventricular assist device, tricuspid valve replacement
			Alive, transplant assessment
Tachy–brady-arrhythmiad	4 (17)	Beta-blocker, declined PPM (1)	Alive and well
Sinus node disease   Atrioventricular block	2 (8)	DCCV, beta-blocker, anti-coagulation, awaiting ablation (1)	Alive. Further admission with IART expediting ablation
	2 (8)
		PPM, anti-coagulation, ablation (1)	Alive and well
		Beta-blocker, PPM (1)	Alive and well

IART, intra-atrial re-entrant tachycardia; ICD, implantable cardiac defibrillator; DCCV, direct current cardioversion; PPM, permanent pacemaker.

^aLimited literature and guidelines suggest potential prognostic implication of shorter pauses in this circulation.¹⁹

^bRegular narrow complex tachycardia lasting less than 30 s.

^cRegular narrow complex tachycardia lasting 30 s or greater.

^dIART and significant sinus or AV node disease recorded within the same patient.

61.5% of the total recordings made in the remaining patients were clinically significant (significant sinus pause, atrioventricular block, or IART). Implantable loop recorder detected rhythms directly changed clinical management in 15/24 patients (64%) (Table 2).

IART was detected in 46% (11/24) of the cohort with half (5 of 11 patients) having no previously documented atrial tachyarrhythmia. Anti-coagulation was newly commenced in five patients.

Significant conduction disease was detected in 33% (8/24), 50% co-existing with IART, with one patient having a previously documented short sinus pause. Six patients had an indication for pacing; five agreed to proceed with this intervention. One patient, with sinus pauses and non-sustained IART, preferred an initial conservative approach with ongoing ILR surveillance rather than pacing. At 1072-day post-ILR implant, they experienced further pauses and atrioventricular block and then agreed to pacemaker implantation. Three patients with tachy–brady-arrhythmia were taking a beta-blocker, which was continued following permanent pacing. None of the patients with atrioventricular block were taking a beta-blocker.

Conduction disease and IART were detected over the entire surveillance period and appeared unrelated to each other in onset.

An additional patient, with pulmonary hypertension and recurrent IART, had a high predicted risk of SCD and was offered a primary prevention ICD. The patient declined the ICD, preferring a conservative approach, and subsequently had a sustained episode of symptomatic IART resulting in a pulseless electrical activity cardiac arrest shortly after presenting to the hospital. Unfortunately, the patient was not a candidate for advanced heart failure therapies, and resuscitation was unsuccessful.

There was no association between the type of surgical repair, right ventricular function, or tricuspid regurgitation severity with detected arrhythmia, adverse outcome, or intervention. In this cohort, those with previously documented IART were not more likely to have further IART compared with those without a documented arrhythmia history (P = 0.08). Moderate or severely impaired left ventricular systolic function was associated with an adverse survival outcome but not with detected arrhythmia. Intra-atrial re-entrant tachycardia and conduction disease were detected in patients despite low sudden death risk scores (Figure 3).

Figure 3.

Implantable loop recorder detected arrhythmia compared with the calculated risk of SCD.⁴ IART, intra-atrial re-entrant tachycardia; SCD, sudden cardiac death.

Three patients had incidental ambulatory ECG monitors during ILR surveillance. None detected any arrhythmia whereas their ILR detected IART in all patients and significant conduction disease requiring eventual permanent pacing in one patient.

Discussion

Empirical ILR implantation in patients with a systemic right ventricle following atrial switch repair for TGA detected a high burden of arrhythmia, which resulted in a change in clinical management in the majority of patients. This is despite our cohort representing a healthy subpopulation with predominately preserved systemic right ventricular function, near normal NYHA functional class, and low predicted risk of SCD (Table 1).

Most arrhythmias were IART with half occurring in patients without a recognized history of arrhythmia. The incidence is consistent with previous reports in cohorts with a heterogenous risk of SCD (38% compared with 14–44%) utilizing intermittent surveillance and symptom-based monitoring,^4,5 as well as ILR studies in unselected populations with congenital heart disease (30–53%).^13,14 As atrial arrhythmias are typically progressive in this population, it is possible that these arrhythmias would have eventually been detected by conventional monitoring. There is, however, a significant potential clinical benefit to the early detection and intervention afforded by an ILR surveillance strategy. This is of particular importance given the strong relationship between IART with heart failure and SCD in this population.^1,5–7 The pathophysiological mechanisms are incompletely understood; however, it is increasingly recognized that the association is robust and early intervention can mitigate adverse outcomes and improve quality of life.²⁰

Strategies to maintain sinus rhythm are limited with significant adverse effects experienced from the few anti-arrhythmic agents suitable for this population.²¹ Our unit, in line with others,¹⁹ prefers early ablation as the strategy to manage sustained IART. However, recurrence is frequently observed and the prognostic value of repeated interventions, the acceptable burden of paroxysmal arrhythmias, and the threshold to pursue advanced heart failure therapies influence clinical decision-making.^20–22

Conduction abnormalities, predominantly sinus node disease, are common following atrial switch and may promote IART, reduce cardiac output, and potentially contribute to SCD.¹⁸ Consequently, there is an accepted lower threshold for pacing¹⁸; however, patient selection and approach remain challenging given the high incidence of device-related complications.²³ Although a significant proportion of our screened cohort had a pre-existing intra-cardiac pacing device (32%), an additional 11% of patients were identified to have a pacing indication over the course of the study. Of note, half of these patients (three out of six) had daytime atrioventricular block, without reversible trigger, which presents a clear risk factor for SCD. In TGA with concordant atrioventricular connection, in contrast to congenitally corrected transposition, disruption of the atrioventricular node conduction is unusual and the mechanism for block is unclear. It is possible that accelerated fibrosis following initial surgical repair and subsequent structural or electrophysiological procedural interventions are contributory. In addition, diseased conduction tissue may be particularly vulnerable to sudden changes in depolarization that occur with IART and/or sinus pauses and result in prolonged tissue refractoriness resulting in atrioventricular block (type 3 and type 4 blocks).²⁴

Conduction abnormalities were not independently associated with SCD in a recent large retrospective analysis,⁴ although the cohort was limited to intermittent rhythm surveillance with ECGs and Holter monitoring. Increasing ventricular conduction delay (QRS duration) and repolarization instability (QTc dispersion) over time can predict SCD across a spectrum of ACHD disease.²⁵ Whether this association is due to an elevated risk of atrioventricular block, re-entrant tachycardia, or purely worsening heart failure is unclear. The correlation of sequential ECG and echo studies with continuous rhythm monitoring could offer important mechanistic insight into arrhythmogenesis and SCD, further guiding timely intervention, and is an important area for further study. The only death, thus far, in our monitored cohort was associated with IART followed by low cardiac output circulatory failure.

The alternative strategy of clinic ECGs and intermittent ambulatory ECG recordings capture short, unselected periods¹² with poor adherence within this population (see Supplementary material online, Table S1) and do not reliably detect clinically relevant arrhythmias. By comparison, ILRs were well tolerated and had a high rate of clinically significant findings. Given these findings, alongside the added value of prospective symptom–rhythm correlation and response to interventions, our unit currently recommends lifelong ILR surveillance in this population who do not have a pre-existing cardiac monitoring device. It is hoped that this will improve the detection and understanding of clinically relevant arrhythmias with the potential to reduce the burden of SCD. How those with ACHD experience monitoring and surveillance is not fully understood and is an important component that requires further evaluation if care is to be delivered holistically as well as effectively from a clinical perspective. At this stage, costs are not fully evaluated. The incremental costs and benefits of ILR surveillance for this small cohort should also be considered in terms of their expected lifetime healthcare costs and relative to the whole cohort of patients referred for ILR implantation. To accurately assess cost-effectiveness, economic evaluation should be incorporated into a prospective clinical study that compares this approach to usual care. Finally, we should consider the rapidly increasing diagnostic utility of wearable smart watches that is already informing clinical decision-making in many clinical fields.²⁶ Future studies are required to investigate their use in ACHD to enable a more patient-centred approach to surveillance and potentially reduce the care burden for patients and healthcare providers.²⁷

Limitations

This was a single-centre study not sufficiently powered for robust subgroup analysis. Nevertheless, the clinical event rate was sufficiently high to be confident about the likely benefit of this approach over routine and ad hoc monitoring based on symptoms and risk scoring. In fact, it clearly identifies a concern over the conventional monitoring approach in ACHD guidelines in missing clinically relevant arrhythmias in low-risk individuals.

Because ambulatory monitoring had poor adherence in our cohort and in view of the high rate of SCD, randomizing patients to this option was felt not to be clinically justifiable, and therefore a control group remaining in usual care was not included.

We also acknowledge an inherent selection bias given a third of potentially eligible patients refused or were not offered ILR implantation. However, based on these results, we have subsequently re-approached this omitted group, and the majority have agreed to future ILR surveillance.

Conclusion

Continuous rhythm monitoring with an ILR in patients with systemic right ventricle following atrial switch is well tolerated and detects a high burden of clinically relevant tachy- and brady-arrhythmias that impact decision-making. The existing SCD risk scoring system did not predict arrhythmias detection in this cohort, and the long-term significance of these findings is unclear.

Supplementary material

Supplementary material is available at Europace online.

Funding

There was no funding for this study.

Data availability

Data will be supplied by direct request to the authors.