Sudden cardiac death in congenital heart disease

Abstract

Sudden cardiac death (SCD) accounts for up to 25% of deaths in patients with congenital heart disease (CHD). To date, research has largely been driven by observational studies and real-world experience. Drawbacks include varying definitions, incomplete taxonomy that considers SCD as a unitary diagnosis as opposed to a terminal event with diverse causes, inconsistent outcome ascertainment, and limited data granularity. Notwithstanding these constraints, identified higher-risk substrates include tetralogy of Fallot, transposition of the great arteries, cyanotic heart disease, Ebstein anomaly, and Fontan circulation. Without autopsies, it is often impossible to distinguish SCD from non-cardiac sudden deaths. Asystole and pulseless electrical activity account for a high proportion of SCDs, particularly in patients with heart failure. High-quality cardiopulmonary resuscitation is essential to improve outcomes. Pulmonary hypertension and CHD complexity are associated with lower likelihood of successful resuscitation. Risk stratification for primary prevention implantable cardioverter-defibrillators (ICDs) should consider the probability of SCD due to a shockable rhythm, competing causes of mortality, complications of ICD therapy, and associated costs. Risk scores to better estimate probabilities of SCD and CHD-specific guidelines and consensus-based recommendations have been proposed. The subcutaneous ICD has emerged as an attractive alternative to transvenous systems in those with vascular access limitations, prior device infections, intra-cardiac shunts, or a Fontan circulation. Further improving SCD-related outcomes will require a multidimensional approach to research that addresses disease processes and triggers, taxonomy to better reflect underlying pathophysiology, high-risk features, early warning signs, access to high-quality cardiopulmonary resuscitation and specialized care, and preventive therapies tailored to underlying mechanisms.

Graphical Abstract

Multidimensional aspects of SCD in CHD. Maximizing outcomes related to SCD in patients with CHD requires comprehensively addressing the many facets involved (blue ellipses). Three key aspects for each dimension are indicated in adjoining boxes. SCD, sudden cardiac death; CHD, congenital heart disease; CPR, cardiopulmonary resuscitation; ICD, implantable cardioverter-defibrillator; S-ICD, subcutaneous implantable cardioverter-defibrillator.

Sudden cardiac death, Cardiac arrest, Congenital heart disease, Implantable cardioverter-defibrillator, Risk stratification

Introduction

Sudden cardiac death (SCD) in a young patient is a devastating event that has a profound impact on families and communities. It is among the leading causes of mortality in patients with congenital heart disease (CHD), accounting for up to 25% of all deaths.¹ Our highest goal is to prevent these catastrophic occurrences. Doing so requires answering critical questions: Who is affected? What are the underlying pathophysiologies and triggers? Are there early warning signs? What therapies are most effective? In addressing these and other fundamental queries, the knowledge acquired over the past few decades has primarily been derived from an epidemiological vantage point. While this perspective has been essential in laying the foundation for our understanding of SCD, shortcomings should be recognized and addressed in order to effectively advance the field. Herein, we provide a critical overview of the current state of knowledge on SCD in CHD across a range of topics including epidemiology, pathophysiology, associated factors, autopsy evidence, acute resuscitation, risk stratification, and implantable cardioverter-defibrillator (ICD) indications and challenges. Knowledge gaps that present avenues for future research are summarized.

Epidemiology of sudden cardiac death in congenital heart disease

Congenital heart disease refers to structural malformations of the heart and/or great vessels that are present at birth, and excludes diagnoses such as accessory pathways in structurally normal hearts and inherited arrhythmia syndromes and cardiomyopathies. Accurate portrayal of the epidemiology of SCD in CHD is obscured by limitations such as geographical variations in how SCD is reported, the potential for death certificates to overestimate SCD by misclassification (e.g. gastrointestinal bleeds, strokes, and pulmonary emboli), lack of autopsies, and inaccuracies in CHD diagnostic codes and denominators. Most studies are from tertiary care centres and are subject to referral bias. They only include cohorts with known CHD before death, whereas SCD may occur in patients with undiagnosed CHD prior to death.² Notwithstanding these limitations, SCD in patients with CHD appears to be a relatively rare event at the general population level, with a yearly incidence rate between 0.07 and 0.40 per 100 000 person-years.^2,3

As noted in Table 1, among CHD cohorts that differed according to age ranges and other criteria, such as the inclusion of non-repaired CHD, 0.28–2.7% succumbed to SCD (or had an aborted SCD) each year.^2,4–7 By historical comparison, these rates are at least 20–30-fold higher than the general population. Identification of the highest risk lesions is not clear-cut, in part due to inter-study variability. Nevertheless, there are some consistencies across studies. The rate of SCD in patients with tetralogy of Fallot ranges from 0.9 to 1.5% per year. Other identified high-risk substrates include transposition of the great arteries (TGA), particularly with atrial switch surgery or congenitally corrected TGA, cyanotic heart disease (both Eisenmenger and non-Eisenmenger), Ebstein anomaly, and a Fontan circulation. Although the incidence of SCD is higher in adults than children, it is unclear whether risk continues to increase linearly with age, plateaus, or exhibits a U-shaped pattern after peaking in the mid-30s to 40s. There is some evidence to suggest that the incidence of SCD is highest in the mid-30s in patients with complex CHD and at an older age, i.e. after 50 years, in those with simple forms of CHD.^4,7,8

Table 1

Incidence of sudden cardiac death per 1000 patient-years

	Silka et al.4	Gallego et al.5	van der Velde et al.6	Moore et al.7	Lynge et al.2
Study design	Single centre	Single centre; includes aborted SCD	Muticentre	Single centre	Nationwide
Study period	1958–1996	1990–2010	2001–2004	2000–2015	2000–2009
Study population	Repaired CHD	Repaired CHD ≥18 years	Repaired and unrepaired CHD	Repaired and unrepaired CHD ≥16 years	Repaired and unrepaired CHD aged 0–35 years
Overall	0.9	2.6	2.7	0.4	0.28
Repaired tetralogy of Fallot	1.5	1.4	1.3	1.0	0.9
Complete TGA	4.9	9.5	4.6	*	1.7
Atrial switch	*	*	3.7	2.4	*
Arterial switch	*	*	5.4	*	*
Congenitally corrected TGA	*	25	1.8	2.1	*
Cyanotic heart disease	*	5.4	16.6	*	*
Eisenmenger syndrome	*	*	17.3	4.8	*
Non-Eisenmenger	*	5.4	15.5	*	*
Fontan circulation	*	2.8	4.4	2.1	1.9
Left-sided lesions	*	*	1.4	*	*
Ebstein anomaly	*	*	5.1	0.7	5.0
Ventricular septal defect	0.2	3.6	*	0.2	0.06
Aortic coarctation	1.3	2.1	*	0.4	0.5
Atrioventricular septal defect	0.9	1.8	*	1.8	1.0
Aortic stenosis	5.4	0	*	0.0	*
Pulmonary stenosis	0.3	0	*	*	0.3
Patent ductus arteriosus	0.0	0	*	*	*
Atrial septal defect	0.0	0.0	1.0	0.2	*

SCD, sudden cardiac death; TGA, transposition of the great arteries.

*

Not available.

Pathophysiology of sudden death in congenital heart disease

Sudden cardiac death has traditionally been defined as an unexpected non-traumatic death due to a cardiac cause that occurs within 1 h of the onset of new or worsening symptoms. It requires the absence of preceding haemodynamic deterioration and includes those who die during sleep or are found dead within 24 h of last being seen alive and stable. It is not a single disease but, rather, a terminal outcome common to several pathophysiological processes and triggers. Underlying lesions include haemodynamic abnormalities, myocardial ischaemia, fibrosis, and/or hypertrophy, and primary arrhythmias, i.e. tachyarrhythmias and atrioventricular conduction block. Triggers for SCD can be diverse and encompass toxic, autonomic, metabolic, neurohormonal, and inflammatory factors.

Important to the discussion of primary prevention ICD, the terminable rhythm is not always a shockable ventricular arrhythmia. In fact, trends indicate that asystole and pulseless electrical activity, when combined, have surpassed ventricular arrhythmias as the more common presenting rhythm among SCD victims in the general population.⁹ Furthermore, without autopsies, histological analyses, and toxicological studies, it is often impossible to distinguish SCD from non-cardiac sudden deaths, which account for 40% of all sudden deaths in the general population.¹⁰ Taken together, general population data indicate that <30% of sudden deaths are due to shockable ventricular arrhythmias. Whether this proportion is higher in the CHD population remains uncertain given the scarcity of data.^11,12 An analysis of 213 sudden death victims with CHD revealed that at least 20% were non-arrhythmic (i.e. 9% aortic dissection or aneurysm rupture, 4% cerebrovascular accident, 4% pulmonary embolism/haemorrhage, 2% myocardial infarction, and 1% upper gastrointestinal bleed).⁸ Of the remaining patients with presumed arrhythmic SCD, only 22% had a documented arrhythmia at the time of arrest: ventricular in 84%, supraventricular in 8%, and bradycardia in 8%. Thus, while <15% of sudden death victims with CHD had a confirmed potentially shockable ventricular arrhythmia, incomplete data render this lower limit estimate imprecise. Other registries have reported pulseless electrical activity in 16% and asystole in 11% of adults with CHD and out-of-hospital cardiac arrests.^11,12

Detailed knowledge of the pathophysiology of SCD according to the type of CHD is essential to tailoring preventive approaches, which may include lifestyle modifications, pharmacological agents, and an array of interventional therapies (e.g. catheter-based haemodynamic interventions, cardiac surgery, sympathectomy, ablation, pacemakers, and ICDs). For example, patients with Ebstein anomaly may be at risk for SCD from rapidly conducting and/or multiple accessory pathways that are amenable to catheter ablation. Moreover, increasing evidence implicates atrial tachyarrhythmias as an important cause of SCD in patients with TGA and atrial switch surgery by virtue of provoking myocardial ischaemia-induced ventricular arrhythmias.^13,14 Limited data suggest that beta-blockers afford some protection¹⁵ and timely ablation of rapidly conducting atrial arrhythmias could theoretically avert events.

Role of autopsies for sudden cardiac death in congenital heart disease

When a patient with CHD dies suddenly, close collaboration between the cardiologist and pathologist is imperative in clarifying the exact cause in order to support the family, learn from the life lost, and provide insights that could impact future care. In an autopsy series that included 46 post-operative cases with CHD, necropsy detected findings in 8.5% which, if known before death, would have had an important impact on clinical management, highlighting its importance in audit and education.¹⁶ The need to carry out autopsies more broadly in the context of SCD, in general, should be emphasized given that a sizeable proportion of SCD events in patients with CHD occur prior to diagnosis of CHD.¹⁷

The approach and protocol will vary depending on whether pathologists are facing native disease or previously palliated or corrected CHD. Direct causes of death include cardiac failure with myocardial fibrosis, pulmonary hypertension, and haemorrhage.¹⁸ Apart from surgical corrections and congenital anomalies, necropsy may not reveal specific new findings to explain the SCD. For the majority of these cases, the SCD is presumed arrhythmic.¹⁷ Common CHD lesions identified in autopsy series of SCD include tetralogy of Fallot, Eisenmenger syndrome, TGA, and Fontan circulation. As most patients with complex CHD have undergone surgical and/or percutaneous interventions, the anatomy can be quite challenging. Expert opinion should be sought by the general pathologist in this situation, with a detailed study of the medical history and procedures carried out during life.

Non-congenital heart disease phenotypes for sudden cardiac death

Several variables beyond the type of CHD and standard cardiovascular risk factors could modulate risk for SCD. Sex-related analyses indicate that 65–75% of SCD events in adults with CHD occur in men.^8,19 Underlying reasons to explain the lower proportion of women remain speculative and may hypothetically include oestrogen protective effects, lesser extent of myocardial fibrosis, disparities in traditional risk factors, and/or sex-related differences in types of CHD. Cardiovascular mortality has been associated with higher levels of brain natriuretic peptide (BNP), endothelin-1, soluble tumour necrosis factor receptor type I, norepinephrine, and interleukin-6.²⁰ High-sensitivity C-reactive protein levels provide incremental prognostic value for risk of death or heart failure beyond N-terminal proBNP.²¹ It is unknown whether markers of inflammation, heart failure, and/or neurohormonal activation are of prognostic value for SCD in patients with CHD in a manner similar to the general population. In CHD patients, impaired cardiac autonomic nervous activity has been linked to a higher risk for SCD.²² Autonomic dysfunction is associated with increased sympathetic activity, neuroendocrine activation, higher cytokine levels, and parasympathetic withdrawal—all markers of poorer prognosis.²²

Genetic determinants of sudden cardiac death in congenital heart disease

Attempts to use genetic factors to predict SCD in the general population have been disappointing. However, the unique genetic milieu in patients with CHD renders this research avenue worth exploring. To date, a genetic cause for CHD could be identified in 35% of patients, with monogenic models and chromosomal abnormalities associated with syndromic and non-syndromic forms of CHD.^23,24

Although speculative, there is the potential for genetic factors that are associated with both CHD and ventricular dysfunction or arrhythmias to modulate risk for SCD. For example: (i) mutations in transcription factors, TBX20 and NKX2.5, may result in septal defects and dilated cardiomyopathy; (ii) RASopathy syndromes, which are caused by mutations in the RAS/mitogen-activated protein kinase pathway that plays a vital role in the development, when combined with pulmonary stenosis are also responsible for cardiac hypertrophy; and (iii) mutations in sarcomeric proteins have been associated with both CHD and myopathies, e.g. MYH6 mutations can cause atrial septal defects, Shone complex, and hypoplastic left heart syndrome but also dilated or hypertrophic cardiomyopathy, and MYH7 mutations have been reported in Ebstein anomaly and left ventricular noncompaction.^23,24 There is also some evidence to support a role for common variants predisposing to ventricular maladaptation that could theoretically be linked to SCD. For example, in patients with a Fontan circulation, common genetic variants associated with over-activation of the renin–angiotensin–aldosterone system and increased catecholamine release have been linked to unfavourable haemodynamics and higher mortality.²³ In tetralogy of Fallot, common variants in the hypoxia-inducible factor gene correlate with right ventricular dilation and dysfunction.²⁵ Much remains to be discovered about genetic determinants of SCD in CHD.

Acute resuscitation in congenital heart disease

Congenital heart disease-specific considerations

Education and research on the science of resuscitation are essential components of a comprehensive approach to improve the outcomes related to cardiac arrest in patients with CHD. High-quality cardiopulmonary resuscitation (CPR), with adequate depth and rate of chest compressions, early defibrillation for shockable rhythms, and administration of epinephrine with concurrent CPR for non-shockable rhythms, could improve success rates in any setting. However, in patients with CHD, consideration must also be given to the unique anatomy and cardiopulmonary physiology while adhering to basic standards of high-quality CPR. For example, in assessing haemodynamic status, blood pressure measurements may be unreliable in the presence of vascular obstructions or anomalies, an ipsilateral Blalock-Taussig shunt, or a subclavian flap repair.²⁶ Oxygen saturation levels should be interpreted with knowledge of baseline values and type of CHD. In the setting of a cardiac malposition (e.g. meso- or dextrocardia) or pacemaker or ICD, placement of defibrillator patches or paddles should be adapted according to the position of the heart and at least 8 cm away from the implanted device.²⁶

The impact of CHD on chest compressions during CPR is summarized in Table 2. Importantly, for patients with Fontan physiology who have no sub-pulmonary ventricle, positive pressure ventilation can result in haemodynamic collapse owing to a reduction in superior and inferior vena caval flow with reduced filling of the single ventricle. It should, therefore, be delivered with the lowest possible mean airway pressure necessary to prevent atelectasis and obtain adequate intra-pulmonary gas exchange.²⁷ If the patient with Fontan surgery has a cardiac arrest in an intensive care unit with indwelling catheters, pressure tracings may be monitored and chest compressions modified to optimize the generated blood pressure. There is currently insufficient data to recommend modifying CPR techniques in patients with Fontan palliation or other forms of CHD based on underlying physiology.

Table 2

Impact of congenital heart disease on chest compressions during cardiopulmonary resuscitation

Type of CHD	Effect on chest compressions
Fontan physiology	Chest compressions produce systemic but not pulmonary blood flow Upon chest recoil, pulmonary arteries are predominantly filled by inferior vena cava flow; superior vena cava flow is dependent on the balance between cerebral and pulmonary vascular resistance
Left-sided heart disease
Severe mitral stenosis or regurgitation	Elevated left atrial pressure and pulmonary vascular resistance limit effective pulmonary blood flow and cardiac output
Aortic stenosis	Reduction in stroke volume and cardiac output Left ventricular hypertrophy limits ventricular filling, further compromising cardiac output Challenging to generate sufficient aortic root pressure to support adequate coronary perfusion to the hypertrophied ventricle
Aortic regurgitation	Stroke volume and cardiac output are limited by the regurgitant flow Coronary perfusion pressure and cardiac output can be compromised
Right-sided heart disease
Restrictive right ventricular physiology	Impaired right ventricular filling upon chest recoil, resulting in inadequate pulmonary blood flow and cardiac output Limited coronary perfusion and blood flow to a hypertrophied right ventricle
Pulmonary regurgitation	Worsening pulmonary regurgitation Limited coronary perfusion and blood flow to a hypertrophied right ventricle
Pulmonary arterial hypertension	Chest compressions and resuscitation drugs generally fail to generate pulmonary blood flow, left ventricular filling, and cardiac output Conventional resuscitation is rarely effective

Extracorporeal cardiopulmonary resuscitation

Extracorporeal cardiopulmonary resuscitation refers to the use of extracorporeal membrane oxygenation in patients where conventional CPR is unsuccessful in achieving a sustained return of spontaneous circulation. Currently, there are no published randomized controlled trials comparing outcomes of ECPR to conventional CPR in patients with or without CHD. Nonetheless, current guidelines state that ECPR may be considered for selected individuals as rescue therapy when conventional CPR efforts are failing in settings in which it can be expeditiously implemented and supported by skilled providers.²⁸ In patients with CHD, numerous challenges can be encountered in cannulation for ECPR as a result of access issues, shunts, and anatomies that may require additional venous cannulas.²⁷ In the presence of a shunt, the arterial cannula should be positioned with care to avoid shunt occlusion or over circulation.

Post-resuscitation sequelae and outcomes

In the general population, <20% of SCD victims with out-of-hospital cardiac arrests have successful restoration of a spontaneous circulation and under 10% survive to hospital discharge.²⁹ Data on CPR outcomes in patients with CHD appear more promising but remain limited. In one study of 38 children and adults with CHD who had in- or out-of-hospital cardiac arrests, 66% survived to hospital discharge and 53% remained alive at 3 years.¹¹ In another study of 62 adults with CHD and out-of-hospital cardiac arrests, 44% survived to discharge and 40% remained alive at 1 year.¹² Pulmonary hypertension and CHD complexity were associated with higher early mortality, whereas factors such as public location of the arrest, witnesses, bystander CPR, shorter CPR duration, and a shockable rhythm were determinants of success (Figure 1).^11,12

Figure 1

Post-cardiac arrest phases for patients with congenital heart disease. The upper boxes summarize the post-cardiac arrest phases, with the immediate phase beginning upon return of a spontaneous circulation, followed by an early and intermediate phase (20 min to 72 h), a recovery phase (72 h until discharge), and a rehabilitation phase (following discharge). Key factors associated with early and late mortality are listed along with directed interventions. The primary causes of, and interventions for, cardiac arrest-related late morbidity are also shown.

The mid-term prognosis appears favourable in patients with a primary arrhythmic SCD discharged from the hospital with an ICD.¹² Nevertheless, as a group, survivors of SCD appear to have a worse prognosis when compared with age-, sex-, and disease-matched controls.¹¹ Reasons remain speculative and may include a higher burden of pre-arrest comorbidities and/or ramifications of the cardiac arrest.¹¹ Neurological sequelae related to hypotensive anoxic brain injury and subsequent reperfusion injury are leading causes of mortality and morbidity following CPR. Residual sequelae such as coma and seizures but also subtler cognitive, emotional, and physical disabilities may cause limitations in daily activities and dependence on caregivers, adding substantial long-term morbidity.

Risk stratification for sudden cardiac death

Key considerations

Predicting SCD is an imperfect science. Although continuing to improve risk stratification algorithms is a laudable goal, it should also be recognized that there is a stochastic component to SCD that will limit the accuracy of any predictive model. Akin to weather forecasting, the further in time one attempts to predict SCD, the greater the ‘noise’ introduced by stochastic elements. Additional challenges to predict SCD in the CHD population include the marked heterogeneity of cardiac defects, surgical approaches, and percutaneous interventions; a population that is in constant flux owing to evolving surgical and medical therapies; and the fact that the prognostic value of any given factor could vary widely according to the type of CHD and other variables, thereby limiting the utility of global risk scores. To provide but one concrete example of the latter, inducible sustained ventricular tachycardia could be helpful in risk stratifying patients with the potential for scar-based reentrant ventricular tachycardia, such as tetralogy of Fallot. In this population, inducible sustained ventricular tachycardia is associated with a nearly five-fold higher risk of clinical ventricular tachycardia or SCD during an average follow-up of 6.5 years, and is most helpful in stratifying patients at intermediate risk.^30,31 In contrast, for other types of CHD not associated with reentrant ventricular tachycardia, such as TGA with atrial switch surgery, inducible ventricular tachycardia appears to be of no prognostic value.¹⁵

Any stratification scheme for primary prevention ICDs should carefully balance anticipated benefits against potential risks. In assessing projected benefits, it is understandable that the literature has primarily focused on estimating risk for SCD. While this is a critical factor, it is not the only element to consider (Figure 2). Competing causes of mortality account for >75% of deaths in patients with CHD,¹ such that they cannot be ignored. Moreover, even if the terminal event is SCD, it may not be due to a shockable rhythm and, hence, will not respond to ICD therapy. Importantly, heart failure is increasing in prevalence in the aging population with CHD. Less than 50% of sudden cardiac arrests in patients with heart failure are due to shockable rhythms, with rates as low as 27% in certain clinical subtypes.³² In addition, ICD-related complications are substantially higher in the CHD population and must also be factored into the balance.³³ Economic considerations are likewise paramount since the cost threshold at which it is economically viable to save one quality-adjusted life year varies considerably among nations.²⁶

Figure 2

Factors to consider in weighing risks and benefits of primary prevention implantable cardioverter-defibrillators. ICD, implantable cardioverter-defibrillator; SCD, sudden cardiac death.

There are different methodological techniques that could provide the framework to weigh risks and benefits in a quantitative fashion.³⁴ Approaches such as estimating numbers needed to treat and harm, along with their relative value-adjusted modifications, require subjective weighting schemes, e.g. weight assigned to an ICD-related complication relative to a year of life saved. Other approaches assign joint distributions of benefits and risk, such as the stated preference method, risk-benefit plane, and multi-criteria decision analyses. Ideally, weights inputted into mathematical models should be derived quantitatively. Regardless of the approach, the ideal candidate for a primary prevention ICD has a high risk for SCD due to a shockable rhythm, low risk of dying from other causes, low risk of ICD-related complications, and is in an environment in which associated costs are affordable.

Risk scores

Two recent studies were designed to identify and validate factors associated with SCD in adults with CHD (Table 3).^19,35 A case–control study combined 165 patients with SCD, along with 310 matched controls, from the Dutch nationwide CONCOR registry and the Toronto Congenital Cardiac Centre for Adults.⁸ Identified factors were combined with a CHD lesion-specific baseline risk of SCD estimated from the CONCOR registry to create a point-based risk score (PREVENTION-ACHD), which was then prospectively validated in a single-centre cohort study of 783 patients followed for 2 years.³⁵ Similarly, a risk prediction model from the Spanish adult CHD network merged retrospective cohort data with a case–control study to identify 278 cases of SCD (n = 163) or non-fatal cardiac arrest (n = 115).¹⁹ Congenital heart lesions were clustered into four risk categories and combined with identified predictors to generate 5-year risk estimates. High-risk features common to both scores include systolic dysfunction of the systemic or sub-pulmonary ventricle, a wide QRS complex, and ischaemic heart disease.

Table 3

Studies of global risk scores for sudden cardiac death in congenital heart disease

	PREVENTION-ACHD	SPANISH ACHD
Year of publication	Risk score 20128; Prospective assessment 202135	202119
Study design for development of risk score	Retrospective multicentre case–control study with matching by age, gender, diagnosis, date and type of surgical repair, treating medical centre	Data combined from single-centre retrospective cohort and multicentre case–control studies
Outcome	In deriving score: SCD of documented or presumed arrhythmic aetiology; in prospective assessment: SCD or ventricular arrhythmia >30 s, aborted SCD, or appropriate ICD intervention	SCD or non-fatal sudden cardiac arrest, including syncopal VT and appropriate ICD shocks
Number of events	165	278 (163 SCD; 115 non-fatal)
Time horizon	Assessed in prospective cohort with 2-year follow-up	5 years
Variables retained in the final model	Type of CHD Coronary artery disease New York Heart Association classes II/III heart failure Supraventricular tachycardia Systemic ejection fraction <40% Sub-pulmonary ejection fraction <40% QRS duration ≥120 ms QT dispersion ≥70 ms	Type of CHD (clustered into four groups) Younger age (per year) Male sex Unexplained syncope Symptomatic ischaemic heart disease Non-life threatening ventricular arrhythmia QRS duration (per ms) Moderate–severe systemic ventricular hypertrophy Moderate–severe sub-pulmonary ventricular hypertrophy Moderate–severe dysfunction of either ventricle (ejection fraction <45%)

Such risk prediction models provide an important step forward towards identifying high-risk features common to a broad spectrum of patients with CHD. However, the factors considered are limited by the availability of data that is routinely collected, and the models are not devised to tease out predictors unique to specific defects. Moreover, the risk scores focus on one of the numerous elements to consider in risk stratification for primary prevention ICDs, i.e. estimates of SCD. Risk prediction scores developed for patients with specific defects, such as tetralogy of Fallot,³⁰ are smaller in scope but carry the advantages of greater homogeneity, enhanced granularity, and ability to adjust for competing causes of mortality. They are, however, subject to their own limitations, particularly regarding applicability (e.g. non-availability of factors that require testing beyond standard metrics), surrogate outcomes, and generalizability.

Implantable cardioverter-defibrillator indications

Unlike risk scores, guidelines and consensus-based recommendations are not risk prediction tools and should not be interpreted as such. Their purpose is not to be all-encompassing, but to draw attention to particular scenarios where evidence-based consensus can be achieved. It is not the objective of guidelines to maximize sensitivity but, rather, to assist practitioners about appropriate decisions under specific clinical circumstances after considering all sources of evidence. Consistent with the Myerburg principle of risk continuum, lower-risk patients with more prevalent conditions not addressed by guidelines may contribute a larger number of overall cases of SCD. Guidelines must tackle the challenge of reconciling imperfect data with complex clinical decision-making. To that end, indications for ICDs in CHD have evolved from initial criteria limited to survivors of repeated sudden cardiac arrest to contemporary recommendations for patients who are considered possibly at risk of future cardiac arrest.

Although ICD guidelines for CHD were included as brief sections in early device guidelines and documents, the first ICD recommendations specific for CHD patients were published in 2014, as an expert consensus statement from the Pediatric and Congenital Electrophysiology Society (PACES)/Heart Rhythm Society (HRS).³⁶ Specific recommendations regarding ICD indications for children and adults with CHD have subsequently been published by the American Heart Association (AHA)/American College of Cardiology (ACC)/HRS in 2017,³⁷ European Heart Rhythm Association (EHRA) in 2018,³⁸ European Society of Cardiology (ESC) in 2020,³⁹ and PACES in 2021.⁴⁰  Table 4 summarizes ICD recommendations consistently addressed in these documents.

Table 4

Implantable cardioverter-defibrillator recommendations in patients with congenital heart disease

	PACES/HRS 201436	ACC/AHA/HRS 201737	EHRA 201838	ESC 202039	PACES 202140
Sudden cardiac arrest due to VF/VT after evaluation to define cause and exclude reversible triggers	I	I	I	I	I
Unstable VT after electrophysiologic and haemodynamic evaluation	I	I	I	I	I
LVEF <35% with biventricular physiology and NYHA Classes II–III	I	IIb	I	IIa
Unexplained syncope with LVEF <35% or VT/VF at EPS	IIb	IIa	II	IIa	IIa
Tetralogy of Fallot with clinical VT or multiple risk factors	IIa	IIa	II	IIa
Single/systemic ventricular dysfunction with NSVT, NYHA Classes II–III	IIb	IIb	II	IIb	IIb
Non-hospitalized CHD awaiting transplant	IIb		II
Life expectancy <1 year; Incessant VT/VF; NYHA Class IV; severe psychiatric illness	III	III	III		III
Class of recommendation	Class I: ICD is recommended	Class IIa: ICD should be considered	Class II: ICD can be considered	Class IIb: ICD may be considered	Class III: ICD is not indicated or contraindicated

PACES, Pediatric and Congenital Electrophysiology Society; HRS, Heart Rhythm Society; ACC, American College of Cardiology; AHA, American Heart Association; EHRA, European Heart Rhythm Association; ESC, European Society of Cardiology; VF, ventricular fibrillation; VT, ventricular tachycardia; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; EPS, electrophysiology study; VT, ventricular tachycardia; NSVT, non-sustained VT.

Class I recommendations (implantable cardioverter-defibrillator is recommended)

There is consistent consensus that an ICD is indicated in (i) CHD patients who are survivors of cardiac arrest due to ventricular arrhythmias following evaluation of the cause of the event and after exclusion of any reversible triggers and (ii) in patients with symptomatic sustained ventricular tachycardia after haemodynamic and electrophysiologic evaluation. The PACES and ESC documents^39,40 also propose that in select cases, ablation, or surgery may offer effective alternatives. Less consistent is the Class I indication by PACES/HRS³⁷ and EHRA³⁸ for adults with a left ventricular ejection fraction <35% and New York Heart Association (NYHA) Classes II–III functional status. AHA/ACC/HRS guidelines³⁷ suggested this as a Class IIb indication, the ESC as a IIa indication,³⁹ and the 2021 PACES consensus statement, which focused on children, issued no formal recommendation due to the lack of sufficient data.⁴⁰

Class IIa recommendations (an implantable cardioverter-defibrillator should be considered)

There is consensus that an ICD should be considered in the CHD patient with unexplained (presumably arrhythmic) syncope in the setting of impaired left ventricular function or inducible ventricular tachycardia or ventricular fibrillation at electrophysiology study. There is a similar consensus for consideration of ICD implantation for the patient with tetralogy of Fallot and multiple risk factors.

Class IIb recommendations (an implantable cardioverter-defibrillator may be considered)

The uncertainties of risk stratification and risk of potential implant complications in the patient with systemic right or single ventricular dysfunction, non-sustained ventricular tachycardia, NYHA functional classes II–III symptoms, and/or severe systemic atrioventricular valve insufficiency are recognized in the ICD recommendations as a Class IIb indication. The role of the subcutaneous ICD in these patients remains to be defined. The non-hospitalized CHD patient awaiting transplant may also be considered for ICD implantation. It is worth noting that for non-sustained ventricular tachycardia, there is a paucity of data in the CHD population on what threshold (e.g. rate, number of consecutive beats, frequency of episodes) confers higher risk, rendering interpretation of single, slow, or brief runs challenging, particularly when incidentally detected by cardiac implantable electronic devices or prolonged monitoring.

Class III recommendations (an implantable cardioverter-defibrillator is not indicated)

There is consistency among the recommendations that an ICD is not indicated in patients with a life expectancy <1 year, incessant episodes of ventricular tachycardia or ventricular fibrillation, significant psychiatric illness which may be aggravated by the ICD implant or preclude consistent follow-up, or NYHA functional class IV symptoms when the patient is not a candidate for cardiac resynchronization therapy.

Implantable cardioverter-defibrillators in congenital heart disease

Transvenous systems

Challenges and complications

Implantation of a transvenous ICD in patients with CHD presents unique challenges. In up to 15%, venous access may be complicated by obstructions or congenitally absent or anomalous veins. Anatomic barriers to stable fixation may exist, such as a pulmonary ventricle of smooth left ventricular morphology or sub-pulmonary atrioventricular valve disease. In TGA with atrial switch, the superior baffle must be patent to be traversed, with care to avoid phrenic capture on the lateral systemic ventricular wall and native left atrial appendage. In patients with intra-cardiac shunts, a high degree of caution should be exercised when considering transvenous leads due to the risk of systemic thrombo-embolism. Options should be thoughtfully considered on a case-by-case basis [e.g. shunt closure prior to transvenous lead placement, subcutaneous ICD (S-ICD), epicardial approach].⁴¹

Despite meticulous planning, device-related complications in CHD remain high and exceed that seen in acquired heart disease.³³ In a meta-analysis of CHD ICD studies, device-related complications occurred in 26% of patients and inappropriate shocks in 25% (mean follow-up 3.7–3.8 years).³³ Given the advancing life expectancy of adults with CHD, cumulative morbidity is likely to be substantial. Complications of transvenous ICDs in patients with CHD and predisposing factors are summarized in Table 5.

Table 5

Complications of transvenous implantable cardioverter-defibrillators in patients with congenital heart disease and pre-disposing factors

Complication	Pre-disposing factors
Inappropriate shocks	Propensity to macroreentrant atrial tachycardia (which may demonstrate 1:1 conduction and aberrancy) Sinus tachycardia in young patients Lead failure or oversensing
Lead dislodgement	Complex anatomy: anomalous veins, baffles, pulmonary ventricle of left ventricular morphology Sub-pulmonary atrioventricular valve disease (e.g. Ebstein anomaly) Multiple reinterventions
Lead failure	Young age at implant and long dwell time Lead strain in physically active younger patients Acquired scar due to surgery, ablation, or remodelling Lead recalls
Infection/endocarditis	Multiple reinterventions Prosthetic material post-surgical repair Impaired immunity (e.g. cyanosis, immunodeficiency)
Bleeding/haematoma	Challenging vascular access Anticoagulation for atrial arrhythmia Impaired coagulation (e.g. cyanosis, liver disease)
Thrombo-embolism	Shunts Comorbid atrial arrhythmias Prosthetic material post-surgical repair Low flow state (e.g. Fontan)

Lead management, follow-up, and ancillary testing

Device extraction is increasingly utilized in CHD to manage lead failure or infection, and may allow magnetic resonance imaging conditionality if no lead is abandoned. Limited transvenous access, prior operations, baffles, conduits, stents, and occluder devices may alter lead position and pose unique challenges to lead extraction.⁴² Several reports have shown procedural success and safety of lead extraction in CHD by using various tools (e.g. laser, mechanical, and rotational sheaths). However, unique complications have been reported such as damage to the sub-pulmonary atrioventricular valve in patients with TGA that appears particularly prone to injury from lead extraction.⁴²

Remote monitoring and automatic device safety alerts have greatly facilitated follow-up. Periodic chest X-rays can be helpful in detecting lead-related problems such as coronary compression, myocardial strangulation, dislodgement, and fracture. Echocardiography is useful in monitoring ventricular function in patients with a high percentage of ventricular pacing, and to identify endocardial lead-related complications affecting sub-pulmonary atrioventricular valves.⁴³ Exercise testing can provide data for programming arrhythmia detection parameters, as well as for detecting rate-related QRS morphology and T wave changes.

The subcutaneous implantable cardioverter-defibrillator in congenital heart disease

Congenital heart disease patients with complex access issues inspired the development of the S-ICD. The S-ICD has since evolved into a system as effective as the transvenous ICD with regard to successful defibrillation testing at 80 J, with similar success rates in patients with and without CHD (100 vs. 98.5%).⁴⁴ In a recent randomized trial, the S-ICD was found to be non-inferior to the transvenous ICD for device-related complications and inappropriate shocks.⁴⁵ While the S-ICD avoids complications associated with transvenous leads, its main limitation is the absence of anti-bradycardia and anti-tachycardia pacing. It is a particularly attractive option for patients with vascular access limitations, prior transvenous device infections, intra-cardiac shunts, and/or absence of a sub-pulmonary ventricle (i.e. Fontan circulation). The implant technique is extra-thoracic and simple. The subcutaneous lead could be placed to the right or left of the sternum along the parasternal space, ipsilateral to the cardiac apex. Moreover, hybrid systems with a S-ICD combined with epicardial or leadless pacing can be feasible if certain conditions are met, thereby expanding potential clinical applications.

Nevertheless, the S-ICD remains a relatively new technology with limited experience in the CHD population.^44,45 Screening with a reconstructed surface ECG reveals that 40% of CHD patients are non-eligible for the S-ICD predominantly owing to a wide bundle branch block.⁴⁶ The initially reported unacceptably high rate of inappropriate shocks due to T wave oversensing has since decreased substantially as a result of an enhanced algorithm (SmartPass^©), implanter experience (e.g. intermuscular implantation, two-incision technique), and improved patient selection, with no difference in the rate of inappropriate shocks in patients with and without CHD (10.5 vs. 10.9%).⁴⁴ Thus, the S-ICD provides a safe and effective option in CHD patients who remain eligible post-screening and do not have a pacing indication.

Implantable cardioverter-defibrillator and sports participation

Fear of SCD has historically led to restrictions on physical activity. However, in patients with CHD, <10% of SCD events occur during exercise.^2,8 In most instances, cardio-protective and beneficial mental effects of physical activity far outweigh risks for SCD.⁴⁷ There are, however, notable exceptions. Over 80% of sudden deaths in patients with TGA and atrial switch occur during exercise.¹⁴ In this subgroup, it is prudent to limit exercise to low to moderate intensity dynamic sports and low intensity static sports.⁴⁷ Restriction of physical activity may also be warranted in patients with congenital coronary anomalies.² Formal exercise testing may help determine the level of exercise that is safe. An individualized tailored approach with a discussion about what is known and not known, followed by ‘shared decision-making’ is key to successful sports counselling. Guidelines recommend that ICD recipients with CHD engage only in non-competitive and non-contact sports.⁴⁸ The protection an ICD offers should not be used as justification for participation in an exercise programme that is more intense than recommended in the absence of an ICD.

Future directions

Over the past few decades, the rapid expansion of knowledge on SCD in CHD has largely been driven by observational studies and real-world experience. Although the profound insights gained from this epidemiological lens are not to be overlooked, drawbacks should be recognized to more effectively target knowledge gaps through future research (Table 6).

Table 6

Knowledge gaps and future directions

Knowledge gaps	Future directions
Imprecise or uncertain estimates of the epidemiology of SCD in CHD due to limitations in observational research including varying definitions, incomplete taxonomy, inconsistent ascertainment of outcomes, limited data granularity, and understudied subgroups	Efforts to better understand SCD in CHD will require innovative approaches to research with enhanced quality, scope, and depth of large-scale data, including standardized definitions and improved taxonomy to better reflect underlying pathophysiology
Incomplete understanding of premonitory signs and risk factors for SCD in CHD	A collaborative multipronged and transdisciplinary approach to research is required. Priorities include: Identifying high-risk features across the gamut of potential anatomic, haemodynamic, surgical, electrophysiological, inflammatory, neuroendocrine, environmental, metabolic, and genetic factors Incorporating advances in ‘big data’, wearable technologies that permit real-time data acquisition, and innovations in computer science and statistical methods that enable rapid synthesis of large volumes of data to adaptively identify and predict patterns
Limited data on optimal approaches for, and acute and long-term outcomes after, resuscitation in CHD	Efforts should be directed towards advancing the science of resuscitation specific to CHD through transdisciplinary research, improving education and access to high-quality CPR and specialized care, and systematically tracking outcomes
No definitive evidence to support a mortality reduction from ICDs in CHD based on specific criteria or a particular risk stratification scheme	Consideration should be given to flexible and efficient randomized adaptive ICD trial designs suitable to smaller targeted populations with CHD deemed to be at clinical equipoise
Paucity of data on non-ICD approaches to SCD prevention, including lifestyle modifications, pharmacological therapies, and catheter and surgical interventions	An in depth understanding of disease processes and triggers for SCD across the various types of CHD through multicentre research that includes detailed histopathological analyses should help guide a more tailored approach to prevention

Conclusion

Sudden cardiac death in patients with CHD is a high-stake high-priority matter. Importantly, SCD is not a unitary diagnosis but a terminal event with diverse causes and triggers such that improving outcomes requires a multipronged interdisciplinary approach that addresses numerous facets including taxonomy, pathology and pathophysiology, phenotypes, early warning signs, resuscitation, and preventive therapies (Graphical Abstract). Much remains to be learned about underlying mechanisms and triggers in order to guide tailored therapies. Implantable cardioverter-defibrillators play an important role in SCD prevention, with risk scores that have emerged to assist in selecting appropriate candidates. Nevertheless, risk stratification is not simply a matter of identifying patients at high risk for SCD. Many events are not due to shockable rhythms. Moreover, some factors associated with SCD confer a similar or even higher risk of death from other causes and do not effectively identify those at the greatest proportional risk of arrhythmic death. In light of these and other complexities, the prevention of SCD in the CHD population carries unique challenges. However, unlike many victims within the general population for whom the first contact with the healthcare system occurs at the time of SCD, most patients with CHD will (or at least should) be under medical surveillance. This major advantage provides the opportunity for robust high-quality multidimensional research to enhance our understanding of SCD, improve risk prediction, and, ultimately, save lives.

Funding

Dr P.K. is supported by the endowed André Chagnon Research Chair in Electrophysiology and Congenital Heart Disease.

Conflict of interest: Dr S.B. received a research grant from Medtronic and consulting fees from Milestone Pharmaceuticals. Dr B.M. received a research grant from Boston Scientific and consulting fees from Medtronic and Biotronik. The other authors have no potential conflict of interest to disclose.