Heart disease is the leading cause of death in the US, with sudden cardiac death (SCD) accounting for most of these fatalities.1 Considering the growing prevalence of coronary heart disease and cardiomyopathies (see Figure 1) that, taken together, cause most SCDs, effective treatments are needed. Since by definition SCD is ‘sudden’ (generally accepted to be within one hour after onset of symptoms), success rates of secondary prevention strategies are limited.2 Several clinical studies have demonstrated that antiarrhythmic therapy in addition to β-blockers confers no benefit for all-cause mortality.3,4 By contrast, a series of studies have shown implantable cardioverter–defibrillators (ICDs) to be effective in the prevention of SCD.3,5 However, for primary prevention strategies to work, accurate risk stratification needs to be performed. This is particularly true against a background of increasing cost constraints in healthcare systems.
This article will focus on the clinical indications, safety and efficacy of implantable device therapy including ICDs. Other therapies, such as antiarrhythmic drugs or simple pacemakers, will not be reviewed.
Risk Stratification for Primary Prevention
It has been well established that patients at lowest risk of SCD constitute the group with the greatest number of SCDs (see Figure 2).6 For this reason it is essential to risk-stratify patients to identify those at highest risk of SCD. These patients, in turn, are most likely to profit from ICD therapy.
The single most important parameter for adequate risk stratification is left ventricular ejection fraction (LVEF), which is most commonly determined by echocardiography. Other modalities such as cardiac magnetic resonance imaging or catheter-based ventriculography are also valid for determination of LVEF.7 The diagnostic value of resting electrocardiograms (ECGs) is limited to the identification of structural diseases such as hypertrophy and ischaemia as well as congenital abnormalities (e.g. channelopathies such as Brugada-syndrome, long-and short-QT syndrome [LQTS and SQTS, respectively]). Microvolt T-wave alternans (MTWA) measures repolarisation abnormalities in the ventricles. In MTWA, alternating T-wave amplitudes and morphologies are detected from beat to beat by computerised signal processing techniques.
The predictive value of abnormalities in MTWA has been controversial and more studies are needed before MTWA can become a routine screening technique in this setting.8 Holter recordings showing non-sustained ventricular tachycardias (NSVTs) without underlying heart disease have not been proved to be of predictive value for SCD. The determining factor for the relevance of NSVTs appears to be the structural heart disease, not the arrhythmia. Positive electrophysiological (EP) testing is a well-established marker of increased risk of SCD. However, there are a high number of false-positive tests and a negative test does not imply a lack of risk.8
Primary Prevention of Sudden Cardiac Death in Cardiomyopathies
The primary prevention of SCD and thus the effectiveness of ICD therapy has been evaluated in a series of randomised, controlled trials studying patients with an ischaemic cardiomyopathy.
The Multicenter Automated Defibrillator Trial I (MADIT I) randomised patients with a history of myocardial infarction (MI), spontaneous NSVT, LVEF ≤35% and inducible VT refractory to intravenous procainamide administration into either an ICD or a conventional therapy group.9 MADIT I was prematurely halted when early analysis revealed a significant reduction of overall mortality by 54% for the ICD-treated patients.
The Multicenter Unsustained Tachycardia Trail (MUSTT),10 which had a similar design to MADIT I, evaluated whether EP-guided antiarrhythmic therapy (antiarrhythmic or ICD) was superior to no antiarrhythmic therapy. After five years of follow-up the ICD group fared best, with a reduction in total mortality of 31% compared with the antiarrhythmic group and 24% compared with the group without antiarrhythmic therapy.
The Multicenter Automated Defibrillator Trial II (MADIT II) randomised patients with previous MI and LVEF ≤30% to receive either an ICD or conventional therapy.5 The ICD group revealed a 31% reduction in total mortality and a 67% reduction in SCD. Post hoc analysis of the MADIT II data indicated that despite improved survival in the ICD group, these patients had an increased risk of heart failure.11 One proposed explanation might be that right ventricular (RV) pacing with dual-chamber ICDs and myocardial damage induced by ICD shocks contributed to the progression of heart failure.
The Defibrillator in Acute Myocardial Infarction Trial (DINAMIT) studied patients with a recent MI (within six to 40 days), LVEF ≤35% and impaired cardiac autonomic function.12 Subjects were randomised to either the ICD or the non-ICD group. During the follow-up period (mean 30 months), the ICD group fared no better in terms of overall mortality compared with the control group.
In conclusion, these studies prove that patients with ischaemic cardiomyopathy and LVEF ≤35% receiving ICD therapy have significantly improved outcomes, provided the ICD is not implanted within 40 days after an MI.
Whether ICD therapy for primary prevention of SCD is of benefit for patients with non-ischaemic cardiomyopathy is less certain. Several studies have not been able to demonstrate a clear advantage for device therapy.
In the Cardiomyopathy Trial (CAT), 104 patients with a recent diagnosis of dilated cardiomyopathy (DCM) and LVEF ≤30% were randomised to receive an ICD or conventional therapy. After a mean follow-up of 5.5 years, no significant difference in cumulative survival could be found between the two groups.13 The Amiodarone versus Implantable Cardioverter Defibrillator Trial (AMIOVIRT) studied 103 patients with DCM, LVEF ≤35% and asymptomatic NSVTs.14 Subjects received either amiodarone or an ICD. Again, there was no improved survival in the ICD group. Possible reasons for the negative results of CAT (published in 2002) and AMIOVERT (published in 2003) were suggested to be the relatively small sample size and the lack of a run-in phase on optimal medical therapy. Had such a therapy been initiated, the results might have been different.
The next step was the Defibrillators in Nonischemic Cardiomyopathy Treatment Evaluation (DEFINITE) trial (published in 2004). In DEFINITE 458 patients with DCM, LVEF ≤35% and frequent premature ventricular complexes or NSVTs were randomised to either best medical therapy or an additional single-chamber ICD.15 After a mean follow-up of 29 months, the results showed a significant 80% reduction of SCD from arrhythmia and a trend towards a reduction in all-cause mortality in the ICD therapy group (p=0.06). In 2005 the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) was published. It enrolled 2,521 patients with LVEF ≤35% regardless of aetiology and New York Heart Association (NYHA) functional class II (70%) or III (30%) symptoms.
Roughly 50% of the patients had ischaemic and non-ischaemic cardiomyopathies. Patients received an ICD, amiodarone or conventional treatment. Over the median follow-up period of 45.5 months, ICD therapy reduced the overall mortality rate by 23% compared with the control group. Interestingly, the amiodarone therapy group had no survival benefit. The results could be confirmed regardless of the aetiology of heart failure; however, the benefit was limited to patients in NYHA functional class II.3
A meta-analysis based on five primary prevention trials with a total of 1,854 patients suffering from non-ischaemic cardiomyopathy suggested that ICD therapy might reduce all-cause mortality by 31% compared with medical therapy alone. Thus, it seems reasonable to advocate ICD therapy to prevent SCD in patients with non-ischaemic cardiomyopathy.
Several studies have shown that ICD therapy is of benefit in hypertrophic cardiomyopathy (HCM).16–19 The largest of these included 506 unrelated patients with HCM. An annual 10.6% reduction of SCD for secondary prevention and a 3.6% reduction for primary prevention was found.19 Apart from secondary prevention, at least one of four major risk factors is sufficient to justify ICD implantation for primary prevention in HCM: a history of premature HCM-related SCD in one or more first-degree or other relatives under 50 years of age; massive LV hypertrophy (maximum wall thickness ≥30mm); NSVTs (≥120/min) on Holter monitoring; and prior unexplained syncope not of neurocardiogenic origin.19
Arrhythmogenic Right Ventricular Cardiomyopathy
Arrhythmogenic RV cardiomyopathy (ARVC) is a genetically (desmosomal mutations) determined disease characterised by a progressive replacement of predominantly RV cardiomyocytes by fibro-fatty tissue. A characteristic of ARVC is an electrical instability that may lead to malignant arrhythmias and SCD in apparently otherwise healthy young individuals (particularly athletes). ARVC has been suggested to account for up to 10% of SCDs in patients under 35 years of age.20 A set of major and minor criteria needs to be met for a diagnosis of ARVC, including ECG, structural, functional, histological and imaging parameters. As in HCM patients, there is an obvious indication for ICD therapy for secondary prevention of SCD and sustained VT or ventricular fibrillation (VF).20,21
For primary prevention, matters are more controversial and decisions should be made on a case-by-case basis. Arrhythmias may respond favourably to β-blockers, sotalol or amiodarone treatment, and catheter ablation should also be considered if recurrent VTs are present. However, whether these measures are sufficient to prevent SCD has not been established. The prognostic role of EP testing is unknown.21 However, in ARVC patients with otherwise unexplained syncope or a positive family history of SCD, an ICD should be considered. Since arrhythmias often arise during physical exercise in these patients, competitive sports should be avoided.
Primary Prevention of Sudden Cardiac Death in Coronary Artery Disease
Coronary artery disease (CAD) is present in 65–85% of patients suffering from SCD.22 Nonetheless, CAD per se is insufficient to warrant primary prophylactic ICD therapy. A reduced LVEF is the single most important risk factor, as discussed above. Additional risk factors are needed to identify those CAD patients who are at increased risk of SCD. Diabetes has been linked to such an increased risk.23 Substantial data have correlated SCD with markers of plaque vulnerability (e.g. heritable alterations of specific matrix metalloproteinases), markers of enhanced thrombogenesis (e.g. elevated D-dimer, increased apo-B, decreased apo-A1 or platelet glycoprotein receptor polymorphisms), genetic variations that predispose to vasospasm (e.g. variations in the vascular endothelial nitric oxide synthetase [eNOS] system) and markers of inflammatory response (e.g. C-reactive protein).22 However, whether any single or combination of these risk factors should mandate ICD therapy in patients with preserved LVEF has not been established.
MI is followed by remodelling processes and, depending on the amount of affected tissue, a decrease of ventricular function and an increase of EP vulnerability.24 Recently, the Immediate Risk- Stratification Improves Survival (IRIS) study addressed the issue of whether these processes mandate early (within 30 days) prophylactic ICD implantation after MI. The results did not reveal an overall reduction in mortality;25 thus, at this time early implantation of an ICD after MI is not indicated.
Primary Prevention of Sudden Cardiac Death in Channelopathies and Congenital Heart Disease
Generally, the term electrical cardiomyopathy refers to LQTS, SQTS, Brugada syndrome and catecholaminergic polymorphic VT (CPVT). All of these diseases put the patient at increased risk of SCD.26 Unfortunately, due to the low incidence of electrical cardiomyopathies there is a lack of large randomised trials studying their treatment. Often, these cardiomyopathies are diagnosed in young patients, with the first clinical manifestation being SCD. For patients with LQTS and CPVT, β-blockers have been shown to significantly reduce SCD.27,28 For other patients, and those at particularly high risk of SCD, only ICD therapy can effectively prevent SCD. Thus, adequate risk stratification is needed to identify patients in whom the benefits of ICD placement outweigh the potential side effects. According to the 2006 American Heart Association (AHA)/American College of Cardiology (ACC)/ European Society of Cardiology (ESC) guidelines, ICD therapy is only indicated for secondary prevention. Primary prevention by ICD therapy should be considered in patients at high risk of SCD.21
In LQTS the degree of maximal QT elongation in repetitive ECG recordings correlates well with the risk of cardiac events,29,30 with a QT time of 500ms regarded as a critical cut-off.31,32 In addition, a recent history of syncope can elevate the risk of SCD by up to 18-fold.27,33,34 Genotyping can further help, with LQTS1 and LQTS2 patients experiencing more non-fatal cardiac events and LQTS3 patients suffering more SCD and sudden cardiac arrests.35
Currently, only about 50 patients worldwide have been diagnosed with SQTS. In baseline ECGs a QTc of 350–360ms is considered short. However, whether, in analogy to the LQTS, an ever shorter QTc correlates to higher SCD risk is unknown. Programmed ventricular stimulation for additional risk stratification is of uncertain value in SQTS. However, case reports from highly symptomatic families with markedly increased prevalence of SCD and syncope revealed very short QTcs. In these families, afflicted patients received ICDs. Thus, it seems reasonable to implant ICDs for primary prevention in patients with a history of syncope.
In Brugada syndrome a type I ECG, a history of syncope and sudden cardiac arrest are predictors of SCD. However, electrophysiological testing for additional risk stratification is again of controversial and uncertain value in these patients. However, the negative predictive value, particularly in asymptomatic patients, is unquestionable at 86–100%.36–39
Central to therapy for CPVT are β-blockers and/or calcium antagonists.40 ICD therapy is limited to high-risk patients.21 However, programmed ventricular stimulation, a previous history of syncope or a positive family history of SCD are of no help in identifying individuals at particular risk of SCD. Genotyping may identify one of two known mutations, with the cardiac ryanodine 2 receptor (RYR2) mutation indicating an increased risk of syncope and SCD. Unfortunately, the penetrance of the RYR2 mutation varies significantly. Most helpful are stress ECG tests documenting the development of bidirectional premature complexes and the eventual degeneration into VT and VF. After therapy initiation, follow-up stress ECG tests should document effective treatment. In patients experiencing multiple ICD discharges, selective LV sympathetic denervation can be considered.41
Congenital heart disease comprises a large group of complex cardiac defects. Advances in surgical treatment have considerably improved the prognosis of affected patients; however, despite much improved survival, mortality rates do not return to those seen in healthy controls. In tetralogy of Fallot (TOF), SCD is responsible for ~30% of late mortality.42 However, controlled trials for ICD therapy in congenital heart disease are lacking, not least because of the great heterogeneity and low incidence of these diseases. For TOF, a higher risk of inappropriate therapies and other device-associated complications coupled with a lower incidence of appropriate discharges compared with patients with DCM mandate particularly careful patient selection.43 In atrial redirection surgery for dextro-transposition of the great arteries, the (systemic) morphological right ventricle may develop severe dysfunction due to chronic pressure overload.44 However, placement of ICD leads can be challenging.45 Whether these patients will benefit from ICD therapy as much as DCM patients, for example, is uncertain and controversial.46,47
Miscellaneous Aspects of Implantable Cardioverter–Defibrillator Therapy
Relevant surgical complications with ICD implantations such as infection and lead dysfunction occur in around 1–2% of cases.48 Some of these complications may actually mandate further surgery. Lead dysfunction can include (micro-) dislocations and lead fractures. This can result in insufficient signal amplitudes or false signals, which in turn may trigger inappropriate therapies.
Other causes can be the addition of drugs, such as amiodarone, that can increase the defibrillation threshold.49 Later complications after implantation include lead and pocket infection in 0.5% of patients, lead problems such as dislodgment, fracture or migration in 8% of patients, generator malfunction in 1.9% of patients and system modification due to threshold elevation in 1.9% of patients.50
Effect of Right Ventricular Pacing
Several trials have suggested that ICD patients are at increased risk of heart failure events.5,12 The Dual-Chamber and VVI-Implantable Defibrillator (DAVID) Trial demonstrated that dual-chamber ICD patients with a decreased LVEF and a lower heart rate programmed at DDD-70 without the need for antibradycardic pacing suffered from increased heart failure morbidity.51 One study suggests that in such patients a cumulative incidence of RV pacing >2% should be avoided to reduce this independent cause of hospitalisation and mortality.52
Quality of Life Issues
The way patients react psychologically to ICD implantation is highly variable. In the Antiarrhythmics Versus Implantable Defibrillators (AVID) trial, patients who had experienced an ICD shock during the follow-up period also reported reductions in their physical functioning and mental wellbeing and increased anxiety.53 In the Canadian Implantable Defibrillator Study (CIDS), quality of life was affected negatively only if patients had received more than four shocks.54 However, in both AVID and CIDS, patients who did not experience ICD discharges reported a similar or superior quality of life to the part of the medical control group that did not experience any side effects either. In MADIT II, appropriate ICD shocks led to a reduction in physical capacity but not mental wellbeing.
ICD use for primary prevention of SCD in ischaemic and non-ischaemic cardiomyopathies has dramatically increased hand in hand with the miniaturisation of devices and the publication of ground-breaking trials. This development owes much to the impressive absolute and relative beneficial effects on survival in high-risk patients. Nonetheless, ICDs require implant surgery, exposing the patient to potential complications, and they are expensive, which limits their use in cost-conscious health systems. Most heart failure patients with a wide range of underlying diseases and a reduced LVEF of ≤35% will profit from ICD therapy. However, the benefit is less clear for some patients with channelopathies and congenital heart disease. For these patients careful and individual weighing of risks versus benefits is mandatory. Finally, it has to be recognised that RV pacing in heart failure patients may actually increase heart-failure-related morbidity and mortality.