Keywords

Introduction

Sudden cardiac death (SCD) is most commonly defined as death from unexpected circulatory arrest occurring within an hour of the onset of symptoms or during sleep [1]. In the majority of cases, SCD is triggered by an arrhythmic event, most frequently ventricular tachycardia, ventricular fibrillation and asystole, although recently, pulseless electrical activity has been noted more frequently.

The leading cause of death in the United States, according to the Centers for Disease Control and Prevention, is cardiovascular disease (611,000 deaths annually) [2]. Other common causes of death include malignancy (584,000), chronic respiratory disease (149,000), accidents (130,000), and strokes (128,000) [2]. Despite a dramatic decline in mortality from heart disease over the past 30 years, SCD remains the leading cause of cardiovascular death. It is estimated that SCD accounts for more than 300,000 deaths in the United States annually and approximately 50% are the first known cardiac event [2, 3]. Even this may be an underestimation, as the majority of patients who suffer out of hospital cardiac arrest never survive the initial event [2].

In order to reduce the burden of SCD, much effort has been directed to identify and better treat those most at risk. It is well described that underlying structural heart disease in the context of ischemia, systolic heart failure and fibrosis often trigger sustained arrhythmias that may lead to cardiovascular collapse and death. However, a broad range of at risk populations for SCD exists, including patients with family history of coronary artery disease, heart failure with reduced ejection fraction (HFrEF), ambient ventricular arrhythmia (PVC, non-sustained ventricular tachycardia (VT), sustained VT), prior cardiac arrest, advanced age, male sex, African American race, left ventricular hypertrophy, congenital heart disease, and cardiac conduction abnormalities, such as bundle branch block. Additionally, patients with underlying channelopathies , including long QT syndrome and Brugada syndrome as well as arrhythmic myopathies, for example arrhythmogenic right ventricular cardiomyopathy, Wolf Parkinson White syndrome and hypertrophic cardiomyopathy, present differently than those with traditional risk from ischemic and non-ischemic cardiomyopathies. Various electrical markers may predict SCD including T-wave alternans, late potentials on signal averaged electrocardiography, inducibility of sustained tachyarrhythmia by programmed electrical stimulation, prolonged QRS and QT intervals, and/or abnormal heart rate variability or turbulence and abnormal baroreflex sensitivity. Additionally, there are further co-morbidities that contribute to and complicate risk stratification including smoking, hypertension, hyperlipidemia, obesity, diabetes mellitus, renal failure, drug abuse, and congenital heart disease. Finally, life style and social factors play a predictive role including activity levels, socioeconomic status, as well as stress (emotional or physical) [1, 4]. Given the complexity and diversity of these factors, it is not surprising that risk stratification of SCD remains a major challenge.

A series of large, multicenter randomized trials demonstrated the benefit of the implantable cardioverter defibrillators (ICD ) for preventing sudden death in selected populations [5,6,7]. This led to a rapid expansion of this therapy, which has likely contributed to a significant decline in incidence of ventricular fibrillation [2, 8, 9] and deaths [8] during out of hospital cardiac arrest. Due to this success and the changing epidemiology of SCD, there has been a renewed focus on risk stratification in this population [3, 10]. However, despite advances in our understanding, SCD remains a significant problem primarily because of a lack of a comprehensive and validated approach to detecting patients at risk. The major risk factors evaluated are presented in Table 6.1 and will be summarized in this chapter.

Table 6.1 Summary of risk stratification tools for sudden cardiac death
Full size table

Genetics and Genomics

Multiple population based studies exist suggesting a strong genetic contribution to individual SCD risk, independent of traditional cardiovascular risk factors. The landmark Paris Prospective Study showed increased risk of SCD in middle aged men with parental sudden death (RR 1.80 (95% CI, 1.11–2.88) [11]. Similarly, the Seattle case-controlled study, demonstrated increased risk of SCD (odds ratio (OR) = 2.69, 95% CI = 1.35–5.36) among patients with a parental history of early onset sudden death (age < 65) [12]. Additionally, family history of sudden death has been shown to be an independent risk factor (OR 2.72, 95% CI 1.84–4.03) for primary ventricular fibrillation in acute myocardial infarction [13]. Despite the multiple studies demonstrating a strong association between family history and SCD, there has yet to be a specific genetic variant or clinical marker identified that has proven effective in predicting individual risk. Furthermore, due to the variability in the mechanism of SCD, there is likely a broad spectrum of heritability of SCD in different populations.

In individuals under 40 years of age, SCD occurs in a Mendelian pattern with cardiomyopathies and electrical disorders being the most prevalent [14]. In individuals above 40 years old, which account for the majority of events, SCD is most commonly caused by ventricular fibrillation in the setting of acute or prior myocardial infarction. To attain a better understanding, genome-wide association studies (GWAS) are now being performed to isolate genetic variants modulating SCD risk, with specific interest in genes that play a role in structural abnormalities, as well as heart rate and ECG indices of slowed conduction and abnormal repolarization [14,15,16,17,18].

GWAS use dense maps of hundreds of thousands of single nucleotide polymorphism (SNP) arrays to identify genotype patterns associated with a particular phenotype [14], in this case SCD. An increasing number of genetic variants are starting to be uncovered with strong associations [19, 20] for SCD, and this approach holds future promise as costs decrease and gene-scanning technologies improve. However, these observations are limited in clinical applicability at this time due to small sample size.

Left Ventricular Ejection Fraction

Left ventricular systolic function, estimated by ejection fraction (LVEF) , is the most common studied marker of SCD risk, and it is clearly a powerful predictor of cardiac mortality [21,22,23]. LVEF is easy to measure, reproducible and predictive in both ischemic and non-ischemic cardiomyopathies. In clinical practice, LVEF has become the primary criterion used for ICD placement. The MADIT II trial demonstrated a significant reduction in SCD and all-cause mortality after ICD placement among patients with previous myocardial infarction and LVEF ≤30% [24]. The SCD-HeFT trial demonstrated decreased all-cause and sudden death mortality after ICD placement in patients with both ischemic and non-ischemic cardiomyopathy, NYHA class II or III functional status and LVEF ≤35% [7]. However, not all studies show a benefit of ICDs among patients with reduced LVEF. The CABG-Patch trial showed no benefit of ICD therapy in patients with EF < 35% undergoing surgical coronary revascularization [25], possibly due to the antiarrhythmic effect or the improved systolic function following revascularization. Similarly, the DINAMIT and IRIS studies showed no benefit of early implantation of ICDs following myocardial infarction despite a reduced ejection fraction [26, 27]. Again, this may be due to improvement in LVEF post MI or competing non-arrhythmic causes of mortality [26]. Hence, the timing of LVEF assessment and of intervention and changes in underlying myocardial status are also important variables to consider. Finally, as noted above, most patients that survive cardiac arrest in more contemporary studies only have mildly depressed or near normal systolic function [5, 28] and the predictive role of LVEF is therefore limited in these populations without significant underlying cardiomyopathy.

Resting Electrocardiogram (QRS and QT Intervals)

The resting electrocardiogram is a non-invasive, inexpensive diagnostic tool that is available in most clinical settings and can provide useful prognostic information. QRS duration represents interventricular conduction time and when prolonged may promote ventricular arrhythmias by altering electrical and mechanical function through abnormal dispersion of depolarization and repolarization and resultant cardiac dyssynchrony. QRS prolongation may also be a marker of more advanced LV dilation and dysfunction. QRS prolongation (≥120 ms) has been shown to predict both overall mortality and SCD in patients with ischemic and non-ischemic cardiomyopathy, independent of LVEF [29]. In a subgroup analysis of the MUSTT trial, QRS duration and the presence of left bundle branch block were found to be independent predictors of overall mortality and SCD in patients with ischemic cardiomyopathy [30]. Despite the above findings, there are a number of conflicting ICD trials that were unable to demonstrate an effect of QRS duration on mortality independent of LVEF [31, 32], limiting its overall predictive value.

The QT interval represents ventricular repolarization, which is routinely corrected for heart rate or RR interval and measured as QTc. A QTc duration greater than 420–440 ms (longer upper limit of normal in females than in males) is associated with a two to threefold increase in cardiac mortality [33,34,35] among patients with and without coronary artery disease. Applied to the general population, a prolonged QT interval and/or increased QT interval dispersion (the maximal inter-lead QT variance in 12 lead electrocardiogram) predicts increased cardiac and total mortality [36]. However, in patients with advance heart failure, QT interval and interval dispersion were unable to predict mortality independent of LVEF [37], also limiting its clinical applicability in patients with significant underlying cardiomyopathy. Of note, these studies did not include subjects with genetic channelopathies including Long QT syndrome in whom a QTc > 500 ms (or the rarer Short QT syndrome with QTC < 350 ms) is associated with increased risk of SCD independent of structural heart disease [38, 39].

Finally, there are a number of ECG abnormalities recognized that carry an increased incidence of SCD in specific situations including Brugada syndrome, Arrhythmogenic Right Ventricular Dysplasia and Wolf Parkinson White Syndrome (Table 6.2).

Table 6.2 Composite risk scores for SCD in Primary Prevention Cohorts
Full size table

Ambulatory Electrocardiogram (Holter Recording of Ventricular Ectopy and Heart Rate Turbulence)

Historically, long term ambulatory electrocardiograph y by Holter monitoring was used to predict SCD in survivors of myocardial infarction. Patients with complex ventricular ectopy defined as more than 10 premature ventricular beats per hour or non-sustained ventricular tachycardia (NSVT) were shown to have increased mortality [40, 41]. However, in one more recent multivariate analysis performed in the thrombolytic/reperfusion era, complex ventricular ectopy was not shown to be an independent predictor of mortality and the presence of post infarct NSVT no longer predicted mortality or arrhythmic events [42]. Studies evaluating complex ventricular ectopy, specifically NSVT among patients with non-ischemic cardiomyopathy have also had conflicting results [43,44,45].

Heart rate turbulence is a non-invasive marker of electrical instability that can be assessed on Holter monitor and has been shown to identify patients at high risk for all-cause mortality and sudden death, specifically in post-infarction and congestive heart failure patients [46,47,48,49]. Heart rate turbulence is believed to reflect baroreflex sensitivity and is a surrogate marker of cardiac autonomic tone. Under normal conditions, a ventricular premature beat (VPB) results in a transient drop in blood pressure triggering baroreceptor activated inhibition of vagal tone and subsequent increase in heart rate. Increased myocardial contractility following a VPB then results in a transient increase in blood pressure with a decrease in sinus node activity. To analyze, the RR intervals following VPBs are assessed for an initial short acceleration followed by a deceleration of heart rate. Absent or diminished biphasic pattern reflects an abnormal response and increased risk of cardiac arrhythmic death [49].

Exercise Electrocardiogram (NSVT and T-wave Alternans)

Exercise stress testing is a commonly performed, non-invasive test used to detect myocardial ischemia and to evaluate exercise capacity and cardiovascular function. Exercise induced NSVT is associated with increased cardiovascular mortal ity within the next de cades [50]. However, exercise induced NSVT rarely occurs in structurally normal hearts (other than in RVOT ventricular ectopy syndromes) and in otherwise healthy individuals is not associated with an increase in cardiovascular or total mortality [51, 52]. NSVT does not predict long term mortality in patients with myocardial infarction treated with coronary artery reperfusion and beta-blockers or dilated cardiomyopathy, independent of LVEF. However, NSVT does have prognostic significance in patients with hypertrophic cardiomyopathy and certain genetic channelopathies [50].

Microvolt T-wave alternans (TWA) is a beat to beat alternation in the morphology and amplitude of the ST-segment or T-wave that has been associated with increased propensity for sustained ventricular arrhythmias [53]. Only rarely is the variability higher than the microvolt range and actually visible to the eye on routine electrocardiographic strips. The effect of TWA is further augmented by increased heart rate, ventricular premature beats, ischemia from coronary artery occlusion or reperfusion, adrenergic stimulation and mental stress [53]. The underlying mechanism is believed to involve enhanced sympathetic nerve activity and abnormal calcium handling that disturbs cardiac repolarization in vulnerable myocardium leading to discordant alternans of repolarization of myocytes. Risk stratification is based on peak TWA measured during symptom limiting bicycle or treadmill exercise testing at a target heart rate range of 105–110 beats/min (and during post-exercise recovery) or during 24 h ambulatory ECG recording. Early studies showed that TWA was a powerful electrocardiographic tool in predicting risk of cardiovascular mortality and SCD in patients with myocardial infarction, as well as in cohorts with ischemic and non-ischemic cardiomyopathy. The predictive value of TWA in patients not utilizing beta-adrenergic blockade agents is significantly less than those on beta-adrenergic therapy [53, 54]. This phenomenon is likely the result of the beta-adrenergic blocker’s effect on the sympathetic nervous system [53], and indicates that beta blockers should not be withheld before the procedure.

Despite considerable enthusiasm over TWA, subsequent large prospective trials failed to validate the role of TWA for risk stratification in patients with reduced LVEF. The TWA SCD HeFT [55] study showed no predictive value to predict mortality or arrhythmic events among patients in the SCD HeFT trial. In addition, the Master study showed that TWA did not predict appropriate ICD therapy in a large cohort of subjects implanted for standard primary prevention indications [56]. These studies have tempered the enthusiasm for this technique in patients with significant underlying cardiomyopathy. More recently, a different analytic approach was developed to assess T-wave alternans in the absence of exercise. This is a somewhat different measure despite the common name. Ongoing studies are assessing its role in stratifying risk of life threatening arrhythmias and SCD among patients with LVEF >35%, specifically to be used in conjunction with other markers of electrical instability such as signal-averaged ECG, heart rate turbulence or heart rate recovery (see CARISMA and AVID studies under Electrophysiology Study). In this capacity, TWA in the presence of one other factor increased hazard ratios up to 212% and positive predictive value by 78% [53]. In this subgroup of patients, its clinical utility remains undetermined.

Signal Averaged Electrocardiography

Signal averaged electrocardiography (SAECG ) is a method of recording and averaging electrocardiographic data with the goal being to detect late potentials [57]. This test was initially found to be a promising marker of sudden cardiac death in patients with ischemic cardiomyopathy and later non-ischemic cardiomyopathy [57,58,59,60]. In a study of 182 patients post acute myocardial infarction from 1992 with 14 month follow up, and primary outcome of sustained ventricular arrhythmia or SCD, SAECG was found to be a significant predictor of SCD with a 2.7 fold increase in risk [61]. El-Sherif et all found similar results when they compared SAECG to clinical acumen, ejection fraction, and ventricular arrhythmia and showed that SAECG provided that best predictive criterion [62]. Despite these findings, some studies that have tried to use SAECG was to predict utility of prophylactic ICD implantation in patients with coronary artery disease and depressed ejection fraction at time of elective coronary bypass surgery have shown no survival benefit of ICD implantation if patients with abnormal SAECG [63].

With regard to non-ischemic cardiomyopathy, Mancini et al., in 1993 were able to show that patients with an abnormal SAECG had a significant increase in their risk for sustained ventricular arrhythmias as compared to patients with normal SAECG or bundle branch block (BBB) [57]. Goedel-Meinen et al. in 2001 demonstrated that SAECG was a useful independent prognostic factor for SCD, mortality, and cardiac events [64]. Subsequent studies have shown conflicting and less promising results. A more recent study performed comparing mIBG to SAECG, heart rate variability, and QT dispersion showed that mIGB and not SAECG, heart rate variability or QT dispersion was a predictor of SCD in patients with mild to moderate ejection fraction [65]. It should also be noted that in a study by Brembilla-Perrot et al. showed, that in 128 patients with dilated non-ischemic cardiomyopathy, that in the presence of BBB the predictive value and specificity of SAECG is significantly diminished and that the SAECG itself did could not predict the risk of sudden cardiac death and mortality [66].

A head to head comparison between TWA and SAECG for predicting arrhythmia risk assessment showed that TWA a was more sensitive predictor of ventricular tacharrhythmia and death. It did validate the ability of SAECG to predict risk of spontaneous ventricular arrhythmias though not as well as TWA [67]. Therefore, while an abnormal SAECG may be a marker of risk for SCD, its value, especially as compared to newer modalities, is limited especially when it comes to predicting need for ICD. However, it may be useful in combination but not in isolation.

Electrophysiology Study

Electrophysiology study (EPS ) is an invasive procedure designed to diagnose disorders of the cardiac conduction by placing catheters in the right atrium, right ventricle and across the triscupid valve. With this tool and the prospect of delineating high risk patients for sudden cardiac death and malignant ventricular arrhythmias, both the MADIT and the Multicenter Unsustained Tachycardia Trial (MUSTT) were designed and implemented [68].

In MUSTT, patients with coronary artery disease, left ventricular ejection fraction less than or equal to 40% and asymptomatic non-sustained ventricular tachycardia were enrolled at 85 study sites [68]. Patients who met inclusion criteria and had sustained VT induced by EPS were randomized to cardiac medications only or a strategy of arrhythmia suppression first with antiarrhythmic drugs to suppress the inducibility of VT and if unsuccessful then an ICD. A survival benefit only among patients treated with ICDs and not antiarrhythmic drugs [68]. There was an absolute reduction in risk of cardiac arrest or death from arrhythmia by 7% after 5 years of follow up [68].

The Multicenter Automatic Defibrillator Implantation Trial (MADIT), which enrolled as a similar population with a lower ejection fraction, equal to or less than 35%. In MADIT, patients with inducible VT were randomized to an ICD or to “conventional care” [69]. These trials were followed up by the Multicenter Automatic Defibrillator Implantation Trial (MADIT) II which implanted ICDs in patients with ischemic cardiomyopathy in the absence of EP study for risk stratification [70]. This study included patients with a left ventricular ejection fraction ≤30%. A substudy of MADIT II performed programmed ventricular stimulation through the implanted ICD. This showed that inducible ventricular tachycardia (VT) did correlate with instances of VT but there was an inverse relationship with ventricular fibrillation requiring defibrillation [70]. Interestingly they showed that patients with non-inducible VT had a substantial burden of VT and a higher ventricular fibrillation burden than inducible patients, which casts major doubts on the utility and predictive value of EPS [70].

The above findings are supported by the Antiarrhythmics Versus Implantable Defibrillators (AVID) trial and the Electrophysiologic Study Versus Electrocardiographic Monitoring (EVSEM) trial [70, 71]. The AVID trial included patients with both ischemic and non-ischemic cardiomyopathies [71]. Investigators found that EPS had a limited predictive value for death though it did predict recurrent VT or ventricular fibrillation in patients with inducible VT with rates less than or equal to 200 beats per minute and tended to predict recurrent VT or ventricular fibrillation in patients with coronary artery disease [71]. However, again, EPS tended to miss patients who were at risk for VT or ventricular fibrillation by classifying them as non-inducible [71].

The CARISMA trial was designed to evaluate the predictive value of both invasive and non-invasive strategies of risk assessment for potentially lethal arrhythmias [72]. The study enrolled 5869 patients who had recent myocardial infarctions in ten European countries [72]. The inclusion criteria included left ventricular ejection fraction less than or equal to 40% measured 3–5 days after acute myocardial infarction not undergoing coronary artery bypass surgery [72]. The primary endpoint, unlike earlier studies, was not mortality but rather fatal or near fatal arrhythmias that could have been treated after consensus agreement by a five member committee reviewing the case [72]. Secondary endpoint was all-cause mortality in this study [72]. Monitoring was performed primarily with implantable loop recorders [72]. Risk factors evaluated included LVEF by echocardiography, heart rate variability by Holter Monitoring, TWA, SAECG, EPS, and QT analysis with 12 lead electrocardiogram at index and 6 weeks post acute myocardial infarction [72]. Investigators found that heart rate variability and turbulence at 6 weeks rather than index was the most significant predictor of potential lethal arrhythmia [72]. EPS was a predictor of the primary endpoint but less so than heart rate variability [72]. Interestingly, LVEF and TWA were not found to be a statistically significant predictor of treatable arrhythmia [72]. Another recent study, the ABCD trial, which was designed to evaluate the utility of TWA, EPS or combination, in selecting patients with ischemic cardiomyopathy with an ejection fraction less than or equal to 40% who may benefit from ICD therapy [73]. This study suggested that TWA and EPS were comparable at 1 year and additive in combination in their relative predictive values and might be a tool to help determine patients who would benefit least from ICD therapy in a time dependent way [73]. The utility of EPS may be in these select populations but even so, is limited and not specific enough to be used in a more broad patient population.

Imaging: Assessing Scar Burden

Contrast enhanced cardiac magnetic resonance (cCMR) is an imaging modality that utilizes the power and spatial resolution of MRI images to understand and evaluate cardiac anatomy more accurately [74, 75]. Kim et al. were one of the first groups to demonstrate the potential power of cCMR in that it can distinguish between reversible and irreversible ischemic injury independent of wall motion or infarct age [74]. There are two temporal MRI images that can be used to understand cardiac anatomy and scar burden [74, 75]. The early contrast enhancement phase images are acquired within seconds and reflect perfusion while delayed/late contrast enhancement phase images are captured after several minutes and reflect a myriad of pathophysiological information, including, most importantly, myocardial scar [74, 76].

In 2007, Roes et al. published data that suggested cCMR delineated myocardial scar may be a superior predictor of mortality than LVEF and LV volume in patients with healed myocardial infarction [77]. Despite a number of studies showing a clear association between risk of malignant arrhythmia and myocardial scar, a quantitative continuous relationship has not yet been demonstrated [78]. This suggests that, while scar is an important substrate for VT in patients with ischemic cardiomyopathy, it alone does not correlate with risk for malignant arrhythmia [78] (Fig. 6.1).

Fig. 6.1
figure 1

Impact of scar on mortality. Kaplan-Meier curve analysis showing difference in mortality when patients are stratified according to a large extent (>6) vs. a small extent (<6) of scar tissue on contrast enhanced MRI (Reproduced with permission from Roes et al. [77])

Full size image

In 2009, Kadish et al. designed The Defibrillators To Reduce Risk By Magnetic Resonance Imaging Evaluation study (DETERMINE) as a prospective, multi-center, randomized, clinical trial to investigate the effects of ICD implantation in patients with coronary artery disease (CAD) and mild to moderate LV dysfunction [2, 79]. Their goal was to test the hypothesis that patients with LV infarct mass greater than or equal to 10% of myocardium, by cCMR, on both medical therapy and a functional ICD had a mortality benefit over medical therapy alone [2, 79]. Unfortunately, because of poor enrollment this study was discontinued [2].

Imaging: Assessing Sympathetic Denervation (MIBG)

The autonomic nervous system consists of the sympathetic and parasympathetic components. Under normal circumstances, there is a balance with shifts from one system to the other depending on physiologic requirements. Chronic imbalances may occur in the setting of decreasing LVEF , heart failure and increasing left ventricular end diastolic pressure that may further worsen ventricular dysfunction [2, 80, 81].

In 1993, Mitrani et al. first showed in a small cohort of 18 patients with sympathetic denervation an increase in risk of ventricular tachycardia, even in the absence of coronary artery disease [82]. Ischemic heart disease is a major risk factor for the development of sympathetic denervation, even if only for short durations (minutes) [83]. As part of the remodeling process, denervation leads to increased sympathetic tone resulting in decreased pre-synaptic norepinephrine uptake and post-synaptic beta-adrenoceptors mass in a tachyphylaxis like response resulting in global left ventricular remodeling [84]. The over-stimulation by the sympathetic nervous system results in both supraventricular and ventricular arrhythmias by increasing automatic, triggering automaticity, and reentrance [85, 86].

Computed tomography using radioiodinated metaiodobenzylguanidine (mIBG ) is now recognized as a powerful tool to identify inhomogeneity of the sympathetic nervous system within the cardiac myocardium, a potentially important substrate for sudden cardiac death [86,87,88]. mIBG was initially developed as a marker for adrenal medulla and related adrenergic tumors [87]. It is an analog of norepinephrine that is able to concentrate in sympathetic neurons within the heart [89]. The concentration of mIBG within cardiac neurons directly correlate with the neuronal integrity and function [89]. In 1979, mIBG was first synthesized and tested in both animals and humans [87]. In the human myocardium, mIBG, is preferentially transported by the cardiac neuronal norepinephrine transporter into cardiac sympathetic neurons [87]. mIBG is not metabolized and therefore can be imaged using planar or single photon emission computed tomography after uptake as it is not metabolized as norepinephrine [84]. In a study of 116 patients who underwent mIBG imaging prior to implantation of ICD, Boogers and colleagues showed that this marker was an independent predictor of ventricular arrhythmias that would require ICD therapy [84].

The late heart to mediastinum ratio (HMR) on mIBG has also been found to be an independent predictor of mortality [84]. Early images are typically obtained within an hour of infusion while late images are acquired after 3 h [84]. Late HMR is calculated dividing the regions of interest of the heart mean counts (H) by the regions of interest of the mediastinum mean counts (M) [84]. In the prospective AdreView Myocardial Imaging for Risk Evaluation of Heart Failure (ADMIRE-HF) study of 961 patients with NYHA functional class II/III CHF with an LVEF less than 35%, 237 subjects experienced functional class progression, life threatening arrhythmic events or cardiac death [90]. These investigators compared patients with a late HMR less than 1.6 to those with values greater than or equal to 1.6 [90]. They observed that patients with HMR less than 1.6 were at significant risk for all endpoints including progression of heart failure, arrhythmic events, cardiac death and all-cause mortality (Fig. 6.2), [90]. Multivariable analysis found that HMR, LVEF, BNP, and NYHA class were independent contributors to risk model [90].

Fig. 6.2
figure 2

Sympathetic denervation and cardiac outcomes. Cumulative event curves comparing subjects with H/M <1.6 vs. >1.6. (a) Composite primary end point; (b) heart failure progression; (c) arrhythmic event; (d) cardiac death; (e) all cause mortality. H/M = heart/mediastinum ratio on Computed tomography using radioiodinated metaiobenzylguanidine (mIBG) (Reproduced with permission from Jacobson et al. [90])

Full size image

Finally, the wash out rate (WR), which assesses the retention of norepinephrine by neurons, has been shown to be a potent predictor of sudden cardiac death [84]. Washout is calculated by subtracting the difference between the early H and M by the difference between the late H and M all divided by the difference between early H and M [84]. Tamaki et al. enrolled 106 patients with congestive heart failure, as defined by Framingham criteria, with an LVEF less than 40% [65]. Patients needed to be stable on angiotensin-converting enzyme inhibitor (ACEI), diuretics, and digoxin for at least 3 months [65]. Their exclusion criteria included significant renal failure, insulin dependent diabetes, autonomic neuropathy and beta blockers use [65]. Multivariate analysis showed that abnormal WR (>27%) and depressed LVEF were both independent predictors of sudden cardiac death. Among patients with LVEF >35%, there was a significantly higher rate of sudden cardiac death in those with abnormal WR [65]. Their results were validated by meta-analysis of 18 studies with a total of 1755 patients by Verberne et al. [91]. WR has also been shown to be a significant independent predictor of MACE in patients with STEMI [92]. The applicability of this approach may be limited as beta blocker use is ubiquitous and a Class I indication among patients with ischemic myocardial infarction.

Imaging: Assessing Sympathetic Denervation (Positron Emission Tomography)

Positron emission tomography (PET) is another imaging modality used in the evaluation of myocardial sympathetic innervation [2]. PET scanners work by detecting the radiation released by isotope emitting positrons after their annihilation within tissues [93]. PET is typically paired with computed tomography or magnetic resonance imaging [93]. In addition to evaluating myocardial sympathetic innervation, PET can also assess myocardial perfusion, metabolism (stunned or hibernating myocardium), and systolic function [93].

[11C]-meta-Hydroxyephedrine (HED) is a radioligand developed for PET to evaluate the sympathetic nervous system. It is a catecholamine analog labeled with 11C with a half-life of approximately 20 min that was developed based on metaraminol [90]. Like mIBG, HED is not metabolized by catechol-O-methyl transferase (COMT) or monoamine oxidase (MAO) [90]. Therefore it can remain in the sympathetic neurons long enough to be imaged. Based on uptake and retention of HED, Luisi et al. were able to show that hibernating myocardium in farm-bred pigs has significant regional reduction in norepinephrine reuptake [88].

The PAREPET trial was a prospective observational trial designed to study the hypothesis that inhomogeneity in human myocardial sympathetic innervation and/or hibernating myocardium could predict risk for arrhythmic death independent of left ventricular function [94]. This study enrolled 204 patients with inclusion criteria of ischemic heart disease and heart failure with LVEF >35% on optimal medical therapy who were not candidates for coronary revascularization and not eligible for primary prevention ICD [94]. Exclusion criteria included patients with prior cardiac arrest or ICD discharge, recent infarction (less than 30 days), or revascularization (PCI within 3 months or bypass grafting within last year) [94]. They also evaluated perfusion, using 13 N–ammonia, and viability using 18F-2deoxyglucose [94].

The primary endpoint of this trial was sudden cardiac arrest, which was defined as arrhythmic death or ICD discharge for ventricular fibrillation or ventricular tachycardia greater than 240 beats per minute over a 4 year follow up period [94]. Infarct volume and LVEF were not predictors of sudden cardiac arrest [94]. However, patients who later suffered from sudden cardiac arrest had a greater sympathetic denervation burden as seen in viable denervated myocardium (33 ± 10% vs. 26 ± % of LV; p = 0.001). Using multivariate analysis they developed a four component prediction system that focused on independent variables that predicted arrhythmic risk [94]. These variables included percentage of denervated myocardium (>37.6%), left ventricular end-diastolic volume index (>99 mL/m2), creatinine (>1.49 mg/dL), and no angiotensin inhibition therapy (Fig. 6.3). More than 40% of the study population had none of the independent risk factors and an annual rate of sudden cardiac arrest of less than 1%, which is lower than the rate of arrhythmic death in patients with known coronary artery disease and LVEF between 35 and 50% [94]. About 36% of the cohort had only one predictive risk factor and an annual risk of sudden cardiac arrest of approximately 4% [94]. The remainder of the population, approximately 20%, had two or more independent risk factors and an annual rate of sudden cardiac arrest of ~12% [94]. Interestingly, infarct size, left ventricular ejection fraction, BNP and other variables did not improve their predictive model.

Fig. 6.3
figure 3

PET Scan and other Clinical Factors to Predict SCD. Kaplan-Meier curves illustrating significant differences in the incidence of sudden cardiac arrest (SCA) in relation to the number of risk factors present (p < 0.0001). Subjects with no risk factors had an annual rate of SCA < 1%; 1 risk factor had an annual rate of ~4% and 2 or more risk factors had an annual risk of ~12% (Reproduced with permission from Fallavollita et al. [94])

Full size image

Despite these encouraging results, the multivariate analysis was post hoc so it needs to be replicated to strengthen this observation. Other limitations include the use of HED, which has a short half-life requiring close proximity to cyclotron, and exclusions of patients with recent myocardial infarction, non-ischemic cardiomyopathy and patients with more preserved left ventricular systolic function. Of note, 18-Flourine labeled norepinephrine analogs may improve clinic utility of PET because of improved half-life as compared to HED.

Summary

SCD remains a major health care problem, particularly in western cultures. Early studies identified LVEF, ischemic heart disease, heart failure and ambient arrhythmias as predictors of events. This led to large multicenter trials establishing the role of the ICD for primary prevention of SCD. However, a changing epidemiology of SCD has confirmed an unmet need for risk stratification. In addition, there is interest in improved risk stratification of ICD eligible patients. In this regard, newer imaging, advanced electrocardiographic and genetic techniques raise hope that specific markers or a combinations of tests will allow identification of high risk subjects who can benefit from specific antiarrhythmic therapy, including those with uncommon arrhythmia substrates. A summary of risk stratification approaches for different groups at risk for SCD is summarized in Table 6.3. However, despite these newer techniques and promising markers of SCD, most studies remain observational with frequent posthoc analysis. Accordingly, contemporary risk stratification and guidelines for therapy have changed little over the past decade to address the unmet need of prevention of SCD.

Table 6.3 Patient subgroup evaluations for primary prevention ICD placement
Full size table

Beyond clinical interest, the cost and practicality of risk stratification strategies and ICD implantation is significant. The cost of genetic testing, ejection fraction evaluation, resting electrocardiograms, exercise electrocardiogram, ambulatory electrocardiogram, SAECG, EPS, imaging of myocardial scar, and innervation studies is significant. It is unlikely that funding aimed at limiting testing and device therapy could be funded any other way than by public means, which is on the decline. However understanding both the cost and benefit of evaluation of SCD should of major importance to governments, institutions, providers and patients as medical costs continue to rise. Thus far there has been no cost benefit and cost effectiveness studies looking at SCD that could help guide a reasonable and appropriate approach to SCD risk stratification wide scale. Studies of this nature would need to compare cost of strategy to cost of devices and patient benefit indices. While this would be a daunting task, it would also be an extremely worthwhile and immensely beneficial enterprise.