The Spectrum of Epidemiology Underlying Sudden Cardiac Death

Abstract

Sudden cardiac death (SCD) from cardiac arrest is a major international public health problem accounting for an estimated 15–20% of all deaths. Although resuscitation rates are generally improving throughout the world, the majority of individuals who suffer a sudden cardiac arrest will not survive. SCD most often develops in older adults with acquired structural heart disease, but it also rarely occurs in the young, where it is more commonly due to inherited disorders. Coronary heart disease (CHD) is known to be the most common pathology underlying SCD, followed by cardiomyopathies, inherited arrhythmia syndromes, and valvular heart disease. Over the past three decades, declines in SCD rates have not been as steep as for other causes of CHD deaths, and there is a growing fraction of SCDs not due to CHD and/or ventricular arrhythmias, particularly among certain subsets of the population. The growing heterogeneity of the pathologies and mechanisms underlying SCD present major challenges for SCD prevention, which are magnified further by a frequent lack of recognition of the underlying cardiac condition prior to death. Multifaceted preventative approaches, which address risk factors in seemingly low risk and known high-risk populations will be required to decrease the burden of SCD. In this Compendium, we review the wide-ranging spectrum of epidemiology underlying SCD within both the general population and in high-risk subsets with established cardiac disease placing an emphasis on recent global trends, remaining uncertainties, and potential targeted preventive strategies.

SCD/SCA: Background, Mechanisms and Risks

Introduction

Sudden cardiac death (SCD)/sudden cardiac arrest (SCA) refers to an unexpected death or arrest from a cardiovascular cause that occurs rapidly outside of the hospital or in the emergency room (ER)^{1, 2}. The presumption based upon epidemiologic studies of SCD and SCA survivors is that such rapid deaths are often due to lethal ventricular arrhythmias in the setting of underlying coronary heart disease (CHD)^3–5. Despite major advances in treatment and prevention of CHD and implantable cardioverter defibrillators (ICDs) for SCD prevention in high-risk patients, SCD remains a major public health problem estimated to account for 15–20% of all deaths^{6, 7}. Reported declines in SCD rates⁸ have not been as steep as for other causes of CHD death^9–12, and the reasons for this disparity are not well understood. There may be a growing fraction of SCDs not due to CHD and/or ventricular arrhythmias, particularly among certain subsets of the population. In addition, SCD preventive strategies are lacking in low-risk individuals without established heart disease that comprise the largest proportion of SCDs^{5, 13,14}. In order to further reduce the incidence of SCD, preventive strategies need to be tailored to diverse populations at varying levels of risk. In this Compendium, we review the broad spectrum of epidemiology underlying SCD, from common to rare forms, with an emphasis on preventive strategies, recent trends, and unanswered questions.

SCA Incidence: Estimates and Definitions

Estimates regarding the annual incidence of SCA and SCD vary widely depending on data sources for case ascertainment, definitions employed, and methods utilized for extrapolation of rates.^{15, 16}. These difficulties in extrapolating SCA and SCD rates are likely magnified further when comparing SCD rates across countries where EMS protocols, autopsy rates, and national recording systems vary. The majority of global comparisons (Figure 1) are based upon rates of emergency medical service (EMS) attended out of hospital cardiac arrests (OHCA), which appear to be much lower in Asia (52.5 per 100,000 person-years) as compared to Europe (86.4 per 100,000 person-years), North America (98.1 per 100,000 person-years), and Australia (111.9 per 100,000 person-years). There also appear to be regional variations within geographic regions¹⁷. For instance, among 10 regions in North America, rates of EMS attended cardiac arrest range from 159 per 100,000 person-years in Dallas, Texas to 71.8 per 100,000 person-years in Ottawa, Ontario.

Figure 1. Incidence rates of EMS attended OHCA (A), EMS treated OHCA (B), and EMS treated OHCA of presumed cardiac cause (C).

Incidence is per 100,000 person-years. Compared to Europe, North America, and Australia, EMS attended OHCA was lower in Asia, and EMS treated OHCA of presumed cardiac cause was higher in North America than in other regions. *P < 0.05; **P < 0.01; ***P < 0.001.

EMS indicates emergency medical service; OHCA, out of hospital cardiac arrest. Adapted from Berdowski et al¹⁷ with permission.

However, the above estimates are crude approximations which at the same time both over- and under- estimate SCA rates. First, EMS is not in attendance for a significant fraction of SCAs, and the proportion of EMS attended deaths is known to vary significantly across countries. Second, a significant fraction of EMS attended OHCAs are not unexpected nor do they occur in a short time frame from the onset of symptoms. Death certificates are also known to overestimate SCD rates for similar reasons¹⁸. To obtain a more precise estimate of SCD/SCA, expert panels have advocated for the establishment of precise and uniform definitions of SCD/SCA and to integrate multiple source methods for case ascertainment ^{2, 15}. Standardized definitions of SCD/SCA have been proposed, which generally define SCD as an unexpected death without obvious extra-cardiac cause that occurs in association with a witnessed rapid collapse or within one hour of the onset of symptoms^{1, 2, 19}. There are no national surveillance mechanisms to record such characteristics of deaths; and therefore, approximations are based on extrapolations from population-based studies. In prospective studies utilizing standardized definitions and multiple sources of surveillance for case ascertainment in the United States¹⁸, Netherlands²⁰, Ireland²¹, and China²², SCD rates range from 40–100 per 100,000 in the general population², with rates being lowest in China²². In individuals of <35 years old, SCD is rare with an incidence of 1 to 3 per 100,000 per year in recent reports^23–25.

Even when a strict definition and multiple sources of ascertainment are used, other non-cardiac conditions that evolve rapidly such as acute cerebral hemorrhage, aortic rupture, and pulmonary embolism cannot be excluded without a carefully performed autopsy. Autopsy rates are generally low and vary widely across countries with rates as low as 10 % of all deaths within the United States²⁶ compared to 23.8% in Finland ²⁷, and the protocols for the performance of autopsies in the cases of suspected SCD vary widely as well, even within regions of countries. These differences in autopsy rates and protocols likely contribute to some of the geographical differences in the incidence and causes of sudden cardiac death.

SCA Trends in Survival and Underlying Rhythm

Several major advances in CPR²⁸ and post resuscitation care have resulted in improved resuscitation rates from OHCA. In a recent report from the Cardiac Arrest Registry to Enhance Survival (CARES), a prospective clinical registry of 70,000 OHCA survivors in the United States, survival rates to hospital discharge increased from 5.7% in 2005 to 8.3% in 2012²⁹. In Denmark, even greater increases in 30 day survival (3.5% to 10.8%) were observed from 2001 to 2010³⁰. Both in-hospital and pre-hospital survival rates contribute to these improved outcomes post OHCA. However, even with these improvements, absolute survival rates remain in the 10% range or less.

Although survival rates are higher for OHCA where ventricular fibrillation (VF) is the initial rhythm (21%), the proportion of cases where VF is found at the time of EMS arrival has been declining over the last three decades^{31, 32}, with a resultant increase in cases where pulseless electrical activity (PEA) and asystole are the initial rhythm³³. This is an unsettling trend since resuscitation rates are much lower for these rhythms, and we currently have no known strategies for prevention of these deaths³³. Part of this changing pattern appears to be explained by a concomitant increase in the proportion of arrests occurring in the home,^{4, 20, 34} where the arrest is less likely to be witnessed. However, even when the arrest is witnessed by a bystander or an AED is applied, VF or pulseless ventricular tachycardia (VT) is less likely to be encountered as the initial rhythm in arrests occurring in the home versus in public⁴. Proposed explanations for the proportional decline in VF as compared to other rhythms include an overall decrease in the prevalence of CHD, and an increased used of beta blockers and ICDs in high risk patients ^27,29. At the same time, the population is aging, and advances in medical treatments have resulted in an increased prevalence of end-stage cardiovascular disease (CVD) and as well as other severe comorbidities. These older, sicker patients may be more likely to have arrests in the home setting and to have acute precipitants leading to PEA (i.e. respiratory, metabolic, vascular) ^{33, 35}, and/or be less likely to sustain VF up to the point of EMS arrival.

Demographics of SCD Victims

The majority of SCDs occur in the adult population, with less than 1% occurring in individuals less than age 35¹⁰. Among adults, the absolute rate of SCD increases markedly with age; however, the proportion of deaths that are sudden appears to be higher in younger age groups^{5, 19, 36}. There are also recognized differences in SCD incidence by sex and race, which are largely unexplained. Women have a lower incidence of SCD and SCA than men³⁷, even when one accounts for the prevalence of other predisposing conditions such as CHD, myocardial infarction (MI), and heart failure (HF)^{13, 38, 39}. Women who suffer OHCA are on average older, more likely to present with PEA and/or experience their arrest at home as compared to men⁴⁰. These demographics may partially explain why the decline in SCD and OHCA rate has been less pronounced among women as opposed to men in recent years ^{10, 12, 40}. On the other hand, women, especially at younger ages⁴¹, appear to have a higher rate of successful resuscitation and survival from shockable rhythms⁴⁰, possibly due to favorable effects of smaller body size and/or estrogen on success of defibrillation and/or post-resuscitation hemodynamics.

With respect to race, black as opposed to white Americans have been documented to have higher rates of OHCA^{42, 43} and SCD^{44, 45}, as well as poorer rates of survival from cardiac arrest⁴⁶. Similar to women, blacks of both sexes are more likely to have an unwitnessed arrest or PEA documented at the time of the arrest^{42, 46, 47}. These unfavorable arrest characteristics do not entirely account for the poorer survival among blacks. Even when limited to OHCAs due to VF/VT, national rates of survival to hospital discharge have been documented to be 27% lower among black patients, and much, but not all, of this disparity appears to be explained by black patients receiving treatment at hospitals with worse outcomes⁴⁸. Blacks may also be less likely to receive pre-hospital resuscitation efforts in the United States. In one recent large cohort study, patients with OHCA in low income black neighborhoods were less likely to receive bystander initiated CPR than those in high income white neighborhoods⁴⁹.

Data are even more limited for other racial and ethnic differences in SCD incidence. Despite having a higher prevalence of cardiac risk factors⁵⁰, Hispanic Americans may have lower SCD rates than non-Hispanic populations based upon limited data from death certificates^{10, 51} and coroner evaluations in the United States⁵². It also appears that the incidence of SCD may be lower among Asian populations in the United States based upon death certificate data ¹⁰. Estimates of SCD incidence in longitudinal population based studies of SCD in China²² and Japan⁵³ are consistently lower than those from studies performed in North America or other regions with predominantly white populations. These racial differences in SCD/SCA incidence and survival are poorly understood, and further studies performed in large-scale population-based cohort studies of diverse ethnicity are needed to determine the origin of these disparities.

Underlying Pathophysiology of SCD

The epidemiology of SCD is directly related to the pathophysiology that underlies the event. Our knowledge regarding the predominant pathologies underlying SCD is primarily dependent on autopsy series and cardiac evaluations in cardiac arrest survivors, the detail nature of which may vary significantly among counties. Variation in the meticulous nature of histologic examinations across countries likely influences the reported proportions of pathologic causes of sudden death worldwide. Despite these limitations, it is generally accepted that CHD is the most common cardiac pathology underlying SCD (Figure 2) in adults over age 35, particularly among white men where it is responsible for approximately 70–75% of SCDs^{7, 16, 20, 54}. In women, the percentage of SCD and SCA due to CHD appears to be lower. In cardiac arrest survivor series⁵⁵ and SCD autopsy series⁵⁴, CHD was found in 45–50% of women versus 80–90% of men⁵⁴. The percentage of SCDs with underlying CHD also appears to be lower in blacks versus whites (47% versus 63%) and left ventricular hypertrophy is more common among older black than white SCD victims⁵⁶. In Japan, CHD is thought to account for a much lower percentage of SCDs⁵³, although the percentage due to CHD appears to be increasing over time⁵⁷.

Figure 2. Causes of SCD and rates (A), and age of SCD onset in each disease (B).

A. Coronary heart disease is the leading cause of SCD, but the rates of baseline heart disease differ between Western countries and Japan.

B. SCDs occur in elderly populations in coronary heart disease and valvular heart disease, whereas most SCDs in CPVT and LQTS develop at age less than 35 years.

ARVC indicates arrhythmogenic right ventricular cardiomyopathy; BrS, Brugada syndrome; CPVT, catecholaminergic polymorphic ventricular tachycardia; ERS, early repolarization syndrome; HCM, hypertrophic cardiomyopathy; LQTS, long QT syndrome; NIDCM, non-ischemic dilated cardiomyopathy; PUFA, polyunsaturated fatty acids; SCD, sudden cardiac death.

Beyond CHD, the causes of SCD are heterogeneous and include cardiomyopathies, valvular heart disease, myocarditis, hypertrophy, and primary electrical heart disease accounting for the remainder. (Figure 2)⁷. On average, approximately 5% of SCDs or cardiac arrests, a significant cardiac abnormality is not found after clinical evaluation in SCA survivors or at autopsy in SCD victims^{55, 58}. This percentage appears to be higher in women, where structurally normal hearts are more commonly encountered.^{54, 55, 59} In Asians, the primary ion channelopathies are estimated to be responsible for 10% of SCDs⁶⁰. In young adults and children less than age 35, CHD accounts for a much smaller proportion of deaths, with hypertrophic cardiomyopathy (HCM), coronary artery anomalies, myocarditis, arrhythmogenic right ventricular cardiomyopathy (ARVC), and primary ion channelopathies accounting for significant proportions⁶¹.

The presumed mechanism underlying an abrupt, unheralded death in these conditions is electrical instability leading to a lethal arrhythmia triggered by ischemia or other arrhythmogenic stimuli resulting in acute hemodynamic collapse. This hypothesis is difficult to prove as most deaths are not monitored, and those that are comprise a highly selected population. Studies in epidemiologic cohorts of men³ and women⁵ from the 1970s to 1990s suggest that 88 to 91 percent of deaths that occur within one hour of symptom onset are arrhythmic in nature. Since VF degenerates to asystole over the course of several minutes, the majority of SCD victims demonstrate asystole or PEA when first examined by rescue teams⁴⁷. In cases of SCD where there has been a relatively short delay between collapse and the initial determination of rhythm, the proportion of cases with documented ventricular tachyarrhythmias increases to 75–80%^{4, 62, 63}. However, as mentioned previously, VF is less often and PEA is more commonly encountered in recent OHCA series³¹. Therefore, a proportion of SCD is likely due to abrupt hemodynamic collapse in the absence of preceding fatal arrhythmia, and this proportion may be growing in the population.

Risk Factors and Predisposing Conditions for SCD in the General Population

The presence of overt structural and/or primary electrical heart disease is associated with major elevations in SCD risk, and separate risk stratification schema exist for the majority of these disorders which will discussed in later sections. However, the majority of SCDs occur among individuals without clinically recognized heart disease^{5, 13,14}. Approximately 44–52% of men and 59–69% of women who suffer SCD will not have had CVD diagnosed prior to the event; and therefore, SCD is the first manifestation of heart disease^{5, 13,14}. Although the absolute incidence among individuals without apparent heart disease is low, the majority of SCD events take place in this segment of the population. For this segment of the population, current efforts directed at preventing SCD are primarily comprised of risk factor and lifestyle modification.

CHD Risk Factors

As described above, CHD underlies a significant proportion of SCD; thus, risk factors for CHD are associated with SCD risk in the population. Hypertension, diabetes, hypercholesterolemia, obesity, and smoking have all been associated with elevated risks of SCD among men and women in prospective cohort studies^{5, 13, 44, 64, 65}. Diabetes is a particularly strong risk factor for SCD^{64, 66}, even in higher risk populations^{67, 68}. Hypertension and resultant left ventricular hypertrophy (LVH) appear to be particularly important markers of SCD risk in blacks^{45, 56}, in whom the prevalence of these conditions is greater⁶⁹. Smoking confers marked elevations in SCD risk, especially among women⁵. Importantly, smoking cessation is associated with a prompt reduction in the elevated risk for SCDs^70–72 (Figure 3), particularly among individuals who have not yet developed overt CHD⁷². Serum cholesterol appears to be more strongly related to SCD at younger ages^{5, 38}, and a recent meta-analysis of randomized trials suggests that cholesterol lowering with statins may confer modest benefits on SCD incidence⁷³.

Figure 3. Reduction in SCD risk associated with smoking cessation among U.S middle-aged women.

The reference category is current smokers. The white diamond represents age-adjusted HR. The black diamond represents multivariable-adjusted HR. P value for trend <0.0001 in age and multivariable adjusted models.

HR indicates hazard ratio; SCD, sudden cardiac death. Adapted from Sandhu et al⁷² with permission.

Family History of SCD

Several studies have demonstrated a familial predisposition to SCD and/or VF^{64, 74–76}. Three separate case-control studies have demonstrate that a history of SCD among a first-degree relative is an independent risk factor for VF^{75, 77} or SCD⁷⁶ in the setting of an acute MI. In the Paris Prospective Study⁶⁴, parental history of SCD was an independent risk factor for occurrence of SCD (RR = 1.80; 95% CI 1.11 to 2.88); but was not associated with fatal MI. Conversely, a parental history of fatal MI had no effect on SCD risk. These data in aggregate suggest that genetic or unknown environmental factors responsible for the familial aggregation of SCD or ischemic VF may predispose to fatal arrhythmia as a discrete trait and/or manifestation of CHD. The consistent associations implicating a family history of arrhythmic death as an independent risk factor for SCD in the general population has led to several studies focused on identifying common genetic variants that predispose to ventricular arrhythmias and SCD in the population^{78, 79}.

Diet

Dietary intake and blood-based measures of selected nutrients have been specifically associated with SCD in epidemiologic studies. In observational studies, consuming fish ~1–2 times per week has been associated with significant 42–50% reductions in SCD risk, with minimal impact of risk of non-fatal MI^80–82. These inverse associations with SCD were more extreme when marine n-3 fatty acids [eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)] were estimated as a proportion of fatty acids from the diet⁸³ or measured directly in blood^{80, 84}. These data in relatively healthy observational cohorts are supportive of experimental data suggesting that n-3 fatty acids may have a selective effect on susceptibility to arrhythmias⁸⁵. However, randomized trial data in post MI populations have not been consistently supportive of this hypothesis^86–88.

Magnesium intake has also been related to SCD rates. In the Nurses’ Health Study, the relative risk of SCD was significantly lower among women in the highest quartile of dietary magnesium intake. The inverse association was stronger for plasma magnesium, where each 0.25 mg/dL (1-SD) increment in plasma magnesium was associated with a 41% reduced risk of SCD⁸⁹. A similar inverse association between serum magnesium and SCD was also found in the Atherosclerosis Risk in Communities study⁹⁰. Finally, there are likely additive and interactive effects of these and other nutrients on SCD incidence⁹¹. Recent data suggest that a Mediterranean-style diet pattern, consisting of higher intake of vegetables, fruits, nuts, whole grains, fish, and low intake of red/processed meat, may also lower SCD risk among women^{92, 93}.

Alcohol Intake

The relationship between alcohol intake and SCD is complex. Prospective U.S. cohort studies comprised of individuals consuming small-to-moderate amounts of alcohol^94–96 have found U-shaped associations between recent alcohol intake and SCD with reduced risks at levels of ½ to 1 drink per day and no reduction at 2 or more drinks per day. Heavy levels of alcohol consumption (> 6 drinks per day) have also been associated with increased risk for SCD in other populations⁹⁷. In contrast, alcohol intake has an inverse linear association with non-fatal MI⁹⁸. Recently, consuming above one drink per day was found to be associated with 2 fold elevations in the risk of experiencing VF during acute ST elevation myocardial infarction⁷⁷. These data in aggregate suggest that the favorable effects of alcohol on atherosclerosis and thrombosis may be offset by potential proarrhythmic effects at higher levels of intake.

AF, Renal Disease, and OSA

Recent data has highlighted the potential link between atrial fibrillation (AF) and SCD. In patients with established AF treated with anticoagulation, SCD accounts for over 20% of all deaths⁹⁹. In recent population based cohort and case-control studies, patients with AF have on average a 2.5 fold increased risk of SCD¹⁰⁰ or VF ^{77, 101} as compared to those without AF. The mechanism underlying this elevation in SCD risk is not completely understood; but it does not appear to be entirely dependent on coexisting CVD or explained by use of antiarrhythmic drugs¹⁰¹; however, in one population-based study, much of the excess SCD risk associated with AF could be accounted for by coexisting HF¹⁰².

Patients with severe chronic kidney disease (CKD) are also at higher risk for SCD, with annualized SCD rates approaching 5.5% in patients undergoing dialysis¹⁰³. There are also recent data to suggest that individuals with more moderate levels of CKD have a higher risk for SCD as compared to persons with normal kidney function^{68, 104}. Presumably, some of this could be due to electrolyte shifts and/or significant degrees of LVH observed in these patients. Recent data also suggest that obstructive sleep apnea¹⁰⁵ and seizure disorders¹⁰⁶ may be contributors to SCD risk in the population. Whether treatment for the above disorders will attenuate the elevated SCD risk is unknown and requires further exploration.

Triggers of SCD

Diurnal/Seasonal Variation

SCD tends to occur more frequently at certain times of the day, week, and year. SCD incidence peaks from 6 AM to noon^{107, 108}, and is highest on Monday and lowest over the weekend^{109, 110}. These morning and Monday peaks in SCD rates appear to be blunted by beta-blockers¹¹¹, suggesting that adrenergic triggers may underlie part of these circadian variations. There also appears to be seasonal variability in SCD incidence, with the highest and lowest rates observed in the winter and summer months, respectively, in both hemispheres^{110, 112}. These relationships observed in the general population may differ in patients with underlying heart diseases. In patients with ARVC and Brugada syndrome (BrS), ventricular arrhythmias tend to peak in the summer months.^{113, 114} These findings suggest that the onset of SCD may be associated with endogenous rhythms and external factors such as activity levels, psychological exposures, sunlight, temperature, and other climatic conditions^{112, 115}.

Physical Activity

Most studies^{65, 116–119}, but not all^{120, 121}, have found protective associations between regular physical activity and SCD or cardiac arrest, particularly for moderate levels of exertion^{65, 68, 117–119}. It is also well recognized that SCD occurs with a higher than average frequency during or shortly after vigorous exertion. The proportion of exertion related SCDs varies widely from 3–13 percent^{116, 119, 121–123} depending on the population surveyed. Case-control and case-crossover analyses have demonstrated that vigorous exertion can trigger cardiac arrest¹¹⁶ and SCD^{119, 121}, and this risk appears to be greater in men versus women^{119, 121}. Habitual exercise lowers this transient excess risk of SCD; however for men, the risk remains significantly elevated even among those who exercise most frequently^{119, 121}.

Despite these risks, the absolute risk of exertion related SCD is low. Recent population-based estimates regarding the frequency of exercise related OHCA and SCD range from 2.1 per 100,000 person years in the Netherlands¹²² to 0.46 per 100,000 person years in France¹²³. The majority of these exertion related SCD events took place in adults over age 35 (Figure 4), and the incidence was 15–20-fold higher in men as compared to women¹²². Even among athletes participating in the same sporting activity, rates of exertion related SCD remain significantly higher among men¹²⁴.

Figure 4. Age distribution of sports-related sudden deaths in France.

Deaths in the overall population (blue) versus young competitive athletes (red)

Among the 820 reported sports-related sudden deaths, only 50 cases (6%) occurred in young competitive athletes.

Adapted from Marijon et al¹²³ with permission.

Psychosocial Determinants

Depression, anxiety, and psychological stress have all been linked to SCD and OHCA risk in diverse populations. Anxiety, particularly phobic anxiety, has been directly associated with SCD, but not non-fatal MI risk, in men¹²⁵ and women¹²⁶. Depression^{127, 128} and other major psychiatric disorders, in particular schizophrenia¹²⁹, have been associated with higher rates of SCD as well. Potential pro-arrhythmic properties of antipsychotic or antidepressant medications^{130, 131} could underlie part of this apparent excess SCD risk observed in patients with psychiatric disorders. In addition to the chronic effects of psychosocial stress, acute mental stress may act as a trigger for SCD.

Acute increases in the incidence of SCD have been documented in populations suffering disasters such as earthquakes or wars^132–135. On the day of the Northridge earthquake, there was a sharp increase in the number of SCDs related to CHD, which was followed by an unusually low incidence of CHD deaths in the week after the earthquake. In contrast, in the recent Japan earthquake and tsunami, where multiple aftershocks occurred and the level of devastation was quite high in comparison, the incidence of SCD ¹³⁵ and OHCA¹³⁴ was increased for up to four weeks after the event, particularly among the elderly, and was significantly associated with level of seismic activity¹³⁵. These disasters demonstrate how severe emotional stress may precipitate cardiac events in vulnerable and/or predisposed populations.

Air Pollution

Several studies have examined the impact of short-term air pollution exposures [most often fine particulate matter (PM_2.5), carbon monoxide, or oxides of nitrogen], and risk of out-of-hospital cardiac arrests¹³⁶. In studies based in metropolitan areas of Europe^{137, 138}, United States^{139, 140} and Australia¹⁴¹, elevated risks of OHCA have been temporally associated with increased levels of PM. However, other studies from Washington State, United States¹⁴² and Copenhagen, Denmark,¹⁴³ did not find consistent associations. Long-term exposures to air pollution have been associated with increased mortality from CHD^{144, 145} and exposure to roadway pollutants may elevate SCD risk¹⁴⁶.

SCD in the Patient Populations with Structural Heart Disease

Coronary Heart Disease (CHD)

CHD underlies a significant proportion of SCD, especially in Western countries, and overt CHD is associated with marked increases in SCD risk¹⁴⁷. In the Framingham Study, pre-existing CHD was associated with 2.8–5.3 fold increases in SCD risk¹³, and women and men have a 4 to 10 fold higher risk of SCD respectively after experiencing an MI^{5, 38}. The absolute rate of SCD appears to be highest in the first 30 days after MI and decreases gradually with time^{148, 149}; although the proportion of patients who die from non-SCD is greater in the first 18 months¹⁵⁰. The incidence of SCD after MI has declined in parallel with CHD mortality over time¹⁴⁹, with rates as low as 1% per year in patients receiving optimal medical therapy and revascularization¹⁵⁰. However, rates remain high in certain subsets of post-MI patients.

There are three general settings where SCD occurs in patients with CHD: (1) during or after acute MI, (2) provoked by coronary ischemia without MI, and (3) in the presence of myocardial structural alterations (fibrosis, scar, left ventricular dilatation) secondary to prior MI or chronic ischemia. Only 19% and 38% of cardiac arrest survivors develop a new Q-wave MI and enzymatic evidence of MI, respectively⁶². The prevalence of acute coronary thrombus or active coronary lesion in autopsy series of SCD varies depending on autopsy protocol and histological techniques, ranging from 19–74%^151–153. With respect to the type of active lesion found at autopsy, approximately 2/3 of coronary thrombi are organizing, and late stage lesions or coronary erosions are more commonly encountered in women¹⁵⁴. In most series^{152, 155, 156}, stable plaques and/or chronic changes alone are found in ~50% of SCD victims with CHD on autopsy. From these data, it appears that plaque rupture with or without associated thrombosis and/or MI is present in some, but not all, CHD patients at the time of SCD.

The potential underlying mechanism precipitating SCD also differs depending upon the setting in which it occurs and the chronicity of disease. The two most common mechanisms are thought to be polymorphic VT/VF precipitated by acute ischemia and/or infarction and monomorphic VT degenerating to VF arising from a reentrant circuit within or surrounding a myocardial scar. In addition to these primary arrhythmic causes, a significant proportion of SCDs in the post-MI population appear to be due non-arrhythmic causes such as myocardial rupture and or extensive re-infarction, and this percentage appears to be highest within the first month after MI¹⁵⁷. In patients with end-stage ischemic cardiomyopathy, other modes of death such as acute pump failure and/or respiratory arrest resulting in PEA, or primary bradyarrhythmias comprise a significant proportion of SCD as well¹⁵⁸.

Risk Factors for SCD in Patients with Established CHD

Left ventricular systolic dysfunction and severity of HF symptoms are currently the strongest predictors of SCD risk among patients with prior MI and/or ischemic cardiomyopathy^158–160. After MI, mortality risk increases gradually until the left ventricular ejection fraction (LVEF) declines to 40%, and then exponentially increases as LVEF decreases further¹⁶¹. SCD rates reach 10% over a median follow-up of approximately 2 years among patients with LVEF<30% and CHF in clinical trials^{148, 162}. Based upon a randomized clinical trials performed in populations with low LVEFs and CHF^162–164, ICD therapy is recommended for patients with ischemic dilated cardiomyopathy, prior MI, New York Heart Association (NYHA) Class II and III HF, and LVEF ≤35%¹⁶⁵. In contrast, ICD therapy does not reduce mortality in the early post-MI period (within 40 days)^{166, 167}, possibly due to a predominance of non-arrhythmic causes of death during this time window¹⁵⁷.

Stratifying SCD risk based solely on LVEF and degree of systolic HF has two major well-recognized limitations. First, LVEF and NYHA class are both strongly associated with other modes of cardiovascular death^{148, 168}, and patients with the greatest functional impairment secondary to systolic HF and/or lowest LVEF are more likely to die from HF as opposed to SCD¹⁵⁸. The inability of these clinical markers to discriminate SCD risk from other competing causes of death has important clinical implications. In a recent prospective study series of 1,100 patients with systolic dysfunction¹⁶⁹, CHD patients who received primary prevention ICDs on the basis of LVEF and CHF were more likely to die than to experience an appropriate ICD therapy from their device. Second, the majority of patients who suffer a SCA or SCD do not appear to have LV systolic dysfunction and/or clinical HF preceding death^{7, 19}. In a prospective registry of cardiac arrests in the Netherlands, only 26% of SCAs with heart disease had HF prior to death, and only 19% of patients had an LVEF 30%¹⁴. In a more contemporary cohort in Multnomah County, Oregon, one-third of SCAs who had an echocardiogram prior to death had an LVEF<35%¹⁷⁰.

In addition to LVEF and CHF, other potential markers of increased SCD risk in patients with CHD include sustained VT induced at electrophysiology study (EP study), left ventricular scar size and heterogeneity on cardiac magnetic resonance (CMR), T-wave alternans, markers of autonomic function such as baroreflex sensitivity and impaired heart rate turbulence, and conventional ECG measures such as left bundle branch block (LBBB), QRS duration, LVH, and QT interval^{161, 171–173}. To date, only inducible sustained VT at EP study has been proven in a randomized clinical trial to identify individuals at a higher risk of SCA versus non-SCA¹⁷¹. However, the sensitivity of this test in isolation is inadequate to guide ICD therapy, especially in patients with LVEF less than 30%^{161, 171}. Recently, sustained VT at EP study was also found to be effective at stratifying arrhythmic death risk among patients with LVEF <35 in the early post-MI period¹⁷⁴.

Risk Factors for SCD in Patients with Preserved LVEF

Although the incidence of SCD is lower in patients with HF with preserved LVEF (HFpEF) as compared to those with reduced LVEF (HFrEF), the ratio of SCD to progressive HF deaths is higher, with SCD comprising 11 to 28% of all deaths^175–177. Relatively little is known regarding SCD risk prediction in CHD patients with preserved LVEFs. Prior history of MI, HF, and history of diabetes are consistent risk factors for SCD in this population^{178, 179}. Other potential clinical risk factors identified in these populations include male sex, AF, physical inactivity, LBBB on ECG, NT-proBNP levels and severity of coronary artery disease (CAD)^{68, 178, 179}.

Cardiomyopathies

Next to CHD, non-ischemic cardiomyopathies are the second most frequent cause of SCD in the United States and European countries, which account for approximately 10% to 15% (Figure 2)^{7, 10, 18}. Further, the prevalence of cardiomyopathies in young autopsied SCD victims aged ≤35 years is higher, and is reported to be 15% to 30%^{24, 25, 180, 181}. On the other hand, non-ischemic cardiomyopathies are more frequently observed as a cause of SCD in Japan (approximately 30% to 35% of SCD victims)¹⁸². The three major etiologically-distinct cardiomyopathies are non-ischemic dilated cardiomyopathy (NIDCM), HCM, and ARVC.

Non-ischemic Dilated Cardiomyopathy (NIDCM)

NIDCM has an estimated prevalence of 1:2500¹⁸³ and is defined by the presence of LV dilatation and LV systolic dysfunction in the absence of abnormal loading conditions (hypertension, valve disease) or CAD sufficient to cause global systolic impairment¹⁸⁴. Causes of NIDCM include gene mutations, myocarditis caused by viral, bacterial, fungal, or parasitic infections, toxicity due to alcohol, chemotherapeutic agents, metals, and autoimmune and systemic disorders. However, the majority of cases remain unexplained despite a thorough evaluation. Inherited NIDCM is reported to occur in up to 40% of cases, mostly in an autosomal dominant fashion¹⁸⁵. To date, mutations in more than 40 genes have been reported, in which TTN, MYH7, TNNT2, and LMNA are the most frequently identified, encoding titin, myosin heavy chain, cardiac troponin T (all in sarcomere), and lamin A/C (in nuclear envelope), respectively¹⁸⁵.

Prior episodes of sustained ventricular tachyarrhythmia, history of syncope, reduced LVEF, HF, and family history of SCD are the primary risk factors utilized to identify patients at a sufficiently high enough SCD risk to warrant ICD therapy¹⁸⁶. Two primary prevention randomized trials of ICD therapy^{187, 188}, included NIDCM patients with LVEF of ≤35% and HF symptoms (NHYA I – III) and demonstrated significant reductions in the SCD rate in patients with NIDCM (hazard ratio of 0.20¹⁸⁸ and 0.34¹⁸⁹) and reductions in total mortality when combined in meta-analysis¹⁸⁹. However, as in patients with ischemic cardiomyopathy, LVEF has a low sensitivity and specificity for predicting SCD and more specific markers are needed¹⁹⁰. Recently, midwall fibrosis detected by late gadolinium enhancement CMR was demonstrated to improve SCD risk prediction beyond LVEF in a large study of patients with NIDCM¹⁹¹.

Hypertrophic Cardiomyopathy (HCM)

HCM, defined by increased LV wall thickness not solely explained by abnormal loading conditions, is considered the most common inherited cardiac disease with an estimated prevalence of 1:500 in the general population¹⁹². In adult patients, the clinical diagnosis of HCM is made by cardiac imaging showing a left ventricular wall thickness of ≥15 mm in one or more segments. HCM can be present with lesser degrees of the wall thickening (13 to 14 mm), but other features of HCM, such as a family history, non-cardiac symptoms and signs, ECG abnormalities, and abnormalities on multi-modality cardiac imaging are required to support the diagnosis¹⁹³. To date, over 1500 mutations in more than 11 genes encoding components of the sarcomere or adjacent Z-disc have been identified, with the most common encoding beta myosin heavy chain and myosin binding protein C^{192, 193}.

The annual incidence of cardiovascular death in HCM is approximately 0.5 – 2% in contemporary series, and SCD from a lethal ventricular arrhythmia remains one of the common modes of death^{192, 193}. SCD is more likely to occur in young patients (<30 years) and is uncommon in older patients (>60 years)¹⁹². Established risk factors for SCD in patients with HCM include a history of unexplained syncope, family history of SCD, a maximal left ventricular wall thickness of ≥30 mm, repetitive non-sustained VT, and abnormal blood pressure response to exercise¹⁹². According to the ACCF and AHA guidelines, the presence of one or more of these risk factors can be used to select patients for primary prevention ICD placement¹⁹⁴. The most recent ESC guidelines¹⁹³ recommend the use of a prediction model which incorporates absolute risk and individual effect sizes of the above and other SCD risk factors (Figure 5)¹⁹⁵ at 1 – 2 year intervals. Implantation of an ICD is recommended in patients with an estimated 5-year SCD risk of ≥6% and a life expectancy of >1 year (Class IIa).

Figure 5. Sudden cardiac death risk prediction model for patients with hypertrophic cardiomyopathy.

A web-based risk calculator is provided on the website of European Society of Cardiology (http://www.doc2do.com/hcm/webHCM.html).

LVOT indicates left ventricular outflow tract; NSVT, non-sustained ventricular tachycardia; SCD, sudden cardiac death. Adapted from O’Mahony et al.¹⁹⁵ and Elliott et al¹⁹³ with permission.

Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC)

ARVC is a genetically determined heart muscle disorder characterized by fibrofatty replacement of the right ventricular myocardium¹⁹⁶. As the disease progresses, the left ventricle may also become involved. The estimated prevalence of ARVC is 1: 2,000 – 5,000¹⁹⁷, and up to 60% of the patients have a mutation with an autosomal dominant trait and incomplete penetrance¹⁹⁸. Among nearly 15 genes reported to cause ARVC, mutations in genes encoding components of cardiac desmosomes (plakophilin 2, desmoglein 2, and desmoplakin) are most frequently identified.¹⁹⁸ The 2010 Task Force diagnostic criteria for ARVC¹⁹⁶ consist of major and minor findings in six different categories: 1) structural alterations, 2) tissue characterization, 3) repolarization abnormalities, 4) depolarization abnormalities, 5) arrhythmias, and 6) family history including genetic testing.

SCD is a common cause of death in patients with ARVC especially in those in the fourth decade of life or younger, and it may be the first arrhythmic event in up to 50% of cases^{199, 200}. Several observational studies^201–204 showed that the annual rate of death or VF in patients treated by ICDs was 1.5% to 4%, and predictors of appropriate ICD therapy include a history of a cardiac arrest or VT with hemodynamic compromise, younger age, LV involvement, unexplained syncope, presence of non-sustained VT and inducibility during EP study^{201, 204}. Current ACCF/AHA/HRS guidelines¹⁸⁶ recommend the prophylactic use of an ICD in those who have one or more risk factors for SCD (Class IIa).

Valvular Heart Disease

Valvular heart disease is reported to be the cause of death in 1% to 5% of the SCD victims (Figure 2)^{10, 18, 205}. Even after surgical procedures, SCD occurs in 15% to 30% of patients, accounting for 0.2% to 0.9%/year, and is most commonly triggered by ventricular arrhythmias ²⁰⁶. Patients with aortic stenosis are at the highest risk of SCD after valve replacement, particularly within two years²⁰⁶. Prior to valve replacement, asymptomatic patients with severe AS have annual SCD rates of 1–3 %^{207, 208}, and recent observational data suggest that this risk may be lowered by early surgery.²⁰⁹ The role of mitral valve prolapse (MVP) in SCD is controversial. The majority of MVP is thought to be benign, but there are certain characteristics such as leaflet thickness, redundancy, and increased LV diameter that appear to be associated with higher risk ²¹⁰, and recent data suggests that women with bileaflet prolapse and complex ventricular ectopy may be at particular risk²¹¹. Overall, data regarding SCD risk stratification and appropriate utilization of ICDs in patients with valvular disease are scarce and further studies in this at-risk subgroup of patients are needed.

SCD in the Absence of Structural Heart Disease

Autopsy-negative SCD/ Sudden Unexplained Death

Autopsy-negative sudden death is more commonly reported in younger individuals. Autopsy series from Ireland²⁴ and Sydney¹⁸¹ reported that 27– 29% of sudden arrhythmic deaths in individuals less than age 35 had no demonstrable structural heart disease on autopsy. In a Danish nationwide study of SCD²³, this proportion was even higher (43%). However, when detailed histologic examinations are performed, the percentage of autopsy-negative SCD is much lower. In one prospective study of 273 consecutive SCD cases aged 1 to 35 years in Italy¹⁸⁰, detailed histologic examination identified concealed pathologic substrates, such as focal myocarditis, regional ARVC, and conduction system abnormalities in 60 out of 76 cases without macroscopic evidence for structural heart disease. After histologic exam, only 16 (6%) had no detectable abnormalities. Although discordances in the frequency of the autopsy-negative SCDs could be due to differences in regional genetic background, it is likely that the frequency of “autopsy-negative” cases would decrease if more detailed histologic examinations were carried out in all SCD victims.

Among patients with autopsy negative SCD, approximately 50% will have inherited arrhythmic syndrome (IAS)^{212, 213}, such as, long QT syndrome (LQTS), BrS, catecholaminergic polymorphic VT (CPVT), and early repolarization syndrome (ERS). Even when structural abnormalities of uncertain significance are found, IAS appear to underlie a significant fraction of SCDs²¹⁴. Taken together, these finding suggest that substantial numbers of SCD may be attributable to IAS in the young. Performing molecular autopsies in cases with autopsy-negative SCD and/or cascade screening of families is important to establish the cause of death and to identify relatives potentially at high SCD risk. In cases of sudden unexplained death, where a diagnosis is not made either by antemortem or postmortem analysis, genetic testing of family members reveals a possible disease causing mutation in 31% of families, and IAS comprise 30% of these mutations ²¹⁵.

Inherited Arrhythmic Disorders and Their Epidemiology

Long QT Syndrome (LQTS)

Congenital LQTS is a hereditary disorder, characterized by delayed myocardial repolarization resulting in prolongation of the QT interval on 12-lead ECG and predisposition to torsade de pointes (TdP) which can result in SCD^{216, 217}. Approximately 75% of patients with LQTS and 95% of genotype-positive LQTS will have a mutation in genes encoding the slow component (KCNQ1, LQT1) and the rapid component (KCNH2, LQT3) of the delayed rectifier potassium current and the cardiac sodium channel (SCN5A, LQT3) ^{218, 219}. Conversely, it is estimated that 25% to 35% of genetically-affected patients have a normal or borderline QTc at rest^{219, 220}, requiring exercise or a catecholamine infusion to disclose the masked QT interval.^{217, 220, 221} In order to directly estimate the prevalence of LQTS, Schwartz et al ²²² carried out 12-lead ECGs in 43,080 white infants. Prolonged QTc intervals of 451 to 460 ms, 461 to 470 ms, and >470 ms were observed in 177 (0.41%), 28 (0.06%), and 31 (0.07%) infants, respectively. Of these, 17 out of 43,080 infants were found to be affected by LQTS on the basis of genetic testing and further clinical evaluation, indicating a prevalence among whites of 1: 2,534. Extrapolating these results to the non-genotyped infants, the authors estimated the prevalence of LQTS was closer to 1: 2,000²²².

The estimated incidence of cardiac arrest or SCD before the age of 40 in untreated patients is estimated to be 0.30%/year, 0.60%/year, and 0.56%/year in LQT1, LQT2, and LQT3 respectively.²¹⁹ Most arrhythmic events developed during exercise or emotional stress in LQT1, at rest or with sudden noises in LQT2, and at rest or during sleep in LQT3²²³. Other risk factors for arrhythmic events in LQTS include prior history of syncope, significant QTc prolongation^{219, 224} and location and number of mutations^{219, 224–227}. Beta-blockers remain the mainstay of therapy for the majority of these patients, and ICDs are generally reserved for patients who have suffered a cardiac arrest²²⁸.

Brugada Syndrome (BrS)

BrS was first described in 1992²²⁹ and is thought to underlie, to a certain extent, the mystery of unexpected nocturnal death, which is colloquially called “Pokkuri” in Japan, “Lai Tai” in Thailand, and “Bangungut” in the Philippines²³⁰. BrS is a primary electrical disorder affecting middle-aged males with their first arrhythmic event typically developing during sleep at a mean age of 40 years²³¹. The clinical phenotype is 8 to 10 times more prevalent in men than in women, which is attributable, at least in part, to the higher testosterone level in men²³². Twelve-lead ECGs at rest are characterized by a coved type ST-segment and J point elevation of ≥ 2 mm (0.2 mV) followed by a negative T wave in the right precordial leads (V1–3), which is referred to as a “type 1 Brugada ECG”²³¹.

In the 2013 HRS/EHRA/APHRS expert consensus statement²²⁸, BrS is diagnosed when a type 1 ST-segment elevation is observed either spontaneously or after the administration of a sodium channel blocking agent in at least one right precordial lead (V1 and V2), which is placed in a standard or a superior position, in which case, documentation of VT/VF, clinical symptoms, or a family history is no longer necessary. Table 1 displays the reported prevalence of a type 1 Brugada ECG across population-based studies. The prevalence of the type 1 ECG pattern in adults is greatest in Japan^{233, 234}, the Philippines²³⁵ and among Japanese-Americans in North America²³⁶ (0.15 to 0.27% [1: 350–700]). Rates in Europe^237–239 (0 to 0.017% [less than 1: 5,000]) and North America^{240, 241} (0.005 to 0.1% [1: 1000–20,000]) appear to be lower. These estimates of BrS do not account for temporal variability of the ECG morphology²³³ or patients who exhibit type 1 ECG only in the superior lead positions or after drug-provocation²²⁸.

Table 1.

The Prevalence of a Type 1 Brugada ECG^* in the Population Studies

Coutry	Authors	Year Published	Individuals Screened, n	Male Sex	Mean Age or Range of Age, y	Type 1 ECG, n (%)
Europe
Finland	Junttila et al	2004	2479	100%	18–30	0
Greece	Letsas et al237	2007	11488	58%	15–98	2 (0.017)
Italy	Gallagher et al238	2008	12012	91%	30 ± 9	2 (0.017)
Germany	Sinner et al	2009	4149	49%	51 ± 14	0
Denmark	Pecini et al239	2010	18974	45%†	52 ± 12*	0
North America
Canada	Lee et al241	2005	3983	100%	31	4 (0.100)
USA (Japanese-American)	Ito et al236	2006	8006	100%	45–68	12 (0.150)
USA	Patel et al240	2009	162590	65%	not described	8 (0.005)
Asia
Japan	Sakabe et al233	2003	3339	79%	> 18	5 (0.150)‡
Japan	Yamakawa et al	2004	20387	51%	10	1 (0.005)
Japan	Oe et al	2005	21944	51%	7	1 (0.005)
Japan	Tsuji et al234	2008	13904	27%	58 ± 10	37 (0.266)
Philippines	Gervacio-Domingo et al235	2008	3907	not described	≥20	7 (0.179)
Taiwan	Juang et al	2011	20562	39%	49 ± 21	1 (0.005)
Korea	Uhm et al	2011	10867	100%	21 ± 5	0

ECG indicates Electrocardiogram; and USA, United States of America.

^*Studies including patients with a coved type ECG and J point amplitude ≥ 0.1 mV were excluded.

^†in the first examination

^‡those with a continuous type 1 ECG

The primary risk factors for SCD in type 1 BrS are prior history of syncope or aborted SCD. In recently-published multicenter registry studies^242–244, the incidence of the cardiac events (SCD, VF, and/or appropriate ICD shocks) in type I BrS ranged from 7.7–10.2%/year, 0.6–3.0%/year, and 0.5–0.8%/year in those with a history of aborted SCD due to VF, syncopal episodes, and no symptoms, respectively.

Catecholaminergic Polymorphic VT (CPVT)

CPVT is a familial arrhythmogenic disorder characterized by polymorphic ventricular tachyarrhythmias or bidirectional VT induced by physical or emotional stress²⁴⁵. The patients show no detectable cardiac morphological abnormalities, and the ECG is normal except for a lower heart rate at rest^{245, 246}. The affected patients usually develop arrhythmic events (syncope, aborted cardiac arrest, or SCD) during adrenergic activity in the first or second decade of life,^{245, 247–249} and the clinical course is considered to be highly malignant. Without proper treatment such as beta blockers, flecainide, and ICDs ^245–251, mortality reaches >30% by the age of 30 years²⁵⁰ and the estimated 8-year fatal or aborted SCA event rate after the diagnosis is 13%²⁴⁷. The population prevalence of CPVT is difficult to estimate since it cannot be detected on resting 12-lead ECG, but is projected to be approximately 1: 10,000²²⁸.

Early Repolarization Syndrome (ERS)

An early repolarization ECG pattern (ERP), which consists of a J wave elevation ≥0.1 mV, either notched or slurred, accompanied by an ST segment elevation, has long been considered to be a benign finding and unrelated to serious cardiac events^{252, 253}. This notion has recently been challenged by studies demonstrating that an ERP in the inferior and/or lateral leads is more commonly found in patients with idiopathic VF as compared to controls^{254, 255}, raising the possibility that the ERP may be a marker of an arrhythmogenic substrate. Considering these data, a recent expert consensus panel defined ERP as the presence of J-point elevation ≥0.1 mV in ≥2 contiguous inferior and/or lateral leads, and ERS is diagnosed in the presence of ERP in a patient resuscitated from otherwise unexplained VF/polymorphic VT or autopsy negative SCD victim with a previous ECG demonstrating ERP²²⁸.

The question of whether an ERP on resting ECG confers an increased risk of SCD in the general population has been examined in several population-based studies^{253, 256–261}, which are summarized in Table 2. The definition of ERP varies widely between these studies. In some studies, ERP had to be present in the inferior and lateral leads^{256, 257, 260}, but others considered J-point ST elevations in all body surface leads to be ERP^{253, 258, 259, 261}. Prevalence estimates in these studies range from 1% to 24% and 0.6% to 6.4% for J point elevation of ≥0.1 mV and 0.2 mV, respectively. Notwithstanding these differences in methodology, ERP is reported to be more prevalent in younger age groups, men, or individuals of African descent^{258, 259, 262}. The majority of European studies^{256, 257} and Japanese studies^{260, 261} found significant associations between the ERP and cardiac or sudden arrhythmic death, while studies conducted in the United States generally did not^{253,258, 259}. One U.S. study suggested that the association between ERP and SCD may be limited to women and/or white individuals²⁵⁸. These results suggest that there may be an ethnic, racial, and/or sex differences in the relationship between the ERP and SCD.

Table 2.

The Prevalence of Early Repolarization ECGs and Their Prognosis in the Population Studies

Coutry	Authors	Year Published	Position of ERP	Individuals Screened, n	Male Sex	Mean Age at Baseline, y	J point elevation		Mean Follow-up Period, y	RR of Death According to the ERP
							≥ 0.1 mV , n (%)	≥ 0.2 mV , n (%)		Cardiac	Sudden or Arrhythymic
Europe
Finland	Tikkanen et al256	2009	Inf or Lat	10864	52%	44 ± 8	630 (5.8)	67 (0.6)	30 ± 11	1.28* in Inf 1.34* in Lat	1.43* in Inf 0.75 in Lat
Germany	Sinner et al257	2010	Inf or Lat	6213	49%	52 ± 10	812 (13.1)	not described	19	3.44*	not described
France	Rollin et al	2012	Inf or Lat	1161	52%	50 ± 9	159 (13.7)	74 (6.4)	14 ± 2	5.28* in Inf 6.27* in Lat†	not described
North America
USA	Klatsky et al253	2003	All	73088	44%	37 ± 13	670 (0.9)	494 (0.7)	14	0.8†	not described
USA	Uberoi et al	2011	Inf or Lat	29281	87%	55 ± 15	664 (2.3)	0	8 ± 4	1.73 in Inf 0.83 in Lat†	not described
USA	Olson258	2011	All	15141	44%	54 ± 6	1866 (12.3)	not described	17 ± 4	not described	1.23 in all 2.03* in whites
USA	Ilkhanoff et al259	2014	All	5039	46%	25	1249 (20.9)‡	not described	23	0.96†	not described
Asia
Japan	Haruta et al260	2011	Inf or Lat	5976	44%	not described	1429 (23.9)	not described	24 ± 15	0.75*	1.83*
Japan	Hisamatsu et al261	2013	All§	7630	41%	52	264 (3.5)	not described	15	2.54*	not described

Ant indicates anterior leads; ECG, electrocardiogram; ERP, early repolarization pattern; Inf, inferior leads; Lat, lateral leads; RR, relative risk; and USA, United States of America.

^*Statistically significant

^†For the cardiovascular death

^‡At baseline

^§≥ 0.2 mV in anterior leads

It is important to note that the calculated relative risks for sudden arrhythmic death associated with a J-point elevation of 0.1 mV are quite modest (Table 2) and the absolute risk of arrhythmic death in asymptomatic individuals with the ERP on ECG is extremely low²⁵⁵. Three fold elevations in sudden/arrhythmic death have been observed when ERP is more strictly defined as a J point elevation of >0.2 mV associated with horizontal/descending ST segment limited to the inferior leads^{256, 263}. However, this pattern was only noted in 0.3% of the population²⁶³.

Conclusions

SCD is a major public health problem all over the world, and although resuscitation rates are improving, the majority of individuals who suffer SCA will not survive, and often the underlying cardiac condition is not recognized prior to death. Behind these tragic events, there are various causes, risks, and predisposing conditions, which differ in the prevalence according to region, age, ethnicity, race and sex. As such, a multifaceted approach, which addresses risk factors both in high and low risk populations, will be required to decrease the burden of SCD. Population wide approaches as well as improved identification of high risk individuals who will benefit from ICDs will be crucial to prevent SCD events and improve patient outcomes. Although substantial progress has been made in this field, further studies addressing SCD prevention across the whole spectrum of disorders, from CHD in the general population to the rarer inherited disorders, are warranted to address many remaining uncertainties regarding the multitude of factors which underlie susceptibility to SCD.

Nonstandard Abbreviations and Acronyms

ARVC

arrhythmogenic right ventricular cardiomyopathy

AF

atrial fibrillation

BrS

Brugada syndrome

CMR

cardiac magnetic resonance

CVD

cardiovascular disease

CPVT

catecholaminergic polymorphic ventricular tachycardia

CAD

coronary artery disease

CHD

coronary heart disease

ERP

early repolarization ECG pattern

ERS

early repolarization syndrome

EP

electrophysiology

EMS

emergency medical service

ER

emergency room

HF

heart failure

HCM

hypertrophic cardiomyopathy

ICD

implantable cardioverter defibrillator

IAS

inherited arrhythmic syndrome

LBBB

left bundle branch block

LVEF

left ventricular ejection fraction

LQTS

long QT syndrome

MVP

mitral valve prolapse

MI

myocardial infarction

NIDCM

non-ischemic dilated cardiomyopathy

NYHA

New York Heart Association

OHCA

out of hospital cardiac arrests

PEA

pulseless electrical activity

SCA

sudden cardiac arrest

SCD

sudden cardiac death

TdP

torsade de pointes

VF

ventricular fibrillation

VT

ventricular tachycardia