Early repolarisation: controversies and clinical implications
- Correspondence to Dr Elijah R Behr, St George's University of London, Division of Cardiovascular Sciences, Cranmer Terrace, London SW17 0RE, UK;
Contributors Both authors contributed equally to this manuscript.
- Received 4 November 2011
- Revised 24 January 2012
- Accepted 21 February 2012
- Published Online First 22 March 2012
Early repolarisation was previously considered a benign variant but in recent years has emerged as a marker of risk for sudden death. In part, this appears to reflect a change in the definition. ECG territory, degree of J-point elevation and ST-segment morphology are associated with different degrees of risk for subsequent ventricular arrhythmia. At present the dataset is insufficient to allow risk stratification in asymptomatic individuals and further epidemiological and mechanistic research is required.
- Early repolarisation
- early repolarisation syndrome
- sudden cardiac death
- idiopathic ventricular fibrillation
- arrhythmic right ventricular dysplasia
- sudden adult death syndrome
- molecular genetics
Early repolarisation (ER) was historically considered a benign ECG variant.1 Recently, it has emerged as a marker of risk for sudden cardiac death (SCD), particularly when seen in the inferior and inferolateral ECG leads. This risk has been associated with idiopathic ventricular fibrillation (IVF), termed early repolarisation syndrome (ERS) and has also been identified in the general population.2–4 Although ER is a common ECG finding, the syndrome remains rare and determining risk in the individual is a major challenge. This review will describe the problems surrounding ER and place this challenge into context.
Definitions of ER: a historical perspective
Slurring or notching of the terminal QRS (J-point elevation) had been described by Shipley and Hallaran in 1936, but ER was first defined by Wasserburger in 1961 as a benign finding more prevalent in men, athletes and individuals of black ethnicity.1 It was found predominantly in the anterolateral ECG leads and consisted of elevated take-off of the ST-segment at the end of the QRS (the J junction) relative to the succeeding TP interval with downward concavity of the ST-segment and symmetrical T-waves (figure 1).1 5 ER was not associated with cardiovascular disease and subsequent investigation indicated that subjects were more likely to be young, male and physically active with lower body mass index (BMI), heart rate, blood pressure and blood cholesterol.6 J-point elevation was noted to vary between different ECG recordings and generally decrease over time.6
The definition has evolved, however, and it was redefined by Haissaguerre et al in 2008 to describe an association with IVF, the new definition being used almost exclusively since.2 It was also found to be associated with increased risk of SCD in the general population.4 This new definition required J-point elevation of at least 0.1 mV in two or more consecutive ECG leads with a slurred or notched appearance, but concomitant ST-segment elevation or T-wave symmetry was not necessary (figure 2). Leads V1–V3 were excluded to avoid confusion with the Brugada syndrome (BrS) and arrhythmogenic right ventricular cardiomyopathy.
Following work by Tikkanen et al in 2011, the ST-segment and T-wave morphology were again identified as important.7 Two ST-segment types were described in addition to the Haissaguerre definition of ER: rapidly ascending with >0.1 mV ST-segment elevation within 100 ms after the J point or persisting through the ST-segment; and horizontal/descending with ≤0.1 mV ST-segment elevation within 100 ms after the J point (figures 1 and 2, respectively). This represents an evolution of the definition into two different forms: one compatible with the original description and seen predominantly in athletes (ER with rapidly ascending ST-segment); and another that historically would not have been considered premature repolarisation (ER with horizontal/descending ST-segment).
Difficulties in measuring ER
Accurate measurement of the notched ER pattern is usually straightforward. When the transition from the terminal QRS into the ST-segment is unclear J-point elevation may be more subjective. This is a problem for slurred ER. The discrepancy between repeated QRS duration measurements has been shown to approach 40 ms and the QRS complex may begin and end at different times in different ECG leads with up to 20 ms difference.8 What appears to be ER in one territory may therefore look like QRS fragmentation and conduction delay in another (figure 3). This methodological issue has not been dealt with as yet and subjective measurement of slurring exposes individual studies to possible observer bias despite laudable attempts at internal validation. This is particularly relevant when making comparisons between studies. It also implies a potentially mechanistic role for delayed cardiac conduction in some individuals displaying an ER pattern.9
ER in the general population
There are conflicting reports from large observational cohort studies about the prevalence of ER, its characteristics, epidemiological features and associated risk. These probably reflect methodological differences such as the different definitions used and difficulties in measuring the slurred pattern, as well as cohort differences such as variation in gender, age, ethnic composition, end points and length of follow-up (table 1). Based on the original definition of ER, its prevalence in the white Caucasian general population was thought to be 1–2%. Using the new definition, prevalence has ranged from 4.5% to 24% across different ethnic groups.1 4 6 10–14
In middle-aged white Europeans (the Finnish Heart Study and the German MONICA/KORA cohorts), the prevalence of ER varied between 6% and 14%.4 10 ER was predominantly slurred, seen most commonly in the inferior territory and tended to be stable over time. There was an association with smoking, diabetes, personal and/or family history of coronary artery disease (CAD) and ECG signs of CAD.4 10 12 This was particularly true for ER in the inferior leads.4 When ER was subgrouped by ST-segment morphology, however, a rapidly upsloping ST-segment was still associated with younger age, lower BMI, lower heart rate and blood pressure, ECG signs of left ventricular hypertrophy and shorter QTc interval. A horizontal/descending ST-segment was associated with older age, ECG signs of CAD and longer QRS duration.7
The presence of inferior ER in white Europeans conferred a two- to fourfold increased risk of cardiac death, greatest in middle-aged men when J-point elevation was ≥0.2 mV and tending to occur at least a decade after ER was detected.4 10 When inferior ER with J-point elevation ≥0.2 mV was stratified according to ST-segment morphology, a horizontal/descending ST-segment was independently associated with a threefold increased risk of SCD while subjects with lateral ER and/or rapidly upsloping ST-segment morphology were not found to be at increased risk.7 Increased all-cause mortality was also associated with ER in a similar fashion to sudden and cardiac death across both cohorts, although the strength of association was weaker and therefore apparently driven by the cardiac aetiology (table 1).4 10
In a Japanese population of Hiroshima survivors studied over 46 years, the incidence of ER on the ECG was 715/100 000 person-years with a peak incidence in the second decade of life.11 Lateral notched ER was most prevalent and this appeared to fluctuate and generally reduce over time.11 Japanese subjects with inferolateral ER also demonstrated an increased risk of unexpected death but overall ER was associated with a decreased risk of cardiac death.11
Among black Americans studied as part of the Atherosclerosis Risk in the Community study the reported prevalence of ER was as high as 24%.12 Consistent with earlier findings, ER was more common in black men with lower BMI and heart rate, suggesting a healthier subgroup, although there was also an association with higher blood pressure.12 A second US study, the largest to date, considered the prevalence of ER with and without ST-segment elevation in a population of mixed ethnicity (the VA Palo Alto Health Care System cohort; white 81%, black 13% and Hispanic 6%). R wave slurring or J waves (compatible with the new definition of ER) were demonstrated in 14% of subjects, but ER with ST-segment elevation (compatible with the original definition of ER) was demonstrated in only 2.3%.13 In the US cohorts ER also appeared more common laterally, but when ECG leads V1–V3 (which have usually been omitted from analysis in other studies) were included prevalence was actually greatest in the anterior territory.12 13
In the Atherosclerosis Risk in the Community study there was no increase in the risk of ER-associated SCD seen in black subjects or men but, unexpectedly, an eightfold increase in risk in white women only. All-cause mortality was, however, unaffected by the presence of ER for any ethnicity or gender.12 Despite its size the VA cohort demonstrated no increase in cardiovascular mortality with ER, irrespective of territory and morphology. The follow-up for this cohort was shorter than for any of the other studies and may partly explain the negative result, particularly for inferior slurring.13 Also, the most prevalent form of ER measured in the US cohorts and among black subjects was anterolateral and associated with ST-segment elevation. This probably represents a more benign form akin to the original definition of ER and may therefore be ‘drowning’ out the effects of inferior slurred ER with horizontal/descending ST-segment morphology. This latter pattern appears to be more relevant in a white middle-aged European population. However the associated risk becomes apparent only after at least 10 years' follow-up and may therefore not yet have been appreciated in the VA cohort.4 10 13
ER in athletes
Athletes appear to have an intermediate prevalence of ER, higher than that of the general population but lower than in IVF survivors (see below), with recent estimates ranging between 22% and 43%.3 7 15 ER is more common in male, black, taller athletes with lower heart rates and greater amounts of previous training.15 Lateral ER with rapidly upsloping ST-segment morphology is most prevalent, although inferior ER with horizontal/descending ST-segments is also seen.7 15 Both types increase with exercise training.15 The majority of evidence suggests that ER in athletes is benign, although one small study has shown increased prevalence of inferior and inferolateral slurred J-point elevation without ST-segment elevation in athletic survivors of IVF.16
Risk associated with ER in the asymptomatic individual
The overall implications of the ER pattern in the asymptomatic general population and athletes are difficult to assess from such mixed data but we may be able to draw some conclusions. The original definition of ER with anterolateral J-point elevation, rapidly ascending ST-segment elevation and symmetrical T-wave morphology is more common in the fit and healthy population with greater prevalence in black men. It does not appear to have significant associated risk for sudden death or all-cause mortality in comparison with a predominantly inferior pattern of J-point elevation accompanied by a horizontal/descending ST-segment. This appears to be relevant mainly to white Caucasians, and further dissection of the phenotype in different ethnic groups will be necessary to tease out its importance in non-Caucasians. These data suggest that the mechanisms underlying these two distinct forms of ER may be different.
ER and IVF (ERS)
J-point elevation is seen more often among survivors of IVF than among healthy controls, and various studies have estimated the prevalence of ER as 23% to 68%.2 3 17–21 This has been a consistent message and has led to the term ERS. Similar to the general population cohorts, there are differences between studies, due to small numbers of cases, differences in definitions of ER and ethnic variation (table 2). In addition, protocols for diagnosis and exclusion of underlying cardiac disease differ between institutions and may have influenced reported prevalence. For example, retrospective studies have usually evaluated IVF survivors using ECG, stress testing, coronary angiography and echocardiography and not all appear to routinely include cardiac MRI and/or sodium channel blocker provocation testing.3 20 One retrospective study that employed a complete investigation algorithm reported a 32% prevalence of ER.19 In a prospective registry of cardiac arrest survivors with preserved cardiac function, the prevalence of ER was 23% in those who received no primary diagnosis after thorough cardiovascular investigation. This compared with a 6% prevalence in those who received subsequent diagnoses, including ion channel disease, cardiomyopathies, myocarditis and coronary spasm.18
Patients with ERS are more likely to be young and male with history of cardiac arrest during sleep.2 19 ER is most commonly seen in the inferior leads with J-point elevation of ≥0.2 mV and horizontal/descending ST-segment morphology.2 3 23 Lateral ER in leads V4–V6 and ER with ST-segment elevation appear to be equally prevalent in patients and healthy controls, lending further credence to this being a more benign finding.3 There are conflicting reports about whether the ER pattern remains static, fluctuates or diminishes over time, but there is general consensus that J-point height increases before an arrhythmic episode.2 17 18 Subsequent implantable cardioverter defibrillator (ICD) recorded ventricular fibrillation (VF) events were initially reported as higher in IVF survivors with ER than in those without ER, but again this has not been consistent in more recent studies.2 20
In Japanese IVF survivors, ER was associated with slower heart rate, longer PR interval and QRS duration and shorter QTc interval than in healthy controls.22 In European and US studies QTc intervals were shorter in one series but the findings in others have been inconsistent or non-significant owing to small numbers.2 3 Only one study has compared ECG characteristics in patients with ERS with those of asymptomatic healthy individuals carrying ER.20 Although there was a trend towards longer PR interval and QRS duration in the ERS group, these results were not significant, but again numbers were small. Although left precordial notched ER was more common in IVF survivors in this group, the finding has not been reproduced by larger cohorts.18 20
Heritability of the ER pattern
The potential for genetic heritability of the ER pattern has been examined in community-based cohorts and a large family-based cohort.14 24 ER, particularly the inferior notched type, appears highly heritable with a reported 2.5-fold increased risk of offspring developing ER when it is present on the parent's ECG and a heritability estimate of 0.49 (h2 0.49, SE 0.14).24 Inferior ER is also more prevalent in the surviving blood relatives of patients who have had an unexplained sudden death (sudden arrhythmic death syndrome (SADS)) than age- and sex-matched controls. This is analogous to a familial form of ERS where the index patient has not survived their cardiac arrest. Pedigree analysis has suggested that an autosomal dominant pattern of inheritance with incomplete penetrance is most likely.25
It is well documented that in patients with ERS, episodes of VF are preceded by transient increases in J-point elevation and ventricular premature beats with short coupling intervals but the underlying mechanism is less clear.2 21 In a perfused canine left ventricular wedge preparation the ER pattern has been shown to result from a transmural voltage gradient created by greater density of Ito in the ventricular epicardium relative to the endocardium. Dispersion of repolarisation facilitates development of ventricular arrhythmia by phase II re-entry.26 Experimentally this can be modulated with attenuation of J-point elevation during isoproterenol infusion and accentuation of J-point elevation during bradycardia, hypothermia, hypercalcaemia and β-blocker administration.26 These experimental findings have not been reproduced in humans as yet and there are limitations to this model that reduce its applicability.
Genetic studies have focused upon candidate genes that might influence phase II of the action potential. A gain of function mutation in KCNJ8, encoding the KATP channel, has been identified in one patient with ERS.27 In one series, mutations in CACNA1, CACNB2 and CACNA2D1, presumed to cause loss of function affecting the L-type calcium current, were detected in 16% of probands with ERS as well as a small minority of probands with BrS.28 Recent data have also identified loss of function SCN5A mutations in 3/50 Japanese patients (6%) with ERS. These would usually be expected to be associated with a BrS phenotype yet sodium channel blockade resulted in augmentation of the ER pattern and/or development of VF without induction of the BrS phenotype.22 28
There are other similarities between ERS and BrS, including male predominance, peak incidence of SCD in the fourth decade and response to quinidine and isoproterenol.2 29 30 This has led to the suggestion that ERS and BrS form part of the same spectrum of repolarisation disorder known as the ‘J-wave syndromes’.31 This descriptor has been criticised for confusing the debate on ERS. J waves historically described characteristic ECG changes during hypothermia, but in ERS the terminology is increasingly used to depict J-point elevation alone.32 In addition, important differences exist between ERS and BrS. J point and ST-segment elevation are usually attenuated or even abolished by ajmaline in ERS17 22 33 and whereas BrS can be exacerbated by fever, ERS has been reported to be exacerbated by hypothermia.34
Evidence in favour of the ER pattern being a disorder of repolarisation includes the shorter associated QTc interval and attenuation of J-point elevation during increases in heart rate. In addition, patients respond favourably to quinidine (an Ito blocker) during a VF storm.29 However, it is possible that not all ER is the same. Although initial reports in ERS described low prevalence of signal-averaged ECG late potentials, a recent study demonstrated abnormal late potentials in 86% of patients with ERS compared with 27% of IVF survivors without ER.2 19 Abnormal late potentials are usually associated with depolarisation abnormalities, and in ERS showed diurnal variation with nocturnal ascendency, reflecting the peak at-risk period for SCD events.19
When ER is subgrouped by ST-segment morphology in population studies the benign form with a rapidly upsloping ST-segment is associated with a shorter QTc interval. The more malignant form with horizontal/downsloping ST-segment is associated with longer QRS duration, which may therefore reflect a disorder of depolarisation due to underlying subtle conduction and/or structural disease.7 This form is also associated with older age and increased ECG signs of CAD. Cardiac imaging studies were not performed in these studies to rule out structural abnormalities and the inferior ER pattern was present in serial ECGs, suggesting a more permanent electrophysiological substrate.7 It therefore appears possible that different underlying mechanisms may result in the ER pattern seen on the ECG: disordered repolarisation, depolarisation or a mixed disease. This is similar to the continuing debate in BrS.35
Unfortunately, the results from population studies may not be that comparable with the results from case–control studies of patients with IVF. In the former, SCD was most commonly due to myocardial infarction and the average age of death was 50–60 years. In the latter group, coronary angiography was normal and SCD occurred on average 10–20 years earlier. It therefore appears likely that these represent two different conditions. However, one unifying theory suggests that ER may reflect increased dispersion of repolarisation that places the individual at increased risk of arrhythmic death only in the presence of additional triggers such as myocardial ischaemia or Purkinje extrasystoles in the rare cases of IVF.36 Indeed ER alone may not be directly causative for ventricular arrhythmia and instead acts as a modifier, which may explain the not insignificant frequency of potentially ‘malignant’ ER in healthy controls.
Patient management and risk stratification
In ERS, survivors of cardiac arrest warrant ICD implantation for secondary prevention. Isoproterenol infusion titrated to increase heart rate to 90–120 bpm can be used acutely to suppress a VF storm, as can overdrive pacing.21 29 Quinidine has been shown to reduce recurrent VF episodes as well as normalise ER changes on the ECG, and targeted ablation of initiating ventricular ectopy has been successful in some patients with refractory disease.29
At present, asymptomatic individuals with an ER pattern without a family history of ERS or a premature or SADS death require no further investigation. The absolute increase in risk associated with J-point elevation in the general population is small and there are no recognised risk stratification strategies to identify those at highest risk. This is of particular concern as the only recognised primary prevention strategy is ICD implantation with its own inherent risks of complication. Using Bayes' law of conditional probabilities, finding J-point elevation in a young adult increases the probability of IVF from 3.4:100 000 to 11:100 000.3 Combining horizontal/descending ST-segment morphology with J-point elevation affords only a modest increase in probability to 30:100 000.23 In addition, the increased risk of SCD associated with inferior J-point elevation of ≥0.2 mV in the general population is not apparent until 10 years after the index ECG.4 The average age of death in the observational cohorts was 10–20 years older than in IVF survivors and the commonest cause of SCD was myocardial infarction.2 4 10 This has led to the suggestion that the most sensible treatment strategy for asymptomatic subjects with ER is modification of risk factors for CAD.36
If there is family history of ERS or premature sudden death or SADS in an immediate blood relative of an asymptomatic individual with an isolated ER pattern, then the a priori risk of further sudden death due to ERS would be expected to be higher. There are, however, few data as yet beyond the greater prevalence of ER seen in immediate blood relatives of patients with SADS to influence management decisions.25 Similarly, if an individual has unexplained syncope, and a vasovagal aetiology appears unlikely, then the relevance of an ER pattern to the underlying aetiology may be important but there are currently no data to guide management. It is important to exclude other conditions that might be responsible for a risk of arrhythmia in these patient groups and their clinical and genetic characterisation should be the subject of prospective research.
In summary, lateral ER with rapidly upsloping ST-segment morphology can be considered benign.7 Individuals with inferior J-point elevation of ≥0.2 mV and horizontal/downsloping ST-segments are likely to be at increased risk of SCD but further risk stratification strategies are required for those who are asymptomatic, have documented syncope or a family history of premature sudden death and/or SADS.7 Late potentials with nocturnal ascendency during 24 h signal-averaged ECG show some promise but thus far neither treadmill testing nor programmed electrophysiological studies have proved useful.2 19 22
In recent years the perception and prevalence of ER has altered drastically and this may in part be due to evolution of the definition. What was once regarded as a benign ECG variant is now considered a marker of risk for arrhythmic death. It appears that the ER pattern originally defined by Wasserburger, and most commonly seen in young athletic men of black ethnicity, is unlikely to signify risk for sudden death in isolation. Establishing the importance of inferior and inferolateral ER with a horizontal/descending ST-segment will require international consensus on common definitions and methodologies for ER measurement as well as collaborative epidemiological and mechanistic research. At present we do not understand the significance of ER in asymptomatic individuals and those without clear evidence of ERS nor do we have the ability to refine the association with unexpected mortality. The risk associated with current definitions is significantly less than the risk associated with the only established preventive treatment, which is ICD placement. Therefore investigation of the asymptomatic individual with an isolated ER pattern and no other features to suggest arrhythmic risk is not indicated.
Early repolarisation (ER) is a common ECG finding, particularly in young adults, men, individuals undertaking regular sporting activity and/or individuals of black ethnicity.
The underlying pathophysiology of ER, depolarisation versus repolarisation, is undecided.
ER was historically a benign finding in the general population and remains benign when found in a typical lateral pattern with rapidly ascending ST-segment morphology and symmetrical T waves.
The definition of ER has evolved and inferior J-point elevation has been associated with risk of sudden cardiac death in survivors of idiopathic ventricular fibrillation and the general population.
Early repolarisation syndrome (ERS), the presence of ER in the context of otherwise idiopathic ventricular fibrillation (VF), is rare.
In ERS secondary prevention implantable cardioverter defibrillator placement is warranted and a VF storm can be managed with isoproterenol and/or overdrive pacing.
There are no established risk stratification or management strategies for ER in other situations.
In an asymptomatic individual with ER there is no indication to investigate unless an immediate blood relative has had a premature or sudden arrhythmic death syndrome death or has ERS.
General disclosures include the receipt of funds for research purposes only from Biotronik and Boston Scientific.
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed.