JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO.

-, 2017
ª 2017 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION ISSN 2405-500X/$36.00

PUBLISHED BY ELSEVIER http://dx.doi.org/10.1016/j.jacep.2016.12.017

The Electrophysiological Substrate

of Early Repolarization Syndrome
Noninvasive Mapping in Patients

Junjie Zhang, PHD,a,b Mélèze Hocini, MD,c Maria Strom, PHD,d Phillip S. Cuculich, MD,a,e Daniel H. Cooper, MD,a,e
Frédéric Sacher, MD, PHD,c Michel Haïssaguerre, MD,c Yoram Rudy, PHDa,b,e

ABSTRACT

OBJECTIVES This study sought to map the epicardial electrophysiological (EP) substrate in early repolarization (ER)
syndrome patients using noninvasive electrocardiographic imaging (ECGI), and to characterize substrate properties that
support arrhythmogenicity.

BACKGROUND The ER pattern is a common ECG finding. Recent studies established a definitive clinical association
between ER and fatal ventricular arrhythmias. However, the arrhythmogenic substrate of ER in the intact human heart
has not been characterized.

METHODS Twenty-nine ER syndrome patients were enrolled, 17 of whom had a malignant syndrome. Characteristics of
the abnormal EP substrate were analyzed using data recorded during sinus rhythm. The EP mapping data were analyzed
for electrogram morphology, conduction, and repolarization. Seven normal subjects provided control data.

RESULTS The abnormal EP substrate in ER syndrome patients has the following properties: 1) abnormal epicardial
electrograms characterized by presence of J waves in localized regions; 2) absence of conduction abnormalities, including
delayed activation, conduction block, or fractionated electrograms; and 3) marked abbreviation of ventricular repolari-
zation in areas with J waves. The action potential duration (APD) was significantly shorter than normal (196
19 ms vs.
235
21 ms; p < 0.05). Shortening of APD occurred heterogeneously, leading to steep repolarization gradients
compared with normal controls (45
17 ms/cm vs. 7
5 ms/cm; p < 0.05). Premature ventricular contractions (PVCs)
were recorded in 2 patients. The PVC sites of origin were closely related to the abnormal EP substrate with J waves and
steep repolarization gradients.

CONCLUSIONS ER is associated with steep repolarization gradients caused by localized shortening of APD. Results
suggest association of PVC initiation sites with areas of repolarization abnormalities. Conduction abnormalities were not
observed. (J Am Coll Cardiol EP 2017;-:-–-) © 2017 by the American College of Cardiology Foundation.

From the aCardiac Bioelectricity and Arrhythmia Center, Washington University, St. Louis, Missouri; bDepartment of Biomedical
Engineering, Washington University, St. Louis, Missouri; cBordeaux University Hospital, LIRYC institute, Pessac, France;
d
CardioInsight Technologies, Cleveland, Ohio; and the eSchool of Medicine, Washington University, St. Louis, Missouri. This study
was supported by National Institutes of Health (NIH)–National Heart, Lung, and Blood Institute grants R01-HL-033343 and R01-
HL-049054 (to Dr. Rudy) and by Washington University Institute of Clinical and Translational Sciences grant UL1-TR000448
from the National Center for Advancing Translational Sciences of the NIH. Dr. Rudy is the Fred Saigh Distinguished Professor at
Washington University. Dr. Hocini has received lecture fees from Medtronic and St. Jude Medical; has served on the advisory
board for Medtronic; and is a stockholder of CardioInsight Technologies. Dr. Strom is a paid employee and stockholder of Car-
dioInsight Technologies. Dr. Cuculich has received research support from NIH and March of Dimes. Dr. Cooper has received
consultant fees and speaker honoraria from Boston Scientific, St. Jude Medical, Medtronic, and Biotronik. Dr. Sacher has received
consultant fees and speaker honoraria from Biosense Webster, St. Jude Medical, Sorin Group, Medtronic, and Biotronik.
Dr. Haïssaguerre is a stockholder of CardioInsight Technologies; and received lecture fees from Biosense Webster and Medtronic.
Dr. Rudy cochairs the scientific advisory board and receives royalties from CardioInsight Technologies. CardioInsight Technolo-
gies does not support any research conducted in Dr. Rudy’s laboratory. Dr. Zhang has reported that he has no relationships
relevant to the contents of this paper to disclose.

Manuscript received June 8, 2016; revised manuscript received December 21, 2016, accepted December 22, 2016.
2 Zhang et al. JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017
EP Substrate in Early Repolarization Patients - 2017:-–-

T
ABBREVIATIONS he early repolarization (ER) pattern electrocardiographic imaging (ECGI) (12,14–21) have
AND ACRONYMS on the electrocardiogram (ECG) is demonstrated its ability to obtain high-resolution
characterized by a J-wave $0.1 mV panoramic EP data of epicardial activation and repo-
AD = activation duration
in inferior and/or lateral leads (1). It resolves larization, and their alteration by disease and in-
AP = action potential
during exercise and fast pacing, but accentu- terventions in humans (17–21). In the current study, we
APD = action potential
ates during bradycardia. The prevalence of characterize the epicardial EP substrate in ERS patients
duration
the ER pattern in the general population is on the basis of high-resolution ECGI data obtained
ARI = activation-recovery
interval
estimated to range between 1% and 13% during sinus rhythm (SR), in an effort to provide in-
(2,3). It is thought to be more common in sights into the substrate properties that support
AT = activation time
males, young athletes, and people of African arrhythmogenicity in these patients.
BrS = Brugada syndrome
descent.
ECG = electrocardiogram
For decades, the ER pattern was consid-
ECGI = electrocardiographic METHODS
ered a benign ECG manifestation. Since the
imaging
1980s, this view has been challenged on the
EGM = electrogram PATIENT POPULATION. ERS patients from Washing-
basis of sporadic observations that linked the
EP = electrophysiological ton University and Bordeaux University Hospital
J-wave with ventricular arrhythmia (4–6). In
ER = early repolarization were enrolled. The clinical diagnosis is ER pattern on
a recent study with a large cohort of patients,
ERS = early repolarization the ECG, defined as an elevation of the J-point
the prevalence of the ER pattern was signifi-
syndrome (J-wave) $0.1 mV in at least 2 contiguous leads. The
cantly higher in patients with idiopathic
LV = left ventricle/ventricular J-wave is manifested either as QRS slurring or
ventricular fibrillation (VF) compared with
PVC = premature ventricular notching in the inferior lead, lateral lead, or both. The
control subjects (1). This was the first study
contraction patients should have at least 1 of the following:
that provided clinical evidence supporting a
RT = recovery time idiopathic VF, unexplained syncope, or familial inci-
definitive association between the ER pattern
RV = right ventricle/ventricular dence of unexplained sudden cardiac death. Patients
and an increased risk of ventricular
SR = sinus rhythm with structural heart disease, coronary artery disease,
arrhythmia. Following this study, additional
VF = ventricular fibrillation or other conditions, including long QT syndrome,
population-based studies provided corrobo-
short QT syndrome, and Brugada syndrome (BrS)
rating evidence (2,7–9). The critical role of the ER
were excluded. All patients had structurally normal
pattern in initiating VF has been supported by ob-
hearts and normal ventricular function. Data from 7
servations of a consistent and marked J-wave accen-
healthy subjects provided normal controls (14).
tuation preceding the onset of arrhythmia (1,10) and
These were healthy adults, ranging in age from 21 to
by electrophysiological (EP) mapping data that sug-
43 years. All control subjects had normal 12-lead ECGs
gested an association between the origin of ectopy
and no known history of heart disease. Protocols were
that initiated VF and the location of repolarization
approved by the institutional review boards at both
abnormalities (1). Meta-analysis on 16 studies
centers; written informed consent was obtained from
involving 334,524 subjects suggests that the ER
all patients.
pattern is associated with an increased risk for sud-
den cardiac arrest, cardiac death, and death from any NONINVASIVE MAPPING. During ECGI, body-surface
cause (11). ECG potentials were acquired simultaneously from
Studies in patients with ER syndrome (ERS) have 256 electrodes using a multichannel data acquisition
been so far confined to investigation of the body- system (Biosemi, Amsterdam, the Netherlands). Next,
surface ECG characteristics and extrapolating the patient underwent thoracic computed tomogra-
possible mechanisms. However, ECG characteristics phy with ECG gating to obtain the epicardial geome-
have been shown to be inadequate measures of un- try and torso electrode positions. Body surface
derlying repolarization properties (12). Reports of potentials were baseline corrected and bandpass
invasive catheter mapping in patients with ERS pro- filtered (0.05 to 400 Hz) to remove high-frequency
vide limited information about the abnormal EP sub- noise and DC component. If necessary, a 60-Hz
strate (1,13). Understanding the mechanism of ER and notch filter was applied to remove power-induced
how it may predispose patients to an increased risk of noise. The pre-processed signals and the patient-
arrhythmias requires detailed characterization of the specific heart–torso geometry were processed with
EP substrate in the intact heart of ERS patients. Simi- ECGI algorithms to reconstruct epicardial potentials,
larly, risk stratification for arrhythmia and differential unipolar electrograms (EGMs), and maps of epicardial
diagnosis between benign and malignant ERS require activation and repolarization. ECGI has been
noninvasive mapping of the EP substrate in individual validated extensively in torso-tank and canine ex-
subjects. Recent developments in noninvasive periments, and in human studies. It provides high-
JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017 Zhang et al. 3
- 2017:-–- EP Substrate in Early Repolarization Patients

accuracy and high-resolution (4 to 6 mm) panoramic applied a multivariable linear regression model with 2
data for noninvasive evaluation of EP properties. In independent, dichotomous variables: study group
order to further evaluate ECGI’s accuracy of deter- (normal/ERS) and healthcare system (Washington
mining the activation-recovery interval (ARI) non- University/Bordeaux University). All tests with
invasively in the in-situ heart of human subjects, an p < 0.05 were considered statistically significant.
additional experiment was performed in patients Statistical analysis was performed by using SPSS
undergoing cardiac surgery. In this experiment, version 19 (IBM, Armonk, New York).
epicardial electrograms were recorded with 240
evenly distributed electrodes, mounted in an epicar- RESULTS
dial sock. Details of the validation method and
results are provided in the Online Appendix Section 1, Twenty-nine ERS patients (26 men, 3 women) were
which includes Online Figures 1–4 and Online Table 1. enrolled in this study. Seventeen (59%) had previous
aborted sudden cardiac death or arrhythmic events
DATA ANALYSIS. Characteristics of the abnormal EP
(idiopathic VF), 10 of which received an implantable
substrate in ERS patients were analyzed by using
cardioverter-defibrillator. Fourteen (48%) experi-
data recorded during SR. The EP mapping data were
enced unexplained syncope. Ten (34%) had a family
analyzed for EGM morphology, conduction, and
history of sudden cardiac death or ERS. Detailed
repolarization. The J-wave on local epicardial uni-
characteristics for individual patients are provided in
polar EGM is defined as J-point elevation of >5% of
Online Tables 2 and 3.
peak-to-peak QRS amplitude. The area of the
epicardium with J waves on the EGMs was measured EGM CHARACTERISTICS AND LOCALIZATION. Figure 1
as a percentage of the total epicardial surface area. shows epicardial EGM characteristics and localiza-
Conduction was evaluated by activation time (AT), tion for representative examples in 3 ERS patients.
activation duration (AD), EGM fractionation, and ECGI data for all patients are shown in Online
voltage. AT was determined by the maximum nega- Figures 8 to 36. Patient ER-11 was an asymptom-
tive slope of the EGM during QRS inscription (22,23). atic patient with a family history of sudden death.
All ATs were referenced to the beginning of the QRS He had an ER pattern in lateral leads. Patient ER-14
in ECG lead II. Epicardial activation isochrone maps experienced previously aborted sudden death, and
were created from ATs. AD was defined as the in- had an ER pattern in both inferior leads and lateral
terval between the earliest and latest AT, consid- leads. Patient ER-20 did not have clinical arrhythmic
ering all epicardial EGMs. Repolarization was events, but experienced unexplained syncope. His
assessed by recovery time (RT) and ARI. Local RT ER pattern was found in inferior leads. The 12-lead
was determined from the maximum positive slope of ECGs of these patients are provided in Online
the EGM T-wave, which reflects the sum of local AT Figures 5 to 7. Data from a normal subject are pro-
and local action potential (AP) duration (APD) vided for reference. Epicardial J-wave was observed
(22,23). Steep RT dispersion has been shown to in EGMs from all 29 ERS patients (0.68
0.25 mV
provide a substrate for unidirectional block and re- vs. 0 mV in control; p < 0.05); 8 in the anterior wall,
entry. ARI was defined as the difference between 19 in the lateral wall, and 23 in the inferior wall
RT and AT. ARI is independent of AT and a surrogate (most patients had an epicardial J-wave in multiple
for local APD (23). From the RT map and ARI map, locations). The proportion of epicardium that pre-
epicardial dispersion of repolarization was measured sented with a J-wave ranged from 23% to 57%, with
as the maximum difference D RT and D ARI between a mean value of 39
9%. EGMs with abnormal low
2 adjacent EGM sites on the epicardium. Epicardial voltage and fractionation, indicative of slow
RT and ARI gradients ( D RT/D x and DARI/ Dx) were discontinuous conduction, were not found in any
computed through division by the distance D x be- ERS patient.
tween the 2 adjacent sites.
EPICARDIAL ACTIVATION. Figure 2 shows ECGI
STATISTICAL ANALYSIS. All continuous data are epicardial activation isochrone maps for a control
presented as mean
SD. Continuous variables were subject and 3 ERS patients (same patients as in
analyzed by unpaired Student t test. The Sat- Figure 1). The epicardial activation pattern during SR
terthwaite modified t test was used for variables with in ERS patients was characterized by a normal
unequal variances. The Mann-Whitney U test was epicardial breakthrough (white asterisks) in the right
used for variables with non-normal distribution. To ventricle (RV). The excitation wave front spread
account for potential influences of combining sub- uniformly and rapidly to activate both ventricles. The
jects from 2 healthcare systems on the results, we left ventricular (LV) base was the latest region to
4 Zhang et al. JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017
EP Substrate in Early Repolarization Patients - 2017:-–-

F I G U R E 1 Epicardial J-Wave Distribution and Magnitude

Abnormal epicardial electrograms (EGMs) with a J-wave that resembles the early repolarization (ER) pattern (terminal-QRS notch or slurring)
on the surface electrocardiogram are observed in ER syndrome (ERS) patients, but not seen in normal subjects. Example maps of epicardial J-
wave magnitude in 3 ERS patients are shown. The maps are shown in anterior view and inferior view for each subject. The locations of epicardial
J waves are ER-11: lateral and apical LV; ER-14: lateral and inferior LV; ER-20: lateral and inferior LV. Inset: representative EGMs with a J-wave
(QRS only) in ERS patients (B to D) and EGMs from corresponding locations in the normal control subject (A). Twelve-lead ECGs for ERS
patients are shown to the right of the epicardial maps. A red asterisk indicates the presence of ER pattern. LA ¼ left atrium; LV ¼ left ventricle;
RA ¼ right ventricle; RV ¼ right ventricle.

activate. The activation pattern was not affected by were present (Figures 3B to 3D, white arrows).
the presence of the J-wave; regions of slow conduc- Figure 4E shows EGMs with prominent J-wave (loca-
tion (isochrone crowding) or conduction block (adja- tion 1) and EGMs without J-wave from an adjacent
cent activation times differ by more than 50 ms) were location (location 2) for 3 ER patients. In patient ER-
not found in ERS patients. A similar activation 11, for example, ATs at location 1 and 2 were similar
sequence was observed in the normal control sub- (AT(1) ¼ 66 ms, AT(2) ¼ 61 ms), but repolarization
jects, as well as in the isolated human heart from in- dispersion was observed between the 2 locations:
dividuals with no history of cardiac disease (24). AD, DRT was 65 ms (RT(1) ¼ 232 ms, RT(2) ¼ 297 ms), and
the time needed for activation of both ventricles, was DARI was 70 ms (ARI(1) ¼ 166 ms, ARI(2) ¼ 236 ms).
54
7 ms for ERS patients, comparable with 47
9 ms This resulted in steep gradients of RT and ARI (D RT/
for normal control. Dx ¼ 43 ms/cm; DARI/Dx ¼ 46 ms/cm), much steeper
than those of control (typically 5 to 8 ms/cm and 4 to
EPICARDIAL REPOLARIZATION. Representative 10 ms/cm, respectively, in the same region). Patients
maps of RT (Figure 3) and ARI (Figure 4) from a normal ER-14 and ER-20 also had shortened RT and ARI, as
subject and 3 ERS patients (same patients as in well as increased gradients of RT and ARI, but the
Figure 1) demonstrate marked local changes in repo- location of abnormal repolarization varied.
larization in ERS patients compared with normal Table 1 summarizes the ECGI-derived parameters,
control. Abnormal repolarization was observed pri- adjusted for differences between the 2 participating
marily in regions with a prominent J-wave. Steep centers. Compared with normal control, ERS patients
epicardial RT gradients and ARI gradients occurred had shortened RT (223
28 ms vs. 265
30 ms;
mostly at the border of regions where J-wave EGMs p < 0.05) and ARI (196
19 ms vs. 235
21 ms;
JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017 Zhang et al. 5
- 2017:-–- EP Substrate in Early Repolarization Patients

F I G U R E 2 Activation Isochrone Maps

Examples of activation during SR are shown for a normal control subject (A) and 3 ERS patients (B to D). The maps are shown in anterior view
and inferior view for each subject. ERS patients and the normal subject have similar activation patterns. After breakthrough in the anterior
right ventricle (asterisk), the wave front propagates uniformly to activate both ventricles. The LV base is the latest region to activate.
Conduction block and slow conduction were not observed in the ERS patients. Abbreviations as in Figure 1.

p < 0.05), increased repolarization dispersion D RT and inferior LV (not shown). In patient ER-11, about
(52
15 ms vs. 18
14 ms; p < 0.05) and D ARI 32% of the epicardium had J waves in the EGMs,
(53
15 ms vs. 16
10 ms; p < 0.05), and increased located in the anterolateral and apical LV. Figure 5B
repolarization gradients D RT/ Dx (48
18 ms/cm vs. shows that in patient ER-1, the anterior RV had the
8
6 ms/cm; p < 0.05) and D ARI/D x (45
17 ms/cm shortest ARI (about 140 ms), and the majority of LV
vs. 7
5 ms/cm; p < 0.05). Online Table 4 shows that had ARI <180 ms, significantly below the normal
there is no significant difference between healthcare range. Patient ER-11 had shortened ARI (about
systems for all parameters. 160 ms) in the anterolateral and apical LV. Figure 5C
shows the PVC activation patterns. In patient ER-1,
VENTRICULAR ARRHYTHMIAS. Premature ventricu- the PVC originated from the apical LV region, then
lar contractions (PVCs) were recorded in 2 patients, propagated towards the basal RV. In patient ER-11,
ER-1 and ER-11. Figure 5 shows substrate maps the PVC initiation site was located in the mid-
(J-wave and ARI) and PVC activation maps for pa- anterior LV near the septum. The inferior LV was
tients ER-1 and ER-11. Figure 5A shows that in patient the latest region to activate during a PVC. In both
ER-1, a large portion of epicardium (54%) was cases, locations of the EP substrate with J-wave
affected by abnormal EGMs with J waves, including EGMs, shortened ARIs, and steep ARI gradients
the anterior RV, anterior LV, lateral LV, apical LV, correlated with the PVC sites of origin.
6 Zhang et al. JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017
EP Substrate in Early Repolarization Patients - 2017:-–-

F I G U R E 3 RT Maps

Examples of epicardial recovery time (RT) during SR are shown for a normal control subject (A) and 3 ERS patients (B to D). Maps are shown in
anterior view and inferior view for each subject. ERS patients have regions with abnormally short RT (dark blue). White arrows in B and C
point to regions with steep RT gradients. Abbreviations as in Figures 1 and 2.

DISCUSSION transitions to the ST segment. These late QRS char-

acteristics could reflect delayed activation or early
On the basis of studies in the canine ventricular repolarization; these 2 possibilities cannot be differ-
wedge preparation (4), differences in the magnitude entiated from the body-surface ECG. The name early
of the transient outward current, I to, in epicardium repolarization syndrome was given on the basis of a
and endocardium result in different configurations hypothesis, not on a confirmed underlying mecha-
of the AP phase-1 notch across the ventricular wall. nism. Therefore, the first obvious question we tried to
The resulting AP transmural gradient is thought answer with ECGI was whether the ECG abnormalities
to be responsible for the J-wave on the ECG. in ERS patients are due to delayed activation in re-
Augmentation of a repolarizing current by mutation gions of the ventricles, or indeed due to early regional
can result in accentuation of the AP notch and loss of ventricular repolarization. The high-resolution
the AP dome, leading to the development of panoramic mapping was essential for characterizing
arrhythmias due to a mechanism termed phase-2 the properties of the EP substrate and answering this
re-entry (4). question. The maps showed no evidence for delayed
The clinical ERS phenotype is defined on the basis activation in any of the patients. J waves were found
of body-surface ECG morphologies. These include in epicardial EGMs, indicative of the presence of a
presence of a J-wave in the form of a positive voltage gradient during phase-1 of the AP at the
deflection or slurring of the QRS waveform as it EGM location, similar to J waves recorded with
JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017 Zhang et al. 7
- 2017:-–- EP Substrate in Early Repolarization Patients

F I G U R E 4 ARI Maps

Continued on the next page

8 Zhang et al. JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017
EP Substrate in Early Repolarization Patients - 2017:-–-

and ARI dispersion, independent of conduction. With

T A B L E 1 Comparison of ECGI Parameters Between ERS Patients
and Normal Subjects
ARI being the surrogate for local APD, shortened ARI
suggests abbreviated AP. The formation of regions
Normal Control ERS Patients
(n ¼ 7) (n ¼ 29)
with steep gradient of repolarization is caused by
J-wave magnitude (mV)* 0 0.68
0.25 spatially heterogeneous abbreviation of the AP over
AD (ms) 47
9 54
7 short distances. By contrast, the apex-to-base ARI
Mean RT (ms)* 265
30 223
28 dispersion (about 42 ms) and interventricular ARI
DRT (ms)* 18
14 52
15 dispersion (about 32 ms) result in much shallower
DRT/Dx (ms/cm)* 8
6 48
18 gradients in normal subjects (14). Results from this
Mean ARI (ms)* 235
21 196
19
study are consistent with a previous case report that
DARI (ms)* 16
10 53
15
presented preliminary data from 2 ERS patients
DARI/Dx (ms/cm)* 7
5 45
17
mapped by ECGI (19).
Values are mean
SD. *p < 0.05. Since the establishment of a link between the ER
DARI ¼ ARI dispersion; DARI/Dx ¼ ARI gradient; DRT ¼ RT dispersion; DRT/Dx ¼ pattern on the ECG and fatal cardiac arrhythmias (1),
RT gradient; AD ¼ activation duration; ARI ¼ activation-recovery interval;
ECGI ¼ electrocardiographic imaging; ERS ¼ early repolarization syndrome; numerous studies have been conducted in an effort
RT ¼ recovery time.
to stratify risk of ventricular arrhythmias in patients
with the ER pattern. Risk stratification in this pop-
ulation is of clinical importance, given the preva-
pseudo-ECG leads from the ventricular wedge prep- lence of the ER pattern in the general population.
aration. By contrast, J waves in body-surface ECG Importantly, young, otherwise healthy individuals
leads are not location specific, because each lead with ER may have increased vulnerability to idio-
records a signal that reflects integrated electrical pathic VF. Surface ECG markers, including the
activity over the entire heart. There was marked amplitude and distribution of the J-wave,
abbreviation of ventricular repolarization in areas morphology of the ST-segment, and the presence of
with J-wave epicardial EGMs, suggesting that loss of ventricular ectopy have very limited success in risk
the AP dome gave rise to a transmural voltage stratification, although they have been shown to
gradient at the EGM location. The results are consis- associate with arrhythmic risk in patients presenting
tent with observations of a recent experimental study with unexplained syncope (2,8,9,26,27). A recent
(25) and provide evidence in support of the early study investigated the role of invasive EP studies in
repolarization mechanism in patients. risk stratification in ERS patients (28). Results indi-
Figures 3 and 4 demonstrate that there is significant cate that VF inducibility does not have a role in risk
regional abbreviation of the AP (based on recon- stratification in ERS patients; it neither predicts
structed ARIs) on the ventricular epicardium of ERS arrhythmic risks, nor correlates with the ECG
patients compared with normal control. As shown in markers listed in the preceding text. Better under-
Figure 4E, in regions with abnormal EGMs, the standing of the EP substrate and a capability for its
epicardial J-wave is followed by shortening of RT. noninvasive mapping can help with the develop-
These regions were located in close proximity to re- ment of effective diagnostic and risk stratification
gions with relatively longer RT. The regional differ- approaches. We will examine this possibility with
ences in RT gave rise to steep repolarization ECGI in a future study.
gradients. Given the fast and uniform conduction, the ER and BrS are often collectively referred to as
difference in ATs between 2 adjacent locations is too “J-wave syndromes” because the 2 conditions share
small to account for the difference in RTs. Thus, RT similar ECG characteristics and a number of clinical
and RT dispersion are primarily determined by ARI features. Both are associated with vulnerability to VF

F I G U R E 4 Continued

(A to D) Activation-recovery interval (ARI) maps for a normal control subject and 3 ERS patients. Maps are shown in anterior view and inferior
view for each subject. ERS patients have regions with abnormally short ARI (dark blue). White arrows in B and C point to regions with steep ARI
gradients. Top 2 rows of E show electrocardiographic imaging (ECGI)-reconstructed EGMs from 2 adjacent locations in each ERS patient.
Location 1 (top row): prominent J-wave and short ARI; location 2 (middle row): absence of J-wave and normal ARI. There is a steep gradient of
repolarization across these 2 locations. Bottom row of E shows 3 EGMs from the normal subject, from locations marked in A. The time instances
of activation (AT) (black square), recovery (RT) (red square), and corresponding ARIs are indicated. Abbreviations as in Figures 1 and 3.
JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017 Zhang et al. 9
- 2017:-–- EP Substrate in Early Repolarization Patients

F I G U R E 5 The EP Substrate in Relation to PVCs

(A) Epicardial J-wave magnitude and (B) ARI maps during SR. (C) Premature ventricular contraction (PVC) activation map and 12-lead electrocardiogram (ECG). Each
panel shows maps for 2 patients, ER-1 (left, PVC coupling interval ¼ 635 ms) and ER-11 (right, PVC coupling interval ¼ 520 ms). The hearts are displayed in anterior and
lateral views. White arrows in B point to regions with steep ARI gradients. In C, J-wave EGMs from the PVC site of origin are shown (black arrow); 12-lead ECG depicting
the PVC is shown to the right of the map. The PVC initiation sites are indicated by plus signs (D). Note that an electrophysiological (EP) substrate with abnormal EGMs,
shortened ARIs, and steep ARI gradients correlated with the PVC site of origin in both cases. Abbreviations as in Figures 1 to 4.

in young adults without apparent structural heart inhibition of VF) (25,30). Nevertheless, there are
disease, and have a higher prevalence in males than several key characteristics that distinguish ERS from
females (29). The J-wave and associated ST-segment BrS, suggesting different underlying mechanisms for
elevation are accentuated before the arrhythmic the ECG phenotype and arrhythmogenesis. Moderate
event, and VF is often initiated during bradycardia. structural abnormalities (RV interstitial derange-
Their responses to quinidine and isoproterenol are ment), though undetectable by noninvasive clinical
ameliorative (normalization of the J-wave and imaging, have been found in endomyocardial biopsies
10 Zhang et al. JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017
EP Substrate in Early Repolarization Patients - 2017:-–-

of BrS patients (31). In a recent study, ECGI was per- STUDY LIMITATIONS. The control data were not ob-
formed in 25 BrS patients (20). Altered or delayed tained at the time of this study; they were recorded
epicardial activation in the RV outflow tract was previously with the same ECGI methodology. The
found in 20 patients. EGMs with abnormally low ability to generalize the results and draw a quantita-
voltage and fractionation were found in all patients. tive conclusion regarding ECGI sensitivity in detect-
The abnormal EGMs in the setting of BrS can reflect ing and localizing J waves is limited by the relatively
slow discontinuous conduction, as well as abnormal small number of 29 patients.
repolarization as demonstrated in a recent experi- In this first study, not all patients had a malignant
mental study (32). Prolonged repolarization and steep form of ERS. Subjects with ECG ER pattern and a
repolarization gradients were also observed. These broad spectrum of evidence for arrhythmogenic sub-
findings demonstrate the coexistence of conduction strate (idiopathic VF, syncope, family history of sud-
and repolarization abnormalities in BrS. The den cardiac death) were included. ER subjects
abnormal EP substrate was confined to the RV without evidence of arrhythmogenesis in them or
outflow tract. By contrast, the widespread distribu- their family were excluded. Properties of the
tion of J waves in ERS patients indicates that the EP abnormal EP substrate (J-wave EGMs and regional
substrate may not be localized to a specific region of APD shortening with steep repolarization gradients)
the heart. The current study identifies additional were consistent across the patient population. Future
differences between ERS and BrS. Conduction delays largescale studies are needed to differentiate be-
and EGMs indicative of slow discontinuous conduc- tween subgroups (e.g., based on symptoms and ge-
tion were not found in ERS patients. The location and notype) and to establish the potential of ECGI for
size of the abnormal EP substrate varied among ERS noninvasive diagnosis and risk stratification in ERS
patients. Instead of ARI prolongation in BrS patients, patients.
shortened ARIs were observed in ERS patients during Changes in heart rate are known to affect the ER
SR. pattern on the ECG, but the present study was con-
There is a paucity of EP mapping data of ERS- ducted only at resting heart rate. Similarly, several
related arrhythmias, given the small number of ERS drugs (including quinidine, isoproterenol, milrinone,
patients who have experienced aborted sudden car- and cilostazol) have been shown to have ameliorative
diac death. Catheter mapping was performed in effects on the ER pattern. It will be constructive to
8 ERS patients with a history of idiopathic VF (1). In characterize the effects of increased heart rate and of
6 patients with ER pattern only in the inferior ECG these drugs on the EP substrate in terms of spatial
leads, all ectopies were mapped to the inferior ven- repolarization patterns and dispersion in future
tricular wall. In 2 patients with ER pattern in both studies using ECGI.
inferior and lateral leads, ectopy originated from It is well established that the J-point amplitude
multiple locations. In the current study, PVCs were increases significantly at the onset of an arrhythmic
recorded in 2 patients. The PVC sites of origin were event. However, polymorphic VT or VF was not
located in an epicardial region with marked J waves, recorded during ECGI in any of the patients in this
short ARIs, and steep ARI gradients. Proximity of study. The 2 patients (ER-1 and ER-11) in Figure 5 had
areas with steep ARI gradients to PVC sites of origin isolated PVCs that were not followed by VT. The
supports the hypothesis that reactivation of early- study characterizes the EP substrate and records iso-
repolarizing cells by local currents could generate lated PVCs, but does not provide direct recordings of
the ectopy (33). The timing of the J-wave did not VT in ERS patients.
change before the onset of PVC compared with ECGI can reconstruct epicardial dispersion of
baseline SR beats and beats following the PVC. In repolarization. Experimental studies (25) have shown
addition, the T-wave morphology of the preceding that in addition to markedly elevated epicardial
SR beat was not altered before the PVC. This sug- dispersion of repolarization, transmural dispersion of
gests lack of or minimal coupling between the repolarization is an important component of the
J-wave and PVC, consistent with the view that the arrhythmogenic substrate. However, transmural
J-wave is a marker of steep phase-1 repolarization dispersion of repolarization throughout the heart
gradients (30,33). The localization of the PVC origin cannot be measured directly with ECGI.
in regions with steep repolarization gradients sug-
gests that local PVC initiation could be a trigger of CONCLUSIONS
re-entrant arrhythmia. A recent study using ECGI in
an ERS patient during VF identified rotors in the Noninvasive ECGI revealed the presence of an
inferior-lateral LV wall (26). abnormal EP substrate in ERS patients,
JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017 Zhang et al. 11
- 2017:-–- EP Substrate in Early Repolarization Patients

characterized by spatially heterogeneous APD Department of Biomedical Engineering, Washington

shortening, steep repolarization gradients, and University, Campus Box 1097, One Brookings Drive,
regional distribution of J-wave EGMs on the St. Louis, Missouri 63130. E-mail: rudy@wustl.edu.
epicardium. Steep repolarization gradients provide a
substrate that is susceptible to the development of
PERSPECTIVES
asymmetrical conduction and re-entrant arrhyth-
mias. Conduction abnormalities in the form of slow
conduction or conduction block were not present. COMPETENCY IN MEDICAL KNOWLEDGE: The arrhythmo-
PVC sites of origin colocalized to the regions of genic substrate of ER syndrome in the intact human heart has not
J-wave presence and steep repolarization gradients been characterized. Noninvasive EP mapping with ECGI demon-
mapped during SR. strates that ER is associated with localized shortening of the
action potential duration, giving rise to steep repolarization
ACKNOWLEDGMENTS The authors thank Drs.
gradients that provide a substrate for re-entrant arrhythmia.
Michele Orini, Pier D. Lambiase, and Peter Taggart
(University College London, United Kingdom) for
TRANSLATIONAL OUTLOOK: Being noninvasive, ECGI is a
providing the intraoperative cardiac mapping data in
promising tool for arrhythmic risk stratification in ER patients on
the Online Appendix Section 1. The authors also
the basis of EP substrate properties; this possibility requires
thank Eric Novak for his expert help with the statis-
evaluation in a larger study. With its single-beat mapping
tical analysis.
capability, ECGI could be used to locate the origin of frequent
arrhythmic PVCs for possible intervention.
ADDRESS FOR CORRESPONDENCE: Dr. Yoram Rudy,
Cardiac Bioelectricity and Arrhythmia Center,

REFERENCES

1. Haissaguerre M, Derval N, Sacher F, et al. Sud- associated with favorable long-term outcome. Cir- substrate in patients after myocardial infarction:
den cardiac arrest associated with early repolari- culation 2011;123:2666–73. noninvasive characterization with electrocar-
zation. N Engl J Med 2008;358:2016–23. diographic imaging. J Am Coll Cardiol 2011;58:
10. Nam GB, Ko KH, Kim J, et al. Mode of onset of
1893–902.
2. Tikkanen JT, Anttonen O, Junttila J, et al. Long- ventricular fibrillation in patients with early repo-
term outcome associated with early repolarization larization pattern vs. Brugada syndrome. Eur Heart 19. Ghosh S, Cooper DH, Vijayakumar R, et al.
on electrocardiography. N Engl J Med 2009;361: J 2010;31:330–9. Early repolarization associated with sudden death:
2529–37. Insights from noninvasive electrocardiographic
11. Cheng YJ, Lin XX, Ji CC, et al. Role of early
3. Kambara H, Phillips J. Long-term evaluation of imaging. Heart Rhythm 2010;7:534–7.
repolarization pattern in increasing risk of death.
early repolarization syndrome (normal variant J Am Heart Assoc 2016;5:e003375. 20. Zhang J, Sacher F, Hoffmayer K, et al. Cardiac
RS-T segment elevation). Am J Cardiol 1976;38: electrophysiologic substrate underlying the ECG
12. Burnes JE, Ghanem RN, Waldo AL, Rudy Y.
157–61. phenotype and electrogram abnormalities in Bru-
Imaging dispersion of myocardial repolarization, I:
4. Gussak I, Antzelevitch C. Early repolarization gada syndrome patients. Circulation 2015;131:
comparison of body-surface and epicardial mea-
syndrome: clinical characteristics and possible 1950–9.
sures. Circulation 2001;104:1299–305.
cellular and ionic mechanisms. J Electrocardiol 21. Vijayakumar R, Silva JNA, Desouza KA, et al.
13. Haissaguerre M, Shoda M, Jais P, et al. Map-
2000;33:299–309. Electrophysiologic substrate in congenital long QT
ping and ablation of idiopathic ventricular fibril-
5. Boineau JP. The early repolarization variant– syndrome: noninvasive mapping with electrocar-
lation. Circulation 2002;106:962–7.
normal or a marker of heart disease in certain diographic imaging (ECGI). Circulation 2014;130:
subjects. J Electrocardiol 2007;40:3.e11–6. 14. Ramanathan C, Jia P, Ghanem R, Ryu K, 1936–43.
Rudy Y. Activation and repolarization of the
6. Kalla H, Yan GX, Marinchak R. Ventricular 22. Haws CW, Lux RL. Correlation between in vivo
normal human heart under complete physiological
fibrillation in a patient with prominent J (Osborn) transmembrane action-potential durations and
conditions. Proc Natl Acad Sci U S A 2006;103:
waves and ST segment elevation in the inferior activation-recovery intervals from electrograms:
6309–14.
electrocardiographic leads: a Brugada syndrome effects of interventions that alter repolarization
variant? J Cardiovasc Electrophysiol 2000;11: 15. Ghanem RN, Burnes JE, Waldo AL, Rudy Y. time. Circulation 1990;81:281–8.
95–8. Imaging dispersion of myocardial repolarization, ii:
23. Coronel R, de Bakker JM, Wilms-Schopman FJ,
Noninvasive reconstruction of epicardial mea-
7. Rosso R, Kogan E, Belhassen B, et al. J-point et al. Monophasic action potentials and activation
sures. Circulation 2001;104:1306–12.
elevation in survivors of primary ventricular recovery intervals as measures of ventricular ac-
fibrillation and matched control subjects: inci- 16. Ramanathan C, Ghanem RN, Jia P, Ryu K, tion potential duration: Experimental evidence to
dence and clinical significance. J Am Coll Cardiol Rudy Y. Noninvasive electrocardiographic imaging resolve some controversies. Heart Rhythm 2006;
2008;52:1231–8. for cardiac electrophysiology and arrhythmia. Nat 3:1043–50.
Med 2004;10:422–8.
8. Derval N, Simpson CS, Birnie DH, et al. Preva- 24. Durrer D, Van Dam RT, Freud G, Janse M,
lence and characteristics of early repolarization in 17. Wang Y, Cuculich PS, Zhang J, et al. Noninva- Meijler F, Arzbaecher R. Total excitation of the
the CASPER registry: Cardiac Arrest Survivors With sive electroanatomic mapping of human ventric- isolated human heart. Circulation 1970;41:
Preserved Ejection Fraction Registry. J Am Coll ular arrhythmias with electrocardiographic 899–912.
Cardiol 2011;58:722–8. imaging. Sci Transl Med 2011;3:98ra84.
25. Koncz I, Gurabi Z, Patocskai B, et al. Mecha-
9. Tikkanen JT, Junttila MJ, Anttonen O, et al. Early 18. Cuculich PS, Zhang J, Wang Y, et al. The nisms underlying the development of the elec-
repolarization electrocardiographic phenotypes electrophysiological cardiac ventricular trocardiographic and arrhythmic manifestations of
12 Zhang et al. JACC: CLINICAL ELECTROPHYSIOLOGY VOL. -, NO. -, 2017
EP Substrate in Early Repolarization Patients - 2017:-–-

early repolarization syndrome. J Mol Cell Cardiol 29. Antzelevitch C, Yan GX. J-wave syndromes: 33. Benito B, Guasch E, Rivard L, Nattel S. Clinical
2014;68:20–8. Brugada and early repolarization syndromes. Heart and mechanistic issues in early repolarization of
Rhythm 2015;12:1852–66. normal variants and lethal arrhythmia syndromes.
26. Mahida S, Derval N, Sacher F, et al. History and
J Am Coll Cardiol 2010;56:1177–86.
clinical significance of early repolarization syn- 30. Antzelevitch C, Yan GX. J wave syndromes.
drome. Heart Rhythm 2015;12:242–9. Heart Rhythm 2010;7:549–58.

27. Rosso R, Glikson E, Belhassen B, et al. Dis- 31. Frustaci A, Priori SG, Pieroni M, et al. Cardiac KEY WORDS early repolarization,
tinguishing “benign” From “malignant early repo- histological substrate in patients with clinical idiopathic ventricular fibrillation, mapping,
larization”: The value of the ST-segment phenotype of Brugada syndrome. Circulation sudden cardiac death
morphology. Heart Rhythm 2012;9:225–9. 2005;112:3680–7.

28. Mahida S, Derval N, Sacher F, et al. Role of 32. Szél T, Antzelevitch C. Abnormal repolariza-
electrophysiological studies in predicting risk tion as the basis for late potentials and fraction- A PPE NDI X For an expanded methods and
of ventricular arrhythmia in early repolariza- ated electrograms recorded from epicardium in results sections as well as supplemental
tion syndrome. J Am Coll Cardiol 2015;65: experimental models of Brugada syndrome. J Am figures and tables, please see the online
151–9. Coll Cardiol 2014;63:2037–45. version of this paper.