Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/284161293

ElectrocardiographicanteriorT-waveinversioninathletesofdifferentethnicities:differentialdiagnosisbetweenathlete'sheartandcardiomyopathy

ARTICLEinEUROPEANHEARTJOURNAL·NOVEMBER2015

ImpactFactor:15.2·DOI:10.1093/eurheartj/ehv591

CITATION

1

READS

100

11AUTHORS,INCLUDING:

NabeelSheikh

StGeorge's,UniversityofLondon

49PUBLICATIONS239CITATIONS

SEEPROFILE

AbbasZaidi

StGeorge's,UniversityofLondon

45PUBLICATIONS233CITATIONS

SEEPROFILE

MaurizioSchiavon

UnitàLocaleSocioSanitariaPadovaULSS16

138PUBLICATIONS2,165CITATIONS

SEEPROFILE

SanjaySharma

StGeorge's,UniversityofLondon

435PUBLICATIONS6,504CITATIONS

SEEPROFILE

Availablefrom:NabeelSheikh

Retrievedon:04January2016

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CLINICAL RESEARCHSports cardiology

Electrocardiographic anterior T-wave inversionin athletes of different ethnicities: differentialdiagnosis between athlete’s heart andcardiomyopathyChiara Calore1†, Alessandro Zorzi1†, Nabeel Sheikh2†, Alberto Nese1,Monica Facci1, Aneil Malhotra2, Abbas Zaidi2, Maurizio Schiavon3,Antonio Pelliccia4, Sanjay Sharma2‡, and Domenico Corrado1‡*

1Inherited Arrhythmogenic Cardiomyopathy Unit, Department of Cardiac, Thoracic, and Vascular Sciences, University of Padova, Via N. Giustiniani 2, 35121 Padova, Italy; 2St George’sUniversity of London, London, UK; 3Center of Sports Medicine, Department of Public Health, Padova, Italy; and 4Center for Sport Medicine and Sciences (CONI), Rome, Italy

Received 10 April 2015; revised 9 July 2015; accepted 12 October 2015

Aims Anterior T-wave inversion (TWI) is a recognized variant in athletes of African/Afro Caribbean origin and some endur-ance athletes; however, the presence of this specific repolarization anomaly also raises the possibility of cardiomyop-athy. The differentiation between physiological adaptation and cardiomyopathy may be facilitated by examining otherrepolarization parameters, notably the J-point and the ST-segment.

Methods andresults

We compared the electrocardiogram pattern of anterior TWI in a series of 80 healthy athletes (median age 21 years,75% males); 95 patients with hypertrophic cardiomyopathy (HCM) (median age 46 years, 75% males), including 26 af-fected athletes; and 58 patients with arrhythmogenic right ventricular cardiomyopathy (ARVC) (median age 32 years,71% males), including 9 affected athletes. Athletes and patients were of either white/Caucasian or black/Afro Caribbeandescent and showed TWI ≥1 mm in ≥2 contiguous anterior leads (V1–V4). We aimed to identify repolarization pat-terns for differentiating physiologic from pathologic TWI. After adjustment for age, gender, and ethnicity, J-point ele-vation ,1 mm (but no ST-segment elevation without J-point elevation) in the anterior leads showing TWI and TWIextending beyond V4 remained independent predictors for both ARVC, with OR ¼ 569 (95% CI ¼ 38–8545;P , 0.001) and OR ¼ 6.0 (95% CI ¼ 1.2–37.8; P ¼ 0.03), respectively, and HCM with OR ¼ 227 (95% CI ¼ 12–1620; P , 0.001) and OR ¼ 331 (95% CI ¼ 20–2752; P ¼ 0.001), respectively. In athletes with anterior TWI, the com-bination of J-point elevation ≥1 mm and TWI not extending beyond V4 excluded a cardiomyopathy, either ARVC orHCM, with 100% sensitivity and 55% specificity.

Conclusion The combination of J-point elevation and TWI confined to lead V1–V4 offers the potential for an accurate differenti-ation between ‘physiologic’ and ‘cardiomyopathic’ anterior TWI, among athletes of both white/Caucasian or black/AfroCaribbean descent. Conversely, ST-segment elevation without J-point elevation preceding anterior TWI may reflectcardiomyopathy.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Keywords Hypertrophic cardiomyopathy † Arrhythmogenic right ventricular cardiomyopathy † Pre-participation screening †

Sports cardiology † Sudden death

† These authors contributed equally to this article and are shared first authors.‡ These authors contributed equally to this article and are shared senior authors.

* Corresponding author. Tel: +39 049 8212458, Fax: +39 049 8212309, Email: domenico.corrado@unipd.it

Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2015. For permissions please email: journals.permissions@oup.com.

European Heart Journaldoi:10.1093/eurheartj/ehv591

European Heart Journal Advance Access published November 17, 2015 by guest on N

ovember 18, 2015

http://eurheartj.oxfordjournals.org/D

ownloaded from

IntroductionAnterior (V1–V4) T-wave inversion (TWI) on an athlete’s 12-leadelectrocardiogram (ECG) is often considered an abnormal findingand warrants thorough clinical investigation to exclude diseases fre-quently implicated in exercise-related sudden cardiac death (SCD),notably the cardiomyopathies.1 – 3 Anterior TWI is detected in2–4% of patients with hypertrophic cardiomyopathy (HCM)4 butmay be present in as many as 80% of patients with arrhythmogenicright ventricular cardiomyopathy (ARVC).5 – 8 However, anteriorTWI is also observed on the ECG of healthy athletes, with a re-ported prevalence of 2–7% in the general population of Caucasian(white) athletes,4,6,9 12–13% in African/Caribbean (black) ath-letes,4,10 and 14–28% in the subgroup of endurance athletes ofmixed ethnicity.11 – 12 Previous studies comparing the prevalenceof anterior TWI in black athletes and black patients with HCM sug-gest that TWI limited to V1–V4 and preceded by convexST-segment elevation is likely to represent an ethnic variant of the‘athlete’s heart’.4,10 The recently published Seattle criteria recom-mend that anterior TWI preceded by a convex ST-segment eleva-tion in black athletes does not require further assessment toexclude a cardiomyopathy in the absence of symptoms, a positivefamily history, or an abnormal physical examination.2 Althoughthis approach offers the potential to lower the high burden of false-positive ECG rates traditionally observed in the black athletic popu-lation,11 – 16 the diagnostic accuracy of this ECG criterion has notbeen validated quantitatively. A further unresolved issue is whetherthe Seattle recommendation is also applicable to white athletes,considering that there are emerging reports that up to 8% of whiteendurance athletes show anterior TWI.11

The present study compared several qualitative and quantitativeelectrocardiographic features in healthy athletes and patients withHCM and ARVC of both black and white ethnicity who showedTWI in the anterior leads. The aim was to identify co-existingECG parameters that may help to differentiate physiological anter-ior TWI from pathological anterior TWI.

Methods

Study populationThis study compared three different cohorts of subjects, each containingindividuals of both white and black ethnicity, as follows: (i) 80 healthycompetitive athletes; (ii) 95 patients with HCM, including 26 affectedathletes; and (iii) 58 patients with ARVC, including 9 affected athletes.For each group, the criterion for inclusion into the study was the pres-ence of TWI of ≥1 mm in at least two contiguous anterior leads (V1 toV4), in the absence of right or left bundle branch block (LBBB) and re-gardless of the presence or absence of TWI in the other leads. In thegroup of healthy athletes, the ECG used for analysis was acquired atthe time of the last pre-participation evaluation. Among non-athlete pa-tients with cardiomyopathy, the ECG used for analysis was recordedduring the last follow-up visit. For athletes diagnosed with a cardiomy-opathy, we used the ECG performed during pre-participation evaluationthat led to the diagnosis of cardiomyopathy following subsequent inves-tigation. The study population consisted of consecutive patients and ath-letes detected with anterior TWI between 2010 and 2012 in threecollaborative centres: the Inherited Arrhythmogenic CardiomyopathyUnit of the Department of Cardiac, Thoracic and Vascular Sciences,

University of Padova, Italy (white patients with HCM, either non-athletes and athletes; white and black patients with ARVC, either non-athletes and athletes); the Division of Cardiovascular Sciences, StGeorge’s University of London, UK (white athletes with HCM, blackhealthy athletes and black patients, either athletes or non-athletes,with HCM); and the Center for Sport Medicine and Science (CONI),Rome, Italy (white healthy athletes).

Healthy athletesAll athletes underwent detailed clinical evaluation to rule out underlyingcardiac disease, including family and personal history, physical examin-ation, basal and exercise ECG, 24-h Holter monitoring, and echocardi-ography. Athletes with a previous history of cardiac or pulmonarydisease, systemic hypertension, diabetes mellitus, family history of car-diomyopathy, or family history of premature (≤40 years) SCD were ex-cluded from the study. Athletes with electrocardiographic evidence ofventricular pre-excitation or complete bundle branch block were alsoexcluded. Sports activity was classified according to the two generaltypes of exercise (dynamic and static), based on the Mitchell classifica-tion.17 Contrast-enhanced cardiac magnetic resonance imaging was re-served for selected cases with findings suspicious of an underlyingcardiomyopathy, including (i) athletes exhibiting ‘grey zone’ left ven-tricular (LV) hypertrophy on echocardiography, suspicious of mildHCM and (ii) athletes with right ventricular (RV) structural or functionalabnormalities or arrhythmias of RV origin, suspicious of ARVC. The dif-ferential diagnosis between athlete’s heart and cardiomyopathy wasbased on current criteria.18

Hypertrophic cardiomyopathy patientsThe diagnosis of HCM was based upon a maximal LV wall thickness of≥15 mm in end-diastole in the absence of a cardiac or systemic causecapable of producing the same degree of hypertrophy or a maximalwall thickness of 13–15 mm in the context of ECG repolarizationanomalies and a family history of HCM in a first-degree relative or posi-tive genotype, in the absence of a cardiac or systemic cause capable ofproducing the same degree of hypertrophy.19 In the 26 athletes withHCM, the ECG was obtained during pre-participation screening andled to a diagnosis at a time when they were still actively involved incompetitive sporting activity.

Arrhythmogenic right ventricular cardiomyopathypatientsAll ARVC patients had an established diagnosis of ARVC in accordancewith the 2010 Revised Task Force diagnostic criteria and fulfilled twomajor criteria, one major criterion, and two minor criteria or four minorcriteria, as reported elsewhere.20 Patients with a diagnosis of ‘border-line’ or ‘possible’ ARVC were excluded from the study. As with athleteswith HCM, the ECG in 9 athletes with ARVC was obtained at the time ofpre-participation screening.

ElectrocardiographyElectrocardiography was performed using conventional methods, withthe subject in supine position and acquisition taken during quite respir-ation with a recording speed of 25 mm/s and 10 mm/1 mV calibration.For digital analysis, hard copies of the ECGs were scanned on a flatbedscanner at a resolution of 300 dpi. Electrocardiogram measurementwith digital callipers and interpretation was performed independentlyby two observers (C.C. and A.Z.) who were blinded to the disease sta-tus. In cases of disagreement, consensus was obtained from a third ob-server who is a highly qualified electrophysiologist and an expert incardiomyopathies (D.C.).

C. Calore et al.Page 2 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

Standard measurements included heart rate; rhythm; PR interval;QRS axis, voltage, and duration; ST-T-wave abnormalities and patho-logical Q-waves. Electrocardiogram abnormalities were defined as fol-lows: (i) voltage criteria for LV hypertrophy according to theSokolow-Lyon index (SV1 + RV5 or RV6, whichever is larger,.35 mm); (ii) pathologic Q-waves as .0.04 s in duration and/or.25% of the ensuing R-wave in depth in at least two leads (except III,aVR, and V1); (iii) pathological ST-segment depression as ST-depressionof .1 mm in depth in at least two adjacent leads; (iv) abnormal TWI asTWI ≥1 mm in at least two adjacent leads (except aVR, V1, and III), inthe absence of conduction disturbance; (v) max TWI ≥5 mm as max-imum amplitude of negative T-wave ≥5 mm in any lead (except aVR,V1, and III); (vi) left atrial enlargement as a negative portion of theP-wave in lead V1 ≥1 mm in depth and ≥0.04 s in duration; (vii) com-plete right bundle branch block (RBBB) as a RSR’ pattern in V1 whereR’ . R and a slurred S-wave in leads I and V6 with a QRS duration of≥0.12 s; (viii) complete LBBB as a QS or rS complex in lead V1 and aRsR’ wave in lead V6 with QRS duration of ≥0.12 s; (ix) frontal-planeleft QRS axis deviation as 2308 to 2908.

The following quantitative ECG parameters of ventricular repolar-ization were measured: (i) the extent of TWI across the precordialleads; (ii) the maximum J-point elevation in the anterior leads (V1–V4) exhibiting TWI; (iii) the maximum ST-segment elevation in the an-terior leads exhibiting TWI; (iv) the maximum depth of anteriorT-wave negativity under the isoelectric line; and (v) the maximum cor-rected QT interval (QTc) calculated using Bazett’s formula. As previ-ously described in detail,21 when the end of the QRS complex wasdifficult to define because of a gradual slope towards the ST-segment,the J-point was extrapolated as the point at which the ST-segment in-tersected the tangent of the slope of QRS terminal part extendedupward.

Statistical analysisContinuous variables were expressed as mean and standard deviationor median and inter-quartile range (IQR) for normally and non-normally distributed variables, respectively. Group differences werecompared using the Student’ t-test or Wilcoxon rank sum test for nor-mally and non-normally distributed variables, respectively. Categoricalvariables were expressed as absolute numbers (n) and percentages(%), and compared with the x2 test or the Fisher exact test, as appro-priate. Univariate and multiple binary logistic regression analysis wasused to determine which of the following variables predicted a cardio-myopathy: (i) J-point elevation ≤1 mm in at least one anterior lead(V1–V4) exhibiting TWI; (ii) ST-segment elevation ≤1 mm in at leastone anterior lead exhibiting TWI; (iii) Max TWI ≥5 mm; and (iv) TWIextending beyond V4 to the lateral (V5–V6, I, aVL) and/or also involv-ing the inferior (II, III, aVF) ECG leads. Variables with a P-value of,0.15 at univariate regression analysis were selected for multiple re-gression analysis. Two different models were built: one unadjusted andone adjusted for age, gender, and ethnicity. Positive and negative pre-dictive values for anterior TWI in the setting of pre-participationscreening were calculated according to the estimated prevalence ofHCM and ARVC among athletes with anterior TWI. The prevalenceof HCM in athletes is 0.06%,22 but only 2% of athletes with HCMshow anterior TWI;16 hence, the prevalence of athletes with HCMwho have anterior TWI is 0.06 × 2% ¼ 0.0012%. In our populationof mixed ethnicity, 8%16 of athletes show anterior TWI; therefore,the estimated number of athletes with HCM and anterior TWI was0.0012/8% ¼ 0.015%. The prevalence of ARVC in screened athletesis 0.03%.23 The prevalence of anterior TWI among individuals withARVC is 50%;7 hence, the number of athletes with ARVC who exhibitanterior TWI is 0.03 × 50% ¼ 0.015%. In our population of mixed

ethnicity, 8%16 of athletes show anterior TWI; therefore, the esti-mated number of athletes with ARVC and anterior TWI was 0.015/8% ¼ 0.2%. Ninety-five per cent confidence interval (CI) of sensitivity,specificity, and negative and predictive values were calculated basedon the binomial distribution. All P-values reported are two-sided,and ,0.05 was considered statistically significant. Data were analysedwith SPSS 17 (SPSS Inc., Chicago, IL, USA).

Results

Demographics and baseline clinicalcharacteristicsHealthy athletesThe median age of athletes was 21 years (IQR 16–27, range 14–37).The majority (75%) were male and of black ethnicity (66%) (see Sup-plementary material online, Table S1). Seventy-seven athletes (98%)were engaged in sport disciplines with moderate or high static anddynamic components (IIB, IIC, IIIB, and IIIC). All athletes trained forat least 6 h per week and competed at regional, national, or inter-national level. None reported symptoms indicative of cardiovasculardisease or showed uncommon/non-training-related ECG abnormal-ities, except for TWI. At echocardiography, the median septum wallthickness was 11 mm (IQR 9–12), the LV end-diastolic diameter51 mm (IQR 40–57), and the posterior LV wall thickness 10 mm(IQR 8–11). The mean LV mass was 183+ 65 g, and the meanLV mass index was 97+ 38 g/m2. Mild RV dilation in the absenceof regional or global systolic dysfunction was observed in 63(79%) athletes.

Hypertrophic cardiomyopathy patientsHypertrophic cardiomyopathy patients had a median age of 46 years(IQR 29–56 years, range 14–62). The majority were males (75%)and of white ethnicity (56%). At echocardiography, the septal wallthickness was 19 mm (IQR 17–33), LV end-diastolic diameter was44 mm (35–52), and the posterior LV wall thickness was 12 mm(IQR 11–16). The median LV mass was 222 g (IQR 154–298),and the median LV mass index was 123 g/m2 (91–152).

Arrhythmogenic right ventricular cardiomyopathypatientsArrhythmogenic right ventricular cardiomyopathy patients hada median age of 32 years (IQR 19–42 years, range 13–53). Themajority were males (71%) and of white ethnicity (95%). Atechocardiography, median RV end-diastolic area was 29 cm2 (IQR26–35 cm2), and median RV fractional area change was 32% (IQR28–36%). Regional RV wall motion abnormalities were found in53 patients (91%). Eleven patients (19%) showed LV involvement,with regional or global LV systolic dysfunction.

Comparison of electrocardiogramfindings between healthy athletes andhypertrophic cardiomyopathy patientsElectrocardiogram findings in healthy athletes and HCM patients areshown in Table 1. Compared with HCM patients, athletes more fre-quently exhibited J-point elevation ≥1 mm in at least one anteriorlead with a negative T-wave (80 vs. 8%, P , 0.001), ST-segment

Anterior T-wave inversion in athlete’s heart vs. cardiomyopathy Page 3 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

elevation ≥1 mm in at least one anterior lead with a negativeT-wave (83 vs. 40%, P , 0.001), and TWI confined to leads V1–V4 (71 vs. 3%, P , 0.001) (Figure 1). In contrast, HCM patientsmore frequently exhibited max TWI ≥5 mm (57 vs. 19%, P ,

0.001) and TWI extending to the lateral (V5–V6, I, aVL) and/orinvolving the inferior (II, III, aVF) leads (Figure 2).

At multiple regression analysis, after adjustment for age, gender,and ethnicity, J-point elevation ,1 mm in leads containing negativeT-waves (OR ¼ 227, 95% CI ¼ 12–1620, P , 0.001), TWI beyondlead V4 (OR ¼ 331; 95% CI ¼ 20–2752, P ¼ 0.001), and max TWI≥5 mm (OR ¼ 10.95% CI ¼ 1.3–100, P ¼ 0.03) remained inde-pendent predictors for HCM (Table 2).

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1 Comparison of ECG findings between healthy athletes and patients with hypertrophic cardiomyopathy bothwith anterior T-wave inversion

HCM, n 5 95 Healthy athletes, n 5 80 P

Age 46 (29–56) 21 (16–27) ,0.001

Gender (male) 71 (75) 60 (75) 1.0

Ethnicity (black) 42 (44) 53 (66) 0.004

Sinus rhythm 94 (99) 80 (100) 1.0

Heart rate (b.p.m.) 62 (56–67) 57 (51–67) 0.01

PR interval duration (ms) 162 (146–180) 160 (138–176) 0.48

QRS interval

Duration (ms) 94 (84–104) 88 (80–100) 0.07

Sokolow-Lyon criteria for LVH (mm) 35 (25–48) 35 (29–42) 0.88

Cornell criteria for LVH (mm) 19 (13–28) 17 (13–24) 0.13

Abnormal Q-wave 15 (16) 0 ,0.001

Abnormal axis 7 (7) 0 0.04

J pointa

Elevation ≥1 mm 8 (8) 64 (80) ,0.001

Number of leads with elevation ≥1 mm 0 (0–0) 1 (1–2) ,0.001

Max elevation (mm) 0 (0–0) 1 (1–1.5) ,0.001

Elevation ≥2 mm 3 (3) 15 (19) 0.001

ST-segmenta

ST-segment elevation ≥1 mm 38 (40) 66 (83) ,0.001

Number of leads with ST elevation ≥1 mm 0 (0–1) 2 (1–2) ,0.001

Max ST-segment elev. (mm) 0.5 (0–1) 1 (1–2) ,0.001

Elevation ≥2 mm 12 (13) 29 (36) ,0.001

Negative T-waves

TWI in V1–V4 onlyb 3 (3) 57 (71) ,0.001

Distribution

V1 and V2 19 (20) 75 (94) ,0.001

V3 and V4 93 (98) 30 (38) ,0.001

V5 and V6 85 (90) 7 (9) ,0.001

D1/aVL 66 (70) 9 (11) ,0.001

DII/aVF/DIII 47 (50) 11 (14) ,0.001

Number of leads with TWI 8 (7–9) 3 (3–4) ,0.001

Max depth (mm) 5 (3.5–9) 3 (2–4.5) ,0.001

Max TWI ≥5 mm 54 (57) 15 (19) ,0.001

QT interval

QTc (ms) 435 (416–453) 402 (388–416) ,0.001

QTc .470 ms 5 (5) 0 0.01

Variables are presented as numbers (%) or median (1st–3rd quartiles).HCM, hypertophic cardiomyopathy; LVH, left ventricular hypertrophy; TWI, T-wave inversion.aIn ≥1 lead V1–V4 with negative T-wave.bTWI in ≥2 contiguous leads V1–V4, but not extending beyond V4 to the lateral (V5–V6, I, aVL) and/or involving the inferior (II, III, aVF) electrocardiogram leads.

C. Calore et al.Page 4 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

Comparison of electrocardiogramfindings between healthy athletes andarrhythmogenic right ventricularcardiomyopathy patientsElectrocardiogram findings in healthy athletes and ARVC patientsare shown in Table 3. Compared with ARVC patients, athletesmore frequently exhibited J-point elevation ≥1 mm in at least oneanterior lead with a negative T-wave (80 vs. 2%, P , 0.001),ST-segment elevation ≥1 mm in at least one anterior lead with anegative T-wave (83 vs. 38%, P , 0.001), TWI confined to leadsV1–V4 (71 vs. 45%, P , 0.001), and max TWI ≥5 mm (19 vs. 5%,P , 0.001). In contrast, ARVC patients more frequently exhibitedTWI extending to the lateral and/or involving inferior leads(Figure 3).

At multiple regression analysis, after adjustment for age, gen-der, and ethnicity, J-point elevation ,1 mm in leads with nega-tive T-waves (OR ¼ 569, 95% CI ¼ 38–8545, P , 0.001) and

TWI extending beyond lead V4 (OR ¼ 6.0; 95% CI ¼ 1.2 –37.8, P ¼ 0.03) remained independent predictors of ARVC(Table 4).

Electrocardiogram of athletes withcardiomyopathyThe ECG characteristics of the subgroups of patients who were di-agnosed with either HCM or ARVC at pre-participation screening(affected athletes) are shown in Supplementary material online, Ta-ble S2 and S3. None of the 26 athletes with HCM (n ¼ 22 male; n ¼22 white) revealed TWI confined to V1–V4. J-point elevation≥1 mm in at least one anterior lead containing a negative T-wavewas present in 3 of 22 (14%) white and in 3 of 4 (75%) black athleteswith HCM. None of the 9 athletes with ARVC (all white males) re-vealed J-point elevation ≥1 mm in at least one anterior lead contain-ing a negative T-wave, while 7 of 9 (78%) demonstrated negativeT-waves confined to V1–V4.

Figure 1 Representative ECG tracings of healthy athletes with anterior T-wave inversion. Top: Anterior T-wave inversion confined to V1–V4 ina 21-year-old white athlete. Note that negative T-waves in V3–V4 are preceded by J-point elevation and ST-segment elevation. Bottom: AnteriorT-waves inversion (V1–V4) extending to infero-lateral leads in a 19-year-old black athlete. Negative T-waves in V2–V4 are preceded by J-pointelevation and ST-segment elevation.

Anterior T-wave inversion in athlete’s heart vs. cardiomyopathy Page 5 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

Figure 2 Representative electrocardiogram tracings of patients with hypertrophic cardiomyopathy showing anterior T-wave inversion. Top:Anterior T-wave inversion extending beyond V4 in a 32-year-old white hypertrophic cardiomyopathy patient. Negative T-waves in V2–V4 arepreceded by an upward ST-segment (maximum displacement ¼ 2 mm) without J-point elevation. Bottom: Anterior T-wave inversion confined toV1–V3 in a 38-year-old black HCM patient. Negative T-waves in V2–V4 are preceded by an upward ST-segment (maximum displacement ¼ 1.5mm) without J-point elevation.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 2 Univariate and multiple regression analysis for predictors of hypertrophic cardiomyopathy among individualswith inverted T-waves in the anterior precordial leads V1–V4

Univariate model Unadjusted multiple model (r2 5 0.89) Adjusted multiple model for age, gender,and ethnicity (r2 5 0.95)

OR (95% CI) P OR (95% CI) P OR (95% CI) P

J-point ,1 mma 44 (18–108) ,0.001 J-point ,1 mma 260 (24–2840) ,0.001 J-point ,1 mma 227 (12–1620) ,0.001

ST ,1 mma 3.6 (1.9–6.9) ,0.001 ST ,1 mma 2.3 (0.2–25) 0.40 ST ,1 mma 8.5 (0.2–33) 0.15

TWI beyond V4b 75 (22–262) ,0.001 TWI beyond V4b 330 (30–3613) ,0.001 TWI beyond V4b 331 (20–2752) 0.001

Max TWI ≥5 mm 31 (2.4–11) ,0.001 Max TWI ≥5 mm 6.3 (1.4–25) 0.014 Max TWI ≥5 mm 10 (1.3–100) 0.03

CI, confidence interval; OR, odds ratio; TWI, T-wave inversion.aIn at least one anterior (V1-V4) lead with TWI.bExtending to the lateral (V5–V6, I, aVL) and/or involving the inferior (II, III, aVF) electrocardiogram leads.

C. Calore et al.Page 6 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

Electrocardiogram diagnostic accuracyfor cardiomyopathyComparing healthy athletes with both patients and athletes withcardiomyopathy (either HCM or ARVC), J-point elevation,1 mm or TWI extending beyond V4 showed an overall sensitivityfor cardiomyopathy of 100%, a specificity for athlete’s heart of 55%,

a positive predictive value of 0.03 and 0.42% for HCM and ARVC,respectively, and a negative predictive value of 100% for bothHCM and ARVC (Table 5, see Supplementary material online, TableS4). In a sub-analysis of ECG findings between healthy athletes andathletes affected by cardiomyopathy, the sensitivity and negativepredictive value of combined ECG criteria for both HCM andARVC were 100% (see Supplementary material online, Table S5).

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 3 Comparison of electrocardiogram findings between healthy athletes and patients with arrhythmogenic rightventricular cardiomyopathy, both with anterior T-wave inversion

ARVC, n 5 58 Healthy athletes, n 5 80 P

Age 32 (19–42) 21 (16–27) 0.01

Gender (male) 41 (71) 60 (75) 0.58

Ethnicity (black) 3 (5) 53 (66) ,0.001

Sinus rhythm 58 (100) 80 (100) —

Heart rate (b.p.m.) 60 (53–70) 57 (51–67) 0.24

PR interval duration (ms) 168 (152–187) 160 (138–176) 0.03

QRS interval

Duration (ms) 99 (87–113) 88 (80–100) 0.005

Sokolow-Lyon criteria for LVH (mm) 17 (12–22) 35 (29–42) ,0.001

Cornell criteria for LVH (mm) 8 (5–12) 17 (13–24) ,0.001

Abnormal Q-wave 0 0 —

Abnormal axis 11 (23) 0 ,0.001

J-pointa

Elevation ≥1 mm 1 (2) 64 (80) ,0.001

Number of leads with elevation ≥1 mm 0 (0–0) 1 (1–2) ,0.001

Max elevation (mm) 0 (0–0.5) 1 (1–1.5) ,0.001

Elevation ≥2 mm 0 15 (19) ,0.001

ST-segmenta

Elevation ≥1 mm 22 (38) 66 (83) ,0.001

Number of leads with elevation ≥1 mm 0 (0–1) 2 (1–2) ,0.001

Max elevation (mm) 1 (0–1) 1 (1–2) ,0.001

Elevation ≥2 mm 1 (36) 29 (36) ,0.001

Negative T-waves

TWI in V1–V4 onlyb 26 (45) 57 (71) ,0.001

Distribution

V1 and V2 58 (100) 75 (94) 0.07

V3 and V4 36 (62) 30 (38) ,0.001

V5 and V6 13 (22) 7 (9) 0.02

D1/aVL 7 (12) 9 (11) 0.77

DII/aVF/DIII 20 (35) 11 (14) 0.004

Number of leads with TWI 5 (4–7) 3 (3–4) ,0.001

Max depth (mm) 2.5 (2–3.5) 3 (2–4.5) 0.10

Max TWI ≥5 mm 3 (5) 15 (19) 0.02

QT interval

QTc (ms) 424 (406–451) 402 (388–416) 0.001

QTc .470 ms 4 (1) 0 0.01

Variables are presented as numbers (%) or median (1st–3rd quartiles).ARVC, arrhythmogenic right ventricular cardiomyopathy; LVH, left ventricular hypertrophy; TWI, T-wave inversion.aIn ≥1 lead V1–V4 with negative T-wave.bTWI in ≥2 contiguous leads V1–V4, but not extending beyond V4 to the lateral (V5–V6, I, aVL) and/or involving the inferior (II, III, aVF) electrocardiogram leads.

Anterior T-wave inversion in athlete’s heart vs. cardiomyopathy Page 7 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

Compared with J-point elevation ,1 mm, ST-segment elevation,1 mm in combination with TWI extending beyond V4 showed alower sensitivity and negative predictive value for cardiomyopathywhile maintaining a similar specificity and positive predictive valuefor athlete’s heart (Table 5 and Figure 4).

Inter-observer variability inelectrocardiogram interpretationThere was excellent inter-observer agreement in the evaluation ofJ-point elevation ,1 mm and TWI extending beyond V4 translating

Figure 3 Representative electrocardiogram tracings of patients with arrhythmogenic right ventricular cardiomyopathy showing T-wave inver-sion in anterior leads. Top: Anterior T-wave inversion in V1–V4 in a 19-year-old white arrhythmogenic right ventricular cardiomyopathy patient.Negative T-wave in V3 is preceded by an upward ST-segment (maximum displacement ¼ 1 mm) without J-point elevation. Bottom: anteriorT-wave inversion in V1–V3 without J–ST elevation in a 35-year-old white arrhythmogenic right ventricular cardiomyopathy patient.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 4 Univariate and multiple regression analysis for predictors of arrhythmogenic cardiomyopathy amongindividuals with T-wave inversion in the anterior precordial leads V1–V4

Univariate model Unadjusted multiple model (r2 5 0.89) Adjusted multiple model for age, gender, andethnicity (r2 5 0.95)

OR (95% CI) P OR (95% CI) P OR (95% CI) P

J-point ,1 mma 228 (29–1773) ,0.001 J-point ,1 mma 401 (38–4273) ,0.001 J-point ,1 mma 569 (38–8545) ,0.001

ST elev ,1 mma 4.1 (2.0–8.3) ,0.001 ST ,1 mma 1.1 (0.3–3.9) 0.89 ST ,1 mma 0.92 (0.17–5.0) 0.82

TWI beyond V4b 3.1 (1.5–6.2) 0.002 TWI beyond V4b 9.6 (2.0–46) 0.005 TWI beyond V4b 6.0 (1.2–37.8) 0.03

Max TWI ≥5 mm 4.2 (1.2–15.4) 0.03 Max TWI ≥5 mm 2.9 (0.3–24) 0.33 Max TWI ≥25 mm 1.9 (0.09–37.9) 0.69

CI, confidence interval; OR, odds ratio; TWI, T-wave inversion.aIn at least one anterior (V1–V4) lead with TWI.bExtending to the lateral (V5–V6, I, aVL) and/or involving the inferior (II, III, aVF) electrocardiogram leads.

C. Calore et al.Page 8 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 5 Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) with 95% confidence intervals of ST-T electrocardiographicmarkers for cardiomyopathies

Specificity(%)

Sensitivity for HCM(%)

Sensitivity for ARVC(%)

PPV for HCM(%)

NPV for HCM(%)

PPV for ARVC(%)

NPV for ARVC(%)

ST-segment elevation ,1 mma 83 (72–90) 60 (49–70) 62 (48–74) 0.05 (0–0.33) 99.99 (99.91–100) 0.68 (0–1.56) 99.91 (99.50–100)

J-point ,1 mma 80(70–88) 92 (84–96) 98 (91–100) 0.07 (0–0.33) 99.99 (99.95–100) 0.91(0–2.02) 99.99(99.91–100)

TWI beyond V4b 71 (60–81) 97 (92–100) 55 (42–68) 0.05 (0–0.26) 99.99 (99.97–100) 0.36 (0–1.16) 99.88 (99.5–100)

ST-segment elevation ,1 mma or TWIbeyond V4b

58 (46–68) 97 (91–99) 85 (73–93) 0.04 (0–0.20) 99.99 (99.96–100) 0.38 (0–1.05) 99.95 (99.65–100)

J-point elevation ,1 mma or TWI beyondV4b

55 (43–66) 100 (96–100) 100 (94–100) 0.03 (0–0.19) 100 0.42 (0–1.08) 100

TWI, T-wave inversion; HCM, hypertophic cardiomyopathy; ARVC, arrhythmogenic right ventricular cardiomyopathy.aIn at least one anterior (V1–V4) lead with TWI.bExtending to the lateral (V5–V6, I, aVL) and/or involving the inferior (II, III, aVF) electrocardiogram leads.

Figu

re4

Close-up

ofleadV

3focusing

onJ-point

fromelectro-

cardiogramtracings

reportedin

Figures1–

3.Top:N

otethe

J-pointelevation

precedingT

-wave

inversionin

leadV

3(both

panelsfrom

healthyathletes);M

iddle:Note

theST

-segmentelevation

(without

J-pointelevation)preceding

T-w

aveinversion

inlead

V3

(bothpa-

nelsfrom

patientsw

ithhypertro

phiccardio

myo

pathy);Bottom

:N

oteST

-segment

elevation(w

ithoutJ-point

elevation)preceding

T-w

aveinversion

inone

patientw

itharrhythm

ogenicright

ven-tricular

cardiomyopathy

(leftpanel)andT

-wave

inversionw

ithoutJ-point

orST

-segment

elevationin

anotherpatient

with

arrhyth-m

ogenicrightventricularcardiom

yopathy(rightpanel).A

rrowsin-

dicatethe

J-point(i.e.the

QR

S–

ST-segm

entjunction).

Anterior

T-w

aveinversion

inathlete’s

heartvs.cardiom

yopathyP

age9

of13

by guest on November 18, 2015 http://eurheartj.oxfordjournals.org/ Downloaded from

to a k (measurement of agreement) of 0.97 (P , 0.001) and 1.0 (P ,

0.001), respectively.

DiscussionThe present study demonstrates that the combined evaluation ofJ-point elevation and TWI confined to lead V1–V4 allows more ac-curate differentiation between ‘physiologic’ and ‘cardiomyopathic’TWI among individuals of both white and black ethnicity.

Differential diagnosisBecause TWI is one of the most striking ECG markers for HCM andARVC, the presence of this repolarization abnormality on an ath-lete’s ECG is considered a ‘red-flag’ sign requiring comprehensiveevaluation.1– 3,24,25 Often, this process entails expensive clinical in-vestigation using a combination of imaging techniques, exercisetests, 24-h Holter monitoring, and, if feasible, genetic testing to ex-clude an underlying cardiomyopathy. However, TWI in V1–V4 isalso observed on the ECG of healthy athletes, particular those ofAfro-Caribbean descent.4,10 The present study identified additionalECG parameters that can be used to allow discrimination betweenphysiological (athlete’s heart) and pathological (either HCM orARVC) anterior TWI, independent of ethnicity.

Hypertrophic cardiomyopathyA previous study comparing the prevalence of anterior TWI in blackathletes and black patients with HCM suggested that TWI limited toleads V1–V4 and associated with an upward ST-segment is likely torepresent an ethnic variant of ‘athlete’s heart’.4 The present studyconfirmed these findings demonstrating that, unlike healthy athletes,both patients and athletes with HCM usually exhibit TWI extendingbeyond V4. However, our study did not confirm the discriminatingvalue of an upward ST-segment preceding TWI, given that it wasdetected in 40% of the overall HCM population compared with83% of trained healthy athletes. Instead, we observed that theJ-point amplitude preceding TWI provided the best accuracy for dis-criminating electrical manifestations of physiological adaptationfrom pathological anterior TWI: anterior TWI without J-point ele-vation (i.e. J point ,1 mm) showed a 92% sensitivity for HCM com-pared with 60% of TWI without ST-segment elevation, despite asimilar specificity.

Arrhythmogenic right ventricular cardiomyopathyT-wave inversion in V1–V4 is the most striking and frequent ECGmarker of ARVC, which is the leading cause of SCD in white athletesin some parts of the world.5 – 8,23,26 Our study demonstrated thatthe extent and distribution of TWI across all precordial leads haslimited value for distinguishing athlete’s heart from ARVC. In fact,

Figure 5 Proposed flow chart for the interpretation of anterior (V1–V4) T-wave inversion on an athletes’ electrocardiogram. * Extending tothe lateral (V5–V6, I, aVL) and/or involving the inferior (II, III, aVF) ECG leads; #In at least one anterior (V1–V4) lead showing T-wave inversion.

C. Calore et al.Page 10 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

TWI extending beyond V4 showed a sensitivity for ARVC of only55% in the overall cardiomyopathy cohort and of only 22% in thesubgroup of athletes with ARVC. An upward ST-segment precedingTWI also appeared of limited value for differential diagnosis. Add-itionally, the current study reveals that J-point elevation precedingTWI was rarely observed in the total ARVC cohort and never foundin the subgroup of athletes with ARVC; indeed, the absence ofJ-point elevation equated to a disease sensitivity of 98 and 100%in these two groups, respectively.

Hypertrophic cardiomyopathy and arrhythmogenic rightventricular cardiomyopathyAlthough no single ECG criterion was accurate enough to differentiateathlete’s heart from either HCM or ARVC, the combined analysis ofthe distribution of TWI across the precordial leads and the amplitudeof J-point elevation preceding anterior TWI allowed identification of acardiomyopathy in all affected individuals, regardless of the ethnicity.None of the patients or athletes with cardiomyopathy, either blackor white, revealed both TWI limited to V4 and J-point elevation pre-ceding TWI (Figure 5). In comparison, the combined analysis of the ex-tent of TWI across the precordial leads and ST-segment elevationpreceding anterior TWI (i.e. the pattern utilized by the Seattle Cri-teria2) showed a lower overall sensitivity for cardiomyopathies pre-dominantly due to ARVC rather than to HCM, as 15% of ARVCpatients revealed anterior TWI confined to V1–V4 preceded byST-segment (but without J point) elevation (Table 5).

The main goal of ECG interpretation in the context of pre-participation evaluation is to avoid any false-negatives, and in this re-gard, our ECG algorithm provided a 100% negative predictive value,which means 0% false-negative results. It is well known that the posi-tive predictive value of broad-based ECG screening is very low be-cause of the low prevalence of cardiomyopathy among athletes withECG abnormalities.1 In detail, the current practice of screening allathletes with anterior TWI translates into an estimated positive pre-dictive value of 1 : 6667 for HCM and 1 : 500 for ARVC. Using ouralgorithm based on the combination of J-point amplitude and thedistribution of TWI in athletes with anterior TWI, the overall speci-ficity for athlete’s heart was 55%, which implies that the number offalse-positives (i.e. healthy athletes with anterior TWI) is reducedapproximately by 50%. Accordingly, our algorithm allowed to im-prove the positive predictive values for cardiomyopathies of theECG pattern of TWI in the athlete, by halving the number of athleteswith anterior TWI who require further evaluation to detect 1 car-diomyopathy (from 6667 to 3333 for HCM and from 500 to 238 forARVC), which is expected to result in a significant screening costsavings (see Supplementary material online, Tables S2–S4).

PathophysiologyJ-point elevationTrained athletes commonly reveal the early repolarization patternthat arises from a physiologic increase of vagal tone and/or with-drawal of sympathetic activity.27 – 29 In our study, the vast majorityof anterior TWI observed among healthy athletes of both ethnicitieswas preceded by J-point elevation, which suggested an underlyingearly repolarization mechanism (Figure 4). This explains why thepresence of J-point elevation preceding negative T-waves was themost accurate ECG marker for ‘physiological’ anterior TWI. Indeed,

this marker was never observed among non-athlete patients withHCM or ARVC, although it was occasionally found among athleteswith HCM, suggesting that signs of training-related ‘physiological’early repolarization may co-exist with ‘pathological’ TWI causedby the underlying cardiomyopathy.

ST-segment elevationWe observed ST-segment elevation without J-point elevation in asizeable proportion of patients with HCM and ARVC. This findingis in agreement with the results of previous studies showing a similarprevalence of ST-segment elevation in association with anteriorTWI in these cardiomyopathies.4,7,30 In addition, Sheikh et al.16 re-ported a 63% prevalence of ST-segment elevation in highly trainedathletes with HCM. Elevation of the ST-segment in these heart mus-cle diseases does not occur in the context of a physiological earlyrepolarization but instead reflects underlying myocardial abnormal-ities such as ischaemic myocardial injury, myocardial degenerationand necrosis, or fibro-fatty myocardial replacement that may inter-fere with the normal transmission of electrical forces across the ven-tricular walls, resulting in ST-T repolarization abnormalities.

Extent of T-wave inversionThe limited overall accuracy of the extent of TWI alone for differ-entiating between athlete’s heart and a cardiomyopathy is explainedby the different ventricular involvement in HCM and ARVC. WhileHCM is predominately a ‘LV’ cardiomyopathy which most often ex-hibits TWI localized or extending to the lateral (V5–V6) and/or in-ferior LV leads (II/aVF/III),19,30 ARVC in its typical form ischaracterized by TWI confined to the RV leads (V1–V3/V4).7

Accordingly, the criterion of TWI limited to V1–V4 is useful in dis-tinguishing anterior TWI representative of athlete’s heart fromthe more diffuse TWI of HCM, but is less helpful for differentiatingathlete’s heart from ARVC.

Study limitationsBecause of the disparities of age and gender between athletes andpatients with cardiomyopathies, it remains to be proved whetherour discriminating ECG algorithm would provide the same accuracyin differentiating athletes and patients with cardiomyopathies of thesame age and gender. However, the sub-analysis of healthy youngathletes (median age 21 years) vs. young athletes with cardiomyop-athies (median age 23 years) confirmed the discriminating power ofour ECG predictors. Furthermore, we cannot certain whetherother confounding factors (unknown or difficult to adjust, includingthe different sources of HCM and ARVC patients/athletes) may haveinfluenced the observed association between TWI patterns anddiagnosis. The predominantly black athletic population of our studymay not represent the ethnic composition of the general athleticpopulation encountered in the European countries. However, ourstudy focused on the subset of athletes who commonly exhibit an-terior TWI, notably the black athletes.12 Although athletes weresubject to a systematic and comprehensive array of investigationsto exclude an underlying cardiomyopathy, based on this cross sec-tional study, we cannot be certain about the possibility of such indi-viduals developing overt features of a cardiomyopathy at a laterdate.24,31 Finally, our excellent sensitivity of ECG repolarization pat-terns for diagnosis of cardiomyopathies should not be generalized

Anterior T-wave inversion in athlete’s heart vs. cardiomyopathy Page 11 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

to the entire population of patients with HCM or ARVC, butaccording to the study design is only applicable to the subset ofpatients with anterior TWI.

ConclusionsThe present study demonstrates that electrocardiographic analysisof J-point amplitude and distribution of TWI across the precordialleads provides a highly effective method for differentiating anteriorTWI reflecting adaptive early repolarization representative of theathlete’s heart from pathological anterior TWI in individuals withHCM or ARVC, regardless of ethnicity (Figure 5). Considerationthat anterior TWI preceded by J-point elevation and TWI limitedto V1–V4 is a normal variant in athletes provides excellent sensitiv-ity for detecting athletes harbouring HCM or ARVC. These com-bined criteria are associated with a halving of the number offalse-positive results among athletes with anterior TWI and, thus,reduce the number of healthy athletes with such an ECG patternbeing subject to further unnecessary and expensive clinical investi-gation to exclude an underlying cardiomyopathy.

According to our results, the perspective that anterior TWI with-out J-point elevation and/or extending beyond V4 is due to cardio-vascular adaptation to physical exercise can only be accepted onceany inherited cardiomyopathy has been excluded by a comprehen-sive clinical work-up, including cardiac magnetic resonance.Conversely, the presence of concomitant pathologic ECG abnor-malities (such as pathological Q-waves or epsilon waves) is an indi-cation for further investigations to exclude a cardiomyopathy also inpatients with J-point elevation preceding anterior TWI.1,16,32

The most significant implication of the present study is that thecurrent Seattle recommendation2 and future updates on the inter-pretation of the athlete’s ECG may be modified on the basis of ourfindings, provided that such data are replicated by other major cen-tres, because of the relatively small cohort of our study patients andathletes with HCM and ARVC.

Supplementary materialSupplementary material is available at European Heart Journal online.

FundingThis work was supported by the TRANSAC research Grant of the Uni-versity of Padua, Italy; and the Cardiac Risk in the Young, London, UK.

Conflict of interest: none declared.

References1. Corrado D, Pelliccia A, Heidbuchel H, Sharma S, Link M, Basso C, Biffi A, Buja G,

Delise P, Gussac I, Anastasakis A, Borjesson M, Bjornstad HH, Carre F,Deligiannis A, Dugmore D, Fagard R, Hoogsteen J, Mellwig KP,Panhuyzen-Goedkoop N, Solberg E, Vanhees L, Drezner J, Estes NA III, Iliceto S,Maron BJ, Peidro R, Schwartz PJ, Stein R, Thiene G, Zeppilli P, McKenna WJ. Re-commendations for interpretation of 12-lead electrocardiogram in the athlete.Eur Heart J 2010;31:243–259.

2. Drezner JA, Fischbach P, Froelicher V, Marek J, Pelliccia A, Prutkin JM, Schmied CM,Sharma S, Wilson MG, Ackerman MJ, Anderson J, Ashley E, Asplund CA,Baggish AL, Borjesson M, Cannon BC, Corrado D, DiFiori JP, Harmon KG,Heidbuchel H, Owens DS, Paul S, Salerno JC, Stein R, Vetter VL. Normal electro-cardiographic findings: recognising physiological adaptations in athletes. Br J SportsMed 2013;47:125–136.

3. Uberoi A, Stein R, Perez MV, Freeman J, Wheeler M, Dewey F, Peidro R, Hadley D,Drezner J, Sharma S, Pelliccia A, Corrado D, Niebauer J, Estes NA III, Ashley E,Froelicher V. Interpretation of the electrocardiogram of young athletes. Circulation2011;124:746–757.

4. Papadakis M, Carre F, Kervio G, Rawlins J, Panoulas VF, Chandra N, Basavarajaiah S,Carby L, Fonseca T, Sharma S. The prevalence, distribution, and clinical outcomesof electrocardiographic repolarization patterns in male athletes of African/Afro-Caribbean origin. Eur Heart J 2011;32:2304–2313.

5. Bhonsale A, James CA, Tichnell C, Murray B, Madhavan S, Philips B, Russell SD,Abraham T, Tandri H, Judge DP, Calkins H. Risk stratification in arrhythmogenicright ventricular dysplasia/cardiomyopathy-associated desmosomal mutation car-riers. Circ Arrhythm Electrophysiol 2013;6:569–578.

6. Migliore F, Zorzi A, Michieli P, Perazzolo Marra M, Siciliano M, Rigato I, Bauce B,Basso C, Toazza D, Schiavon M, Iliceto S, Thiene G, Corrado D. Prevalence of car-diomyopathy in Italian asymptomatic children with electrocardiographic T-wave in-version at preparticipation screening. Circulation 2012;125:529–538.

7. Steriotis AK, Bauce B, Daliento L, Rigato I, Mazzotti E, Folino AF, Marra MP,Brugnaro L, Nava A. Electrocardiographic pattern in arrhythmogenic right ven-tricular cardiomyopathy. Am J Cardiol 2009;103:1302–1308.

8. Zorzi A, Migliore F, Elmaghawry M, Silvano M, Marra MP, Niero A, Nguyen K,Rigato I, Bauce B, Basso C, Thiene G, Iliceto S, Corrado D. Electrocardiographicpredictors of electroanatomic scar size in arrhythmogenic right ventricular cardio-myopathy: implications for arrhythmic risk stratification. J Cardiovasc Electrophysiol2013;24:1321–1327.

9. Le VV, Wheeler MT, Mandic S, Dewey F, Fonda H, Perez M, Sungar G, Garza D,Ashley EA, Matheson G, Froelicher V. Addition of the electrocardiogram to thepreparticipation examination of college athletes. Clin J Sport Med 2010;20:98–105.

10. Basavarajaiah S, Boraita A, White G, Wilson M, Carby L, Shah A, Sharma A. Ethnicdifferences in left ventricular remodeling in highly-trained athletes: relevance to dif-ferentiating physiologic left ventricular hypertrophy from hypertrophic cardiomy-opathy. J Am Coll Cardiol 2008;51:2256–2262.

11. Brosnan M, La Gerche A, Kalman J, Lo W, Fallon K, MacIsaac A, Prior DL. Compari-son of frequency of significant electrocardiographic abnormalities in enduranceversus nonendurance athletes. Am J Cardiol 2014;113:1567–1573.

12. Wasfy MM, DeLuca J, Wang F, Berkstresser B, Ackerman KE, Eisman A, Lewis GD,Hutter AM, Weiner RB, Baggish AL. ECG findings in competitive rowers: normativedata and the prevalence of abnormalities using contemporary screening recom-mendations. Br J Sports Med 2015;49:200–206.

13. Berge HM, Gjesdal K, Andersen TE, Solberg EE, Steine K. Prevalence of abnormalECGs in male soccer players decreases with the Seattle criteria, but is still high.Scand J Med Sci Sports 2015;25:501–508.

14. Bessem B, de Bruijn MC, Nieuwland W. The ECG of high-level junior soccerplayers: comparing the ESC vs. the Seattle criteria. Br J Sports Med 2015;49:1000–1006.

15. Pickham D, Zarafshar S, Sani D, Kumar N, Froelicher V. Comparison of three ECGcriteria for athlete pre-participation screening. J Electrocardiol 2014;47:769–774.

16. Sheikh N, Papadakis M, Ghani S, Zaidi A, Gati S, Adami PE, Carre F, Schnell F,Wilson M, Avila P, McKenna W, Sharma S. Comparison of electrocardiographic cri-teria for the detection of cardiac abnormalities in elite black and white athletes.Circulation 2014;129:1637–1649.

17. Mitchell JH, Haskell WL, Raven PB. Classification of sports. J Am Coll Cardiol 1994;24:864–866.

18. Corrado D, Basso C, Pelliccia A, Thiene G. Sports and heart disease. In: Camm J,Luscher TF, Serruys PW (eds). The ESC Textbook of Cardiovascular Medicine.New York: Oxford University Press; 2009. p1215–1237.

19. Elliott PM, Anastasakis A, Borger MA, Borggrefe M, Cecchi F, Charron P,Hagege AA, Lafont A, Limongelli G, Mahrholdt H, McKenna WJ, Mogensen J,Mihoyannopoulos P, Nistri S, Pieper PG, Pieske B, Rapezzi C, Rutten FH,Tillmanns C, Watkins H. 2014 ESC Guidelines on diagnosis and management ofhypertrophic cardiomyopathy. Eur Heart J 2014;35:2733–2779.

20. Marcus FI, McKenna WJ, Sherrill D, Basso C, Bauce B, Bluemke DA, Calkins H,Corrado D, Cox MG, Daubert JP, Fontaine G, Gear K, Hauer R, Nava A,Picard MH, Protonotarios N, Saffitz JE, Sanborn DM, Steinberg JS, Tandri H,Thiene G, Towbin JA, Tsatsopoulou A, Wichter T, Zareba W. Diagnosis of arrhyth-mogenic right ventricular cardiomyopathy/dysplasia: proposed modification of thetask force criteria. Circulation 2010;121:1533–1541.

21. Zorzi A, Leoni L, Di Paolo FM, Rigato I, Migliore F, Bauce B, Pelliccia A, Corrado D.Differential diagnosis between early repolarization of athlete’s heart and coved-type Brugada electrocardiogram. Am J Cardiol 2015;115:529–532.

22. Basavarajaiah S, Wilson M, Whyte G, Shah A, McKenna W, Sharma S. Prevalence ofhypertrophic cardiomyopathy in highly trained athletes: relevance to pre-participation screening. J Am Coll Cardiol 2008;51:1033–1039.

23. Corrado D, Basso C, Pavei A, Michieli P, Schiavon M, Thiene G. Trends in suddencardiovascular death in young competitive athletes after implementation of a pre-participation screening program. JAMA 2006;296:1593–1601.

C. Calore et al.Page 12 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from

24. Pelliccia A, Di Paolo FM, Quattrini FM, Basso C, Culasso F, Popoli G, De Luca R,Spataro A, Biffi A, Thiene G, Maron BJ. Outcomes in athletes with marked ECGrepolarization abnormalities. N Engl J Med 2008;358:152–161.

25. Schnell F, Riding NA, O’Hanlon R, Axel Lentz P, Donal E, Kervio G, Matelot D,Leurent G, Doutreleau S, Chevalier L, Guerard S, Wilson MG, Carre F. Recognitionand significance of pathological T-wave inversion in athletes. Circulation 2015;131:165–173.

26. Corrado D, Basso C, Rizzoli G, Schiavon M, Thiene G. Does sports activity enhancethe risk of sudden death in adolescents and young adults? J Am Coll Cardiol 2003;42:1959–1963.

27. Quattrini FM, Pelliccia A, Assorgi R, Di Paolo FM, Squeo MR, Culasso F, Castelli V,Link MS, Maron BJ. Benign clinical significance of J-wave pattern (early repolariza-tion) in highly trained athletes. Heart Rhythm 2014;11:1974–1982.

28. Serra-Grima R, Donate M, Alvarez-Garcıa J, Barradas-Pires A, Ferrero A,Carballeira L, Puig T, Rodrıguez E, Cinca J. Long term follow-up of early repolariza-tion pattern in elite athletes. Am J Med 2015;128:192.e1–e9.

29. Muramoto D, Yong CM, Singh N, Aggarwal S, Perez M, Ashley E, Hadley D,Froelicher V. Patterns and prognosis of all components of the J-wavepattern in multiethnic athletes and ambulatory patients. Am Heart J 2014;167:259–266.

30. Delcre SD, Di Donna P, Leuzzi S, Miceli S, Bisi M, Scaglione M, Caponi D,Conte MR, Cecchi F, Olivotto I, Gaita F. Relationship of ECG findings to phenotypicexpression in patients with hypertrophic cardiomyopathy: a cardiac magnetic res-onance study. Int J Cardiol 2013;167:1039–1045.

31. Schnell F, Riding N, O’Hanlon R, Axel Lentz P, Donal E, Kervio G, Matelot D,Leurent G, Doutreleau S, Chevalier L, Guerard S, Wilson MG, Carre F. Recognitionand significance of pathological T-wave inversions in athletes. Circulation 2015;131:165–173.

32. Zaidi A, Sheikh N, Jongman JK, Gati S, Panoulas VF, Carr-White G, Papadakis M,Sharma R, Behr ER, Sharma S. Clinical differentiation between physiological remod-eling and arrhythmogenic right ventricular cardiomyopathy in athletes with markedelectrocardiographic repolarization anomalies. J Am Coll Cardiol 2015;65:2702–2711.

Anterior T-wave inversion in athlete’s heart vs. cardiomyopathy Page 13 of 13

by guest on Novem

ber 18, 2015http://eurheartj.oxfordjournals.org/

Dow

nloaded from