Introduction. CLINICAL RESEARCH Electrocardiology and risk stratification

Transcription

1 Europace (2013) 15, doi: /europace/eus311 CLINICAL RESEARCH Electrocardiology and risk stratification T-wave integral: an electrocardiographic marker discriminating patients with arrhythmogenic right ventricular cardiomyopathy from patients with right ventricular outflow tract tachycardia Alexander Samol 1, Christian Wollmann 1, Christian Vahlhaus 1}, Joachim Gerss 2, Hans-Jürgen Bruns 1,Günter Breithardt 1, Eric Schulze-Bahr 1,3, Thomas Wichter 1, and Matthias Paul 1 * 1 Department of Cardiovascular Medicine, Division of Cardiology, University Hospital Münster, Münster, Germany; 2 Institute for Biostatistics and Clinical Research, University of Münster, Münster, Germany; and 3 Department of Cardiovascular Medicine, Institute for Genetics of Heart Diseases (IfGH), University Hospital Münster, Münster, Germany Received 23 May 2012; accepted after revision 20 August 2012; online publish-ahead-of-print 1 October 2012 Aims Clinical and electrocardiographic (ECG) presentation of patients with arrhythmogenic right ventricular cardiomyopathy (ARVC) and idiopathic right ventricular outflow-tract tachycardia (RVOT) may be similar. The aim of the study was to assess the validity and utility of T-wave integral measurement as an ECG discriminator of patients with ARVC and RVOT using a body surface mapping (BSM).... Methods A 120-channel BSM with quantitative signal analysis of the T-wave integral was performed in 10 patients with ARVC. and results Results were compared with those obtained from 13 patients with RVOT and a control group of 12 healthy subjects (controls). Age, body mass index, and QRS-axis on surface ECG were not significantly different between the groups. Arrhythmogenic right ventricular cardiomyopathy patients showed a significantly negative T-wave integral in the right lower anterior region of the torso when compared with RVOT (P, 0.001). There was no statistically significant difference between RVOT patients and controls. At a cut-off level of 20.3 mv ms, sensitivity and specificity were 83% [area under curve (AUC) for the comparison of ARVC and RVOT]. These differences were pronounced in ARVC patients with a plakophlin-2 mutation (P, 0.001).... Conclusion Quantitative analysis of the BSM T-wave integral in distinct anatomical regions discriminates ARVC patients from RVOT patients and controls and may serve as an additional diagnostic tool Keywords Arrhythmogenic right ventricular cardiomyopathy Electrocardiography Body surface mapping Repolarization Introduction Arrhythmogenic right ventricular cardiomyopathy (ARVC) is among the leading causes of ventricular tachyarrhythmias and sudden cardiac death in the young. Replacement of right ventricular (RV) myocardium by fibrous-fatty tissue provides the structural basis for ventricular tachycardias (VT) and for typical alterations detectable on 12-lead surface electrocardiogram (ECG). 1 3 The These authors contributed equally to the manuscript. Present address: Department of Cardiology, Landesklinikum St. Pölten, St. Pölten, Austria. } Present address: Department of Cardiology and Angiology, Klinikum Leer, Leer, Germany. Present address: Department of Cardiology, Niels-Stensen-Kliniken, Marienhospital Osnabrück, Osnabrück, Germany. * Corresponding author. Department für Kardiologie und Angiologie, Klinik für Kardiologie, Universitätsklinikum Münster, Albert-Schweitzer-Campus 1, Gebäude A1, Münster, Germany. Tel: ; fax: +(49) , Matthias.Paul@ukmuenster.de Published on behalf of the European Society of Cardiology. All rights reserved. & The Author For permissions please journals.permissions@oup.com.

2 T-wave integral: an ECG marker discriminating patients with ARVC from patients with RVOT 583 What s new? Body surface potential mapping can be utilized to discriminate patients with arrhythmogenic right ventricular cardiomyopathy and right ventricular outflow-tract tachycardia. T-wave integral as measured by body surface potential imaging identifies repolarization abnormalities in torso areas outside the standard lead position. latter may manifest as a prolongation/dispersion of precordial QRS-complexes, epsilon-potentials, as well as T-wave inversions in right precordial leads. 4,5 Albeit VT arising from the RV frequently occurs in patients with ARVC, it is also observed in patients without identifiable structural heart disease, e.g. in idiopathic right ventricular outflow-tract tachycardia (RVOT), which is one important differential diagnosis of ARVC. In contrast, patients with RVOT mostly show a normal surface ECG and have a favourable long-term survival. 6 It is essential to establish the correct diagnosis because treatment strategies and long-term prognosis differ significantly. Intriguingly, typical signs of ARVC on conventional 12-lead surface ECG are often subtle or even absent and, thus, may hamper diagnosing the underlying disease correctly. 7 In addition, right precordial T-wave inversions can be identified in a smaller portion (14%) of RVOT patients. 8 Body surface mapping (BSM) is a dependable tool for the detection of heterogeneities of ventricular repolarization unrevealed by conventional ECG recording. 9 So far, there have been two reports on repolarization abnormalities detectable by body surface QRST integral mapping. 10,11 Calculation of the QRST integral reflects the combination of both, the upstroke and downstroke in the action potential. 12 To eliminate potentially confounding effects of the QRS width, we assessed potential differences in repolarization in patients with ARVC and RVOT measuring the T-wave integral as an objective mathematical calculus with the help of a 120-channel BSM recording. Since mutations in the cardiac plakophilin-2 (PKP2) have been reported as the most common cause of autosomal-dominant ARVC (ARVC-9), 13 we additionally investigated a possible impact of the genotype on T-wave integral alterations. Materials and methods Study patients Patients with arrhythmogenic right ventricular cardiomyopathy and right ventricular outflow-tract tachycardia Ten consecutive male patients with ARVC (mean age years) and 13 patients with RVOT (mean age years) were prospectively enrolled in this study after written informed consent was obtained. All patients underwent detailed investigations including left ventricular and RV angiography (n ¼ 23), coronary angiography (n ¼ 23), and programmed ventricular stimulation (n ¼ 23) according to a stimulation protocol published earlier. 14 In addition, all patients met the past 15 and current criteria for definite ARVC. 16 Applying a scoring system with major criteria counting as 2 and minor criteria as 1 score point, all patients had at least four points which were required for the diagnosis of definite ARVC. 16 Right ventricular angiograms were evaluated in consensus by two experienced cardiologists (T.W. and M.P.) as to the extent of global RV dilatation and regional RV dysfunction (absent, moderate, severe). A moderate regional RV dysfunction was defined as the presence of severe hypokinesia in up to two different RV segments (RV outflow-tract, free wall, subtricuspid area, or apex). Evidence of severe hypokinesia or akinesia in more than two different RV segments lead to the classification of a severe regional RV dysfunction. Arrhythmogenic right ventricular cardiomyopathy patients were regularly followed up to evaluate a potential impact of T-wave integral on their arrhythmia profile. Follow-up started with the date of BSM recording and the first occurrence of any sustained VT/VF (i.e. lasting.30 s or requiring termination due to haemodynamic deterioration, respectively) served as the primary clinical endpoint. Follow-up data were obtained during scheduled visits in our outpatient clinic, regular telephone calls with patients and/or referring physicians at regular intervals, and in case of an adverse event (e.g. VT recurrence). In those patients with an implanted cardioverter defibrillator (ICD), interrogations of the device were scheduled every 3 months. An ICD therapy was classified as appropriate if the stored ECGs and/or RR intervals retrieved from the ICD confirmed that the tachyarrhythmia before the first treatment by the device was sustained and of ventricular origin. None of the ICD patients was pacemaker-dependent, the overall percentage of stimulation within the month before BSM recording was zero to exclude a potential confounding effect of temporary cardiac pacing (i.e. cardiac memory). Patients with RVOT had no evidence of any underlying structural heart disease. Global left ventricular function was normal in all patients (ARVC and RVOT), and no patient had coronary artery disease. Control group Twelve healthy volunteers (mean age years) with no history, signs, or symptoms of coronary artery disease or underlying cardiomyopathy served as a control group. Data acquisition and data analysis Body surface mapping Body surface mapping measurements were performed in the absence of any concomitant antiarrhythmic or sedative medication. In ARVC patients with an ICD, no stimulated QRS complexes were observed before or during BSM recording. As reported earlier, BSM electrodes were applied to the chest in vertical strips (Figure 2A and B). 17 Each electrode (Foxmed GmbH, Idstein, Germany) had a 10 mm diameter Ag/AgCl sensor embedded in epoxy housing with a 2 mm gel cavity. In vertical direction, the inter-electrode distance on the strips was 50 mm. The 120 unipolar ECG leads referred to Wilson central terminal and were recorded simultaneously. The ECG signals were amplified, bandpass-filtered (spectrum: Hz), and A/D converted to 16 bit samples (0.5 mv least significant bit) using a sampling rate of 1 khz (Mark-6, Biosemi, Amsterdam, Netherlands). Leads that were excessively noisy or which contained artefacts were excluded from the data set for further off-line analysis. Measurements were performed during stable sinus rhythm and at resting conditions (5 min). Recordings with high signal-to-noise ratio and without any offset variation due to respiratory motion were selected for single-beat analysis as reported before. 17 T-wave integral in each channel including standard chest leads V 1 V 6 was measured in patients with ARVC and compared with those in RVOT patients and controls. Both the on- and the offset of the T-wave were manually defined in consensus by three experienced cardiologists (A.S., C.V., and M.P.) according to peak-slope-intercept technique, 18 marked by digital callipers, and the PQ-interval served as reference zero.

3 584 A. Samol et al. Twelve-lead surface electrocardiogram A 12-lead surface ECG at a speed of 50 mm/s in the absence of any concomitant anti-arrhythmic medication at the time of the BSM recording was available in all patients and controls. Blinded for clinical details, these ECGs were quantitatively analysed in consensus by two experienced cardiologists (M.P., C.W.) using a digital calliper. Analyses comprised the measurement of QRS durations in precordial leads: maximum (QRS max ), minimum (QRS min ) and average QRS duration as well as the dispersion of QRS durations across V 1 -V 6 (i.e. QRS max QRS min ), presence/absence of characteristic epsilon potentials, and the presence/extent of T-wave inversions in leads V 1 -V 6. Statistical analysis Differences of metric target variables between groups were assessed by non-parametric ANOVA with post-hoc testing adjusted for multiplicity applying the closed test principle. In the case of binary target variables, the x 2 test was used. Results are intended to be exploratory (hypothesis generating), not confirmatory, and were interpreted accordingly. P values 0.05 were regarded significant. Receiver operating characteristic (ROC) analysis of the T-wave integral was performed for each lead. Values for the area under curve (AUC) are presented as mean + standard error of the mean; all other data are expressed as mean + standard deviation (where applicable). For correlation of target variables the Spearman s rank correlation coefficient was chosen. To account for variable lengths of follow-up, the probability of remaining arrhythmia/event-free was analysed by the Kaplan Meier method, and differences in event-free survival between groups were evaluated by Cox regression analyses. Statistical analyses were performed using IBM SPSS Statistics (version 20 for Windows, IBM Corporation, Somers, NY, USA). Results Clinical characteristics Demographic characteristics of the study patients are depicted in Table 1. There were no differences as to the age or body mass index at BSM recording between the groups (Table 1). Three patients with ARVC (30%) had received an automatic cardioverterdefibrillator (ICD) years prior to the BSM recording. Among these patients was one patient with a survived episode of sudden cardiac arrest, one with an extensive disease manifestation and sustained VT; and one patient with moderate RV dysfunction, non-sustained VT, and a highly malignant family history (two brothers succumbed to sudden cardiac death). During follow-up of years an ICD was implanted in another three patients after VT recurrences (cycle length ms) Table 1 Clinical characteristics of study patients ARVC... RVOT Controls P value All PKP2-pos PKP2-neg... Clinical characteristics Patients, n Age, years ns Sex (M/F) 10/0 a 5/0 b 5/0 b 4/9 a,c 6/6 c a,0.01; b,c ns Body mass index ns ICD, n (%) 6 (60) a 3 (60) c 3 (60) c 0 a a,0.01; c ns RV dilatation Normal, n (%) 0 a 0 c 0 c 13 (100) a ND Moderate, n (%) 6 (60) a 3 (60) c 3 (60) c 0 a ND Severe, n (%) 4 (40) a 2 (40) c 2 (40) c 0 a ND RV wall motion abnormalities Normal, n (%) 0 a 0 c 0 c 13 (100) a ND Moderate, n (%) 3 (30) a 1 (20) c 2(40) c 0 a ND Severe, n (%) 7 (70) a 4 (80) c 3 (60) c 0 a ND Programmed-ventricular stimulation Patients with PVS, n (%) 10 (100) 5 (100) 5 (100) 13 (100) ND ns Inducible sustained VT, n (%) 6 (60) a 3 (60) c 3 (60) c 0 a ND Task Force Score a c c a # major criteria d a c c a # minor criteria d a c c a a,0.01; c ns Follow-up, years c c c ns VT recurrence, n (%) 5 (50) 2 (40) c 3 (60) c c ns Time interval between BSM and VT recurrence, year c c c ns Values are expressed as mean + SD where applicable. a,b,c denote for different levels of significance between the groups compared in the line they appear; d from different diagnostic categories [16]. ARVC, arrhythmogenic right ventricular cardiomyopathy; BSM, body surface mapping; F, female; ICD, implantable cardioverter defibrillator; M, male; n, number; ns not significant; ND, not done; PKP2, plakophilin-2; pos, positive; PVS, programmed-ventricular stimulation; RVOT, right ventricular outflow-tract tachycardia; VT, ventricular tachycardias; ICD, implanted cardioverter defibrillator; RV, right ventricular.

4 T-wave integral: an ECG marker discriminating patients with ARVC from patients with RVOT years after BSM recording. Overall, five patients experienced life-threatening ventricular tachyarrhythmias years following BSM. Genetic analyses revealed mutations in the PKP2 encoding gene in five ARVC patients (50%). However, there were no significant differences in clinical disease manifestation as reflected in the Task Force Score between ARVC patients with PKP2-mutations and those without (PKP2-neg; P ¼ ns; Table 1). In PKP2-neg patients, additional screening as to potential mutations in other genes involved in ARVC (desmoglein-2 and desmocolin-2) was negative. Twelve-lead standard surface electrocardiogram in the study population BSM and conventional 12-lead surface ECGs were recorded during sinus rhythm in all patients. There were no significant differences in cycle length or QRS-axis between patients with ARVC in comparison with RVOT patients and controls (Table 2). In addition, QRS durations in leads V 1 V 6 and QRS dispersion were comparable between these groups. In three ARVC patients (30%) with a severe disease manifestation, epsilon potentials were detectable. T-wave inversion in precordial leads were present in nine patients with ARVC (90%), in 8 of 13 patients with RVOT (62%), and in 6 of 12 controls (50%; P ¼ ns; Table 2). The number of leads with T-wave inversion yet varied significantly and was highest in ARVC (2.6 vs. 0.8 vs. 0.5, respectively, between ARVC, RVOT, and controls; P, 0.01; Table 2). In lead V 1, 9 of 10 ARVC patients (90%), 8 of 13 RVOT patients (62%), and 6 of 12 controls (50%) had a negative T-wave. In leads V 1 and V 2, 7 of 10 patients with ARVC (70%), 2 of 13 patients with RVOT (15%), and none of the controls showed T-wave inversions. In the ARVC group, 6 of 10 patients (60%) had T-wave inversions in lead V 1 V 3 (n ¼ 4 patients) or beyond (n ¼ 2 patients) compared with no RVOT patient and no control subject. Between RVOT and controls, the number of leads with T-wave inversion was comparable ( vs ; P ¼ ns; Table 2). The presence of a PKP2 gene mutation made no difference on QRS-dispersion or the presence of epsilon potentials (Table 2). Mutation carriers tended to show more leads with T-wave inversions than non-carriers (P ¼ 0.086; Table 2). Table 2 Electrocardiographic characteristics of study patients ARVC... RVOT Controls P value All PKP2-pos PKP2-neg... Patients, n lead surface ECG Rhythm SR SR SR SR SR Heart rate, bpm ns QRS-axis, a b b a a a ns; b,0.05 QRS-durations QRS-V 1, ms ns QRS-V 2, ms a b b a a a ns; b,0.05 QRS-V 3, ms ns QRS-V 4, ms ns QRS-V 5, ms ns QRS-V 6, ms ns QRS-dispersion V 1 V ns Epsilon potential, n (%) 3 (30) a 1 (10) b 2 (20) b 0 a 0 a b ns; a,0.02 T-wave inversion V 1 V 6 Patients with TWI, n (%) 9 (90) 5 (100) 4 (80) 8 (62) 6 (50) ns Precordial leads with TWI, n a b b a a b ns; a,0.01 BSM-ECG and T-wave integrals Lead V 1 (mv ms) ns Lead V 2 (mv ms) ns Lead V 3 (mv ms) ns Lead V 4 (mv ms) ns Lead V 5 (mv ms) ns Lead V 6 (mv ms) ns Values are expressed as mean + SD. a,b denote for different levels of significance between the groups compared in the line they appear. ARVC, arrhythmogenic right ventricular cardiomyopathy; bpm, beats per minute; BSM, body surface mapping; ms, milliseconds; n, number; neg, negative; ns, not significant; PKP2, plakophilin-2, pos, positive; RVOT, right ventricular outflow-tract tachycardia; SR, sinus rhythm; TWI, T-wave inversion.

5 586 A. Samol et al. Body surface mapping in study population Calculating the T-wave integral in leads V 1 V 6, no statistically significant difference between the groups was found (P ¼ ns; Table 2). However, the values for T-wave integrals were negative in a distinct area in patients with ARVC compared with those with RVOT and controls (P, 0.001; Table 3). Best results to discriminate between ARVC and RVOT patients were obtained in channel #6. To account for inter-individual differences in torso shape and distances from the heart to body surface the four most significant adjacent channels were picked and the mean values of T-wave integral in this area were calculated for each patient. This area was located in the right lower anterior region of the torso and reflected recordings of the BSM channels #6, #7, #13, and #14 (Figure 2A and B). In ROC analyses, the AUC was [mean + standard error of the mean (SEM); P, 0.001]. At a cut-off level of 20.3 mv ms ROC analyses comparing ARVC and RVOT revealed a sensitivity and specificity of 83% (Table 3). Comparing ARVC with controls, patients with ARVC showed significant lower T-wave integrals in an adjacent area localized in the right lower anterior region of the torso reflecting recordings of the channels #20, #21, #27, and #28. Mean values of T-wave integrals of these four most significant channels for each patient was calculated and ROC analysis performed [AUC (mean + SEM), sensitivity 73%, specificity 72% at cut-off level of 0.21 mv ms]. In contrast, the value of the T-wave integrals in RVOT patients and controls was not different (P ¼ ns). Reciprocal changes were found in the posterior region on the left upper back of the torso comprising the four most significant adjacent BSM channels #79, #86, #93, and #94 (Figures 1 and 2A, B; Table 3). When compared with both RVOT and controls, patients with ARVC had a significantly higher mean T-wave integral of these four adjacent channels resulting in an AUC of (compared with RVOT; P, 0.001; Table 3) and (compared with controls; P, 0.001; Table 3) with a sensitivity and specificity of 75 of 85% at a cut-off value of 23.1 mv ms (RVOT) and 74 of 85% at a cut-off value of 23.0 mv ms (controls; Table 3), respectively. The difference in T-wave integral between RVOT and controls remained non-significant (P ¼ ns; Table 3). Analysing potential differences in T-wave integral in the above mentioned regions between PKP2-pos and PKP2-neg ARVC patients in comparison to RVOT and controls, it became apparent that PKP2-pos patients had the lowest T-wave integral in the anterior region of all groups (P, 0.001). ROC analyses revealed an AUC for the comparison of PKP2-pos and RVOT of (mean + SEM) with a sensitivity of 90% and specificity of 85% at a cut-off level of 23.9 mv ms. Similar results were obtained when T-wave integral of PKP2-pos patients were compared to that of controls [AUC (mean + SEM); sensitivity 82%, specificity 85% at cut-off level 23.9 mv ms]. Between PKP2-neg patients ( mv ms) and RVOT ( mv ms) these differences were less pronounced [AUC (mean + SEM); sensitivity 77%, specificity 70% at cut-off level 0.4 mv ms]. No statistically significant differences in T-wave integral were found between PKP2-neg patients and controls ( mv ms vs mv ms; mean + SD, P ¼ 0.28). In the posterior region, T-wave integral of both ARVC genetic subgroups was significantly more positive when compared with RVOT and controls (P, 0.001) but, different to that in the anterior region, insignificant in the direct comparison of PKP2-pos and PKP2-neg patients ( mv ms vs mv ms; P ¼ ns). Table 3 Receiver operating characteristic analyses of the T-wave integral in the study population ROC-analyses of T-wave-integral... ARVC vs. RVOT ARVC vs. controls RVOT vs. controls... Anterior region (channels #6, #7, #13, #14) T-wave integral (mv ms) vs vs vs Area under curve Cut-off level (mv ms) Sensitivity (%) Specificity (%) P value,0.001,0.001 ns Posterior region (channels #79, #86, #93, #94) T-wave integral (mv ms) vs vs vs Area under curve Cut-off level (mv ms) Sensitivity (%) Specificity (%) P value,0.001,0.04 ns Values are depicted as mean + standard deviation except for the area under curve which is shown as mean + standard error of the mean. ARVC, arrhythmogenic right ventricular cardiomyopathy; ms, milliseconds; mv, millivolt; ns, not significant; RVOT, right ventricular outflow-tract tachycardia.

6 T-wave integral: an ECG marker discriminating patients with ARVC from patients with RVOT 587 Figure 1 Three-dimensional-projection of T-wave integral. Three-dimensional-projections of the T-wave integral distribution on a standard torso model (frontal view; upper row) with corresponding recordings from the back (lower row) are depicted for the study population. Impact of body surface mapping findings on clinical presentation In patients with ARVC, T-wave integral was significantly correlated with the angiographic disease extent (anterior region: r ¼ 0.42; P, 0.044; posterior region: r ¼ 20.51; P, 0.02) and with the overall inducibility of sustained VT during programmed ventricular stimulation (n ¼ 6 patients; anterior region: r ¼ 20.49; P, 0.017; posterior region: r ¼ 0.64; P, 0.001). T-wave integral was not correlated with age, sex, or BMI. In addition, T-wave integrals in both regions were not associated with a higher propensity of VT recurrences during follow-up (P ¼ ns). Discussion Electrocardiographic diagnosis of ARVC in delineation to RVOT is challenging. The present BSM study focussed on repolarization abnormalities. Apart from the standard ECG leads, novel signal patterns could be identified which allowed a reliable electrocardiographic differentiation between patients with ARVC and RVOT/ controls. Our results suggest that maximum discrimination between ARVC- and RVOT-patients using T-wave integral analysis is localized in an area 10 cm caudal and 10 cm lateral from standard chest lead V 1. The prevalence of T-wave inversion beyond lead V 1 is known to change from infancy to adolescence and was estimated to be still present in 1 3% of a healthy population between 19 and 45 years of age. 19 In ARVC, divergent figures of T-wave inversion prevalence in precordial leads have been reported (36; 6 47; 20 54; 21 and 96%; 22 ). In the present study, 6 of 10 patients (60%) with ARVC showed T-wave inversion in leads V 1 V 3 or beyond which is in line with findings by Piccini et al. 23 (63%). It was hypothesized that characteristic ECG changes in patients with ARVC due to the progressive nature of the disease will change over time. 24 Different extents of disease manifestation may therefore contribute to the hitherto published low sensitivity but good specificity of T-wave inversions as an ECG discriminator of ARVC from RVOT. 20,25 To fill this gap, earlier reports concentrating on QRST-integral mapping with the help of BSM in patients with ARVC and RVOT identified negative areas on the right anterior and inferior chest independent of the prevalence of T-wave inversion on the corresponding standard 12-lead surface ECG. 10,11 Our analysis focussed on the T-wave integral to exclude a potentially confounding influence of the depolarization with an almost two-fold larger set of BSM-electrodes (120 vs. 62 electrodes 10,11 ) allowing a higher spatial resolution. Arrhythmogenic right ventricular cardiomyopathy patients showed a significantly lower T-wave integral in a circumscribed region of the right lower anterior torso when compared with both RVOT and controls with a comparably high sensitivity and specificity (P, 0.001). The observed significant abnormalities in repolarization in patients with ARVC may reflect local disturbances in refractoriness and conduction velocity caused by the interspersed fibro-fatty tissue. A thereby facilitated susceptibility to ventricular tachyarrhythmias, as indirectly expressed by the inducibility of sustained VT during programmed ventricular stimulation, could account for more adverse disease progression. 4 In our study, VT inducibility during programmed-ventricular stimulation was significantly correlated with the T-wave integral both in the anterior and in the posterior region (P, 0.02). Cox-regression analyses yet failed to show an impact on VT occurrence during follow-up. Study limitations This study included a relatively small number of patients with a rare cardiac disease. The patients are yet well characterized and

7 588 A. Samol et al. Figure 2 (A) Original body surface mapping recording in arrhythmogenic right ventricular cardiomyopathy and right ventricular outflow-tract tachycardia. Depiction of an original body surface mapping recording in a patient with arrhythmogenic right ventricular cardiomyopathy (black curve) and superimposed that of an right ventricular outflow-tract tachycardia patient (red curve). Each channel is visualized in a 1000 ms window. The regions of interest are highlighted in yellow and those with optimal discrimination encircled in darker blue. (B) Body surface mapping recording zoomed-in display of regions of interest. The region of interest with channel number on the left top of each rectangle, standard chest leads are marked on the right top where applicable, area-under-curve of the receiver operating characteristic analysis are listed on the left bottom, and P values are provided on the right bottom. Each channel is visualized in a 1000 ms window.

8 T-wave integral: an ECG marker discriminating patients with ARVC from patients with RVOT 589 underwent detailed diagnostic evaluation. However, due to the sample size, a potentially age- or disease-extent-dependent cut-off level of T-wave integral remains to be elucidated. Sensitivity and specificity results as well as ROC areas reported in this paper have been obtained on the same learning data sets than the ones that have been used to find the relevant discriminating measurements and to establish the discrimination cut-off levels. There is no independent test set for assessing the performance of our method. In addition, since the correlation of BSM parameters and the risk of sudden cardiac death may change with age, disease manifestation, and progression, variables which did not show a significant difference in our study could gain a significant and independent influence. Gender-specific differences in patients with ARVC could not be addressed in this study. However, no higher prevalence of T-wave inversion in leads V 1 V 3 were detectable in a large single-centre study [overall: 60 of 171 patients (45%); male: 44 of 121 patients (36%); female: 16 of 50 patients (32%); P ¼ ns 25 ]. Changes of T-wave inversion/t-wave integral in the clinical cause were beyond the aim of our study. In addition, the proposed cut-off values warrant prospective validation in a larger and independent group of patients. Our findings may yet initiate translation into a more convenient clinical routine application as standard ECG recording systems steadily develop and future enhanced software may allow the measurement of T-wave integral in alternative lead position. Conclusion Analysis of T-wave integral discriminates patients with ARVC from those with RVOT and from controls. Body surface mapping measurements detected torso areas outside the standard lead positions with significantly altered T-wave integral, in particular in the anterior right lower torso. Further studies using T-wave integral analysis should address the potential clinical value in identifying those patients in whom despite detailed invasive investigations the diagnosis of ARVC otherwise would have remained inconclusive (i.e. borderline ARVC 16 ). Acknowledgements The authors are grateful to Mrs Petra Gerdes, and Thomas Schawe for their excellent technical support. Conflict of interest: none declared. Funding This work was supported in part by grants from the Deutsche Forschungsgemeinschaft (DFG), Bonn, Germany, Sonderforschungsbereich 656 (Projects C1). References 1. Marcus FI, Fontaine GH, Guiraudon G, Frank R, Laurenceau JL, Malergue C et al. Right ventricular dysplasia: A report of 24 cases. Circulation 1982;65: Fontaine G, Fontaliran F, Frank R. Arrhythmogenic right ventricular cardiomyopathies: clinical forms and main differential diagnoses. Circulation 1998;97: Thiene G, Nava A, Corrado D, Rossi L, Pennelli N. Right ventricular cardiomyopathy and sudden death in young people. N Engl J Med 1988;318: Turrini P, Corrado D, Basso C, Nava A, Bauce B, Thiene G. Dispersion of ventricular depolarization-repolarization: a noninvasive marker for risk stratification in arrhythmogenic right ventricular cardiomyopathy. Circulation 2001;103: Corrado D, Fontaine G, Marcus FI, McKenna WJ, Nava A, Thiene G et al. Arrhythmogenic right ventricular dysplasia/cardiomyopathy: need for an international registry. Circulation 2000;101:E101 E O Donnell D, Cox D, Bourke J, Mitchell L, Furniss S. Clinical and electrophysiological differences between patients with arrhythmogenic right ventricular dysplasia and right ventricular outflow tract tachycardia. Eur Heart J 2003;24: Kazmierczak J, De Sutter J, Tavernier R, Cuvelier C, Dimmer C, Jordaens L. Electrocardiographic and morphometric features in patients with ventricular tachycardia of right ventricular origin. Heart 1998;79: Markowitz SM, Litvak BL, Ramirez de Arellano EA, Markisz JA, Stein KM, Lerman BB. Adenosine-sensitive ventricular tachycardia: right ventricular abnormalities delineated by magnetic resonance imaging. Circulation 1997;96: Mitchell LB, Hubley-Kozey CL, Smith ER, Wyse DG, Duff HJ, Gillis AM et al. Electrocardiographic body surface mapping in patients with ventricular tachycardia. Assessment of utility in the identification of effective pharmacological therapy. Circulation 1992;86: De Ambroggi L, Aime E, Ceriotti C, Rovida M, Negroni S. Mapping of ventricular repolarization potentials in patients with arrhythmogenic right ventricular dysplasia: principal component analysis of the ST-T waves. Circulation 1997;96: Peeters HA, SippensGroenewegen A, Schoonderwoerd BA, Wever EF, Grimbergen CA, Hauer RN et al. Body-surface QRST integral mapping. Arrhythmogenic right ventricular dysplasia versus idiopathic right ventricular tachycardia. Circulation 1997;95: Berger T, Hintringer F, Fischer G. Noninvasive imaging of cardiac electrophysiology. Indian Pacing Electrophysiol J 2007;7: Gerull B, Heuser A, Wichter T, Paul M, Basson CT, McDermott DA et al. Mutations in the desmosomal protein plakophilin-2 are common in arrhythmogenic right ventricular cardiomyopathy. Nat Genet 2004;36: Wichter T, Hindricks G, Lerch H, Bartenstein P, Borggrefe M, Schober O et al. Regional myocardial sympathetic dysinnervation in arrhythmogenic right ventricular cardiomyopathy. An analysis using 123I-meta-iodobenzylguanidine scintigraphy. Circulation 1994;89: McKenna WJ, Thiene G, Nava A, Fontaliran F, Blomstrom-Lundqvist C, Fontaine G et al. Diagnosis of arrhythmogenic right ventricular dysplasia/cardiomyopathy. Task Force of the Working Group Myocardial and Pericardial Disease of the European Society of Cardiology and of the Scientific Council on Cardiomyopathies of the International Society and Federation of Cardiology. Br Heart J 1994;71: Marcus FI, McKenna WJ, Sherrill D, Basso C, Bauce B, Bluemke DA et al. Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia: proposed modification of the task force criteria. Circulation 2010;121: Bruns HJ, Eckardt L, Vahlhaus C, Schulze-Bahr E, Haverkamp W, Borggrefe M et al. Body surface potential mapping in patients with Brugada syndrome: right precordial ST segment variations and reverse changes in left precordial leads. Cardiovasc Res 2002;54: McLaughlin NB, Campbell RW, Murray A. Comparison of automatic QT measurement techniques in the normal 12 lead electrocardiogram. Br Heart J 1995; 74: Marcus FI. Prevalence of T-wave inversion beyond V1 in young normal individuals and usefulness for the diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia. Am J Cardiol 2005;95: Morin DP, Mauer AC, Gear K, Zareba W, Markowitz SM, Marcus FI et al. Usefulness of precordial T-wave inversion to distinguish arrhythmogenic right ventricular cardiomyopathy from idiopathic ventricular tachycardia arising from the right ventricular outflow tract. Am J Cardiol 2010;105: Peters S, Trümmel M. Diagnosis of arrhythmogenic right ventricular dysplasiacardiomyopathy: value of standard ECG revisited. Ann Noninvasive Electrocardiol 2003;8: Ma KJ, Li N, Wang HT, Chu JM, Fang PH, Yao Y et al. (2009) Clinical study of 39 Chinese patients with arrhythmogenic right ventricular dysplasia/cardiomyopathy. Chin Med J (Engl) 2009;122: Piccini JP, Nasir K, Bomma C, Tandri H, Dalal D, Tichnell C et al. Electrocardiographic findings over time in arrhythmogenic right ventricular dysplasia/cardiomyopathy. Am J Cardiol 2005;96: Marcus FI, Zareba W. The electrocardiogram in right ventricular cardiomyopathy/ dysplasia. How can the electrocardiogram assist in understanding the pathologic and functional changes of the heart in this disease? J Electrocardiol 2009;42: 136.e1 136.e Bauce B, Frigo G, Marcus FI, Basso C, Rampazzo A, Maddalena F et al. Comparison of clinical features of arrhythmogenic right ventricular cardiomyopathy in men versus women. Am J Cardiol 2008;102: