OUP user menu

Beat-to-beat T-wave amplitude variability in the risk stratification of right ventricular outflow tract-premature ventricular complex patients

Tomohide Ichikawa, Yoshihiro Sobue, Atsunobu Kasai, Ken Kiyono, Junichiro Hayano, Mayumi Yamamoto, Kentarou Okuda, Eiichi Watanabe, Yukio Ozaki
DOI: http://dx.doi.org/10.1093/europace/euu404 138-145 First published online: 1 March 2015


Aims Premature ventricular complexes (PVCs) originating from the right ventricular outflow tract (RVOT) may occasionally trigger monomorphic ventricular tachycardia (MVT), polymorphic ventricular tachycardia (PVT), or ventricular fibrillation (VF). We examined whether an analysis of the ventricular repolarization instability could differentiate PVT/VF triggered by RVOT-PVCs from benign RVOT-PVCs or MVT.

Methods We evaluated the ventricular repolarization instability as assessed by the beat-to-beat T-wave amplitude variability (TAV) using Holter recordings in patients with RVOT-PVCs but with no structural heart disease. We determined the prematurity index, defined as the ratio of the coupling interval of the first ventricular tachycardia (VT) beat or isolated PVC to the preceding R–R interval just before the VT or isolated PVC in the Holter recordings. The study patients were classified into RVOT-PVCs/MVT (n = 33) and PVT/VF (n = 10).

Results The two groups did not differ with respect to the age, sex, and left ventricular ejection fraction. There was no significant difference in the prematurity index between the two groups (RVOT-PVCs/MVT 0.66 ± 0.16 vs. PVT/VF 0.61 ± 0.13, P = 0.60). The patients with PVT/VF had a significantly larger maximum TAV than those with RVOT-PVCs/MVT (31 ± 13 vs. 68 ± 40 µV, P < 0.001). Patients with a higher than median value of the TAV (33 µV) were at increased risk of PVT/VF vs. those with a lower than median value, after adjusting for the age and sex [9.25 (95% confidence interval: 1.27–19.2); P = 0.03].

Conclusions The TAV analysis is a useful measure to identify the subset of usually benign RVOT-PVC/MVT patients prone to PVT/VF.

  • Sudden cardiac death
  • Arrhythmia
  • Repolarization
  • Catheter ablation
  • Holter electrocardiogram

What's new?

  • Premature ventricular complexes (PVCs) originating from the right ventricular outflow tract (RVOT) are one of the most common arrhythmias and are generally considered benign. Nevertheless, a subset develops into lethal polymorphic ventricular tachycardia (PVT) or ventricular fibrillation (VF).

  • Up to now, coupling interval of the triggering PVC, prematurity index, and QT index was useful markers to identify the subset of patients with RVOT-PVCs prone to PVT/VF.

  • We found that for the first time an increased beat-to-beat T-wave amplitude variability in the Holter electrocardiogram might identify the subset of patients with RVOT-PVCs prone to PVT/VF.


Premature ventricular complexes (PVCs) originating from the right ventricular outflow tract (RVOT) are one of the more common ventricular arrhythmias.1 Right ventricular outflow tract-premature ventricular complexes are frequently observed in patients who have no structural heart disease and are considered benign even when they develop into monomorphic ventricular tachycardia (MVT).2 Nevertheless, one occasionally comes across lethal polymorphic ventricular tachycardia (PVT) or ventricular fibrillation (VF) triggered by RVOT-PVCs.37

The mechanism of RVOT-ventricular tachycardia (VT) is typically triggered activity caused by a calcium overload via a cyclic adenosine monophosphate mediated process.8 Therefore, various investigators have focused on the role that short RVOT-PVC coupling intervals might play in triggering PVT/VF; however, the triggering coupling interval values reported vary widely.36 Recently, Igarashi et al.7 reported that the prematurity index or QT index4 could differentiate PVT/VF triggered by RVOT-PVCs from MVT.

The pathophysiology of ventricular tachyarrhythmias is believed to require both a triggering event, and underlying temporal and spatial dispersion of the repolarization. The T-wave alternans (TWA) phenomenon, beat-to-beat alteration in the morphology, and amplitude of the T-wave have long been recognized and linked to arrhythmogenesis.9,10Among the non-invasive risk stratifiers, the frequency-domain technique for the assessment of TWA is an established measure for the prediction of VT/VF in diverse heart diseases.10 Recently, Couderc et al.11 developed a novel approach to determine the beat-to-beat T-wave amplitude variability (TAV) obtained from the Holter electrocardiogram (ECG). They demonstrated that an increased TAV was associated with an increased risk for VT assessed by the frequency of implantable cardioverter-defibrillator (ICD) anti-tachycardic therapies in post-infarction patients with severe left ventricular dysfunction. Subsequent studies have shown that an increased TAV is also observed in patients with various heart diseases.12,13 The aim of this study was to test the hypothesis that the TAV is a useful parameter for the identification of the subset of patients with RVOT-PVCs prone to PVT/VF and differentiate them from the benign RVOT-PVC and MVT patients.


Study population

We prospectively enrolled 43 consecutive RVOT-PVC patients without any structural heart disease who underwent 24 h Holter ECG recordings from May 2008 to December 2013. We classified the patients into two groups: RVOT-PVC/MVT and PVT/VF. We first screened the RVOT-PVC patients who had PVCs appearing in all 12 ECG leads simultaneously to determine the origin based on the algorithm shown in the previous report.14 We excluded the patients with structural heart disease on the basis of the medical history, physical examination, laboratory data, echocardiography, and coronary angiography. Arrhythmogenic right ventricular cardiomyopathy was excluded in the PVT/VF patients according to the task force criteria.15 Brugada syndrome was excluded by the absence of a Brugada-pattern ECG after flecainide provocation and the absence of a family history of sudden death at a young age in the all PVT/VF patients. We also excluded patients with long-QT syndrome and J wave syndrome. We excluded patients with atrial fibrillation because the utility of the TAV analysis was untested in such patients. The study protocol was approved by the Institutional Review Board and all patients gave their written informed consent.

Electrocardiogram analysis

A single electrocardiographic technician blinded to any of the clinical information of the patients analysed the ECG measurements. We determined the following ECG parameters from both the 12-lead ECG at rest and Holter recording during isolated PVCs or episodes of VT: (i) the sinus cycle length just before isolated PVCs or VT; (ii) coupling interval, defined as the R–R interval between isolated PVCs or PVCs initiating VT and the preceding sinus complex; (iii) prematurity index,4 defined as the ratio of the coupling interval to the preceding sinus R–R interval [i.e. value of (ii)/value of (i)]; and (iv) the QT index,4 defined as the ratio of the coupling interval to the QT interval of the preceding sinus complex. We examined the polarity of the PVC in lead I and defined it as positive if there was a positive deflection exceeding the negative component by an amplitude of >0.1 mV.16 Ventricular tachycardia was considered present when at least three consecutive PVCs occurred at a rate of >120 beats/min. Polymorphic ventricular tachycardia was defined as VT with more than five consecutive beats with different QRS morphologies.5 Ventricular fibrillation required a disorganized ventricular rhythm with no discrete QRS complexes. The PVT/VF episodes were obtained from Holter recordings or automated external defibrillators during cardiopulmonary resuscitation, and monitor ECGs recorded on the wards.

Holter electrocardiogram recordings and analysis of the T-wave amplitude variability

Digital 24 h Holter ECG recordings (SpiderView, Ela Medical, Sorin Group, Le Plessis Robinson, France) were sampled at 1 kHz with a resolution of 2.5 µV. The electrodes (Blue Sensor, Ambu, Ballerup, Denmark) were placed in an orthogonal X, Y, and Z lead configuration. The Holter data were manually edited to eliminate any ectopic beats and signal noise and then analysed for the conventional heart rate variability parameters, dynamic heart rate variability indices including the deceleration capacity,17 non-Gaussian index,18 heart rate turbulence,19 and signal-averaged ECG.20 The detailed method for measuring the TAV has been reported elsewhere.11,12 SyneTVar 3.10b software (Ela Medical, Sorin Group) was used for the analysis of the TAV based on the vector magnitude of VM = √X2 + Y2 + Z2, where VM was used as the primary lead for the analysis. The software automatically selects a cluster of 60 QRS-T consecutive complexes for the analysis (Figure 1). The clusters were excluded from analysis in case of (i) the presence of atrial and ventricular ectopy, (ii) high R–R interval variability defined as the presence of R–R intervals >20 or <20% from the mean, and (iii) a noise level exceeding 10 µV. The variance of the TAV (in µV), defined as the average of the squared deviations from the mean, was assessed on each of eight consecutive 50 ms T-wave segments (TAV 1–TAV 8) following the QRS offset (defined as the QRS onset +120 ms) for a given cluster. The mean TAV was defined as the average TAV from T-wave segments 1–8, and the max TAV as the maximum TAV from T-wave segments 1–8. Finally, each patient had a mean TAV that was the mean value of the TAV from all clusters of the recording and had a max TAV that was defined as the maximum TAV value of all the clusters. Uninterpretable clusters were carefully deleted by visual inspection. To assess the reproducibility of the TAV analysis, all 43 recordings were interpreted at an interval of 1 month, resulting in a correlation coefficient of 0.95. Holter ECGs were recorded before the administration of any antiarrhythmic drugs.

Figure 1

T-wave amplitude variability. A set of 60 continuous sinus beats is identified and aligned based on beginning of the QRS and the T wave is windowed using a 50 ms duration interval. Upper panels: the oblique view of 60 consecutive T waves. Lower panels: transverse views of the 60 consecutive T waves. The control subject exhibited a small variation in the T-wave amplitude (10 µV in this case), but the ventricular fibrillation patient had a large variation in the T-wave amplitude (81 µV in this case).

Electrophysiologic evaluation and radiofrequency catheter ablation

An electrophysiologic evaluation and radiofrequency catheter ablation (RFCA) targeting the triggering PVC was attempted before the administration of amiodarone or after withdrawal of other antiarrhythmic drugs for more than five times the half-life. The surface ECGs and bipolar endocardial electrograms were continuously monitored and stored on a computer-based digital amplifier/recorder system (Labsystem™ Pro, Bard Electrophysiology, Lowell, MA, USA). Intracardiac electrograms were filtered from 30 to 500 Hz. The Carto™ system (Biosense-Webster, Diamond Bar, CA, USA) or EnSite™ Array™ mapping system (St. Jude Medical, St. Paul, MN, USA) was used for the 3D electroanatomical mapping. Steerable quadripolar electrode catheters were introduced into the His bundle region and RV apex, and a decapolar catheter (5 mm electrode spacing, St. Jude Medical) was positioned within the coronary sinus via the right femoral vein. A 7-Fr quadripolar electrode catheter with a 4 mm distal electrode and a deflectable tip (Safire™ BLU™ or Therapy™ Cool Flex™, St. Jude Medical) was used for mapping and ablation. Programmed stimulation was performed with up to three extrastimuli from the RV apex and RVOT, and isoproterenol was used in an effort to induce VT. The site of the PVC origin was defined as the site where the earliest ventricular activation was recorded and/or a perfect pace map was obtained, and radiofrequency energy was delivered at that site. The radiofrequency energy was a maximum power of 30 W, maximum temperature of 40°C, and duration of up to 60 s for each delivery. Successful RFCA was defined as the absence of any spontaneous or induced clinical VT/PVCs, both in the absence and presence of isoproterenol, at the end of the procedure.

Statistical analysis

Differences between the two groups were evaluated using a Student's t-test for continuous variables and the χ2 test for categorical data. A multiple regression analysis was used to test the relationship between the heart rate variability index, heart rate turbulence, and TAV. A multivariate logistic analysis adjusted by the age and sex was performed to test the predictive value of the TAV for PVT/VF. Data are presented as the mean ± SD. A P-value of < 0.05 was considered as statistically significant. The statistical analyses were performed using JMP10.0.2 software.


Patient characteristics

The baseline characteristics classified according to the type of arrhythmia are summarized in Table 1. No significant differences were found in regard to the age, sex, comorbidities, and left ventricular ejection fraction between the RVOT-PVC/MVT patients and PVT/VF patients. No patients had a family history of sudden death. We found no significant difference in the prevalence of a positive QRS in lead I between the two groups. There were no significant differences in the coupling interval, prematurity index, and QT index in the 12-lead ECG between the two groups. Also, there were no significant differences in the QRS duration, R wave duration index, and R/S wave amplitude index, and the QRS morphology of the PVCs met the RVOT origin criteria. The clinical profiles of the 10 PVT/VF patients are presented in Table 2. Two patients with PVT/VF also had episodes of MVT with a cycle length of ≤300 ms (Patients #4 and #6, Table 2), and they were analysed as PVT/VF patients. A representative case of PVT/VF recorded on the Holter ECG is shown in Figure 2.

View this table:
Table 1

Baseline characteristics of the patients

RVOT-PVC/MVT (n = 33)PVT/VF (n = 10)P-value
Age (years)54 ± 1847 ± 140.25
Sex: male/female14/198/20.07
RVOT-PVC/MVT, n19/14
PVT/VF, n7/3
 Syncope, n (%)2 (6)4 (40)
 Presyncope, n (%)0 (0)4 (40)
 Cardiac resuscitation, n (%)0 (0)3 (30)
 Hypertension, n (%)6 (18)3 (30)0.41
 Diabetes, n (%)3 (9)2 (20)0.57
 Dyslipidemia, n (%)7 (21)1 (10)0.66
 Chronic kidney disease, n (%)4 (12)1 (10)1.00
Twelve-lead ECG
 Positive QRS in lead I, n (%)23 (70)8 (80)0.70
 QT interval (ms)388 ± 37383 ± 370.73
 QTc interval (ms)388 ± 79416 ± 170.28
 Coupling interval (ms)517 ± 81477 ± 580.16
 Prematurity index0.59 ± 0.110.65 ± 0.230.26
 QT index1.30 ± 0.341.25 ± 0.110.62
 QRS duration of PVCs (ms)145 ± 34142 ± 240.79
 R wave duration index (%)37 ± 1536 ± 170.83
 R/S wave amplitude index (%)21 ± 1924 ± 250.82
LVEF (%)59 ± 560 ± 50.43
Medications, n (%)
 Class I antiarrhythmic drug8 (24)1 (10)0.66
 Amiodarone0 (0)2 (20)0.05
 β-Blocker9 (27)6 (60)0.07
 PVC origin (septum/lateral-free wall)5/35/31.00
 Success/partial success8/07/11.00
ICD1 (3)5 (50)0.001
  • Data represent the means ± SD or frequency.

  • RVOT, right ventricular outflow tract; PVC, premature ventricular complex; MVT, monomorphic ventricular tachycardia; PVT, polymorphic ventricular tachycardia; VF, ventricular fibrillation; LVEF, left ventricular ejection fraction; RFCA, radiofrequency catheter ablation; ICD, implantable cardioverter-defibrillator; QTc, rate-corrected QT interval by Bazett's formula.

View this table:
Table 2

Clinical profile of the PVT/VF patients

CaseAge (years)SexType of ventricular tachyarrhythmiaSymptomCPRLVEF (%)Max TAV (µV)Mean TAV (µV)PVC originRFCA resultsAmiodarone/β-blockerICD implantationOutcome
561MalePVTPresyncope588855SeptalPartial successA/B+Alive
839MalePVTSyncope+524316Lateral/free wallSuccess−/−Alive
945MalePVTSyncope639928Lateral/free wallSuccess−/BAlive
1047FemaleVFSyncope661612Lateral/free wallSuccess−/−+Alive
  • CPA, cardiopulmonary arrest; CPR, cardiopulmonary resuscitation; TAV, T-wave amplitude variability; A, amiodarone; B, β-blocker.

  • The other abbreviations are as in Table 1.

Figure 2

Polymorphic ventricular tachycardia/ventricular fibrillation recording from the Holter ECG. (A) Twelve-lead ECG with an RVOT-PVC. (B) Ventricular fibrillation triggered by an RVOT-PVC self-terminates after ∼90 s (Case #10 in Table 2). CM5 and NASA are modified bipolar leads corresponding to V5 and V1 in 12-lead ECG, respectively.

Holter electrocardiogram parameters

Table 3 summarizes the Holter ECG data. The heart rate variability index and heart rate turbulence were determined in all patients. The TAV results were interpretable in 43 of 52 patients (83%), whose atrial (n = 1) or ventricular (n = 4) premature complexes, or paroxysmal atrial fibrillation (n = 4) precluded the TAV measurement. We analysed 1696 ± 95 clusters in the study patients. There were no significant differences in the heart rate, conventional heart rate variability parameters, dynamic heart rate variability indices including the deceleration capacity, non-Gaussian index, heart rate turbulence, or signal-averaged ECG. A multiple regression analysis revealed that there were no significant associations between the heart rate variability index, heart rate turbulence, and TAV (see Supplementary material online, Table S1). The PVT/VF patients tended to have fewer PVCs (P = 0.07) and a relatively higher number of PVC morphologies (P = 0.07) compared with the RVOT-PVC/MVT patients. The coupling interval in the PVT/VF patients was shorter than that in the RVOT-PVC/MVT patients (P = 0.03). There were no significant differences in the prematurity index and QT index between the two groups. The PVT/VF patients exhibited a significantly higher value of the max TAV than patients with RVOT-PVC/MVT (P < 0.001). The distribution of the max TAV across the repolarization intervals is presented in a Supplementary material online, file. The max TAV was frequently recorded at the peak of the T-wave.

View this table:
Table 3

Holter ECG recording parameters

RVOT-PVC/MVT (n = 33)PVT/VF (n = 10)P-value
Heart rate (beats/min)70 ± 867 ± 100.24
PVC (beats/day)6033 ± 76191506 ± 2300.07
Characteristics of the PVC, VT, PVT
 Number of PVC morphologies, n1.2 ± 0.51.7 ± 1.30.07
 R–R interval before the PVC or VT (ms)808 ± 187742 ± 2220.36
 Cycle length (ms)a429 ± 75272 ± 70<0.001
 Coupling interval (ms)500 ± 74428 ± 840.03
 Prematurity index0.66 ± 0.160.61 ± 0.130.60
 QT index1.24 ± 0.231.20 ± 0.180.68
Heart rate variability
 SDNN (ms)139 ± 48124 ± 420.37
 pNN50 (%)12 ± 156.6 ± 5.70.25
 HF (ms2)1593 ± 5334170 ± 1420.41
 LF (ms2)682 ± 730553 ± 5030.60
 LF/HF2.5 ± 1.73.5 ± 2.00.13
 Deceleration capacity (ms)6.1 ± 1.87.4 ± 2.50.09
 Acceleration capacity (ms)−7.1 ± 2.5−7.7 ± 2.70.52
λ25s0.65 ± 0.210.59 ± 0.180.45
Heart rate turbulence
 Onset (%)−1.6 ± 2.0−1.7 ± 2.00.97
 Slope (ms/beat)6.3 ± 4.67.3 ± 5.80.59
 Abnormal18 (60)5 (55)1.00
Signal-averaged ECG
 Positive, n (%)7 (22)4 (44)0.22
 fQRS (ms)93 ± 10101 ± 220.11
 RMS40 (µV)45 ± 3343 ± 290.86
 LAS40 (ms)34 ± 1039 ± 170.25
 Noise (µV)0.5 ± 0.40.6 ± 0.50.58
 Max TAV (µV)31 ± 1368 ± 40<0.001
 Mean TAV (µV)18 ± 6.432 ± 15<0.01
 Noise (µV)4.5 ± 1.35.3 ± 1.70.12
  • Data represent the means ± SD.

  • SDNN, SD of the normal-to-normal interval; pNN50, the proportion of successful intervals that differ by more than 50 ms; HF, high frequency; LF, low frequency; fQRS, filtered QRS duration; RMS40, root mean square voltage of the terminal 40 ms of the filtered QRS; LAS40, the amount of time that the filtered QRS complex remains below 40 µV.

  • The reference values were as follows: deceleration capacity >2.5 ms, heart rate turbulence: turbulence onset <0% and turbulence slope >2.5 ms/beat, and non-Gaussian index (λ25 s) <0.6. If the signal-averaged ECG filled two or more of an fQRS >114 ms, RMS40 <20 µV, and LAS40 >38 ms, it was defined as positive. TAV, T-wave amplitude variability.

  • aMVT only (n = 14).

Risk prediction

When patients were dichotomized by a median value of the TAV of 33 µV, patients with a higher than median value were at increased risk of PVT/VF vs. those with a lower than median value after adjusting for the age and sex [9.25 (95% confidence interval: 1.27–19.2); P = 0.03]. Thus, the sensitivity and specificity for predicting PVT/VF were 90.0 and 60.6%, respectively. The coupling interval was not predictive of PVT/VF [4.76 (95% confidence interval: 0.83–39.1); P = 0.08].

Treatment and follow-up

Radiofrequency catheter ablation targeting the triggering PVCs was performed (Table 1). Ventricular fibrillation was induced in one patient with PVT/VF (Patient #8, Table 2) with three extrastimuli from the RV apex in the absence of isoproterenol, and an isoproterenol infusion alone induced MVT in one MVT patient. There was no significant difference in the location of the origin of the triggering PVC between the two groups. Endocardial mapping presented no local abnormal ECGs including fragmentation or delayed potentials. After the RFCA for the initial targeted PVC, a PVC with a different QRS morphology appeared in one patient in the PVT/VF group (Patient #4, Table 2). A further energy delivery ∼1 cm above this site abolished another form of PVCs. Partial success was achieved in Patient #5 since the triggering PVC occurred very infrequently, and the induction of the PVC or PVT/VF was difficult. We also confirmed that no perfect pace maps were obtained in the left ventricular outflow tract, pulmonary artery, or aortic cusps. The mean bipolar local activation time at the successful RFCA site was 35 ± 12 ms before the surface QRS onset. No significant difference was observed in the RFCA success rate between the two groups. Implantable cardioverter-defibrillators were implanted in five patients with PVT/VF, but the remaining five patients refused to undergo an ICD implantation. An appropriate shock for VF was delivered in Patient #2, in whom RFCA was refused. No patients died during a mean follow-up of 28 ± 10 months.


The electrophysiological mechanisms responsible for enhanced susceptibility to life-threatening arrhythmias in RVOT-PVC patients without structural heart disease are not well understood. We observed that the prematurity index and QT index failed to differentiate the subset of patients with RVOT-PVCs prone to PVT/VF from the RVOT-PVC/MVT patients. We found that for the first time that the increased repolarization lability as assessed by the TAV was predictive of PVT/VF.

Previous studies have suggested that short coupling intervals of triggering PVCs are responsible for the development of PVT/VF, regardless of the location of the focus.2 Leenhardt et al.3 showed that short-coupled PVCs (a coupling interval of 245 ± 28 ms) might be related to PVT/VF, which was referred to as a ‘variant of torsade de pointes.’ Subsequent authors have reported that PVT/VF can be triggered by intermediately coupled RVOT-PVCs,4,6 or even longer RVOT-PVC coupling intervals5,7 trigger the development of PVT/VF. Therefore, a ‘short’ coupling interval alone may be a poor predictor of whether an RVOT-PVC is benign or possibly lethal. In this study, the coupling interval in the PVT/VF patients was shorter than that in the RVOT-PVC/MVT patients, but the predictive accuracy of the coupling interval was low compared with the max TAV.

The smaller prematurity index and the QT index have been shown to be associated with the initiation of PVT or VF.4 Igarashi et al.,7 who examined the prematurity index between patients with MVT and PVT arising from RVOT-PVCs, found that a smaller prematurity index was able to differentiate between MVT and PVT. In this study, however, we found that there were no significant differences in the prematurity index and the QT index between the two groups.

The PVT/VF patients tended to have a greater number of PVC morphologies than the RVOT-PVC/MVT patients. In addition, two patients with PVT/VF had both episodes of PVT/VF and MVT, suggesting we may not have been able to exclude an undiagnosed cardiomyopathic process contributing to the arrhythmias in the PVT/VF patients.

In this study, the average number of PVCs during the Holter recordings was relatively smaller in the RVOT-PVC/MVT patients than those with PVT/VF. We have no clear explanation for our observation, but the study by Viskin et al.6 showed that the total number of PVCs was relatively greater in those with benign RVOT-VT than those with short-coupled RVOT-VT or idiopathic VF. Moreover, Leenhardt et al.3 reported that the total number of PVCs in the eight patients with a short-coupled variant of torsades de pointes was 846 ± 1473 per day, which was less than our results.

We found that an increased TAV was associated with the PVT/VF patients, suggesting that an arrhythmogenic substrate exists in the ventricles, but we could not diagnose whether there was a repolarization abnormality localized to the RVOT area or not. It may be helpful if we could assess the repolarization process in the RVOT region.

The TAV, in addition to TWA9 and the QT interval variability,21 has been described as a promising new technique for the quantification of ventricular repolarization instability and the stratification of the arrhythmic risk in various heart diseases.1113 We previously examined the TAV in the survivors of VT/VF without evidence of structural heart disease.22 The cut-off value for the max TAV in that study was 38 µV,22 which was similar to this study (median 33 µV). Extramiana et al.12 reported that the max TAV in long-QT patients was 46 µV, lower than the 58 µV cut-off value determined by Couderc et al.11 for the MADIT II study population with an ICD therapy endpoint. These observations suggest that the optimal cut-off value for the TAV for the risk stratification should be determined by the underlying heart disease.

Recently, newly developed time-domain techniques such as the modified moving average TWA,23 in addition to conventional spectral methods,9 have been proven to be useful measures to stratify high-risk patients among the diverse underlying heart diseases. The strength of the TWA is that it has experimental background and extensive clinical evidence. The TAV, a conceptually similar measure, has been developed as a quantitative approach that assesses subtle ventricular repolarization abnormalities. Although much clinical evidence has been provided,1113,22 no experimental study has clarified the electrophysiologic mechanism of the TAV. In addition, some technical difficulties are associated with the TAV measurements. The TAV should be measured by the orthogonal X, Y, Z leads, and interpretable when patients have a lower noise level, stable R–R interval, and no premature complexes in the 60-beat clusters. Future study is required to test whether the TAV and TWA are independent or complementary markers for risk stratification.

Endocardial repolarization alternans was shown to have good concordance with the TWA measured on the surface ECG.24 Recently, Maury et al.25 examined the beat-to-beat variations in the T-wave morphology before the spontaneous VT/VF onset in ICD-stored intracardiac electrograms. They found that several parameters characterizing the T-wave (amplitude, shape, or duration) varied greater immediately before VT/VF than during the reference. In addition to estimate the ventricular repolarization instability of the baseline using TAV, short-term prediction of VT/VF onset is an important area for the development of future preventive therapeutic options in this setting.

A previous report by Leenhardt et al.3 suggests that patients with short-coupled PVCs have a depressed heart rate variability; however, very limited data are available on the autonomic modulation in RVOT-PVC patients. In this study, we showed that there were no significant differences in the heart rate variability, and heart rate turbulence parameters among the patient groups, suggesting that autonomic modulation may have a lesser influence on the occurrence of PVT/VF in RVOT-PVC patients.

In this study, RFCA was performed in 8 of 10 patients with PVT/VF to eliminate the triggering PVC. We observed that RFCA of one focus was followed by the onset of a new PVC morphology in Patient #4 (Table 2). This finding was consistent with the previous report.5 Indeed, Noda et al.5 have shown that rapid pacing from the responsible RVOT area gives rise to a polymorphic morphological change in the QRS configuration. Therefore, they proposed that functional block and/or delayed conduction by rapid firing arising from a local focus at the RVOT lead to fibrillatory conduction, causing PVT or VF without an organic delayed conduction zone. If PVT/VF is caused by a trigger acting on a substrate with dispersion of the repolarization, RFCA alone which removes the trigger but not the substrate, may not guarantee the long-term prevention of PVT/VF. We also believe that in the future, those RVOT-PVCs in patients who exhibit a greater TAV may guide the ICD implantation.

Study limitations

Our analysis was based on observations from a small sample size at only two institutions. A larger study designed to evaluate the TAV cut-off values for discriminating between benign and possibly lethal RVOT-PVCs would be required. The RVOT-PVCs/MVT may also be observed in the patients with arrhythmogenic right ventricular cardiomyopathy or Brugada syndrome, but we did not rule out these possibilities. The RVOT-PVCs are generally considered as a benign type of arrhythmia and more prevalent. In this study, however, the prevalence of RVOT-PVCs was relatively lower than PVT/VF. This probably represents a referral bias, because patients with PVT/VF are more likely to be hospitalized or referred for aggressive treatments such as RFCA or ICD implantations, whereas RVOT-PVC patients are more likely to be treated conservatively as outpatients and have a lesser chance of undergoing Holter recordings.


The TAV, an analysis of ventricular repolarization instability, may be a useful measure for differentiating patients prone to PVT/VF triggered by RVOT-PVCs from benign RVOT-PVCs or MVT.


This study was supported by the Suzuken Memorial Foundation and JSPS KAKENHI grant numbers 23700544 and 26461094.


The authors thank Ms Akemi Yamauchi for analysing the ECG data and Dr Mari Alford Watanabe and Mr John Martin for preparing the manuscript.

Conflict of interest: none declared.


View Abstract