OUP user menu

Effect of cardiac resynchronization therapy on beat-to-beat T-wave amplitude variability

David Žižek, Marta Cvijić, Jerneja Tasič, Matevž Jan, Sabina Frljak, Igor Zupan
DOI: http://dx.doi.org/10.1093/europace/eus055 1646-1652 First published online: 15 March 2012

Abstract

Aims T-wave amplitude variability (TAV) is a promising non-invasive predictor of arrhythmic events in patients with dilated cardiomyopathy. We aimed to evaluate the effect of cardiac resynchronization therapy (CRT) on native TAV, its relation with left ventricular (LV) reverse remodelling and the occurrence of ventricular tachyarrhythmias (VTs).

Methods and results In this prospective study, we included 40 heart failure patients with left bundle branch block in sinus rhythm (25 male; 16 with ischaemic aetiology; aged 62.7 ± 9.5 years; New York Heart Association class II–IV). Echocardiographic parameters and TAV were evaluated at baseline and 6 months after implantation of CRT device combined with an implantable cardioverter-defibrillator. T-wave amplitude variability was determined by a 20-min high-resolution electrocardiogram Holter recording during native conduction. After TAV assessment, patients were monitored for 15.7 ± 5.2 months for the occurrence of VTs. Decrease in median TAV [from 40.45 μV (24.75–56.00) to 28.15 μV (20.93–37.95), P = 0.004] was observed after 6 months of CRT. However, decrease of median TAV was only noticed in patients with LV reverse remodelling [46.9 μV (27.5–70.0) to 25.8 μV (20.2–32.4), P < 0.001] and in patients without VTs [40.5 μV (27.5–55.9) to 24.4 μV (17.1–31.5), P < 0.001]. Native median TAV > 35.4 µV after 6 months of CRT had an 83% sensitivity and 93% specificity for predicting the occurrence of VTs.

Conclusions Decrease of TAV after CRT is associated with LV reverse remodelling and indicates a reduction of the intrinsic arrhythmogenic substrate. Median TAV after CRT had a good predicting value for VT occurrence in long-term follow-up.

  • Cardiac resynchronization therapy
  • Reverse remodelling
  • T-wave variability
  • Ventricular tachyarrhythmias

Introduction

Cardiac resynchronization therapy (CRT) has emerged as an effective treatment strategy for patients with advanced heart failure.13 Cardiac resynchronization therapy reduces morbidity and mortality as a pacemaker alone (CRT-P)4 or in combination with an implantable cardioverter-defibrillator (CRT-D).5,6 However, despite several favourable results of this therapeutic option, proarrhythmic and antiarrhythmic effects remain controversial. Several clinical trials have demonstrated a decreased incidence of ventricular tachyarrhythmias (VTs) associated with left ventricular (LV) reverse remodelling7,8 and native electrical remodelling.9 In contrast, some experimental studies and case reports have indicated that increased dispersion of repolarization (DR) after the initiation of CRT could induce malignant VTs.1012 Assessment of DR with beat-to-beat T-wave amplitude variability (TAV) is a promising non-invasive predictor of arrhythmic events in post-myocardial infarction patients with reduced LV systolic function13 or dilated cardiomyopathy.14 The aim of our study was to evaluate the long-term effect of CRT on native TAV, its relation to LV reverse remodelling and occurrence of VTs.

Methods

The study complies with the Declaration of Helsinki. The study protocol was approved by The National Medical Ethics Committee and all patients gave a written informed consent before entering the study.

Study population

In this prospective study, we included 40 patients with standard indications for CRT1 [QRS width ≥120 ms, New York Heart Association (NYHA) class II–IV, LV ejection function (LVEF) ≤35%, including both ischaemic and nonischaemic cardiomyopathy] and successful placement of the device at our centre between September 2008 and January 2010. All patients had clinical and laboratory examinations, medical therapy, chest X-ray, electrocardiographic, and echocardiographic parameters recorded at baseline and 6 months after CRT-D implantation. Exclusion criteria were (i) pacemaker dependency, atrial fibrillation/flutter, or any other non-sinus cardiac rhythm; (ii) inherited channelopathies; (iii) concomitant conditions other than cardiac diseases associated with a higher mortality; (iv) β-blocker agents up-titration or additional antiarrhythmic drugs during follow-up (post-implantation); (v) device malfunction/lead dislodgement; (vi) long-term follow-up <12 months; and (vii) <90% of biventricular pacing.

Device implantation

All patients underwent CRT-D implantation using standard techniques under local anaesthesia. The preferred LV lead position was a lateral or a posterolateral vein. Patients received pulse generators either from Medtronic (Concerto I and II) or Sorin Group (Paradym), and were programmed in the DDD(R) mode. For each patient, atrioventricular (AV) delay15 and interventricular (VV) delay optimization16 was performed at implantation. After optimization, the majority of our patients had sequential interventricular pacing (LV lead first—delay of 24 ms).

Echocardiography methods

Transthoracic two-dimensional echocardiography was performed at baseline and at least 6 months after device implantation. All echocardiographic assessments were made on Vivid Systems Four ultrasound equipment. Standard images were examined by two cardiologists blinded to the study outcomes. Left ventricular end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), and ejection fraction (LVEF) were quantified using manual planimetry of two- and four-chamber views and Simpson's technique.17 Decrease in LVESV ≥15% 6 months after CRT was defined as LV reverse remodelling.18

Electrocardiogram recordings and measurement of T-wave amplitude variability

The electrocardiogram (ECG) recordings were performed using SpiderView digital Holter recorder (ELA Medical, SORIN Group, Paris, France) at a sampling rate of 1000 Hz with a resolution of 2.5 µV. The electrodes were placed in a pseudo-orthogonal X, Y, and Z lead configuration. Electrocardiograms recordings were acquired within 1 week before and 6 months after CRT implantation. All 20-min Holter recordings were performed in the afternoon after regular working hours in our out-patient clinic for pacemakers. To avoid excess daytime interference during recordings patients were in supine position in an isolated environment (room with dim light).19 At 6-month follow-up the devices were temporarily reprogrammed to VVI 40 bpm to allow native conduction. QRS duration was determined with build-in software and verified by experienced cardiologist. T-wave amplitude variability was measured using SyneTVar 3.10b software (ELA Medical, SORIN Group, Paris, France) based on the vector magnitude VM = √X2 + Y2 + Z2, as previously described.13,14,19 T-wave amplitude variability, defined as the average of squared deviations from the mean, was assessed on 60 consecutive sinus beats within a given cluster. Three different filters were applied and clusters were excluded from analysis in case of (i) ventricular or atrial premature beat, (ii) high RR interval variability (heart rate varies >20% around its mean), and (iii) noise level exceeding 10 mV. T-wave amplitude was computed on 10 consecutive 50 ms T-wave segments (T1–T10) following QRS offset (defined as QRS onset +150 ms according to median baseline QRS duration) and expressed in microvolts (Figure 1). All 60-beat clusters within the 20-min recording were tabulated and TAV data were averaged. Median TAV was defined as the median values among segments T1–T10, and max TAV as the maximum value from T1 to T10. A single experienced cardiologist blinded to this study analysed the TAV.

Figure 1

T-wave amplitude variability (in microvolts) was assessed on the 60 consecutive sinus beats within a given cluster. T-wave amplitude variability was computed on 10 consecutive 50 ms T-wave segments (T1–T10) following QRS offset (defined as QRS onset +150 ms according to median baseline QRS duration).

Also heart rate variability (HRV) was obtained from 20-min Holter recording and analysed according to the actual guidelines.20 Owing to short-term recording only frequency-domain indexes [very-low-frequency (VLF), low-frequency (LF), and high-frequency (HF) components] were calculated.

Assessment of ventricular tachyarrhythmias during long-term follow-up

Device interrogations were performed in our device outpatient clinic at implant, 1 and 6 months post-implantation, and every 6 months thereafter. Additional follow-up was made in the case of a device shock or new symptoms onset. Ventricular tachyarrhythmia that occurred in the long-term follow-up after 6 months of biventricular pacing was considered for analysis. Ventricular tachyarrhythmia episodes and appropriateness of antitachycardia pacing (ATP) or shock were validated by two independent electrophysiologists blinded to the study outcome. Inappropriate therapy was excluded from our analysis. Ventricular tachyarrhythmia episodes occurring during the first month after device implantation were considered as a patient's baseline arrhythmic status.

Statistical analysis

The Kolmogorov–Smirnov test was used to verify normal distribution. Normally distributed continuous variables were expressed as means and standard deviations. In non-normal distributed continuous variables, data were expressed as median together with the 25th and 75th percentiles (inter-quartile range). Categorical data were summarized as frequencies and percentages. For comparisons of continuous variables paired and unpaired Student's t-test was used for normally distributed variables and the Wilcoxon matched-pair test or the Mann–Whitney U test for non-normally distributed variables. The data for categorical variables were analysed by using the Fisher's exact test. Correlations were performed by Spearman's rank correlation methods. The sensitivity and specificity for the occurrence of VTs were calculated using cut-off value determined by the receiver-operating characteristic (ROC) curve. For all tests, a two-tailed P value of ≤0.05 was considered statistically significant. Data were analysed using SPSS version 16 (SPSS Inc, Chicago, IL, USA).

Results

Study population characteristics

Our study population consisted of 40 patients, predominantly male (25 patients; 62.5%) with a mean age of 62.7 ± 9.5 years. Ischaemic aetiology was present in 18 patients (45%). A total of 30 patients (75%) were in NYHA class III. Median LVEF was 27.5% (20.0–34.7). All patients had a left bundle branch block with median QRS duration of 150 ms (140–160). There were 12 patients (30%) with a history of ventricular fibrillation or ventricular tachycardia before CRT-D implantation. The LV lead was inserted in the lateral or posterolateral vein in 33 patients (82.5%). All patients received optimal medical therapy, if tolerated. Angiotensin-converting enzyme inhibitors or angiotensin II receptor blocker received 39 patients (97.5%), β-blockers also 39 patients (97.5%), spironolactone 37 patients (92.5%), diuretics 25 patients (62.5%), and amiodarone 3 patients (7.5%). Impact of CRT on native TAV and echocardiographic parameters was evaluated 6 months after CRT-D implantation. After 6 months of biventricular pacing, patients were further followed for 15.7 ± 5.2 months for ventricular tachyarrhytmia (VT) occurrence. In the long-term follow-up two patients died due to cardiac pump failure. No ablation procedures due to VTs were performed.

Effect of cardiac resynchronization therapy on clinical, echocardiographic, electrocardiographic parameters and T-wave amplitude variability

At 6-month follow-up, significant improvements in clinical and echocardiographic parameters were registered (Table 1). There was a significant reduction in QRS duration, whereas no significant differences in heart rates were noticed. Also HRV parameters did not significantly change after CRT [HF: 83.5 ms2 (36.2–123.2) vs. 55.5 ms2 (37.5–114.2); LF: 104.5 ms2 (48.0–251.0) vs. 99.7 ms2 (66.0–301.3); VLF: 485.0 ms2 (401.0–900.0) vs. 407.0 ms2 (206.5–800.5)]. After 6 months of biventricular pacing, lower values of both max TAV (P < 0.001) and median TAV (P = 0.004) were observed. Distribution of median TAV for each of the 10 T-wave segments before and after CRT is presented in Figure 2. Median TAV is unevenly distributed across the repolarization interval. At 6-month follow-up, the highest median TAV values were observed in the central part of the T-wave (TAV 3–5), whereas before CRT higher median TAV values occurred earlier (TAV 2–4). Median TAV is significantly higher in most T-wave segments before CRT when compared with follow-up (TAV1–4, 7–10).

View this table:
Table 1

Clinical, echocardiographic, and electrocardiographic parameters and T-wave amplitude variability at baseline and after 6 months of cardiac resynchronization therapy

Before CRTAfter CRTP value
Functional characteristics
 NYHA class I/II/III/IV0/6/30/47/23/10/0<0.001
 6-MWT (min)330 ± 110379 ± 114<0.001
Echocardiographic characteristics
 LVESV (mL)164.0 (130.5–208.5)140.5 (87.0–190.7)0.001
 LVEDV (mL)220.0 (188.5–267.2)211.0 (156.2–277.7)0.020
 LVEF (%)27.5 (20.0–34.5)35.0 (25.0–44.7)<0.001
Electrocardiographic characteristics
 Heart rate (bpm)65.4 ± 9.763.8 ± 8.70.284
 QRS duration (ms)150 (140–160)140 (130–150)<0.001
T-wave amplitude variability
 Max TAV (µV)94.70 (68.08–137.78)58.95 (51.02–76.00)<0.001
 Median TAV (µV)40.45 (24.75–56.00)28.15 (20.93–37.95)0.004
  • CRT, cardiac resynchronization therapy; 6-MWT, 6-min walking distance; LVEF, left ventricular ejection fraction; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; TAV, T-wave amplitude variability; NYHA, New York Heart Association.

Figure 2

Distribution of the median T-wave amplitude variability for each of the ten consecutive 50 ms T-wave segments (T-wave amplitude variability 1–10) before (blue bars) and after 6 months of cardiac resynchronization therapy (red bars). TAV, T-wave amplitude variability; CRT, cardiac resynchronization therapy.

Left ventricular reverse remodelling and T-wave amplitude variability

At 6-month follow-up, 20 (50%) patients demonstrated significant LV reverse remodelling, whereas 20 (50%) patients did not. There were no significant differences in baseline clinical characteristic and heart failure therapy between patients with and without LV reverse remodelling. Baseline echocardiographic, ECG parameters, and TAV between these two groups were similar. In group of patients with LV reverse remodelling, there was significant improvement in echocardiographic parameters after 6 months of CRT [LVEF: 27.0% (20.0–33.5) vs. 43.5% (35.3–50.0), P < 0.001; LVESV: 171.5 mL (130.5–213.0) vs. 109.0 mL (74.0–153.8), P < 0.001; and LVEDV: 226.0 mL (188.5–275.8) vs. 183.5 mL (146.5–266.0), P = 0.001]. Conversely, patients without reverse remodelling had no improvement in these parameters [LVEF: 29.0% (20.5–35.0) vs. 27.5% (20.25–35.0), P = 0.649; LVESV: 156.5 mL (120.0–190.0) vs. 170.0 mL (124.3–224.0), P = 0.390; and LVEDV: 219.5 mL (184.3–261.5) vs. 227.5 mL (206.8–295.8), P = 0.140]. In addition, significant decrease of TAV [max TAV: 94.5 µV (69.8–128.6) vs. 54.5 µV (45.0–72.0), P < 0.001; median TAV: 46.9 µV (27.5–70.0) vs. 25.8 µV (20.2–32.4), P < 0.001] was observed only in group with LV reverse remodelling. In group without LV reverse remodelling, there were no significant changes in TAV [max TAV: 89.15 µV (62.1–144.4) vs. 66.3 µV (53.4–91.0), P = 0.053; median TAV: 35.3 µV (21.6–48.0) vs. 31.7 µV (22.5–48.1), P = 0.852].

There was a significant positive correlation between absolute change of median TAV and absolute change of LVESV (ρ = 0.65; P < 0.001). Furthermore, absolute change of median TAV was positively correlated with absolute changes of LVEDV (ρ = 0.55; P < 0.001) and negatively correlated with absolute change of LVEF (ρ = −0.51; P = 0.001).

Ventricular tachyarrhythmias and T-wave amplitude variability

During the long-term follow-up period (15.7 ± 5.2 months) after TAV assessment, 12 patients (30%) experienced VTs with appropriate therapy. Eleven patients had ventricular tachycardia (six ATP alone and five shocks) and one patient had ventricular fibrillation (shock). One patient experienced a ventricular tachycardia 1 month after CRT implantation.

No significant differences in clinical characteristics (age, aetiology, NYHA class, 6-min walking distance), heart failure, and antiarrhythmic therapy were observed between patients with and without VTs (Table 2). Baseline heart rate, QRS duration, HRV parameters, and echocardiographic parameters in both groups were similar. In the group of patients who experienced VTs LVEF was 25.0% (20.0–30.0) and 29.0% (20.3–35.0) in group without VTs (P = 0.39), LVEDV was 255.0 mL (209.0–373.5) vs. 217.0 mL (185.8–257.3) (P = 0.09), LVESV 185.0 mL (154.3–270.0) vs. 158.0 mL (122.6–199.0) (P = 0.08). The comparison of baseline TAV in patients with and without VTs showed no differences [max TAV: 95.6 µV (68.1–141.9) vs. 91.3 µV (67.6–136.4), P = 0.87; median TAV: 46.9 µV (27.5–70.0) vs. 40.5 µV (27.5–55.9), P = 0.75].

View this table:
Table 2

Baseline characteristics of patients with and without occurrence of ventricular tachyarrhythmias in long-term follow-up

No VTs (n= 28)VTs (n = 12)P value
Age (years)62.5 ± 10.263.3 ± 8.10.797
Male individuals (%)15 (53.6)10 (83.3)0.152
Ischaemic cardiomyopathy (%)10 (35.7)8 (66.7)0.094
Secondary prevention (%)8 (28.6)4 (33.3)1.00
NYHA class I/II/III/IV0/6/20/20/0/10/20.174
6-MWT (min)335 ± 109316 ± 1150.633
Heart rate (bmp)65.8 ± 10.664.4 ± 7.40.644
QRS duration (ms)150 (140–160)160 (142–175)0.320
ACE/ARB (%)28 (100)11 (91.7)0.300
β-Blockers (%)28 (100)11 (91.7)0.300
Spironolactone (%)25 (89.3)12 (100)0.541
Diuretics (%)15 (53.6)10 (83.3)0.152
Amiodarone (%)1 (3.6)2 (16.7)0.209
  • 6-MWT, 6-min walking distance; ACE, angiotensin-converting enzyme inhibitors; ARB, angiotensin II receptor blocker; VTs, ventricular tachyarrhythmias; NYHA, New York Heart Association.

Patients without VTs in long-term follow-up showed significant improvement in LV volumes and LVEF (Table 3). Conversely, patients who experienced VTs had no changes in echocardiographic parameters 6 months after CRT implantation. Moreover, in patients without VTs TAV significantly decreased after 6 months of biventricular pacing, whereas no changes were observed in patients with VTs (Table 3). However, there were no significant changes after 6 months in HRV parameters neither in group with VTs nor in group without VTs.

View this table:
Table 3

Echocardiographic characteristics and T-wave amplitude variability at baseline and after 6 months of cardiac resynchronization therapy in patients with and without occurrence of ventricular tachyarrhythmias in long-term follow-up

No VTs (n = 28)VTs (n = 12)
Before CRTAfter CRTP valueBefore CRTAfter CRTP value
LVESV (mL)158.0 (122.6–199.0)115.5 (72.3–159.3)<0.001185.0 (154.3–270.0)213.5 (170.0–252.5)0.58
LVEDV (mL)217.0 (185.8–257.3)195.5 8132.0–225.5)0.002255.0 (209.0–373.5)265.5 (231.0–363.3)0.51
LVEF (%)29.0 (20.3–35.0)40.5 (35.0–45.0)<0.00125.0 (20.0–30.0)25.0 (18.8–30.0)1.00
Max TAV (µV)91.3 (67.6–136.4)54.5 (47.8–72.0)<0.00195.6 (68.1–141.9)73.5 (56.8–103.4)0.31
Median TAV (µV)40.5 (27.5–55.9)24.4 (17.1–31.5)<0.00139.5 (21.6–57.2)45.3 (36.2–52.4)0.53
  • CRT, cardiac resynchronization therapy; LVEF, left ventricular ejection fraction; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; TAV, T-wave amplitude variability; VTs, ventricular tachyarrhythmias.

Figure 3 shows an ROC curve of native median TAV after 6 months of biventricular pacing to predict the occurrence of VTs (area under the curve = 0.857, 95% CI 0.71–1.00). A median TAV of 35.4 µV had a sensitivity of 83% and a specificity of 93%.

Figure 3

Receiver-operating characteristic curve analysis on native median T-wave amplitude variability at 6 months of biventricular pacing and occurrence of ventricular tachyarrythmias with good predicting value (area under the curve = 0.875). Small numbers next to the line indicate value of median T-wave amplitude variability in microvolts.

Discussion

Our study showed a decrease of native TAV after 6 months of biventricular pacing. However, decrease of this DR index was only observed in patients with LV reverse remodelling. In addition, decrease of TAV after CRT was associated with decreased incidence of VTs in long-term follow-up.

In patients with advanced heart failure there are several arrhythmogenic substrates that includes increased DR.21 Among several methods assessing DR, QT interval variability22 and particularly microvolt T-wave alternans (TWA) could be useful in predicting life-threatening arrhythmias.23 However, because the spectral algorithm for the TWA detection (a repeating sequence in T-wave morphology) needs a relatively high sustained rate of 100–110 bpm through a controlled exercise or with cardiac pacing, its applicability could be limited in patients with advanced heart failure and high doses of β-blocking agents.24 TWA could also be obtained from Holter recordings,25 but stable periods of sustained TWA are rarely found in resting conditions. Repolarization variability is a more common finding in heart failure and it could be easier to identify in Holter recordings.22 As the majority of the patients included in our study had heart failure on optimal therapy and did not have a pacemaker before CRT, we used a method focusing on the variability of T-wave amplitude in the scalar ECGs, recently reported by Couderec et al.13 In our study, we used similar parameters for TAV measurement, apart from extended QRS offset (QRS onset +150 ms), as all our patients had a wide QRS (median duration 150 ms). Focusing on intrinsic myocardial electrophysiology and to avoid potential fusion between intrinsic and biventricular-paced beats; TAV at 6-month follow-up was measured during native conduction. An altered pathway of ventricular activation induced by biventricular pacing develops repolarization changes following resumption of sinus rhythm, known as cardiac memory effect.26 Cardiac memory develops and reaches steady state within 3 weeks of right ventricular pacing and 60 days of biventricular pacing.27 These repolarization changes completely resolve within hours to months, depending on the duration of pacing. However, this cardiac memory is believed to be produced by ventricular electrical remodelling.28

A number of studies have been published in an attempt to interpret the arrhythmogenic effect of CRT. A few smaller studies showed that biventricular pacing was associated with decrease of different DR parameters.29,30 In contrast, reversal of the normal direction of activation during LV epicardial pacing could augment intrinsic transmural heterogeneity of repolarization and potentially contribute to the occurrence of VTs.10,12 Furthermore, sequential biventricular pacing (LV lead first) could have an effect on ECG markers of DR.31 There is lack of data to propose that this early pro-arrhythmic effects persist during long-term pacing. Our study showed a decrease of native TAV after 6 months of biventricular pacing. Focusing on intrinsic myocardial electrophysiology, our results suggest that native decrease of DR after CRT could play a role in the amelioration of the arrhythmogenic substrate. Recent studies which report evidence of native electrical remodelling support present findings.9,32 Furthermore, experimental studies have demonstrated that CRT partially restores heart failure-induced ion current changes, consequently reducing action potential duration heterogeneity and improving intraventricular conduction.33

The relationship between LV reverse remodelling and different ventricular repolarization indexes after CRT is unclear. Our results indicate that there is a strong correlation between improvement of LV function and intrinsic TAV. A significant decrease of TAV was found only in patients with LV reverse remodelling. Our findings reinforce a similar repolarization-based study by Lellouche et al.,34 who reported a marked reduction in DR parameters (Tp-e, QTc) after 1 year of CRT in patients with LV reverse remodelling. In contrast to their results, patients without LV reverse remodelling in our study did now show an increase of intrinsic TAV. This discrepancy could be attributed to the effect of LV epicardial pacing,1012 as all measurements of TAV in our study were obtained during native conduction.

Our study results showed that only patients without VTs demonstrated LV reverse remodelling and significant decrease of DR. Therefore, antiarrhythmic effects of CRT seem to be associated with LV reverse remodelling. Similar conclusions were made in the extension phase of CARE-HF trial, where it was speculated that improvement of cardiac function contributed to the beneficial arrhythmic status.4 Moreover, the retrospective analysis of the InSync ICD Registry7 and InSync III Marquis Study,8 demonstrated a significant reduction of VTs in the responders to CRT (defined as ≥10 and ≥15% decrease in LVESV, respectively). However, relatively low sensitivity of this echo-based cut-off value for the prediction of VTs would implicate that response to CRT reduces but does not eliminate the occurrence of life-threatening arrhythmic episodes. In a similar study, patients with CRT-D for primary indication, whose LVEF improved above 35%, were at low risk for VTs beyond the first year.35 In addition, only patients with LV reverse remodelling demonstrated significant reduction in VT occurrences after up-grading from an ICD to a CRT-D device.36

In our study, decrease of TAV was only observed in the patient group without VTs. Furthermore, our study results seem to provide additional evidence in support of the potential clinical value of this repolarization-based methodology: a median TAV value of >35.4 µV at 6 months of biventricular pacing after ROC curve analysis had a good predicting value (sensitivity 83%, specificity 93%) of VT occurrence in long-term follow-up. T-wave amplitude variability cut-off value obtained in our study is similar to that proposed by Ribeiro et al.37 and Extramiana et al.38 and lower than that used by Couderec et al.13 Differences in study population characteristics (non-ischaemic dilated cardiomyopathy patients, wide QRS) may account for the different values.

Although this study provides new information regarding the course of the intrinsic cardiac electrophysiology after CRT, the relationship and the timeline between LV reverse remodelling and ventricular repolarization require further investigation. However, if TAV is strongly linked to the arrhythmic events, it could provide a basis for possible downgrading of a CRT-D to a CRT-P at the time of battery depletion in an individual patient with CRT-D for primary prevention. Further prospective studies in larger patient cohorts and longer follow-up are warranted to assess this issue.

Conclusion

Decrease of native TAV after 6 months of CRT indicates a reduction of the intrinsic arrhythmogenic substrate associated with depressed LV function. Decrease of TAV was detected only in patients with LV reverse remodelling and without VTs. Median TAV after 6 months of CRT had a good predicting value for the VT occurrence in long-term follow-up.

Study limitations

This study has some limitations. The relatively small number of patients and lack of randomization with placebo group limits the strength of our findings. Moreover, our mean follow-up after TAV assessment was relatively short (15 months) and it is conceivable that with longer follow-up more patients would experience VTs.

Transient repolarization instability following the re-initiation of biventricular pacing12 after assessment of native TAV (at 6-month follow-up) may have had an impact on VT occurrences. However, a relatively short time of CRT cessation probably did not induce significant repolarization changes to be clinically relevant in our patients. Furthermore, after interrogation of the pacemaker memory data, no VTs were recorded immediately after re-initiation of biventricular pacing.

Repolarization-based methodology of TAV assessment has been reported only recently and applied in few other clinical conditions. The method needs more extensive evaluation to define parameters of TAV measurement (number of clusters, T-wave window, QRS offset, noise level) and predictive values.

Conflict of interest: none declared.

References

View Abstract