Cardiac resynchronization therapy reduces T-wave alternans in patients with heart failure

Introduction

Systolic heart failure describes a state of weak myocardial contractility and low left ventricular ejection fraction (LVEF), resulting in an increased incidence of arrhythmias and decompensated heart failure necessitating hospitalization. Methods that can predict myocardial electrical instability and mechanical deterioration are needed so that patients can be better managed. Recently, T-wave alternans (TWA) has been recognized as a predictor for sudden cardiac death, arrhythmic events, and mortality (sudden and non-sudden cardiac death) in patients with heart failure.^1–7

T-wave alternans is an electrocardiogram (ECG) phenomenon with beat-to-beat alternation in the morphology and/or amplitude of the T-wave during a normal rhythm.^6–10 T-wave alternans is the result of alternations in repolarization of the action potential duration caused by abnormalities in intracellular Ca⁺⁺ cycling (Ca⁺⁺ release, reuptake, and transport to the junctional sarcoplasmic reticulum).^10–12 Thus, TWA represents myocardial electrical instability, and is associated with an increased incidence of ventricular tachyarrhythmia [ventricular tachycardia (VT)/ventricular fibrillation (VF)].^5,13–17 On the other hand, weak myocardial contractility during heart failure is also primarily caused by abnormal intracellular Ca⁺⁺ cycling, causing mechanical alternans, a sign of acute deterioration of heart failure.^9,18,19 Thus, TWA is potentially associated with not only VT/VF, but also failure of myocardial mechanics. While TWA amplitude increase before VT/VF events in humans has been reported,^{14–16,20,21} a few animal experiments have demonstrated that TWA exists during myocardial mechanical failure.^22–24 Patients with heart failure also show a high incidence of TWA.^3,25–27

Previous studies in implantable cardioverter-defibrillator (ICD) patients have demonstrated that device electrogram (EGM)-based TWA is related to the occurrence of sudden cardiac death.^15,16,20,21 The present study used EGM-based TWA method to determine the impact of cardiac resynchronization therapy (CRT) on TWA and cardiac function in heart failure patients who received CRT-ICD.

Methods

The present study was a single-centre, prospective study with 6-month follow-up post-implantation of CRT-ICD in patients with heart failure. The study protocol was approved by the hospital Investigational Research Board and in compliance with the Declaration of Helsinki and with the laws and regulations in China. All patients completed written informed consent.

Selection of patients

Twenty-seven patients were enrolled between November 2011 and February 2013. The inclusion criteria included: (i) patients had systolic heart failure; (ii) patients were clinically indicated for the therapy with resynchronization ICD (Medtronic Models 7277, 7279, and 7289), e.g. LVEF ≤ 35%, a QRS duration ≥120 ms, New York Heart Association (NYHA) Classification III or IV, and under optimal heart failure therapy with medications; (iii) the baseline rhythm was sinus with 1 : 1 atrioventricular (AV) conduction; and (iv) patients were 18 years of age or older. Patients were excluded if (a) measurement of TWA was not feasible (e.g. permanent atrial fibrillation, frequent premature ventricular contractions); (b) patients had a medical condition that would limit study participation as assessed by physicians; (c) patients would be inaccessible for follow-up at the study centre; (d) patients were enrolled or intended to participate in other clinical studies during the study period that would confound the results of the present study; and (e) patients were pregnant.

Study procedures

Device data collection

The first data collection occurred shortly after device implantation, and the second data collection was conducted at 3 months after CRT. All enrolled patients received CRT-ICD implantation and stayed in the hospital for 7 days post-implant during which CRT was not turned on. Five to seven days after the implantation of CRT-ICD, the device was interrogated with Medtronic Programmer 2090 and the lead configurations of two-channel device EGMs were programmed for left ventricular (LV) near-field EGM using LVtip-to-LVring of LV lead and for far-field EGM using device Can-to-RVcoil. The LV near-field EGM was used to assess the possibility of detecting TWA in a local LV region. The device EGM range was programmed at 2 or 4 mV as described previously.^16,21 The programmer was connected with a data acquisition system (DATAQ Instruments) for EGM data recording at a sampling rate 900 Hz. Then, each patient underwent TWA tests using AAI pacing and CRT pacing (DDD mode and AV delay: paced AV interval, 137.0 ± 27.3 and sensed AV interval: 108.1 ± 21.8 ms) with a rest interval of 5 min between AAI and CRT pacing tests. The procedures of EGM data collection and TWA pacing tests at 3-month follow-up were identical to those at baseline.

T-wave alternans pacing tests

After the device was interrogated and two-channel EGM recording was in place, data collection was performed in each patient for: (i) a baseline of 5 min, (ii) AAI up-titration pacing rate from 90 to 105 b.p.m. with 90 s for each pacing rate, (iii) 5 min of rest, and (iv) CRT up-titration pacing rate from 90 to 105 b.p.m. with 90 s for each pacing rate. Subsequently, EGM data at each pacing rate of each pacing mode were analysed offline for TWA in the frequency domain.

T-wave alternans measurements

Clinically, surface ECG TWA is measured in the frequency domain by applying the fast Fourier transform.^1,2,8 To measure TWA in device EGMs, a custom software program using Matlab (Release 7.0 version 14, Matlab Inc.) was developed and validated using the fast Fourier transform based on the public domain⁸ and as described previously.²⁸ Briefly, the ventricular sensing or pacing marker was used as a fiducial point of QRS complex. T-waves within the 90 s TWA pacing test were measured during a 150–200 ms wide interval after the fiducial point of each QRS complex. The amplitudes of corresponding points of consecutive T-waves were measured to construct a beat series. The power spectrum of this beat series was then calculated using the fast Fourier transform. T-wave alternans was indicated at the peak of the spectrum at 0.5 cycles/beat.^8,28 The magnitude of TWA (μV) was then calculated as the square root of the alternans power. The K score was defined as the ratio (signal-to-noise ratio) of the alternans power divided by the standard deviation of the power in the noise band. T-wave alternans test was positive only when its K score was ≥3 and the TWA amplitude was documented. Otherwise, negative TWA test result was assigned if the K score was <3. T-wave alternans pacing test results were discarded if EGM quality was inappropriate for TWA analysis, which was mostly caused by inconsistent capture, loss of 1 : 1 AV conduction.

Clinical data

Patients' clinical data of cardiac disease history, echocardiogram (Echo), NYHA classification, ECG, and medications were collected at baseline and 3-month follow-up visit. Another follow-up at 6-month post-implant was conducted to assess cardiac events including arrhythmic events with appropriate ICD therapy and the mortality.

Data analysis

Data are presented as mean ± SD or percentage wherever applicable. The number or percent of positive TWA tests was compared using the method of Chi-square test between AAI pacing test and CRT pacing test and between baseline test and test at 3-month follow-up, respectively. Paired or unpaired t-test was used to compare numeral values at baseline and follow-up and EGM TWA amplitudes by different pacing modes, lead configurations, and test timepoints. A statistically significant difference required a probability of P < 0.05.

Results

Patient clinical data

Major clinical characteristics of the 27 enrolled patients are summarized in Table 1. All patients were indicated for CRT-ICD with low LVEF (27.3 ± 5.8) and wide QRS duration (158.9 ± 32.9 ms). The large LV end-diastolic volume (mean 265 mL, Table 1) was indicative of dilated heart failure decompensation. All patients underwent optimal medications with diuretics, beta-receptor blockers, angiotensin-converting enzyme inhibitors, and/or angiotensin II receptor blocker. Six patients (22.2%) had a history of VT/VF prior to the enrolment. During follow-up post-CRT, two patients died from complications of decompensated heart failure.

Overall T-wave alternans tests

Each patient underwent TWA AAI and CRT pacing tests at two timepoints (baseline and 3-month follow-up visit), e.g. AAI pacing test at 90 b.p.m. followed by AAI pacing test at 105 b.p.m., and CRT pacing test at 90 b.p.m. followed by CRT pacing test at 105 b.p.m. Thus, a total of 108 TWA tests at baseline and 108 TWA tests at the 3-month follow-up were performed. Eight TWA tests at the baseline and 14 TWA tests at the 3-month follow-up were excluded for analysis because of loss of follow-up or inappropriate EGM recordings due to fast heart rate, inconsistent capture, or AV block during the pacing. Figure 1 shows an example of EGM recordings and corresponding spectral TWA measurements during AAI pacing test and CRT pacing test at 105 b.p.m. Visible EGM TWA with beat-to-beat alternation in T-wave peak-to-peak amplitude can be seen in this example (Figure 1).

Figure 1

Example of device-recorded EGMs and corresponding spectral TWA measurements at baseline pacing tests. Top: EGM and 102 µV TWA during AAI pacing at 105 b.p.m. Bottom: EGM and 91 µV TWA during CRT pacing at 105 b.p.m. Visible TWA (peak-to-peak T-wave amplitude between two small bars) presents in EGM recording during pacing test.

Open in new tabDownload slide

T-wave alternans at baseline and 3-month follow-up visit

AAI pacing test vs. cardiac resynchronization therapy pacing test

Figure 2 illustrates the results of TWA tests at baseline and 3-month follow-up. At baseline, the number of tests yielding TWA was significantly higher for AAI pacing than for CRT pacing either at the pacing rate of 90 or 105 b.p.m. (Figure 2A and E). At 3-month follow-up, there was no significant difference in the number of positive TWA test results between AAI pacing and CRT pacing (Figure 2B and F) though there was a trend of reduction in positive TWA test results by CRT pacing at pacing rate 105 b.p.m. (P = 0.196, Figure 2F). Of the 29 positive TWA tests by CRT pacing (combining both 90 and 105 b.p.m. pacing tests), 21 were matched with the positive TWA tests by AAI pacing at the same pacing rates in the same patients, a 72.4% concordance with AAI pacing tests. The overall TWA amplitude was 86.1 ± 57.1 µV during AAI pacing test and 71.8 ± 50.3 µV during CRT pacing test (P = 0.179 between two pacing modes).

Figure 2

Comparisons of TWA test results by AAI and CRT pacing at baseline and 3-month CRT. Left panels are for pacing tests at 90 b.p.m. while the right panels are for pacing tests at 105 b.p.m. (A, E) Comparisons between AAI pacing and CRT pacing at baseline (BL); (B, F) at 3-month (3M) CRT; (C, G) between the baseline and 3-month follow-up by AAI pacing tests; and (D, H) between the baseline and 3-month follow-up by CRT pacing tests. P values were obtained by χ² test.

Open in new tabDownload slide

AAI pacing test at baseline and 3-month follow-up

Following 3 months of CRT, the number of positive TWA test results during AAI pacing, either 90 or 105 b.p.m., was significantly reduced compared with that at baseline (Figure 2C and G). Overall, there was a 30% reduction of positive TWA test results for AAI pacing at 90 b.p.m. and 31% reduction for AAI pacing at 105 b.p.m. after 3 months of CRT.

Cardiac resynchronization therapy pacing test at baseline and 3-month follow-up

The number of positive TWA test results by CRT pacing at 90 b.p.m. at baseline was not significantly different from that at 3-month follow-up (Figure 2D). There was a trend of reduction (22% from the baseline) in the positive TWA test results by CRT pacing at 105 b.p.m. at 3-month follow-up visit when compared with the baseline (P = 0.108, Figure 2H).

T-wave alternans detected by different lead configurations

As shown in Figure 3, the number of positive TWA test results at the baseline was higher for LV near-field EGMs (44% of all TWA tests) than for far-field EGMs (30%, P = 0.040, Figure 3A) while the number of positive TWA test results at 3-month follow-up showed no difference between two lead configurations (22.6% for near-field EGMs vs. 21.3% for far-field EGMs, Figure 3B). The number of positive TWA test results detected by the LV near-field EGMs was significantly reduced at 3-month follow-up (22.6%) from the baseline (44.0%, P = 0.002). There was no significant change in the number of positive TWA test results by the far-field EGMs between baseline and at 3 months (Figure 5D). There was no significant difference between the overall TWA amplitude in LV near-field EGMs (87.2 ± 59.1 µV) and that in far-field EGMs (74.3 ± 48.1 µV, P = 0.213 between two EGMs).

Figure 3

Comparisons of the percentage of patients with positive TWA detected by near-(LVtip–LVring) and far-field (Can–RVcoil) lead configurations. Test results of positive TWA are pooled together for both AAI pacing and CRT pacing. (A) Per cent of positive TWA test results between near-field vs. far-field EGMs at the baseline; (B) near-field vs. far-field at 3-month CRT; (C) comparison of TWA in near-field EGM between baseline and 3-month CRT; and (D) comparison of TWA in far-field between baseline and 3-month CRT.

Open in new tabDownload slide

T-wave alternans and the assessment of cardiac function

Left ventricular ejection fraction and NYHA classification were significantly improved after 3 months of CRT (LVEF: from 27.3 ± 5.8 to 35.9 ± 10.5, P < 0.001; NYHA: from 2.8 ± 0.6 to 1.6 ± 0.6, P < 0.001), with a 31.5% improvement in LVEF and average reduction of 1.2 in NYHA classification. The LV end-diastolic volume significantly decreased from baseline 265.4 ± 92.1 to 212.0 ± 82.4 mL (P = 0.015), suggesting an LV reversal remodelling. Figure 4 shows LVEF in individual patients based on TWA test results. Left ventricular ejection fraction at baseline was significantly higher (32.8 ± 4.8%) for patients who had negative TWA test by AAI pacing than those who had positive TWA test (26.1 ± 5.0%, P = 0.013, Figure 4A). At 3-month follow-up, there was a trend of higher LVEF (40.1 ± 12.0%) for patients who had negative results of TWA tests by AAI pacing than that for patients who had positive results of similar TWA tests (32.8 ± 6.0%; Figure 4A). Furthermore, the percent improvement in LVEF after 3 months of CRT over the baseline value was higher for patients whose baseline TWA test results were negative than those whose baseline TWA test results were positive (Figure 4B). There was no correlation between the amplitude of TWA and LVEF. Two patients who had TWA and low LVEF (25 and 20%) at baseline died of heart failure deterioration, and one patient who also had TWA and low LVEF (23%) at baseline received a heart transplant. Two of these three patients had positive TWA test results at 3-month follow-up and one patient died before the 3-month follow-up.

Figure 4

Association of TWA test results with LVEF. (A) The ordinate gives the value of LVEF (%); from the left to right are individual measurements of LVEF in patients without and with TWA at the baseline (BL) and at 3-month (3M) CRT. (B) The ordinate is the percent changes in LVEF after 3-month CRT from the baseline based on TWA test results. The abscise is the change of TWA test results from baseline to (→) 3-month CRT (N, no TWA; and P, positive TWA).

Open in new tabDownload slide

T-wave alternans and cardiac events at baseline and follow-up

Six patients experienced VT/VF events prior to enrolment. Five of these six patients had TWA and one did not have TWA at the baseline. Of these six patients, five did not have VT/VF during the follow-up period, while one patient died of heart failure (Figure 5). Three patients who did not have a history of VT/VF before enrolment experienced VT/VF episodes with successful ICD therapies during the 6-month follow-up period, and all these three patients had TWA at baseline and 3-month follow-up visit. Thus, of 21 patients who had positive TWA tests, 9 patients (42.9%) had either a history of VT/VF events or VT/VF events with ICD therapies during the 6-month follow-up period, and 8 of these 9 patients (88.9%) who had VT/VF events had positive TWA during pacing tests. In these nine patients who either had a VT/VF history at the enrolment or experienced VT/VF events during the follow-up, TWA at the baseline was detected by the near-field EGM in eight patients (88.9%) and by far-field EGM in four patients (44.4%, P = 0.046 vs. the near-field EGM).

Figure 5

The number of patients with TWA test results, the VT/VF incidence and mortality at baseline, 3-month, and 6-month follow-up. *One patient without appropriate EGM at 3 months; ^†Same patient.

Open in new tabDownload slide

Discussion

Main findings

Consistent with previous reports that TWA of device EGM can be measured by spectral analysis of the fast Fourier transform and a predictive measure of the risk for VT/VF occurance,^20,28,29 the present study also showed an association between TWA and the occurrence of VT/VF events. Furthermore, the present study demonstrates two phases, acute and chronic, of the reduction of TWA by CRT in patients with heart failure and an association between TWA and mechanical function in terms of LVEF. In addition, the LV near-field EGM is more sensitive in detecting TWA than the far-field EGM.

T-wave alternans has been thought to represent myocardial electrical instability and be related to the occurrence of VT/VF. Computer simulations and basic research in animals identify spatiotemporal TWA as an immediate causal antecedent of VT/VF.^30–33 T-wave alternans occurrence or TWA amplitude increase occurring before VT/VF events in humans have been reported.^{14–16,20,21} While many clinical studies have been performed to establish ECG TWA test as an independent predictor of long-term risk of VT/VF events,^1–9 the potential value of TWA assessed in implantable-device EGMs in ICD patients is as an immediate antecedent of VT/VF. Christini et al.²⁹ observed TWA of unipolar EGMs recorded from temporary electrodes in patients who underwent an electrophysiological study. Armoundas et al.²⁰ reported TWA immediately prior to an event of VT in the device EGM in one patient with an ICD. Recently, the multi-centre study by Swerdlow et al.¹⁶ demonstrates that high-amplitude TWA of device EGMs could yield a sensitivity of 61% of the predictive power for VT/VF event prediction in ICD population, demonstrating a temporal relationship between TWA and the occurrence of VT/VF in patients with a high risk of sudden cardiac death. The present study also demonstrates an association of positive TWA test results with a history of VT/VF incidence and with VT/VF occurrence during the follow-up period. However, the present study was not designed to record device EGMs right before each VT/VF episode. Thus, it is unknown whether TWA amplitude of the present study increased prior to VT/VF events as observed in the previous study.^{14–16,20,21}

In the present study, the finding of lower LVEF in patients with TWA than in those without TWA suggests that TWA may be associated with the process of mechanical instability of heart failure. Three patients who died of heart failure or received heart transplant had TWA and low LVEF. Previous animal experiments have demonstrated that TWA in large amplitude exists during myocardial mechanical failure.^22–24 Human clinical observations also confirmed existence of TWA related to heart failure, especially in patients with cardiomyopathy.^3,25–27 The mechanism of the association of TWA and the progression of heart failure is not well established. However, it has been well recognized that the underlying mechanisms of TWA are associated with abnormal intracellular calcium cycling that leads to alternations in repolarization (electrical alternans) of the action potential.^10–12 On the other hand, the abnormal intracellular Ca⁺⁺ cycling is also associated with weakening myocardial contractility in heart failure, leading to alternative weakening contraction.^18,19 Therefore, TWA is potentially related to not only the risk of the occurrence of ventricular tachyarrhythmia, but also with progression of worsening myocardial mechanics.

Cardiac resynchronization therapy pacing in the present study reduced the number of positive TWA test results, which is consistent with a previous report.³⁴ Reduction of TWA by CRT in the present study occurred in two phases, e.g. the acute phase at the baseline CRT pacing and the chronic phase after 3 months of CRT. As shown in Figure 2, there is approximately a 27% reduction in positive TWA test results by CRT pacing from 77% positive TWA test results by AAI pacing at baseline. A further reduction of positive TWA test results by either AAI pacing or CRT pacing was observed after 3 months of CRT, suggesting a reverse remodelling of myocardial electricity by CRT. Furthermore, there was a 31.5% increase in LVEF after 3 months of CRT in the present study, indicating the reserve mechanical remodelling. The reverse remodelling post-long-term CRT may be caused by CRT-related improvement in ventricular function with a possible concurrent improvement in intracellular calcium cycling. The potential explanation for acute-phase reduction of TWA by CRT pacing in the present study may be related to the resynchronization effect of CRT, i.e. some of positive TWA test results in CRT-indicated patients might be related to dysschronization and/or an altered pattern of action potential propagation that could be corrected instantly by CRT pacing, leading to an acute reduction in TWA. More research is needed to confirm this theory.

Clinical TWA tests have been conducted using body surface ECG as essentially a far-field cardiac signal. Previous studies in which TWA of device EGM was assessed mostly used far-field lead configurations.^16,20,28 The present study used the LV lead to collect EGM generated in the LV and observed more positive TWA test results than in far-field EGM collected by device Can-to-RVcoil lead configuration. It is unknown whether TWA in near-field EGM is truly a local TWA in the region surrounding the recording electrodes, or if the locally detected TWA represents part of a large ventricular area that has TWA not detected by far-field EGM. This finding interestingly raises the possibility of local TWA and thus regional ventricular instability that can be detected with most sensitivity by a local electrode. More research is warranted to clarify the clinical relevance of locally detected TWA.

Limitations

First, the sample size of the present study is relatively small with 27 CRT-ICD patients. Four patients developed VT/VF incidence during the 6-month follow-up period—this small number of VT/VF occurrences cannot allow for an accurate determination of the sensitivity and specificity of using TWA for prediction of VT/VF events. Furthermore, the usefulness of TWA for VT/VF event prediction has been confirmed in several studies in ICD patients. The present study lacks an ICD control group to compare the predictive value of TWA between CRT-ICD patients and ICD-only patients. Moreover, while the present study demonstrates an association of pacing-induced TWA with VT/VF events in history or during follow-up, the study does not address the temporal relationship between TWA and VT/VF events, e.g. whether TWA immediately precedes spontaneous VT/VF as observed in the previous studies.^15,16,20,21 In addition, in the present study, the reduction in TWA and the improvement in LVEF after CRT are used to suggest reverse electrical and mechanical remodelling of a failing heart in CRT-ICD patients. Such a correlation between reverse remodelling and the reduced TWA and improved LVEF by CRT needs to be further confirmed in basic mechanism investigation and clinical studies with longer follow-up and more parameters relating to reverse remodelling.

Conclusions

T-wave alternans has been recognized as a predictive measure of either short-term or long-term occurrence of ventricular tachyarrhythmia. The present study demonstrates an association of TWA with cardiac electrical and mechanical events and reduction of TWA by acute and long-term CRT that leads to improvement of LVEF and NYHA classification, suggesting reverse electrical and mechanical remodelling by CRT. With advancements in the technology of implantable devices, device-based TWA assessment is possible, which can potentially be used to predict risk of ventricular arrhythmia, and to assess the progression of heart failure and guide medical therapy.

Conflict of interest: X.Z., J.C., and C.L. are employees of Medtronic that makes pacemaker products, and they receive salary from the company and hold company stocks. All other authors have no conflicts to declare.

Funding

This work was partly supported by a research grant of Medtronic (Shanghai) Inc. Shanghai, China. X.Z., J.C., and C.L. are Medtronic employees and had a role in study design, data collection, and analysis.

References