Time-dependent risk reduction of ventricular tachyarrhythmias in cardiac resynchronization therapy patients: a MADIT-RIT sub-study

Abstract

Aims

Data on the time-dependent benefit of cardiac resynchronization therapy with defibrillator (CRT-D) compared with a dual-chamber implantable cardioverter-defibrillator (ICD) to reduce death or ventricular tachycardia (VT) or ventricular fibrillation (VF) are limited. We aimed to evaluate the time-related risk of death or sustained VT or VF in patients receiving CRT-D vs. ICD in the MADIT-RIT trial.

Methods and results

Kaplan–Meier survival analyses and multivariate Cox regression models were utilized to compare the incidence and the risk of death or sustained VT/VF in the CRT-D and ICD subgroups by the elapsed time after device implantation (6 months). Of the ICD (n = 742) and CRT-D (n = 757) patients enrolled, the risk of death was lower in CRT-D vs. in ICD early after device implantation [hazard ratio (HR) = 0.42, 95% confidence interval (CI): 0.17–1.03, P = 0.058] and beyond 6 months of follow-up (HR = 0.39, 95% CI: 0.21–0.73, P = 0.004), with the 6-month interaction P = 0.899. The overall risk of sustained VT/VF was reduced in CRT-D vs. ICD patients (HR = 0.73, 95% CI: 0.52–1.03, P = 0.07). However, the risk was similar in the first 6 months (HR = 1.00, 95% CI: 0.62–1.62, P = 0.988), and a lower risk emerged 6 months after CRT-D implantation (HR = 0.58, 95% CI: 0.38–0.88, P = 0.011), with the 6-month interaction P = 0.059.

Conclusion

The reduced mortality risk of CRT-D compared with an ICD alone began early after device implantation and was sustained during long-term follow-up; the reduced risk for ventricular tachyarrhythmias did not emerge until 6 months after device implantation.

Clinical trial registration

http://clinicaltrials.gov/ct2/show/NCT00947310.

Introduction

Over the past decade, cardiac resynchronization therapy (CRT) has been established as an effective treatment for patients with heart failure (HF), left ventricular (LV) dysfunction, and a wide QRS.^1,2 Cardiac resynchronization therapy has been shown to improve cardiac function,³ to induce LV reverse remodelling,⁴ to alleviate HF symptoms,¹ and to improve survival.² Cardiac resynchronization therapy also reduces the risk of ventricular tachycardia (VT) or ventricular fibrillation (VF),^5,6 and the reduction is related to the degree of LV reverse remodelling.^7–9 Left ventricular reverse remodelling was recorded at 3 months after CRT pacing.⁴ However, the temporal association of risk reduction with CRT for mortality and for VT or VF has not yet been investigated.

We investigated in this MADIT-RIT sub-study, the chronology of the risk reduction in death and in VT/VF episodes in patients with a CRT-D vs. ICD implant. We pre-specified that LV reverse remodelling through CRT-D predominantly occurs within a 6-month time interval after CRT-D implantation. The aim of this study was to determine, if the beneficial reduction in death and in ventricular tachyarrhythmias was similar or different before and after this 6-month time interval following CRT-D vs. ICD implantation.

Methods

MADIT-RIT

The design¹⁰ and primary endpoint results of the MADIT-RIT study have been published previously.¹¹ In short, this randomized multicentre clinical trial investigated the comparative effectiveness of different ICD programming strategies for the reduction of inappropriate ICD therapy and found that innovative ICD programming with high-rate therapy cut-off therapy or long delay therapy is associated with a significant reduction in inappropriate therapy.¹¹

Patient population

The MADIT-RIT trial enrolled 1500 patients, aged at least 20 years, with ischaemic or non-ischaemic systolic HF. All patients met the guideline criteria for primary prevention implantation of a dual-chamber ICD (DDD ICD) or CRT-D.^12,13 Patients were excluded from the trial if they had permanent atrial fibrillation, if they had a history of sustained VT or VF, if they had interventional or surgical coronary revascularization, or had an enzyme-defined myocardial infarction within 3 months before enrolment. Further exclusion criteria comprised second- or third-degree atrio-ventricular block, New York Heart Class IV, and the foreseeable future need of coronary revascularization. Patients were to be taking stable guideline-conforming medication for their cardiac condition, including a beta-blocker, an angiotensin-converting enzyme (ACE)-inhibitor, or an angiotensin receptor-blocking agent.

Device programming and follow-up schedule

In MADIT-RIT, the three programming arms differed in the rate cut-off for treatment of a VT/VF episode and the required duration of an episode with treatment rendered. Arm A involved a VT zone ≥170 bpm [detection delay 2.5 s, treatment with adenosine triphosphate (ATP) and shocks] and a VF zone ≥200 bpm (detection delay 1 s, treatment with ATP and shocks). In contrast, Arm B (high-rate cut-off arm) comprised a monitor only VT zone ≥170 bpm, and only ventricular rates ≥200 bpm (VF zone, delay 2.5 s) were followed by antitachycardia treatment (Quick Convert ATP and shocks). Arm C (long delay arm) included a VT zone ≥170 bpm (detection delay 60 s, treatment with ATP and shocks), a second VT zone ≥200 bpm and <250 bpm (detection 12 s, treatment with ATP and shocks), and a VF zone (detection delay 2.5 s, treatment with ATP and shocks).

Anti-bradycardia pacing was recommended to prevent unnecessary pacing in dual-chamber ICD's and to provide maximum resynchronization with biventricular pacing in CRT-D devices. Patients had subsequent planned follow-up visits every 3 months during the first year, and every 6 months during later follow-up. During each visit, a brief interim history was taken, and a physical examination and a comprehensive device interrogation were carried out.

Data were transmitted to the study Coordination and Data Center (CDC) at the University of Rochester, NY, USA, via a coded connection using electronic case report forms. In addition, a disc with the recorded interrogation was sent directly to the CDC via a courier service for further detailed analysis.

Definitions, study endpoints

In this formally pre-specified MADIT-RIT sub-study, patients were divided into subgroups based on their implanted device (DDD ICD or CRT-D). The analysis was performed on an intention-to-treat basis and models were adjusted for the randomized ICD programming arm.

The primary endpoint of the study was all-cause mortality. A pre-specified Mortality Review Committee, consisting of three experienced clinicians, appraised all available information on lethal events during the study from patient records and device recordings.^14,15

The secondary endpoint was a surrogate ventricular tachyarrhythmia endpoint, combining the first lasting non-treated VT (defined as >30 beats of fast ventricular rhythm as detected and recorded by the device), and ICD-treated VT or VF episodes.

The tertiary endpoint was the combination of the first lasting monitored non-treated VT (defined as >30 beats of fast ventricular rhythm as detected and recorded by the device), or ICD-treated VT or VF episode, or death.

The Electrogram and Device Interrogation core laboratory, consisting of experienced physicians and dedicated staff, was charged with the review, analysis, and adjudication of arrhythmia episodes, based on pre-specified criteria.¹⁰

Ventricular arrhythmia episodes were classified as VT or VF according to the electrogram-derived morphology rather than relying on the device definition of VT or VF. The core laboratory formed an independent view on each episode and assorted the events to a final classification (VT or VF, or monitored non-treated lasting VT).

Statistical analysis

Continuous variables are expressed as mean ± SD. Categorical data are summarized as frequencies and percentages. Baseline clinical characteristics were compared between the ICD and CRT-D subgroups, and stratified by treatment arm, using Wilcoxon rank-sum test for continuous variables and χ² test or Fisher's exact test for dichotomous variables, as appropriate.

The univariate endpoint analysis was carried out using the Kaplan–Meier method and log-rank testing for differences between the groups. Multivariate Cox proportional hazards regression analyses were used for multivariate endpoint analyses. Baseline parameters with significant differences between the study groups, and those that were predictive of the specific endpoint were included as covariates into the multivariate analyses, including randomized ICD programming arm.

The multivariate models were adjusted for treatment Arm B vs. Arm A, treatment Arm C vs. Arm A, age at enrolment, gender, ischaemic aetiology of cardiomyopathy, prior myocardial infarction, prior interventional coronary revascularization, New York Heart Association (NYHA) class equal or greater III, diabetes mellitus, heart rate at baseline, and systolic blood pressure at baseline. The interaction of CRT-D treatment and follow-up time of 6 months has been tested, and hazard ratios (HRs) were reported for the first 6 months in the trial and for the follow-up time after the first 6 months, as pre-specified in the analysis.

Adjusted HR's with 95% confidence intervals (CIs) are reported. A two-tailed P-value of <0.05 was considered statistically significant. Interaction P-values were computed and reported. Analyses were carried out with SAS software (version 9.3, SAS institute, Cary, NC, USA).

Results

Baseline clinical characteristics

Relevant clinical characteristics of the MADIT-RIT subgroups with ICD vs. CRT-D are shown in Table 1. Patients with CRT-D implants had lower left ventricular ejection fraction (LVEF) and a higher proportion of NYHA class III compared with conventional ICD patients. Furthermore, CRT-D patients were older, more often females, and had less often ischaemic aetiology of HF, prior myocardial infarction, and prior interventional coronary revascularization. Distribution of ICD vs. CRT-D treatment was balanced among the MADIT-RIT programming Arms A, B, and C (P-value = 0.870) (Table 2).

The proportion of patients treated with beta-blockers and ACE-inhibitors/angiotensin receptor blockers was high in both patients with ICD or CRT-D. In the CRT-D group, beta-blocker treatment was less frequent, whereas the prescription of amiodarone, aldosterone antagonists, and diuretics was more common.

The risk of all-cause mortality in cardiac resynchronization therapy with defibrillator patients compared with implantable cardioverter-defibrillator alone, stratified by the elapsed time after device implantation

During the median follow-up of 17 months, 29 (3.8%) patients died with an implanted CRT-D and 42 (5.7%) patients died with an ICD alone. The cumulative probability of all-cause mortality for CRT-D vs. ICD patients is shown in Figure 1. Patients with an implanted CRT-D had a trend towards a lower incidence of mortality compared with those implanted with an ICD despite the relatively short follow-up time (P = 0.107). Non-cardiac mortality was similar in CRT-D vs. ICD study subgroups (CRT-D: 9 of 29 deaths, ICD: 14 of 42 deaths).

The survival advantage was significant after adjustment for the relevant baseline characteristics. Patients implanted with a CRT-D had an overall significant, 60% lower risk of all-cause mortality compared with those with an ICD-only (HR = 0.40, 95% CI: 0.23–0.69, P = 0.001) (Table 3).

The lower mortality was evident within the first 6 months after CRT-D implantation (HR = 0.42, 95% CI: 0.17–1.03, P = 0.058), and it was sustained beyond 6 months after device implantation (HR = 0.39, 95% CI: 0.21–0.73, P = 0.004). The effect size was similar when stratified by time (HR's 0.42 and 0.39 within and after the first 6 months), and there was no significant interaction revealed with follow-up time (interaction P-value = 0.899) suggesting an early reduction in all-cause mortality with an implanted CRT-D.

The risk of ventricular tachyarrhythmias or death in cardiac resynchronization therapy with defibrillator patients compared with implantable cardioverter-defibrillator alone, stratified by the elapsed time after device implantation

In the trial, there were 90 (11.9%) CRT-D patients and 113 (15.2%) ICD patients developing VT or VF according to the endpoint definition. The cumulative probability of VT or VF episodes was significantly lower in CRT-D patients than in ICD patients (P = 0.038) (Figure 2). The Kaplan–Meier graph is suggestive of no difference in VT/VF in the first 6 months in CRT-D patients when compared with an ICD alone.

In the multivariate models, the risk of sustained VT/VF episodes showed a trend to be lower among CRT-D patients compared with conventional ICD patients (HR = 0.73, 95% CI: 0.52–1.03, P = 0.073).

However, the risk of VT/VF in CRT-D patients compared with ICD patients was significantly lower after the first 6 months of device implantation (HR = 0.58, 95% CI: 0.38–0.88, P = 0.011). In the first 6 months, there was no significant difference in VT or VF in CRT-D vs. ICD patients (HR = 1.00, 95% CI: 0.62–1.62, P = 0.988). The interaction P-value with the follow-up time of 6 months was borderline significant, suggesting a difference in CRT-D effect to reduce ventricular tachyarrhythmias over time (interaction P-value = 0.059) (Table 3).

Evaluating the combined endpoint of VT/VF or death revealed consistent findings. Patients with an implanted CRT-D compared with ICD did not have a lower risk of VT/VF or death in the first 6 months after device implantation; however, after the first 6 months the significant reduction became evident (Figure 3 and Table 3).

Discussion

The main findings of this MADIT-RIT sub-study evaluating the time-dependent effects of CRT-D vs. ICD on mortality and ventricular tachyarrhythmias are that there is an early reduction in all-cause mortality with an implanted CRT-D; however, the reduction in VT or VF episodes is delayed and emerging only 6 months after CRT-D implantation. This effect is evident after adjustment for relevant clinical covariates in the study. These findings imply that even patients with an implanted CRT-D are at substantial risk of VT or VF in the first 6 months after device implantation and should be carefully monitored and treated when seen with VT or VF in the clinic.

The MADIT-RIT trial was the first large scale randomized defibrillator study that included conventional ICD and CRT-D devices according to a guideline-based indication and not with the objective to evaluate CRT-D effects in patients with a hypothetical CRT indication. This gave us the unique opportunity to evaluate the differences in the risk of all-cause mortality and VT/VF in patients with implanted ICD vs. a CRT-D implant in a ‘real life’ setting in these different but equally important HF patient sub-populations. Our results are, therefore, novel and remarkable because CRT-D indication implies the presence of severe systolic HF and delayed LV electrical activation; the combination of which had to be considered particularly arrhythmogenic and life-threatening in view of previous HF and defibrillator studies.¹⁶

However, the relative reduction in mortality risk and in the risk of VT/VF in CRT-D vs. ICD patients was similar. Interestingly, the mortality advantage was early and sustained over the follow-up period, whereas the favourable prevention of sustained ventricular tachyarrhythmias was time-dependent and emerged 6 months after device implantation.

This association is thought provoking and raises the question, why is there an immediate reduction in death but not in VT/VF? It is hard to answer. We know from previous studies that there is an immediate reduction in LV dyssynchrony, mitral regurgitation, and LV contractility after CRT implantation.¹⁷ However, the structural changes in the myocardium, including the reduction in scar tissue, fibrosis, apoptosis, as well as changes in genomics may develop over time.¹⁸ We can, therefore, hypothesize that ventricular tachyarrhythmias are more related to the structural changes in the myocardium, ‘arrhythmogenic substrate’, and less so to the acute changes in haemodynamics. In contrast, the early beneficial reduction in all-cause mortality may emerge from the improvement in haemodynamics, an immediate effect after CRT implantation. But this needs further investigation. We may also hypothesize that the lower risk of VT/VF relates to LV reverse remodelling, which is known to occur predominantly during the first 6 months of CRT.¹⁹

In the MADIT-II trial, 24% of the patients experienced VT/VF over the mean follow-up period of 17.2 months, a follow-up duration very similar to MADIT-RIT. Different causes are likely to have contributed to have only half of the risk of appropriate ICD therapy in MADIT-RIT compared with MADIT-II. As previously published, the conservative programming in the high-rate cut-off arm and the delayed therapy arm not only reduced inappropriate ICD therapy, but also significantly reduced appropriate ICD therapy, in particular appropriate ATP.¹¹ In addition, improved medical treatment including the frequent use of beta-blockers may have contributed to the lower risk of VT/VF or sudden cardiac death (SCD).

However, in line with the recently published data from the MADIT-CRT trial in patients with mild HF^6,8 and a matched pair analysis,⁷ the present data from the MADIT-RIT trial add more substantial evidence that the treatment of delayed electromechanical ventricular activation by CRT not only lowers mortality from progressive HF, but also lowers the incidence of sustained ventricular tachyarrhythmias, that is, however, emerging only 6 months after device implantation.

It is important to note that there is an equal risk of ventricular tachyarrhythmias in CRT-D patients in the first 6 months as in ICD patients. This highlights the importance of treating VT/VF in CRT-D patients early, and therefore, improve clinical outcome.²⁰ It is, however, not known whether patients with VT/VF in the first 6 months may have been CRT non-responders, or whether VT/VF is associated with more frequent hospitalization for HF or death in this particular cohort.

Interestingly, CRT-D effect on mortality was similar in all programming arms; however, the reduction in VT/VF was most pronounced in the high-rate cut-off VT therapy ICD programming arm, Arm B. We do not have information on LV remodelling in MADIT-RIT, or left bundle branch block (LBBB) status in both ICD and CRT-D patients, and therefore, we do not know if this is in relation with echocardiographic outcomes. Another explanation could be that patients in Arm B have less monitored non-treated VT episodes stored in the ICD device due to overwriting. Even if this would be the case, monitored non-treated VT's do not seem to affect outcome,¹¹ and the pronounced beneficial effects of CRT-D on mortality and VT/VF in Arm B further stresses the everyday clinical implications of a simple one-zone VT therapy programming in patients with primary prevention ICDs.

Our study has certain limitations. Effects of unmeasured confounders cannot be ruled out with certainty, since CRT-D and ICD treatments were homogeneously distributed, but not randomized in MADIT-RIT. We did not collect data on LV reverse remodelling after CRT-D implantation, and the data on LBBB were partial.

It is important to note, that despite the significant reduction in death and VT/VF with CRT-D, more than 10% of CRT-D patients still received an appropriate ICD therapy for VT/VF during the 17 months of follow-up and some of them might have actually died without the defibrillator option. There are no randomized data appropriately powered available on the prevention of SCD through the defibrillator option in CRT patients when compared with a CRT without a defibrillator.² However, considering the above-mentioned residual crude event rates, we do not uniformly advocate refraining from the defibrillator option in CRT patients, as a non-negligible number of patients will still carry the risk of SCD. Instead, the decision to implant a CRT-P or a CRT-D system must still be made on an individual basis taking into account co-morbidity and the patients' concept on dying.

Conclusion

We showed that treatment with a primary prevention CRT-D based on the current guidelines is associated with an early and sustained survival advantage and a time-dependent reduction in the risk of VT/VF when compared with a primary prevention ICD. The difference in the risk of VT/VF emerges 6 months after device implantation. Patients with CRT-D should still be carefully monitored for VT/VF, especially in the first 6 months after device implantation.

Funding

The MADIT-RIT study was supported by a research grant from Boston Scientific, St. Paul, Minnesota, to the University of Rochester School of Medicine and Dentistry.

Conflict of interest: M.S.: research support and fees for educational activities from Biotronik, Boston Scientific, Medtronic, and Sorin Group. B.O.: consulting and/or speaking fees from Medtronic, Boston Scientific, Boehringer Ingelheim, BioControl, and Amarin. H.K.: research grant and speaker honoraria from Boston Scientific. C.S.: research grant from Boston Scientific. J.P.D.: grant support from Boston Scientific, Biosense-Webster, Medtronic, honoraria Boston Scientific, Medtronic, Biosense-Webster, St. Jude Medical, Biotronik, and Sorin. I.G.: research grant from Boston Scientific. A.-C.H.R.: Mirowski-Moss Awardee, unrestricted grants from Falck Denmark and The Lundbeck-Foundation. B.M.: research grant from Boston Scientific. W.Z.: research grant from Boston Scientific. A.J.M. research grant from Boston Scientific. V.K. and S.M.: none declared.

Acknowledgements

We would like to thank the work of Bronislava Polonsky from the University of Rochester, Rochester, NY for the assistance in the statistical analyses, creating the databases, and the macros used in the present manuscript.

References