Combined Cardiac Resynchronization and Implantable Cardioversion Defibrillation in Advanced Chronic Heart Failure: Title and subTitle BreakThe MIRACLE ICD Trial FREE

An important subset of patients who have chronic heart failure (HF) also have cardiac dyssynchrony. Delays in interventricular or intraventricular electrical activation cause marked abnormalities in the sequence of global and segmental right and left ventricular (LV) activation and impair mechanical performance.^{1 - 3} The ability of new methods of cardiac stimulation to resynchronize ventricular function, improve overall cardiac performance, and increase exercise capacity has been shown by the results of several observational and controlled studies.^{4 - 10} The Multicenter InSync Randomized Clinical Evaluation (MIRACLE), the first randomized, double-blind, controlled trial of cardiac resynchronization therapy (CRT) through atrial-synchronized biventricular pacing demonstrated a significant improvement in New York Heart Association (NYHA) functional class, an increase in exercise duration, and improvements in quality of life in patients with moderate to severe HF and wide QRS interval. These patients did not have an indication for pacing or defibrillator therapy.¹¹

Despite increases in survival rates achieved by new pharmaceutical agents, such as angiotensin-converting enzyme inhibitors and β-blockers, the 5-year mortality following the diagnosis of HF is about 50%, with most of these deaths occurring suddenly and unexpectedly.^{12 - 13} Although an increasing number of implantable cardioverter defibrillator (ICD) recipients have dilated hearts and marked conduction abnormalities, most reported studies of CRT excluded patients with standard indications for an ICD. One randomized study included patients with symptomatic HF with a wide QRS complex interval and indications for an ICD but was of short duration (3 months) and did not assess the symptomatic benefit of CRT.¹⁴ In these patients, HF often represents the main factor limiting quality of life and survival. On the other hand, the improved survival conferred by ICDs is not limited to patients with life-threatening ventricular arrhythmias¹⁵ and includes survivors of myocardial infarction and LV dysfunction without prior symptomatic ventricular arrhythmias.¹⁶ Furthermore, even though ICD implantation alone decreases HF mortality, it does not appear to ameliorate symptoms. These observations have prompted development of implantable devices that combine cardiac resynchronization with the electrical treatment of tachyarrhythmias.¹⁷

We hypothesized that patients with moderate to severe HF symptoms, a wide QRS interval, LV systolic dysfunction, and an established indication for an ICD would benefit from CRT, and that CRT would not be proarrhythmic or compromise ICD therapy.

The Multicenter InSync ICD Randomized Clinical Evaluation (MIRACLE ICD) study was a randomized, double-blinded, parallel-controlled clinical trial to evaluate the efficacy of CRT in a large number of patients with moderate to severe systolic HF, ventricular dyssynchrony, and an indication for an ICD.¹⁵ More specifically, indications for an ICD at study entry generally were cardiac arrest (manifest by loss of consciousness) due to ventricular tachycardia or ventricular fibrillation without a transient, reversible cause; patients having recurrent, poorly tolerated, and sustained ventricular tachycardia that occurs spontaneously or can be induced. Except for the ICD indication and the timing of baseline tests described later, the patient inclusion criteria and study design were identical to those previously reported for the MIRACLE study.¹¹ Enrollment began after October 1, 1999, and was completed by August 31, 2001. The investigational review board of each participating institution reviewed and approved the study protocol, and all patients granted their written informed consent before entering the trial.

The Article lists the study inclusion and exclusion criteria. Eligible patients received a stable and appropriate drug regimen, which included an angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker, if tolerated, for at least 1 month. If a patient was taking a β-blocker, it had to have been initiated at least 3 months before enrollment. Initiation of β-blockade was not permitted during the trial period.

Inclusion

Age ≥18 years
Cardiac arrest due to ventricular fibrillation or ventricular tachyarrhythmia, or spontaneously sustained ventricular tachyarrhythmia, or inducible ventricular fibrillation or sustained ventricular tachyarrhythmia
New York Heart Association functional class III or IV congestive heart failure
Left ventricular ejection fraction ≤35%
QRS duration ≥130 ms Left ventricular end diastolic diameter ≥55 mm
Left ventricular end diastolic diameter ≥55 mm
Stable drug regimen for ≥1 month

Exclusion

Estimated survival <6 months
Baseline 6-minute walk test >450 m
Bradycardia requiring pacemaker
Unstable angina, myocardial infarction, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, cerebral vascular accident, or transient ischemic attack within previous 3 months
>2 infusions of inotropic drug per week
Systolic blood pressure <80 mm Hg or >170 mm Hg
Resting heart rate >140/min
Serum creatinine >3 mg/dL (>265 µmol/L)
Hepatic enzymes >3-fold upper normal values
Severe lung disease
Chronic atrial arrhythmias, or cardioversion or paroxysmal atrial fibrillation within previous 1 month
Heart transplant recipient
Severe valvular heart disease

Patients who met the criteria for entry into the study underwent the following evaluations within 7 days of system implantation: estimation of NYHA functional class¹⁸ ; 6-minute walking test¹⁹ ; quality of life evaluation using the Minnesota Living with Heart Failure Questionnaire²⁰ ; 2-dimensional Doppler-flow echocardiography, which included measurement of LV ejection fraction, internal LV diastolic dimensions, end diastolic and systolic volumes, and degree of mitral regurgitation; plasma neurohormonal concentrations; and QRS width on 12-lead surface electrocardiogram. Patients with mild symptoms classified as NYHA.

In patients who agreed to participate, an implant was attempted with the Model 7272 InSync ICD (Medtronic Inc, Minneapolis, Minn), a standard right atrial pacing lead, a standard right ventricular (RV) pacing/defibrillator lead, and a choice of several LV transvenous leads (Medtronic Inc) positioned in a coronary sinus tributary. The InSync ICD delivers atrial-synchronized biventricular pacing for cardiac resynchronization, antitachycardia pacing through RV or RV and LV leads, and cardioversion and defibrillation to treat ventricular tachyarrhythmias delivered through the RV lead only.

Within 7 days of a successful implant, but before randomization, patients underwent a cardiopulmonary exercise test to measure peak oxygen consumption per unit time (O₂) and exercise duration. Patients were then randomly assigned, in blocked groups for each center, to active CRT, including optimal medical treatment and active ICD therapy (CRT group), or optimal medical treatment and active ICD therapy (control group) (Figure 1). Patients and the physicians from the HF team, who continued to follow patients after implantation of the CRT/ICD system but were not involved in the programming of the device, remained unaware of the randomization assignment until after the 6-month visit.

For patients in the CRT group, the device was programmed to a mode that paced both ventricles simultaneously following atrial-sensed events at rates of 130/min or lower. Atrial pacing occurred only for sinus rates of less than 35/min. For patients in the control group, the device was programmed to a mode that inhibited atrial or ventricular pacing unless the intrinsic rate was less than 35/min. Implantable cardioverter defibrillator therapy was activated in all patients.

Patients returned at 1, 3, and 6 months for full interrogation of the CRT/ICD system, reassessment of quality of life, follow-up 6-minute walking test, estimation of NYHA functional class, and monitoring of background drug regimen. Echocardiogram, cardiopulmonary exercise testing, and measurements of plasma neurohormones were repeated at the 6-month visit, after which the blinded phase of the study was completed and CRT was activated in patients initially randomized to the control group. Standard protocols were used to perform cardiopulmonary exercise tests and echocardiograms and to collect plasma neurohormones. Independent core laboratories, unaware of the patient randomization assignment, interpreted the data.

As in the previously reported MIRACLE study,¹¹ the 3 primary efficacy end points of MIRACLE ICD were NYHA functional class, quality of life score, and distance covered during the 6-minute walking test. In addition, several secondary end points were examined, including peak O₂, treadmill exercise duration, LV ejection fraction, LV end-systolic and end-diastolic volumes, LV end-diastolic dimension, severity of mitral regurgitation, QRS duration, neurohormone concentrations, and a clinical composite response that assigned all randomized patients to 1 of 3 response groups (worsened, improved, or unchanged).

A patient was classified as worsened if he or she died, was hospitalized due to worsening HF, permanently discontinued double-blind treatment due to or associated with worsening HF, permanently discontinued double-blind treatment because of withdrawal of consent or other administrative reason, had worsening HF at the time of study discontinuation, demonstrated worsening in NYHA class at last-observation-carried-forward (LOCF) or had moderate to marked worsening of patient global assessment score at LOCF. A patient was said to have improved if he or she had not worsened (as defined above) and demonstrated improvement in NYHA class at LOCF or had a moderate to marked improvement in patient global assessment score at LOCF. Finally, a patient was unchanged if he or she was neither improved nor worsened.^{21 - 22} All randomized patients contributed to all analyses with the following exceptions: only patients with data available at both baseline and follow-up were included in efficacy analyses and all patients undergoing an implant attempt were included in the adverse event analysis.

SAS software version 8.2 (SAS Institute, Cary, NC) was used to generate the random allocation sequence. The method of randomization was not disclosed to participating centers and was accomplished in blocked groups of 4 for each to ensure balance of CRT and control assignments at each participating institution. Randomization occurred following a successful implant. The randomization assignment for each patient was provided to the unblinded electrophysiology staff in a consecutively numbered and opaque (folded paper inside an envelope inside a second envelope) sealed envelope that was opened at the time of randomization. The HF staff was blinded to the randomization schedule and each patient's randomization assignment throughout the 6-month follow-up visit.

All end points were analyzed according to the intention-to-treat principle. Data are presented as median changes between baseline and 6 months. Confidence limits for medians were computed using a distribution-free approach.²³ Mean values are presented as mean (SD). For continuous variables, including NYHA, changes from baseline to the 1, 3, or 6-month visit in the control group vs the CRT group and demographic characteristics were compared with the Wilcoxon rank sum test. For categorical end points, differences in the distribution of responses to treatment at 6 months in the 2 groups were compared by using the Fisher exact test. Survival curves were constructed according to the Kaplan-Meier method with time zero being the date of implant, and differences between curves were examined by the log-rank test statistic. Confidence intervals (CIs) for survival were computed on the log-log survival scale.²⁴

For the primary efficacy variables, prespecified objectives were considered reached if differences between the groups in all 3 end points had P≤.05, if 2 had P≤.025, or if 1 had P≤.017, by using the Hochberg criterion.²⁵ The sample size (112 patients per treatment group) was estimated on the basis of the assumption that the study would have 80% power (2-sided α = .017) to detect a difference in NYHA class of 0.75, quality of life of 13 points, or distance walked in 6 minutes of 50 m. For secondary end points, P<.05 was considered statistically significant. All P values were calculated using 2-sided tests. In an analysis that was not prespecified, potential clinically relevant covariates were analyzed by using analysis of variance with randomization assignment, the covariates, and the interactions between the covariates and randomization assignment as independent variables.

A complication was defined as a sign, symptom, illness, or other medical event that was resolved invasively or that resulted in the death of or serious injury to a patient. The termination of a significant device function was also considered a complication. An invasive procedure was one that penetrated the skin, including the administration of parenteral fluids or drugs. The clinical events review committee reviewed and classified adverse events without knowledge of the randomization assignment. Investigators had full access to all data and performed analyses without restrictions or limitations from the sponsor.

Through August 31, 2001, 639 patients were enrolled in the MIRACLE ICD study program as detailed in Figure 1. Of those, 210 patients had mild HF symptoms (NYHA functional class II) and, by a priori protocol design, had a separate primary end point and are not included in this analysis. Sixty NYHA class III or IV patients underwent an implant attempt but did not proceed to the randomized therapy phase. A total of 369 patients included in the MIRACLE ICD study underwent successful implantation and were randomized (control group, n = 182; CRT group, n = 187). The baseline characteristics of the randomized patients, summarized in Table 1, were consistent with patients with moderate to severe HF and candidates for ICD therapy. Except for a higher percentage of patients with ischemic heart disease in the control group, the baseline clinical characteristics of the 2 groups were similar.

In the control group, 14 patients (8%) crossed over to CRT before the end of the randomized phase of the study. Biventricular pacing was activated early because of worsening symptoms of HF in 11 patients, bradycardia in 2 patients, and programming errors in 1 patient. In the CRT group, 10 patients (5%) crossed over from active biventricular pacing to no pacing before the end of the randomized phase. Biventricular pacing was deactivated due to LV lead dislodgement in 2 patients, diaphragmatic stimulation in 2 patients, and programming errors in 6 patients.

The results of the efficacy end points analyses are summarized in Table 2. An improvement in quality of life was observed in both study groups (Figure 2). However, the median decrease (improvement) in quality of life score between baseline and the 6-month visit was significantly higher in the CRT group compared with the control group (P = .02). Similarly, a significantly greater median decrease in NYHA functional class was measured in the CRT group compared with the control group (P = .007) (Figure 3). A similar increase between baseline and 6 months in median distance covered during 6-minute walking test was measured in both groups (Figure 2). When a LOCF analysis for surviving patients was performed, both the magnitude of changes in the primary end points and the P values for comparison between groups were nearly identical to those previously described in the analysis of paired data at 6 months (Table 2).

The treatment effect on quality of life score and NYHA functional class was not influenced by the use of a β-blocker, underlying heart disease (ischemic vs nonischemic), morphology of the QRS complex (left vs right bundle branch block), or the baseline duration of the QRS interval (P>.10 for all interactions with randomization assignment). This analysis was not preplanned, however, and may have been underpowered.

At 6 months, patients randomized to treatment with CRT exercised on the treadmill for a longer duration than at baseline, whereas treadmill exercise duration decreased in the control group (P<.001) (Table 2). A higher median increase in peak O₂ was measured in the CRT than in the control group (P = .04). By echocardiographic analysis, there was a trend toward greater reductions in LV systolic and diastolic volumes in the CRT group (P = .06 for both), although changes in other measures of LV size and function were similar in both groups. Median changes in plasma neurohormones were similar in both groups, except for a greater decrease in epinephrine in the control than the CRT group, a difference of borderline statistical significance (P = .05). A larger proportion of patients with CRT improved in their composite clinical response than in the control group and a smaller proportion worsened during the study period, but this was only a statistical trend (P = .07) (Figure 4).

During the 6-month randomization period, 47 patients (26%) in the control group and 42 patients (22%) in the CRT group experienced at least 1 spontaneous episode of ventricular tachycardia or fibrillation (P = .47). Of the spontaneous and treated episodes, outcomes of ICD therapy were recorded for 233 episodes in the control group and 678 episodes in the CRT group. Four episodes (1.7%) were not successfully terminated within the interval determined by device criteria in the control group vs 1 episode (0.1%) in the CRT group. These 5 episodes all eventually terminated spontaneously. There was no difference between the study groups in the detection times of ventricular fibrillation episodes. Furthermore, there was no difference in the number of patients receiving either appropriate or inappropriate ICD treatment, when comparisons are made by randomization assignment and by whether CRT was activated (Table 3). A total of 15 patients in the control group and 14 patients in the CRT group died during the 6-month follow-up. In each group, 3 of these deaths were characterized as sudden deaths. Cumulative survival at 6 months was 92.2% in the control group (95% CI, 87.2%-95.3%) vs 92.4% (95% CI, 87.5%-95.4%) in the CRT group (log rank P = .96). Of the 429 enrolled patients with at least 1 implant attempt, 5 (1.2%) died within 30 days of their latest implant attempt.

The median duration of the successful implantation procedures was 2.75 hours (interquartile range, 2.2-3.6). Between randomization and the 6-month visit, 78 patients in the control group (42.9%) and 85 patients in the CRT group (45.5%) were hospitalized. The mean (SD) length of hospital stay was 5.4 days (4.7) in the control group vs 4.8 days (4.9) in the CRT group (P = .06). During the 6-month follow-up, the probability of hospitalization for worsening HF or death from any cause was 25.9% (95% CI, 19.8%-32.5%) for the control group vs 25.7% (95% CI, 19.6%-32.3%) for the CRT group (P = .69). The risk of death or all-cause hospitalization was 48.3% (95% CI, 40.6%-55.6%) for the control group vs 47.4% (95% CI, 40.0%-54.4%) for the CRT group (P = .88).

In the control group, 86% of patients received no RV pacing during the randomization period. In the CRT group, 94% of patients were ventricular paced for 90% or more of the time during the randomization period.

Of the 429 patients undergoing an implant attempt, 120 patients (28%) experienced 159 complications from implant to hospital discharge. Of these 159 complications, 37 (23%) were related to the LV lead, including 15 coronary sinus dissections and 4 cardiac perforations. Other perioperative complications included HF decompensation in 6 patients, all treated with intravenous medications; heart block in 3 patients, all requiring bradycardia pacing support; muscle stimulation in 4 patients, treated by either a lead repositioning or lead replacement; pericardial effusion in 2 patients treated with a pericardiocentesis; pericarditis in 1 patient treated with intravenous medications; hemo/pneuomothorax in 3 patients treated with the placement of a chest tube; ventricular tachycardia and ventricular fibrillation in 5 patients, in which 3 patients were treated with external defibrillation and 2 patients were treated with intravenous medications; and elevated pacing thresholds or loss of capture in 7 patients, in which 6 patients were treated with a lead repositioning or lead replacement and 1 patient had a set screw tightened in the connector block.

Fifty patients had an unsuccessful CRT system implant but a successful placement of an ICD-only system. Of those 50 patients, 20 experienced a total of 35 complications from the time of hospital discharge through 6 months. Heart failure decompensation was the most common complication, accounting for 19 events (Table 4).

From hospital discharge to the end of the 6-month randomization period, 175 (46%) of the 379 patients with successful implants experienced 398 complications. The rate of device-related events was substantially lower than the rates anticipated in the prespecified criteria of the original study protocol. The frequency of adverse events unrelated to the device or to HF did not differ significantly between the 2 groups.

The improvements in quality of life and NYHA functional class in patients with moderate to severe HF, LV systolic dysfunction, wide QRS interval, and indications for an ICD were similar to those observed in comparable patient populations without indications for ICDs.^{7 - 8 ,10 - 11} However, the absence of a positive treatment effect on the 6-minute walking test contrasts with these earlier trials and with the improvements observed in this study with the more objective measurements of peak O₂ and treadmill exercise duration. Whether these discrepancies are due to differences between patient populations or to the different timing of the 6-minute walking test (performed before, instead of after, CRT system implantation) remains uncertain. Furthermore, although there was a trend of decreased LV volumes and increased ejection fraction in MIRACLE ICD (Table 2), these observations were not as compelling as those in the MIRACLE study.¹¹ Perhaps this is related to the fact that MIRACLE ICD patients, on the whole, were more ill (ie, having an ICD indication) with less chance for morphometric remodeling benefits that might be associated with cardiac resynchronization.

Although complications associated with biventricular pacemaker and ICD insertion were not trivial, the procedure was generally well tolerated with acceptable morbidity. Of particular interest was the incremental risk associated with the placement of the LV electrode and the impact of biventricular pacing on the incidence of ventricular arrhythmias and ICD functions, because all patients in this study were a priori candidates for an ICD. Of the 19 patients (4.4%) with either coronary sinus dissection or cardiac perforation, 11 ultimately underwent a successful CRT system implantation procedure, suggesting that the incremental risk was quite low. After successful CRT system implantation, 47 patients (12%) underwent LV lead repositioning or replacement during follow-up, which is comparable with the rates observed in previously commercialized LV lead models.^{11 ,26}

In the control group, crossovers to CRT were prompted by an inability to medically manage congestive HF in 11 of 14 patients, whereas in the CRT group, all crossovers to the alternate treatment were due to programming errors, lead dislodgment, or diaphragmatic pacing. This low incidence of crossovers had no effect on the results of the analyses, which were performed on an intention-to-treat basis.

Small uncontrolled studies have suggested that CRT may prevent ventricular arrhythmias.^{27 - 29} However, a larger randomized crossover study found no survival benefit conferred by biventricular pacing in HF patients with a preexistent indication for an ICD.³⁰ In this study, the number of patients experiencing ventricular arrhythmias was similar in both treatment groups (Table 3). Perhaps the 6-month duration of the randomization phase was too brief to allow expression of the full therapeutic effects of CRT. Considerably longer follow-up has often been required to demonstrate the benefits in pharmaceutical trials in HF. Another important observation in this trial was that detection of ventricular arrhythmias and their successful treatment by the ICD were unimpeded by the presence of biventricular pacing.

There were no differences in survival or rates of hospitalization between the control and the CRT groups. However, this study was not designed with the power to detect differences in survival and the 6-month randomization period may have been too brief to detect differences in hospitalization rates between the treatment groups. A meta-analysis of randomized controlled trials of cardiac resynchronization in symptomatic patients has suggested a reduction in mortality from congestive HF with biventricular pacing strategies.³¹

Right ventricular pacing was infrequent in the control group, whereas the CRT group was predominately biventricular paced. The Dual Chamber and VVI Implantable Defibrillator (DAVID) trial³² showed that, in patients with life-threatening ventricular arrhythmias, the combined risk of death and hospitalization for worsening failure was higher in patients exposed to dual-chamber pacing with RV apical stimulation than in patients left unpaced in their spontaneous rhythm. Although patients in the control groups of both studies were treated similarly, only approximately one third of the patients in the DAVID trial had a QRS duration of 130 ms or more and only 12% were in NYHA functional class III or IV HF and therefore were not eligible for randomization in the MIRACLE ICD trial. Also, differences in rates of hospitalizations for HF in the DAVID trial became apparent at 6 months, corresponding with the duration of the randomized phase of this study. These results raise the question of a treatment-limiting effect of RV added to LV stimulation vs a duration of follow-up too brief for the demonstration of differences in survival and rehospitalization.

The major limitation of the MIRACLE ICD trial is that it was not designed or powered to detect a mortality or morbidity difference between control and CRT groups and the follow-up was relatively short.

In conclusion, CRT without interfering with ICD functions improved the quality of life, functional capacity, and cardiopulmonary exercise test performance of patients with moderate to severe HF, a wide QRS interval, and life-threatening ventricular arrhythmias. These therapeutic effects were observed in the context of appropriate preexistent and continuing vigorous medical management of these patients.