Study Reveals Optimal Use of Cardiac Resychronization

Abstract & Commentary

By John P. DiMarco, MD, PhD

Source: Cardiac-resynchronization therapy for mild-to-moderate heart failure. Tang ASL, et al. N Engl J Med. 2010; 363:2385-2395.

This paper reports results from the resynchronization-Defibrillation for Ambulatory Heart Failure Trial (RAFT). This was a multicenter, double-blind, randomized trial testing the hypothesis that addition of resynchronization therapy (CRT) to an ICD, compared to an ICD only in patients with New York Heart Association (NYHA) class II or class III symptoms, left ventricular systolic dysfunction, and a widened QRS complex would improve survival and reduce hospitalizations. The study was conducted in 34 centers outside the United States, with 24 of these centers in Canada. Initially, eligible patients could have either NYHA class II or class III symptoms of heart failure despite receiving optimal medical therapy. After the trial was underway, class III heart failure became an accepted criterion for CRT. The protocol was then modified such that only patients with NYHA class II symptoms could be enrolled. All patients had a left ventricular ejection fraction of 30% or less. Patients with either ischemic or nonischemic heart disease were eligible. All had an intrinsic QRS duration of 120 m/sec, or more of a paced QRS duration of 200 m/sec or more, and were scheduled for ICD implantation for either primary or secondary prevention of sudden cardiac death. Baseline evaluations included history and physical examination, a six-minute walk test, a quality-of-life assessment, and a medication evaluation to ensure optimal medical therapy. Patients enrolled were randomly assigned in a 1:1 ratio to receive either an ICD only or an ICD with resynchronization therapy (ICD-CRT). At the time of implant, programming of the device was standardized to minimize ventricular pacing in the ICD-only group and to maximize ventricular pacing in the ICD-CRT group. Uniform arrhythmia detection and therapy settings were employed. Patients were then seen periodically in follow-up until the end of the trial. The primary outcome was death from any cause or a hospitalization related to heart failure. Admissions for other medical problems that then were complicated by worsened heart failure during the hospitalization were not classified as heart-failure hospitalizations. Events were adjudicated by a committee that was blinded to study-group assignments. Secondary outcomes included all-cause mortality and cardiovascular mortality. Preplanned subgroup analyses of patients with class II vs. class III heart failure were performed.

Over a six-year period from 2003 to 2009, a total of 1,798 patients were enrolled in RAFT. The clinical characteristics were similar to those for earlier ICD and CRT studies. Of the total group, 80% had NYHA class II and 20% NYHA class III symptoms. The mean left ventricular ejection fraction was 22.6%. Patients were treated with appropriate heart-failure regimens, with 90% receiving beta blockers, 96% receiving an ACE inhibitor or angiotensin receptor blocker, and 42% receiving spironolactone. The two groups were also well-matched with regard to QRS duration, six-minute walk test time, estimated glomerular filtration rate, and other clinical parameters. After randomization, five patients in the ICD group and six patients in the ICD-CRT group did not receive device implants. In the ICD group, a left ventricular lead could be successfully implanted in 841 of 888 patients (94.7%). In 39 of these patients, a second procedure was required to achieve a successful implant.

During the course of the study, five patients in the ICD group either withdrew or were lost to follow-up, compared to 10 patients in the ICD-CRT group. Twelve patients underwent cardiac transplantation during the course of the study (five in the ICD group and seven in the ICD-CRT group). Crossovers from ICD to ICD-CRT therapy occurred in 96 patients. In the ICD-CRT group, 53 patients (6%) did not receive CRT either because of left ventricular lead failure (47) or malfunction (6).

There were 364 deaths or hospitalizations for heart failure among the 904 patients in the ICD group (40.3%) as compared with 297 deaths or heart-failure hospitalizations among the 894 patients in the ICD-CRT group (33.2%). By life-table analysis, the hazard ratio (HR) for the primary endpoint was 0.75, with a 95% confidence interval (CI) of 0.64 to 0.87; p < 0.001. There were 422 deaths in the two study groups. The time to death was significantly prolonged in the ICD-CRT group (HR 0.75, 95% CI, 0.62 to 0.91; p = 0.003). Using a number needed to treat analysis, this would translate to 14 patients who would need to be treated for five years with ICD-CRT in order to prevent one death. Interestingly, the number of hospitalizations for any cardiac cause was similar between the two groups (404 ICD; 423 ICD-CRT), but the number of hospitalizations for heart failure was significantly reduced in the ICD-CRT group (19.5% compared to 26.1%). This was partly attributable to the higher number of device-related hospitalizations in the ICD-CRT group with 179 such hospitalizations (20%) as compared to 110 in the ICD group. When patients were analyzed according to heart-failure class, similar patterns were seen in both the NYHA class II and class III groups. ICD-CRT therapy was more effective in patients with longer QRS durations (equal or above 150 m/sec or more), with a hazard ratio of 0.59 compared to a hazard ratio of 0.99 in those with shorter QRS durations. In addition, patients with a left-bundle branch-block morphology experienced more benefit than those with either nonspecific intraventricular conduction defects or a right-bundle branch block. Similar benefits were seen with ICD-CRT in patients with either ischemic or nonischemic causes of heart failure. Complications were more common in the ICD-CRT group. During the first 30 days after device implantation, there were 118 device- or implantation-related complications among the 888 patients in the ICD-CRT group (13.3% compared to only 61 such problems among the 899 patients in the ICD group (7.5%). This was due to an increase in the frequency of pneumothorax, hematoma, lead dislodgement requiring intervention, device pocket problems, and coronary sinus dissection in the ICD-CRT group.

The authors concluded that addition of CRT to the use of an ICD in the setting of optimal medical therapy reduced rates of death and hospitalization for heart failure in patients with NYHA class II and class III heart failure, left ventricular systolic dysfunction, and a wide intrinsic QRS complex. This survival benefit should be considered in light of the decreased battery longevity and the increased rate of complications seen with the more complicated devices.


This excellent study points out for us the risks and benefits of extending CRT indications in heart-failure patients scheduled to receive an ICD. Although both mortality and heart-failure morbidity were improved, these improvements came at the price of increased device-related complications and decreased battery longevity. How then can we use these data to guide our clinical practices? I find the subgroup analyses in this study particularly helpful. Patients with baseline QRS durations at or above 150 m/sec and those with a left-bundle branch morphology were most likely to benefit. In patients with these findings at baseline, the increased cost and risk for complication associated with CRT is certainly justifiable. If these findings are not present, the risks remain, but the probability of benefit is low. In the latter situation, my preference is to implant an ICD only if the patient has Class II symptoms and then consider an upgrade if the heart failure worsens. We also carefully re-evaluate the clinical status of our patients prior to ICD generator changes and consider CRT in those we think likely to benefit. This relatively conservative approach seems to me to be the best response to the data presented here in RAFT and in the earlier MADIT-CRT trial (N Engl J Med. 2009;361:1329-1338).