Resynchronization-Defibrillation for Ambulatory Heart Failure Trial - RAFT

Contribution To Literature:

Highlighted text has been updated as of January 19, 2024.

The RAFT trial showed that CRT-D was superior to ICD in reducing mortality in patients with mild to moderate symptoms (NYHA class II/III), LV systolic dysfunction (EF ≤30%) and wide QRS complex.


Earlier trials have established the superiority of cardiac resynchronization therapy (CRT)-D over implantable cardioverter-defibrillator (ICD) alone in patients with severely symptomatic chronic heart failure (CHF) (New York Heart Association [NYHA] class III or IV), left ventricular (LV) systolic dysfunction, and wide QRS complex. In addition, another trial (MADIT-CRT) noted a reduction in death or CHF hospitalization with CRT-D, as compared with ICD alone in patients with NYHA class I and II symptoms with LV systolic dysfunction and wide QRS complex. 

The current trial sought to study the safety and efficacy of CRT-D, as compared with ICD alone in patients with mild to moderate symptoms (NYHA class II/III), LV systolic dysfunction (ejection fraction [EF] ≤30%), and wide QRS complex.

Study Design

  • Blinded
  • Parallel
  • Randomized
  • Stratified

Inclusion Criteria:

  • NYHA class II or III
  • EF ≤30% (ischemic and nonischemic)
  • Intrinsic QRS duration ≥120 msec, or ≥200 msec if paced rhythm
  • Sinus rhythm or permanent atrial fibrillation or flutter with a controlled ventricular rate
  • Planned ICD implantation for indicated primary or secondary prevention of sudden cardiac death
  • Optimal heart failure pharmacologic therapy
  • Number of enrollees: 1,798
  • Duration of follow-up: 6 years
  • Mean patient age: 66.2 years
  • Percentage female: 17%
  • Ejection fraction: 22.6%
  • NYHA class: II (80%), III (20%)

Exclusion Criteria:

  • Intravenous inotropic agent in the past 4 days
  • Life-expectancy <1 year due to noncardiac causes
  • Expected to undergo cardiac transplantation within 1 year (status I)
  • In-hospital patients needing intensive care
  • Uncorrected or uncorrectable primary valvular disease
  • Restricted, hypertrophic, or reversible forms of cardiomyopathy
  • Severe primary pulmonary disease
  • Tricuspid prosthetic valve
  • Patients with an existing ICD
  • Coronary revascularization within 1 month, if previously determined LVEF >30%

Primary Endpoints:

  • All-cause mortality or hospitalization for worsening CHF

Secondary Endpoints:

  • All-cause mortality
  • Cardiovascular mortality
  • CHF hospital admission

Drug/Procedures Used:

Patients were randomized in a 1:1 fashion to receive either ICD or CRT-D. Medtronic transvenous devices and leads were used in both arms. Standard implantation techniques were employed, with an emphasis on placing the LV lead in the lateral or posterolateral wall of the left ventricle whenever possible. Programming was set to minimize right ventricular pacing in the ICD arm, and to maximize biventricular pacing in the CRT-D arm.

Concomitant Medications:

Angiotensin-converting enzyme inhibitors (97%), beta-blockers (90%), spironolactone (42%), Coumadin (34%), and diuretic (84%)

Principal Findings:

A total of 1,798 patients were randomized, 894 to CRT-D and 904 to ICD. Baseline characteristics were fairly similar between the two arms. About 34% had diabetes mellitus, and 66% had ischemic cardiomyopathy. The majority of patients had mildly symptomatic (NYHA class II, 80%) systolic LV dysfunction (mean LVEF, 22.6%). About 13% of the patients were in permanent atrial fibrillation or flutter. The mean nonpaced QRS complex was about 158 msec, and the mean paced QRS complex was 208 msec; a left bundle branch block (LBBB) was noted in about 72% of the patients. An LV lead was successfully placed in 95% of the patients in the CRT-D arm. There was a 10% cross-over to the CRT-D arm from the ICD arm.

The primary endpoint of all-cause mortality or CHF hospitalization was significantly reduced over the course of the study in the CRT-D arm compared with ICD (33.2% vs. 40.3%, hazard ratio [HR] 0.75, 95% confidence interval [CI] 0.64-0.87, p < 0.001). All-cause mortality was also significantly lower (20.8% vs. 26.1%, HR 0.75, 95% CI 0.62-0.91, p = 0.003). This corresponded to a number needed to treat (NNT) of 14 patients for 5 years to prevent one death with CRT-D, as compared with ICD. CHF hospitalization was similarly reduced (19.5% vs. 26.1%, p < 0.001). However, device-related hospitalizations were higher in the CRT-D arm (20.0% vs. 12.2%, p < 0.001). Results for the primary and secondary endpoints were similar for NYHA class II or III patients.

Among other subgroup analyses, there seemed to be a greater benefit with CRT-D in patients with wider QRS complex at baseline (≥150 msec for native, ≥200 msec for paced). LBBB morphology also seemed to derive greater benefit. There was a trend towards greater efficacy in women, although they constituted only 17% of the study population.

Device- or implantation-related complications were significantly higher in the CRT-D arm, as compared with the ICD arm at 30 days (13.3% vs. 6.8%, p < 0.001). These included hemothorax or pneumothorax, device-pocket hematomas, device-pocket infections needing intervention, and lead dislodgement requiring intervention. Coronary sinus dissection was noted in 1.2% of patients in the CRT-D arm.

At 18-month follow-up, all-cause hospitalizations were similar between the CRT-D and ICD arms (37% vs. 38.8%, p = 0.44). However, cardiovascular-related hospitalizations (18.6% vs. 23.8%, p = 0.0008) and CHF-related hospitalizations (11.3% vs. 15.6%, p = 0.0009) were lower in the CRT-D arm. Hospitalization for device-related events was numerically higher in the CRT-D arm (16.4% vs. 13.9%, p = 0.15). Lead repositioning, lead failure, a reattempt to implant an LV pacing lead, and ICD pulse generator replacement for expected battery depletion were the major causes of the increased rate of hospitalization in the ICD-CRT group.

Long-term follow-up: 58.4% (1,050 patients) were enrolled in the long-term follow-up cohort (median 7.7 years). All-cause mortality was lower with CRT-D vs. ICD: 71.2% vs. 76.4%, acceleration factor 0.80, 95% CI 0.69-0.92, p = 0.002. The composite of all-cause mortality, cardiac transplantation, implantation of ventricular assist device for CRT-D vs. ICD: 75.4% vs. 77.7%; acceleration factor 0.85, 95% CI 0.74-0.98.


The results of the RAFT trial indicate that in patients with NYHA class II or III LV systolic dysfunction (EF <30%) and a wide QRS complex, who were optimally medically treated, CRT-D is superior to ICD alone in reducing the rates of all-cause mortality, CHF hospitalization, and their composite endpoint. There is, however, a significant increase in the risk of device- or procedure-related complications at 30 days in the CRT-D arm, including a higher risk of hospitalization for device-related causes. On subgroup analysis, the results of the primary endpoint seemed to be enhanced in patients with a wider QRS complex at baseline and those with LBBB morphology. The benefit of CRT-D over ICD appeared sustained on long-term follow-up, although data were available for 58.4% of the total population.

This trial thus extends the findings of other sentinel trials in this field. The COMPANION trial established the superiority of CRT-D over optimal medical management alone in reducing all-cause mortality in a similar group of patients, but with NYHA class III or IV symptoms. While improvements in LV dimensions with CRT in patients with mildly asymptomatic CHF were noted in the MIRACLE ICD-II and REVERSE trials, there was no significant difference in clinical endpoints in these trials. MADIT-CRT demonstrated a significant reduction in the composite endpoint of mortality or CHF events in similar patients with NYHA class I or II symptoms over a mean follow-up of 2 years. All-cause mortality was only reduced in the CRT-D arm in patients with LBBB morphology on long-term follow-up.

The results of the RAFT trial in patients with mostly class II symptoms over a long duration of follow-up indicate that CRT-D is superior to ICD in reducing all-cause mortality, with a NNT of only 14 patients at 5 years. Cardiovascular readmissions and CHF readmissions were lower with CRT-D, but device-related hospitalizations were numerically higher, resulting in no significant difference between the two arms for all-cause hospitalizations. The higher complication rates in this trial are similar to those noted by other CRT-D trials. Thus, although a lot more patients may now be eligible for CRT-D, careful surveillance for these complications is essential. Further studies are necessary to identify optimal LV lead placement for maximal benefit, and the cost-effectiveness of extending CRT-D to mildly symptomatic patients, since CRT-D is significantly more expensive than ICD implantation alone.


Sapp JL, Sivakumaran S, Redpath CJ, et al., on behalf of the RAFT Long-Term Study Team. Long-Term Outcomes of Resynchronization–Defibrillation for Heart Failure. N Engl J Med 2024;390:212-20.

Editorial: Stevenson LW, Montgomery JA. Seeking More Time With Synchrony. N Engl J Med 2024;390:269-70.

Presented by Dr. Anthony Tang at the American Heart Association Scientific Sessions, Chicago, IL, November 14, 2010.

Tang AS, Wells GA, Talajic M, et al. Cardiac-resynchronization therapy for mild-to-moderate heart failure. N Engl J Med 2010;363:2385-95.

Gillis AM, Kerr CR, Philippon F, et al. Impact of Cardiac Resynchronization Therapy on Hospitalizations in the Resynchronization-Defibrillation for Ambulatory Heart Failure Trial (RAFT). Circulation 2014;129:2021-30.

Clinical Topics: Arrhythmias and Clinical EP, Heart Failure and Cardiomyopathies, Implantable Devices, SCD/Ventricular Arrhythmias, Acute Heart Failure

Keywords: Cardiac Resynchronization Therapy, Heart Failure, Defibrillators, Implantable

< Back to Listings