Importance of QRS Duration in Resynchronization Therapy

Abstract & Commentary

By John P. DiMarco, MD, PhD, Professor of Medicine, Division of Cardiology, University of Virginia, Charlottesville.

Source: Sipahi I, et al. Impact of QRS duration on clinical event reduction with cardiac resynchronization therapy: Meta-analysis of randomized controlled trials. Arch Intern Med 2011 Jun 20. [Epub ahead of print.]

Cardiac resynchronization therapy (CRT) is now accepted as a disease-modifying therapy in patients with heart failure, left ventricular dysfunction, and intraventricular conduction defects. In this paper, Sipahi and colleagues review the data on the benefits of CRT in relationship to baseline QRS duration.

The authors performed a meta-analysis after reviewing all studies involving CRT. The meta-analysis could only include trials that reported clinical outcomes stratified by QRS duration, were randomized, and had a non-CRT control group. Although 412 studies were initially identified, there were only a total of five randomized controlled trials enrolling 5813 patients that met all criteria for inclusion in the meta-analysis. These trials, however, were the large, long-term studies on CRT and, therefore, they included much of the published data on the effects of CRT. The studies were COMPANION, CARE-HF, REVERSE, MADIT-CRT, and RAFT. COMPANION and CARE-HF included patients with only class III or class IV heart failure. REVERSE and MADIT-CRT included patients with only class I or II heart failure. RAFT included mostly patients with class II, but also some with class III, heart failure. Where possible, patients were then stratified as having either moderately prolonged QRS durations (120 to 149 msec) compared to those with greatly prolonged QRS durations (greater than or equal to 150 msec). Accommodations were made if the trial used different QRS cut-off points when reporting outcomes.

In these five studies, most patients were male. About 50% had nonischemic heart failure. The mean left-ventricular ejection fractions ranged from 21% to 27%. Between 69% and 90% of the patients had left bundle branch block morphology. All studies required patients in both the control and the CRT groups to receive optimal medical therapy. In COMPANION and CARE-HF, the CRT arm involved biventricular pacing (CRT-P) only without associated defibrillation capability. In REVERSE, MADIT-CRT, and RAFT, the large majority of patients had devices with defibrillation (CRT-D) in both arms.

The meta-analysis showed that patients with a severely prolonged QRS duration randomized to CRT had a 40% risk reduction in death and heart failure hospitalization. In contrast, there was no statistically significant benefit in patients with moderately prolonged (120 msec to 149 msec) QRS durations in any of the trials and only a trend toward benefit in CARE-HF. In this latter trial, the moderately prolonged QRS duration subgroup, however, included many patients with a QRS duration between 150 msec and 158 msec. When data from all the trials were plotted, there was a statistically significant relationship between QRS duration and the log of the risk-reduction slope. Patients with QRS durations below 150 msec did not receive benefit from CRT. The magnitude of benefit became more prominent as the QRS duration became more prolonged. The same results were seen when several scenarios were tested using sensitivity analysis. The importance of baseline QRS duration was not affected by the patient's heart failure classification.

The authors conclude that the effectiveness of CRT is only demonstrable in patients with a QRS of 150 msec or greater. They urge that QRS duration should be an important criteria for selecting patients for CRT.


There are a number of factors that affect the probability of benefit with CRT. Higher response rates to CRT have been noted in patients with left bundle branch block as opposed to right bundle branch block, among women as opposed to men, in those with a nonischemic cardiomyopathy as opposed to a ischemic cardiomyopathy, and in those with longer QRS durations. However, for the most part, government agencies and guideline writers have taken the entry criteria for the large CRT trials and used those criteria when writing guidelines for insertion. This means that a significant number of patients for whom CRT is currently recommended will have QRS durations below 150 msec. Unfortunately, as shown here, this group seems to have relatively little short-term benefit. We also must remember that there is significant increased risk of complications as well as expense associated with adding the left ventricular lead. The average rate of left ventricular lead complications, either failure to insert or late dislodgement or phrenic nerve stimulation, may be as high as 15%. The CRT trials usually did not count these LV lead problems as primary endpoints even if another surgical procedure or added hospitalization was required. Finally, we must also remember that many of the potential benefits of CRT may be due to a favorable influence on ventricular remodeling. These benefits may take years to become apparent and were probably not detectable in the studies reviewed here.

This meta-analysis is quite useful in helping us structure our thinking, particularly in patients with class I or II heart failure who receive relatively little early benefit. For patients with an ICD indication but only moderate QRS prolongation, my preference has been to implant an ICD and then upgrade those patients who go on to either require right ventricular pacing or develop progressive widening of their QRS complex. Among patients with a borderline QRS prolongation with class III or IV heart failure who need early benefit, I discuss with the patient and the referring physician the likelihood that they will improve with CRT. For patients with shorter QRS durations (120-130 ms) or right bundle branch block, I often discourage use of the more complicated CRT device.