QRS Duration Does Not Predict Occurrence of VT in Patients with ICDs

Abstract & Commentary

By John P. DiMarco, MD, PhD Dr. DiMarco is Professor of Medicine, Division of Cardiology, University of Virginia, Charlottesville. Dr. DiMarco is a consultant for Novartis and does research for Medtronic and Guidant.

Synopsis: QRS duration does not predict occurrence of VT or VF necessitating ICD therapy.

Source: Buxton AE, et al. QRS Duration Does Not Predict Occurrence of Ventricular Tachyarrhythmias in Patients with Implanted Cardioverter-Defibrillators. J Am Coll Cardiol. 2005;46:310-316.

Buxton and his colleagues from the painfree Rx II trial examined the impact of QRS duration on the occurrence of arrhythmias in patients who received an ICD. The PainFREE Rx II study was a randomized trial comparing empiric antitachycardia pacing vs initial shocks as therapy for episodes of rapid ventricular tachycardia (VT). The study enrolled 634 patients and, of these, 431 had a history of coronary artery disease. Data from these patients were included in the current analysis. All patients were programmed to 3 detection zones, a slow VT zone (cycle length cut-off 320 to 360 m/sec), a fast VT zone (cycle length 240 to 320 m/sec), and a ventricular fibrillation (VF) zone (cycle length less than 240 m/sec). Therapy was randomized between antitachycardia pacing and shocks in the fast VT zone. The QRS duration was measured at baseline and classified as left bundle branch block, right bundle branch block, or nonspecific intraventricular conduction delay by standard criteria. An expert panel analyzed all available stored electrograms of VT and VF episodes. The QRS duration was then correlated with VT or VF occurrence and it's value as a predictor assessed.

Four hundred and thirty one patients were included in the study. Two hundred and twenty five patients received their device after an episode of sustained or induced VT (secondary prevention group), whereas, 206 patients received the device for primary prevention. The QRS duration was ≥ 120 m/sec in 291 patients and < 120 m/sec in 140 patients. During follow-up, a total of 95 patients experienced at least one ICD therapy. There was no difference in the incidence of spontaneous VT based on QRS duration. Twenty three of 138 (17%) primary prevention and 40 of 153 (26%) secondary prevention patients with a QRS duration ≥ 120 m/sec experienced an episode. In comparison, 13 of 68 (19%) primary prevention patients and 19 of 72 patients (26%) secondary prevention patients with a QRS duration < 120 m/sec experienced an episode. There was also no difference in time to first VT event between the QRS duration groups. Similar results were seen among only the patients in the group randomized to receive shock therapy in their second VT zone, as opposed to antitachycardia pacing. Utility analysis of the various QRS duration cut-offs showed that there was no combination of QRS duration cut-offs with clinically useful sensitivity and specificity for predicting tachyarrhythmic events.

Buxton and colleagues conclude that QRS duration does not predict occurrence of VT or VF necessitating ICD therapy. They argue that there is no basis for using QRS duration to stratify patients to be considered for ICD implantation.


Until this past January, the Center of Medicare Services had included QRS duration of ≥ 120 m/sec as a major criterion for ICD insertion for primary prevention of sudden death. This opinion was primarily based on observations made in the Multicenter Automatic Defibrillator Implantation Trial II (MADIT II). However, in that trial, the observation related to QRS duration was a substudy analysis result. Subsequently, the Sudden Cardiac Death-Heart Failure Trial showed no consistent relationship between QRS duration and VT risk. The data in this study confirm that QRS duration is not a reliable predictor. This sequence of events reemphasizes some of the problems inherent in basing a decision on substudy results from a single trial.

I have some reservations about the current study. The study group analyzed had only a small proportion of patients with a left bundle branch block, and the largest group was classified as having a nonspecific intraventricular conduction defect. This seems unusual. In addition, it appears that QRS duration was only looked at in a univariate fashion. If patients with more severe heart failure and a left bundle branch block were not included in the trial for some reason, this may have skewed the current observations.