Background: Although QRS duration is known to be a predictor of mortality in patients with left ventricular dysfunction, our purpose was to evaluate the prognostic power of computer-measured QRS duration in a general medical population.
Methods: Analyses were performed on the first electrocardiogram digitally recorded on 46,933 consecutive patients at the Palo Alto Veterans Affairs Medical Center between 1987 and 2000. Patients with electrocardiograms exhibiting Wolff-Parkinson-White were excluded (n = 44), and those with bundle branch block or electronic pacing were considered separately, leaving 44,280 patients for analysis (mean age 56 +/- 15 years; 90% were males). There were 3659 (8.3%) cardiovascular deaths (mean follow-up of 6.0 +/- 3.8 years).
Results: A survival plot showed significant separation according to a QRS duration score. After adjustment in the Cox model for age, gender, and heart rate, the QRS duration score was a strong independent predictor of cardiovascular mortality. For every 10-ms increase in QRS duration, there was an 18% increase in cardiovascular risk. The results were similar in patients with an abnormal electrocardiogram, a bundle branch block, and a paced rhythm.
Conclusion: Quantitative QRS duration was a significant and independent predictor of cardiovascular mortality in a general medical population.