Aims: Patients with left ventricular systolic dysfunction and electrocardiographic QRS duration (QRSd) >or=120 ms may obtain symptomatic and prognostic benefits from cardiac resynchronization therapy (CRT). However, clinical trials do not describe the methods used to measure QRSd. We investigated the effect of electrocardiogram (ECG) display format and paper speed on the accuracy of manual QRSd assessment and concordance of manual QRSd with computer-calculated mean and maximal QRSd.
Methods and results: Six cardiologists undertook QRSd measurements on ECGs, with computer-calculated mean QRSd close to 120 ms. Display formats were 12-lead, 6-limb, and 6-precordial leads, each at 25 and 50 mm/s. When the computer-calculated mean was used to define QRSd, manual assessment demonstrated 97 and 83% concordance at categorizing QRSd as < and >or=120 ms, respectively. Using the computer-calculated maximal QRSd, manual assessment demonstrated 83% concordance when QRSd was <120 ms and 19% concordance when QRSd was >or=120 ms. The six-precordial lead format demonstrated significantly less intra and inter-observer variabilities than the 12-lead, but this did not improve concordance rates.
Conclusion: Manual QRSd assessments demonstrate significant variability, and concordance with computer-calculated measurement depends on whether QRSd is defined as the mean or maximal value. Consensus is required both on the most appropriate definition of QRSd and its measurement.