Beat-to-beat three-dimensional ECG variability predicts ventricular arrhythmia in ICD recipients

Background Methodological difficulties associated with QT measurements prompt the search for new electrocardiographic markers of repolarization heterogeneity. Objective We hypothesized that beat-to-beat 3-dimensional vectorcardiogram variability predicts ventricular arrhythmia (VA) in patients with structural heart disease, left ventricular systolic dysfunction, and implanted implantable cardioverter-defibrillators (ICDs). Methods Baseline orthogonal electrocardiograms were recorded in 414 patients with structural heart disease (mean age 59.4 ± 12.0; 280 white [68%] and 134 black [32%]) at rest before implantation of ICD for primary prevention of sudden cardiac death. R and T peaks of 30 consecutive sinus beats were plotted in 3 dimensions to form an R peaks cloud and a T peaks cloud. The volume of the peaks cloud was calculated as the volume within the convex hull. Patients were followed up for at least 6 months; sustained VA with appropriate ICD therapies served as an end point. Results During a mean follow-up time of 18.4 ± 12.5 months, 61 of the 414 patients (14.73% or 9.6% per person-year of follow-up) experienced sustained VA with appropriate ICD therapies: 41 of them were white and 20 were black. In the multivariate Cox model that included inducibility of VA and use of beta-blockers, the highest tertile of T/R peaks cloud volume ratio significantly predicted VA (hazard ratio 1.68, 95% confidence interval 1.01 to 2.80; P = .046) in all patients. T peaks cloud volume and T/R peaks cloud volume ratio were significantly smaller in black subjects (median 0.09 [interquartile range 0.04 to 0.15] vs. median 0.11 [interquartile range 0.06 to 0.22], P = .002). Conclusion A relatively large T peaks cloud volume is associated with increased risk of VA in patients with structural heart disease and systolic dysfunction.