Superior diagnostic performance of beat-to-beat T-wave angle monitoring for ventricular arrhythmia prediction during cardiac rehabilitation
Abstract
Risk stratification for ventricular arrhythmias in ischemic heart disease (IHD) traditionally relies on left ventricular ejection fraction (LVEF) and exercise-derived variables such as blood pressure response, Duke treadmill score, Veterans score, and rate–pressure product. However, these static parameters may not adequately capture beat-to-beat electrical instability. Dynamic T-wave angle variation, reflecting periodic repolarization instability, may provide superior predictive value. To determine whether dynamic T-wave angle variation during cardiac rehabilitation is superior to conventional pre-rehabilitation risk markers in predicting ventricular arrhythmias. We conducted a prospective cohort study including 78 patients with IHD (31 women). LVEF was <35% in 27 patients, 35–49% in 15, and >50% in the remainder. Fifty-two underwent percutaneous coronary intervention and the rest coronary bypass surgery, with no difference in arrhythmic outcomes between revascularization strategies. All patients completed a 12-session cardiac rehabilitation program with continuous 12-lead telemetry. Beat-to-beat T-wave angle was analyzed in real time. A sudden change >10° was predefined as abnormal. The primary endpoint was ventricular arrhythmias (paired premature ventricular complexes, ventricular bigeminy, or non-sustained ventricular tachycardia). Predictive performance was assessed using sensitivity, specificity, positive and negative predictive values, and comparative diagnostic accuracy analysis. Between-method comparisons were performed using chi-square testing, with p<0.05 considered significant. A T-wave angle change >10° predicted ventricular arrhythmias with sensitivity 89% and specificity 96% (p<0.001 vs no angle change). False-negative cases were attributable to supraventricular arrhythmias (17 atrial tachycardia, 4 atrial flutter), in which T-wave angle variation was not statistically significant. Among conventional predictors, only LVEF <35% (sensitivity 77%, specificity 94%; p=0.01) and exercise-induced hypotension (sensitivity 78%, specificity 97%; p=0.008) were significant, but both showed lower sensitivity compared with T-wave angle variation (p<0.05 for sensitivity comparison). Other markers demonstrated inferior diagnostic performance: Duke score (44%/86%), Veterans score (47%/89%), flat and hypertensive blood pressure responses (28%/74% and 35%/68%), and rate–pressure product (56%/87%), none reaching comparable discriminative capacity. Dynamic T-wave angle variation >10° during supervised cardiac rehabilitation showed superior sensitivity and high specificity for predicting ventricular arrhythmias compared with conventional pre-rehabilitation risk markers. Beat-to-beat repolarization monitoring may enhance electrical risk detection beyond traditional stratification strategies in patients with ischemic heart disease.Comparative ROC CurvesForest Plot – Diagnostic Accuracy