T wave alternans (TWA) induced by pacing predicts arrhythmia vulnerability in animals and humans. To assess the predictive value of TWA measured non-invasively with exercise, we performed a prospective multicenter study in 148 pts undergoing electrophysiology study (EP). The relationships between TWA and the induction of sustained ventricular tachycardia at EP as well as arrhythmia free survival were assessed. TWA was defined as positive if sustained alternans occurred with onset heart rate ± 110 bpm. Events were death, sustained ventricular tachyarrhythmias, or appropriate defibrillator discharges. The patient population was 64% male with a mean age of 54 ± 17 years and a mean ejection fraction of 41 ± 18%. Coronary artery disease was present in 48% of pts. TWA was positive in 20% of pts, negative in 55% and indeterminate or missing in 25%. The results of EP were positive in 23%, negative in 61%, and nonspecific (ventricular fibrillation) or missing in 16%. In pts with determinate TWA and EP results (n = 92), TWA predicted EP outcome with a sensitivity of 62%, specificity of 86% and relative risk of 4.3 (p<0.0001). Follow-up was available on 111 pts with a mean duration of 303 days. There were 13 endpoint events. Kaplan-Meier analysis revealed that TWA was a significant predictor of event-free survival (EFS, Relative risk 10.6, p = 0.011). At one year, EFS was 0.817 for TWA positive patients and 0.983 for TWA negative patients. EP also predicted events (Relative risk 3.7, p = 0.029). EFS at one year was 0.771 for EP positive patients and 0.939 for EP negative patients. We conclude that non-invasive assessment of TWA with exercise is a moderately sensitive but specific predictor of EP. Further follow–up data are needed to confirm the relative predictive value of TWA and EP for arrhythmic events in this patient population.

1 Oak Park Drive

Bedford, MA 01730

781-271-1200