Session S92.2

Electrocardiographic Prediction of Arrhythmias

Z Syed*, BM Scirica, CM Stultz, JV Guttag

Massachusetts Institute of Technology
Cambridge, MA, USA

Information in electrocardiographic (ECG) signals may have value in the short-term prediction of arrhythmias. This study evaluates the use of morphologic variability (MV), a recently proposed metric measuring subtle variability in the shape of ECG signals over long periods, to predict arrhythmias from continuous Holter ECG data. We assessed the relationship between MV and the composite occurrence of ventricular tachycardia and cardiac pause (150 events) in 2302 patients hospitalized following non-ST-elevation acute coronary syndrome. MV was measured on the first 24 hours of data for each patient and was studied for an association with endpoints occurring subsequently (median duration 5.8 days) during hospitalization. On univariate analysis, high MV was strongly associated with the occurrence of arrhythmias (HR 2.02; 95% CI 1.42-2.87; p<0.001). The relationship between MV and arrhythmias was consistent even after adjusting for heart rate variability and baseline clinical characteristics (adjusted HR 1.90; 95% CI 1.26-2.86; p=0.002). Our results suggest that patients with high MV are at an increased risk of arrhythmias.

(Abstract Control Number: 53)