Quantifying abnormal QRS peaks using a novel time-domain peak detection algorithm: Application in patients with cardiomyopathy at risk of sudden death

Abstract

Abnormal components in the QRS complex on the surface electrocardiogram have been used to predict sudden cardiac death in patients with heart disease. We propose a novel method to automate detection of abnormal peaks within the QRS complex. The approach involves identification of such peaks from consecutive unfiltered 10-beat QRS averages. A simulation using synthetic QRS peaks is conducted to assess the methods robustness to noise. The performance of the method is tested using high-resolution precordial lead electrocardiograms recorded from normal subjects and patients with cardiomyopathy. The 10-beat average performance is compared to a 100-beat average, as is commonly used in other state-of-the-art QRS component algorithms, and shown to be more sensitive in detecting abnormal QRS peaks. The clinical performance is tested amongst the cardiomyopathy patients and the method is shown to discriminate those at risk of sudden cardiac death with high sensitivity and specificity.