Wavelet analysis of bipolar endocardial electrograms for morphology based detection of ventricular tachycardias

Morphology based detection criteria may increase the specificity of the detection of ventricular tachycardia (VT) in implantable cardioverter-defibrillators. The aim of the study was to create a VT detection algorithm using the wavelet transform based morphology of bipolar right ventricular electrograms (EGM). EGM were obtained during an electrophysiological study in patients (P) with inducible monomorphic VT. The wavelet description of a template EGM during normal sinus rhythm (NSR) or atrial fibrillation (AF) characterized by 7 wavelet parameters (WP) was compared to EGM under analysis. For discrimination of NSR/AF and VT, the difference of WP was considered. Discrimination of VT from NSR/AF was achieved in all P with 100% accuracy. The description of the morphology of EGM in the wavelet domain seems to be a feasible tool for discrimination of NSR or AF from VT. The incorporation of this algorithm in modern ICD-systems might improve automated arrhythmia recognition