High amplitude T-wave alternans precedes spontaneous ventricular tachycardia or fibrillation in ICD electrograms

BACKGROUND T-wave alternans (TWA) increases acutely prior to ventricular tachycardia (VT) or fibrillation (VF) in animal studies, suggesting that it may provide a warning for VT/VF in implantable cardioverter defibrillator (ICD) patients. Clinically, measurement of surface ECG TWA requires preprocessing the input signal to reduce noise and/or analyzing more sinus beats than are recorded in ICDs as pre-onset, stored intracardiac electrograms (EGMs) before VT/VF. Our objective was to measure TWA from the few sinus EGMs stored in ICDs before spontaneous VT/VF in humans. OBJECTIVE The purpose of this study was to evaluate the technical feasibility of measuring TWA from pre-onset ICD EGMs and to measure EGM TWA before spontaneous VT/VF in humans. METHODS We developed a method to measure EGM TWA as a simple average (AVE) of peak-to-peak alternans. Using simulation, we determined the effect of ICD signal processing on EGM TWA for durations comparable to those in pre-onset EGMs. We then applied this method to pre-onset ICD EGMs that preceded 101 episodes of sustained VT/VF in 10 patients. In 6 of these patients, EGM recordings in atrial pacing and sinus rhythm provided control data. RESULTS In simulation, the AVE method discriminated input TWA differences >= 15 AV. In patients, EGM TWA was 78 +/- 62 mu V prior to VT/VF vs. 13 +/- 10 mu V in control recordings (p<.0001). Eighty percent of pre-onset measurements exceeded 30 mu V, while 95% of control measurements were Less than 30 mu V. CONCLUSIONS A simple averaging method can measure TWA preceding VT/VF in stored ICD EGMs. Pilot data indicate that high-amplitude EGM TWA usually precedes spontaneous VT/VF and is infrequent in control recordings. They provide a rationale for developing ICD technology to measure EGM TWA continuously, both to warn patients and to initiate pacing algorithms to prevent VT/VF.