Automatic measurement of short-term variability of repolarization on human intracardiac electrograms predicts VT/VF events.

Background: Implantable cardioverter-defibrillators (ICDs) have proven highly effective in reducing sudden cardiac death by terminating life-threatening arrhythmias. However, their reactive nature underscores the need for preventive strategies to further improve patient outcomes. Short-term variability (STV) of repolarization has emerged as a promising predictive marker for ventricular arrhythmias. Despite its potential, the feasibility of implementing STV monitoring in real-world ICD settings remains uncertain owing to challenges such as noise, abnormal depolarizations, and device-specific constraints. Objective: This study aimed to evaluate the feasibility of STV monitoring in clinical practice using anonymized electrogram recordings from Medtronic Cobalt/Crome ICD and cardiac resynchronization therapy defibrillator devices. Methods: The dataset includes 636 monomorphic ventricular tachycardia episodes, 107 ventricular fibrillation events, and 6000 baseline recordings. Cardiac beats with abnormal repolarization patterns or noise were excluded to ensure reliable STV estimation. STV values were then compared within individuals (baseline vs arrhythmic events) and across groups (patients with and without arrhythmic events). Results: STV values were significantly elevated before arrhythmic events compared with baseline 5.8 ms (interquartile range, 3.3-7.6) vs 4.3 ms (interquartile range, 2.8-7.6) ( Conclusion: This study supports the role of STV as a marker for arrhythmic risk, evidenced by significant elevations in STV preceding arrhythmic events. It also emphasizes the importance of signal preprocessing to mitigate confounding factors and enhance the predictive accuracy of the STV parameter.