Background: Antitachycardia pacing (ATP) success rates as low as 50% for fast ventricular tachycardias (VTs) have been reported providing an opportunity for improved ATP to decrease shocks.
Objective: The purpose of this study was to determine how a new automated antitachycardia pacing (AATP) therapy would perform compared with traditional burst ATP using computer modeling to conduct a virtual study.
Methods: Virtual patient scenarios were constructed from magnetic resonance imaging and electrophysiological (EP) data. Cardiac EP simulation software (CARPEntry) was used to generate reentrant VT. Simulated VT exit sites were physician adjudicated against corresponding clinical 12-lead electrocardiograms. Burst ATP comprised 3 sequences of 8 pulses at 88% of VT cycle length, with each sequence decremented by 10 ms. AATP was limited to 3 sequences, with each sequence learning from the previous sequences.
Results: Two hundred fifty-nine unique ATP scenarios were generated from 7 unique scarred hearts. Burst ATP terminated 145 of 259 VTs (56%) and accelerated 2.0%. AATP terminated 189 of 259 VTs (73%) with the same acceleration rate. The 2 dominant ATP failure mechanisms were identified as (1) insufficient prematurity to close the excitable gap; and (2) failure to reach the critical isthmus of the VT. AATP reduced failures in these categories from 101 to 63 (44% reduction) without increasing acceleration.
Conclusion: AATP successfully adapted ATP sequences to terminate VT episodes that burst ATP failed to terminate. AATP was successful with complex scar geometries and EP heterogeneity as seen in the real world.
Keywords: Antitachycardia pacing; Computational cardiac electrophysiology; Monomorphic ventricular tachycardia; Self-adapting algorithms; Virtual patient.
Copyright © 2020 The Authors. Published by Elsevier Inc. All rights reserved.