Premature Ventricular Contraction-Mediated Ventricular Fibrillation: Clinical characteristics, Application of Machine-Learning Algorithm and Outcomes of Catheter Ablation: Multicentric Case Series
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Premature ventricular contractions (PVCs) are common in patients with and without structural heart disease. In a subset of patients, PVCs are associated with malignant ventricular arrhythmias including ventricular fibrillation (VF). There are limited clinical tools available to identify which patients with PVCs are at risk for VF.
Methods
We analyzed data from an international, multi-center cohort of 61 patients who underwent catheter ablation for PVCs (41 with PVC-triggered VF, 20 controls). We evaluated the prevalence of routine 12 lead ECG characteristics in patients with PVC-triggered VF including (a) early repolarization (ER) in inferior/ lateral leads and (b) QRS notching of the sinus beat or PVC. We evaluated whether a machine learning (ML) ECG algorithm (Factor ECG) could discriminate between individuals with PVC-triggered VF and individuals with PVC and no history of VF. Explainability analyses were performed to identify which components of the ECG waveform were associated with risk prediction.
Results
In 41 patients with PVC-triggered VF, there were a median of 8 ICD shocks/per patient prior to index PVC ablation. The mean coupling interval of the PVC to the antecedent sinus beat was 313±130 ms. When compared to controls, early repolarization (39% vs. 20%) and QRS notching (71% vs. 25%) were significantly more prevalent in individuals with PVC-triggered VF. After a median ablation of 1 [IQR: 1-3]), 82% of patients remained free of VT/VF and ICD shocks over a median follow up of 400 [90–2490] days. The ML ECG algorithm demonstrated reasonable discrimination of patients with PVC-triggered VF compared to PVC without VF (AUROC 0.85 [0.56-1.0]). Anterior ST segment deviation and left bundle branch like delay of the ECG waveform were salient contributors to ML-based prediction.
Conclusions
In patients with PVC-triggered VF, routine ECG parameters including early repolarization and QRS notching were present in up to two-thirds of patients and were more prevalent compared to individuals with PVC without VF. An ML-based ECG algorithm effectively distinguished between PVC-triggered VF compared to PVC without a history of VF.
