ECG-based prediction of postoperative atrial fibrillation with the support of artificial intelligence

Objectives:
This study investigates the association between perioperative conduction disorders and atrial fibrillation (AF) following cardiac surgery, condition linked to long-term heart failure, stroke, and mortality with the help of artificial intelligence (AI) methods.
Methods:
Perioperative 12-lead electrocardiograms (ECGs) from 3,405 patients undergoing myocardial revascularization, valve surgery, aortic surgery, or combinations thereof were analysed. Clinical and electrocardiographic parameters, including interatrial, bundle branch, and atrioventricular blocks, were compared between patients with and without AF recurrence during hospitalization. Variables showing significant differences were further analysed using multivariate logistic regression models. Statistical significance was defined as p