Prediction of postoperative atrial fibrillation with postoperative epicardial electrograms. (31st December 2022)

• Record Type: Journal Article
• Authors: Feilberg Rasmussen, Louise; Andreasen, Jan Jesper; Riahi, Sam; Lip, Gregory Y. H.; Lundbye-Christensen, Søren; Melgaard, Jacob; Graff, Claus
• Is Part Of: Scandinavian cardiovascular journal. Volume 56:Number 1(2022)
• Page Start: 378
• Page End: 386
• Publication Date: 2022-12-31
• Subjects: Cardiac surgery -- new-onset postoperative atrial fibrillation -- atrial electrocardiogram -- ECG -- prediction model; Cardiovascular system -- Diseases -- Periodicals; 617.41
• DOI: 10.1080/14017431.2022.2130421
• Languages: English
• ISSNs: 1401-7431

Abstract

Abstract: Objectives. New-onset postoperative atrial fibrillation (POAF) is a common complication after cardiac surgery. The arrhythmia often entails a longer hospital stay, greater risk of other complications, and higher mortality both short- and long-term. An investigation of the use of early atrial electrograms in predicting POAF in cardiac surgery was performed. Design. In this prospective observational study, a total of 99 consecutive adult patients undergoing coronary artery bypass grafting, valve surgery or both were included. On the first postoperative morning, standard 12-lead electrograms (ECG), unipolar atrial electrograms (aEG), and vital values were recorded. The outcome was new-onset POAF within one month postoperatively. Results. Three multivariable prediction models for POAF were formed using measurements derived from the ECG, aEG, and patient characteristics. Age, body mass index, and two unipolar electrogram measurements quantifying local activation time and fractionation were strongly associated with the outcome POAF. The performance of the POAF prediction models was assessed through receiver operating curve characteristics with cross-validation, and discrimination using the leave-one-out-method to internally validate the models. The cross-validated area under the receiver operating characteristic curve (AUC) was improved in a prediction model using atrial-derived electrogram variables (AUC 0.796, 95% CI 0.698–0.894), compared with previous ECG and clinical models (AUC 0.716, 95% CI 0.606–0.826 and AUC 0.718, 95% CI 0.613–0.822, respectively). Conclusions. This study found that easily obtainable measurements from atrial electrograms may be helpful in identifying patients at risk of POAF in cardiac surgery.