Most remote monitoring alerts about atrial fibrillation are not relevant to identify clinically significant events: proposal of a new approach

Abstract

Abstract Background Cardiac Implantable Electronic Devices (CIEDs) are an important tool for detecting Atrial Fibrillation (AF) in implanted patients. However AF burden values and notifications emitted by the manufacturer's platforms are not directly related to the standard classification of AF types (paroxysmal, persistent or permanent) that are used in daily practice. Moreover, AF alerts represent the most frequent notifications for implanted patients resulting in a time-consuming review for healthcare professionals. Purpose This study intends to compare the manufacturers' atrial burden related notifications in remotely monitored (RM) patients to the detection of clinically significant events with a new proprietary algorithm. Methods From 2017 to 2020, all RM patients from 57 centers with daily atrial burden measurements available for at least 30 days and at least one atrial burden related alert were enrolled. All atrial burden related alerts emitted by the manufacturers' platforms were compared to the following clinically significant events (based on the standard classification) detected by a new proprietary algorithm: “1st recorded AT/AF episode”, “paroxysmal AF”, “increasing paroxysmal AF”, “persistent AF”, and “end of persistent AF”. Results This multicentric retrospective study analyzed, between 01/2017 and 10/2020, 2 463 RM patients with a Biotronik, Boston Scientific or Medtronic CIED (implantable defibrillator, pacemaker or implantable loop recorder), with a mean follow-up of 490 days [33–1386]. A total of 22 345 manufacturers' atrial burden related alerts were emitted while only 4 826 clinically significant events were detected by the algorithm: 1770 “1st recorded AT/AF episode”, 620 “Paroxysmal AF”, 252 “Increasing paroxysmal AF”, 1373 “Persistent AF”, and 811 “End of persistent AF”. These clinically significant events represent only 22% of the total number of atrial burden related alerts emitted by the manufacturers' platforms. Conclusion A new AF alert algorithm could have the potential to identify clinically significant AF status change in remotely monitored implanted patients while reducing the total number of alerts generated and thus the review burden for healthcare professionals. Funding Acknowledgement Type of funding sources: None.