The additional value of an algorithm for atrial fibrillation at the stroke unit

Area of Science:

• Cardiology
• Neurology
• Medical Technology

Background:

• Detecting paroxysmal atrial fibrillation (AF) during inpatient cardiac telemetry is challenging.
• Existing methods rely on routine assessment by stroke unit teams, which may have limitations.

Purpose of the Study:

• To evaluate the diagnostic yield of an automated AF detection algorithm.
• To compare automated detection with routine stroke unit team assessment in patients with recent ischemic stroke or TIA.

Main Methods:

• Prospective inclusion of patients with acute ischemic stroke or TIA undergoing telemetry.
• Utilized the ST segment and arrhythmia monitoring (ST/AR) algorithm with AF alarms deactivated.
• Compared algorithm detections with routine stroke unit team evaluations after 24 hours.

Main Results:

• The ST/AR algorithm suggested AF in 4.7% of patients, compared to 1.2% detected by routine analysis.
• The algorithm correctly diagnosed AF in 2.6% of patients, identifying cases missed by routine assessment.
• Interrater reliability between the algorithm and routine analysis was low (κ = 0.388).

Conclusions:

• Automated AF detection offers additional value for identifying paroxysmal AF in ischemic stroke patients.
• The ST/AR algorithm can detect cases missed by routine stroke team analysis.
• Automated AF detections require careful clinical evaluation to distinguish true positives from false positives.