Atrial Fibrillation Detection With an Analog Smartwatch: Prospective Clinical Study and Algorithm Validation.

Area of Science:

• Cardiology
• Digital Health
• Medical Devices

Background:

• Atrial fibrillation affects 4% of the global population, increasing stroke and heart failure risk.
• Diagnosis can be challenging, especially for asymptomatic or paroxysmal cases.
• Wearable devices offer potential for improved atrial fibrillation detection.

Purpose of the Study:

• To validate an algorithm for automatic atrial fibrillation detection using single-lead ECGs from a smartwatch.
• To assess the diagnostic performance and reliability of smartwatch-based ECG analysis.

Main Methods:

• Recruited 262 patients across 4 French sites for simultaneous 12-lead and single-lead ECG capture.
• Independent cardiologists reviewed ECGs; algorithm performance (sensitivity, specificity) was calculated.
• Assessed single-lead ECG quality (wave visibility, intervals, heart rate) against the 12-lead gold standard.

Main Results:

• The algorithm achieved high sensitivity (0.963) and specificity (1.000) for atrial fibrillation detection, excluding noise and other arrhythmias.
• Single-lead ECGs showed good accuracy for P waves (96.9%), QRS complexes (99.2%), and T waves (91.2%).
• Minimal differences in PR duration (<3 ms) and QRS width (<40 ms), and heart rate (<0.55 bpm) were observed compared to cardiologists.

Conclusions:

• The algorithm demonstrates excellent diagnostic performance for detecting atrial fibrillation.
• Smartwatch-derived single-lead ECGs provide sufficient quality for clinical use in routine care.