Wrist band photoplethysmography in detection of individual pulses in atrial fibrillation and algorithm-based detection of atrial fibrillation

Aims Atrial fibrillation (AF) is the most common tachyarrhythmia and a significant cause of cardioembolic strokes. Atrial fibrillation is often intermittent and asymptomatic making detection a major clinical challenge. We evaluated a photoplethysmography (PPG) wrist band in individual pulse detection in patients with AF and tested the reliability of two commonly used algorithms for AF detection. Methods and results A 5-min PPG was recorded from patients with AF or sinus rhythm (SR) with a wrist band and analysed with two AF detection algorithms; AFEvidence and COSEn. Simultaneously registered electrocardiogram served as the golden standard for rhythm analysis and was interpreted by two cardiologists. The study population consisted of 213 (106 AF, 107 SR) patients. The wrist band PPG achieved individual pulse detection with a sensitivity of 91.7 ± 11.2% and a positive predictive value (PPV) of 97.5 ± 4.6% for AF, with a sensitivity of 99.4 ± 1.5% [7.7% (95% confidence interval, 95% CI 5.5% to 9.9%); P