Photoplethysmography-based atrial fibrillation detection in patients after crytpogenic stroke.

Introduction: Undiagnosed atrial fibrillation (AF) is a potential underlying cause of cryptogenic stroke. Prolonged screening for AF using a photoplethysmography (PPG) smartwatch might offer a solution for detecting AF in patients with cryptogenic stroke. In this study, we aim to investigate this strategy by comparing AF detection rates using a PPG-smartwatch and 48 h Holter monitor. Methods: From December 2019, patients with cryptogenic stroke were included to undergo 28 days of semi-continuous AF monitoring using a Fitbit smartwatch with a PPG-based FibriCheck algorithm, with simultaneous Holter monitoring during the first 48 h. From April 2021, a detailed screening log was installed to characterize potential study participants. Results: After logged screening of 1,312 patients, enrollment was prematurely halted due to slower-than-expected inclusion rates. 40.8% of the screened patients had cryptogenic stroke, of which 92.5% were non-eligible for inclusion due to logistical, technological, and study-related challenges. Of the 43 patients enrolled, 37 completed PPG monitoring using a smartwatch. 43% of patients had PPG-detected AF in the 28 days after cryptogenic stroke. During the first 48 h, PPG-based screening detected AF in 2 patients, whereas no AF was detected using concurrent Holter monitoring. Conclusion: The PPG-smartwatch detected AF in 43% of the participants after cryptogenic stroke. However, discrepancies with concurrent Holter monitoring raise major concerns about the accuracy of the detected PPG-based AF. Moreover, the feasibility of a PPG-based screening strategy is limited due to logistical and technological challenges, partly inherent to cryptogenic stroke patients.