LAUSR

Obstructive sleep apnea (OSA), the most common upper-airway disease, is closely associated with cardiovascular risk. However, the early detection of OSA is challenging, relying on polysomnography (PSG) or portable monitor (PM) in hospitals. A novel technology, photoplethysmography (PPG), has been developed for OSA screening. But there has been limited validation of PPG-based smart devices compared to PSG or PM. This study aimed to investigate the feasibility and verify the validity of PPG-based smart devices in the detection of OSA in real-world settings. A total of 119 consecutive outpatients, were recruited from the Chinese PLA General Hospital from Sep 29 to Nov 10, 2019. Participants were simultaneously tested with a smart watch, and PSG or PM for a whole night sleep. In all, 102 patients (48.5±13.7 years old, female 23.5%) were into the final analysis after excluding 17 patients with poor quality of PPG signals. Among them, 83 patients (81.3%) were diagnosed with OSA. Compared to PM, the corresponding accuracy, sensitivity and specificity of PPG-based smart device in predicting moderate to severe OSA patients ( Apnea Hypopnea Index, AHI ≥15) were 87.9% (95% Confidential Interval, CI: 78.8%–94.1%), 89.7% (95% CI: 75.8%–97.1%) and 86.0% (95% CI: 72.1%–94.7%), respectively. Compared to PSG, the accuracy, sensitivity and specificity in predicting all the sleep apnea patients (AHI ≥5) were 81.1% (95% CI: 72.1%–88.2%), 76.5% (95% CI: 50.1%–93.2%) and 100% (95% CI: 29.2%–100%), respectively. Moreover, for moderate to severe OSA patients (AHI ≥15), the predictive ability of PPG-based smart device in OSA was of no significant difference compared to PM (P=0.75) or PSG (P=0.52). The PPG-based smart device demonstrated good performance in detecting OSA. Nevertheless, it requires further validation in a large-scale population on screening OSA. Type of funding source: None