Consumer-Led Screening for Atrial Fibrillation

Background There are limited data on mobile health detection of prevalent atrial fibrillation (AF) and its related risk factors over time. Objectives This study aimed to report the trends on prevalent AF detection over time and risk factors, with a consumer-led photoplethysmography screening approach. Methods 3,499,461 subjects aged over 18 years, who use smart devices (Huawei Technologies Co.) were enrolled between October 26, 2018, and December 1, 2021. Results Among 2,852,217 subjects for AF screening, 12,244 subjects (0.43%; 83.2% male, mean age 57 ± 15 years) detected AF episodes. When compared with 2018, the risk (adjusted HRs, 95% CI) for monitored prevalent AF increased significantly for subjects when monitoring started in 2020 (adjusted HR: 1.34; 95% CI: 1.27-1.40; P < .001) or in 2021 (adjusted HR: 1.67; 95% CI: 1.59-1.76; P < 0.001). Of the 961,931 subjects who screening for both AF and OSA, 18,032 (1.9%, 97.8% male, mean age 44 ±17 years) were identified as high risk for OSA, which resulted in a 1.5-fold increase (95% CI: 1.30-fold to 1.75-fold) in the prevalent AF. A total of 5,227 (53.3%, 5,227/9,797) subjects were effectively followed up, from which 4,903 (93.8%, 4,903/5,227) subjects were confirmed with the diagnosis of AF, by the mAFA Telecare Team health providers. Conclusions Photoplethysmography-based smart devices can facilitate screening for AF with >93% confirmation of detected AF episodes even for the low-risk general population, highlighting the increased risk for detecting prevalent AF and the need for modification of OSA that increase AF susceptibility. (Mobile Health [mHealth] Technology for Improved Screening, Patient Involvement and Optimizing Integrated Care in Atrial Fibrillation [mAFA (mAF-App) II study]; ChiCTR-OOC-17014138)

and untreated AF episodes would still lead to adverse outcomes, given a similarly poor prognosis of symptomatic and asymptomatic AF. 5 The proliferation of mobile health (mHealth) and smart devices permits much earlier detection for AF, especially subclinical AF, in general population. 6,7 Single/multilead electrocardiograms (ECGs), photoplethysmography (PPG), and oscillometry devices can be employed into the wearables to detect AF, with a validated diagnostic ability comparable to standard 12-lead ECGs. 8,9 Greater AF detection has been associated with more prolonged, frequent monitoring. 9  including the pre-mAFA phase of AF screening, also called the Huawei Heart Study, using Huawei smart devices. This phase investigated the incidence of AF identified with PPG-based screening strategy among the general population. 6 Those with identified AF would be considered for entry into the mAFA II trial to validate the integrated ABC (avoid stroke with anticoagulants, better symptom management, cardiovascular and other comorbidities risk management) care supported by mHealth technology in the management of AF.
The present ancillary analysis from the mAFA-II Trial Long-Term Extension Cohort aimed to describe trends on prevalent AF detection in the general population over time with consumer-led screening for AF.

METHODS
The design and principal findings from the Huawei Heart Study have been previously reported. 6 AF screening was conducted using PPG-based Huawei smart devices (Huawei Technologies Co.) in the general population. The monitored suspected AF cases were further confirmed by health providers in the mAFA Telecare center and network hospitals, with clinical evaluation, ECG, or 24-hour Holter monitoring. This screening approach has been previously reported in prior reports from this program. 6,10 In brief, the subjects aged over 18   vi. Diabetes was defined as a fasting blood sugar level of 126 mg/dL (7 mmol/L) or higher, a random blood sugar level of more than 200 mg/dL (11.1 mmol/L).
When the study participants filled in the questionnaire, these risk factors were required to be confirmed by doctors.

SLEEP APNEA SCREENING WITH PPG SMARTER.
A machine-learning model with PPG signals, including green light, infrared light, and red light sources, has been developed to monitor blood oxygen saturation. 12    Data with a normal distribution were presented as mean AE SD. Data with a non-normal distribution were presented as median (IQR).      Figure 3).
OSA RISK AND DETECTED AF. There were 961,931 subjects (86.9% male, mean age 37 AE 14 years) who were screened for both AF and OSA risk using PPG smart devices. Among these, 6,120 subjects (0.6%,  Abbreviations as in Table 1.

DISCUSSION
In this large, prospective, population-based, consumer-led screening study using smartwear conducted over 3 years, our main findings are as follows: Our prior validation studies reported that the diagnostic sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of mobile phones with PPG for AF detection were over 94%, compared with 12-lead ECG. 8,9 In this present study involving 3 million subjects with a mean age of 37 years over 3 years, the confirmation of suspected AF by PPG devices remained over 93%, consistent with the previous report from the Mafi-II trial Long-term Extension Cohort. 13 In the WATCH AF trial (Smart-WATCHes for Detection of Atrial Fibrillation) using a smartwatch-based PPG algorithm, there was a sensitivity of 93.7%, a specificity of 98.2%, a positive predictive value of 97.8%, and a negative predictive value of 94.7%, respectively, as well as an overall accuracy of 96.1% for AF detection among subjects, with a mean age 76 years. 14 Thus, affordable, easyfor-use, consumer-led PPG-based smartwear can be a good screening tool, not only for an elderly population with comorbidities, but also for mass screening in the low-risk general population, with more prolonged, frequent monitoring. The increased secular trends on detected prevalent AF are perhaps affected by increasing cardiovascular risk factors that enhance AF susceptibility in the Chinese population. Over the last decade, obesity has increased by 67.6% in the Chinese adult population, whereas the mean physical activity has reduced from 385.9 MET h/7 d to 212.8 MET h/7 d. 16 Growing trends are also seen in the prevalence of hypertension, which has increased to 27.9% in adults over 18 years, 16 whereas the all-age prevalence of diabetes rose from 3.7% to 6.6% from 1990 to 2016. 17 On the other hand, higher levels of cardiovascular health are associated with decreased risk of developing AF. 18 Moreover, OSA was the most popular user-reported risk factor in the present screening study. A recent meta-analysis of the global prevalence and burden of OSA demonstrated that the prevalence was highest in China, followed by the United States, Brazil, and India, with an estimated 936 million (95% CI: 903-970 million) men and women aged 30 to 69 years having mild-to-severe OSA (AHI $5), and 425 million (95% CI: 399-450 million) having moderate-to-severe OSA (AHI $15) globally. 19 The present large, populationbased screening study found a 1.5-fold increase in the prevalent AF with high-risk OSA (detected via smartwear), suggesting the need to control risk factors that increase AF susceptibility.  of paroxysmal AF. 20 The first detection time of AF burden of <50% per 24 hours was 4 days by active measurement and 2 days by periodic measurement. 9 Instead of one-off ECGs, PPG-based smart devices allow continuous monitoring and would permit a much earlier detection of paroxysmal AF or asymptomatic AF, allowing the timely introduction of therapies to protect patients, not only from the consequences of the arrhythmia, but also from progression of AF from an easily treated condition to an utterly refractory problem.
Indeed, the lifestyle and risk factor modification interventions are increasingly associated with reduced AF burden. 21 However, these lifestyle factors cannot be considered in isolation; for example, OSA would contribute to hypertension, diabetes, and HF.
Hence, an integrated care approach would be required to fully implement clustered risk management in the AF patient, not just focusing on individual risk factors. The use of smart technology may support implementation of the integrated approach aimed at both primary and secondary prevention. 22 Also, the possibility of smartwear to detect risk factors such as OSA and patient-centered risk factor mitigation would be consistent with the move toward a more holistic or integrated care approach to AF management 23 that is now recommended in guidelines. 24  In the present largest screening study involving 2,852,217 subjects over 3 years for atrial fibrillation (AF) and 979,013 subjects for both AF and obstructive sleep apnea (OSA), photoplethysmography (PPG)-based smart devices can facilitate screening for AF with >93% confirmation of detected AF episodes, even for the low-risk general population, with more prolonged monitoring. High risk for OSA resulted in a 1.5-fold increase in the prevalent AF. This consumer-led AF screening approach highlights the increased risk for detecting prevalent AF episodes over time and the need for modification of OSA and other risk factors that increase AF susceptibility.