Home > Cardiology > AHA 2021 > Atrial Fibrillation > New developments in remote diagnostics and monitoring of AF

New developments in remote diagnostics and monitoring of AF

Presented by
Dr Steven Lubitz, Massachusetts General Hospital, USA
AHA 2021
A novel algorithm for Fitbit wearables using photoplethysmogram (PPG) software was able to detect undiagnosed atrial fibrillation (AF) with a high positive predictive value in a large population study. Another study demonstrated that individual AF trigger testing did not lead to improved AF-related quality of life compared with symptom surveillance alone. However, fewer AF episodes were reported in the experimental arm of the study. Moreover, a significant association was observed between alcohol use and the occurrence of AF.

Dr Steven Lubitz (Massachusetts General Hospital, MA, USA) explained that early detection of AF may prevent morbidity of this condition [1]. Since smartwatches and fitness trackers are often equipped with optical PPG sensors, software algorithms that analyse PPG data can detect AF on these wearable devices. Dr Lubitz and colleagues developed a software algorithm with overlapping PPG pulse tachogram sampling and assessed its positive predictive value for undiagnosed AF in various wearable Fitbit devices.

Participants of the Fitbit Heart Study (NCT04380415) were 22 years or older, possessed a compatible Fitbit device, and did not have a known AF diagnosis. The software reported an irregular heart rhythm detection (IHRD) if the algorithm registered ≥30 minutes of irregular rhythm. Patients with an IHRD received an ECG patch to wear for 1 week. In total, 455,669 participants enrolled in the study (median age 47 years, 71% women), of whom 4,278 (1%) received an IHRD notification. Of these participants, 1,057 (24.7%) completed the 1-week ECG patch period, illustrating the limited engagement of participants in this remote clinical trial. The primary endpoint was the positive predictive value of the first IHRD for AF during ECG monitoring.

The percentage of participants receiving an IHRD was higher among men and participants ≥65 years old. The positive predictive value of IHRD for AF was 98% and similar in the predefined age and sex subgroups. IHRD showed a sensitivity of 68% for AF measurement during ECG monitoring. In addition, participants who had an IHRD notification and completed ECG patch monitoring displayed AF in 32% of the cases. This percentage was equally divided across sex and age categories. Dr Lubitz added that the ≥65 years subgroup is an important category to consider since AF in these patients is associated with an increased risk of stroke.

Dr Gregory Marcus (University of California, CA, USA) presented another study on the remote management of AF [2,3]. He argued that the risk factors for AF, such as hypertension, age, male sex, and coronary disease, are mostly chronic and immutable, and that the acute triggers of AF are less well documented. Thus, the I-STOP-Afib trial (NCT03323099) aimed to test individual AF triggers using an innovative N-of-1 study design. Eligible adult participants were randomised to a data-tracking control group (n=248) or an experimental arm (n=251) and received a KardiaMobile to track AF episodes. Participants in the experimental arm could select a presumed trigger from a menu of triggers or add a customised trigger. These participants received instructions on when to avoid or expose themselves to the selected trigger during a 6-week period. Hereafter, they received individual (N-of-1) results of the enhanced risk of AF via exposure to the selected trigger. Subsequently, participants followed a 4-week lifestyle-changing period, in which they could adjust their behaviour in response to the results. The primary outcome was the change in the Atrial Fibrillation Effect on QualiTy-of-Life (AFEQT) questionnaire in the intention-to-treat population after 10 weeks. The most commonly selected triggers were caffeine (n=53), alcohol (n=43), reduced sleep (n=31), and exercise (n=30).

The average improvement on the AFEQT in the trigger-testing arm (+1.7) was not significantly larger than the average improvement in the control arm (+0.5; P=0.17). However, patients in the experimental arm documented 40% fewer self-reported AF episodes during the 4-week lifestyle-changing period compared with patients in the control arm (P<0.0001). This effect was driven by participants who selected alcohol, dehydration, or exercise as triggers. Per-protocol analysis of the N-of-1 trials displayed significant near-term effects of alcohol exposure on AF (OR 1.77) and customised triggers on AF (OR 4.09).

Discussant Dr Mina Chung (Cleveland Clinic, OH, USA) argued that conventional randomised controlled trials deliver average results for a pre-defined population, whereas the current N-of-1 study design has the potential to address individual patients. Although subgroup analyses in standard randomised controlled trials offer stratification of the results, the N-of-1 approach promises to surpass their ability to provide a truly personalised approach. According to Dr Chung, the significant secondary outcomes of the current study support the value of N-of-1 studies for individual patients.


    1. Lubitz SA, et al. Detection of Atrial Fibrillation in a Large Population using Wearable Devices: the Fitbit Heart Study. LBS04, AHA Scientific Sessions 2021, 13–15 November.
    2. Marcus GM, et al. The Individualized Studies of Triggers of Paroxysmal Atrial Fibrillation Trial. LBS04, AHA 2021 Scientific Sessions, 13–15 November.
    3. Marcus GM, et al. JAMA Cardiol. 2021 Nov 14. Doi: 10.1001/jamacardio.2021.5010.


Copyright ©2021 Medicom Medical Publishers

Posted on