Investigation of monitoring technologies to deliver ‘pill-in-the-pocket’ oral anticoagulation in Atrial Fibrillation

Abstract

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia and AF-related strokes account for approximately a quarter of all strokes. Current guidelines recommend continuous, lifelong oral anticoagulation (OAC) based on co-morbidities, irrespective of AF burden or pattern. Nevertheless, emerging evidence suggests that stroke risk is dynamic and modulated by AF episodes. A rhythm-guided ‘pill-in-the-pocket’ OAC, where anticoagulation is initiated only during and shortly after AF episodes, may be a safe and effective alternative. This approach requires accurate, reliable real-time AF detection with timely patient notification.
This thesis evaluates the diagnostic performance, adherence and participants’ engagement across three closed-loop AF monitoring systems: implantable cardiac monitors (ICMs) integrated with the bespoke SMART-ALERT software, the Apple Watch and the CART Ring.
This thesis includes four complementary studies. A systematic review and meta-analysis of rhythm-guided OAC identified major limitations in detection and patient notifications, resulting in delays in OAC initiation, highlighting the need for better integrated solutions. The multicentre UK Confirm Rx study showed that the Confirm Rx ICM diagnostic accuracy improved with longer AF duration, from a positive predictive value of 74% for AF ≥6 minutes to 100% for episodes ≥24 hours. It identified sex-based differences in R-wave amplitude and AF detection.
The WEAR-TECH ECG study evaluated single-lead electrocardiograms from the Apple Watch and the CART Ring in three UK centres. Automated detection was suboptimal with the Apple Watch missing approximately one in three AF episodes and the CART Ring missing one in eight. However, physician interpretation for both sinus rhythm and AF was excellent, reinforcing the importance of a hybrid model for clinical decision-making. Finally, the SMART-ALERT study demonstrated feasibility of integrating the SMART-ALERT software with an ICM. The software sent automatic SMS following AF episode >30 minutes, with 74% of SMS sent within 24 hour of AF onset. Participant engagement was excellent (99.6%) with a median acknowledgement time of 4.5 minute. In contrast, both the Apple Watch and the CART Ring had poor AF detection and acknowledgement rates, and suboptimal adherence.
This thesis presents a comprehensive framework for using continuous AF monitoring and automated patient notification for rhythm-guided OAC. While some challenges remain, the SMART-ALERT platform represents an important step towards an alternative stroke prevention strategy based on ‘pill-in-the-pocket’ OAC guided by continuous rhythm monitoring.
 

More information

Identifiers

ORCID for Andre Rodrigues Pontes Briosa E Gala: orcid.org/0000-0002-9145-6374

Catalogue record

Export record