# Clinical practice of continuous rhythm monitoring after embolic stroke of undetermined source

## TL;DR

This study examines the use of implantable cardiac monitors to detect atrial fibrillation in patients with undetermined stroke causes and finds that while effective, it requires significant healthcare resources.

## Contribution

The study quantifies the healthcare workload associated with ICM remote monitoring for AF detection in ESUS patients.

## Key findings

- 48 out of 172 patients were diagnosed with AF using ICM remote monitoring.
- On average, 20.3 trained physician workhours are needed to diagnose one AF case.
- ICM alarms were frequent, even in high-risk patients.

## Abstract

Embolic stroke of undetermined source (ESUS) accounts for up to 20% of ischemic strokes annually. Undetected atrial fibrillation (AF) is one important potential underlying cause. For AF, oral anticoagulation has evolved as the most preferable means of secondary stroke prevention. To detect unrecognized paroxysmal AF, long-term ECG monitoring is required, and implantable cardiac monitors (ICM) appear most suitable. Yet, ICMs are particularly costly, implantation is invasive, and remote monitoring places a personnel burden on health care providers. Here, we use data from a large cohort of ESUS patients to systematically analyze the effort of ICM remote monitoring for AF diagnosis and the strain on health care providers.

From a prospective, single-center, observational ESUS registry, we analyzed all ICM-equipped patients post-ESUS (n = 172) between January 1st, 2018, and December 31st, 2019. Through January 2nd, 2023, 48 patients (27.9%) were diagnosed with AF by ICM remote monitoring. During follow-up, a total of 29,180 remote monitoring episodes were transmitted, of which 17,742 were alarms for AF. A systematic estimation of workload revealed that on average, 20.3 trained physician workhours are required to diagnose one patient with AF.

ICM remote monitoring is useful to diagnose AF in cohort of post-ESUS patients. However, the number of ICM alarms is high, even in a cohort at known high risk of AF and in whom AF detection is therapeutically consequential. Improved automated event classification, clear recommendations for ICM interrogation after AF diagnosis, and a careful patient selection for ICM monitoring are warranted.