Contemporary Practice and Optimising Referral Pathways for Implantable Cardiac Monitoring for Atrial Fibrillation after Cryptogenic Stroke.

OBJECTIVES

Diagnosing atrial fibrillation (AF) in patients following Cryptogenic stroke (CS) has therapeutic implications that can reduce the risk of further strokes. However, it's indolent and paroxysmal nature makes this challenging. Prolonged rhythm monitoring using implantable loop recorders (ILRs) can significantly increase the AF detection rate in the clinical trial paradigm. Whether this can be translated to real-world practice is unknown. An evaluation of referral pathways, workload and real-world efficacy may help select patients and inform service development.

MATERIALS AND METHODS

Retrospective review of all patients with CS referred to a tertiary electrophysiology referral hospital for ILR implantation between February 2017 and October 2020 for AF detection was conducted. The electronic health record was used to determine demographic and mortality data. Remote monitoring was used to identify AF occurrence.

RESULTS

107 patients were included. The average time from stroke to ILR implantation was 10.5 (5.9-18.6) months. The average monitoring duration was 18.1 ± 11.2 months with 15 (14.0%) patients diagnosed with AF and commenced on anticoagulation. One diagnosis were made in the first 30 days whereas 11 (73%) were made within 12 months. Paroxysmal AF episodes ranged from 6 min to 13 h. Patients with CHADS-VASc >3 were more likely to have AF (20.3% vs 4.7%, p = 0.02). Age was independently associated with AF detection after multi-variate regression. 352 ± 1171 unique events were recorded per patient, 75% of which were for suspected AF. External manufacturer-led triage of transmissions reduced transmission volume by 33%.

CONCLUSIONS

ILR-based AF detection rate was high among referred CS patients, despite implantation occurring relatively late. Older patients may be less likely to be referred despite positive correlation between age and AF detection. Although recording algorithms and external triage reduced transmission volume, specialist analysis was required to manage the ILR event burden.