Implantable loop recorders (ILRs) are small cardiac devices inserted into subcutaneous tissue to detect and record arrhythmic events for up to 3 years. The most common indications for ILR implantation are unexplained syncope, palpitation, and cryptogenic stroke (CS). In two real-world registries, ILRs were most frequently implanted because of syncope [1,2], but ILR also has proven its efficacy to detect atrial fibrillation in CS patients [3,4].

ILR implantation is currently reimbursed for 3 different indications in Korea: recurrent syncope, palpitation of unexplained cause, and CS. On November 2018, the reimbursement coverage for CS was expanded from “recurrent” to the first stroke episode. However, no nationwide reports have demonstrated whether this policy change has affected ILR utilization trends. We analyzed time-series trends in ILR implant before and after reimbursement policy change using the Korean Health Insurance Review & Assessment Service (HIRA) database. We also evaluated the diagnostic yield of ILR to detect atrial fibrillation or flutter (AF/AFL) in CS patients.

This nationwide cohort study investigated the ILR utilization in Korea between July 2016 and October 2021. The main findings were: 1) total ILR implant number gradually increased for both CS and non-CS indications; 2) about a third of patients underwent ILR implantation for CS indication and the number of ILR implantation for CS indication significantly increased after the expanded reimbursement policy was implemented in November 2018; and 3) the diagnostic yield of ILR for AF/AFL detection was 26.3% in CS patients over 3 years, which was significantly higher among early implants compared to the late implant group.

In this study, we first evaluated whether implementation of a policy to allow ILR implant for the first stroke instead of only after recurrent stroke resulted in an increase in ILR usage in CS patients in Korea. The number of ILR implants for CS indication significantly increased after the policy change, and CS recently accounted for almost a third of all ILR implantation indications, which is a three-fold increase from when reimbursement began. The stroke-to-ILR implant interval also significantly decreased after the policy change, which possibly resulted in younger patients with ILR implants after the policy change than before.

There is a clinical significance of AF/AFL detection in CS patients because it has an impact on secondary stroke prevention, which is possibly associated with higher anticoagulation incidence, although controversial [6,7]. The AF/AFL detection rate in our cohort was similar or relatively higher than in previous studies, which showed that most AF detections by ILR occurred during the first few months [8–12]; however, the cumulative incidence of AF detection continued to increase over a year, albeit slowly. Our study results were also consistent with these findings, but the most distinct observation from our study population was that the interval from the index stroke to ILR implant was remarkably long [3,4,9–13]. This is not only because of the different study design and population in our study, but also because ILR implants were reimbursed for only recurrent stroke until November 2018 in Korea. The reason why the interval was substantially long even after the policy change is likely because “recurrent” stroke still accounted for a large portion of CS indications, especially in the early period after the policy change. This might explain why the AF/AFL detection rate was not different despite the fact that the expanded policy led to a significant increase in ILR implant for CS. Considering that the detection rate was significantly higher in the early than in the late implant group, it is possible that the policy change impacted the AF/AFL detection rate in CS patients. However, ILR still seems useful to detect AF/AFL in patients after a long interval between stroke and ILR implant, presumably those who experienced recurrent stroke, even though the diagnostic yield in AF detection was higher in patients who were implanted with ILR within 2 months after their stroke. The interval from the stroke to ILR implant was still long in patients with non-recurrent events (mean 1,289 d). We attribute this to the relatively late timing of ILR insertion for CS in Korea, partly because of possible time lag between policy change and implementation in clinical practice.

A significant portion of CSs is attributed to cardioembolism. Cardioembolic stroke has a worse prognosis than strokes of other etiologies, and has higher risk of recurrence, severe disability, and mortality; thus, CS warrants comprehensive evaluation, particularly cardiac diagnostics [14–18]. However, echocardiography was performed in just over half of CS patients against our expectations, which was one of the major differences in our study from previous studies. Nevertheless, those studies included a relatively small number of patients from a single or limited number of centers or they had a prospective study design following a recommended protocol, while our study is a nationwide population-based retrospective cohort study [4,9,12,13,19]. Therefore, our study likely reflects the real-world practice for CS work-up in Korea. There is another possibility that other imaging modalities such as cardiac magnetic resonance imaging or computed tomography were performed as an alternative to echocardiography. For example, transcranial Doppler exam was performed in 29.6% of CS patients.

This study has several limitations. First, stroke etiology could not be confirmed because of the inherent limitations of claims data. The etiology of the study population is most likely to be cryptogenic considering the indications for ILR implantation coverage in Korea. We also incorporated 2 other criteria, including no prior history of AF/AFL and anticoagulation, to increase validity. However, patients with prior stroke, whether cryptogenic or not, could have a non-CS indication for ILR implantation. Second, we presumed that atrial fibrillation or flutter diagnosed after ILR implant was ILR-driven. However, we cannot confirm that all those events were detected by ILR nor which criteria were used for clinical diagnosis of AF/AFL in terms of duration [4]. Third, there was a chance of missing diagnoses of AF/AFL in patients without regular ILR interrogation. Lastly, we could not evaluate whether AF/AFL detection by ILR is beneficial for patient improving clinical outcomes including secondary stroke prevention.

This study investigated temporal trends and changes in ILR utilization following the expanded coverage policy for ILR implantation. We identified that the number of ILR implantations for CS indication significantly increased after the expanded reimbursement policy and CS accounted for a third of all indications. We also showed that the incidence of AF/AFL detection was higher when ILR was implanted within 2 months from the index stroke than after more than 2 months. We hope that our findings can be the basis for further studies to demonstrate the clinical benefit of AF/AFL detection using ILR and to determine the best timing of ILR implantation.