Impact of Neuroimaging Patterns for the Detection of Atrial Fibrillation by Implantable Loop Recorders in Patients With Embolic Stroke of Undetermined Source

ObjectivesAtrial fibrillation (AF) is a well-known etiology of embolic stroke of undetermined source (ESUS), although the optimal detection strategy of AF was not been fully evaluated yet. We assessed AF detection rate by implantable loop recorder (ILR) in patients with ESUS and compared the clinical characteristics and neuroimaging patterns between the patients with AF and AF-free patients.MethodsWe reviewed clinical characteristics and neuroimaging patterns of consecutive patients with who were admitted to our comprehensive stroke center for ESUS and underwent ILR insertion between August 1, 2019, and January 31, 202. The inclusion criteria were (1) 18 years of age or older; (2) classified as having cryptogenic stroke extracted from the group with undetermined stroke according to ESUS International Working Group; and (3) underwent ILR insertion during or after admission due to index ischemic events. Ischemic stroke pattern was classified as (1) tiny-scattered infarction, (2) whole-territorial infarction, (3) lobar infarction and (4) multiple-territorial infarction. Interrogations of data retrieved from the ILR were performed by cardiologists in every month after the implantation.ResultsIn this study, 41 ESUS patients who received an ILR implantation were enrolled (mean age, 64 years; male sex, 65.9%). The rate of AF detection at 6 months was 34% (14 patients), and the mean time from ILR insertion to AF detection was 52.5 days [interquartile range (IQR), 45.0–69.5]. The median initial NIH stroke scale scores were significantly greater in patients with AF than those without AF (6.5 vs. 3.0, p = 0.019). Whole-territorial infarction pattern was significantly more frequent in patients with AF than in those without AF (64.3% vs.11.1%, p = 0.002).ConclusionsHigher covert AF detection rates within the ESUS patients were most often associated with higher NIHSS and whole-territorial infarction patterns on brain imaging