Abstract 13293: Detection of Atrial Fibrillation by Implantable Loop Recorders Following Cryptogenic Stroke: A Retrospective Study of Predictive Factors and Outcomes

Abstract

Introduction: Cryptogenic strokes account for up to 40% of ischemic strokes. Atrial fibrillation (AF) is a known cause of ischemic stroke. Current data shows that occult AF can be detected by implantable devices at higher rates than conventional cardiac rhythm monitoring. There are, however, limited data available on risk factors and outcomes associated with AF detection by implantable loop recorders (ILRs). Objective: To investigate the risk factors and outcomes associated with occult AF detected by ILR in patients with cryptogenic stroke. Methods: We conducted a retrospective chart review of patients admitted with cryptogenic stroke at Ascension St John Hospital and Ascension Macomb-Oakland Hospital in Michigan who had ILRs placed from 1/1/2016 to 1/31/2020. Data were collected on demographics, comorbidities, treatment and outcomes. AF detection was defined as continuous AF for 30 seconds. Data were analyzed using Student’s t-test, the χ2 test and logistic regression. Results: We reviewed 172 patients, 52.3% male, 56.4% white, mean age 62.7 ± 13.6 years. The incidence of AF detection by ILR was 14% (24/172) over a mean follow-up of 12.75 ± 10.71 months. The mean duration of monitoring prior to AF detection was 4.5 months (range:1 day to 14 months). The median duration of AF was 6 minutes (range: 37.2 seconds to 11.3 hours). From univariable analysis, older age (p=0.03), male sex (p=0.09), embolic stroke pattern on imaging (p=0.06), and lack of AF symptoms (p=0.001) were associated with AF detection by ILR. From multivariable analysis, patients with detected AF were more likely to be older (OR=1.04, p=0.04), male (OR=3.6, p=0.03), asymptomatic (OR=6.3, p=0.01), and have an embolic stroke pattern on imaging (OR=3.3, p=0.04). 95.7% of patients with confirmed AF were started on anticoagulation for secondary stroke prevention. There was no difference in the incidence of stroke post-hospitalization between those with AF detection and those without (16% vs. 16.4%, p=0.96). Conclusions: In patients with cryptogenic stroke, age, gender, stroke pattern, and lack of AF symptoms are independent predictors of occult AF detection by ILR. Most patients with confirmed AF were started on anticoagulation for secondary stroke prevention and had low stroke recurrence rates.