Abstract WP231: Risk Factors for Paroxysmal Atrial Fibrillation and Atrial Flutter Detection Following Cryptogenic Ischemic Stroke

Abstract

Introduction: Long-term post-stroke insertable cardiac monitor (ICM) implantation has resulted in increasing frequencies of paroxysmal atrial fibrillation and atrial flutter (PAF) detection in a significant proportion of cryptogenic ischemic stroke (CIS) patients. Determining risk factors for PAF detection specific to the CIS population could allow for better selection of patients for ICM implantation. Methods: A retrospective study of CIS patients (n=95; mean age 65.9 years; 56.8% female) implanted with the Reveal LINQ ICM between September 2013 to July 2015. Device implantation was performed during, or soon after, index stroke admission. The study cardiac electrophysiologist confirmed PAF diagnosis. Univariate and multivariate logistic regression analyses compared clinical, laboratory, ECG and echocardiographic variables between patients with and without PAF. Results: PAF was detected in 22/95 (23.2%) patients. Factors associated with PAF detection include older age (mean (yr) 74.0 vs. 63.4; p=0.003), prolonged PR-interval (PR>175 ms; OR 3.88, 95% CI 1.28-11.7), mitral regurgitation (MR; OR 3.36, 95% CI 1.24-9.08), left atrial diameter (LA diameter mean (cm) 3.80 vs. 3.51; p=0.039), and left atrial volume index (LAVI mean (cc/m2); 32.8 vs. 25.0; p=0.037). Controlling for age in the multivariate logistic regression model, obesity (BMI>30; OR 4.80, 95% CI 3.50-9.60) and MR (OR 4.47, 95% CI 3.15-5.79) are independently associated with PAF detection. Conclusion: Long-term cardiac monitoring identified PAF in a substantial number of CIS patients. Older age, prolonged PR, MR, and larger LA diameter and LAVI are significantly associated with PAF diagnosis. Controlling for age, significant independent risk factors include obesity and MR. A larger prospective study is warranted to confirm these findings and to identify other possible risk factors. Ultimately, a risk factor based algorithm could enable better selection of CIS patients for ICM implantation.