Predictors of Paroxysmal Atrial Fibrillation in Patients With a Cryptogenic Stroke: Selecting Patients for Long-Term Rhythm Monitoring

Abstract

After a cryptogenic stroke, patients will often require prolonged cardiac monitoring; however, the subset of patients who would benefit from long-term rhythm monitoring is not clearly defined. Using significant predictors of AF using age, sex, comorbidities, baseline 12-lead electrocardiogram, short term rhythm monitoring and echocardiogram data, we created a risk score and compared it to previously published risk scores. Patients admitted to Montefiore Medical Center between May 2017 and June 2022 with a primary diagnosis of cryptogenic stroke or TIA who underwent long-term rhythm monitoring with an implantable cardiac monitor were retrospectively analyzed. Variables positively associated with a diagnosis of clinically significant atrial fibrillation include age (p < 0.001), race (p = 0.022), diabetes status (p = 0.026), and COPD status (p = 0.012), the presence of atrial runs (p = 0.003), the number of atrial runs per 24 hours (p < 0.001), the total number of atrial run beats per 24 hours (p < 0.001) and the number of beats in the longest atrial run (p < 0.001), LA enlargement (p = 0.007) and at least mild mitral regurgitation (p = 0.009). We created a risk stratification score for our population, termed the "ACL score." The ACL score demonstrated superiority to the CHA2DS2-VASc score and comparability to the C2HEST score for predicting device-detected AF. The ACL score enables clinicians to better predict which patients are more likely to be diagnosed with device-detected AF after a cryptogenic stroke.