Abstract 16515: Atrial Strain Predicts Atrial Fibrillation in Patients With Cryptogenic Cerebrovascular Events
Introduction: Atrial fibrillation (AF) may account for a substantial number of patients with cryptogenic ischaemic cerebrovascular accidents (CVA), and has significant therapeutic implications. Echocardiography (Echo) is often performed in this setting and we sought whether the assessment of atrial strain would improve the prediction of AF.
Methods: In this case control study (n= 114) we followed 700 patients with cryptogenic CVA, and included all patients who had an echo following the CVA. Patients were excluded from analysis if an alternate cause for the CVA was identified or if they were known to have AF. Patients were followed for 1 year, over which time 57 were identified with AF (using a combination of admission codes, medical records, outpatient correspondence and ECGs). A 1:1 match was performed with 57 non AF cases matched for age, gender and duration of follow up. Patients underwent standard echocardiographic measurements and had Reservoir (εR) and contractile strain (εCt) measured using R-R gating.
Results: There were no significant differences in baseline characteristics between AF and non AF cases which respect to: Age, gender, blood pressure, diabetes, heart failure, previous myocardial infarction and smoking history. There were significant differences between AF and non AF cases for E/e’ (13.9±6.5, 10.8±3.5; p= 0.01), εR (%) (22±8.5, 31.3±7.3; p <0.001) and εCt (%) (10.3±5.3, 15.1±3.0; p<0.001). All other echo characteristics were similar including: LV mass, left atrial volume indexed (LAVi), LV EDV and LVEF. Conditional logistic regression using nested models revealed incremental value of E/e’ and εR over the Charge AF score (Chart). LAVi did not significantly improve the model. εCt was not included in the model due to collinearity with εR.
Conclusion: Following a cryptogenic CVA, traditional (E/e’) and novel (εR) echocardiographic parameters add incremental and independent predictive value to the established clinical model for predicting AF.
Author Disclosures: F. Pathan: None. K. Negishi: None. E. Sivaraj: None. S. Neilson: None. R. Rafiudeen: None. J. Galligan: None. T.H. Marwick: None.
- © 2016 by American Heart Association, Inc.