Abstract 3648: Detection Of Paroxysmal Atrial Fibrillation In Patient With Acute Brain Ischemia Combining Cardiac And Holter Monitoring: Prevalence And Predictors

Abstract

INTRODUCTION: The yield of serial electrocardiograms (ECG) plus 72-hour cardiac monitoring and Holter monitoring in detection of paroxysmal atrial fibrillation (PAF) is not well known. METHODS: consecutive patients with stroke or transient ischemic attack (TIA) admitted to a Stroke Unit (SU) during January 2009-June 2010 were studied. Patients underwent serial ECG and cardiac monitoring in the first 72 hours. Furthermore, 24-hours Holter monitoring were performed in patient with brain ischemia of unknown origin or if cardiac embolism was suspected. The presence of atrial fibrillation (AF) and PAF was recorded. RESULTS: 537 patients were included, 59.6% males. Mean age 69.1 (SD 13.5) years. Previous AF was present in 15.8% patients. ECG on admission showed not previously known AF in 22 (4%) patients. Cardiac monitoring and serial ECG in the SU detected PAF in 12 (2.2%) cases more. 24-hours Holter monitoring was completed in 156 patients, in 42 cases PFA was detected, of them 6 had been previously detected by serial ECG/cardiac monitoring. One case of PAF detected by serial ECG was not confirmed by Holter monitoring. In total, 10.8% of patients were diagnosed with new PAF. In multivariate analyses, NIHSS on admission (OR 1.1 for each 1 point increase; 95% CI, 1-1.2), enlarged left atrial (OR 5.8; 95% CI, 2.8-12), absence of carotid plaques by duplex ultrasound (OR 2.1; 95% CI 1-4.4) and hyperlipidemia (OR 2.4; 95% CI 1.119-5) were predictors of PAF. CONCLUSIONS: 24-hours Holter monitoring increased by 6 fold the detection of PAF as compared to ECG plus cardiac monitoring in acute stroke patients. Stroke severity, enlarged left atrial, absence of carotid plaques and hyperlipidemia are associated with PAF.