Abstract 209: Final Results of the Surprise Study:Long-term Monitoring For Atrial Fibrillation (AF) in Cryptogenic Stroke
Background The true frequency of AF in patients with cryptogenic stroke or TIA is not well defined. The aim was to estimate the frequency and burden of AF in patients with apparent cryptogenic minor stroke or TIA by long term monitoring providing complete data on arrhythmia occurrence. Methods Patients with minor stroke or DWI-positive TIA were included if stroke causation remained unknown during standardized work up including 24 hours telemetry. A Reveal XT®, an atrial fibrillation sensitive loop-recorder, was implanted subcutaneously allowing continuous monitoring for up to 3 years. Arrhythmia episodes were adjudicated by senior consultant cardiologist. Endpoints include episodes of AF, time and burden of AF. A total of 84 patients were included and had a minimum of three months of monitoring before final analysis. Five patients were explanted due to local infections or discomfort. (fig.2) Results In 13 patients (15.5 %) AF was documented by long term monitoring. The mean burden of AF was 2 hours pr day monitored, varying from less than a minute to 17 hours pr day monitored. (median 20 minutes pr day monitored) Kaplan Meier (fig.1) presents time from stroke onset to first AF event, mean time was 106,0 days (SD 47,9 days) Time from stroke onset to implantation was at a median of 56 days. (Mean 80 days, SD 74,9). Logistic regression analysis including all elements of CHADS2VAS found increasing risk of AF with an OR = 1.096 (p=0.015) with increasing age in years. CHADS 2 VAS score was 4.14 in the AF group vs. 3.24 (p=0.03). Conclusion Paroxysmal AF is frequent and brief in patients with cryptogenic stroke. Long term monitoring resulted in change of treatment in one out of 6 patients in this cohort.