PC058. A Novel Aortic Arch Flow Model to Test the Efficacy of Cerebral Embolic Protection Devices Using Transcranial Doppler

Abstract

Clinically apparent stroke occurs in up to 6% of patients undergoing thoracic endovascular aortic repair and up to 4% of patients undergoing carotid artery stenting. However, clinically silent cerebral infarction is seen on postoperative diffusion-weighted magnetic resonance imaging in >60% of patients undergoing thoracic endovascular aortic repair and leads to an increased risk of future stroke and dementia. Embolic protection devices are associated with lower stroke rates and smaller volume of ischemic lesions on diffusion-weighted magnetic resonance imaging. We created a novel aortic arch flow model and tested the efficacy of the FilterWire EZ (Boston Scientific, Marlborough, Mass) using transcranial Doppler ultrasound. By employing a three-dimensional printed aortic arch, we created a closed-circuit flow model that can simulate continuous and pulsatile blood flow (Fig, A-C). Dissolved cornstarch particles simulate circulating red and white blood cells, which typically range in size from 6 to 15 μm, to provide a background signal for detection of solid and gaseous emboli on transcranial Doppler ultrasound (Fig, D and E). A 38-μm filter placed in series ensures that injected microparticles do not recirculate and the circuit is deaired. We then prepared and deployed a FilterWire EZ, a uniform 110-μm pore filter commonly used during carotid artery stenting, according to the instructions for use. Embozene Microspheres (Boston Scientific) of known sizes were diluted in saline to 1 mL of particles per 10 mL of solution. High-intensity transient signals (HITS) were recorded during filter deployment, injection of 1 mL of solution, and filter retrieval. This process was repeated for Embozene particles of 75-, 100-, 250-, and 400-μm diameters. Each continuous 1-second shower of emboli was counted as 15 HITS. Despite preparation of the FilterWire EZ according to instructions for use, there were on average 41 ± 7 (range, 31-49) HITS during filter deployment and 7 ± 7 (range, 1-17) HITS during filter retrieval. Average number of HITS during injection was 87 ± 130 (range, 4-315), and more HITS were seen distal to the filter on injection of microparticles that were smaller than the 110-μm pore size (142 ± 150 vs 5 ± 1; P = .31). Mean velocity (cm/s) of the fluid decreased by 18% ± 2% after filter deployment and 50% ± 34% during injection of microparticles and returned to baseline after filter retrieval. We created and validated a novel aortic arch flow model. HITS were seen during all stages of the procedure using the FilterWire EZ. This model can be used to test the efficacy of all available embolic protection devices and methods of deairing stent grafts.