In Vitro Comparison of Four Cerebral Protection Filters for Preventing Human Plaque Embolization during Carotid Interventions

Abstract

Purpose: To evaluate in an in vitro bench-top model the efficacy of 4 filtration devices designed for cerebral protection during carotid angioplasty. Methods: Embolization during carotid angioplasty was simulated with human plaque material (8 to 12 particles weighing 6.02 ± 0.10 mg) in an open flow model with the aortic arch and carotid bifurcation made from elastic silicone tubes and saline used as the fluid medium. The 500 to 1500-µm particles were injected into the internal carotid artery (ICA) in front of the test protection device, which was deployed 5 cm distal to the bifurcation. Particles getting past the protection device or flowing into the external carotid artery (ECA) were caught in 100-µm filters and weighed. Ten trials were made on each of 4 devices: Angioguard, FilterWire EX, TRAP, and NeuroShield. All were deployed according to the manufacturers' directions except the FilterWire, which was manually repositioned in all tests to eliminate a gap between the filter and the tube wall. Results: The lowest weight of missed particles in the effluent was obtained with the NeuroShield (0.05 ± 0.04 mg, 0.8% of injected particle weight) and FilterWire (0.08 ± 0.05 mg, 1.3%; p=0.254 compared to NeuroShield, p=0.006 versus TRAP). The weight of particles missed by the TRAP device (0.16 ± 0.06 mg, 2.6%; p<0.001 versus NeuroShield and Angioguard, p<0.05 compared to Filter Wire) was higher, and the largest amount of missed particles was observed with the Angioguard filter (0.27 ± 0.06 mg, 4.4%; p<0.001 compared to all). NeuroShield and FilterWire were significantly different (p<0.001) compared to Angioguard and TRAP in a pairwise analysis. No embolization into the ECA occurred. Conclusions: In vitro, none of the tested devices had the ability to completely prevent embolization into the ICA. Comparing current designs, the NeuroShield filter and the FilterWire EX captured the highest percentage of human particles in this in vitro model, probably due to their larger filter volume.