PubMed1994
Implantation of caval filters to prevent thrombo-embolism is a common medical option for twenty years, since the standard Greenfield model. The numerous new models with various geometries have been the subject of many studies aimed at testing their effectiveness in clot capture and the stability of their fixation to the vein, but very few ones are dedicated to the influence of their shape on blood flow, and to the flow disturbances induced by trapped clots shape and location. So, we have designed a caval blood flow simulation test bench of different filters and clot models. Opcetra and LGM filters were furnished with a reference regular input flow of physiological viscosity fluid. Rigid thrombus model, trapped in filter middle, was a one centimetre diameter sphere; and a recent soft one, which could model itself on the final part of the capture zone, was a two centimetres high cone. Perturbations, eddies, trajectories and velocities were recorded by particles visualization and ultrasound velocimetry. Downstream the Opcetra, velocity vectors were practically axial, no eddies could be seen; the central cylinder had changed the input parabolic flow into a one with two symmetrical maxima. For LGM, on the one hand, from the middle, the sheet branches perturbed the flow and induced several jets. On the other hand, the end ring had created an helicoïdal wake which extended up to three centimetres. With thrombi models, for Opcetra, a central spherical clot induced eddies inside the filter, but smoothed the two maxima of the profile, which vanished more rapidly. For the LGM, a central spherical clot increased the perturbations, whereas a final conical one decreased them. Finally, it was clear that filter shape, capture location and clot aspect were the leading causes of the hydrodynamic induced disturbances and might be important for the trapped clots evolution.