Experimental embolic capture by asymmetric Greenfield filters

An uncommon complication of percutaneous insertion of the titanium Greenfield filter is angulation at discharge with clustering of the limbs to one side of the vena cava. Potential adverse effects of this asymmetry were tested in a perfusion system with cadaver venae cavae. The stainless-steel Greenfield filter (SGF) was compared with the modified-hook titanium model (TGF-MH) and a percutaneous modified-hook model of the SGF (SGF-MH) with flexible plastic emboli in four sizes and varied sequences. There were 458 runs comprised of 2203 embolic events, with an overall successful capture rate of 81% despite tilting, asymmetry, and accelerated flow rates from 1.0 to 2.5 L/min. Significant differences were observed in vena cava with diameters greater than 22 mm where capture rate was 76% as opposed to 86% in venae cavae less than 22 mm (p = 0.01). Larger emboli (8 x 25 mm and 8 x 100 mm) were trapped uniformly, whereas smaller emboli (5 x 10 mm) had 59% capture (p = 0.001). Similarly, the capture rate was less when smaller emboli were released before the larger ones (71% vs 87%; p = 0.001). Capture in small versus large venae cavae was not significantly different for the SGF (75%) or TGF-MH (84% vs 87%) but differed for the SGF-MH (98% vs 67%; p less than 0.001). Logistic regression was used to develop a mathematic model for interaction of the variables and determined two first-order interactions: caval size and limb position and caval size and type of filter.(ABSTRACT TRUNCATED AT 250 WORDS)