Evolution of hook design for fixation of the titanium Greenfield filter

Abstract

The titanium Greenfield filter was developed to facilitate operative and percutaneous insertion by a carrier (12F) smaller than that required for the standard stainless steel filter (24F). In preliminary clinical studies however, it proved to have an unacceptable rate of distal slippage and caval wall penetration. Therefore four modifications in hook design and base diameter have been tested in 20 sheep by implantation above and below the renal veins. In comparison to control titanium Greenfield filters (n = 8), neither changes in hook angle with smaller diameter base (type I, n = 4), elimination of limb flare (type II, n = 4), or 80 degree hook angle (type IV, n = 10) prevented migration and entry into venous tributaries or penetration. Penetration scores on the basis of autopsy findings did not correlate with antemortem predictions based on cavograms and showed 1.5 ± 2.0/filter for controls, 1.0 ± 1.15/filter for type I, 1.75 ± 0.96/filter for type II and 1.67 ± 1.37/filter for type IV. It was necessary to change the configuration of the hook completely to a recurved hook design providing a base pad for the 80 degree hook (type VI, n = 6) to limit migration and penetration with a score of 0.33 ± 0.82/filter. Effects of experimental embolism were studied in six sheep (two acute, four chronic) and showed effective trapping by both type IV and type VI filters with no evidence of filter migration or damage to the caval wall. Persistent thrombus appeared to be caused by the remnants of oxidized cellulose used to induce thrombi, but there was no evidence of propagation proximal to the filter. The recurved hook design appears to correct the tendency for the titanium Greenfield filters to migrate distally and should inhibit caval wall penetration. General use of the device, however, cannot be recommended until the results of clinical trials with careful follow-up are known.