Evaluation of a new percutaneous stainless steel Greenfield filter by intravascular ultrasonography.

The standard stainless steel Greenfield filter was modified by extending the limbs at the apex to permit loading into a 12F carrier for percutaneous insertion. This model was tested for fixation and resolution of experimental embolism in dogs (six filters in four dogs) and sheep (eight suprarenal filters in eight sheep). Serial radiographs and venacavography at 2 and 4 weeks allowed measurement of migration and suspected caval penetration, confirmed by direct examination at autopsy. Caval penetration was graded according to visibility of point (0), hook (1+), or limb (2+) for each limb of the filter. Oxidized cellulose-induced thrombi were injected and trapped in two dogs and four sheep without proximal filter displacement or thrombus propagation. Thrombi showed incomplete resolution during 4 weeks' observation by interval venacavography and subsequent autopsy but no tendency for proximal propagation. Distal migration occurred in four of six filters in dogs, and caval wall penetration was seen (score 1.8/filter). Less migration was seen in sheep (2/8), but there was evidence of hook penetration in four animals (score 0.88/filter). Intravascular ultrasonography, 20 MHz, with a 6.5 F catheter assembly successfully visualized all 10 filters in which it was attempted and showed thrombi attached to the filter or caval wall, all of which were confirmed at explant. No thrombi were missed; however, intravascular ultrasonography was limited in assessment of caval wall penetration by lack of position control and reverberation artifact. Although it performs well as a filter, the elasticity of the percutaneous stainless steel greenfield filter allows distal migration and caval penetration comparable to the titanium Greenfield filter and will require further modification of hook design to improve performance.