In vivo evaluation of vena caval filters: can function be linked to design characteristics?

PURPOSE: To compare the five vena caval filters marketed in the United States and one investigational vena caval filter and to determine whether there is an association between their design and their in vivo function.

METHODS

Four of each type of filter--Simon Nitinol (SN), Bird's Nest (BN), Vena Tech (VT), Greenfield stainless steel (PSGF), Greenfield titanium (TGF), and the investigational stent cone filter (NGF)--were studied for 60 days in 12 sheep. Radiographic and pathologic outcomes to be assessed included clot capture and resolution, vena caval penetration, position of the filter, thrombogenicity, and vessel wall reaction.

RESULTS

Filters differed with respect to the number of clot-trapping levels and the interdependence of the legs. All devices were successfully placed. Intentionally embolized clot was captured. One VT and two SN filters migrated in response to clot capture. Resolution of thrombus was variable, and related to the design of the device. Fibrin webbing was widely present with the VT, BN, and SN filters but limited in the others. The VT and NGF filters demonstrated the most stable filter base diameter.

CONCLUSIONS

The performance of vena caval filters differs with respect to clot resolution and mechanical stability. Interdependent filter limbs and single-stage conical capture sites appear to result in more favorable performance in in vivo studies.