Comparison of caval filters in the management of venous thromboembolism.

Over a seven-year period, 38 Mobin-Uddin MU and 33 Kim-Ray Greenfield KG vena cava filters were inserted in 71 patients with no substantial complications or deaths directly related to the surgical procedure. Thirteen of the 16 early deaths were due to progression of associated diseases. The remaining three deaths were caused by pulmonary emboli and occurred in each instance after insertion of a MU filter. None of the 14 late deaths were related to the caval filter or sequellae of venous thrombosis. Inferior vena cava patency was objectively assessed in 36 patients by cavagram (22) or autopsy (14). Caval patency of 95% (18 of 19) in those with KG filters was significantly greater than the 47% patency (eight of 17) seen with MU devices (P less than .01). Gray scale ultrasound, when successful in visualizing the vena cava, was found to be a reliable indicator of patency. Venous stasis phenomena were noted in 50% of those with a patent cava and in 70% of those with an occlusion. Thus, it is not surprising that edema was found in only 38% of the patients as compared with 75% of the patients with the MU filter. Both of these devices are safe to insert. However, on the basis of superior patency rate, lower incidence of stasis phenomena, and the absence of recurrent pulmonary emboli in our series, we prefer insertion of a KG filter when caval interruption is required.