Twelve-year clinical experience with the Greenfield vena caval filter.

Optimal protection from pulmonary thromboembolism (PE) by a mechanical device requires long-term freedom from complications as well as from recurrent PE. From 1974 to 1986 the Greenfield filter was implanted surgically in 469 patients, most commonly when anticoagulation was contraindicated (38%) or failed to prevent recurrent PE (27%). The jugular vein was used for access in 83% and the femoral vein in 16%; in the remainder the filter was inserted via the right atrium or the left axillary vein. Technical complications included air embolism in 4 patients (0.9%) and misplacement in 19 (4%), with no sequelae. There were 7 deaths from PE within 48 hours postoperatively, primarily in patients with pulmonary hypertension. An additional 9 later deaths were attributed to PE, 5 confirmed by autopsy, but in only 1 of these was there trapped thrombus in the filter. An additional 9 patients had nonfatal PE, suspected and confirmed in 2 by angiogram, for an overall recurrent PE rate of 4% over 12 years. During this follow-up there were 133 deaths, and 146 patients were followed up for an average of 43 months (190 were lost to follow-up). Ten patients had thrombosed inferior venae cavae at the time of insertion of the Greenfield filter, and 3 were recanalized late to a caval patency rate of 96%. Three initially patent filters occluded, for a long-term filter patency rate of 98%, which was independent of anticoagulation. Clinically silent embolism into the Greenfield filter was seen in 18 patients, half of whom had dissolution on subsequent cavogram. Venous stasis sequelae developed in 44% of the patients, compared with 52% with preoperative edema, and ulceration was seen in 3%. Of 32 patients who had suprarenal filters, 11 died and 12 returned for follow-up study, which showed that all were patent. Favorable results with suprarenal filter placement have made this an appropriate choice in pregnant women and in young women anticipating pregnancy.