Endovascular delivery of vein segments with valves versus direct anastomosis.

PURPOSE: To demonstrate an experimental model designed to allow placement of a vein segment containing a valve into a femoral vein without the need for anastomoses or surgical incision on the target limbs. METHODS: Segments of jugular veins containing valves were harvested bilaterally from 7 female dogs weighing 30 to 40 kg. One of the vein segments was then interposed into one femoral vein, using a standard surgical end-to-end anastomosis. Another segment was loaded into a self-expanding nitinol stent, everted, and secured outside of the stent. This vein-coated stent was then deployed from the jugular approach into the opposite common femoral vein. The endpoint of the study was thrombosis of one of the veins segments or patency at 120 days. The function of the valves was tested prior to completion of the study by instilling dye under incremental pressures and recording the pressure necessary to cause reflux. Light microscopy and electron microscopy were utilized to evaluate structure of the valves at study endpoint. RESULTS: Both grafts in one dog occluded during the second week after surgery. Anastomotic and vein-stented sites remained patent in the remaining 6 animals for 120 days. Retrograde phlebography demonstrated valve competency bilaterally in all cases. The reflux pressures on the stented side averaged 70 mmHg while the transplanted vein segment averaged 55 mmHg. Light microscopy and electron microscopy showed intact vein valve structure in all segments at the completion of the study. CONCLUSIONS: This experimental work demonstrates that endovascular delivery of a native canine vein containing a competent valve is feasible and effective.