Creation of a transrenal arteriovenous dialysis shunt: feasibility study in a swine model

To investigate the feasibility of percutaneous renal artery and vein access for the creation of a transrenal arteriovenous hemodialysis graft.

Renal-artery-to-ipsilateral-renal-vein conduits were constructed with use of entirely percutaneous techniques in seven swine. Renal artery and vein access was performed in six animals with use of a retrograde (inside-out) technique and in one animal with use of an antegrade (outside-in) technique. Modified 8-F sheaths were used in the first three animals and Wallgrafts were used in the final four animals to form the arterial and venous limbs of each shunt. The arterial and venous limbs were joined together by a subcutaneous segment of 6-mm reinforced polytetrafluoroethylene (PTFE) in five animals and by external conduits in two animals. Wallgrafts were deployed from the renal artery and vein into the segments of PTFE. The free ends of each conduit were tunneled and joined together to close the arteriovenous circuit. Post-shunt angiography was used in all animals to document successful shunt creation and demonstrate rapid arteriovenous shunting as a determinant of technical feasibility. Two of the seven animals received additional anticoagulation therapy and/or antiplatelet therapy to prevent shunt thrombosis during the follow-up period. The three initial animals were killed within 2 hours of shunt creation, and two of the remaining four animals returned for angiographic follow-up, one on day 2 and one on day 9. All animals underwent a complete necropsy to assess for potential complications including hemorrhage and vascular or bowel injury.

Retrograde renal arterial and venous access was successful in all six animals in which it was attempted. Five of six arterial accesses and four of six venous accesses traversed the peritoneum with two arterial accesses and one venous access penetrating a loop of large bowel. Antegrade access was performed and successfully accomplished in the final animal. Brisk arteriovenous shunting was demonstrated on completion angiography in all animals. Graft occlusion was present in the two animals that returned for follow-up and two animals died before follow-up as a result of graft leakage and subsequent hemorrhage. Minimal perinephric and intrarenal hemorrhage was demonstrated at necropsy after shunt insertion in the remaining five animals. Renal infarction was present in all kidneys used for transrenal access.

The transrenal approach for the creation of a percutaneous arteriovenous shunt is feasible after renal artery and vein access by either the retrograde or antegrade technique. Additional technical refinements of the procedure and the devices used will be necessary before follow-up studies are conducted.