Improved Patency of Transjugular Intrahepatic Portosystemic Shunts in Humans

Abstract & Commentary

Synopsis: Haskal reports the results of transjugular intrahepatic portosystemic shunts (TIPS) in humans using PTFE-covered Wallstents placed in 13 patients. Seven covered stents were placed at shunt creation and seven during revision of TIPS. Persistent biliary-TIPS fistulas were demonstrated in six cases, prior to stent-graft placement despite repeated shunt revisions with additional metallic stents. Results showed that all but one stent-graft TIPS were patent by portal venography at a mean follow-up duration of 19 months. Only one patient developed TIPS thrombosis due to a persistent biliary-TIPS fistula.

Source: Haskal ZJ. Improved patency of transjugular intrahepatic portosystemic shunts in humans: Creation and revision with PTFE stent-grafts. Radiology 1999;213(3):759-766.

One of the well-documented problems with tips is their limited and unpredictable patency. Stenoses greater than 50% and recurrent portal hypertension develop in 25-50% of cases within 6-12 months of shunt creation. Shunt surveillance and revisions are necessary to maintain long-term patency. The purpose of this study was to expand on stent-graft TIPS data by reporting results with PTFE stent-grafts used for both creation and revision of TIPS. Thirteen patients (8 men, 5 women; mean age, 54 years) with portal hypertension were included in the study. Seven shunts were created de novo with PTFE grafts, and seven pre-existing shunts were revised by using the stent-grafts. One patient had two revisions with stent-grafts due to early failure of one revision. The indications for TIPS were variceal bleeding in 10 patients, refractory ascites and a refractory large right hepatic hydrothorax in one, refractory ascites in one, and Budd-Chiari syndrome in one. Seven patients had undergone shunt revision because of TIPS thrombosis, and biliary-TIPS fistulae were documented in six cases by means of gentle injection of iodinated contrast material into the occluded shunts. The fistulae persisted despite shunt thrombectomies and reopening of the lumina with additional Wallstents. The stent-grafts were put together on a sterile tabletop at the time of TIPS revision or initial shunt creation. The length of the segment to be lined with PTFE was from the portal venous entry site to the caval ostium. A 3- or 4-mm-diameter standard thin-wall PTFE graft was dilated using 10- or 12-mm dry high-pressure balloons. The pre-expanded graft material was cut to length and sutured solely at the leading end of the Wallstents. The device was introduced into the patient through a 30-cm-long 16-F sheath. Results showed that stent-grafts were placed accurately without incidents and no associated complications. The mean duration of venographic follow-up was 19 months and none of the patients developed recurrence of the symptoms that led to the initial TIPS placement. All but one graft-lined TIPS in both new and revision groups were widely patent at follow-up. Asymptomatic stent-graft thrombosis occurred at three weeks of graft implantation in one patient. A parallel transcaval shunt was formed between the retrohepatic cava and the left portal vein and was immediately formed with a new stent-graft.

Comment by Moni Stein, MD

Randomized clinical trials comparing TIPS and endoscopic sclerotherapy have shown that average rate of variceal rebleeding after TIPS was approximately 30%—less than sclerotherapy. In almost all cases, rebleeding after TIPS formation is related to shunt malfunction. Solving the problem of TIPS durability could broaden its clinical application and benefit and restore its credibility as a long-lasting solution. Animal work has shown that histologic and venographic study of animals with TIPS lined with stent-grafts revealed near-absence of any tissue within the shunt. The overall patency of the graft group was good, in marked contrast to that in control animals (with conventional stents), which developed occlusions or marked stenoses (40-72%) within four weeks of TIPS.1 TIPS shunts with severe stenoses or occlusions develop severe neointimal hyperplasia, which appears as a thick rind of myofibroblasts and extracellular collagen matrix. This process is worse in the intrahepatic shunt and venous outflow and is thought to limit long-term patency. In acute TIPS occlusions, TIPS-biliary fistulas are a common factor. In both acute and long-term occlusions, PTFE-lined stent-grafts are considered the most promising development that will likely prolong TIPS durability. A new multicenter FDA trial has been launched with a new generation device that will likely revolutionize this procedure. Andrews et al evaluated the potential benefits of placing a PTFE-covered stent-graft during initial creation of a TIPS in clinical practice.2 De novo TIPS were created with a PTFE stent-graft in four men and four women with symptomatic portal hypertension awaiting liver transplant. Patients were followed with TIPS ultrasound (US) and/or venography until liver transplantation or death. Six recovered specimens underwent gross and microscopic evaluation. All TIPS placements were successful and six shunts were primarily patent, with a mean patency of 289 days. Five were found to be patent at transplant and one was found to be patent at autopsy. Three patients developed a total of four stenoses (one tandem lesion) during follow-up, leading to revision in two patients. Only one (nonsignificant) stenosis clearly developed in an area covered by PTFE. Haskal concludes that placement of a de novo PTFE stent-graft during TIPS formation is feasible and may extend primary shunt patency. It is my prediction that in 2-3 years most TIPS will be performed with stent-grafts, resulting in improved patency rates and clinical credibility.

References

1. Haskal ZJ, et al. PTFE-encapsulated endovascular stent-graft for transjugular intrahepatic portosystemic shunts: Experimental evaluation. Radiology 1997; 205:682-688.

2. Andrews RT, et al. Stent-grafts for de novo TIPS: Technique and early results. J Vasc Interv Radiol 1999;10(10):1371-1378.