Chapter 19: Patency Outcomes of Transjugular Intrahepatic Portosystemic Shunts in the Stent-Graft Era

Twenty years elapsed between the original animal description of the transjugular intrahepatic portosystemic shunt (TIPS) in 1969 and the first human case. Those early human shunts were created with balloon expandable Palmaz stents, later moving to self-expanding stents. In 1992, the Wallstent became the first FDA-approved device for TIPS; thousands of patients were treated thereafter as the procedure became widely disseminated.1–9 With these cases came increasing reports documenting frequent TIPS stenoses and occlusions, developing in as many as 25% to 50% of patients within 6 months of shunt creation.^3,10–13 This shunt dysfunction was associated with recurrent variceal bleeding, affecting 9.8% to 24% of patients in controlled trials.^14–21 Extensive literature developed, validating sonographic means of follow-up, the need for follow-up programs and repeated shunt revisions.^22–28 The patency question clouded the wider application of TIPS, particularly in ‘healthier’ Childs A patients.

These issues spurred a decade of investigations into both characterizing the histopathologic processes causing stenoses and solving them. Approaches included anticoagulation and antiplatelet regimens, radiation, and brachytherapy.^29–34 In the late1990’s, both animal and human feasibility studies described marked patency improvements in TIPS lined with expanded polytetraflouoroethylene (ePTFE).^35–39 These led to commercial development of an FDA-approved TIPS ePTFE endograft, the VIATORR (W. L. Gore & Associates, Flagstaff, AZ). With this, most TIPS patency issues were resolved, and TIPS research turned increasingly to reassessment of prior clinical outcomes, with covered stents. Shunt creation for portal hypertension remains the mainstay treatment, with reduced mortality, morbidity, and hospitalization secondary to their creation in many patients electively rather than emergently.^40,41 This chapter reviews this history, as well as the current state of TIPS patency in the era of covered stents.

Fig. 19.1 Human explant specimen 9 months after TIPS creation using a Wallstent demonstrates a thick layer of fibroblasts that has grown through the stent wires, narrowing the shunt lumen. (Reprinted with permission of the SIR.)

LaBerge et al reported the first human histology of seven TIPS explants in 1991.42 Within a week of TIPS creation, metal stent wires were shown pressing back the surrounding liver tract, lined with clot and fibrin. By 3 weeks, pseudointimal tissue composed of myofibroblasts grew from the surrounding liver, enveloping the stents. By 3 months, dense collagen layers lined the lumina of the TIPS, with shunt stenoses ( Fig. 19.1). Subsequent post mortem and transplant donors further confirmed these findings.^42–46 This pseudo-endothelium appeared to be derived from the liver parenchyma (rather than dropout endothelialization).⁴⁷ In contrast, TIPS lined with stent-grafts only develop a thin pseudo-endothelium ( Fig. 19.2).

The outflow hepatic vein represents a second site of distinct shunt stenosis ( Fig. 19.3). After TIPS formation, its caliber has been shown to reduce to half its original diameter.³ These stenoses appeared as early as 90 to 180 days after initial shunt placement ( Fig. 19.4). This intimal hyperplasia appears similar to that seen in other veins or anastomoses, as with dialysis graft venous stenoses or other surgical vascular anastomoses.^10,48,49

A third, unique cause of TIPS failure is the biliary-to-TIPS fistula. As part of creating the shunt, biliary structures can be naturally traversed. In some cases, a sufficiently large bile duct may be punctured in close proximity to the eventual TIPS tract, allowing bile leakage into the tract. The thrombotic effect of bile can lead to recurring shunt occlusion within hours after creation.⁵⁰ The phenomenon is unique to bare stents. The inhibitory effect of bile on smooth muscle cells formation prevents the gradual pseudoendothelial proliferation that encompasses the bare stent’s wires and leads to “TIPS healing.”^51,52 This issue was resolved with ePTFE TIPS endografts that were designed to be relatively bile impermeable due to an outer fluoroethylene polymer wrap.

Fig. 19.2 Photograph of a gross specimen with a VIATORR-lined TIPS. This 14-month explant post liver transplant demonstrates only a thin white pseudo-endothelium (asterisk) lining parts of the stent, with no tract thrombus or stenosis.

Fig. 19.3 Four-month venogram of a bare stent TIPS demonstrates diffuse stenosis of both the intraparenchymal and hepatic vein outflow portions of the TIPS. (Reprinted with permission of the SIR.)

Adjunct medical therapy has been studied for its potential to reduce TIPS occlusion in bare stent shunts. In a randomized trial of 49 TIPS patients treated for acute and subacute variceal hemorrhage, there were no shunt occlusions in the anticoagulation group compared with the control group, which had 5 shunt occlusions within the first 3 months.29 The role of biliary leaks was not described. Subsequent follow-up demonstrated no significant difference in shunt dysfunction between the treated and control groups. Further, the routine use of anticoagulation in patients treated for variceal hemorrhage can create therapeutic dilemmas. Ultimately, anticoagulation has, in the stent graft era, little role in TIPS, with the exception of patients requiring anticoagulation because of underlying hypercoagulable syndromes. The majority of these patients are ones treated for Budd-Chiari syndrome or noncirrhotic patients treated for acute porto-spleno-mesenteric thrombosis.

In another trial, Siegerstetter et al randomized 84 patients to antiplatelet-derived growth factors and anti-aggregation agents vs. controls.³⁰ While they reported some evidence of reduced hepatic outflow vein stenosis in the experimental group, the therapy has not found widespread approval, partly presumably due to the development of endografts, the lack of widespread availability of some of the agents tested, and reluctance to expose patients with prior bleeding to prolonged clotting times.

The Development of TIPS Endografts

Animal Studies

The search for an alternative to the bare metal stent for lining the TIPS tract began with animal TIPS models. The porcine model proved to be an excellent substrate because it allowed the creation of shunts using human tools and techniques. Further, porcine TIPS histology mimics that seen in humans, with both intrashunt and hepatic vein stenosis, though at a very accelerated rate: within a few weeks of TIPS creation.³⁵ These reliable effects are seen without the need for inducing portal hypertension or cirrhosis in the animal, a daunting task in growing swine.^53,54

Multiple materials have been studied within this model, among them polyethylene terephthalate (PET), silicone, polycarbonate urethanes, and expanded polytetrafluoroethylene (ePTFE).^55–57 In 1995, Nishimine et al described the results of the first studies of ePTFE in swine TIPS.³⁵ Shunts created with handmade PTFE-lined stent grafts demonstrated a 69% patency rate at 4 weeks, compared to 8% for shunts utilizing bare stents. Additionally, stenoses in the patent PTFE-lined stents were less than 50%. At 3 months, 46% of the stent grafts were still patent. Shunt occlusion was attributed to draining vein stenosis in 38% of the failed shunts. This was echoed in a subsequent study by Haskal et al in 1997, using an encapsulated PTFE-stent graft designed for TIPS.³⁶ In 8 shunts, seven demonstrated wide patency by 5 months, while 85% of the bare stent control group demonstrated occlusions and stenoses by 4 weeks. Histologic reports of the explants paralleled the appearance of those described in prior human studies.^44,45 In contrast, TIPS created with silicone-covered endografts,⁵⁷ polycarbonate urethanes,⁵⁶ and polyethylene terephthalate/polyester,⁵⁸ showed no improvement or worse results, compared with bare stent controls ( Fig. 19.5).

Early Human Experience

With the promising animal results of TIPS, early human feasibility studies in both revision and de novo TIPS applications were reported ( Fig. 19.6).^38,49 In 1997, a Dotter Institute pilot study reported the use of a handmade ePTFE endograft used in revising failing TIPS.⁴⁹ Six patients were treated with a custom-made PTFE graft mounted on a Z-stent. This stent graft was dilated to 14 mm and held in place within the endothelial layer by Wallstents. Some shunts remained patent to 315 days. The results of this preliminary study indicated that the PTFE coating of the revised grafts was effective in prolonging shunt patency in patients with failing TIPS.

Ferral et al reported a series of 13 TIPS created with polyester fabric-covered nitinol stents (modified Cragg Endopro System I stent graft).⁵⁹ While the long-term follow-up and shunt patency assessments were mixed and somewhat incompletely described, this study focused on the technical success of shunt creation with nondedicated stent grafts and the immediate patency results. The thrombotic effect of polyethylene terephthalate graft material within the TIPS tracts, as with swine, was confirmed in a subsequent series using PET-covered Wallstents.⁵⁸ This series described salvage of PET-lined TIPS using PTFE-covered stent grafts. The reason for the thrombogenic effect of polyester in the de novo TIPS tract is specific to acute shunt creation and its exposure to a fresh parenchymal tract. Luo et al described similar findings with higher 1-year and 2-year patency rates of 91% and 85% for PTFE stents versus 78% and 63% for PET-covered stents.⁶⁰ Later studies and experience have shown that PET is not as thrombotic, when used to reduce mature endothelialized bare stent or ePTFE-lined shunts.

Fig. 19.4 A 60-year-old woman with nonalcoholic liver cirrhosis who underwent VIATORR TIPS creation for treatment of portal gastropathy and refractory ascites. (a) Digital subtraction portal venography after TIPS creation. The portosystemic gradient dropped from 21 mm Hg to 10 mm Hg. Her ascites resolved over several weeks. (b) Fourteen months later, she returned with recurrent ascites requiring frequent paracenteses and diuretic therapy. TIPS venography demonstrates an unstented segment of outflow hepatic vein (double arrow). The shunt is otherwise widely patent. In retrospect, the hepatic vein was incompletely covered at the time of initial shunt creation. (c) Hepatic venogram with catheters in both the TIPS and the inferior vena cava demonstrates the secondary stenosis of the outflow vein (arrow). (d) After placement of a coaxial VIATORR, the portosystemic gradient was reduced from 18 mm Hg to 12 mm Hg. At the 4-week follow-up, her ascites had spontaneously diuresed.

Fig. 19.5 Two-week explant from a porcine TIPS created with a non-porous prototype polycarbonate urethane endograft demonstrates shunt thrombosis without evidence of endothelialization.

The encouraging animal and preliminary human study data led to numerous reports of custom-made PTFE stent grafts. These further confirmed the marked shunt patency improvements.37,38,58,61–64 A commercially designed TIPS endograft was manufactured by W. L. Gore and Associates and was marketed as the “VIATORR TIPS endoprosthesis.” The randomized multicenter U.S. trial comparing it to the Wallstent was begun in 1999, and led to FDA on label approval of the device for both de novo and revision TIPS applications. This study enrolled 253 patients and shunt stenosis was evaluated by venography. There was a marked decrease in shunt stenosis at 6 months in the VIATORR group, to only 16% as compared to 42% (p < 0.001) in the Wallstent group, with an associated significantly longer time to revision in the stent-graft group.

Fig. 19.6 This is an example of an early hand-sewn TIPS endograft. Preexpanded ePTFE was sutured onto the leading end of a Wallstent. It was then compressed into a loading cartridge using a surrounding lacing suture (not shown).38

These findings fostered the exponential growth in use of TIPS stent grafts ( Table 19.1). Hausegger et al utilized sonography and venography for follow-up evaluation of such patients, reporting patency rates at 6 months and 12 months of 87% and 80%, respectively.⁶⁵ In the subgroup of patients evaluated by venography, the portosystemic gradient remained essentially unchanged at 6 months. In 2007, Bureau et al reported a multicenter study in France, Spain, and Canada in which 80 patients were randomized to VIATORR or bare stents.⁶⁶ Patients were followed by ultrasound every 3 months and angiography every 6 months, or at the time of dysfunction. At 2 years, the VIATORR group patency rate was 76% versus 36% in the bare stent group (p = 0.001), and absence of encephalopathy in 67% and 51% (p < 0.05), respectively. The probability of survival at 2 years for VIATORR and bare stent patients was not statistically significant at 58% and 45%, respectively. Ultrasound surveillance was deemed insufficiently accurate enough to predict shunt dysfunction in this study. However, a retrospective review of 126 patients post TIPS surveillance with ultrasound demonstrated that shunt dysfunction detection rates are equivalent for stent graft and bare stent shunts.⁶⁷ In 2003, Angermayr et al confirmed improved survival in a retrospective comparative review of patients treated with stent grafts versus bare stents in the creation of TIPS, reporting a survival rate of 88% versus 73% at 1 year and 76% versus 62% at 2 years, respectively (p = 0.01).⁶⁸

The new predictable durability of shunts fostered further comparisons with different shunt diameters. In 2010, Riggio et al reported randomization of cirrhotic patients with variceal bleeding and/or refractory ascites to 8-mm or 10-mm diameter VIATORR stent grafts.⁶⁹ Evaluated outcomes included recurrent symptoms and encephalopathy. The trial was stopped after enrolling 45 patients, as patients treated with the 8-mm device were found to have persistent ascites or varices after TIPS. The overall decrease in pressure gradient was significantly lower in the 10-mm group, explaining the observed differences, without decreased hepatic encephalopathy. While the findings are notable, it is arguable that the question of optimizing stent diameters has not been resolved, as ascites and hemorrhage cohorts were mixed together. The goal is to create the smallest portosystemic shunt that will treat the given indication in a patient ( Fig. 19.7). To that end, a diametrically adjustable TIPS endograft could be a great advantage to patients. At present, this is achieved by deployment of a 10-mm VIATORR without balloon expansion, or initial dilation only to 8-mm. Later, continued radial expansion of the endograft can be performed, as needed, and may lead to a fully expanded stent graft in many patients.

In a 2010 New England of Journal Medicine study, Garcia-Pagan et al reported a randomized comparison of patients with acute variceal hemorrhage treated with early TIPS (with endografts) compared with continued endoscopic therapy.⁷⁰ Sixty-three patients were included: 32 patients treated with TIPS, 31 with pharmacologic and endoscopic therapy. At a mean 16 months follow-up, rebleeding or inability to control bleeding was far lower in the TIPS group vs. endoscopic therapy (3% vs. 50%, p < 0.001) and 1-year survival was 86% and 61%, respectively (p < 0.001). Days in intensive care units were 8.6 ± 9 vs. 3.6 ± 4 in TIPS patients (p = 0.01) and overall hospitalization time (median 15 vs. 4 days, p = 0.014) were significantly higher in the non-TIPS group. This study was the first to show that early TIPS creation with endografts led to prolonged transplant-free survival, reduced bleeding, and shorter hospitalization.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Clinical Outcomes of Balloon-Occluded Retrograde Transvenous Obliteration and Balloon-Occluded Antegrade Transvenous Obliteration
2. Paracentesis and the LeVeen and Denver Shunts
3. Management of Duodenal Varices
4. Endoscopic Classification and Management of Varices

Stay updated, free articles. Join our Telegram channel

Join