Endovascular caudal retraction of the cranial end of a misplaced Viatorr TIPS prior to liver transplantation

Abstract

Transjugular intrahepatic portosystemic shunt (TIPS) extension far into the inferior vena cava (IVC) or the right atrium may complicate or preclude orthotopic liver transplantation depending on the space available for placement of a hemostatic clamp in the suprahepatic IVC. Until 2004, most TIPS were performed with bare metal stents, which integrate into the vessel wall, making percutaneous or intraoperative repositioning uncertain. Most TIPS are currently created with stent grafts that have an outer fabric to increase shunt patency and prevent endothelial ingrowth. We describe the first known manipulation of a covered stent graft prior to transplantation. The stent graft, which extended well into the IVC, was snared from a femoral approach and deflected caudally in order to document feasibility and nonadherence to the vein wall prior to definitive surgical planning of liver transplantation. Provisions were made for endovascular retraction during actual transplant surgery 9 weeks later, but this became unnecessary when manual retraction of the exposed liver enabled suprahepatic IVC clamping. Due to the nonadherent nature of the outer graft material, compared with a bare metal stent, extension of a stent graft into the IVC or right atrium may not preclude transplantation, and intraoperative endovascular retraction may be considered.

Due to the high morbidity and mortality associated with esophageal variceal hemorrhage and refractory ascites, a transjugular intrahepatic portosystemic shunt (TIPS) is commonly used as a bridge to liver transplantation (1–4). Proper TIPS placement is essential for successful orthotopic liver transplantation (OLT). Extension too far into the inferior vena cava (IVC) or right atrium may leave inadequate room for placement of a suprahepatic vena cava clamp during conventional liver transplantation, where the recipient retrohepatic IVC is resected and a donor IVC interposed. Shunt extension too far into the IVC may be even more critical when preserving the recipient IVC in a “piggy-back” technique. Until 2004, most TIPS were performed using bare metal stents with interstices through which endothelial growth would occur. While pretransplant imaging provides information regarding TIPS placement and allows operative planning, the ability to reposition or remove a metallic TIPS shunt percutaneously or during transplantation is uncertain and dependent on the degree of endothelial ingrowth (5–8).

TIPS shunts are now more commonly performed with covered stents to prevent the endothelial ingrowth and bile leaks that compromised the patency of bare metal stents. The Viatorr stent (W. L. Gore and Associates, Flagstaff, AZ) is covered with expanded tetrafluoroethylene (ePTFE) graft material that prevents endothelial ingrowth and increases patency. We present the first known description of an endovascular Viatorr shunt manipulation to document proof of the stent's nonadherence to the IVC before the patient was listed for a liver transplant and to plan the surgery accordingly.

CASE REPORT

A 56-year-old man with a history of primary sclerosing cholangitis, cirrhosis, and ulcerative colitis presented to an outside hospital in December 2009 with massive esophageal variceal hemorrhage, ascites, and a Model for End-Stage Liver Disease score of 12. A TIPS procedure was performed. The patient was subsequently sent to our facility for liver transplant evaluation. During the workup, it was noted that the patient had recurrent ascites with velocity changes on TIPS ultrasound consistent with a stenosis at the hepatic vein end of the shunt. TIPS revision was performed with balloon dilatation and extension of the hepatic vein end of the shunt with a 10 × 40 mm Viatorr. The revised shunt inadvertently extended too far cephalad. Although the appearance on digital subtraction angiography was of possible extension to the right atrium (Figure 1), nonenhanced computed tomography (CT) proved extremely valuable in proving its proximal extent to be in the IVC close to the junction with the right atrium (Figure 2). Due to concerns about the ability to place a suprahepatic clamp during transplantation, an endovascular test run of manipulation of the stent graft was performed 15 days after it had been placed. In the interventional suite, a 25 mm Nitinol goose neck loop snare (Microvena, White Bear Lake, MN) was used to engage the cranial tip of the TIPS shunt, demonstrating free mobility and caudal retraction and therefore providing ample room for cross-clamping (Figure 3). No attempt was made to leave the stent graft in a more caudal or angulated position due to the potential resultant risk of thrombosis.

Figure 1.

Digital subtraction image demonstrating the cranial transjugular intrahepatic portosystemic shunt end near the junction of the right atrium and suprahepatic inferior vena cava.

Figure 2.

(a) Axial and (b) coronal CT images at the level of the cranial TIPS demonstrate the precise location to be in the distal suprahepatic vena cava without breech of the right atrial ostium. This was very helpful since the angiographic appearance suggested a right atrial extension.

Figure 3.

(a) A Nitinol goose neck snare grasps the cranial end of the expanded tetrafluoroethylene (Viatorr) transjugular intrahepatic portosystemic shunt. (b) After caudal retraction, the cranial transjugular intrahepatic portosystemic shunt is mobile.

During OLT 9 weeks later, interventional radiology staff and fluoroscopy staff were present to provide loop snare caudal deflection of the TIPS if needed. A Mercedes incision was performed and the suprahepatic vena cava was palpated, demonstrating the presence of the Viatorr shunt and inadequate vena cava clamp room. Gentle retraction on the native liver resulted in retraction of the Viatorr into the hepatic vein/IVC junction, providing ample room for cross-clamping as demonstrated fluoroscopically by the position of the clamp cephalad to the cranial extent of the stent graft. Although an intraoperative endovascular procedure was not needed, it is conceivable that a more cephalad position of the stent graft in other patients might necessitate manipulation at that time. The TIPS was easily dissected from the venous wall and hepatectomy was completed, followed by an uneventful transplantation with a replacement of a total of 8 units of packed red blood cells.

The patient's postoperative course was complicated by portal vein thrombus requiring surgical thrombectomy on day 1 and bile duct revision for stricture on day 7. The patient was discharged home in good health on postoperative day 14 and remained well for the duration of our center's 2-month post-OLT follow-up.

DISCUSSION

TIPS extension into the IVC or right atrium can complicate transplantation by hindering vena cava clamp placement during hepatectomy, potentially resulting in hemorrhage and increased transfusion requirements (9). Radical transplant procedures, including partial transplantation of the right atrium and open cardiotomy, have been described due to misplaced and migrated shunts (10, 11).

Attempts to retract or remove the bare metal Wallstent (Boston Scientific, Natick, MA) may be confounded by the inability to separate the metallic shunt from the vein due to neointimal hyperplasia between the interstices of the bare metal stent and scar tissue formation from the braided metal moving and irritating the vein wall. Unfortunately, the degree of stent adherence is only known intraoperatively.

Previous reports have discussed manipulation of TIPS metallic shunts. Slonim et al reported successful percutaneous removal of six TIPS Wallstents extending into the superior mesenteric vein, IVC, or right atrium. Five of the six Wallstents were retrieved immediately after deployment. In one case, a Wallstent extending into the right atrium was successfully removed after 30 days. Shunt retrieval was predominantly performed with snare devices from the groin or neck since the flexibility and compressibility of the self-expanding Wallstent allows the stent to be snared, folded on itself, and pulled into a sheath for removal (6, 7). Pulling the Wallstent into a sheath attempts to minimize endovascular trauma resulting from dragging a bare metal stent across the endothelium with possible unraveling or fraying at the stent ends during retrieval (7). In another case where two Wallstents were deployed with partial overlap, it was not technically possible to remove the Wallstent from the hepatic vein and IVC; only the more cephalad Wallstent could be snared from the heart and repositioned into the femoral vein, where it was left (5).

Although less invasive than surgery, percutaneous removal has associated procedural risks, including further migration to a more hazardous location, vessel dissection, perforation, valvular damage, induction of conduction abnormalities, possible cardiac perforation, and unraveling of the stent components with worsening luminal compromise (7, 10, 12). Difficulties can be encountered as described by Te et al, whereupon open cardiotomy for TIPS removal was performed secondary to ventricular tachycardia during attempted percutaneous removal. If a misplaced or migrated TIPS shunt is intracardiac in location, valve function should be documented prior to attempted percutaneous removal (11).

Numerous reports have described manipulation of migrated or malpositioned TIPS, all of which have been self-expanding metallic Wallstents or balloon expandable Palmaz stents (Johnson & Johnson Intervention System, Warren, NJ) (5–7). The improved patency of the ePTFE-covered stents has decreased the frequency of revision, reducing the risk of migration or malpositioning which increases when additional stents are used during revision. The ePTFE-lined proximal end of the Viatorr will ideally extend within 5 mm of the hepatic vein IVC junction (13).

The current case highlights the loop snare retraction of a Viatorr stent graft preoperatively to demonstrate feasibility during future transplantation. Secondary benefit may be obtained through the disruption of any early adhesions to the vein wall that could theoretically occur while waiting for transplantation. These findings contrast with reports of the Wallstent, which undergoes luminal neointimal hyperplasia within approximately 3 weeks (14).