- Email: [email protected]
Management of Shunt Dysfunction in the Era of TIPS Endografts
Management of Shunt Dysfunction in the Era of TIPS Endografts Timothy W.I. Clark, MD Significantly improved long-term patency can be achieved with transjugular intrahepatic portosystemic shunt (TIPS) endografts compared to conventional bare stents. In the USA, approximately 80% of TIPS procedures are performed using these devices. The phenomenon of early shunt thrombosis with TIPS created with bare stent TIPS, attributed to biliary fistulae, is seldom observed in patients with TIPS endografts. Intrashunt stenoses within the polytetrafluoroethylene-lined conduit are also rare. However, as with shunts created with bare stents, distinct patterns of dysfunction can occur with TIPS endografts. Some of these are inherent to the learning curve of placing these devices and others are secondary to device design. The interventional radiologist needs to be aware of these patterns of shunt dysfunction and have a systematic approach to their management. Tech Vasc Interventional Rad 11:212-216 © 2008 Elsevier Inc. All rights reserved.
T
he only Food and Drug Administration approved endograft currently available with a specific TIPS indication is the Viatorr (WM. Gore, Flagstaff, AZ), although TIPS endografts may be also created through off-label use of the Fluency (C.R. Bard, Tempe, AZ), Wallgraft (Boston Scientific, Natick, MA), and other endograft devices. The Viatorr device is available in various lengths, and device length is determined by performing a portogram using a calibrated pigtail catheter. Recognition and management of shunt dysfunction relates to an awareness of imaging characteristics and patterns of device deployment (anatomic mismatch of device length, bucking in the portal vein, need for shunt reduction), which may occur in a small but important subset of patients.
Ultrasound Assessment Assessment of TIPS endografts in the immediate post placement period is hampered by trapped air microbubbles between the layers of encapsulated polytetrafluoroethylene (PTFE). The Viatorr device (WM Gore) is constructed of 3 layers of PTFE and, in the 48-72 hours that it takes for these microbubbles to be displaced from the sandwiched layers of fabric, the device is relatively impermeable to insonation with ultrasound. Therefore, early assessment of TIPS endografts with ultrasound can be difficult during the perioperative pe-
Section of Vascular and Interventional Radiology, Department of Radiology, New York University School of Medicine, New York, NY. Address reprint requests to Timothy W.I. Clark, MD, Section of Vascular and Interventional Radiology, Department of Radiology, New York University School of Medicine, 560 First Avenue, HE-221, New York, NY 10016. E-mail: [email protected]
212
1089-2516/08/$-see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1053/j.tvir.2009.04.003
riod. The TIPS can appear occluded simply because of a reflection of the ultrasound beam. For instances where early shunt thrombosis is suspected, such as a patient who develops variceal rebleeding after an otherwise uncomplicated TIPS, shunt assessment can be performed with multidetector computed tomography (Fig. 1A) or magnetic resonance venography (Fig. 1B). The nitinol alloy within the TIPS endograft produces minimal magnetic susceptibility artifact. Most interventionalists obtain a baseline ultrasound within 1 to 4 weeks of shunt creation. Thresholds velocities that are too high (⬎200-250 cm/s) or too low (⬍50 cm/s) are associated with high specificity and sensitivity (⬎90%) for shunt dysfunction.1,2 A progressive increase or decrease in shunt velocities approaching these thresholds should also prompt shunt venography and portosystemic pressure measurements.
Unstented Shunt Outflow At the time of initial TIPS creation, the outflow should be extended to the junction of the hepatic vein with intervening cava.3 The concern of covering hepatic vein outflow by the endograft producing an isolated Budd–Chiari syndrome appears to be more of theoretical concern in that this is a subclinical event. Shortening does not occur at the time of device deployment; however, minor changes in endograft position can occur over time. Insufficient length of the endograft, related either to the operator’s learning curve or to the portal anatomy, will eventually result in TIPS dysfunction and will require revision (Fig. 2). This may be performed using balloon angioplasty with placement of an additional stent (Fig. 3). Sufficient overlap needs to be provided between the interleaved stent with
TIPS endograft dysfunction
213 without the support of PTFE fabric. This portion of the device can become intercalated and buckle at the point of contact with the portal vein if an abrupt transition with the portal vein is present. Resulting turbulence can cause an increase in the portosystemic gradient. Altered shear stress mechanics at this site appear to induce a hyperplastic response with resultant narrowing of the TIPS inflow (Fig. 4). This situation can be managed by smoothing of the transition with the portal vein while preserving intrahepatic portal flow. This is accomplished using a bare stent to further extend the TIPS into the portal vein with a length just sufficient enough to decrease the angle with the TIPS inflow. Care must be taken not to extend the bare stent too far into the portal vein to produce
Figure 1 (A) Coronal multiplanar reconstruction (MPR) image from a CT scan performed for suspected TIPS endograft occlusion. The TIPS is widely patent. (B) Coronal multiplanar reconstruction (MPR) image from a Gd-enhanced T1-weighted MR performed for suspected TIPS endograft occlusion. The TIPS is widely patent.
the endograft to avoid stent migration. In general, extending a shunt with a stent less than 40 mm in length can be associated with stent migration. Extension of the shunt into the inferior vena cava can produce caval thrombosis or interfere with placement of the suprahepatic caval clamp during subsequent orthotopic liver transplant.4,5
Shunt Inflow The goal of endograft placement is to preserve intrahepatic portal flow by confining the covered portion of the device to the hepatic parenchymal tract. The uncovered portion of the Viatorr device is constructed of a single wire of nitinol wire
Figure 2 (A) Initial portogram showing the origin of the parenchymal tract (arrow) at the portal venous end. The covered portion of the endograft should be positioned at this location. (B) Portogram post TIPS creation showing malposition of TIPS, with the covered portion of the Viatorr device extending into the portal vein (arrowheads) beyond the origin of the parenchymal tract (black arrow). The distal outflow of the TIPS is unsupported by the endograft (white arrow).
T.W.I. Clark
214
Figure 3 (A) TIPS venogram 2 months post TIPS creation showing termination of the superior portion of the endograft approximately 2 cm from the junction of the right hepatic vein and inferior vena cava. This intervening segment has developed a significant outflow stenosis (arrows). (B) TIPS venogram following revision with balloon angioplasty and placement of a 10 ⫻ 40 mm self-expanding stent, the superior end of which now extends to the junction of the right hepatic vein with the inferior vena cava (arrow).
interference with portal vein anastomosis during hepatic transplantation. The bare stent portion of the Viatorr device can also interfere with advancement of devices through the TIPS during
revision or shunt venography. Use of a looped guidewire or a looped catheter with a diameter exceeding the interstices of the stent can prevent devices from becoming entangled within the bare stent during manipulation.
Figure 4 (A) TIPS venogram in a patient with an atrophic right hepatic lobe in whom a left hepatic vein to left portal vein TIPS was created. The shunt has already been extending into the portal vein to produce a smoother transition with the endograft inflow; however, the patient returns with recurrent ascites and a hyperplastic narrowing along the inflow (arrows). (B) Angioplasty of this segment required use of an ultrahigh pressure balloon. (C) Completion venogram showing restoration of widely patent inflow within a target portosystemic gradient of 5 mmHg.
TIPS endograft dysfunction When the right hepatic vascular anatomy is unsuitable for TIPS, the portal vein transition of a left portal vein–left hepatic TIPS can pose a particular challenge to optimizing shunt inflow. The 2 cm length of the uncovered portion of the Viatorr device is not sufficiently long to provide a smooth transition with the portal vein. Extending this interface with the TIPS to further support the bare stent portion is needed. We routinely use a bare self-expanding stent for this purpose to enable preservation of intrahepatic portal flow through the interstices of the stent.
Shunt Reduction The smooth lining of TIPS endografts has implications for shunt reduction. When patients present with refractory encephalopathy or worsening liver failure, the endograft will need shunt reduction. Multiple techniques have been described for reduction of bare stent TIPS and are beyond the scope of this article. For a complete review of techniques for shunt reduction, the reader is referred to the article by Madoff et al.6 However, it is generally acknowledged that these techniques may be less suited to the smooth lumen of endografts. Two emerging techniques that have been described in for use with the Viatorr device are the parallel stent technique and the hourglass balloon expandable covered stent technique. Sze and colleagues described the use of a coaxial Viatorr device placed within the existing Viatorr, delivered via 1 of 2
215 parallel guidewires placed within the existing TIPS.7 Becausee the Viatorr device currently requires a 10-French sheath, an oversized 12- to 14-French sheath is used to provide adequate space for the additional 0.035-in. guidewire. Following deployment of the second Viatorr device within the original TIPS, the parallel guidewire becomes compressed between the first and second Viatorr devices. This guidewire is used to deliver a balloon expandable stent. The stent is deployed between the newly deployed stent and the original stent, thereby extrinsically narrowing the newly deployed Viatorr stent into an hourglass configuration. This technique has the advantage of being able to serially increase the narrowing of the TIPS if the portosystemic gradient remains too low. This is achieved through an incrementally larger balloon to further dilate the balloon expandable stent. The second emerging technique for shunt reduction involves the deployment of a PTFE covered balloon expandable stent, such as the Atrium iCast device within the TIPS (Atrium Medical, Hudson, NH). Only the leading edge of the stent is dilated, thereby mounting the stent eccentrically along a balloon or by constraining the delivery system within a sheath during deployment. To optimize contact of the device with the shunt, we place the inferior aspect aligned within the uncovered portion of the Viatorr shunt, 3-4 mm from the edge of the covered portion of the shunt. This provides metal-PTFE contact between the 2 stent systems as an
Figure 5 (A) Initial portogram in a TIPS patient with refractory encephalopathy. The portosystemic gradient was 3 mm Hg. (B) Following deployment of the iCast covered stent, the proximal and distal ends of the stent are sufficiently dilated to enable a stable stent position. Note the narrowed midportion of the stent (arrows). (C) Completion portogram following shunt reduction (arrows) with an increase in the portosystemic gradient from 3 to 9 mm Hg. The patient’s hepatic encephalopathy resolved.
T.W.I. Clark
216 extra measure of stability; there have been anecdotal reports of the iCast device migrating from the TIPS when the device is placed in the straight outflow portion of the shunt. To continue placement of the iCast, the balloon is then deflated, repositioned along the superior outflow portion of the stent, and reinflated to produce an hourglass configuration to the stent. This technique also has the advantage of being able to serially enlarge the aperture in the shunt if the portosystemic gradient is too high (Fig. 5).
Conclusions Interventional radiologists need to be familiar with the patterns of shunt dysfunction among TIPS endografts and have an approach to managing shunt dysfunction. Prompt identification and treatment of shunt dysfunction can enable sustained and durable portal decompression, both as a bridge to liver transplantation or for patients who are not transplant candidates.
References 1. Zizka J, Elias P, Krajina A, et al: Value of Doppler sonography in revealing transjugular intrahepatic portosystemic shunt malfunction: A 5-year experience in 216 patients. AJR Am J Roentgenol 175:141148, 2000 2. Middleton WD, Teefey SA, Darcy MD: Doppler evaluation of transjugular intrahepatic portosystemic shunts. Ultrasound Q 19:56-70, 2003 3. Clark TW, Agarwal R, Haskal ZJ, et al: The effect of initial shunt outflow position on patency of transjugular intrahepatic portosystemic shunts. J Vasc Interv Radiol 15:147-152, 2004 4. Hoxworth JM, LaBerge JM, Gordon RL, et al: Inferior vena cava thrombosis after transjugular intrahepatic portosystemic shunt revision with a covered stent. J Vasc Interv Radiol 15:995-998, 2004 5. Wilson MW, Gordon RL, LaBerge JM, et al: Liver transplantation complicated by malpositioned transjugular intrahepatic portosystemic shunts. J Vasc Interv Radiol 6:695-699, 1995 6. Madoff DC, Wallace MJ, Ahrar K, et al: TIPS-related hepatic encephalopathy: Management options with novel endovascular techniques. RadioGraphics 24:21-36, 2004, discussion, 36-37 7. Sze DY, Hwang GL, Kao JS, et al: Bidirectionally adjustable TIPS reduction by parallel stent and stent-graft deployment. J Vasc Interv Radiol 19:1653-1658, 2008
T
he only Food and Drug Administration approved endograft currently available with a specific TIPS indication is the Viatorr (WM. Gore, Flagstaff, AZ), although TIPS endografts may be also created through off-label use of the Fluency (C.R. Bard, Tempe, AZ), Wallgraft (Boston Scientific, Natick, MA), and other endograft devices. The Viatorr device is available in various lengths, and device length is determined by performing a portogram using a calibrated pigtail catheter. Recognition and management of shunt dysfunction relates to an awareness of imaging characteristics and patterns of device deployment (anatomic mismatch of device length, bucking in the portal vein, need for shunt reduction), which may occur in a small but important subset of patients.
Ultrasound Assessment Assessment of TIPS endografts in the immediate post placement period is hampered by trapped air microbubbles between the layers of encapsulated polytetrafluoroethylene (PTFE). The Viatorr device (WM Gore) is constructed of 3 layers of PTFE and, in the 48-72 hours that it takes for these microbubbles to be displaced from the sandwiched layers of fabric, the device is relatively impermeable to insonation with ultrasound. Therefore, early assessment of TIPS endografts with ultrasound can be difficult during the perioperative pe-
Section of Vascular and Interventional Radiology, Department of Radiology, New York University School of Medicine, New York, NY. Address reprint requests to Timothy W.I. Clark, MD, Section of Vascular and Interventional Radiology, Department of Radiology, New York University School of Medicine, 560 First Avenue, HE-221, New York, NY 10016. E-mail: [email protected]
212
1089-2516/08/$-see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1053/j.tvir.2009.04.003
riod. The TIPS can appear occluded simply because of a reflection of the ultrasound beam. For instances where early shunt thrombosis is suspected, such as a patient who develops variceal rebleeding after an otherwise uncomplicated TIPS, shunt assessment can be performed with multidetector computed tomography (Fig. 1A) or magnetic resonance venography (Fig. 1B). The nitinol alloy within the TIPS endograft produces minimal magnetic susceptibility artifact. Most interventionalists obtain a baseline ultrasound within 1 to 4 weeks of shunt creation. Thresholds velocities that are too high (⬎200-250 cm/s) or too low (⬍50 cm/s) are associated with high specificity and sensitivity (⬎90%) for shunt dysfunction.1,2 A progressive increase or decrease in shunt velocities approaching these thresholds should also prompt shunt venography and portosystemic pressure measurements.
Unstented Shunt Outflow At the time of initial TIPS creation, the outflow should be extended to the junction of the hepatic vein with intervening cava.3 The concern of covering hepatic vein outflow by the endograft producing an isolated Budd–Chiari syndrome appears to be more of theoretical concern in that this is a subclinical event. Shortening does not occur at the time of device deployment; however, minor changes in endograft position can occur over time. Insufficient length of the endograft, related either to the operator’s learning curve or to the portal anatomy, will eventually result in TIPS dysfunction and will require revision (Fig. 2). This may be performed using balloon angioplasty with placement of an additional stent (Fig. 3). Sufficient overlap needs to be provided between the interleaved stent with
TIPS endograft dysfunction
213 without the support of PTFE fabric. This portion of the device can become intercalated and buckle at the point of contact with the portal vein if an abrupt transition with the portal vein is present. Resulting turbulence can cause an increase in the portosystemic gradient. Altered shear stress mechanics at this site appear to induce a hyperplastic response with resultant narrowing of the TIPS inflow (Fig. 4). This situation can be managed by smoothing of the transition with the portal vein while preserving intrahepatic portal flow. This is accomplished using a bare stent to further extend the TIPS into the portal vein with a length just sufficient enough to decrease the angle with the TIPS inflow. Care must be taken not to extend the bare stent too far into the portal vein to produce
Figure 1 (A) Coronal multiplanar reconstruction (MPR) image from a CT scan performed for suspected TIPS endograft occlusion. The TIPS is widely patent. (B) Coronal multiplanar reconstruction (MPR) image from a Gd-enhanced T1-weighted MR performed for suspected TIPS endograft occlusion. The TIPS is widely patent.
the endograft to avoid stent migration. In general, extending a shunt with a stent less than 40 mm in length can be associated with stent migration. Extension of the shunt into the inferior vena cava can produce caval thrombosis or interfere with placement of the suprahepatic caval clamp during subsequent orthotopic liver transplant.4,5
Shunt Inflow The goal of endograft placement is to preserve intrahepatic portal flow by confining the covered portion of the device to the hepatic parenchymal tract. The uncovered portion of the Viatorr device is constructed of a single wire of nitinol wire
Figure 2 (A) Initial portogram showing the origin of the parenchymal tract (arrow) at the portal venous end. The covered portion of the endograft should be positioned at this location. (B) Portogram post TIPS creation showing malposition of TIPS, with the covered portion of the Viatorr device extending into the portal vein (arrowheads) beyond the origin of the parenchymal tract (black arrow). The distal outflow of the TIPS is unsupported by the endograft (white arrow).
T.W.I. Clark
214
Figure 3 (A) TIPS venogram 2 months post TIPS creation showing termination of the superior portion of the endograft approximately 2 cm from the junction of the right hepatic vein and inferior vena cava. This intervening segment has developed a significant outflow stenosis (arrows). (B) TIPS venogram following revision with balloon angioplasty and placement of a 10 ⫻ 40 mm self-expanding stent, the superior end of which now extends to the junction of the right hepatic vein with the inferior vena cava (arrow).
interference with portal vein anastomosis during hepatic transplantation. The bare stent portion of the Viatorr device can also interfere with advancement of devices through the TIPS during
revision or shunt venography. Use of a looped guidewire or a looped catheter with a diameter exceeding the interstices of the stent can prevent devices from becoming entangled within the bare stent during manipulation.
Figure 4 (A) TIPS venogram in a patient with an atrophic right hepatic lobe in whom a left hepatic vein to left portal vein TIPS was created. The shunt has already been extending into the portal vein to produce a smoother transition with the endograft inflow; however, the patient returns with recurrent ascites and a hyperplastic narrowing along the inflow (arrows). (B) Angioplasty of this segment required use of an ultrahigh pressure balloon. (C) Completion venogram showing restoration of widely patent inflow within a target portosystemic gradient of 5 mmHg.
TIPS endograft dysfunction When the right hepatic vascular anatomy is unsuitable for TIPS, the portal vein transition of a left portal vein–left hepatic TIPS can pose a particular challenge to optimizing shunt inflow. The 2 cm length of the uncovered portion of the Viatorr device is not sufficiently long to provide a smooth transition with the portal vein. Extending this interface with the TIPS to further support the bare stent portion is needed. We routinely use a bare self-expanding stent for this purpose to enable preservation of intrahepatic portal flow through the interstices of the stent.
Shunt Reduction The smooth lining of TIPS endografts has implications for shunt reduction. When patients present with refractory encephalopathy or worsening liver failure, the endograft will need shunt reduction. Multiple techniques have been described for reduction of bare stent TIPS and are beyond the scope of this article. For a complete review of techniques for shunt reduction, the reader is referred to the article by Madoff et al.6 However, it is generally acknowledged that these techniques may be less suited to the smooth lumen of endografts. Two emerging techniques that have been described in for use with the Viatorr device are the parallel stent technique and the hourglass balloon expandable covered stent technique. Sze and colleagues described the use of a coaxial Viatorr device placed within the existing Viatorr, delivered via 1 of 2
215 parallel guidewires placed within the existing TIPS.7 Becausee the Viatorr device currently requires a 10-French sheath, an oversized 12- to 14-French sheath is used to provide adequate space for the additional 0.035-in. guidewire. Following deployment of the second Viatorr device within the original TIPS, the parallel guidewire becomes compressed between the first and second Viatorr devices. This guidewire is used to deliver a balloon expandable stent. The stent is deployed between the newly deployed stent and the original stent, thereby extrinsically narrowing the newly deployed Viatorr stent into an hourglass configuration. This technique has the advantage of being able to serially increase the narrowing of the TIPS if the portosystemic gradient remains too low. This is achieved through an incrementally larger balloon to further dilate the balloon expandable stent. The second emerging technique for shunt reduction involves the deployment of a PTFE covered balloon expandable stent, such as the Atrium iCast device within the TIPS (Atrium Medical, Hudson, NH). Only the leading edge of the stent is dilated, thereby mounting the stent eccentrically along a balloon or by constraining the delivery system within a sheath during deployment. To optimize contact of the device with the shunt, we place the inferior aspect aligned within the uncovered portion of the Viatorr shunt, 3-4 mm from the edge of the covered portion of the shunt. This provides metal-PTFE contact between the 2 stent systems as an
Figure 5 (A) Initial portogram in a TIPS patient with refractory encephalopathy. The portosystemic gradient was 3 mm Hg. (B) Following deployment of the iCast covered stent, the proximal and distal ends of the stent are sufficiently dilated to enable a stable stent position. Note the narrowed midportion of the stent (arrows). (C) Completion portogram following shunt reduction (arrows) with an increase in the portosystemic gradient from 3 to 9 mm Hg. The patient’s hepatic encephalopathy resolved.
T.W.I. Clark
216 extra measure of stability; there have been anecdotal reports of the iCast device migrating from the TIPS when the device is placed in the straight outflow portion of the shunt. To continue placement of the iCast, the balloon is then deflated, repositioned along the superior outflow portion of the stent, and reinflated to produce an hourglass configuration to the stent. This technique also has the advantage of being able to serially enlarge the aperture in the shunt if the portosystemic gradient is too high (Fig. 5).
Conclusions Interventional radiologists need to be familiar with the patterns of shunt dysfunction among TIPS endografts and have an approach to managing shunt dysfunction. Prompt identification and treatment of shunt dysfunction can enable sustained and durable portal decompression, both as a bridge to liver transplantation or for patients who are not transplant candidates.
References 1. Zizka J, Elias P, Krajina A, et al: Value of Doppler sonography in revealing transjugular intrahepatic portosystemic shunt malfunction: A 5-year experience in 216 patients. AJR Am J Roentgenol 175:141148, 2000 2. Middleton WD, Teefey SA, Darcy MD: Doppler evaluation of transjugular intrahepatic portosystemic shunts. Ultrasound Q 19:56-70, 2003 3. Clark TW, Agarwal R, Haskal ZJ, et al: The effect of initial shunt outflow position on patency of transjugular intrahepatic portosystemic shunts. J Vasc Interv Radiol 15:147-152, 2004 4. Hoxworth JM, LaBerge JM, Gordon RL, et al: Inferior vena cava thrombosis after transjugular intrahepatic portosystemic shunt revision with a covered stent. J Vasc Interv Radiol 15:995-998, 2004 5. Wilson MW, Gordon RL, LaBerge JM, et al: Liver transplantation complicated by malpositioned transjugular intrahepatic portosystemic shunts. J Vasc Interv Radiol 6:695-699, 1995 6. Madoff DC, Wallace MJ, Ahrar K, et al: TIPS-related hepatic encephalopathy: Management options with novel endovascular techniques. RadioGraphics 24:21-36, 2004, discussion, 36-37 7. Sze DY, Hwang GL, Kao JS, et al: Bidirectionally adjustable TIPS reduction by parallel stent and stent-graft deployment. J Vasc Interv Radiol 19:1653-1658, 2008