| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia (Z.J.H.), and Pathology Associates International, Frederick, MD (L.H.B.). Received November 23, 1998; revision requested December 30; revision received March 24, 1999; accepted July 1. Supported in part by a research grant from Boston Scientific, Natick, Mass. Address reprint requests to Z.J.H., Department of Radiology, New York Presbyterian Hospital (Columbia Presbyterian), MHB 4-100, 177 Fort Washington Ave, New York, NY 10032.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Thirteen TIPS were created in 13 minipigs: eight with PETP-covered Wallstents, five with standard Wallstents. Shunt venography was performed at 5–8 weeks, and necropsy was performed at 7–8 weeks. Histopathologic, immunohistochemical, and scanning electron microscopic examinations were performed.
RESULTS: Mean shunt stenoses of the control and graft groups were 45% and 53%, respectively. Graft stenoses involved the entire graft-bearing segment, whereas bare stent stenoses were localized within the liver tract. Myofibroblast and extracellular collagen matrix proliferation encompassed both control and graft-covered stents. There was one graft TIPS occlusion. One control TIPS stenosis was due to transstent proliferation of normal porcine hepatic tissue. A small focus of bile staining was seen on the abluminal surface of one TIPS, which was a patent PETP-lined shunt.
CONCLUSION: PETP graft TIPS provided equal, but not superior, patency to that of bare stent TIPS. The pattern of PETP TIPS graft healing differed from that of bare stents but was similar to that reported with other polyester graft vascular implants and consisted of diffuse transmural penetration and paving of the graft surface by extracellular collagen matrix and myofibroblasts.
Index terms: Grafts, 95.1268 • Hypertension, portal, 95.711 • Interventional procedures, experimental studies, 95.1268 • Liver, interventional procedures, 761.1269, 95.1268 • Shunts, portosystemic, 95.453 • Stents and prostheses, 95.1268 • Veins, grafts and prostheses, 95.1268
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The authors created the intrahepatic shunts by using techniques identical to those described in detail elsewhere (4). In brief, the animals were sedated with intramuscular injections of ketamine hydrochloride (Phoenix Pharmaceuticals, St Joseph, Mo), xylazine hydrochloride (Phoenix Pharmaceuticals), and atropine sulfate (Phoenix Pharmaceuticals), and, in some cases, acepromazine maleate. They were endotracheally intubated, and intravenous catheters were inserted in a right or left auricular vein. General anesthesia was induced and maintained by using inhaled isoflurane. A single 500-mg dose of cefazolin sodium was injected intravenously at the start of each procedure. External jugular venous access was achieved by means of blind venous puncture by using anatomic landmarks or fluoroscopic guidance during injection of iodinated contrast material through the ear vein. A 40-cm-long 10-F sheath (Cook, Bloomington, Ind) was advanced into the inferior vena cava. A suitable left or right hepatic vein was catheterized, and a sheathed 16-gauge Colapinto needle set (Cook) was introduced into that vein. A 50-cm-long 20-gauge needle was passed through the larger needle into the parenchyma toward the anticipated region of the portal vein. One to four needle passes generally were required to puncture a branch of the portal vein, typically the left portal vein. A 0.018-inch nitinol guide wire (Cook) was passed into the superior mesenteric vein and used to guide passage of subsequent diagnostic catheters. We administered 1,000 U of heparin sodium into the portal vein. Portal venography was performed with hand injection of contrast material. The parenchymal tract was dilated to 10 mm in diameter, and the length of the shunt tract to be covered with the stents was measured by using a kinked wire technique. Metal markers were positioned over the animal's skin to mark the cephalocaudal extent of the shunt to be lined with stents.
Control Group
Five control stent TIPS were created in five animals by using one or two overlapping 10-mm-diameter, 68-mm-long standard Wallstents (Boston Scientific). The initial protocol was designed to have four animals in the control group; however, one animal died of an unrelated bacterial pneumonia at 28 days after TIPS creation, so an additional control TIPS was created in another animal.
Stent-Graft Description and Deployment
Eight stent-graft TIPS were created in eight animals by using one or two overlapping 4- or 7.5-cm-long devices. These stent-grafts consisted of Wallstents covered with a sheath of woven PETP (Fig 1a). A radiopaque wire was braided into the graft-bearing portion of the stent to aid its in vivo positioning (Fig 1b). The water entry pressure of the PETP material was 750–2,000 mL/cm2/min; the thickness of the graft material and stent was 0.28–0.38 mm. The graft material was not gelatin impregnated and did not cover the leading 1.5 cm of each stent-graft. This bare segment was intended for positioning within the portal vein. The leading edge of the graft material was placed at the start of the intrahepatic parenchymal tract, or, in cases wherein a very smallintrahepatic porcine portal vein branch was entered, a short distance into the portal vein. The trailing 1 cm of the stent was also bare.
|
|
The cephalic ends of both control and stent-graft TIPS were extended a short distance beyond the hepatic vein ostium into the inferior vena cava. This ensured that the graft material would reach the hepatic vein ostium and entirely fill the hepatic vein with the graft-containing portion of the stent-graft. All devices were dilated to a diameter of 10 mm. Portal venography was performed to confirm satisfactory stent positioning. All catheters were removed, hemostasis was achieved with manual compression of the neck, and the animals were awoken and extubated and recovered. No additional anticoagulation or antiplatelet agents, aspirin, or antibiotics were administered. All animals were fed a low-protein diet to prevent encephalopathy.
Transjugular shunt venography was performed in multiple projections at 5–8 weeks after implantation. Animals with occluded or nearly occluded shunts at 7 weeks were sacrificed immediately after venography. In all other cases, follow-up was continued to 8 weeks, when repeat venography was performed and immediately followed by necropsy and specimen retrieval. The percentage of shunt stenosis was measured by one physician (Z.J.H.). It was calculated by taking the quotient of the minimal lumen diameter (MLD) divided by the reference (stent) lumen diameter (NLD), subtracting from one, and multiplying by 100%, (ie, percentage of stenosis = [1 - (MLD/NLD)] x 100%). The MLD was the narrowest point in the TIPS lumen.
Histopathologic Examination
At the conclusion of the study, each animal was killed with an intravenous injection of pentobarbital sodium solution (Beuthanasia-D; Schering-Plough Animal Health, Kenilworth, NJ). The liver, suprahepatic inferior vena cava, and extrahepatic portal vein were removed en bloc. A suture ligature was tied to the inferior vena caval wall to aid the pathologist in identifying this segment. The shunts were gently flushed with saline solution, part of the surrounding liver was trimmed away, and the samples were placed into a 10% neutral buffered formalin solution. At pathologic examination, the shunts were cut longitudinally, and photographs of the gross specimen were obtained. The wires of both stent-grafts and control stents were dissected free with use of microscopic guidance. Specimens were stained with hematoxylin-eosin and trichrome stain. Multiple cross sections were obtained from the portal venous portion, the shunt tract, hepatic vein, and inferior vena cava. Representative specimens were imaged by means of scanning electron microscopy.
Immunohistochemical analysis of both control and stent-graft specimens was performed to help characterize the tissue responses. Paraffin-embedded sections of 13 resected TIPS were studied immunohistochemically for Bandeiraea (Griffonia) simplicifolia lectin I (BSL I)—reactive endothelial cells and immunoreactive proliferating cell nuclear antigen (PCNA), which is a marker for proliferating cells. BSL I specifically binds to galactose residues on nonprimate endothelial cells. PCNA staining was performed by using murine immunoglobulin G2a anti-PCNA reagent. Internal positive controls for both immunoperoxidase stains were provided by endothelium within preexisting hepatic vessels.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Stent-grafts.—One graft appeared free of venographically demonstrable stenosis. Six graft stenoses were present (Figs 3, 4). The mean maximum percentage of shunt stenosis was 53% (range, 0%–90%). The stenoses extended over the length of the graft lumen. One graft occlusion was present. The venographic pattern of stenosis differed from that seen in the control group in that it was more homogeneous and diffuse rather than centered on the short parenchymal tract.
|
|
|
|
One 28-day resected TIPS was occluded within the parenchymal portion of the TIPS. A thick collar of fibroblasts lined this section of the shunt with thick thrombus and completed the obstruction.
Scanning electron microscopy of a 61-day resected TIPS demonstrated that the stent wires lay beneath a relatively thin pseudointima with the roughened and irregular luminal surface typical of a fibrous thrombus (Fig 7).
|
Stent-grafts.—At gross examination, the luminal surfaces of the patent grafts revealed a relatively smooth, glistening, white opaque surface (Fig 8). The occluded graft contained a firm, dark red to tan thrombus that filled virtually the entire shunt. Bile staining was noted in only one location—on the abluminal side of a thin area of pseudointima. The tissue covering the PETP and wires in the stent-grafts consisted of layers of fibrous connective tissue: fibroblasts and extracellular collagen matrix (Figs 9, 10). The thickness of tissue varied from a relatively thin uniform pseudointima to a uniformly thick layer lining the entire tract. The tissue layer was much thicker over the stent-graft sections than over the bare wire sections. The one occluded shunt was similarly filled with an inner rind of this tissue, with a large thrombus completing the occlusion. In one case wherein multiple PETP grafts were placed to line the tract, the pseudointima was essentially double the thickness of other areas, because thick pseudointima was present between the overlapping segments of the grafts, as well as on the luminal aspect of the inner stent-graft. As with the control TIPS, the majority of the pseudointimal tissue was found on the luminal surfaces of the stent-grafts.
|
|
|
Scanning electron microscopy of PETP stent-graft specimens was performed at 47 and 49 days. This revealed luminal surface irregularity due to adherence of fibrous or fibrinous material to the surface. A cut surface demonstrated the thick layer of fibrous connective tissue beneath the luminal surface (Fig 11).
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
A number of graft coatings have been tested for TIPS applications in both porcine and human applications, including PTFE, silicone, and modified PETP. In 1995, Nishimine et al (1) compared the results of 13 porcine TIPS lined with PTFE sewn onto a Wallstent and a Gianturco Z stent (Cook) scaffold with 13 TIPS lined with conventional bare stents. At 1 month, 69% of the grafts showed a stenosis of less than 50%, compared with only 8% of the controls. At 3 months, 46% of the grafts still had a stenosis of less than 50%. In 1997, we reported results of using PTFE-encapsulated stents in eight porcine TIPS (4). With the exception of one shunt occlusion related to rapid growth of the animal, the grafts remained free of detectable stenosis at up to 5 months, while control bare stent TIPS developed occlusions or 45%–85% stenoses by 6 weeks. The improved results of that series compared with those of Nishimine et al may partly reflect that the greater thickness of the PTFE in the encapsulated devices provides more of a barrier to transmural cellular ingrowth. In both series, the PTFE was well tolerated by the surrounding tissue and host animal. In 1997, Tanihata et al (3) created TIPS in 14 growing pigs by using Wallstents coated with an impermeable silicone. By 3 weeks, only 14% of shunts were patent; these two shunts had midshunt stenoses greater than 50%. The marked thrombotic response and foreign body reaction incited resulted in patency much poorer than that seen with bare stents.
These studies have helped guide experimental human implantation of custom-built TIPS stent-grafts. At early follow-up, the success of PTFE-lined TIPS in pigs has been borne out in humans. Saxon et al (2) reported a pilot study of six patients treated with PTFE attached to Wallstents and Gianturco Z stents, similar to the porcine TIPS graft of Nishimine et al (1). Half of these patients had biliary-TIPS fistulae. They estimated an increase in shunt patency from a mean of 50–229 days. Our results with PTFE-covered Wallstents for both TIPS creation and failing shunt revision have been similar (16). In our experience with PTFE-covered Wallstents constructed in-house, similar patencies were achieved: 13 of 14 grafts were widely patent at a mean 19-month follow-up (stenosis of 10% or less) (16). Ferral et al (5) recently reported preliminary results for creating de novo TIPS in 13 patients by using a polyester-coated stent (modified Cragg Endopro; Mintec, Bahamas). Using ultrasonographic surveillance, they detected two shunt occlusions and one shunt stenosis at 2-and 3-month follow-up. Mid- or long-term venographic documentation of percentage of shunt stenosis in the cohort was not described in this technical note.
Our experiment showed that the modified PETP-covered Wallstent, similar to the uncovered Wallstent endoprosthesis used for TIPS that is approved by the U.S. Food and Drug Administration, could be delivered and deployed precisely and accurately in a TIPS application. Our results with PETP-lined porcine TIPS failed to show the same prolongation of patency seen with PTFE-lined TIPS. Simply comparing mean percentage of shunt stenosis, it would seem that the PETP stent-graft results were equivalent to bare stent TIPS at follow-up. However, a different pattern of healing distinguished the two groups. The control group (bare stent) showed the typical prompt development of focal stenoses centered in the parenchymal tract, with the remaining, intravenous portions of the stents relatively free of intimal tissue. This proliferative tissue developed in the absence of bile staining. In contrast, the PETP grafts showed a response characteristic of porous polyester grafts placed in other vascular applications: diffuse encapsulation of the graft within a fibrous tissue lining associated with a foreign body reaction and inflammation (10,17,18).
In our experiment, the tissue proliferation was similar to that seen by Schürmann et al (19) in their comparison of Dacron-coated stent-grafts with noncovered stents in iliac arteries in sheep. It was more pronounced than that seen by Dolmatch et al (10) with PETP-covered Wallstent placed in canine arteries; this may be partly attributable to the more demanding nature of the porcine model, as well as the specific TIPS application. A slight focal increase in inflammation was present at the ends of the grafts where silicone glue was used to adhere the graft to the stents. Notably, the tissue growth stopped relatively abruptly at the transition to the bare, uncovered portion of the stents.
Several investigators have implicated the thrombogenic role of biliary-TIPS fistulae in the formation of stenoses in both human and porcine TIPS. LaBerge et al (20) suggested this in 1993, when they reported on bile duct proliferation within the tissue lining several stenotic TIPS. Saxon et al (21) compared stenotic TIPS in humans and pigs and found bile staining in seven (77%) of nine stenotic or occluded porcine TIPS. Jalan et al (22) performed biopsy in stenotic TIPS and found bile incorporated within thrombus in several cases. In three cases, major bile duct transection was closely related to shunt stenosis. In contrast, Sanyal et al (23) characterized the histologic findings in 35 shunts and found smooth muscle cell proliferation in both stenotic and nonstenotic TIPS, independent of the gross morphology or presence of biliary fistulae within the shunts. In our experience, bile staining played a visible and potent role in rapid and recurrent shunt thrombosis in a prominent but small portion of the human patients with stenotic TIPS but in none of the pigs in our studies. In more than 80 porcine TIPS lined with a variety of bare or graft-coated stents, bile staining has been conspicuously absent, perhaps because of the breed of minipigs that we use; the sole case of bile staining was seen in this experiment in association with a PETP-lined shunt.
A recent investigation by Teng et al (24) may shed some light on this controversy. They cultured smooth muscle cells with bile, serum and bile, and serum alone and found that bile was a powerful inhibitor of smooth muscle cell proliferation. These findings support the concept that TIPS stenosis is multifactorial. Early, rapid shunt thrombosis may be related to bile and its inhibitory effect on graft and stent healing. Both liver-related factors and blood-borne elements may cause later, nonthrombotic parenchymal tract and hepatic vein stenoses. The development of a perigraft collar of myofibroblasts surrounding porcine TIPS lined with 30-µm internodal distance PTFE suggests that these myofibroblasts represent a hepatic response to the injury of balloon dilation and stent implantation during TIPS formation (4). On the other hand, the hepatic vein stenoses that affect most shunts during follow-up probably are related to blood-borne or fluid dynamics phenomena implicated in intimal hyperplasia, such as shear stress (25,26).
It may be argued that an ideal TIPS graft must address all these factors to provide durable shunt patency. As such, we cannot address the efficacy of the PETP TIPS graft in preventing bile-related shunt thrombosis because bile staining was seen in only one instance—a shunt in which minimal stenosis was found. On one hand, the high porosity of the PETP graft may prove an ineffective barrier for bile passage, perhaps even serving as a bile wick and resulting in TIPS thrombosis similar to that seen with uncovered stents. On the other hand, the PETP inflammatory response may cause rapid fibroblast aggregation, overcoming the inhibitory thrombotic effect of bile and potentially sealing that site.Practical application: These experimental results indicate that there is no clear benefit to the use of PETP-covered stent-grafts for prolongation of newly created TIPS; the pattern of PETP TIPS graft healing differed from that of bare stents but was similar to that reported with other polyester graft vascular implants—diffuse transmural penetration and paving of the graft surface by extracellular collagen matrix and fibroblasts.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, Z.J.H.; study concepts and design, Z.J.H.; definition of intellectual content, Z.J.H.; literature research, Z.J.H.; clinical studies, Z.J.H.; experimental studies, Z.J.H., L.H.B.; data acquisition, Z.J.H.; data analysis, Z.J.H., L.H.B.; manuscript preparation and editing, Z.J.H.; manuscript review, Z.J.H., L.H.B.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Hoppe, S. L. Wang, and B. D. Petersen Intravascular US-guided Direct Intrahepatic Portocaval Shunt with an Expanded Polytetrafluoroethylene-covered Stent-Graft Radiology, January 1, 2008; 246(1): 306 - 314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Maleux, C. Verslype, S. Heye, G. Wilms, G. Marchal, and F. Nevens Endovascular Shunt Reduction in the Management of Transjugular Portosystemic Shunt-Induced Hepatic Encephalopathy: Preliminary Experience with Reduction Stents and Stent-Grafts Am. J. Roentgenol., March 1, 2007; 188(3): 659 - 664. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Vignali, I. Bargellini, M. Grosso, G. Passalacqua, F. Maglione, F. Pedrazzini, P. Filauri, R. Niola, R. Cioni, and P. Petruzzi TIPS with Expanded Polytetrafluoroethylene-Covered Stent: Results of an Italian Multicenter Study Am. J. Roentgenol., August 1, 2005; 185(2): 472 - 480. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. T. Lau, M. Scott, S. W. Stavropoulos, M. C. Soulen, J. A. Solomon, and T. W. I. Clark Dacron-covered Stent-Grafts in Transjugular Intrahepatic Portosystemic Shunts: Initial Experience Radiology, August 1, 2005; 236(2): 725 - 729. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G.-J. Teng and Q. Lu Bile Leakage during Transjugular Intrahepatic Portosystemic Shunt Creation: In Vitro Effect of Bile on Growth and Function of Human Umbilical Vein Endothelium Radiology, June 1, 2005; 235(3): 867 - 871. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Rossi, F. M. Salvatori, F. Fanelli, M. Bezzi, M. Rossi, G. Marcelli, D. Pepino, O. Riggio, and R. Passariello Polytetrafluoroethylene-covered Nitinol Stent-Graft for Transjugular Intrahepatic Portosystemic Shunt Creation: 3-year Experience Radiology, June 1, 2004; 231(3): 820 - 830. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. C. Madoff, M. J. Wallace, K. Ahrar, and R. R. Saxon TIPS-related Hepatic Encephalopathy: Management Options with Novel Endovascular Techniques RadioGraphics, January 1, 2004; 24(1): 21 - 36. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Wallace, K. Ahrar, L. C. Stephens, and K. C. Wright Transvenous Extrahepatic Portacaval Shunt: Feasibility Study in a Swine Model Radiology, July 1, 2003; 228(1): 119 - 125. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J D G Rose, S Pimpalwar, and R W Jackson A new stent-graft for transjugular intrahepatic portosystemic shunts Br. J. Radiol., October 1, 2001; 74(886): 908 - 912. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. J. Haskal Improved Patency of Transjugular Intrahepatic Portosystemic Shunts in Humans: Creation and Revision with PTFE Stent-Grafts Radiology, December 1, 1999; 213(3): 759 - 766. [Abstract] [Full Text]
|
|
|
|
|
|
M. Cejna, M. Peck-Radosavljevic, S. A. Thurnher, K. Hittmair, M. Schoder, and J. Lammer Creation of Transjugular Intrahepatic Portosystemic Shunts with Stent-Grafts: Initial Experiences with a Polytetrafluoroethylene-covered Nitinol Endoprosthesis Radiology, November 1, 2001; 221(2): 437 - 446. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
