Stenoses of Vascular Anastomoses After Hepatic Transplantation: Treatment with Balloon Angioplasty

167 Stenoses of Vascular Anastomoses After Hepatic Transplantation: Treatment with Balloon Angioplasty Nigel Raby1 Vascular complications after liver transplantation include occlusion or stenosis at the J. Karani1 sites of anastomosis in the hepatic artery, portal vein, and vena cava. From our S. Thomas1 experience with more than 600 liver transplants, vascular stenoses have been identified J O Gradv2 in 10 patients and treated by balloon angioplasty in nine. Three patients with hepatic Roer Williams2 artery stenosis and deteriorating graft function were treated by balloon angioplasty with, a coaxial technique. A specially designed catheter facilitated a successful femoral artery approach. Portal vein stenoses in three patients resulted in portal hypertension. These were treated by balloon dilatation via transhepatic catheterization of the portal vein. Stenoses of the suprahepatic caval anastomosis were dilated in three patients with severe lower limb edema. Technical success was achieved in all three cases of hepatic artery stenosis with improvement in graft function. Recurrent stenoses in two patients were successfully treated with repeated dilatations. Portal hypertension resolved in two of three patients after portal venoplasty. Dilatation of a caval stenosis resulted in the resolution of leg edema in all three cases. Repeated dilatation was required in one case. No reduction in the portal venous pressure gradient occurred after venoplasty in one case, and an ultimately fatal caval thrombosis developed in one patient with caval stenosis before venoplasty could be performed. Our experience suggests that balloon angioplasty of arterial and venous stenoses complicating hepatic transplantation carries little risk and is a useful procedure for the treatment of these problems. AJR 157:167-171, July 1991 During orthotopic liver transplantation, four vascular anastomoses are fashioned. The donor hepatic artery and portal vein are anastomosed to the recipient vessels. The intrahepatic portion of the inferior vena cava requires two anastomoses of donor to recipient cava: one at the level of the diaphragm (the suprahepatic cava) and one below the liver above the level of the renal vessels (the infrahepatic cava). Stenosis or occlusion of any of these anastomoses may occur postoperatively. In this report, we describe nine cases in which anastomotic stenoses were treated by percutaneous angioplasty with improvement in graft function or relief of other significant complications. Another patient with caval stenosis who developed fatal caval thrombosis proximal to a suprahepatic stenosis is also described. Materials and Methods Received October 29, 1990; accepted after re- The 1 0 cases included in the study were culled from our experience with more than 600 vision January 11, 1991. liver transplants. Posttransplantation angiography was done to assess arterial patency in 1 Department of Radiology, Kings College Hos- cases of graft failure and suspected chronic rejection in which liver biopsy had failed to ptal, Denmark Hill, London SE5 9RS, England. identify a specific cause and in patients with biliary strictures or hepatic abscesses. Hepatic 2 Liver Unit, Kings College Hospital, Denmark angiography, arterioportography, and inferior venacavography were all performed via a Hill, London SE5 9RS, England. transfemoral approach, with images obtained on conventional film and with digital subtraction 0361-803X/91/1571-0167 imaging. Arterial stenosis was defined as a solitary short narrowing of the hepatic artery at American Roentgen Ray Society the site of anastomosis just distal to the origin of the gastroduodenal artery. Narrowing of

--- 168 RABY ET AL. AJR:157, July 1991 A more than 70% of the transverse diameter was considered to be significant. Hepatic artery angioplasty was done with a coaxial balloon and a Thomas catheter (Schneider, London, U.K.), specifically designed for this purpose. This is a 7-French catheter that is introduced into the femoral artery with a sheath. The sidewinder configuration is formed at the aortic bifurcation because the catheter is too short to allow formation in the aortic arch. Once formed, the catheter can be used selectively to catheterize the common hepatic artery, thus allowing delivery of the balloon catheter coaxially at the site of stenosis (Fig. 1). A satisfactory result after angioplasty was defined as reduction of the narrowing to no more than 30% of the arterial transverse diameter. Portal vein studies were indicated when evidence of raised portal venous pressure, manifested as variceal bleeding or splenomegaly, was detected clinically and sonographically. Further investigation consisted of measurement of hepatic wedge pressure and splenic pulp pressure. Hepatic wedge pressures were obtained by using cobra or sidewinder catheters with balloon occlusion of the hepatic vein. Splenic pulp pressures were measured after direct splenic puncture with a 23-gauge needle. A difference of 5 mm Hg between these two measurements was taken as evidence of a significant pressure gradient across the stenosis that warranted venoplasty. A pressure gradient across the stenosis was confirmed during transhepatic balloon venoplasty. Portal venoplasty has been described in detail elsewhere [1 ]. The portal vein was punctured via an intercostal transhepatic route and catheterized with a fine-needle introducer. A 4 L ) 5-French sheath was inserted over a guidewire into the donor portal vein. A 5-mm precurved balloon catheter was passed coaxially across the stenosis and inflated. In one case, two balloons were used (Fig. 2) to obtain the necessary effective diameter in a large portal vein. The pressure gradient across the stenosis was measured before and after inflation. A gradient across the stenosis of more than 5 mm Hg was considered abnormal, and reduction of the gradient to below this level after venoplasty was considered to indicate technical success. Patients were examined by venacavography because of bilateral lower limb edema. Pressure studies across the anastomoses in the inferior vena cava were obtained with a single end-hole catheter. Caval venoplasty was achieved via a right femoral vein approach. The stenosis was traversed with a guidewire, and a conventional balloon angioplasty catheter was inflated across the stenosis. The pressure gradient was measured before and after the procedure. A gradient of 1 0 mm Hg was considered to be abnormal, and reduction of the gradient to less than this was considered to indicate success. All procedures were done with the patients taking prophylactic antibiotics. Results Anastomotic stenoses of the hepatic artery were successfully dilated in three adult patients. These stenoses occurred, Fig. 1.-Hepatic angiograms of 36-year-old man with deteriorating liver function. A, Anastomosis has 90% stenosis just distal to origin of gastroduodenal artery. B, Thomas guide catheter in common hepatic artery. A fine guidewire has been passed across stenosis. C, Balloon catheter passed coaxially through Thomas catheter and inflated across stenosis. D, After angioplasty, severity of anastomotic narrowing is markedly reduced to 30%.

AJA:157, July 1991 ANGIOPLASTY OF STENOSES OF VASCULAR ANASTOMOSES 169 Fig. 2.-Angiograms of 6-year-old boy with anemia and splenomegaly. A, Aortoportogram shows anastomotic narrowing of portal vein (arrows). B, Transhepatic portal vein puncture confirms anastomotic stenosis (arrows) and allows direct measurement of pressure gradient. C, Two balloons inflated across stenosis. Pressure gradient measured when catheter was withdrawn across stenosis confirmed reduction of gradient. at 1 0 days, 2 months, and 8 months after transplantation. All patients had abnormal liver function as evidenced by elevated levels of bilirubin and liver enzymes. These showed marked improvement after angioplasty. Repeated arteriography showed no evidence of stricture recurrence at 6 months in one patient who underwent a second transplantation shortly afterward because of vanishing bile duct syndrome. One patient reexamined on two occasions because of deterioration in liver function showed restenosis at 6 and 9 months after the initial procedure. On each occasion, the degree of stenosis, although still abnormal, was progressively less. Results of liver function tests became normal after each angioplasty. This patient remains well, with normal liver function 9 months after the last angioplasty. The third patient had an intrahepatic abscess and strictures of the common hepatic duct. These were managed by endoscopic placement of an 8-French biliary stent and percutaneous drainage of the abscess. Repeated arteriography at 4 and 8 months after angioplasty showed restenosis on each occasion, but these were also progressively less severe and were successfully redilated. Twelve months after initial presentation, the liver parenchyma and biliary tree are normal on sonographic examination, and the patient is well, with normal findings on liver function tests. Portal vein stenosis was diagnosed in three patients (two children aged 2 and 4 and one adult) 2, 4, and 5 years after transplantation, respectively. Gradients across the stenoses of 24 and 12 mm Hg in the children were reduced to 2 mm Hg after venoplasty. Both have reduction in splenomegaly on sonographic examination, and no evidence of variceal bleeding has been seen in 1 year of follow-up. The adult patient had a stenotic gradient of 1 4 mm Hg, which did not alter after venoplasty. Spleen size has not altered, however, and the patient has not had any episodes of variceal bleeding. Four cases of suprahepatic caval stenosis were encountered at 8 and 1 5 months in two cases and at 2.0 and 4.5 years in the other two cases. In one case, caval thrombosis occurred 1 week after demonstration of the stenosis, and pressure studies were not obtained. In this case, despite intracaval streptokinase (Hoechst, Hounslow, U.K.) and subsequent surgical thrombectomy, the patient died as a result of recurrent caval thrombosis. The presence of a significant pressure gradient was confirmed in the three remaining cases. In one of these, venacavography showed reversed flow down the inferior vena cava, mimicking a caval thrombosis (Fig. 3). Superior venacavography from the arm showed a suprahepatic caval stenosis with reversed flow in the inferior vena cava, with the hepatic and renal veins decompressing retrogradely into the inferior vena cava and hemiazygous collaterals. After dilatation of the stenosis, flow within the inferior vena cava was in a normal direction and the pressure gradient was reduced from 24 to 6 mm Hg. Leg edema resolved after the procedure but recurred 9 months later. Recurrent stenosis confirmed by venacavography was treated by another balloon dilatation with a similar good result, which has been maintamed for a year. The other two patients remain free of symptoms 1 0 and 1 4 months after treatment. Discussion Vascular complications of liver transplantation that can be corrected by nonsurgical methods are rare. The 1 0 cases described in this paper represent our experience with more than 600 transplantations. Hepatic artery occlusion in the immediate posttransplantation period requires urgent surgical intervention, usually retransplantation. In some patients, slowly progressive graft ischemia occurs, usually due to arterial thrombosis. In a few instances, however, ischemia is due to an anastomotic stenosis. The usual reason for examining these patients is the development of abnormal liver function; sonography is the primary method of examination. If biliary obstruction is ex- ;

170 RABY ET AL. AJR:157, July 1991 A -C a Lt cluded, a liver biopsy is performed and if the diagnosis is not clear cut, assessment of graft vascularization is indicated. Doppler sonography often can exclude occlusion of the hepatic artery and may suggest the presence of an arterial stenosis, but angiographic evidence is required for confirmation in such cases [2]. Additionally, the detection by sonography of focal areas of altered reflectivity within the hepatic parenchyma strongly suggests graft ischemia [2]. Attempts at treatment without ensuring that the graft is adequately vascularized are generally unsuccessful. lschemia also is regarded as a factor in the development of bile duct strictures [3], although the association is not as strong as with parenchymal abnormalities. One of our patients with arterial stenosis had both focal liver lesions and biliary stricturing in keeping with the above observations. Stenosis of the hepatic artery is uncommon, although occlusion is encountered more frequently [4]. Only four previous cases of percutaneous hepatic artery angioplasty have been described [4-6]. In two of these cases, a transfemoral approach failed and a brachial artery had to be used. This problem was overcome in our series by the use of a guide catheter specially developed for angioplasty of renal and hepatic transplants. The Thomas catheter allows Fig. 3.-Angiograms of 25-year-old woman with edema in both legs. A and B, Anteroposterior (A) and lateral (B) views of inferior venacavogram show nonfilling of intrahepatic cava, mimicking thrombosis. A superior caval injection showed a suprahepatic caval anastomotic stenosis with retrograde flow in inferior vena cava and hemiazygous collaterals. C, Balloon dilatation of caval stenosis. Note waisting of balloon (arrow) at site of stenosis. D, Postangioplasty appearance. Some residual stenosis persists (arrows), but pressure gredient is gone, and normal direction of flow reestablished. selective catheterization of the common hepatic artery and the subsequent passage of an angioplasty balloon through the guide catheter and across the stenosis. Although longterm experience is limited, it appears that, as with arterial stenoses encountered in the transplanted kidney, restenosis after angioplasty is highly probable. However, as in renal transplantation patients, repeated angioplasty of the stenosis may be beneficial [7] and ultimately may result in satisfactory long-term patency, as has occurred in two patients described here. To our knowledge, the use of percutaneous venoplasty of the portal vein after liver transplantation has not been described previously. According to Wozney et al. [4], significant portal vein stenosis is an uncommon complication after transplantation. Our experience agrees with that of Wozney et al., who noted that this complication usually occurs several years after transplantation and most often in children. Increasing spleen size and the development of varices with subsequent gastrointestinal hemorrhage and anemia are signs of this complication. Narrowing of the site of the portal vein anastomosis is seen commonly on real-time sonography [8], and Doppler studies often show turbulent flow at the anastomosis, but such abnormalities do not appear to be of functional

AJA:157, July 1991 ANGIOPLASTY OF STENOSES OF VASCULAR ANASTOMOSES 171 significance. When patients are being considered for portal vein angioplasty, it is essential to confirm with pressure studies that the stenosis is hemodynamically significant by showing a pressure gradient across the stenosis. After a patent but narrowed portal vein is demonstrated, hepatic wedge pressure and direct splenic pulp pressure measurements enable patients with a significant gradient to be selected without traversing the graft parenchyma. Once a pressure gradient has been identified, it can be confirmed during transhepatic portal vein puncture. Stenosis of the suprahepatic caval anastomosis also is rare. Graft function is not impaired unless hepatic venous outflow is severely compromised; thus, caval stenosis manifests as lower limb edema. The diagnosis can be suggested by sonography, but angiographic confirmation with evidence of a significant pressure gradient must be obtained. Percutaneous dilatation is relatively straightforward, and in the three cases in which this was achieved, the symptoms have resolved. Repeated dilatation at 9 months was successful in one patient, and all three patients have had long-term relief of symptoms. The finding of a significant stenosis should be dealt with urgently, because in one patient a delay of 1 week after the initial diagnosis of stenosis resulted in the development of caval thrombosis proximal to the stenosis. The occurrence of acute caval thrombosis carries a very poor prognosis, with both thrombolytic therapy and surgery failing to prevent death in this case. The association of thrombus proximal to caval stenoses has been noted before [9]. In his series, Cardella et al. [9] noted several cases of caval thrombosis that extended into the hepatic veins. We have not encountered this complication. Figure 3 shows how abnormal flow in the inferior vena cava proximal to a stenosis may lead to nonopacification of blood in the inferior vena cava, thus leading to an erroneous diagnosis of thrombus formation. Venacavography from above and below the stenosis may be required to define the situation accurately. If a flow artifact is not recognized as such, thrombolytic therapy may be instituted needlessly. Other than the first case in this series, which has been previously reported [1 0], we are aware of only one other balloon dilatation of the suprahepatic caval anastomosis [11]. Rose et al. [12] reported the dilatation of an infrahepatic caval stenosis, but we have not seen any examples of this complication. With the improving survival rate of patients after liver transplantation, the number of patients with late complications of vascular anastomoses is likely to increase. The addition of successful nonsurgical measures to deal with these complications is therefore welcome. Long-term follow-up is, however, required to define fully the role of these techniques. Short-term technical success is achievable in most cases. The techniques described are safe, and the only significant complication has been sepsis after portal vein angioplasty. No serious sequelae occurred. REFERENCES 1. Ollift 5, Pain J, Karani J, Mowat AP, Williams A. Percutaneous transhepatic dilatation of late portal vein stenosis following orthotopic liver transpiantation. Intervent Radiol (in press) 2. Segel M, Zajko A, Delbert A, Bron K, Skolnick M, Starzl T. Hepatic artery thrombosis after liver transplantation. AJR 1986:146: 137-1 41 3. Zajko A, Campbell W, Logsdon G, et al. Cholangiographic findings in hepatic artery occlusion after liver transplantation. AJR i987;149: 485-489 4. Wozney P, Zajko AB, Bron KM, Point 5, Starzl TE. Vascular complications after liver transplantation. Radiology 1986:147:657-663 5. Abed J, Hidalgo EG, Cantarero JM, et al. Hepatic artery anastomotic stenosis after transplantation: treatment with percutaneous transluminal angioplasty. Radiology 1989;171 :661-662 6. Zajko AB, Bron KM, Starzl TE, et al. Angiography of liver transplantation patients. Radiology i985;1 57:305-311 7. Raynaud A, Bedrossian J, Remy P. Brisset J-M, Angel C-Y, Gaux J-C. Percutaneous transluminal angioplasty of renal transplant arterial steno- 505. AJR i986;146:853-857 8. Raby N, Meire H, Forbes A, Williams A. The role of ultrasound scanning in management after liver transplantation. Clin Radiol 1988:39:507-510 9. Cardella JF, Castaheda-Z#{252}#{241}iga WF, Hunter D, Young A, Amplatz K. Angiographic and interventional radiologic considerations in liver transplantation. AJR i986;146:143-153 10. Reading NG, O Grady JG, Williams A, Nunnerly H. Balloon dilatation of an anastomotic stenosis of the inferior vena cava. Intervent Radio! 1987:2: 81-82 11. Cardella JF. Evaluation and treatment of vascular problems after liver transplantation. Semin Intervent Radiol 1986:3:131-138 12. Rose BS, Van Aman EM, Simon DC, et al. Transluminal balloon angioplasty of infrahepatic caval anastomotic stenosis following liver transplantation (case report). Cardiovasc Intervent Radiol i988;1 1 :79-81