Kenneth L. Croutch, * R q L. Gordon, * Ernest J. Ring, * Robert K. Kerlan Jr., * Jeanne M. LclBerge, * andjohn P. Roberts t. Patients and Methods

Transcription

1 Superselective Arterial Embolization in the liver Transplant Recipient: A Safe Treatment for Hemobilia Caused by Percutaneous Transhepatic Biliary Drainage Kenneth L. Croutch, * R q L. Gordon, * Ernest J. Ring, * Robert K. Kerlan Jr., * Jeanne M. LclBerge, * andjohn P. Roberts t Transcatheter arterial embolization is widely used to treat life-threatening iatrogenic hemobilia, although in the transplanted liver its use has only been reported in two cases. To evaluate more fully whether transcatheter embolization is safe and effective in transplant recipients, we retrospectively reviewed eight cases of severe hemobilia. These occurred after 128 percutaneous transhepatic biliary drainage procedures performed during a 6-year period. In each case, angiography localized the bleeding to a specific intrahepatic branch that was then subselectively catheterized and occluded by transcatheter embolization. Bleeding was successfully controlled by this method in all eight patients with no immediate complications. The main, right, and left hepatic arteries were shown to be patent immediately after embolization by angiography in all patients. Duplex sonography performed in each case (1 to 4 months) after the procedure confirmed that patency was maintained in all patients. No patients developed liver abscesses, sepsis, or clinical liver infarctions after the embolization. No patients underwent retransplantation after embolization. Our experience shows that superselective transcatheter embolization is a safe, effective therapy to correct iatrogenic hemobilia in the liver transplant recipient without threatening the patency of the major hepatic arteries, the viability of the liver, or the integrity of the biliary tree. Copyright by the American Association for the Study of Liver Diseases ver the last few years, liver transplantation has 0 become an increasingly common treatment for various forms of liver failure. Percutaneous transhepatic biliary drainage (PTBD) is used to treat complications such as cholangitis, biliary obstruction, and anastomotic leakage that occur in transplant recipients. Hemobilia is a well-recognized but uncommon complication of PTBD.2.3 When the bleeding is of venous origin, it can usually be managed conservatively by repositioning the drainage catheter or by changing it for a tube of larger diameter.* However, massive hemobilia caused by injury to an intrahepatic artery may necessitate extensive transfusions and may even lead to exsang~ination.~,~ Subselective percutaneous transcatheter arterial embolization is now the accepted method of choice for management of hemobilia in the nontransplanted liver that does not respond to catheter From the Departments of *Radiology and tsurgeiy, University of California, San Francisco, CA. Address reprint requests to: Roy L. Gordon, MD, Department of Radiology, Box 0628, University o/ Caljfomia, Sun Francisco, Sun Francisco, CA Copynght hy the American Association for the Study of Liver Diseases /96/ $3.00/0 However, the use of transcatheter embolization to correct hemobilia from PTBD in patients who undergo transplantation has only been reported in two cases.h Some authors have drawn attention to the potential complications of liver infarction or bile duct ischemia caused by hepatic artery (HA) embolization. %* In general, the reluctance to Ifianipulate and embolize the HA in the posttransplantation setting stems from the fear of inducing postoperative hepatic artery thrombosis. Arterial thrombosis, particularly in the early posttransplantation period, has devastating consequences to the transplanted liver. This is manifested primarily by biliary ischemia with subsequent strict~ring.~~ ~ In this retrospective report, we discuss our experience with angiographic transcatheter embolization in eight liver transplantation recipients who developed hemobilia from PTBD. Our experience shows that subselective transcatheter embolization is a safe, effective therapy to correct hemobilia from PTBD in the liver transplantation patient without threatening the viability of the transplanted liver. Patients and Methods Between January 1, 1989 and June 1, 1995, 580 patients underwent 625 liver transplantations. Of these 625 liver transplan- 118 Livev Transplantation and Surgq, Vol2, No 2 (March), 1996: pp

2 Arterial Embolimation for Hemobilia 119 tations, a total of 74 patients (24 females and 50 males; mean age, 45.5 years; range, 5 months to 67.1 years) underwent 128 PTBDs Retrospective chart rewew of this group identified 8 patients (two females and six males; mean age, 34.5 years; range, 27 to 64.2 years) who developed severe hemobilia. Transcatheter arterial embolization was used to treat the bleeding in each case. Table 1 summarizes the clinical information on these eight patients. An initial diagnostic percutaneous transhepatic cholangiogram (PTC) was performed on all patients with a 22- gauge Chiba needle (Cook, Bloomington, IN) from a right-sided intercostal approach in six patients and a left- sided anterior approach in two patients In 5ix subsequent PTBDs, a peripheral duct (tertiary or quartenary) was punctured with a Lunderquist PTC needle (19-gauge) sheath (4.8-French; Meadox Surgimed, Oakland, NJ) under fluoroscopic guidance. The ductal system was opacified by contrast injection through the retained 22-gauge Chiba needle. In the remaining two patients, a wire was passed through the originally placed Chiba needle to gain access to the biliary tree, and then the tract was enlarged using a Neff percutaneous access set (Cook). Selective hepatic angiography was performed via a transfemoral approach using a 5-French (F) Cobra catheter (MediTech, Watertown, MA) to show the bleeding artery. Once an area of extravasation was identified, the specific bleeding branch vessel was superselectively catheterized using an inner coaxial 2.7-F catheter (Target Therapeutics, Mountain View, CA) (Figure la), Microcoils (Cook, Bloommgton, IN), and/or gelatin sponge pledgets (Gelfoam; Upjohn, Kalamazoo, MI) were then deployed superselectively to occlude the bleeding branch (Fig. 18). A completion angiogram was performed to confirm cessation of hemorrhage and patency of the remainder of the hepatic artery and its branches. Clinical cessation of hemobilia was confirmed by evaluating the vital signs (blood pressure and pulse rate) and bile color, as well as available pre-embolization and postembolization hematocrits. All patients underwent follow-up Doppler ultrasound examination of the main, right, and left hepatic arteries at an average of 106 days after embolization. Viability of the transplanted liver postembolization was evaluated by review of the patients charts in collaboration with the transplant service Results Of the eight patients with hemobilia requiring embolization, six patients had right-sided PTBD, and two patients had left-sided PTBD. Catheter sizes ranged from 8.5-F to 10-F. Analysis of the radiographs from the PTBD procedures showed a puncture of a tertiary or of a more peripheral duct, in all cases. Table 1. Summary of Clinical Information Cause of Liver Source of Em bolization Failure Reason for PBD Hemobilia Aqent Follow-UP Alcohol Autoimmune Alcohol Autoimmune Cryptogenic Hepatitis C and alcohol Heoatitis B Alcohol Fever/bile leak Bile leak Ischemic bile duct stricture Obstruction: increased bilirubin/alkaline phosphatase Obstruction: increased bilirubin/alkaline phosphatase Fever with increased bilirubin/alkaline p hosphatase Obstruction: increased bilirubin/al kaline p hosp hatase Bile leak Right HA branch to bile Left HA branch to bile Right HA branch to bile Left HA branch to bile Right HA branch to right portal vein fistula Right HA branch pseudoaneurysm Right HA branch pseudoaneurysm Right HA branch pseudoaneurysm and gelfoam and gelfoam and gelfoam and gelfoam Alive: 31 months 8, angio day 59 Alive: 35 months 132 Dead: 15 months (bile duct ischemia) 324, angio day 148 Alive: 59 months 4 Dead: 93 days (rejection) 75 Alive: 7 months 90 Dead: 87 days (rejection) 74 Alive: 67 months 143

3 120 Croutch et nl Figure 1. (A) Subselective angiogram shows pseudoaneurysm (fat arrow) and extravasated contrast (small arrows). Percutaneous biliary drain is shown faintly (curved arrows). (B) After superselective embolization with platinum coils (arrow), the bleeding hepatic artery has been occluded and no further bleeding is shown. The onset of hemobilia from the time of PTBD was a mean of 17.8 days (range, 1 to 60 days). From the onset of hemobilia, the time until diagnostic angiographic embolization was 1 to 11 days (mean, 4.8 days), and the number of units of packed red blood cells (PRBC) required before embolization was 0 to 13 U (mean, 4 U). In the eight patients with hemobilia diagnosed by angiography, embolization was performed with either coils (four patients) or coils and gelfoam (four

4 Arterial Embolizution for Hemobilia 121 patients). Subselective embolization via a coaxial microcatheter was achieved in all patients. In seven of the eight patients, the mean prothrombin time (PI-), partial thromboplastin time (PTT), and platelet count were 12.1 seconds (range, 10.4 to 14 seconds), 31.5 seconds (range, 25 to 39 seconds), and 244,000 (range, 55,000 to 422,0001, respectively. One patient had elevated PT and PTT measuring 24.8 seconds and greater than 100 seconds, respectively, along with a platelet count of 50,000. In this patient with elevated PT and PTT and low platelet count, repeat angiography and transcatheter embolization with deployment of additional coils and gelfoam was required 1 day after the initial therapy to stop rebleeding. The bleeding source was a pseudoaneurysm in three patients, a hepatic artery to bile in four patients, and an arterial-portal fistula in one patient. Five patients required no blood transfusions after embolization. Three patients were transfused with a mean of 3.3 U of PRBC (range, 2 to 4 U) after embolization. No patients developed liver abscesses, sepsis, clinically apparent liver infarctions, or ishemic bile duct changes related to the embolization. In all patients, the biliary drainage catheters were removed without further bleeding. No complications related to the procedure occurred. Main, right, and left hepatic artery patency was shown in all patients by Doppler ultrasound from 4 to 324 days (mean, 106 days) postembolization. Two patients also underwent hepatic angiography from 59 to 148 days (mean, 104 days) postembolization, which confirmed patency of the hepatic artery. In the postembolization period, three patients underwent PTC, but no cholangiographic evidence of bile duct ischemia was observed. No patients underwent retransplantation after embolization. Five patients still have their original transplanted livers from 7 to 67 months (mean, 40 months) after embolization. Two patients died from histologically proven rejection 87 and 93 days after embolization with normal hepatic arterial flow confirmed in both patients by sonography on days 74 and 75, respectively. The remaining patient sustained irreversible ischemic damage to his bile ducts as a result of a tight stenosis at the hepatic arterial anastomosis. Although the arterial narrowing was successfully treated by balloon angioplasty, his bile ducts remained narrowed. The damaged bile ducts were the reason for his PTBD, which then lead to his bleeding and subsequent embolization. He died 15 months after embolization without improvement of his bile duct ischemia; however, flow was normal in his hepatic arteries as shown by follow-up angiography on the 148th day and Doppler sonography on the 324th day. Discussion In the native liver, intrahepatic arterial bleeding caused by PTBD can be safely and effectively treated by subselective transcatheter emboli~ation.~,~ This is the treatment of choice, because it is more precise and less invasive than surgical ligation of the hepatic artery or one of its branches. Surgery may not control the bleeding and, in some cases, has lead to ischemia and liver fail~re.~ The diagnostic hepatic angiogram localizes the bleeding site precisely allowing selective embolization of the damaged vessel and thus minimizes liver ischemia. If the angiogram does not show the bleeding artery, the biliary catheter should be removed over a guide wire and the angiogram repeated to exclude the possibility that the catheter is tamponading the leak. 5,6 It is important to use a microcatheter to achieve precise subselective localization and embolization, minimize spasm, and avoid occlusion of noninvolved arterial branches. and gelfoam pledgets were chosen as embolic agents to avoid the pronounced peripheral embolization that results from the use of gelfoam powder or very small polyvinyl alcohol particles. The position of the coils is easily seen fluoroscopically, aiding precise positioning. Central nonselective embolization can threaten the viability of the liver by blocking intraorgan collateral blood flow.6 It was not used in any of our patients and should be avoided. On the basis of our previous nondocumented observations, we attempted, where possible, to straddle the arterial injury site by placing coils immediately proximal and immediately distal to the point of arterial injury, not only to prevent bleeding from antegrade flow in the hepatic branch artery but also to prevent possible back- bleeding from intrahepatic arterial collaterals. Embolization did not cause liver failure, nor was there any damage to the arterial anastomosis that was clinically evident. Objective follow-up by sonography in all patients and repeat angiography in two patients confirmed that blood flow in the major hepatic arteries was not compromised. In addition to damage to the hepatic arteries and subsequent ischemic damage to the liver parenchyma, we were concerned about possible ischemic

5 122 Croutch et al injury to the bile ducts. Ischemia is known to cause narrowing of the bile ducts. l2 However, rejection also causes narrowing of the bile ducts. The cholangiographic appearances associated with both these causes are identical. It is difficult to know whether embolization caused bile duct ischemia in any of our patients. The intrahepatic bile ducts have a rich arterial supply from numerous twigs arising from the intrahepatic arteries, and so this rich arterial supply should not be at great risk to coils or pledgets selectively positioned exclusively at the bleeding site. l3 Three of the five surviving patients had PTCs in the follow-up period after embolization and showed no cholangiographic evidence of ischemic duct narrowing. The other two survivors are well and have not undergone cholangiography. One patient already had ischemic ducts from a previous hepatic artery narrowing before embolization, and the final two patients died of rejection, so would likely have had abnormal ducts. Five of eight patients are alive with patent hepatic arteries. Two patients died at 87 and 93 days after embolization with biopsy-proven rejection and sonographic demonstration of normal artery blood flow. The final patient had ischemic narrowing of his ducts because of an anastomotic stricture of his hepatic artery. This was successfully treated by balloon angioplasty. Two months later, he underwent PTBD, resulting in arterial injury requiring embolization. He died 15 months postembolization with a patent HA shown using follow-up angiography on day 148 and reconfirmed using sonography on day 324. The ischemic changes observed at biopsy before his death are thus most likely to have arisen from his original arterial narrowing. Hemobilia is more frequently a result of injury to a hepatic or a portal vein branch rather than to an arterial branch. These venous leaks almost always stop over time and can usually be managed by replacing the biliary catheter with one of a larger diameter to tamponade the leak. Injection of contrast through the biliary catheter sometimes opacifies the leaking vessels, which helps in the identification of the bleeding vessel and differentiation between portal veins, hepatic veins, and hepatic arteries. From the time of onset of hemobilia to the time of embolization, there was an average delay of 4.8 days (range, 1 to 11 days). As a consequence, an average of 4 U of PRBC (range, 0 to 13 units) were required in the period before embolization. To reduce this delay, there should be a heightened awareness to the possibility of a treatable arterial injury. Arterial injury should be considered when hemobilia persists for more than 24 hours after PTBD, occurs during biliary catheter exchange, or develops unexpectedly in a patient with an established uncomplicated PTBD or when the hematocrit falls progressively without other explainable causes. Hemobilia usually ceases after successful embolization, which is confirmed using a completion angiogram. One patient required a repeat angiogram 1 day after the initial treatment because of recurrent bleeding. Deployment of additional coils and gelfoam pledgets directed at the same bleeding source was required to stop the continued bleeding. Abnormal bleeding parameters, including an elevated PT, PTT, or depressed platelet count, can impair the effectiveness of both coils and gelfoam in binding platelets into a matrix to stop bleeding. In this patient, the PT was 24.8 seconds, PTT was greater than 100 seconds, and the platelet count was 50,000 during the first embolization attempt. Transfusion of 2 U of fresh frozen plasma and a six-pack of platelets, before the second embolization, corrected the coagulation profile, and no further hemobilia occurred. It is suggested that if a coagulopathy exists, fresh frozen plasma and/or platelets should be used to maximize the benefit of transcatheter embolization. At our institution, the initial diagnostic percutaneous transhepatic cholangiogram is performed with a 22-gauge Chiba needle. Then, in the majority of subsequent PTBDs, a peripheral duct is punctured with a 19-gauge sheathed needle under fluoroscopic guidance at the same time as the ductal system is opacified by contrast injection through the retained 22-gauge Chiba. This method uses the easily steerable sheathed needle to insure an accurate peripheral duct entry, which aids smooth future access for potential stenting or balloon dilatation therapy. Analysis of the radiographs from the PTBD procedures in our eight patients with hemobilia showed a peripheral duct puncture in all cases. Central punctures can lead to an increased risk of hemobilia as a result of injuring larger branches of the HA or portal veins that lie concentrated in the hilum. Conclusion Our experience shows that transcatheter arterial embolization is a safe and effective method for stopping hemobilia caused by PTBD in the liver transplant patient.6 After identification of the arterial bleeding source using angiography, superselective

6 Arterial Embolimation for Hemobilia 123 arterial embolization will effectively stop hemobilia without compromising liver function or viability. References 1. Lake JR. Changing indications for liver transplantation. Gastroenterol Clin North Am 1993;22: Mueller PR, van Sonnenberg E, Ferrucci JT. Percutaneous biliary drainage: Technical and catheter-related problems in 200 procedures. AJR 1982;138: Gunther RW, Schild H, Thelen M. Percutaneous transhepatic biliary drainage: Experience with 31 1 procedures. Cardiovasc Intervent Radiol 1988;ll : Uflacker R, Mourao GS, Piske RL, Souza VC, Lima S. Hemobilia: Transcatheter occlusive therapy and longterm follow-up. Cardiovasc Intervent Radiol 1989; 12: Savader SJ, Trerotola SO, Merine DS, Venbrux AC, Osterman FA. Hemobilia after percutaneous biliary drainage: Treatment with transcatheter embolotherapy. J Vasc lnterv Radiol 1992;3: Zajko AB, Chablani V, Bron KM, Jungreis C. Hemobilia complicating transhepatic catheter drainage in liver transplant recipients: Management with selective embolization. Cardiovasc Intervent Radiol 1990;13: Doppman JL, Girton M, Vermess M. The risk of hepatic artery embolization in the presence of obstructive jaundice. Radiology 1982;143: Trojanowski J, Harrist TJ, Athanasoulis CA, Greenfield AJ. Hepatic and splenic infarctions: Complications of therapeutic transcatheter embolization. Am J Surg 1980; , 9. Ward EM, Kiely MJ, Maus TP, Wisner RH, Krom RA. Hilar biliary strictures after liver transplantation: Cholangiography and percutaneous treatment. Radiology 1990;177: Jindal RM, Train J, Meneses P, Emre S, Schwartz ME, Miller CM. Salvage of liver allograft after hepatic artery thrombosis and hilar biliary structure. Transplantation 1994;57: Charnsangavej C, Chuang VP, Wallace S, So0 CS, Bowers T. Angiographic classification of hepatic arterial collaterals. Radiology 1982;144: Terblanche J, Allison HF, Northover JMA. An ischemic basis for biliary strictures. Surgery 1983;94: Cho KJ, Lunderquist A. The peribiliary vascular plexus: The microvascular architecture of the bile duct in the rabbit and in clinical cases. Radiology :