Etiology and Management of Hepatic Artery Thrombosis After Adult Liver Transplantation

Transcription

1 LIVER TRANSPLANTATION 20: , 2014 ORIGINAL ARTICLE Etiology and Management of Hepatic Artery Thrombosis After Adult Liver Transplantation Moustafa Mabrouk Mourad, 1 Christos Liossis, 1 Bridget K. Gunson, 2 Hynek Mergental, 1 John Isaac, 1 Paolo Muiesan, 1 Darius F. Mirza, 1 M. Thamara P. R. Perera, 1 and Simon R. Bramhall 1 1 Liver Unit, Queen Elizabeth Hospital, Birmingham, United Kingdom; and 2 National Institute for Health Research Biomedical Research Unit and Centre for Liver Research, University of Birmingham, Birmingham, United Kingdom Hepatic artery thrombosis (HAT) represents a major cause of graft loss and mortality after liver transplantation. It occurs in up to 9% of adult recipients. The early diagnosis of HAT decreases septic complications, multiorgan failure, and graft loss, and there are better outcomes after treatment. In this study, we reviewed 102 episodes of HAT, which were classified as early hepatic artery thrombosis (E-HAT) when they were diagnosed within the first 21 days after transplantation. The overall incidence of HAT was 7%: 31 episodes (30.4%) were identified as E-HAT, and 71 episodes (69.6%) were identified as late hepatic artery thrombosis (L-HAT). Graft dysfunction was the commonest presentation (30 cases or 29%). Most E-HAT cases were managed with retransplantation (74%), whereas early revascularization was carried out for only 13% with a 75% success rate. The incidence of retransplantation for L-HAT was only 41%, whereas 32% were too ill for relisting and eventually died. Successful conservative management was noted for 13 of the 102 patients (13%) with collateralization and good hepatic perfusion, with biliary complications encountered in 7 cases (54%) subsequently. A multivariate analysis showed that previous episodes of HAT, the number of arterial anastomoses, and a low donor weight were independent risk factors for E-HAT, whereas a history of upper abdominal operations (non-hat), a previous history of HAT, a low donor weight, and a recipient age < 50 years were independent risk factors for L-HAT. The graft survival rates for HAT patients were 52%, 36.6%, and 27.4% at 1, 3, and 5 years, whereas the corresponding rates were 81.4%, 81.2%, and 76.4% for non-hat patients. In conclusion, prompt revascularization for E-HAT patients decreases the incidence of serious, irreversible septic complications and graft loss and improves overall outcomes. A significant number of L-HAT patients do not require further intervention despite the high incidence of ischemic cholangiopathy. Liver Transpl 20: , VC 2014 AASLD. Received August 12, 2013; accepted March 1, Liver transplantation (LT) is the ideal therapy for endstage liver disease. Vascular complications after LT are not uncommon despite the marked improvements and innovations in anastomotic vascular techniques; they frequently result in hepatic insufficiency and graft failure and thus the need for retransplantation. 1 Hepatic artery thrombosis (HAT) is the most serious vascular complication after LT with an overall incidence that varies from 2% to 9%. 2,3 The exact mechanism of HAT development remains unclear and is Abbreviations: CHA, common hepatic artery; CI, confidence interval; DCD, donation after circulatory death; E-HAT, early hepatic artery thrombosis; GDA, gastroduodenal artery; HA, hepatic artery; HAT, hepatic artery thrombosis; LHA, left hepatic artery; L-HAT, late hepatic artery thrombosis; LT, liver transplantation; OR, odds ratio; OLT, orthotopic liver transplantation; RHA, right hepatic artery; SA, splenic artery; US, ultrasonography. Simon R. Bramhall conceptualized the study. Moustafa Mabrouk Mourad, Bridget K. Gunson, and Chris Liosis were responsible for the data collection. Moustafa Mabrouk Mourad and Simon R. Bramhall contributed to the study design. Moustafa Mabrouk Mourad and Chris Liosis performed the statistical analysis. Moustafa Mabrouk Mourad and Chris Liosis prepared the manuscript. Hynek Mergental, John Isaac, M. Thamara P. R. Perera, Paolo Muiesan, Darius F. Mirza, and Simon R. Bramhall contributed to the data interpretation and the critical review of the manuscript. All the authors approved the final manuscript. The investigators did not receive any financial support for this study. None of the authors had any potential financial conflicts of interests relevant to this article. Address reprint requests to Simon R. Bramhall, M.D., F.R.C.S., Liver Unit, Queen Elizabeth Hospital, Nuffield House, 3rd Floor, Edgbaston, Birmingham B15 2TH, United Kingdom. Telephone: 144 (0) ; FAX: 1 44 (0) ; simon.bramhall@uhb.nhs.uk DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2014 American Association for the Study of Liver Diseases.

2 714 MOURAD ET AL. LIVER TRANSPLANTATION, June 2014 likely multifactorial. Many risk factors, both operative and nonoperative, have been implicated. 4 Close postoperative monitoring of these patients and clinical judgment are mandatory. Bile ducts are exclusively dependent on the blood supply from the hepatic artery (HA). HAT usually presents with biliary ischemic lesions such as necrosis with biliary leakage, ischemic stricture(s), cholangitis, and ultimately multiorgan failure and graft dysfunction. 5 Routine surveillance of HA anastomoses with Doppler ultrasonography (US) facilitates the early detection and treatment of these complications before irreversible graft dysfunction and allows therapeutic modalities other than retransplantation to be applied in this era of donor shortages. Various therapeutic options for managing HAT, including arterial reconstruction, surgical thrombectomy, and radiologically guided thrombolysis, are available, but retransplantation is still the goldstandard treatment for HAT. Retransplantation is restricted by both graft availability and the patient s general condition. We have retrospectively reviewed a large cohort of consecutive adult LT patients and carried out univariate and multivariate analyses of graft and surgical technique related factors predisposing patients to HAT, and we report here our protocol for managing early hepatic artery thrombosis (E-HAT) and late hepatic artery thrombosis (L-HAT). PATIENTS AND METHODS All adult LT procedures performed at the Liver Unit of Queen Elizabeth Hospital (Birmingham, United Kingdom) between January 2000 and December 2011 were reviewed. Retrospective data for all recipients were identified, and data gathered on technical aspects related to the arterial reconstructive techniques, donor and graft characteristics, and intraoperative details such as the cold ischemia time, transfusion requirements, and duration of surgery were recorded. Transplant Procedure The selection criteria for liver grafts and donors were based on the indication, blood group compatibility, recipient/graft size matching, clinical circumstances, urgency, and hemodynamic parameters. The transplant procedure was standard and involved a modified piggyback technique. Flushing of the graft was routinely performed after the completion of caval and portal vein anastomoses with 500 to 1000 ml of 5% dextrose. The type of arterial reconstruction depended on the graft s vascular anatomy as well as the arterial inflow/quality of the recipient s arteries. Several arterial reconstruction techniques, depending on the aforementioned factors, were recorded. The arterial anastomoses were fashioned with continuous 6-0 or 7-0 polypropylene sutures (Ethicon, Norderstedt, Germany) at different sites between the graft and donor arterial systems. An anastomosis was most commonly performed between the celiac axis stem or the common hepatic artery (CHA) stem of the graft and the HA of the recipient, usually at the level of the artery bifurcation or at the origin of the gastroduodenal artery (GDA). An interposition conduit using an iliac artery between the graft HA and the infrarenal part of the abdominal aorta was indicated when the diameter of the recipient HA was small and/or there was stenosis of the celiac trunk and for some HATrelated retransplants. Reconstructions of any accessory HAs were performed with end-to-end anastomoses to the donor splenic artery (SA) or GDA branch, and in a few cases, a split vascular graft (the lower part of the aorta connected to the common iliac arteries) was used for reconstruction. Multiple arterial anastomoses were defined as more than 1 arterial anastomosis to establish the continuity between the graft and the recipient hepatic vasculature or as any additional anastomosis to reconstruct any donor HA. Biliary Reconstruction Biliary reconstruction was accomplished with a ductto-duct anastomosis or Roux-en-Y hepaticojejunostomy. A T-tube for free drainage was used for duct-toduct anastomosis patients who received split liver grafts (right lobe grafts). Roux-en-Y hepaticojejunostomy was indicated if primary sclerosing cholangitis was the indication for transplantation or if the bile ducts had small diameters. HAT HAT was defined as E-HAT when it was diagnosed within the first 21 days after transplantation. This definition was based on a recommendation from the UK Liver Advisory, and it allowed these patients to be relisted for superurgent retransplantation. Posttransplant hemoglobin levels were maintained above 8 g/dl and preferably below 10 g/dl. Similarly, platelet counts were maintained at approximately 100,000/lL to reduce the risk of HAT. Before 2004, routine prophylaxis for the prevention of HAT was not standard practice at our unit, and an anticoagulant was prescribed only for high-risk recipients, such as those with multiple vascular anastomoses, back-table reconstruction, interposition grafts, technically demanding HA anastomoses or reconstruction, or previous HAT. Since that time, our practice has been to initiate prophylaxis on day 1 after transplantation with aspirin and enoxaparin for all patients. Specific anticoagulants (eg, warfarin) were prescribed only in addition to previously mentioned drugs for the highrisk recipients listed previously. Anticoagulation was not involved in this study because of the limited data. Doppler ultrasound scans were not routine for all recipients after LT. They were performed only for recipients who had abnormal liver function, delayed graft function, acute graft failure, or suspicious arterial or biliary compromise (suspected vascular or/and biliary complications) so that the HA patency could be

3 LIVER TRANSPLANTATION, Vol. 20, No. 6, 2014 MOURAD ET AL. 715 evaluated as a first diagnostic step. In cases without any Doppler arterial flow, the HA patency was confirmed by computed tomography angiography. The clinical presentation, radiological diagnostic methods, treatment, and outcomes of HAT were recorded. Demographic data, indications for LT, preoperative comorbidities, donor and graft characteristics, and intraoperative details such as the cold ischemia time, transfusion requirements (red blood cells, platelets, and fresh frozen plasma), duration and starting time of LT, and type of arterial reconstruction were compared between patients with HAT and patients without HAT. The management of HAT at our unit was determined at a multidisciplinary meeting, which included surgeons, hepatologists, anesthetists, radiologists, and pathologists. The management was subject to many factors: type of HAT (E-HAT versus L-HAT), time of presentation, clinical presentation, serum transaminitis, and hepatic parenchymal effects (ischemic changes) on contrast computed tomography. In the statistical analysis of risk factors for E-HAT and L-HAT, we considered the recipient s underlying liver disease, age, sex, cytomegalovirus status, and medical comorbidities; previous LT and the number of previous grafts; the donor s age, sex, and cause of death; the type of preservation solution; the cold ischemia time; the back-table donor artery reconstruction; the type and number of arterial anastomoses; the intraoperative transfusion requirements (red blood cells, platelets, and fresh frozen plasma); the use of aprotinin; the operative time; and the starting time of transplantation. Immunosuppression was based on corticosteroids, tacrolimus, mycophenolate mofetil, and azathioprine. 6 Statistical Analysis Statistical Program for Social Statistics 21.0 (IBM, United States) was used. A univariate analysis was performed for categorical variables with the use of Pearson s chi-square test and Fisher s exact test. Continuous variables were analyzed with the Mann- Whitney U test when we were comparing 2 groups and with the Kruskal-Wallis test when we were comparing 3 groups. When 3-group comparisons were significant, post hoc tests were performed to compare E-HAT and L-HAT patients with the non-hat cohort with a Bonferroni-corrected critical P value of For all other analyses, P < 0.05 was considered to be significant. Factors found to be significant in the univariate analysis were considered for inclusion in a backwards stepwise multivariate binary logistic regression analysis in order to identify independent predictors of HAT. In order to account for the different lengths of followup in the cohort and confirm the risk factors for HAT, the analysis was also repeated with Cox regression models. Survival curves were estimated with the Kaplan-Meier method and were compared with logrank tests. RESULTS Adult LT was performed 1507 times during the study period, but 20 of these transplants were excluded: 19 recipients died during either the operation or within the first 24 hours after transplantation, and 1 recipient received a living related transplant. One hundred sixtytwo of the 1487 orthotopic liver transplantation (OLT) cases included in this study (11%) were split LT, and 117 grafts (8%) were retrieved from donation after circulatory death (DCD) donors. Retransplantation in our series was carried out for 119 (8%) of all patients, and duct-to-duct biliary anastomosis was performed for 1210 transplants (81%). All allografts were preserved with University of Wisconsin solution. The piggyback technique was used for 1213 recipients (82%). The median age of the recipients was 53 years (range years); 836 recipients (56%) were male. The median body mass index was 26 kg/m 2 (range kg/m 2 ). The median follow-up was 57 months (range months). HAT One hundred two episodes of HAT were documented for an overall incidence of 7%. The diagnosis of HAT was made at a mean of 81 days (range days). E-HAT and L-HAT were reported for 31 (30.4% of all HAT cases and 2% overall) and 71 transplants (69.6% of all HAT cases and 4.8% overall), respectively. L-HAT developed on top of HA stenosis in 2 recipients, even after dilatation and vascular stenting after 67 and 156 days. In a patient with E-HAT (postoperative day 7), thrombosis of the right hepatic artery (RHA) was observed with a patent left hepatic artery (LHA), and the patient was managed conservatively. The indications for transplantation in the study were equally distributed (P ) between patients with HAT and patients without HAT (Table 1). Vascular Reconstruction An end-to-end anastomosis between the celiac axis stem or CHA stem of the graft and the recipient HA was performed in 807 transplants (54%; Table 2). Forty-eight episodes of HAT occurred in this group (6% versus 8% for other techniques). The graft RHA (in right lobe grafts) was anastomosed to the recipient HA in 72 transplants (5%), and 5 episodes of HAT (7%) were diagnosed in that group. Among the 459 transplants in which the donor HA with a splenic patch was used, 34 HAT events were observed (7.4% versus 6.6% for other techniques). HAT was diagnosed in 5 of 96 transplants (5.2%) in which a patch at the division of the donor GDA was used. The donor HA with an aortic patch was anastomosed to the recipient HA in 23 transplants, and HAT occurred in 3 recipients (13%). Various other arterial anastomoses were fashioned for 30 transplants (2%), and a diagnosis of HAT occurred for 1 recipient (3.3%). An infrarenal aortic conduit was fashioned for 63 transplants (4% overall). It was used primarily in 45 recipients (45/

4 716 MOURAD ET AL. LIVER TRANSPLANTATION, June 2014 TABLE 1. Transplant Indications for the HAT and Non- HAT s TABLE 2. Different Arterial Reconstructive Anastomotic Techniques Used in This Study Indication (n 5 102) (n ) Alcoholic cirrhosis 17 (17) 217 (16) Hepatitis C cirrhosis 3 (3) 121 (9) Primary sclerosing 11 (11) 122 (9) cholangitis Primary biliary cirrhosis 14 (14) 203 (15) Retransplantation 13 (13) 92 (7) Acute liver failure 9 (9) 176 (13) Hepatitis B cirrhosis 3 (3) 20 (1) Hepatocellular carcinoma 1 14 (14) 169 (12) viral hepatitis B/C Other indications 18 (18) 265 (19) 1368 or 3% of primary transplants). There were 105 retransplants (8%) in this series, and aortic conduits were used in 30 retransplantations cases (18/119 or 15%). HAT was diagnosed in 6/30 transplants in this group (9% versus 7% of recipients without a conduit). No particular arterial reconstruction technique showed a statistically higher incidence of HAT in comparison with the other techniques combined. Multiple anastomoses were performed for 384 recipients (26%). Sixty-three aortic conduits, 31 interposition arterial grafts between the donor and recipient HAs (right lobe grafts), and 310 reconstructions were performed for accessory arteries. Two-hundred eightythree of the 310 reconstructions (91%) were anastomosed to the donor SA; 10 (4.5%) were anastomosed to the donor GDA; and in 10 accessory arteries (4.5%), a split interposition graft was used for reconstruction (Tables 2 and 3). Some recipients underwent more than 1 of the aforementioned reconstructions. Three anastomoses were performed for 34 of the 384 recipients. Thirty-nine of the HAT episodes (38%) were observed in this group (15 E-HAT patients and 24 L- HAT patients), as shown in Table 3. Clinical Presentation HAT Non-HAT NOTE: The data are presented as numbers and percentages. The clinical presentation of HAT varied between E-HAT and L-HAT (Table 4). Graft dysfunction was the commonest presentation for E-HAT (19/31 or 61%) and was followed by fulminant hepatic necrosis (5/31 or 16%). L-HAT was associated with a milder clinical course and a more insidious onset. L-HAT presented mainly with ischemic biliary lesions (cholangitic attacks and/or biliary leak; 16/71 or 22%), fulminant hepatic necrosis (13/71 or 18%), and abnormal serum transaminases (12/71 or 17%). The relationship between the time of L-HAT and the type of clinical presentation is illustrated in Fig. 1. OLT Cases Arterial Reconstructive Technique (n ) Recipient anastomotic site GDA bifurcation 703 (47) CHA bifurcation 574 (39) Main HA trunk 51 (3) Aorta 64 (4) Patch at origin of SA* 20 (1) Other 75 (5) Donor anastomotic site CHA stem 471 (32) Patch at origin of SA 459 (31) Celiac axis stem 336 (23) Patch at origin of GDA 96 (6) RHA (split transplant) 72 (5) Aortic patch 23 (2) Other 30 (2) Aortic conduit 63 (4) Interposition vascular graft 31 (2) Accessory HAs (multiple arteries) Site 317 (21) Accessory LHA 164 (11) Accessory RHA 143 (10) Accessory RHA and LHA 10 (1%) Reconstruction (overall) 310 (21%) Ligated (small) 7 (1%) Anastomosed to donor SA 283 (19) Anastomosed to donor GDA 10 (1%) Reconstruction using split 10 (1%) interposition graft (iliac arteries) NOTE: The data are presented as numbers and percentages. *Patch at the origin of the SA from the celiac trunk. Patch at the origin of the GDA from the CHA. Risk Factors and Management of HAT E-HAT There were 31 E-HAT recipients. First, 4 of these recipients were explored early within the first 24 hours: 3 patients underwent successful arterial reconstruction, and an aortic conduit was employed in 1 patient. In the fourth recipient, thrombectomy was performed, but unfortunately, the patient died after 3 days from multiorgan failure. These 4 patients were considered to be at high risk for HAT after the original transplant because they had multiple arterial anastomoses, and according to our protocol, they underwent Doppler US after the operation. This showed HAT, and the decision was re-exploration (no hepatic parenchymal effect). Second, superurgent retransplantation was performed for 23 of the 31 recipients (74%), and this succeeded for 18 of the 23 patients (78%). The main indication for retransplantation in E- HAT patients was graft dysfunction (19 recipients or 83%), and this was followed by fulminant hepatic necrosis (4 recipients or 17%). Third, 1 patient with

5 LIVER TRANSPLANTATION, Vol. 20, No. 6, 2014 MOURAD ET AL. 717 TABLE 3. Incidence of E-HAT and L-HAT With Respect to the Number of Arterial Anastomoses Non-HAT E-HAT L-HAT Multiple Anastomoses (n ) (n 5 31) (n 5 71) 1 anastomosis (n or 74%) 1040 (75) 16 (52) 47 (66) 2 anastomoses (n or 24%) 316 (23) 13 (42) 21 (30) 3 anastomoses (n 5 34 or 2%) 29 (2) 2 (6) 3 (4) NOTE: The data are presented as numbers and percentages. TABLE 4. thrombosis of the RHA and abnormal serum transaminases presented with a patent LHA, well-developed collaterals, and good hepatic perfusion on computed tomography, and the patient was managed conservatively. Fourth, 3 patients were very sick and unfit for retransplantation, and they died eventually from multiorgan failure. A univariate analysis revealed that previous HAT (P ), the number of arterial anastomoses (0.019), and a low donor weight (P ) were significantly associated with E-HAT (Tables 5 and 6). According to a logistic regression, previous HAT [odds ratio (OR) 5 276, 95% confidence interval (CI) , P ], the number of arterial anastomoses (OR , 95% CI , P ), and a low donor weight (OR , 95% CI , P ) were independent predictors of E-HAT (Table 7). L-HAT Clinical Presentation of E-HAT and L-HAT E-HAT L-HAT Clinical Presentation (n 5 31) (n 5 71) Graft dysfunction 19 (61) 11 (15) Ischemic biliary lesions 2 (6) 16 (23) Fulminant hepatic necrosis 5 (16) 13 (18) Bile stasis 2 (6) 7 (10) Abnormal serum 3 (10) 12 (17) transaminases Liver abscess 0 9 (13) Incidentally discovered 0 3 (4) in imaging NOTE: The data are presented as numbers and percentages. There were 71 L-HAT recipients. First, retransplantation was performed for 29 of the 71 recipients (41%). Successful LT was noted for 22 of the 29 recipients (76%). Forty-one percent (12/29) of these L-HAT recipients underwent retransplantation with ischemic cholangiopathy as an indication. Second, 7 patients were relisted for transplantation, but 1 patient was removed from the list because of new-onset renal cell carcinoma found during the transplant assessment. Third, collateralization with a satisfactorily perfused liver was observed in 12 of the 71 patients (17%). This was discovered incidentally for 3 patients on routine imaging, and the rest of the patients presented only with mildly deranged serum transaminases; these patients were managed conservatively without any need for further intervention. Fourth, 23 patients (32%) were high-risk cases for retransplantation, and death eventually occurred within 1 month. In 7 of the 13 patients (54%) who had collateralization (1 E-HAT patient and 12 L-HAT patients), ischemic biliary complications were diagnosed. Six patients had ischemic biliary strictures 2.5 to 94 months after the diagnosis of HAT, and only 1 patient had biliary leakage. A recipient age < 50 years (P ), previous HAT (P ), previous upper abdominal operations (non-hat; P ), a low donor weight (P ), and a longer operative time for transplantation (P ) were associated with L-HAT in a univariate analysis (Tables 5 and 6). Moreover, previous upper abdominal operations (non-hat; OR , 95% CI , P ), previous HAT (OR 5 124, 95% CI , P ), a recipient age < 50 years (OR , 95% CI , P ), and a small donor size (a low weight; OR , 95% CI , P ) proved to be independent predictors of L-HAT according to a logistic regression (Table 7). Survival Rates The overall mortality rate for HAT was 45% (46/102), with 31 deaths (30.4%) related directly to HAT. The median overall patient survival times for non-hat and HAT transplants were 120 and 81 months, respectively (P ; Fig. 2). The graft survival rates were 81.4% and 52% at 1 year and 81.2% and 36.6% at 3 years for non-hat and HAT transplants, respectively. The 5-year graft survival rates for non-hat and HAT transplants were 76.4% and 27.4%, respectively (P ; Fig. 3). DISCUSSION Graft function after LT is dependent on sufficient arterial perfusion. In contrast, the native liver has the ability to withstand even a major obstruction to its

6 718 MOURAD ET AL. LIVER TRANSPLANTATION, June 2014 Figure 1. Dot graph showing the relationship between the time of L-HAT and the type of clinical presentation. arterial hepatic blood flow through collaterals arising from the intrahepatic LHA or RHA, superior mesenteric artery, SA, inferior phrenic artery, left gastric artery, and arteries of the omentum and duodenum. 7,8 Consequently, HAT after LT is devastating and requires urgent revascularization. It is associated with increased morbidity and graft loss and more than 50% mortality with a 75% incidence of retransplantation. 9 A retrospective analysis of 4000 consecutive transplants from the University of Pittsburgh showed that HAT was the indication for nearly onethird of the retransplants performed. 10 Although the exact mechanism of HAT was not always certain, technical failure was considered the principal cause of HAT in that study. The incidence of L-HAT (0.8%) was reported to be less than that of E-HAT (3.4%). 11 However, in our series, the occurrence of L-HAT (4.8%) was twice that of E-HAT (2%) with an overall incidence of 7% for all HAT patients. Although many risk factors, surgical 4 and nonsurgical, 12 have been implicated in the development of HAT, the exact pathogenesis of this complication is still unclear. The timing and clinical manifestations of E-HAT and L-HAT were different, and the predisposing factors were also different. Arterial reconstruction with more than 1 anastomosis (extra anastomoses) has been described as a risk factor for E-HAT. 13,14 Numerous anastomoses imply a higher number of surgical sutures with an abnormal morphology of the artery lumen and a greater risk of endoluminal surface damage and subsequent platelet adhesion and thrombosis. 13 The results of our study have confirmed the higher risk of E-HAT with multiple anastomoses. Each extra anastomosis was associated with an approximately 2-fold increased risk of E-HAT. The other findings from the logistic regression concerning previous HAT and a small donor weight appear to be logical and in agreement with the literature We advise routine daily monitoring of the arterial flow by Doppler US in the early postoperative period in the presence of 1 or both of these risk factors. Recipients who underwent transplantation with a liver from a small donor more frequently suffered E- HAT, and this complication could potentially be prevented through the careful selection of recipients, with a recipient/donor body weight ratio less than 1.25 appearing to be desirable. 12 Secondary or tertiary transplants have been reported to be associated with a high incidence of conduit use for arterial reconstruction because of an unsuitable recipient HA, poor inflow, or intimal dissection. 15 Oh et al. 12 favored the use of aortic conduits in patients with a history of HAT and the replacement of the old conduit with a new one for retransplantation to decrease the incidence of HAT. In this report, a previous history of HAT and a low donor weight were independent risk factors for L- HAT. Miyagi et al. 16 showed that a young recipient agewasariskfactorforhatinpatientswhosearterial anastomosis was performed with a microsurgical

7 LIVER TRANSPLANTATION, Vol. 20, No. 6, 2014 MOURAD ET AL. 719 TABLE 5. Donor and Recipient Characteristics: Comparison of the E-HAT, L-HAT, and Non-HAT s P Value* Non-HAT (n ) E-HAT (n 5 31) L-HAT (n 5 71) Value P Value Value P Value Recipients Age (years) (17-74) 52 (18-71) (18-68) Female sex [n (%)] (43) 20 (65) 33 (46) Body mass index (kg/m 2 ) (15-59) 24 (17-34) 25 (17-37) Body mass index > 30 kg/m 2 [n (%)] (25) 8 (26) 14 (20) Medical comorbidities Diabetes mellitus [n (%)] (20) 5 (16) 11 (15) Cardiac diseases [n (%)] (4) 3 (10) 5 (7) Renal impairment [n (%)] (7) 3 (10) 8 (11) Smoking [n (%)] (36) 7 (23) 25 (35) Cytomegalovirus status [n (%)] (60) 18 (58) 41 (58) Previous abdominal operation [n (%)] (9) 5 (16) (25) Retransplantation as indication [n (%)] (7) 3 (10) (14) >2 liver regrafts [n (%)] (0.9) 1 (3) 1 (1) Previous HAT [n (%)] (0.07) 4 (13) (8) Pretransplant sepsis [n (%)] (8) 1 (3) 8 (11) Pretransplant international (0.7-5) 1.2 (1-4) (1-6) normalized ratio Pretransplant hemoglobin (5-19) 11.2 (7-15) 11 (5-16) Pretransplant platelets: (11-942) 121 (37-557) (6-371) Wait-list time (months) (0-52) 2 (0-19) 1.6 (0-16) Intensive therapy unit stay (days) (0-150) 3 (1-26) 3 (1-97) Hospital stay (days) (1-190) 12 (2-218) 13 (2-337) Acute rejection [n (%)] (38) 12 (39) 31 (44) Median follow-up (months) Donors Age (years) (5-81) 49 (7-67) 47 (11-72) Age > 60 years old [n (%)] (17) 9 (29) 10 (14) Male sex [n (%)] (54) 15 (48) 36 (51) Median body weight (kg) Cytomegalovirus status [n (%)] (48) 16 (52) 35 (49) Cause of death: intracranial (84) 26 (84) 61 (86) versus other [n (%)] DCD [n (%)] (8) 4 (13) 9 (13) DCD donor > 40 years old [n (%)] (6) 2 (6) 6 (8) Domino donor [n (%)] (0.2) 0 1 (1.4) *Three-group comparison (Kruskal-Wallis/Pearson s chi-square test). Post hoc test versus the non-hat group (Mann-Whitney/Pearson s chi-square test) with a Bonferroni-corrected critical P value of The data are presented as medians and ranges. approach (P ). This study also found a trend toward an increased incidence of L-HAT in younger recipients (specifically those < 50 years old). A previous history of upper abdominal operations without a history of HAT was an independent prognostic determinant of L-HAT in our series; it has not been mentioned by others, and this is the first time that this evidence has been supported by a statistical analysis. We could not adjust for this correlation in the current study. The immediate initiation of aspirin therapy in recipients with a low hematocrit after LT may reduce the rate of HAT. 17 Since 2004, we have put all patients on aspirin on the first day after transplantation. In this report, the effect of antiplatelet prophylaxis on the incidence of HAT could not be assessed because many of the HAT episodes observed in our series developed before the prophylaxis began. However, no significant difference in the incidence of HAT was noted between the periods before and after 2004 in our study (7.6% and 6.5%, respectively). The use of microvascular surgical techniques and the regular use of Doppler ultrasound screening could be considered for the possible prevention or early detection of HAT after LT. 18 Unfortunately, routine postoperative Doppler monitoring is not feasible in all transplantation units, and long-term screening cannot happen for all transplant recipients. Knowledge of the risk

8 720 MOURAD ET AL. LIVER TRANSPLANTATION, June 2014 TABLE 6. Operative Characteristics: Comparison of the E-HAT, L-HAT, and Non-HAT s P Value* Non-HAT (n ) E-HAT (n 5 31) L-HAT (n 5 71) Value P Value Value P Value Urgent transplant [n (%)] (12) 3 (10) 6 (8) Transplant started after 18:00 [n (%)] (16) 4 (13) 13 (18) Split graft [n (%)] (11) 4 (13) 10 (14) Cold ischemia time (minutes) ( ) 528 ( ) 565 ( ) Warm ischemia time (minutes) (20-85) 41 (30-63) 41 (26-67) Intraoperative aprotinin use [n (%)] (37) 10 (32) 31 (44) Total blood products (U) (0-208) 24 (0-98) 22 (0-129) Red blood cells (U) (0-200) 4 (0-32) 4 (0-32) Fresh frozen plasma (U) (0-92) 10 (0-32) 8 (0-40) Platelets (U) (0-55) 10 (0-20) 10 (0-40) Operative time (minutes) ( ) 343 ( ) ( ) Posttransplant renal impairment [n (%)] (28) 10 (32) 16 (23) Biliary reconstruction Roux-en-Y hepaticojejunostomy [n (%)] (18) 9 (29) (28) Arterial reconstruction Aortic conduit [n (%)] (4) 2 (6) 4 (6) Interposition arterial graft [n (%)] (2) 2 (6) 1 (1.4) Number of arterial anastomoses Multiple HAs [n (%)] (21) 11 (35) (30) *Three-group comparison (Kruskal-Wallis/Pearson s chi-square test). Post hoc test versus the non-hat group (Mann-Whitney/Pearson s chi-square test) with a Bonferroni-corrected critical P value of The data are presented as medians and ranges. The data are presented as means and standard deviations. TABLE 7. Multivariate Analyses of Variables Predicting the Development of E-HAT and L-HAT Variable OR 95% CI P Value E-HAT Previous HAT Number of arterial anastomoses Donor weight (kg) L-HAT Recipient age (years) Previous HAT Previous upper abdominal operations (non-hat) Donor weight (kg) Transplant operative time (minutes) NOTE: A multivariate analysis was performed with a binary logistic regression via a backward stepwise entry method. All factors found to be significant (at P < 0.025) in the post hoc univariate tests were considered for inclusion in the model. factors for HAT is important in tailoring the screening protocol to the patients at risk. The clinical presentation, recipient condition, graft changes, and comorbidities determine the therapeutic modalities for HAT. E-HAT usually has a severely progressing clinical course with florid symptomatology. At our unit, the management of E-HAT depends on hepatic ischemic changes detected on scans and the degree of deranged liver transaminases. Early cases with an unaffected hepatic parenchyma and normal transaminases were considered for urgent relaparotomy and revascularization, and there were excellent outcomes for both adults and children 5-18 (4 cases in our series with a success rate of 75%). De Gaetano et al., 19 Langnas et al., 20 and Pinna et al. 21 have reported that prompt revascularization (either surgical or radiological) decreases the need for retransplantation in E-HAT cases. This requires routine Doppler US after transplantation, which decreases the high HATrelated mortality rate 5 ; this is not used routinely at

9 LIVER TRANSPLANTATION, Vol. 20, No. 6, 2014 MOURAD ET AL. 721 Figure 2. Overall survival of HAT and non-hat groups. Figure 3. Graft survival of HAT and non-hat groups. our unit and is requested for symptomatic or highrisk recipients. Among patients with ischemic changes and altered liver function, superurgent retransplantation was the treatment for fit recipients because revascularization was likely to produce a significant reperfusion injury that could compromise the grafts

10 722 MOURAD ET AL. LIVER TRANSPLANTATION, June 2014 and jeopardize the patients. 18 This policy succeeded in 78% of the cases (3 successful urgent revascularizations and 18 successful retransplants among 27 recipients with surgical interventions). However, 3 patients were unfit for any surgical intervention and eventually died. A late diagnosis of E-HAT should be managed conservatively with retransplantation for recipients developing ischemic complications and severely distorted hepatic function, but retransplantation for all cases is restricted by the shortage of donors. 18 The management of L-HAT is conservative for cases of collateralization with good hepatic perfusion, and rescue retransplantation is used for patients with ischemic changes and/or complications. In our study, 13 HAT patients had good collateralization with sufficient hepatic perfusion, and they were managed conservatively. However, L-HAT usually presents with equivocal symptomatology such as raised transaminases, ischemic biliary stricture with cholestasis, and liver abscesses. This ambiguity raises the chance of septic complications before the establishment of a HAT diagnosis, which can become critically delayed. Conservative treatment using antibiotics or radiological interventions such as endoscopic retrograde cholangiopancreatography and abscess drainage can be employed in such cases, with the chance for retransplantation for recipients showing an improvement in their general condition Margarit et al. 23 showed that donor age was the only significant predisposing factor for the development of biliary complications secondary to L-HAT in adult LT patients. In our series, 13 HAT patients (12 L-HAT patients and 1 E-HAT patient) were successfully managed conservatively without the need for any intervention. These cases had good collaterals with sufficient hepatic perfusion. Ischemic biliary complications were noticed in 7 of the 13 patients afterward: they had a higher median donor age (48 years, range years) in comparison with the patients without complications (40 years, range years, P ). Our experience, in conjunction with the shortage of donor livers and the high morbidity and mortality rates related to retransplantation, suggests that prompt arterial revascularization to avoid graft loss should be considered. Retransplantation is the goldstandard treatment for HAT, but it has a high mortality rate, and the selection of L-HAT cases that can be managed conservatively with acceptable long-term outcomes is possible despite the high incidence of ischemic cholangiopathy. Further research is required to understand and reduce the incidence of HAT and subsequent biliary complications following the conservative management of these patients. ACKNOWLEDGMENT The authors give many thanks to James Hodson, the medical statistician at University Hospitals Birmingham who reviewed the statistical analysis of this article. REFERENCES 1. Steinbr uck K, Enne M, Fernandes R, Martinho JM, Balbi E, Agoglia L, et al. Vascular complications after living donor liver transplantation: a Brazilian, single-center experience. Transplant Proc 2011;43: Proposito D, Loinaz Segurola C, Garcia Garcıa I, Jimènez C, Gonzalez Pinto I, Gomez Sanz R, et al. Assessment of risk factors in the incidence of hepatic artery thrombosis in a consecutive series of 687 liver transplantations [in Italian]. Ann Ital Chir 2001;72: Abou Ella KA, Al Sebayel MI, Ramirez CB, Rabea HM. Hepatic artery thrombosis after orthotopic liver transplantation. Saudi Med J 2001;22: Pastacaldi S, Teixeira R, Montalto P, Rolles K, Burroughs AK. Hepatic artery thrombosis after orthotopic liver transplantation: a review of nonsurgical causes. Liver Transpl 2001;7: Bekker J, Ploem S, de Jong KP. Early hepatic artery thrombosis after liver transplantation: a systematic review of the incidence, outcome and risk factors. Am J Transplant 2009;9: Shaw BW Jr, Stratta RJ, Donovan JP, Langnas AN, Wood RP, Markin RJ. Postoperative care after liver transplantation. Semin Liver Dis 1989;9: Wozney P, Zajko AB, Bron KM, Point S, Starzl TE. Vascular complications after liver transplantation: a 5-year experience. AJR Am J Roentgenol 1986;147: Wolf R, Porte RJ, van der Vliet TM, Kok T. Development of intrahepatic arterial shunts in a transplanted liver: a potential pitfall for Doppler sonography. J Clin Ultrasound 2001;29: Stange BJ, Glanemann M, Nuessler NC, Settmacher U, Steinm uller T, Neuhaus P. Hepatic artery thrombosis after adult liver transplantation. Liver Transpl 2003;9: Jain A, Reyes J, Kashyap R, Dodson SF, Demetris AJ, Ruppert K, et al. Long-term survival after liver transplantation in 4,000 consecutive patients at a single center. Ann Surg 2000;232: van Dam-Mieras MC, Muller AD, van Hinsbergh VW, Mullers WJ, Bomans PH, Bruggeman CA. The procoagulant response of cytomegalovirus infected endothelial cells. Thromb Haemost 1992;68: Oh CK, Pelletier SJ, Sawyer RG, Dacus AR, McCullough CS, Pruett TL, Sanfey HA. Uni- and multi-variate analysis of risk factors for early and late hepatic artery thrombosis after liver transplantation. Transplantation 2001; 71: Piscaglia F, Vivarelli M, La Barba G, Morselli-Labate AM, Taddei S, Cucchetti A, et al. Analysis of risk factors for early hepatic artery thrombosis after liver transplantation. Possible contribution of reperfusion in the early morning. Dig Liver Dis 2007;39: Soliman T, Bodingbauer M, Langer F, Berlakovich GA, Wamser P, Rockenschaub S, et al. The role of complex hepatic artery reconstruction in orthotopic liver transplantation. Liver Transpl 2003;9: Todo S, Makowka L, Tzakis AG, Marsh JW Jr, Karrer FM, Armany M, et al. Hepatic artery in liver transplantation. Transplant Proc 1987;19(pt 3): Miyagi S, Kawagishi N, Nakanishi W, Fujio A, Miyazawa K, Maida K, et al. Risk factors for hepatic artery thrombosis after microsurgical vascular reconstruction in liver transplantation. Transplant Proc 2013;45: Shay R, Taber D, Pilch N, Meadows H, Tischer S, McGillicuddy J, et al. Early aspirin therapy may reduce

11 LIVER TRANSPLANTATION, Vol. 20, No. 6, 2014 MOURAD ET AL. 723 hepatic artery thrombosis in liver transplantation. Transplant Proc 2013;45: Heaton ND. Hepatic artery thrombosis: conservative management or retransplantation? Liver Transpl 2013; 19(suppl 2):S14-S De Gaetano AM, Cotroneo AR, Maresca G, Di Stasi C, Evangelisti R, Gui B, Agnes S. Color Doppler sonography in the diagnosis and monitoring of arterial complications after liver transplantation. J Clin Ultrasound 2000;28: Langnas AN, Marujo W, Stratta RJ, Wood RP, Shaw BW Jr. Vascular complications after orthotopic liver transplantation. Am J Surg 1991;161: Pinna AD, Smith CV, Furukawa H, Starzl TE, Fung JJ. Urgent revascularization of liver allografts after early hepatic artery thrombosis. Transplantation 1996;62: Valente JF, Alonso MH, Weber FL, Hanto DW. Late hepatic artery thrombosis in liver allograft recipients is associated with intrahepatic biliary necrosis. Transplantation 1996;61: Margarit C, Hidalgo E, Lazaro JL, Murio E, Charco R, Balsells J. Biliary complications secondary to late hepatic artery thrombosis in adult liver transplant patients. Transpl Int 1998;11(suppl 1):S251-S Bhattacharjya S, Gunson BK, Mirza DF, Mayer DA, Buckels JA, McMaster P, Neuberger JM. Delayed hepatic artery thrombosis in adult orthotopic liver transplantation a 12-year experience. Transplantation 2001;71: