Liver Transplantation in Patients With PortalVein Thrombosis

Gerardo Manzanet, Fernando Sanjuan, Paco Orbis, Rafael Lopez, Angel Moya,Manuel Juan, Juanjo Vila, Josep Asensi, Pedro Sendra, Jose Ruız,

Martın Prieto, and Jose Mir

The aim of this study is to analyze the incidence, riskfactors, management, and follow-up of patients with por-tal vein thrombosis (PVT) undergoing primary ortho-topic liver transplantation (OLT). Four hundred fifteenOLTs were performed in 391 patients. In 62 patients,partial (group 1; n 5 48) or complete (group 2; n 5 14)PVT was found at the time of surgery. Portal flow wasreestablished by venous thrombectomy. In this study, wecompare 62 primary OLTs performed in patients withPVT at the time of OLT with a group of 329 primaryOLTs performed in patients without PVT (group 3) andanalyze the incidence of PVT, use of diagnostic methods,surgical management, and outcome. We found no signif-icant differences among the 3 groups for length of surgery,cold and warm ischemic times, and postoperative stay inthe intensive care unit. With the piggyback technique,groups 1 and 2 had greater blood losses and required moreblood transfusions than group 3. The early reoperationrate was greater in group 2. The incidence of rethrombosiswas 4.8% (group 1, 2%; group 2, 14.3%). Reexplorationand thrombectomy (2 patients) and retransplantation(1 patient) had a 100% mortality rate. In particular, themortality rate of patients with complete PVT with exten-sion into the splanchnic veins is high (33%). Three-monthand 4-year patient survival rates were statistically similarin the 3 groups. The presence of PVT at the time of OLTis not a contraindication for OLT. However, if PVT ex-tends into the splanchnic veins, the outcome is guarded.(Liver Transpl 2001;7:125-131.)

Although portal vein thrombosis (PVT) is not cur-rently an absolute contraindication for orthotopic

liver transplantation (OLT), it is considered high riskbecause of the complexity of OLT, resulting in surgicalmorbidity and mortality. PVT is a common complica-tion of liver cirrhosis, with an incidence of 2% to 14%at the time of OLT. The incidence can be markedlygreater in patients with autoimmune cirrhosis, porta-systemic shunts, cirrhosis with tumors, Budd-Chiarisyndrome, or post–necrotic cirrhosis.1-3

Because a liver allograft requires adequate portalflow to be viable, patency of the portal vein must beanalyzed pre-OLT. To deal with extensive venous col-laterals in the hepatoduodenal ligament and the exten-sion of thrombosis within the splanchnic veins requiressignificant surgical experience of the transplant team.Previous knowledge of the presence and extent of PVT

may lead to modification or even contraindication ofthe surgical procedure.

In this study, we compare 62 primary OLTs per-formed in patients with PVT at the time of OLT with agroup of 329 primary OLTs performed in patientswithout PVT and analyze the incidence of PVT, use ofdiagnostic methods, surgical management, and out-come.

Patients and Methods

Between January 1991 and April 1998 at our hospital, 415OLTs were performed in 391 adult patients. Sixty-two pa-tients (44 men, 18 women) were found to have PVT at thetime of OLT. Their mean age was 50 6 11 years. The inci-dence of PVT was determined in the 391 primary OLTs andrelated to the underlying primary liver disease and type ofimplantation technique.

The 62 patients with PVT were divided into 2 groups:48 patients with partial PVT (group 1) and 14 patients withcomplete PVT (group 2). Three hundred twenty-nine pa-tients without PVT (group 3) functioned as healthy controls.Perioperative and postoperative parameters (volume of trans-fused blood and recovery from the surgical field, surgical time,intensive care unit stay, early reoperation rate, and postoper-ative mortality rate) were compared among these 3 groups.

Primary Diagnoses

Primary diagnoses for patients with and without PVT arelisted in Table 1. Viral and alcoholic liver diseases were themost common, whereas more than 1 diagnosis was present in134 patients (data not shown).

Preoperative Radiological Studies

Until 1995, evaluation of the portal vein was performed usingDoppler ultrasound, which was followed by visceral angio-

From the Department of Surgery, Liver Transplant Unit, HospitalLa Fe, Valencia, Spain.

Address reprint requests to Gerardo Manzanet Andres, MD, PhD,Plaza San Pascual, 15, 12540 Vila-real (Castellon), Spain. Telephone/FAX: 34-9638-62754; E-mail: jos_mir@gvn.es

Copyright © 2001 by the American Association for the Study ofLiver Diseases

1527-6465/01/0702-0007$35.00/0doi:10.1053/jlts.2001.21295

125Liver Transplantation, Vol 7, No 2 (February), 2001: pp 125-131

graphy (venous phase) when thrombosis was suspected. Since1995, Doppler ultrasound and vascular magnetic resonanceimaging (MRI) have been performed to evaluate the patencyof the portal vein.

Surgical Findings

PVT was classified as partial if the occlusion of the portallumen was less than 90% (n 5 48) and as complete if theocclusion was greater (n 5 14). Complete PVT was found inthe portal vein only (n 5 8) or extended into the superiormesenteric and/or splenic vein (n 5 6). The wall of the portalvein often presented with thickening and fibrosis. In completePVT, portal hypertension was often severe, with large venouscollaterals in the hepatoduodenal ligament, mainly along theextrahepatic bile duct, whereas in some cases, a large-calibercoronary vein was noted.

Surgical Technique

Hepatectomy technique in the transplant recipient was notdependent on the presence or absence of PVT. Between 1991and 1993, we performed classical hepatectomy (n 5 76), butin 1993, we changed to the piggyback (vena cava–sparing)technique (n 5 315), using the 2 or 3 suprahepatic veins ofthe recipient (Table 2).

To determine the spread of complete PVT along thesplanchnic axis, the portal vein was dissected as far as thesplenomesenteric confluence. A Fogarty balloon catheter wasinserted into the lumen and carefully advanced toward thesuperior mesenteric vein and splenic vein. The patency of the2 venous trunks was verified by inflating the balloon, followedby gentle retraction of the catheter. Additional intraoperativeportography was not necessary. Portal vein thrombectomycould be performed in all cases, either with the Fogarty bal-loon or the aid of a Bakes forceps. In case this surgical ap-proach was not sufficient, thrombectomy was performed,with eversion of the vein and sharp dissection of the thrombusuntil adequate blood flow was obtained. When a large coro-

nary vein was found, it was ligated to prevent steal syndromefrom the portal flow. Portal anastomosis was always end toend, whereas a venous interposition graft was never necessary.With the exception of 1 case, no prophylactic treatment wasadministered with heparin.

Follow-Up

Post-OLT arterial and venous patency was evaluated rou-tinely using Doppler ultrasound at 48 hours, 10 days, and3 months. When deemed necessary, reexploration was per-formed. All patients have been followed up to date. Postop-erative morbidity and mortality rates were defined accordingto the criteria of the European Liver Transplant Registry,which included every event within the first 3 months post-OLT.

Statistical Analysis

Comparison between ratios was performed using chi-squaredand Fisher’s exact tests, and comparison of the means wasperformed using one-way ANOVA and Kruskal-Wallis test.Actuarial survival curves were determined for the patients andcompared using Wilcoxon (Gehan) test. Findings are consid-ered significant for P less than .05.

Results

The incidence of PVT at the time of OLT was 15.9%(62 of 391 OLTs). Although 48 of 62 PVTs occurred inpatients with viral and alcoholic cirrhosis, it was morefrequent in patients with autoimmune cirrhosis (40%)and cryptogenic cirrhosis (28.5%), whereas it was lesscommon in those with hepatocarcinoma (9.1%) andprimary biliary cirrhosis (5.3%; P 5 .02). The inci-dence of PVT in patients with a transjugular intrahe-patic portasystemic shunt (TIPS) was 31%, mainly ofthe partial type (9 of 10 patients; Table 1).

The results of the preoperative radiological studies inthe 3 groups are listed in Table 3. Sensitivity and spec-ificity of preoperative Doppler ultrasound in detectingportal thrombosis were 29% and 97.5%, respectively.Visceral angiography and vascular MRI had sensitivitiesof 57.1% and 65.4% and specificities of 85% and

Table 2. Hepatectomy Technique in the 3 DefinedPatient Groups

ClassicalTechnique Piggyback

Partial PVT (group 1) 10 38Complete PVT (group 2) 2 12Normal (group 3) 64 265Total 76 315

Table 1. Indications for Primary OLT in PatientsWith PVT

Primary Diagnosis Total PVT %

Viral cirrhosis 195 35 17.9Alcoholic cirrhosis 77 13 16.9Hepatocarcinoma 77 7 9.1Cryptogenic cirrhosis 14 4 28.5Autoimmune cirrhosis 5 2 40Primary biliary cirrhosis 1 1 5.3Others 22 0 0Total 391 62 15.6TIPS 32 10 31

Abbreviations: HCV, hepatitis C virus; HBV, hepatitis Bvirus.

126 Manzanet et al

88.3%, respectively. Because the time between radio-logical tests and OLT varied, ranging from 6 days to 10months, sensitivity and specificity of the radiologicalstudies were independent of the time factor.

The 3 groups were not different with regard to pa-tient age, surgical time, or cold and warm ischemictime. In the OLTs performed using the classical tech-nique, blood loss, characterized by the mean volume ofblood recovered by autotransfusion apparatus (Haemo-netic; Haemonetics Corp, Massachusetts) and thenumber of blood transfusions, was similar in the 3groups. However, in OLTs performed using the piggy-back technique, patients with PVT experienced in-creased blood loss (P 5 .008) and were administered alarger number of blood transfusions than patients with-out PVT (P 5 .056; Table 4).

The early reoperation rate, excluding retransplanta-tions, was significantly greater in group 2 (3 of 14patients; 21.4%) than in groups 1 (3 of 48 patients;6.3%) and 3 (13 of 329 patients; 4%; P 5 .04). Post-operative stay in the intensive care unit was not differ-ent among the 3 groups: 5.8 6 10 days in group 1; 7 67 days in group 2; and 6.5 6 11 days in group 3.

Portal thrombectomy was successful in 59 cases.There were 3 cases of postoperative rethrombosis(4.8%); 2 cases with complete PVT (14.3%), and 1 casewith partial PVT (2%). Two patients with a clinicallyviable liver underwent reoperation within 24 to 48hours, and intraoperative portography showed com-plete PVT. Rethrombectomy was not successful in 1patient (complete PVT), who died 9 days later; theother patient (partial PVT) died postoperative day 28 ofsepsis. The third patient (complete PVT) presentedwith complete necrosis of the graft and died at 7 daysduring a re-OLT.

Six patients died in the postoperative period (6 to 72days). Mortality rates at 30 days and 3 months showedno significant differences among the 3 groups (group 1,4.3% and 6.3%; group 2, 16.7% and 21.4%; group 3,8.6% and 10.6%, respectively; P . .3). The postoper-ative mortality rate for complete PVT with extensioninto the splenic and superior mesenteric veins was 33%(2 of 6 patients) caused by rethrombosis of the portalvein.

Four-year actuarial survival rates (Fig. 1) were 76%in group 1, 66% in group 2, and 61% in group 3, withno significant differences among the 3 groups (P 5 .1).

Discussion

The incidence of PVT diagnosed at the time of OLTwas 15.9% in our series, but it was more frequent inpatients with autoimmune or cryptogenic cirrhosis andTIPS and less frequent in those with hepatoma or pri-mary biliary cirrhosis. Although the explanation for thisthrombotic tendency is not clear, it is likely related toincreased intrahepatic resistance to portal flow caused

Table 3. Preoperative Radiological Findings in the 3Defined Patient Groups

RadiologicalImage

DopplerUltrasound Arteriography

VascularMRI

Group 1No. of patients

(n 5 48) 48 5 18Normal 36 3 6Partial PVT 10 1 10Complete PVT 1 1 1Difficult to

evaluate 1 0 1Group 2

No. of patients(n 5 14) 14 2 8

Normal 8 0 3Partial PVT 5 0 4Complete PVT 1 2 1

Group 3No. of patients

(n 5 329) 329 20 120Normal 321 17 106Partial PVT 8 3 14Complete PVT 0 0 0

Table 4. Volume of Recovered and Transfused Blood in the 3 Defined Patient Groups

Portal Vein Status

Volume of Recovered Blood (mL) Units of Transfused Blood

Classical Technique Piggyback Classical Technique Piggyback

Partial PVT (group 1) 3,042 6 1,949 1,473 6 1,174 7.5 6 4.4 4.2 6 3.4Complete PVT (group 2) 1,505 6 1,053 2,047 6 1,435 7.0 6 7.0 5.3 6 3.3Normal (group 3) 3,566 6 3,893 1,151 6 1,348 7.2 6 4.3 3.5 6 2.8P .505 .008 .916 .056

127Liver Transplantation and PVT

by the atrophied cirrhotic liver in combination withcoagulation abnormalities. In addition, a greater inci-dence of PVT has also been reported in patients withhematologic disorders (e.g., polycythemia, Budd-Chiari, antithrombin III deficit) and surgery involvingthe portal vein.1,2,4

We routinely evaluate the patency of the portal veinpreoperatively to plan the operation and obtain goodvenous grafts from the donor in case an alternativemethod of portal inflow is necessary. As also reported byothers,3,5,6 we found a low sensitivity of Doppler ultra-sound (29%) in preoperative screening for the diagnosisof PVT. However, with 2.5% of false-positive diag-noses, the specificity was greater than with visceral an-giography or vascular MRI. Although color Dopplerultrasound has been reported to yield a greater sensitiv-ity,7 we cannot comment on this because we used thistechnique in only a few cases. Vascular MRI proved tobe the most sensitive test for detecting PVT (34.6%).Venous-phase visceral angiography was the most accu-rate test for detecting complete PVT, with 100% sen-sitivity and 4% false-positive diagnoses. However, inpartial PVT, its efficiency was less because it only de-tected 40% of the cases. Although there are numerousstudies on the application of radiological modalities inOLT, only a few compare intraoperative portal veinfindings with the preoperative images. Based on ourexperience, we believe that for now, both color Dopplerultrasound and vascular MRI should be used for theevaluation of portal vein patency until we know the

sensitivity of the first modality. Visceral angiographyshould be reserved exclusively for when complete PVTis suspected but cannot be proven. In the event of aprolonged waiting time before OLT, noninvasive stud-ies should be repeated every 3 months to evaluate pos-sible changes.

Because approximately 75% of the blood flow to anormal liver originates from the portal vein, portal flowneeds to be adequate for OLT to be successful. Portalvein reconstruction is usually provided by end-to-endanastomosis between the recipient and donor portalveins. The surgical technique for reestablishing portalflow in OLT with PVT depends on the extent ofthrombosis and the experience acquired by the trans-plant team (Table 5). Partial PVT less than 25% isusually local and has no clinical repercussions because itcan be treated with resection, whereas that of more than25% requires extensive thrombectomy.8

Initially, complete PVT was considered an absolutecontraindication to OLT until good results were ob-tained with thrombectomy or a venous interpositiongraft between the donor portal vein and splenomesen-teric confluence.9 However, extra-anatomic mesenteric-portal venous grafts, technically less difficult, have oftenreplaced the former technique because it is not neces-sary to dissect the superior pancreatic area, which is richin venous collaterals, and results in less mortality and alow incidence of post-OLT pancreatitis.10-12 In somecases with complete PVT, a large coronary vein or a big

Figure 1. Patient actuarialsurvival curves in primaryOLT with partial PVT, com-plete PVT, and normal portalvein at the time of OLT.

128 Manzanet et al

periportal varix have been used for anastomosis with thedonor portal vein.13-15

Portal flow can be improved by ligating the coronaryvein to prevent the steal of portal blood flow. In addi-tion, to improve portal blood flow, arterialization of theportal vein with the recipient splenic artery has beenreported, as well as interposition of a venous graft be-tween the coronary vein and recipient portal vein.3

Unlike other investigators (Table 5), we never usedvenous grafts for portal vein reconstruction in our pa-tients with PVT because we revascularized all graftsusing thrombectomy in combination with ligation ofthe coronary vein when it was judged necessary to im-prove portal flow.

Although PVT currently is not a contraindication toOLT, the extent of thrombosis may influence the re-sults. Mortality has been reported to be greater in pa-tients with PVT with splanchnic vein involvement(45.5% v 36.1%)3 or associated periphlebitis (83.3% v9.4%),16,17 which may be attributed to greater technical

difficulties and related blood loss.3 Most PVTs in ourseries were partial and only involved the portal vein,whereas complete PVT with spread into the superiormesenteric or splenic vein was less frequent. Althoughcomplete PVT resulted in a greater rate of reoperationfor hemoperitoneum than partial PVT, no significantdifference was found in postoperative survival betweenthese 2 groups. However, complete PVT with involve-ment of the splanchnic veins had a mortality rate of33%.

Some investigators6 routinely perform intraopera-tive portography in all cases of complete PVT by can-nulating a peripheral mesenteric vein. They evaluate theextent of the thrombosis and the result of thrombec-tomy after performing vascular reconstruction. In case asignificant hepatofugal shunt is detected, they ligate themost obvious collaterals, generally the coronary vein orconnections between the splenic vein and left renalvein, to increase portal blood flow. We consider the useof a Fogarty balloon catheter to evaluate splanchnic vein

Table 5. Literature Review of Portal Thrombosis and Liver Transplantation

ReferenceNo. ofCases

Portal VeinReconstruction Rethrombosis (%) Mortality (%) Follow-Up

Stieber et al3 (1991) 32 13 VT or VGSC 38.5 32.4 —14 MPBP 21.45 VT 1 AC 0

Lerut et al16,17 (1997) 38 24 VT 4.2 21.1 —8 MPBP 03 CD 33.31 VGSC 02 Varix 0

Shaked and Busuttil2 (1991) 23 14 VT 4.2 35 —6 MPBP 11.11 VGSC 02 CV or Varix 0

Moreno et al8 (1993) 14 13 VT 23 42.6 6 Alive .1 yr1 MPBP 100 2 Alive .5 mo

Figueras et al6 (1997) 10 CPVT 10 MPBP 10 0 87% Alive at 1 yrDavidson et al1 (1994) 14 13 VT 23 14.3 8 Alive at 37 mo

1 CV 0Langnas et al19 (1992) 16 8 VGSC 12.5 18.8 81% Alive at 12.5 mo

5 VT 203 MPBP 0

Cherqui et al5 (1993) 3 CPVT 10 VT 0 9.1 8 Alive at 4-39 mo8 PPVT 1 Varix 0

Kirsh et al24 (1990) 8 6 MPBP 0 12.5 63% Alive at 1 yr2 VGSC 0

Abbreviations: VT, venous thrombectomy; VGSC, vein graft to splenomesenteric confluence; MPBP, mesoportal venous bypass; CV,anastomosis with recipient stomachal coronary vein; Varix, anastomosis with varicose vein of hepatic hilum; AC, systemic anticoagula-tion; CD, confluence dissection.

129Liver Transplantation and PVT

patency a good alternative to intraoperative portogra-phy. The latter may be reserved for problematic cases orsituations in which adequate portal flow is not achievedto identify and ligate overlooked venous collaterals thatmay cause a vascular steal phenomenon.

The main risk for liver transplant recipients withPVT is early rethrombosis. The incidence varies amonginvestigators (Table 5). A greater risk for rethrombosishas been reported in grafts that were not preserved withUniversity of Wisconsin solution or when venovenousbypass was used.8 Theoretically, this can be attributedto endothelial damage related to the use of less effectivepreservation fluids and hypercoagulability caused bythe use of bypass. Moreover, an outflow problem in thesuprahepatic veins can also lead to rethrombosis of theportal vein.18

Some recommend the use of therapeutic8 or prophy-lactic6 anticoagulation for 3 months to prevent re-thrombosis. Stieber et al3 achieved 100% survival in5 patients with PVT extending into the splanchnicveins by using extensive thrombectomy of the portalvein and its tributaries and anticoagulation. However,others do not use treatment with heparin.5,17,19

In our series, the rate of rethrombosis varied with thedegree of previous thrombosis (complete PVT, 14.3%;partial PVT, 2%). Two of the 3 cases of rethrombosisappeared in patients with extensive complete PVT inwhom thrombectomy was considered incomplete be-cause of the initial inexperience (OLT no. 60 and 62),resulting in suboptimal portal flow. We believe thatobtaining adequate portal flow is the best guarantee forpreventing rethrombosis, and we avoid systemic anti-coagulation because of the risk for postoperative bleed-ing complications.

Partial rethrombosis might manifest with acute de-teriorating liver function and complications secondaryto portal hypertension, such as ascites or gastrointesti-nal bleeding. If detected early, it can be treated effec-tively with rethrombectomy. However, a delay in thisdiagnosis may lead to graft loss and retransplantation.Frequent Doppler ultrasonography in the post-OLTperiod may prevent delay in the diagnosis of PVT andtherefore the need for retransplantation and is recom-mended every 1 to 3 days during the first 2 weeks afterOLT.20,21

Although in some cases, rethrombosis of the portalvein is well tolerated without acute changes in hepato-cellular function, there is a risk for progressive hepaticatrophy and bleeding from varices caused by portalhypertension. Some investigators8 recommend a selec-tive distal splenorenal shunt, whereas others22 recom-mend sclerotherapy as the treatment of choice. For the

treatment of portal vein stenosis in the late post-OLTperiod, percutaneous balloon angioplasty through thetransjugular, transfemoral, or transhepatic approachmay be an alternative option to surgery. This requiresexact location of the stenosed area with angiographicstudies. Placement of a TIPS is not recommended be-cause it reduces effective portal flow and may deterio-rate liver function over the long term.18

Patients with complete PVT are in general poorcandidates for OLT; thus, each transplant team mustevaluate the indication for OLT individually. Becausethe mortality rate for these patients for OLT rangesfrom 0% to 80% (Table 5), OLT must be comparedwith other treatment modalities. Rikkers et al23 per-formed a distal splenorenal shunt in 22 patients and anonselective shunt in 5 patients with variceal bleedingwho were not considered transplant candidates, with asurgical mortality rate of 7.4% and 1-year and 5-yearsurvival rates of 72% and 42%, respectively. Some in-vestigators8 consider OLT the only option for patientswith complete PVT because a mesocaval shunt has highmorbidity and mortality, whereas a distal splenorenalshunt, carrying a lower risk, may be contraindicatedbecause of unmanageable ascites.

Although differences exist between partial and com-plete PVT, the most important risk factor for rethrom-bosis is the extent of thrombus within the portal venousbed. We believe that patients with complete PVT withextension throughout the splanchnic veins should beexcluded from OLT because of the high rates of re-thrombosis and mortality, which are probably greaterthan those of the natural course of their liver disease.However, the outcome in less extensive complete PVTis certainly acceptable.

References1. Davidson BR, Gibson M, Dick R, Burroughs A, Rolles K. Inci-

dence, risk factors, management, and outcome of portal veinabnormalities at orthotopic liver transplantation. Transplanta-tion 1994;57:1174-1177.

2. Shaked A, Busuttil RW. Liver transplantation in patients withportal vein thrombosis and central portacaval shunts. Ann Surg1991;214:696-702.

3. Stieber AC, Zetti G, Todo S, Tzakis AG, Fung JJ, Marino I, et al.The spectrum of portal vein thrombosis in liver transplantation.Ann Surg 1991;213:199-206.

4. Nonami T, Yokoyama I, Iwatsuki S, Starzl T. The incidence ofportal vein thrombosis at liver transplantation. Hepatology1992;16:1195-1198.

5. Cherqui D, Duvoux C, Rahmouni A, Rotman N, Dhumeaux D,Julien M, et al. Orthotopic liver transplantation in the presenceof partial or total portal vein thrombosis: Problems in diagnosisand management. World J Surg 1993;17:669-674.

6. Figueras J, Torras J, Rafecas A, Fabregat J, Ramos E, Moreno G,

130 Manzanet et al

et al. Extra-anatomic venous graft for portal vein thrombosis inliver transplantation. Transpl Int 1997;10:407-408.

7. Grant EG, Tessler FN, Gomez AS, Holmes CL, Perrella RR,Duerickx AJ, et al. Color Doppler imaging of portosystemicshunts. Am J Radiol 1990;154:393-397.

8. Moreno Gonzalez E, Garcıa Garcıa I, Gomez Sanz R, Gonzalez-Pinto I, Loinaz Segurola C, Jimenez Romero C. Liver transplan-tation in patients with thrombosis of the portal, splenic or supe-rior mesenteric vein. Br J Surg 1993;80:81-85.

9. Shaw BW, Iwatsuki S, Bron K, Starzl TE. Portal vein grafts inhepatic transplantation. Surg Gynecol Obstet 1985;161:66-68.

10. Burdick JF, Pitt HA, Colombani PM, Perler BA, Klein AS,Merritt W, et al. Superior mesenteric vein inflow for liver trans-plantation when the portal vein is occluded. Surgery 1990;107:342-345.

11. Sheil AGR, Thompson JF, Stevens MS, Eyers AA, Graham JC,Bookallil MJ. Mesoportal graft for thrombosed portal vein inliver transplantation. Clin Transplant 1987;1:18-20.

12. Tzakis A, Todo S, Stieber A, Starzl TE. Venous jump grafts forliver transplantation in patients with portal vein thrombosis.Transplantation 1989;48:530-531.

13. Czerniak A, Badger I, Sherlock D, Buckels J. Orthotopic livertransplantation in a patient with thrombosis of the hepatic portaland superior mesenteric veins. Transplantation 1990;50:334-336.

14. Hiatt JR, Quinones-Baldrich WJ, Ramming KP, Lois JF, Busut-til RW. Bile duct varices. An alternative to portoportal anasto-mosis in liver transplantation. Transplantation 1986;42:85.

15. Spira R, Widrich WC, Keusch KD, Jackson BT, Katzman HE,Coello AA. Bile duct varices. Arch Surg 1985;120:1194-1196.

16. Lerut J. Portal vein thrombosis and liver transplantation. TransplInt 1998;11:70.

17. Lerut JP, Mazza D, Laterre PF, Donataccio M, de Goyet JV, Van

Beers B, et al. Adult liver transplantation and abnormalities ofsplanchnic veins: Experience in 53 patients. Transpl Int 1997;10:125-132.

18. Fujimoto M, Moriyasu F, Someda H, Nada T, Okuma M,Uemoto S, et al. Recovery of graft circulation following percuta-neous transluminal angioplasty for stenotic venous complica-tions in pediatric liver transplantation: Assessment with Dopplerultrasound. Transpl Int 1995;8:119-125.

19. Langnas AN, Marujo WC, Stratta RJ, Wood RP, Ranjan D,Ozaki C, et al. A selective approach to preexisting portal veinthrombosis in patients undergoing liver transplantation. Am JSurg 1992;163:132-136.

20. Hellinger A, Roll C, Stracke A, Erhard J, Eigler FW. Impact ofcolour Doppler sonography on detection of thrombosis of thehepatic artery and portal vein after liver transplantation. Langen-becks Arch Chir 1996;381:182-185.

21. Kok T, Slooff MJH, Thijn CJP, Peeters PMJG, Bijleveld CMA,van den Berg AP, et al. Routine Doppler ultrasound for thedetection of clinically unsuspected vascular complications in theearly postoperative phase after orthotopic liver transplantation.Transpl Int 1998;11:272-276.

22. Henderson JM, Kutner MH, Millikan WJ Jr, Galambos JT,Riepe SP, Brooks WS, et al. Endoscopic variceal sclerosis com-pared with distal splenorenal shunt to prevent recurrent varicealbleeding in cirrhosis: A prospective, randomized trial. Ann In-tern Med 1990;112:262-269.

23. Rikkers LF, Jin G, Langnas AN, Shaw BW. Shunt surgery duringthe era of liver transplantation. Ann Surg 1997;226:51-57.

24. Kirsch JP, Howard TK, Klintmalm GB, Husbberg BS, Gold-stein RM. Problematic vascular reconstruction in liver transplan-tation. Part II. Portovenous conduits. Surgery 1990;107:544-548.

131Liver Transplantation and PVT