Objectives: Portal vein thrombosis is no longer a contraindication for liver transplantation. However, varied outcomes are still reported with regard to patients with complete portal vein thrombosis.
Materials and Methods: We analyzed data of 505 patients who underwent liver transplant between January 2014 and June 2017. Of 505 patients, 55 (10.9%) had portal vein thrombosis, which included 37 patients (67%) who had living-donor and 18 patients (33%) who had deceased-donor liver transplant. According to Yerdel classification, 5 (9%), 28 (51%), 12 (22%), and 10 patients (18%) had grade I, II, III, and IV portal vein thrombosis, respectively.
Results: Thrombectomy was successful in 46 patients (84%). Forty-three patients had patent portal inflow (93.5%), but 3 patients (6.5%) required a second thrombectomy on day 1 posttransplant. Of 10 patients with grade IV portal vein thrombosis, 9 (16%) underwent extra-anatomic portal vein anastomosis. Of these 9 patients, 7 (77.7%) showed patent portal inflow during the early postoperative period. Median follow-up time was 15.5 ± 13.1 months. During this time, portal vein stenosis was detected in 5 patients. Risk of repeat thrombosis was not significantly different between patients with deceased-donor and patients with living-donor liver transplant. Overall patient survival rates for patients with and without portal vein thrombosis were 73% (40/55) and 84% (376/450), respectively, which showed significant difference (P = .05).
Conclusions: In contrast to early-grade portal vein thrombosis, complete portal vein thrombosis still provides a treatment challenge with high risk of morbidity and poor early outcomes in liver transplant recipients.
Key words : Deceased donor, Liver transplantation, Living donor
Portal vein thrombosis (PVT) remains a serious complication of end-stage liver disease, with an incidence ranging from 7% to 16%. Portal vein thrombosis is not considered to be an absolute contraindication to liver transplant. It can be suc-cessfully performed with the provision of advanced surgical techniques.1
Treatment of PVT is surgically challenging, and the optimum way to manage complete PVT remains controversial. Extra-anatomic portal vein recon-struction can be achieved using various techniques, including renoportal and varicoportal anastomoses.2 There is considerable debate over the management of PVT in living-donor liver transplant (LDLT), which, unlike deceased-donor liver transplant (DDLT), is associated with surgical challenges.
In this study, we retrospectively collected and analyzed data of patients who underwent liver transplant with preexisting PVT at a single center. Patient sociodemographic characteristics, surgical techniques used, and PVT grade and outcomes were considered in our analyses.
Materials and Methods
This study was approved by the Institutional Review Board of our university (ATADEK: 2017/9). We retro-spectively analyzed medical records of 505 patients who underwent liver transplant at our institution between January 2014 and June 2017. Of total patients, 55 (11%) had PVT before surgery. Of these patients, 18 (33%) had DDLT and 37 (67%) had LDLT.
Doppler ultrasonography and computed tomo-graphy (CT) angiography were used to evaluate the portal vein of transplant recipients 2 months before surgery. The surgical approach to radiologically detect PVT was established according to the degree of PVT, which was determined using Yerdel classification.3 The final surgical approach was decided based on the PVT grade identified following hilar dissection.
Patients with PVT and hepatocellular carcinoma (HCC) underwent liver transplant after the nature of PVT was established using contrast-enhanced imaging techniques. Magnetic resonance imaging was also used in this group to determine whether the PVT was benign or malignant.
Blood flow was confirmed in the recipient’s portal vein following the successful reconstruction of throm-bectomy and portal anastomosis, with or without an interpositional graft. Complicated surgical techniques, such as extra-anatomic portal vein anastomosis (renoportal/varicoportal anastomosis), were used for patients in whom portal flow could not be maintained by portoportal anastomosis. Six patients required cryopreserved iliac vein grafts. Deceased-donor iliac vein grafts from different blood groups were preserved in a conserving solution at -80°C. The composition of this conserving solution was as follows: 80 mg gentamicin, 55 mL allogenic fresh frozen plasma, and 5 mL dimethyl sulfoxide. Vein grafts that were cryopreserved in this fashion were presumed to be viable for a period of 1 month, after which they were regarded as expired or unusable.
Portal vein patency and flow characteristics were monitored daily posttransplant by perioperative Doppler ultrasonography during hospitalization. During follow-up, patients received Doppler ultraso-nography every 2 months and CT angiography was performed if there was suspicion of thrombus or stenosis.
Postoperative bleeding, PVT, ascites, and primary graft dysfunction, considered to be PVT-related complications, were evaluated. Patient and graft survival at 1 month and at 1 year after liver trans-plant were also analyzed.
Statistical analyses were performed with Statistical Program for Social Sciences (SPSS Inc., Chicago, IL, USA) version 15.0. Descriptive statistics are shown as means and standard deviation for continuous variables. To evaluate differences between groups, we used chi-square tests. To evaluate disease-free and overall survival, we performed survival plot analyses with the Kaplan-Meier method. P < .05 was considered statistically significant.
Of 505 patients who underwent liver transplant, 55 (11%) had preoperative PVT. Twenty-one were women (38%) and 34 were men (62%), with a mean age of 51.3 ± 14.2 years. Cirrhosis, secondary to hepatitis B virus infection (16%), was the most frequently detected primary reason for transplant; 22 patients (40%) had cryptogenic cirrhosis. Thirty-seven patients (67%) had LDLT, with right-lobe LDLT performed in 35 patients and left-lobe LDLT performed in 2 patients. The remaining patients (n = 18) underwent DDLT.
On the basis of Doppler ultrasonography and CT angiography findings during surgery, 5 (9%), 28 (51%), 12 (22%), and 10 (18%) patients had Yerdel classification grade I, II, III, and IV PVT, respectively. There were no false-positive radiologic findings on preoperative assessments. However, PVT was detected by perioperative evaluation in 1 patient with grade II and in 1 patient with grade III, in contrast to the respective preoperative imaging findings of grade I and grade II PVT, respectively. The sociodemographic data and clinical characteristics of patients with PVT are shown in Table 1.
Hepatocellular carcinoma was documented in 13 patients. Of these, 7 and 3 had grade II and III PVT, respectively. Two patients with HCC had grade IV PVT, and 1 patient with HCC had grade I PVT. A diagnosis of benign PVT was based on the contrast-enhanced imaging findings. Pathologic examinations after surgery confirmed benign PVT in all patients with HCC, except for one. This patient had malignant thrombus on pathologic examination, which were in contrast to findings at preoperative radiologic examination. At 3 years posttransplant, this patient is alive without HCC recurrence.
Thrombectomy was successful in ensuring portal vein inflow in 46 patients (84%); 45 patients had grade I, II, or III PVT and 1 patient had grade IV PVT. This patient was on anticoagulant therapy for a lengthy duration before surgery. After thrombectomy, an interpositional vein graft was required in 1 patient with grade III PVT.
Of 10 patients with grade IV PVT, 9 (16%) under-went extra-anatomic portal vein anastomosis (16%). Five had varicoportal anastomosis (9%), and the other 4 patients (7%) had renoportal anastomosis. A left gastric vein varix was used to provide portal vein inflow for 4 patients with varicoportal anastomosis, and a hilar varix was utilized in the 5th patient. Inadequate portal inflow was demonstrated on perioperative Doppler ultrasonography for the 5th case. Arterialization was added to the varicoportal anastomosis to increase blood flow.
Of 5 patients with varicoportal anastomosis, a cryopreserved iliac vein graft was used in 3 patients, but an interpositional vein graft was not needed for the remaining 2 patients. After a decision to perform renoportal anastomosis before liver transplant, spon-taneous splenorenal collaterals were demonstrated in patients with grade IV PVT, showing adequate portal inflow after the renoportal anastomosis was reconstructed. Portal vein anastomosis was achieved using the left renal vein in all of the renoportal anastomosis cases. A cryopreserved iliac vein graft was used to restore portal vein anastomosis in 3 renoportal anastomosis patients who underwent LDLT, and a fresh iliac vein graft was utilized in the remaining patient who underwent DDLT. The features of portal vein flow were determined using perioperative Doppler ultrasonography following anastomosis. The PVT classification and surgical techniques performed to ensure portal inflow are shown in Table 2.
Low-molecular-weight heparin was utilized to prevent recurrence of thrombus in the portal vein in the early postoperative period. Patients were discharged and treated with acetylsalicylic acid. Rivaroxaban was used as maintenance anticoagulant treatment in 2 patients with Budd-Chiari syndrome, both with MTHFR A1298C homozygous and JAK2-V617F heterozygous mutation.
Postoperative portal vein patency was examined routinely daily for 1 week by Doppler ultrasonography and then weekly until discharge. Computed tomo-graphy angiography was performed if there was suspicion of portal inflow insufficiency. Of 46 patients who underwent thrombectomy, patent portal inflow was established in 43 patients (93%); the remaining 3 patients (7%) underwent a second thrombectomy on day 1 posttransplant. Of these 3 patients, 2 had LDLT and 1 had DDLT. Patent portal inflow was observed in the latter at 1 year follow-up. A portal vein stent was applied to counter the portal vein stenosis at 13 months, with sufficient portal inflow then consistently demonstrated. The other 2 patients died after the second thrombectomy due to thrombosis recurrence that led to graft dysfunction and sepsis.
Patent portal inflow was demonstrated in 7 of 9 patients (78%) with extra-anatomic portal vein anastomosis in the early postoperative period. Of the remaining 2 patients, 1 patient with renoportal anastomosis experienced repeat thrombosis on day 1 posttransplant. It was not possible to establish adequate portal inflow in this patient because the patient had 2 renal veins; the patient died due to graft dysfunction. In the other patient, the right hepatic artery was used for arterialization, in conjunction with varicoportal anastomosis, to increase the portal inflow. Unfortunately, a postoperative increase in the portal inflow caused serious damage to the graft. Despite retransplant, the patient died due to diffuse thrombosis in both the portal vein and hepatic artery.
Portal vein stenosis was detected in 5 patients during follow-up. Repeat thrombectomy was needed on day 1 posttransplant in a patient with grade III PVT. A portal vein stent was placed in this patient at 1 year after liver transplant. Three patients with grade IV PVT underwent extra-anatomic anastomosis using cryopreserved vein grafts. A patient with renoportal anastomosis needed a portal vein stent in month 3 postoperatively, and a stent was placed by the interventional radiology team. Two other patients with varicoportal anastomosis had portal vein stents placed in months 8 and 9 posttransplant. Post-operative bleeding occurred in 6 patients (11%): 3 after thrombectomy, 1 after varicoportal anastomosis, and 2 after renoportal anastomosis. Neither patient with renoportal anastomosis had acute renal failure or any other renal problem during follow-up. Four patients had resistant ascites (2 with renoportal and 2 with varicoportal anastomosis) approximately 2 months after liver transplant, for which diuretic treatment was required.
Median follow-up duration was 13.1 ± 15.5 months (range, 1-55 mo). Fifteen patients (27%) were lost to follow-up. Of these, 11 patients died in the early postoperative period. Two of these patients had extra-anatomic portal vein anastomosis and died due to graft dysfunction as a result of inadequate portal inflow. One patient with renoportal anastomosis also died due to upper gastrointestinal bleeding on day 2 posttransplant, although adequate portal inflow was established. The other patients died for various reasons, including biliary sepsis and pneumosepsis. Only 1 patient died due to HCC recurrence in month 8 posttransplant. The other 3 patients died 2 years after transplant for extrahepatic reasons. The post-operative complications are shown in Table 3.
The risk of repeat thrombosis after thrombectomy was compared in LDLT versus DDLT patients. Of 46 patients who had thrombectomies, 29 (63%) under-went LDLT and 17 (37%) had DDLT. Three patients (7%) had repeat thrombosis, with 2 (7%) in the LDLT group and 1 (6%) in the DDLT group. Differences in risk of repeat thrombosis for DDLT versus LDLT patients were not statistically significant (Table 4).
Overall patient survival rates for patients with and without PVT were 73% (40/55) and 84% (376/450), respectively (Figure 1), which was significant (P = .05). Survival rates for patients with extra-anatomic portal vein and anatomic anastomosis were 56% (4/9) and 76% (11/35) (Figure 2), respectively, which was not significant (P = .29).
Extensive surgical expertise is necessary to overcome the operative challenges related to PVT. The ratio of patients with PVT who had liver transplant in our series (11%) was similar to that previously reported.4
Although the debate on how to care for patients with PVT after transplant is ongoing,5,6 most centers recommend the use of anticoagulant therapy before surgery, although others believe that there is insufficient evidence to support this.7 The objective in our practice is to use low-molecular-weight heparin in patients with PVT at diagnosis until the time of surgery. Our observation is that throm-bectomy is easier in patients who have used low-molecular-weight heparin before surgery for at least 2 months, although data are insufficient to support this premise.
Patients with PVT and malignancy are still considered to be contraindicated for liver transplant owing to the high risk of HCC recurrence. Malignant or benign PVT should be established in patients with cirrhosis and HCC, and it is essential to determine the PVT-related features. The use of well-defined criteria, based on imaging techniques, is recom-mended to establish the nature of PVT.7,8 Taking a biopsy of the thrombus before transplant is recommended as a safe option. However, a biopsy is an invasive procedure and is associated with a high risk of complications, including bleeding and seeding of the tumor. A false-negative malignant PVT diagnosis was reported for only 1 of our patients on radiologic examination. This patient did not have HCC recurrence during the 3 years of follow-up and is still alive to the best of our knowledge.
The use of Doppler ultrasonography preop-eratively is the preferred method of assessing the patency of portal veins. However, the addition of CT angiography has been suggested in recent reports to enhance the quality of the assessment. We routinely use both radiologic techniques early before surgery. Although false-negative results have been previously reported,7,9 we did not encounter this in our present series. Portal vein imaging is recommended imme-diately before surgery. This will result in the most informed, up-to-date evaluation of portal vein patency status. It is important for surgeons to be optimally prepared, especially before LDLT procedures. The use of CT angiography assists the surgeon with deciding on the best surgical strategy. The collateral circulation should also be taken into consideration. If renoportal anastomosis is warranted, careful ligation of the collateral veins that drain the left renal vein should be performed to maintain sufficient flow from the splanchnic area, while protecting the collateral veins from the spleen.
The preoperative identification of PVT grade is key in determining the surgical strategy to be applied, particularly for patients classified with Yerdel grade III or IV PVT who require innovative surgical techniques to restore portal vein flow. Thrombectomy is usually sufficient for Yerdel grade I and II PVT. End-to-end portal anastomosis was successful after thrombectomy in our patients, similar to that reported elsewhere.6,10 Thrombectomy was performed in patients with grade III PVT in our study, and the veins were used for portal vein anastomosis, with a high success rate. In contrast, a graft is usually considered necessary in patients with grade III PVT who undergo LDLT.5 Extra-anatomic portal vein anastomosis is thought to be obligatory when utilizing a vessel graft in patients with complete (grade IV) PVT undergoing LDLT.9,11 Several types of vessel grafts (cryopreserved vein, synthetic, and autologous veins) can be used for LDLT. We had the opportunity to preserve vessel grafts from other deceased donors to use for LDLT. Cryopreserved vein grafts were used to establish portal vein anastomosis in 6 of the patients with grade IV PVT undergoing LDLT.
Repeat thrombosis is a major risk in patients with PVT after liver transplant, with an incidence that varies from 4% to 39%. Frequent Doppler ultra-sonography is critical to avoid delays in diagnosis of repeat thrombosis, which has the potential for graft loss.5,12 Cho and associates12 reported an overall incidence of repeat thrombosis in PVT patients of 18% and up to 29% in patients with grade IV PVT. Kim and associates13 did not report repeat thrombosis after thrombectomy; however, they reported an incidence of 17% in cases where grafts were used for anastomosis. In our series, 4 patients (9% overall and 22% in the grade IV group that required vein grafts) experienced repeat thrombosis. Successful rethrom-bectomy was performed in 2 patients on day 1 posttransplant. The other 2 patients with extra-anatomic anastomosis achieved with a vein graft also had rethrombectomy on day 1 posttransplant, with the procedure being completed successfully in one of them. Unfortunately, the other patient (with renoportal anastomosis) died due to portal insufficiency as a result of recurrent thrombosis. In this case, it must be noted that 2 separate left renal veins, as well as the ovarian and lumbar veins, drained the renal vein to the caval system after thrombectomy. An attempt was made to ligate them, but this attempt was unsuccessful. The use of vein grafts is a risk factor for early repeat thrombosis. The quality of the cryopreserved veins is also important, and it is thought that the risk of coagulopathy in these veins is high. Objectives should be to prevent the risk of repeat thrombosis while using a vein graft, to obtain the shortest grafting route possible, and to avoid kinking. There is a risk that abdominal pressure may compromise portal flow in overweight patients, in particular, when renoportal anastomosis is being performed. Thus, the use of an aortic fresh conduit may be more acceptable for renoportal anastomosis.
Our comparison of risk of repeat thrombosis in the DDLT versus LDLT groups after thrombectomy showed that portal vein anastomosis after thrombectomy for grade I, II, and III PVT led to comparable outcomes in LDLT.
Variable results have been obtained for patients with grade IV (complete) PVT. Good PVT outcomes, compared versus those without PVT, have been previous demonstrated by Koh and associates7 and Kim and associates.13 However, early and late outcomes have been shown to be influenced by PVT grade in recent studies. Patients with grade IV PVT experience significantly worse outcomes, particularly in the early postoperative period.5,12,14 Ghabril and associates15 demonstrated that PVT (especially the complete form) represented an independent risk factor for 90-day mortality and graft failure. The outcomes in the present study support these study findings. Most patients were lost in the first month posttransplant, but 2 patients with renoportal anastomosis (2/4, 50% mortality) died in the first week posttransplant.
Patients with early-grade PVT are candidates for liver transplant with a good prognosis. However, complete PVT remains a challenge, with a high risk of morbidity and poor early outcomes. More studies are warranted to refine the results in this group.
DOI : 10.6002/ect.2018.0260
From the 1Organ Transplantation Department, the 2Anesthesiology and Reanimation
Department, and the 3Radiology Department, Acibadem Mehmet Ali Aydinlar
University Atakent Hospital, Istanbul, Turkey
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Ali Ozer; Acibadem Mehmet Ali Aydinlar University Atakent Hospital, Turgut Ozal Bulvarı, Halkalı, Küçükçekmece, Istanbul 16303, Turkey
Phone: +90 533 212 2324
Table 1. Demographics and Clinical Features
Table 2. Surgical Procedures Applied According to Grade of Portal Vein Thrombosis
Table 3. Postoperative Complications
Table 4. Comparison of Deceased-Donor and Living-Donor Liver Transplant Recipients With Regard to Repeat Thrombosis Risk After Thrombectomy
Figure 1. Kaplan-Meier Analysis of Patients With and Without Portal Vein Thrombosis (P = .05)
Figure 2. Kaplan-Meier Analysis of Patients With Anatomic and Extraanatomic Portal Vein Anastomosis (P = .29)