Objectives: Adequate venous outflow is one of the most important factors responsible for optimal graft function in liver transplantation. Thrombosis of the inferior vena cava in cases of Budd-Chiari syndrome poses a major challenge to a transplant surgeon in establishing proper graft outflow. In deceased donor liver transplant, this problem can be dealt with relative ease as the liver graft includes donor inferior vena cava. However, this is not the case in living donor liver transplant. We present our findings of living donor liver transplant for Budd-Chiari syndrome and discuss techniques that have helped overcome this unique problem without the need for complete inferior vena cava replacement.
Materials and Methods: Our retrospective analysis included living donor liver transplant recipients from November 2006 to March 2020 at our center and selected patients who underwent this transplant for Budd-Chiari syndrome. We studied the extent and severity of inferior vena cava involvement in these cases. We developed a classification that not only helped to stratify patterns of venacaval disease but also helped to plan the surgical technique. The role of interventional radiology combined with surgery in management of extensive inferior vena cava stenosis was studied.
Results: Among 2952 cases of liver transplant in our unit from November 2006 to March 2020, 36 patients had Budd-Chiari syndrome; 21 had significant level of inferior vena cava thrombosis, which was managed with inferior vena cava thrombectomy with either patchplasty (n = 20) or segmental replacement (n = 1). None of our patients showed recurrence of primary disease during the median follow-up of 36 months (range, 8-158 mo).
Conclusions: Establishment of adequate venous ouflow in thrombosed inferior vena cava is possible with proper planning of surgical technique and timely involvement of interventional radiology-guided interventions in patients with Budd-Chiari syndrome.
Key words : Inferior vena cava reinforcement, Inferior vena cava thrombosis in Budd-Chiari syndrome, Living donor liver transplant for Budd-Chiari syndrome, Surgical classification of Budd-Chiari with inferior vena cava involvement
Budd-Chiari syndrome (BCS) is defined as obstruction to the hepatic venous outflow tract in the absence of any cardiac or pericardial disease.1 By definition, veno-occlusive or sinusoidal obstruction syndrome, as it is referred to today, are not included.2 The etiology of BCS can be either a prothrombotic disorder or can be idiopathic in nature.3 Prothrombotic disorders can be both inherent and acquired.4 Geographical variations in the anatomic site of venous outflow obstruction are frequently seen between cases reported in Western populations and those reported in Asian populations. The level of obstruction is more commonly at the hepatic vein in Western population, whereas studies from Asian countries have shown the obstruction to be more prevalent at the level of inferior vena cava (IVC).5
Clinical manifestations of this syndrome may vary from an asymptomatic course to a florid manifestation, ranging from fulminate liver failure to decompensated cirrhosis, such as ascites, jaundice, and liver failure. A low index of suspicion is necessary to make the diagnosis in late cases. Most early cases will resolve with anticoagulation, interventional radiology, or shunt surgery. Orthotopic liver transplant is reserved for those cases that have progressed to cirrhosis. In regions where only living donor transplant is feasible, the surgery tends to be more technically challenging as the liver graft does not come with the IVC.6 Synthetic grafts for caval replacement are possible but generally avoided because of the risk of thrombosis in patients with predisposition. Direct anastomosis to the right atrium may be possible on the assumption that caval obstruction does not need correction on account of collateral development. In this study, we have shared our experiences of living donor liver transplant (LDLT) for BCS without caval replacement.
Materials and Methods
Between November 2006 and March 2020, 2952 LDLTs were performed at the Centre for Liver and Biliary Sciences (New Delhi, India). Of these, 36 LDLTs were undertaken for BCS. These 36 cases were analyzed for demographic profiles, predisposing factors, pretransplant interventions, operative techniques, and perioperative events. The need for long-term anticoagulation was also studied. Graft and patient survival rates were also calculated.
All patients received livers from living related donors: 29 were from first-degree relatives or spouses (6 fathers, 2 mothers, 3 sons, 3 daughters, 9 brothers, 5 sisters, 1 wife) and 7 were from second-degree relatives (4 cousins, 2 nephews/nieces, 1 uncle). Local ethics committee approval was not required because this was a retrospective clinical study.
Thrombophilia screening tests (protein C and S deficiencies, factor V Leiden mutation, antithrombin 3 deficiency, JAK2 mutation, MHTFR gene mutation) were conducted in all recipients and in donors if they were first-degree relatives. Liver biopsy was carried out if the indication for transplant was not clear. Computed tomography scans of patients were reviewed to assess site and extent of block.
The following types of BCS were seen on computed tomography imaging: type I (stenosed hepatic veins alone), type II (hepatic vein stenosis plus caval stenosis; Figure 1), type III (obliterated cava above the level of renal vein), and type IV (caval obliteration/narrowing extends below renal vein ostia; Figure 2).
After recipient hepatectomy, the cava was cross clamped just above the level of renal veins and in the suprahepatic area. The cava was circumferentially dissected and opened at multiple points, depending on the type of block, and repaired.
For type I, the cava was opened at the level of hepatic vein insertion and the opening enlarged. For type II, the cava was opened by a long cavotomy, thrombectomy was done, and the loss of caval diameter was replaced with a vein patch/PTFE patch (Figure 3). For type III, IVC replacement was done with autologous graft or anastomosis to suprahepatic cava. For type IV, patients underwent combined surgery and interventional radiology (Figure 4).
Postoperatively, all patients were started on anticoagulation with low-molecular-weight heparin, which was later switched to oral anticoagulation before discharge from the hospital. During follow-up, international normalized ratio was monitored, which was managed to remain between 2.0 and 3.
Our study included 36 patients (20 males and 16 females), with age ranging from 3 to 57 years (mean and SD of 27.70 ± 16.45 years). The most common clinical symptoms were ascites (n = 25) followed by gastrointestinal bleed (n = 10) and jaundice (n = 7). Hepatopulmonary syndrome (HPS) was seen in 4 patients. Observed mean sodium Model for End-Stage Liver Disease score was 13.6 ± 4.6.
No etiologic factors could be found in 26 patients. Among remaining patients, JAK2 mutation was found in 3 patients, isolated protein C deficiency in 2 patients, combined protein C and S deficiencies in 2 patients, factor V Leiden and protein C deficiency in 1 patient, MTHFR mutation and protein C deficiency in 1 patient, and hepatic cyst in 1 patient.
Six patients had pretransplant interventions, 3 underwent transjugular intrahepatic portosystemic shunt, 2 had interventional radiology-guided hepatic vein stenting, and 1 had IVC stenting. Two patients underwent a second procedure before transplant (2 shunt procedures).
In 5 patients, diagnosis was made at time of surgery. Hepatopulmonary syndrome was the reason for transplant in 4 cases. Eight patients had complete portal vein thrombosis, which precluded any interventional procedure. The remaining patients who were considered for transplant either had progressed to advanced cirrhosis (n = 8) or had severe IVC obstruction that the interventional radiologist felt could not be recanalized (n = 2); in 3 patients, there was suspicion of hepatocellular carcinoma.
Different surgical procedures
Type II BCS was the most common variety and was seen in 19 patients (52.7%). The second most common was type I, in which the ostia of hepatic veins were dilated; this type was seen in 15 patients (41.6%). Type III BCS with obliterated IVC limited to suprarenal was seen in 1 patient (2.7%); treatment required IVC segmental replacement with autologous graft. Type IV was also seen in 1 patient (2.7%), with IVC obliteration extending below renal veins. Inferior vena cava dilation with thrombectomy and patchplasty was done for the suprarenal segment. During the postoperative period, the patient had persistent bilateral lower limb swelling, and this was recanalized by interventional radiology. The patient had complete resolution of lower limb swelling after the procedure. Eight patients had complete portal vein thrombosis, which required eversion thrombectomy.
All patients underwent LDLT. Twenty-four patients received a modified right lobe graft, 4 patients had left lobe graft, and 5 patients received left lateral lobe grafts. An extended right lobe was used in 2 patients, and a right posterior sector was used in 1 patient. Mean graft weight for right lobe grafts was 761.7 ± 87.72 g, with minimum graft weight of 445 g and maximum graft weight of 1119 g. Mean graft-to-recipient weight ratio was 1.41 ± 0.62. For right lobe grafts, mean graft-to-recipient weight ratio was 1.2 ± 0.616, with minimum 0.6 and maximum 3.63. Mean cold ischemia time was 94.47 ± 43.34 minutes, and mean warm ischemic time was 29.26 ± 9.21 minutes. Middle vein reconstruction was done in all modified right lobe grafts. Autologous portal vein grafts were used in 20 patients; 3 patients required deceased donor vein grafts, and recanalized umbilical vein grafts were used in 2 patients. Venoplasty for multiple V5 and V8 veins in close proximity was done for 3 patients. Inferior right hepatic vein (IRHV) reconstruction was done in 10 patients, with IRHV directly implanted over IVC in 7 patients. Three grafts had multiple IRHVs; in these patients, 2-to-1 venous cuffs were created, which were joined to the IVC.
Postoperative complications were seen in 7 patients. One patient had postoperative bleeding that was conservatively treated with blood products. Reexploration with lavage for suspicion of sepsis was done in 2 patients, with 1 patient having a biliary leak. Postoperative portal vein thrombosis was seen in 2 patients, with 1 patient requiring exploration with portal vein thrombectomy and the other treated with interventional radiology-guided portal vein thrombolysis and stenting. Reexploration with splenic artery ligation was done in 1 patient because of poor graft function. One patient required interventional radiology-guided hepatic venous pressure gradient with right hepatic vein (RHV) and middle hepatic vein (MHV) ostial dilatation for persistent high drain output.
Posttransplant ascites was assessed in the form of days required for removal of abdominal drain. In 11 patients, the abdominal drain was removed by day 10 posttransplant, whereas, in 15 patients, it was removed after posttransplant day 10. Eight patients were discharged with drain in situ (1 patient had chylous ascites); their abdominal drains were removed during outpatient follow-up.
Four patients died posttransplant (2 from primary graft dysfunction, 1 from acute respiratory distress syndrome after thrombotic thrombocytopenic purpura, and 1 from sepsis). Two patients died during follow-up for reasons not related to recurrence of BCS. In 2 patients lost to follow-up, there was no evidence of disease recurrence at the last follow-up. Median survival in our study was 36 months. The longest survival was 158 months (Figure 5).
Our study highlighted the fact that BCS is an uncommon indication for liver transplant, as we had only 36/2952 cases of the disease in our LDLT series. Our interventional radiology department, which is a busy department, could only refer 6 cases that had failure of interventional radiology procedures. We prefer LDLT over interventional radiology procedures in cases where the portal vein is completely thrombosed, which was seen in 8 of our cases. There was a higher incidence of portal vein thrombosis in our series, which may have been the result of delayed diagnosis and intervention in Central and East Asia populations. Similarly, in those with extensive IVC blockage, LDLT was preferred as it gave an opportunity to clear the block, and interventional radiology procedures may have been difficult or may have required multiple attempts.
Our report is one of few large series of LDLT reported for BCS. Deceased donor liver transplant has an advantage of allowing caval replacement and therefore is preferable. However, in many parts of the world, LDLT is the only option; as shown in our series, LDLT was possible without caval replacement, with caval replacement only needed in 1 case. In that case, although we replaced the cava, it would have been possible to anastomose to the suprahepatic cava as collaterals were draining the lower cava and renal veins.
In our case series, which mainly included patients from Asia, an identifiable thrombotic disorder could be found only in 10 of 36 patients (27%). In 26 patients (73%), no underlying thrombotic condition was shown. Even in those with portal vein thrombosis, the incidence of procoagulant state was low, suggesting that portal vein thrombosis was also a result of low flow state. Protein C deficiency was the most common prothrombotic abnormality, in 16.2% (6 patients), followed by JAK 2 mutation in 8% (3 patients). As described by Okuda,7 an obstructive hepatocavopathy with primary membranous obstruction of IVC was the cause of BCS in 60% of the study’s Asian population group, which is similar to our own experience. Multiple hepatic cysts with previous surgeries were the cause of secondary BCS in 1 patient. In contrast, an inherent or acquired thrombotic disorder is the main causative factor for BCS in the Western world, with factor V Leiden, primary myeloproliferative disorders, and protein C deficiency accounting for 25%, 20%, and 25%, respectively, of causes for hepatic vein thrombosis.8,9 Classical presentation of BCS depends on the stage of presentation. Features of abdominal pain and hepatomegaly may not be clinically evident in patients with chronic BCS with established collaterals to decompress liver.10 The most common presentation observed in our series were ascites (67.5%) and gastrointestinal bleed (27%); jaundice was noted in only 19% of cases.
Hepatopulmonary syndrome has been reported with a prevalence of 5% to 29% in cirrhotic patients.11 Occurrence of HPS in BCS has been studied by de Binay and colleagues.12 The group showed a higher occurrence of HPS in patients with cirrhotic portal hypertension secondary to BCS. They also showed about a 58% incidence of positive bubble contrast echocardiography, with 27.6% having clinical features and arterial blood-gas analysis suggestive of HPS. Our study showed a prevalence of HPS in 10.8% of cases, with most belonging to the pediatric age group (age <15 years), making the prevalence of HPS in up to 40% of pediatric patients with BCS (4 of 10 patients). Although interventional radiology may be successful in restoration of blood flow, there seems to be a higher incidence of late HPS in this group, which then becomes an absolute indication for liver transplant.
Angioplasty with or without stenting and transjugular intrahepatic portosystemic shunt have shown promising results with long-term patency rates at 5 years of 88% and 74%, respectively.13,14 In our series, of 6 patients who failed interventional radiology, 2 patients subsequently underwent a surgical shunt procedure. Mean duration from intervention to requirement for liver transplant was 1.13 ± 0.8 years (minimum to maximum of 0-2 years), suggesting that interventional radiology procedures are probably the wrong option in this subgroup. If cirrhosis is advanced, the best option may be liver transplant at the outset. Because our interventional radiology and liver transplant teams work in tandem in our center, this is rarely a problem for us.
All patients in our study received LDLTs. In right lobe grafts, all veins were anastomosed to separate openings (RHV and MHV to respective opening). A separate opening on IVC was made for IRHV implantation. The mean cold ischemia time was found to be relatively higher in modified right lobe grafts (104 ± 41.10 min) compared with other grafts (74 ± 34.56 min). This difference can be attributed to time required in benching for MHV reconstruction in modified right lobe grafts. Similarly, warm ischemic time was slightly higher in grafts with IRHVs compared with no IRHV (38.25 ± 6.14 min vs 27.56 ± 6.33 min). No difference in graft outcome was seen among these groups.
In LDLT, patients with associated IVC disease require additional IVC preparation to provide optimal graft outflow. Yamada and associates15 described dissection of stenotic IVC with application of venous patch to enlarge the luminal caliber of IVC. Inferior vena cava thrombectomy with patchplasty was successful in 19 of 37 patients (51.3%), whereas only 1 patient (2.7%) required segmental replacement graft for IVC. Our approach in dealing with IVC thrombus differed from what was shown by Choi and associates16 and Akamatsu and associates,17 where either interposition IVC graft or replacement of native IVC was done. However, our experience showed that a long cavotomy to achieve adequate thrombectomy with venous patch to dilate the IVC provides optimal luminal caliber for graft outflow. In patients with long segment IVC obstruction extending up to below the renal vein, lower limb edema may persist despite a successful liver transplant, owing to poor venous outflow from lower limbs. Such cases can be best managed with combined surgery followed by interventional radiology-guided staged dilation of stenotic IVC.
Immediate postoperative mortality in our case series was 11%. Median follow-up was 3 years, with minimum of 2 months and maximum of 13 years. Two patients died for reasons not related to recurrence of BCS after 4 and 66 months of transplant. We did not observe any long-term recurrence of primary disease in our study.
Liver transplant could be curative for patients in whom the primary pathology resides in liver, such as those with protein C and protein S deficiencies.18 However, identifying 1 etiological abnormality does not rule out the presence of others.19 Therefore, anticoagulation was continued in all transplant recipients in our series. Cruz and colleagues20 demonstrated high incidence of recurrence in transplant recipients of BCS and also noticed much higher bleeding incidences related to anticoagulation. Their study group included patients mainly from Western populations with inherent thrombotic disorders. Another large study of 248 patients from Europe did not show similar results, and recurrence rates were quite low (<5%).21 Monitoring of international normalized ratio was done at regular intervals in addition to other follow-up investigations such as complete blood counts, liver function test, renal function test, and immunosuppression levels every 3 months, as well as lipid profile, stress echocardiography, and electrocardiography annually. Long-term follow-up did not show any major clinical events, suggesting complications were related to anticoagulation.
The early mortality rate of liver transplant for BCS-related cirrhosis has been shown to range from 4% to 21%, and the 1- and 5-year survival rates have been shown to range from 80% to 95% and 65% to 95%, respectively.17 This was acceptable in comparison to survival rates for other diseases requiring liver transplant.17 Although other studies evaluated and compared deceased donor liver transplant for BCS, we saw similar results in our study. However, 5-year survival could not be calculated, as all patients in our study group had not completed their 5 years posttransplant. In our study, the survival probability of the population at the time of this study plateaued around 78% at the end of 5.5 years.
With increasing modalities to treat early BCS and further delaying the progression of disease, requirements for liver transplant for this cause of cirrhosis is expected to fall over time. However, in countries like India and other Asian countries, which have limited deceased donor pools, LDLT can be an acceptable alternative, even in the presence of profound involvement of IVC. Surgical techniques to reconstruct venous outflow in thrombosed IVC and combination of surgery with interventional radiology techniques have shown encouraging results.
Volume : 19
Issue : 8
Pages : 799 - 805
DOI : 10.6002/ect.2020.0541
From the Centre for Liver and Biliary Sciences, Max Super Speciality Hospital, Saket, New Delhi, India
Acknowledgements: The authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no declarations of potential conflicts of interest.
Corresponding author: Yuktansh Pandey, Centre for Liver and Biliary Sciences, Max Super Speciality Hospital, Saket, New Delhi, India
Phone: +91 9630922894
Figure 1. Coronal Computed Tomography Image of Type 2 Budd-Chiari Syndrome Showing Obliterated Hepatic Inferior Vena Cava
Figure 2. Axial Magnetic Resonance Image of Type 4 Budd-Chiari Syndrome Showing Completely Obliterated Supra- and Infrarenal Inferior Vena Cava
Figure 3. Inferior Vena Cava Reinforcement with PTFE Graft Patchplasty for Type 2 Budd-Chiari Syndrome
Figure 4. Angiogram Showing Obstruction Above Suprarenal Inferior Vena Cava
Figure 5. Survival Probability of the Entire Population, Plateauing at >78% After About 5.5 Years