Article Info

Author Affiliations: Institute of Hepatobiliary and Pancreatic Diseases, Department of Hepatobiliary and Pancreatic Surgery (Sun YL, Fu Y, Zhou L and Ma XX), Department of Digestive Medicine (Zhou L), Department of Radiological Intervention (Wang ZW), Department of Radiology (Wu Y), First Affiliated Hospital, Zhengzhou University, School of Medicine, Zhengzhou 450052, China

Corresponding Author: Yu-Ling Sun, MD, PhD, Institute of Hepatobiliary and Pancreatic Diseases, Department of Hepatobiliary and Pancreatic Surgery, First Affiliated Hospital, Zhengzhou University, School of Medicine, 1 Jianshe Road, Zhengzhou 450052, China (Tel: 86-371-67697128; Fax: 86-371-66964927; Email: ylsun@zzu.edu.cn)

Contributors: SYL proposed the study. SYL and FY performed research and wrote the first draft. ZL and MXX collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. SYL is the guarantor.

BACKGROUND:Collateralized intra- and extra-hepatic routes in patients with Budd-Chiari syndrome (BCS) were important. This study aimed to investigate the feasibility and clinical outcomes of the staged management of BCS based on the degree of compensation provided by intra- or extra-hepatic collateral circulations.

METHODS:A total of 103 adult patients with BCS caused by co-obstruction of the inferior vena cava (IVC) and main hepatic veins (MHVs) between March 2001 and October 2009 were enrolled in this study. Based on the pathological classification and degree of hemodynamic compensation by collateral circulations, treatment priority for IVC hypertension was determined in the first-stage treatment. Patients were deemed eligible for second-stage treatment when the first-stage treatment failed to relieve.

RESULTS: Imaging results revealed that most patients had collateral circulations to different extents. Based on the degree of compensation provided by these collateral circulations, 74 patients underwent single-stage treatment for IVC hypertension, i.e., radiologic intervention (RI) for 61 patients and surgical procedures (SPs) for 13. One patient was treated for portal hypertension. Twenty-nine patients underwent second-stage treatment (25 underwent RI and SP, and 4 only SP). The general morbidity and mortality after all procedures were 8.3% and 1.5%, respectively. After a median follow-up of 35 months, 4 patients underwent second-stage treatment and 7 underwent recanalization of the IVC/MHVs. Two patients died of hepatocellular carcinoma and 1 died of graft obstruction.

Budd-Chiari syndrome (BCS) is caused by the obstruction of the main hepatic veins (MHVs) and/or inferior vena cava (IVC), and can present as portal hypertension (PHT) and IVC hypertension.^[1] The spontaneous mortality of BCS approaches 70% in 1 year and 90% in 3 years.^[2] Therefore, different therapeutic modalities such as surgical shunting and bypass procedures have been proposed.^[3-7] Surgical decompression for PHT included cavoatrial, mesoatrial, mesojugular, and splenoatrial shunting. However, these surgical procedures (SPs) have a series of shortcomings, such as severe complications, limited long-term patency, unfavorable perioperative mortality and long-term outcomes.^{[4, 8]} Therefore, management strategies for these cases need to be improved.

During the course of BCS, the chronic blockage of the MHVs increases the sinusoidal pressure which leads to fibrosis, nodular regenerative hyperplasia, and cirrhosis.^[9] The formation of collateral circulation is the main characteristic of imaging results.^{[10, 11]} The compensated circulation improves liver function and silences BCS for many years and thus, renders the patients asymptomatic. The established intra- and extra-hepatic collateral circulations have great impact on choosing the treatment of BCS.^[12-14] However, collateral circulation is now only for diagnostic purpose, its effect of management remains unclear. Hence, we designed a novel algorithm for BCS treatment. This algorithm was based on collateral circulation and other clinical data available at the time of diagnosis.

BCS treatment generally follows a least invasive to most invasive strategy.^[15] Accordingly, a staged management strategy was applied in patients with BCS caused by IVC and MHV obstruction. In these patients, PHT can be caused by IVC obstruction. Therefore, we hypothesized that sufficient collateral circulations can compensate well for clinical manifestations, and that the response to the first-stage treatment for IVC hypertension was pivotal in deciding the next-stage therapeutic procedure for PHT.

A total of 103 adult patients with BCS admitted in our unit from March 2001 to October 2009 were enrolled in this study. The onset of the disease was chronic and the disease course was at least 7 months. The criteria for inclusion were as follows: 1) IVC and MHV obstruction was confirmed and liver diseases were ruled out; 2) the patients had PHT and IVC hypertension; 3) patients' portal veins were patent and there was no malignant tumor; 4) acute variceal hemorrhage was stopped either spontaneously or by medication; 5) patients were younger than 65-year-old; 6) there were no life-threatening comorbidities and other chronic diseases (i.e., tumors, diabetes, or cardiovascular and respiratory diseases); and 7) patients agreed to return for follow-up.

The indications for invasive treatment were as follows: 1) obvious clinical manifestations which included bleeding, abdominal distension, jaundice, signs of the low extremities, anemia, distended abdominal veins, infertility, etc.; 2) biochemical abnormality (high serum bilirubin and alanine aminotransferase, hypoproteinemia, low hemoglobin, etc.); 3) persistent manifestations or unimproved biochemical test results (change in the Child-Pugh grade) after at least 12 weeks of first-stage therapy. BCS was considered asymptomatic and had no indication for invasive treatment when there was no evidence of the aforementioned conditions. Patients without symptoms of IVC hypertension and PHT were also considered as completely compensated by intra- or extra-hepatic collateral circulations.

To investigate the pathologic changes in the BCS patients, we collected liver samples during operation or by liver biopsy during the follow-up period. Due to the inhomogeneous distribution of the lesions (congestion, necrosis, fibrosis, and cirrhosis of the liver),^{[9, 16]} two tissue samples were collected by biopsy from both lobes of the liver not only during operation but also in the follow-up in some patients. HE staining was performed and the following characteristics were evaluated: ischemic necrosis and fibrosis which was graded on a scale from 0 to 4.^[17]

The therapeutic strategy was based on the clinical data of collateral circulations, manifestations, and biochemical tests. In general, the indications for invasive treatment were as follows: 1) identification of intra- and/or extra-hepatic collateral circulations; 2) obvious clinical presentation; 3) biochemical abnormality; 4) persistent manifestations or unimproved biochemical test results after at least 12 weeks of first-stage therapy. In other words, two-stage invasive treatment was not indicated for patients who did not have apparent manifestation and who had imaging results demonstrating sufficient collateral circulations.

The treatments included SP and radiologic intervention (RI). The selection of surgical or endovascular intervention was determined according to our previous study (Table 1).^[4] For patients who needed RI, percutaneous hepatic venography including balloon angioplasty with or without stenting was also performed. At the same time, the pressures of the IVC and free hepatic vein were measured. Fig. 1 shows our working algorithm for the investigation and management of the patients.

The therapeutic strategy for the 103 patients was as follows. Priority treatment (first-stage treatment) was determined for IVC hypertension based on prospective assessments of clinical data (especially for the patients with rich portal-systemic collateral circulations). After the first-stage procedure, all patients were subjected to anticoagulation therapy which included low-molecular-weight heparin or warfarin targeting an international normalized ratio (INR) of 2:3. Then, the anticoagulant was replaced with aspirin within 2 weeks when the patients were discharged. Two-stage management was based on the evaluation of the clinical and biochemical parameters after the first-stage treatment. Second-stage treatment was considered necessary for all patients whose first-stage treatment failed to alleviate PHT or improve biochemical parameters. The manifestation of patients can be significantly improved if the first-stage treatment is effective. Therefore, second-stage treatment was performed at least 12 weeks after the first-stage treatment.

The patients were followed up after the first-stage treatment. Each follow-up included physical examination, biochemical tests, and Doppler US. For those whose disease course was difficult to determine, double-contrast upper gastrointestinal radiography and/or electronic endoscopy for esophageal varicose, catheterization for pressure measurement, as well as angiographic studies of IVC were performed. Liver biopsy was carried out whenever feasible to help determine further treatment.

Quantitative variables were expressed as mean±SD, or median and range. Data in normal distribution were analyzed using unpaired, two-tailed t test, whereas non-normally distributed data were evaluated by the Mann-Whitney U test. Categorical variables were compared by the Chi-square or Fisher's exact test. A P value of <0.05 was considered statistically significant.

Of the 103 patients, 36 were female and 67 were male (median age 35 years, range 19-64). The course of the disease ranged from 7 months to 24 years. Refractory ascites was present in 24 patients. Bleeding from esophageal varices occurred before treatment in 11 patients. At the time of diagnosis, 15, 68, and 20 patients were graded as A, B, and C under the Child-Pugh grading system, respectively. The detailed data of the patients are shown in Table 2.

BCS was diagnosed when Doppler US, 3D-CTA, or venography showed obstructed MHVs and IVC. Gross liver changes were present in all patients, as characterized by hypertrophy of segment I associated with variable atrophy of one lobe and hypertrophy of another lobe. A mottled appearance related to inhomogeneous perfusion of the liver was noted by 3D-CTA. Images clearly depicted the pathologic classifications of the IVC and MHVs (Table 3).

In most patients, intrahepatic collateral circulations were identified either by a typical "spider web" sign or by large connecting veins between the accessory and MHVs (Fig. 2A, B). Except for 4 patients with IIIb, 87 (84%) displayed intrahepatic collateral circulations to different extents. Among all patients, 45 (44%) showed widened accessory hepatic veins; 27 of the 45 patients had inferior right hepatic veins and 4 had small hepatic veins.

Almost all patients demonstrated extra-hepatic collateral circulations to different extents. Various extra-hepatic collateral circulations were found in the abdominal wall as well as in the peritoneal and retroperitoneal areas. Prominent azygos and hemiazygos veins were the most commonly collateralized extrahepatic routes, which were demonstrated in 67 (65%) patients (Fig. 2C). Deep and central tributaries of the systemic circulation (i.e., ascending lumber veins and vertebral venous plexus) were the most frequently collateralized routes (Fig. 2D, E). Superficial collateral circulations of the abdominal wall rose from the external iliac vein and anatomized with the lateral thoracic vein (Fig. 2F). The infrequent renal-inferior phrenic-pericardiophrenic collateral circulations were displayed in four patients. 3D-CTA imaging revealed the anatomic relationship among the blocked hepatic vein, IVC, and collateral circulations from different orientations invasively between the accessory hepatic vein and IVC. Extra-hepatic collateral circulations were also identified from the liver capsule to the heart in five patients.

Liver parenchymal congestion, necrosis, sinusoidal dilatation, and hemorrhage were identified in 17 patients. Moderate to severe necrosis on initial tissue specimens was noted in 23 patients. Heterogeneously distributed fibrosis was observed in 72 patients. Based on the global grading evaluation, 3, 45, 6, and 18 patients can be classified into grades 1, 2, 3, and 4, respectively. Otherwise, in these patients, the fibrosis alternated between regional congestion and necrosis.

Seventy-four patients underwent one-stage treatment, and among them 53 had balloon angioplasty because of IVC membranous and short-segment obstruction. Eight patients with long-segment obstruction of the IVC underwent simultaneous balloon angioplasty and stenting. Three patients with IVC thrombosis underwent radical resection. Among 9 patients with long-segment IVC obstruction, 3 underwent cava-caval bypass, 4 caval-atrial bypass, and 2 splenectomy and spleno-atrial shunting. Meso-caval shunting was performed in one patient with long-segment IVC obstruction because of complete compensation by the extrahepatic collateral circulations.

Two-stage treatment was given to 29 BCS patients. 1) RI was the first-stage treatment and SP was the second-stage treatment. Eight patients with IIIa BCS underwent balloon angioplasty of the IVC with or without stenting for the first-stage treatment. Two months later, these patients had meso-caval shunting for the second-stage treatment. 2) SP was the first-stage treatment and RI was the second-stage treatment. Seventeen patients underwent cava-caval or caval-atrial bypass for the first-stage treatment, followed by percutaneous revascularization for complete MHV occlusion in the second-stage treatment because of persistent ascites. Of these patients, 6 experienced caval-atrial bypass and underwent balloon angioplasty as well as metal stent implantation 9 weeks later. The others only underwent balloon angioplasty for MHV obstruction. 3) SP was the treatment for both stages. Four patients with IIIb had incomplete compensation by the third hepatic pedicle. After caval-atrial bypass for the first-stage treatment, portal hypertension was still persistent and the moderate volume of ascites was measured after three months. Therefore, meso-caval shunting was performed in the second-stage treatment.

All patients developed collateral circulations to some extent. Besides the aforementioned collateral circulations, the peri-esophageal veins, sub-capsular veins, and large accessory hepatic veins were identified more often in the first-stage patients than in the second-stage patients. We believed that this phenomenon completely compensated for PHT after IVC recanalization. For these patients, laparotomic surgery was not needed for PHT, not to mention cavoatrial, mesoatrial, mesojugular, and splenoatrial shunting.

The complications after the procedures are shown in Table 5. The general morbidity and mortality after all procedures were 8.3% and 1.5%, respectively. One patient required emergency operation for thoracic hemorrhage after radical resection. Two patients died from liver failure and multiple organ dysfunction syndromes. Other patients were cured conservatively. The patients who had received the one-stage treatment only had significantly lower rates of complications and mortality compared with those who had received two-stage treatment (6.8% vs 12.1%, P<0.01; 1.4% vs 3.4%, P<0.01). The differences might be due to more surgeries in the two-stage treatment group.

In the patients who underwent only the one-stage treatment, liver function improved significantly after 12 weeks (P<0.05). Considering the incomplete compensation of the collateral circulations, the liver function did not substantially improve in the patients who required two-stage treatment (data not shown). After invasive treatment, the IVC and free hepatic pressure were decreased significantly (32.1±7.7 vs 12.4±3.6 mmHg, P<0.01).

The follow-up in 7, 16, 25, 24, and 31 patients was 7 to 8, 5 to 7, 3 to 5, 2 to 3, and less than 2 years, respectively. The median follow-up was 35 months (range 4-96). Four patients died during the follow-up: two patients died of HCC, one died of graft obstruction and another died from traffic accident. Four patients underwent the second-stage treatment with porto-systemic shunt because the disease was aggravated after the first-stage treatment. Seven patients experienced angioplasty and/or stenting because of recurrent obstruction. Apart from these patients, all patients survived with stable condition and hypertension of the IVC and portal vein was significantly decreased. Liver biopsy in 39 patients showed that 3 to 6 years after operation, the liver had mild to moderate fibrosis, and that there was no congestion or necrosis. Fibrotic grade was significantly improved after operations compared with that before the procedures in patients with lower grade of liver fibrosis (P<0.05).

BCS is a rare and life-threatening disease caused by obstruction of hepatic venous outflow. The prognosis of these patients has markedly improved in recent decades.^{[5, 15, 18-20]} However, the management has not been standardized. In the present study, we first identified intra- and extra-hepatic routes of collateral circulations. Based on the compensated situation by these collateral circulations, we then proposed staged management. Our strategies demonstrated fewer complications and lower mortality in the perioperative period. The results indicated the effectiveness of our treatment strategy in the management of BCS caused by IVC and MHV obstruction.

The natural disease course of BCS is poorly understood because most patients have received some form of treatment.^[5] Traditionally, early treatment is suggested as soon as the diagnosis is established.^[7] However, the slow obstruction of two or three major veins produces a chronic presentation and develops extensive collateral circulations.^{[16, 18]} For patients with asymptomatic conditions, therapy may not be necessary.^[21] In the present study, the development of intra- and extra-hepatic collateral circulations were also identified.

The management of BCS with IVC and MHV obstruction is quite difficult. In past clinical practice, direct decompression of PHT was the treatment of choice. SP, such as mesoatrial, mesojugular, and splenoatrial, was performed. Effective short-term outcomes were obtained, but the perioperative mortality and morbidity were very high.^{[4, 7, 8, 22]} In the present study, for the first time, the treatment priority for IVC hypertension was determined. We generally solved IVC hypertension, and changes in the hemodynamics and biochemical tests were assessed to determine whether second-stage treatment was needed for PHT. Considering the compensation by collateral circulations, if PHT was alleviated to some extent after IVC recanalization, no further treatment was needed. Our new algorithm effectively avoided the traditional surgery for PHT and achieved significant long-term clinical outcomes.

For BCS patients with IVC and MHV obstruction, assessing the compensation of PHT by collateral circulation is difficult. Rich collateral circulations are only observed in a few BCS patients, and the diameter of each collateral vein is distinct. The blood flow velocity in collateral circulation is difficult to quantify. The volume of collateral circulations that shunt the venous blood to the systemic veins cannot be measured accurately. Therefore, in the present study, parameters such as the response of manifestations to the first-stage treatment and change in liver function were used to assess the necessity of the second-stage treatment for PHT. Clinical manifestations may be subtle, and liver biopsy usually provides nonspecific information on BCS patients.^[23] Therefore, the histological lesions observed in liver biopsy specimens are not good prognostic indicators.^[16] To facilitate a complete evaluation of patients who needed further treatment, we used both liver biopsy and clinical manifestations to assess the patient condition.

Percutaneous angioplasty is considered a safe and effective interventional procedure for patients with IVC obstruction, and is also used in short-segment hepatic vein stenosis.^{[19, 24]} In this study, 5 of 7 patients with recurrent obstructions in the IVC and/or MHVs showed long-segment obstructions, which confirmed our conclusion that this procedure is indicated for short-segmental obstruction.^[4] For these patients, angioplasty plus stenting may lead to better patency than meso-atrial shunting or other procedures above the diaphragm. Complications could be found in 2% of patients, which is much lower than that in patients undergoing traditional surgical procedures.^{[4, 6]}

The relationship between clinical presentation and hepatic parenchymal or hepatic venous lesions is not straightforward. On anticoagulation therapy alone, some patients with acute obstruction may recover spontaneously, at least partially, as evidenced by rapidly decreased serum aminotransferase, disappearance or easy control of ascites, and improved liver function.^[25] However, aggravation may occur unpredictably manifested as refractory ascites, wasting, recurrent gastrointestinal bleeding, or development of liver failure. Hence, our staged management allowed us to follow up the patients closely. Four patients underwent second-stage treatment after the first-stage treatment in the follow-up period.

Although encouraging outcomes were achieved using the staged treatment strategy, this study had some limitations. First, a detailed comparison between ultrasound imaging and X-ray angiography was not performed because not all patients underwent hepatic venography or inferior vena cavography. However, the imaging results demonstrated various findings of BCS, which provided clues not only to the correct diagnosis but also to the invasive treatment. Second, we cannot obtain IVC pressure because of ethical reasons. Third, due to the difficulties in determining the blood quantity of collateral circulations, we had to assess the compensative situation of the patient by their symptoms and laboratory tests. Accordingly, the pressure measurements of the IVC and MHVs are needed to guide further invasive procedures.

8 Zhang Y, Zhao H, Yan D, Xue H, Lau WY. Superior mesenteric vein-caval-right atrium Y shunt for treatment of Budd-Chiari syndrome with obstruction to the inferior vena cava and the hepatic veins--a study of 62 patients. J Surg Res 2011;169:e93-99. PMID: 21529832

19 Li T, Zhai S, Pang Z, Ma X, Cao H, Bai W, et al. Feasibility and midterm outcomes of percutaneous transhepatic balloon angioplasty for symptomatic Budd-Chiari syndrome secondary to hepatic venous obstruction. J Vasc Surg 2009; 50:1079-1084. PMID: 19703746