BMJ Case Reports 2009; doi:10.1136/bcr.07.2009.2057
  • Unusual presentation of more common disease/injury

Rapid development of a hepatocellular carcinoma in isolated thrombosis of hepatic veins (classic Budd–Chiari syndrome): case report and review of literature

  1. Jens Walldorf1,
  2. Andrea Tannapfel2,
  3. Hans Jürgen Holzhausen3,
  4. Christian Wittekind4,
  5. Thomas Seufferlein1,
  6. Utz Settmacher5,
  7. Wolfgang E Fleig6,
  8. Matthias M Dollinger1
  1. 1
    University of Halle-Wittenberg, Department of Internal Medicine I, Ernst-Grube-Strasse 20, Halle, 06120, Germany
  2. 2
    Berufsgenossenschaftliche Universitätsklinik Bergmannsheil, Institute of Pathology, Bürkle-de-la-Camp-Platz 1, Bochum, 44789, Germany
  3. 3
    University of Halle-Wittenberg, Institute for Pathology, Magdeburger Str. 14, Halle, 06112, Germany
  4. 4
    University of Leipzig, Institute for Pathology, Liebigstrasse 26, Leipzig, 04103, Germany
  5. 5
    Friedrich-Schiller-University, Department of General and Visceral Surgery, Erlanger Allee 101, Jena, 07747, Germany
  6. 6
    University of Leipzig Hospitals and Clinics, Philipp-Rosenthal-Str. 27, Leipzig, 04103, Germany
  1. Matthias M Dollinger, matthias.dollinger{at}
  • Published 29 November 2009


Budd–Chiari syndrome and membranous obstruction of the inferior vena cava frequently result in the development of mostly benign hepatic lesions. In cases of membranous obstruction of the inferior vena cava, which is prevalent mostly in the East, these lesions often progress to hepatocellular carcinoma. In contrast, malignant transformation has not yet been recognised in patients with isolated hepatic vein thrombosis. We report the case of a 37-year-old male Caucasian who presented with acute Budd–Chiari syndrome without involvement of the inferior vena cava. Despite porto-caval shunting, a hepatocellular carcinoma developed within several months. Three hepatic lesions were treated by radiofrequency thermal ablation until liver transplantation was performed. This report emphasises the possibility of malignant transformation of regenerative nodules in patients with disturbed hepatic perfusion in general. Physicians must be aware of this when assessing regenerative nodules, especially as no unambiguous predictors for the development of hepatocellular carcinoma have been identified so far.


The common definition of Budd–Chiari syndrome (BCS) includes both isolated thrombosis of the hepatic veins (classic BCS) and the (membranous) obstruction of the inferior vena cava (MOVC),1,2 despite distinct aetiologic, epidemiologic and prognostic features.35 Only MOVC regularly gives rise to malignancy—that is, hepatocellular carcinoma (HCC). MOVC occurs predominantly in black South Africans and Japanese, rarely in western countries.68 In contrast, development of HCC in classic BCS is a singular event, except for cases with longstanding cirrhosis caused by thrombosis of the hepatic veins9 with an incidence of HCC similar to other types of chronic liver disease.10

We report the case of a 37-year-old male Caucasian with classic BCS followed by the rapid development of liver cirrhosis and HCC. The case will be discussed with particular reference to actual developments in the knowledge regarding the differences between classic BCS and MOVC and their potential to cause malignant transformation.

Case presentation

A 37-year-old male Caucasian presented in 2004 with increasing abdominal girth, mild epigastric pain and painless jaundice, which had developed within 1 week. After 2 weeks in another hospital he was transferred to our unit for evaluation for liver transplantation because of suspected alcoholic cirrhosis. Before the onset of the symptoms the patient felt well and healthy, and he was not taking any medication. His past medical history was unremarkable, in particular regarding previous icteric episodes or hepatic disorders; he denied smoking and past or current alcohol abuse. The family history revealed no thrombotic, hepatic or haematological disease or malignancy.

Besides jaundice, the physical examination revealed ascites. The liver edge was palpable 8 cm below the costal arch. The results of relevant routine laboratory studies are shown in table 1. Extended laboratory studies (including antiphospholipid antibodies, lupus anticoagulant, antinuclear antibodies (ANA), antimitochondrial antibodies (AMA), smooth muscle antibodies (SMA), liver membrane antibodies (LMA), hepatitis (A,B,C) serology, and FACS analyses for paroxysmal nocturnal haemoglobinuria) demonstrated only a protein C and S deficiency.

Table 1

Relevant laboratory parameters at admission to our unit

Oesophago-gastroduodenoscopy revealed grade 3 oesophageal varices and gastro-oesophageal varices (Sarin type I). Ultrasound examination showed an enlarged liver (16 cm cranio-caudal, 17.3 cm sagittal diameter) with irregular surface, but homogenous parenchyma without focal lesions. Bile ducts were normal. The spleen measured 11.8×4 cm. Generalised moderate ascites was detected. In Duplex ultrasound, hepatic veins were not detectable, while the liver arteries appeared normal. Low intermittent hepatopetal flow (maximal 6 cm/s) was detected in the portal veins. The diagnosis of BCS without involvement of the inferior vena cava was confirmed by magnetic resonance imaging (fig 1).

Figure 1

Initial magnetic resonance image (MRI) showing an enlarged liver without focal lesions, but inhomogeneous and delayed contrast medium uptake. No thrombosis or other flow obstruction of the vena cava is evident, the hepatic veins are not visible (as depicted, angio-MR). Splenic and portal veins are perfused with a delayed contrast medium uptake. Moderate ascites is present.

A transjugular intrahepatic portosystemic shunt (TIPS) could not be deployed because the angiography catheter could not be passed into the liver veins. A surgical portocaval side-to-side interposition shunt (8 mm polytetrafluoroethylene (PTFE) prosthesis) was applied successfully, as confirmed by Duplex sonography and MRI. Intraoperatively, liver and bone marrow biopsies were performed. Liver histology is presented in fig 2. Bone marrow histology ruled out a myelodysplastic syndrome or monoclonal proliferation of plasma cells.

Figure 2

Haematoxylin and eosin stain of the initial liver biopsy: advanced pericentral necrosis with modest cholestasis and mild inflammatory infiltration; beginning of portal fibrosis; regenerative hyperplasia of hepatocytes in periportal areas.

After shunt surgery, the patient recovered quickly and his liver function improved. During follow-up, abdominal ultrasound examinations and laboratory tests were performed every 4 months. At 13 months after the initial diagnosis, ultrasound revealed four solid tumours of the liver. α-Fetoprotein (AFP) was not increased (2.3 ng/ml). MRI showed several circular lesions of the liver up to 28 mm in diameter. Three of these lesions appeared to be compatible with the diagnosis of HCC (fig 3A). One of the suspicious lesions was biopsied under ultrasound guidance. A well differentiated HCC (G1) was diagnosed at histology (fig 4). No expression of AFP was detected immunohistologically, and CK8/18 was focally positive.

Figure 3

Thirteen months after surgical portocaval shunting, T1 weighted MRI showing multiple hyperintense nodules with homogenous enhancement following contrast medium (panel A). Three of these nodules were also hyperintense in T2 weighted echo planar imaging (EPI-MRI, panel B), thus supporting the diagnosis of hepatocellular carcinoma. Panel C (computed tomography) and panel D (MRI) display the result after successful radiofrequency thermal ablation.

Figure 4

Ultrasound guided biopsy of one of three suspicious lesions: liver fibrosis stage 3–4 with formations of pseudolobuli; glandular structures as well as trabecular formations with several cell layers compatible with well differentiated hepatocellular carcinoma (G1).


The patient was evaluated for liver transplantation and placed on the waiting list at Eurotransplant, Leiden. To bridge the time to transplantation, three lesions were treated by two sessions of ultrasound guided percutaneous radiofrequency ablation (RFA). MRI and computed tomography scans after RFA showed necrotic areas compatible with the treated lesions (fig 3B).

Respecting the patient’s demand for additional treatment and considering its potentially beneficial anti-angiogenetic effect, the patient was treated with thalidomide in doses increasing up to 200 mg daily.11,12 Treatment was well tolerated without any side effects.

Twenty-two months after the initial diagnosis of BCS and 7 months after diagnosing HCC, orthotopic liver transplantation was performed. During the operation, the portocaval shunt appeared open. Histology of the explanted liver showed complete cirrhosis and multiple macroregenerative nodes. Necrotic liver tissue was found at the RFA sites, but no further malignant lesion could be found in the explant. A further course was complicated by primary graft failure. Re-transplantation was performed 10 weeks after initial liver transplantation and the patient recovered well.

Outcome and follow-up

Three years after re-transplantation, no recurrence of the malignant tumour was detected on regular follow-up. Liver function and perfusion are normal, as is the AFP value (3.9 ng/ml). The patient is well and in a good general condition.


Aetiology of BCS

In about 50% of cases of BCS primary myeloproliferative disorders are found; other frequent causes are antiphospholipid syndrome, protein C deficiency, factor V Leiden mutation, and hyper-homocysteinaemia.13 In MOVC, a pathogenic role of local infections, phlebothrombosis or endothelial microlesions through respiratory movements of the diaphragm is discussed.5,14 Interestingly, in a very recent publication by Eapen et al15 the incidence of MOVC is reported to decrease. At present, the most common type of hepatic outflow obstruction seems to be an obstruction of both the inferior vena cava and hepatic vein. These changes are most likely related to improvements in hygiene and sanitation.

In our patient, the low protein C and S in the presence of normal factor II levels are suggestive of a primary deficiency even considering the impaired hepatic protein synthesis in the acute setting.16 Due to the multiple mutations responsible for protein C and S deficiency,17 genotyping is not practical. Thus, a definitive classification of the aetiology in the presented case is not possible.

Liver biopsy is not effective in diagnosing BCS in every case, but can provide information on the possible underlying primary disease.18 The histological patterns of the parenchymal response to outflow obstruction vary from veno-portal cirrhosis to pericentral necrosis, depending on the affected vessels.18,19 The biopsy in our case, taken 8 weeks after the onset of symptoms, correlated with the clinically diagnosed subacute form of BCS with occlusion of the hepatic veins (fig 2).

In our case, it might be speculated that the thrombotic event could represent the result of malignant disease rather than its cause. In fact, HCC is complicated by thrombotic events in advanced, large tumours, which compress or invade vascular structures, typically the portal or hepatic veins. The most frequent paraneoplastic syndromes of HCC are hypercholesterolaemia (incidence in HCC patients 11.4%), hypoglycaemia (2.8%), hypercalcaemia (1.8%), thrombocytosis (2.7%), and erythrocytosis (2.5%).2022 The presence of a paraneoplastic syndrome predicts a poor response to therapeutic efforts and poor survival.20 These syndromes characteristically are found in patients with high serum AFP concentrations, and large tumour volume (>10 cm diameter, >30% of the liver volume or bi-lobal tumour involvement and metastasis).20,21 In our patient, no suspicious lesions of the liver were seen on ultrasound or MRI, both initially and during 1 year follow-up. AFP was completely normal and there was no evidence for metastatic disease. The tumour responded well to RFA and, thereafter, in the explanted liver, no evidence of remaining malignant hepatocellular lesions was found histologically. Thus, the hypothesis of a pre-existing HCC as the reason for hepatic vein thrombosis seems very remote.

Treatment of BCS

The main treatment objective in BCS is to avoid the development of cirrhosis and its complications by decompression of the liver.23 Porto-systemic shunting is performed in >90% of patients with BCS, either by surgical shunting,24 percutaneous angioplasty25 or by TIPS.26,27 In the present case, however, surgical shunting did not prevent the progress of mild fibrosis to complete cirrhosis.

Liver transplantation is an alternative to surgical shunting or may be necessary after failure of liver decompression by surgical shunt or TIPS.28 The 10 year survival rate of transplanted patients is about 70% with a low risk of recurrence of thrombotic events, when adequate long term anticoagulation is initiated early on after transplantation.29,30

Hepatic lesions in BCS in MOVC

Hepatic lesions are diagnosed in 40% of patients with BCS. Histologically, they are characterised as large regenerative nodules or nodular hyperplasia-like focal lesions.18

These lesions result from focal hyper-regeneration and develop independently from the grade of fibrosis.31 Varying grades of focal regeneration result from inhomogeneous perfusion of the liver in BCS.32 Venous outflow obstruction leads to an early decrease of portal flow, resulting in nodular regenerative hyperplasia or, if the perfusion is obstructed by large thrombi, infarcts of liver tissue. In chronic courses of BCS, additional portal vein thrombosis may occur in >50% of cases.17 Since low portal venous perfusion is compensated by increased arterial flow, large regenerative nodules of a rather focal nodular hyperplasia (FNH)-like type are observed.19

In European countries these lesions are generally regarded as benign. They may lead to gross changes in the size and shape of the liver, but not to HCC. In the East and on the Indian subcontinent, obstruction of the inferior vena cava is observed with high frequency.5,33 It regularly results in changes of hepatic venous outflow similar to classic BCS. Therefore, this entity is often included in the term Budd–Chiari syndrome. Combined obstruction of the inferior vena cava and the hepatic veins leads to the subsequent development of HCC in up to 47% of cases,6,8,33 especially in patients with longstanding obstruction.34 In a study from Turkey on BCS as a complication of Behçet’s disease,35,36 three out of 56 patients with hepatic vein thrombosis developed HCC. All three of these patients had combined thrombosis of the inferior vena cava and the hepatic veins as typical for Behçet’s disease.37

In a very recent report on the development of HCC on the background of BCS, including inferior vena cava obstruction, Moucari et al had followed a cohort of 97 patients with mostly Caucasian (69%) or African Jamaican background (22%) for at least 12 months.10 An obstructed inferior vena cava was described in 13% of cases. Roughly 60% of these patients were Caucasian, although cavla obstruction is considered a rare disease in this ethnic group. While this study confirms a high risk to develop HCC in patients with obstruction of the inferior vena cava (nine out of 11 patients were diagnosed with HCC), it remains unknown whether caval obstruction was the cause or rather the consequence of HCC.

This different risk of HCC in isolated hepatic vein thrombosis as compared to combined hepatic and caval vein obstruction has not been fully understood. Obviously, obstruction of the inferior vena cava is a more severe disease. In patients with MOVC, thrombosis of the hepatic veins is a rule5 and might lead to more pronounced liver damage, finally resulting in cirrhosis and HCC. However, in their cohort of 97 patients, Moucari et al10 reported an African Jamaican who had developed HCC in a non-fibrotic liver within 2 years of being diagnosed with classic BCS, and a Caucasian with obstruction of the inferior vena cava without hepatic fibrosis, who had presented with HCC at the time of initial diagnosis. The absence of fibrosis in these patients underlines the implication of impaired hepatic blood flow in the development of HCC.

Another explanation stresses the epidemiological differences between the two entities: in countries with a high incidence of MOVC, other causes of malignant liver disease are frequent including viral hepatitis, aflatoxine exposition, poor sanitation, poor nutrition and bacterial infections. Therefore, a superimposed exposition to carcinogenic influences might be assumed.5,38 However, no association between hepatitis B infection and HCC was detected in patients with BCS by Okuda et al.7 In our case, exogenous factors for the development of HCC could not be identified.

In conclusion, the development of HCC on the background of isolated hepatic vein thrombosis is a rare event, which is mostly related to longstanding cirrhosis caused by unresolved hepatic outflow obstruction. In the present case of BCS, the rapid progression to cirrhosis and HCC despite successful portocaval shunting is noteworthy. To the best of our knowledge, only one similar case with an HCC developing on the background of an isolated hepatic vein thrombosis and no fibrosis at the time of diagnosis of BCS has been reported in the literature.10 While the causal relationship between BCS and HCC remained largely speculative in previous case reports, in our case and the case of Moucari et al,10 a close follow up documented the course from acute BCS to HCC within several months.

In Eastern countries, the development of HCC in patients with membranous obstruction of the inferior vena cava is widely recognised,39 while in guidelines developed by western groups, this complication is mentioned in a footnote at best.2,40 With increasing global mobility, diseases and complications formerly restricted to “remote” countries—such as the development of HCC in MOVC—have to be added onto our list of diagnostic considerations.

Malignant transformation of regenerative nodules may also occur against the background of BCS in general. Although this might be a unique one-of-a-kind case, it calls for careful assessment of any hepatic lesions detected in such patients. In addition, it might be speculated that changes in hepatic perfusion in BCS including de novo angiogenesis could be relevant for the pathogenesis of HCC35,41,42

Learning points

  • Careful assessment of any hepatic lesions detected in the course of BCS is required in all patients.

  • Development of HCC in patients with membranous obstruction of the inferior vena cava is widely recognised.

  • Malignant transformation of regenerative nodules may occur against the background of isolated thrombosis of hepatic veins (classic BCS)


We thank Prof. R. P. Spielmann and Prof. C. Behrmann, University Clinic and Polyclinic for Diagnostic Radiology, Martin-Luther-University Halle-Wittenberg, for the excellent assistance in interpreting the CT- and MRI-images and the generous provision of appropriate copies for this report.


  • Competing interests: none.

  • Patient consent: Patient/guardian consent was obtained for publication


No Related Web Pages

Register for free content

The full text of all Editor's Choice articles and summaries of every article are free without registration

The full text of Images in ... articles are free to registered users

Only fellows can access the full text of case reports (apart from Editor's Choice) - become a fellow today, or encourage your institution to, so that together we can grow and develop this resource

Don't forget to sign up for content alerts so you keep up to date with all the case reports as they are published, and let us know what you think by commenting on the Editor's blog