The Budd–Chiari syndrome is a loose term used to designate disorders characterized by hepatic venous outflow obstruction. The pathological features of hepatic vein thrombosis were first described by Budd in 1845. In 1899, Chiari reported the clinical syndrome of hepatomegaly, ascites, and abdominal pain, and coined the eponym Budd–Chiari syndrome. The clinical features and treatment of the disorder depend on the site of obstruction. Hepatic venous outflow obstruction can result from thrombotic or non-thrombotic occlusion of the major hepatic veins or the inferior vena cava. Occlusion at either site interrupts the blood flow from the liver, leading to a marked elevation of sinusoidal pressure, intense congestion of the liver, and destruction of hepatic parenchyma.

Although the Budd–Chiari syndrome is rare in the United States, it is common in Northern India, South Africa, and the Orient. The syndrome can occur at any age, but generally presents in the third or fourth decade of life and is slightly more common in women. Budd–Chiari syndrome can present either as an acute, rapidly evolving illness, or as a disorder that progresses slowly and insidiously over a period of months to years. The majority of patients have rapidly developing liver disease, with signs and symptoms of less that 3 months' duration: the classic triad is the sudden appearance of ascites, hepatomegaly, and abdominal pain. Other common symptoms such as nausea, anorexia, vomiting, and diarrhoea are fairly non-specific. Clinical jaundice is rare, and splenomegaly is found in fewer than one-third of patients. Rarely, patients may present with fulminant hepatic failure and shock. In patients with long-standing Budd–Chiari syndrome, portal hypertension develops, and patients may present with progressive ascites and wasting, bleeding oesophageal varices, hepatic encephalopathy, and hepatorenal syndrome.

The Budd-Chiari syndrome most often occurs as a result of an underlying illness known to be associated with thrombotic complications, although there may be no apparent predisposing factor in up to one-third of cases (Table 1) 362. In large compilations of patients from the United States, the most common predisposing conditions are myeloproliferative disorders, polycythaemia rubra vera, and paroxysmal nocturnal haemoglobinuria. Polycythaemia rubra vera is the single most common cause of the Budd–Chiari syndrome in the United States. Many patients at initial presentation of the condition have a normal haemoglobin and haematocrit but a markedly elevated red cell mass. Studies have suggested that patients with paroxysmal nocturnal haemoglobinuria may have a progressive disease, with early thrombosis of only small hepatic veins and minimal or no symptoms, followed later by complete occlusion of large hepatic veins and a life-threatening course. Other myeloproliferative disorders such as essential thrombocythaemia, agnogenic myeloid metaplasia, and chronic myelogenous leukaemia have also been reported in association with the Budd–Chiari syndrome.

A number of tumours in and around the liver, particularly hepatocellular, adrenal, and renal carcinomas, but also Wilms' tumour, leiomyosarcoma of the inferior vena cava, right atrial myxoma, and carcinoma of the lung, pancreas, and stomach, may cause the Budd–Chiari syndrome either by direct invasion of the inferior vena cava or by increasing the tendency toward thrombosis.

Budd–Chiari syndrome may develop during pregnancy or in the postpartum state, within a few weeks or months after delivery. Many cases occur in women taking oral contraception. Although a definite causative link has not been shown, a multicentre case-control study has suggested that the relative risk of hepatic vein thrombosis among recent users of oral contraceptives compared with that of non-users was 2.37 (p<0.02), a risk similar to that of stroke, myocardial infarction, or venous thromboembolism.

A number of infectious processes may be associated with the Budd–Chiari syndrome, the most common being amoebic abscesses, aspergillosis, hydatid cyst, and syphilis. Trauma, most commonly blunt abdominal trauma, is also a common cause of Budd–Chiari syndrome. Hypercoagulable states that have been reported in patients with Budd–Chiari syndrome include antithrombin III deficiency, protein C deficiency, and the presence of lupus anticoagulant (antiphospholipid antibody). Most of the connective tissue disorders, including systemic lupus erythematosus, mixed connective tissue disorder, and scleroderma have been reported as possible causes. Rare causes of the Budd–Chiari syndrome include Behçet's syndrome, ulcerative colitis, &agr;&sub1;-antitrypsin deficiency, sarcoidosis, and idiopathic hypereosinophilic syndrome.

In the Orient, the Middle East, and South Africa, a membranous web or a long segmental obstruction of the inferior vena cava is the most common cause of the Budd–Chiari syndrome. The origin and pathogenesis of these webs remains controversial. Hirooka and Kimura conducted a detailed study of 205 patients with Budd–Chiari syndrome reported in the Japanese literature, and found that a membranous web accounted for 73 (35 per cent) of cases. They classified the membranes into seven types, based on embryological studies, concluding that the webs resulted from a developmental anomaly or congenital malformation. In 1986, however, a patient had a mural thrombus that evolved over 2 years into a membranous web, suggesting that membranous webs can be a sequel of thrombosis. Although membranous webs have been considered rare in the United States, a series of 35 patients with Budd–Chiari syndrome, out of whom 8 patients (23 per cent) had idiopathic membranous obstruction of the inferior vena cava was reported from Los Angeles; most of these were immigrants from the Far East.

The cause of the Budd–Chiari syndrome is unknown in approximately 30 per cent of cases. However, a latent myeloproliferative disorder, despite the absence of peripheral blood changes, may be demonstrated in a majority of these patients by the spontaneous formation of erythroid colonies in the absence of erythropoietin.

Many disorders can be mistaken for the Budd-Chiari syndrome and should be considered in the differential diagnosis. Liver disorders with features similar to the Budd–Chiari syndrome include alcoholic hepatitis, viral hepatitis, cirrhosis, and hepatic malignancy. Disorders outside the liver that resemble the Budd–Chiari syndrome include constrictive pericarditis, tricuspid insufficiency, right atrial myxoma, and right-sided heart failure. All of these conditions, however, rarely present a diagnostic problem. Hepatic veno-occlusive disorder, on the other hand, mimics the Budd–Chiari syndrome and may be part of the same disease spectrum. Veno-occlusive disorder is characterized by non-thrombotic obstruction at the level of the sublobular and terminal branches of the hepatic veins in the liver. Although initially described as a complication of poisoning with pyrrolizidine alkaloids present in Jamaican bush teas, the disease is now seen as a complication of high dose chemotherapy and radiation therapy in patients receiving bone marrow transplantation.

The diagnosis is established by demonstrating radiographically the patency of the major hepatic veins and inferior vena cava, and/or by liver biopsy.

Budd–Chiari syndrome is generally difficult to diagnose on the basis of clinical criteria alone, and routine laboratory testing is not particularly helpful. The aminotransferases are usually only slightly elevated and are of little diagnostic value. These enzymes may reach concentrations greater than 1000 IU in acute, severe Budd–Chiari syndrome or when hepatic vein thrombosis occurs in association with thrombosis of the portal vein. Serum alkaline phosphatase activity is usually elevated by a factor of 1.5 to 2; the serum bilirubin level is usually less than 5 mg/dl. The characteristics of the ascites fluid are highly variable: ascites fluid protein usually ranges from 1.5 to 3.0 g/dl, although exudative ascites with a protein concentration of greater than 3.0 g/dl is not uncommon.

Because the physical findings and routine liver chemistry tests are non-specific in suspected Budd–Chiari syndrome, imaging studies play an important role in the diagnostic evaluation. Hepatic scintigraphy has been used extensively in the past; some patients show decreased uptake in the left and right lobes, with retained uptake in the caudate lobe, resulting in the so-called ‘central hot spot.’ More often, however, hepatic scintiscanning shows a diffuse, inhomogeneous pattern of tracer uptake, and other imaging modalities, such as ultrasonography, computed tomography, and magnetic resonance imaging, are therefore used. All three of these techniques can reveal non-specific abnormalities such as hepatomegaly, ascites, inhomogeneity of the liver parenchyma, as well as the more specific findings of absence of hepatic veins and ‘common-shaped collaterals.’ Magnetic resonance imaging has certain advantages in evaluating the hepatic vasculature because of its depiction of blood vessels as regions of absent signal, the ‘flow-void phenomenon’, but the technique is expensive and not widely available. Compared with other modalities, real-time Doppler duplex ultrasound has the advantage of low cost and easy availability and at present is probably the screening test of choice. In addition to showing thrombosis or occlusion of the major hepatic veins or inferior vena cava, Doppler ultrasound can often demonstrate the actual level of obstruction. Moreover, recent studies suggest that colour-flow Doppler ultrasound is superior to duplex imaging and can give a real-time overview of all vascular structures within the liver. Both colour-flow Doppler ultrasound and magnetic resonance imaging (MRI) are useful in evaluating shunt patency in patients with the Budd–Chiari syndrome who have undergone portasystemic shunting; both compare favourably with venography. In the future, colour-flow Doppler imaging may become the initial screening technique in patients with a suspected diagnosis of Budd–Chiari syndrome.

For most patients, venography and liver biopsy remain the standards for diagnosis. Inferior vena cavography, performed by introducing a catheter into the femoral vein, can demonstrate the site of obstruction as well as provide access for pressure measurements. The venogram can show occlusion or marked narrowing of the major hepatic veins; characteristically, there is difficulty in catheterizing the hepatic veins in patients with Budd–Chiari syndrome, but the catheter can often be advanced a short distance, and injection of dye in the wedged position may reveal a spider web pattern of intrahepatic collaterals. The inferior vena cava may be shown to be patent, or occluded by tumour, thrombus or web, or compressed and narrowed by an enlarged caudate lobe. Pressure measurements should be taken in the right atrium and the inferior vena cava to determine whether a pressure gradient is present across the diaphragm. Portal venography, from superior mesenteric arteriography with examination of the venous phase, should also be performed if portal vein thrombosis is suspected. Concomitant portal vein thrombosis occurs in up to 20 per cent of patients with hepatic vein thrombosis.

The diagnosis of Budd–Chiari syndrome is supported by a percutaneous needle biopsy of the liver. The classic pathological findings include sinusoidal dilation and centrilobular congestion, followed later in the disease course by centrilobular necrosis, cell dropout and atrophy, scant inflammation, and extravasation of red blood cells into the space of Disse. Eventually, in patients with long-standing Budd–Chiari syndrome, fibrosis of a ‘nutmeg’ liver develops. Blind percutaneous liver biopsy is often difficult or dangerous because of the presence of ascites or coagulopathy; however, ultrasound or CT directed biopsy can be performed with relative safety in most patients.

The suspicion of the Budd–Chiari syndrome demands rapid and aggressive diagnosis. Patients should undergo urgent evaluation, with non-invasive imaging followed by venography. Some authorities urge prompt liver biopsy examination in patients with suspected Budd–Chiari syndrome in order to assess the degree of hepatic injury and to identify patients with Rappaport Zone III necrosis who may require immediate surgery. However, the diagnosis can be established with venography alone in most patients.

The natural history of the Budd–Chiari syndrome is variable, depending on the rapidity and extent of the hepatic venous outflow occlusion. For the majority of patients who do not undergo surgical or anigiographic intervention, however, the prognosis is poor; relentless deterioration and death ensue. Although spontaneous resolution is described, these cases are rare and unlikely to occur except in the very early stages of the disease. On the other hand, some patients with mild disease may be clinically stable for years. These patients often respond to diuretics for a while, especially if an underlying hypercoagulable state (e.g. myeloproliferative disorder) can be treated as well. A small proportion of patients with very early hepatic vein thrombosis respond to transluminal angioplasty or thrombolytic therapy. The majority of patients, however, if treated non-operatively die within months of the onset of symptoms. Survivors of the acute phase almost invariably develop cirrhosis of the liver and die within a few years from hepatic failure, bleeding oesophageal varices, or other complications of chronic liver disease. The overall mortality rate exceeds 50 per cent at 2 years.

The therapeutic approach to patients with the Budd–Chiari syndrome depends on the severity, duration, and cause of the obstruction. Patients with suspected Budd–Chiari syndrome should undergo an urgent evaluation, and those with severe, acute disease or a liver biopsy that shows severe Zone III necrosis or fibrosis should immediately be considered for surgical intervention. Medical therapy is useless in patients with severe disease and in those in whom hepatic vein thrombosis is fully established. However, patients with mild, non-acute Budd–Chiari syndrome, and those in whom a liver biopsy shows no necrosis or fibrosis, can be considered for a trial of medical therapy and close observation.

Diuretic therapy is sometimes useful in relieving the intense ascites (Table 2) 363. Although anticoagulant therapy is probably not effective in promoting clot resolution, it may be useful in preventing further clot formation. The underlying disorder should be treated: for example, patients with a chronic myeloproliferative disorder can be treated with hydroxyurea or &agr;-interferon. Women taking oral contraceptives in whom the Budd–Chiari syndrome develops often do well with simple discontinuation of their medication.

In patients with a possible reversible component to their thrombosis, a trial of a thrombolytic agent is worthwhile. Many patients with acute hepatic vein or inferior vena cava thrombosis are treatable with streptokinase or urokinase. Thrombolytic therapy, however, is probably effective only if given within 72 h of clot formation; it is not useful in patients with established obstruction. Furthermore, thrombolytic therapy carries a risk of severe haemorrhagic complications.

Another medical alternative is the use of percutaneous transluminal angioplasty. Initially, Fogarty balloon catheters were employed in this technique, which involves pulling the inflated balloon through the web. The Gruentzig catheter is now the device of choice. Balloon angioplasty is especially useful for relieving membranous obstruction of the inferior vena cava or hepatic veins and is probably the treatment of choice in this setting. Transluminal angioplasty has also been used to treat hepatic vein thrombosis, although there may be a high rate of reocclusion.

Surgery plays an important role in the treatment of the Budd– Chiari syndrome. Most uncontrolled studies suggest that surgically treated patients do better than those treated medically. The liver damage in Budd–Chiari syndrome results from pressure necrosis secondary to a marked elevation in sinusoidal pressure. Most surgical approaches to reverse the blood flow in the portal vein, transforming the portal vein into an outflow tract for the congested liver. Portal decompression reduces sinusoidal pressure, prevents further injury to the hepatocyte, and allows liver regeneration. Serial liver biopsies following portasystemic shunts show complete or nearly complete return to normal liver architecture and histology; however, the reversibility of the hepatic damage is related directly to the extent and duration of hepatic venous outflow obstruction. Therefore, early surgical decompression would be expected to result in more successful reversal of parenchymal abnormalities.

If the patient has severe, acute Budd–Chiari syndrome and/or the liver biopsy reveals significant centrilobular necrosis, but not severe fibrosis, the patient is a good candidate for an immediate portasystemic shunt. The exact approach depends on the results of angiography and the status of the inferior vena cava. Pressure measurements should be made in the right atrium and the suprahepatic and infrahepatic inferior vena cava. In the absence of caval obstruction, the traditional surgical approach has been a side-to-side portacaval shunt or one of its variants, such as a mesocaval shunt or mesocaval H graft (Table 2) 363. These shunt operations are very effective and carry a low operative mortality. Side-to-side portacaval shunt for hepatic vein thrombosis effectively lowers portal pressure, relieves ascites, returns liver chemistry to normal, relieves hepatosplenomegaly, improves liver architecture, and improves long-term survival. Portasystemic encephalopathy is rarely a complication. The mesocaval shunt and the mesocaval H graft are haemodynamically equivalent but may be associated with a higher incidence of thrombosis or occlusion.

Patients with the Budd–Chiari syndrome may suffer obstruction of the inferior vena cava by thrombosis, tumour or web, or compression by an enlarged caudate lobe. Portal decompression in these patients must be performed by means other than the standard portacaval shunt. As discussed earlier, patients with isolated membranous obstruction of the inferior vena cava can be treated with percutaneous transluminal angioplasty or, failing that, can undergo a surgical transcardiac membranotomy. For those with hepatic vein thrombosis and infrahepatic occlusion of the inferior vena cava, a mesoatrial shunt is the therapy of choice. This technique, first described by Cameron and Maddrey in 1978, involves anastomosis of a long Dacron or Gore-Tex graft to the superior mesenteric vein, which is tunnel led anterior to the liver and into the chest (bypassing the obstructed inferior vena cava, and its anastomosed to the right atrium. Results with the mesoatrial shunt have been encouraging but, over the long term, a 25 to 30 per cent occlusion rate occurs. Warren and others have suggested a two-stage approach for obstruction of the inferior vena cava resulting from compression by a hypertrophied caudate lobe. Patients first undergo a mesoatrial shunt for hepatic and caval decompression; several months later, the patients undergo a second operation, in which the mesoatrial shunt is taken down and a definitive side-to-side portacaval shunt is performed.

Another alternative is the retrohepatic inferior vena cavoplasty or reconstruction and side-to-side portocaval shunt; this is also useful in the setting of recurrent thrombosis of a mesoatrial shunt. A portoatrial shunt is another alternative to the mesoatrial shunt and is useful in patients with Budd–Chiari syndrome with an obstructed inferior vena cava and superior mesenteric vein. Other reported variations on the standard shunt operation include the cavosplenoatrial shunt, and transposition of the spleen into the pleural cavity (to create a portapulmonary shunt). Efficacy of the splenopleural shunt is sporadic.

The peritoneoatrial (LeVeen) shunt does little to alter the underlying pathophysiology of the ascitic syndrome. Its use is limited to patients who are not candidates for either a portasystemic shunt or hepatic transplantation.

For patients with the Budd–Chiari syndrome and hepatic decompensation, liver transplantation is the treatment of choice. A number of patients with advanced hepatic failure resulting from the Budd–Chiari syndrome have undergone transplantation. Early results with orthotopic transplantation of the liver indicated 1- and 3-year survival rates of 71 per cent and 54 per cent respectively; however, five patients have had recurrence of thrombosis in the hepatic veins or the portal vein of the hepatic graft. More recent attempts at hepatic transplantation for treatment of the Budd–Chiari syndrome have included early anticoagulant therapy by using low-dose heparin, followed by coumadin. A 3-year survival of 88 per cent has been observed. The presence of a chronic myeloproliferative disorder is not believed to be a contraindication to liver transplantation, as long as there is adequate therapy for the underlying condition (e.g. hydroxyurea).

A major area of controversy is the question of shunt surgery in patients with Budd–Chiari syndrome who may need a liver transplant later. For patients who have end-stage liver failure orthotopic liver transplantation is clearly required. For patients who have less severe fibrosis and liver damage, the decision is less clear, because liver transplantation is technically difficult and the operative mortality is higher after a portacaval shunt. Practically, however, most patients with early Budd–Chiari syndrome respond well to a portasystemic shunt, and this is therefore the surgical treatment of choice. Patients treated with surgical decompression and those treated non-operatively should be followed closely with serial biochemical studies and non-invasive vascular imaging. Patients who have undergone portasystemic shunting and who have progressive disease should be considered for transplantation (Table 3) 364. Likewise, patients who have been followed closely on medical therapy who show signs of progressive hepatic vein occlusion or centrilobular necrosis should undergo portasystemic shunting.

Campbell DA, et al. Hepatic transplantation with perioperative and long-term anticoagulation as treatment for Budd–Chiari syndrome. Surg Obstet Gynecol 1988; 166:511–17.

Grant EG, Perella, R, Tessler FN, Lois J, Busuttil R, Budd–Chiari syndrome: the results of duplex and color doppler imaging. Am J Roentgenol 1989; 152: 377–81.

Henderson JM, et al. Surgical options, hematologic evaluation, and pathologic changes in Budd–Chiari syndrome. Am J Surg 1990; 159: 41–50.

Hirooka M, Kimura C. Membranous obstruction of the hepatic portion of the inferior vena cava. Arch Surg 1970; 100: 656–63.

Maddrey WC. Hepatic vein thrombosis (Budd–Chiari syndrome): possible association with the use of oral contraceptives. Semin Liver Dis 1987; 7: 32–8.

Millikan WJ, et al. Approach to the spectrum of Budd–Chiari syndrome: which patients require portal decompression? Am J Surg 1985; 149: 167–76.

Mitchell MC, Boitnott JK, Kaufman S, Cameron JL, Maddrey WC. Budd– Chiari syndrome: etiology, diagnosis, and management. Medicine 1982; 61: 199–218.

Murphy FB, Steinberg HV, Shires, GT, Martin LG, Bernardino ME. The Budd–Chiari syndrome: a review. Am J. Roentgenol 1986; 147: 9–15.

Orloff MJ, Girard B. Long term results of treatment of Budd–Chiari syndrome by side to side portacaval shunt. Surg Obstet Gynecol 1989; 168: 33–41.

Sparano J, Chang, J, Trasi S, Bonanno C. Treatment of the Budd–Chiari syndrome with percutaneous transluminal angioplasty. Case report and review of the literature. Am J Med 1987; 82: 821–8.

Terabayashi H, Okuda K, Nomura F, Ohnishi K, Wong P. Transformation of the inferior vena cava thrombosis to membranous obstruction in a patient with the lupus anticoagulant. Gastroenterology 1986; 91: 219–24.

Valla D, et al. Primary myeloproliferative disorder and hepatic vein thrombosis. A prospective study or erythroid colony formation in vitro in 20 patients with Budd–Chiari syndrome. Ann Intern Med 1985; 103: 329–34.

Valla D, Le MG, Poynard T, Zucman N, Rueff B, Benhamou JP. Risk of hepatic vein thrombosis in relation to recent oral contraceptives: a case-control study. Gastroenterology 1986; 90: 807–11.

Warren WD, Potts JR, Fulenwider JT, Millikan WJ, Henderson JM. Two stage surgical management of the Budd–Chiari syndrome associated with obstruction of the inferior vena cava. Surg Obstet Gynecol 1984; 159: 101–7.