Glisson's capsule is to the liver what the tunica albuginea is to the testis and the dura mater is to the brain: an enveloping, tough, collagenous sac that maintains the form of the organ, and perhaps some elements of function, but which strangulates its contents if they expand.

Litres of hepatic artery and portal vein blood normally flood the hepatic sinusoids and fill the hepatic veins on their way out of the liver to the inferior vena cava just below the diaphragm. In the Budd–Chiari syndrome (Fig. 1) 1214, when the normal outflow pathways are lost ( Figs 2 1215, 3 1216), hepatic compliance decreases. So does the flow of blood, frequently leading to self-perpetuating ischaemia of hepatic tissue and to cell death. In severe cases the entire centrilobular architecture of the portal unit collapses, causing what is called zone III necrosis.

Dogs, raccoons, and seals normally have smooth muscle around the exits of their hepatic veins. Minor degrees of hypoxia and other deviations from homeostasis cause these veins to be throttled by the smooth-muscle ‘valves’, which stop the egress of blood and often provoke a fracture of the liver. These peculiarities of the species explain why so little written about the canine liver in shock has any relevance to the human situation.

The most frequent idiopathic cause of the Budd–Chiari syndrome is clotting within the lumen of the hepatic veins (Table 1) 365. Therefore speculation is rife that loss of natural anticoagulants might be causal. In fact, three natural anticoagulants are present in human liver: antithrombin III, protein C, and protein S. Antithrombin III is so potent that 1 ml is enough to antagonize all the thrombin in the circulating blood ( Tables 2 366 and Table 3 367). Cigarette smoking and the use of oestrogens, which are known aetiological agents of the Budd–Chiari syndrome, antagonize antithrombin III rapidly.

In its most virulent form, the victim of acute Budd–Chiari syndrome, usually a woman, is seized with abdominal pain, followed by abdominal swelling with bloody ascites. Hepatomegaly is detectable if the abdomen is lax. Figure 1 1214 and Table 1 365 describe the causes of the syndrome.

As soon as the diagnosis is suspected, and provided that renal function is good enough to sustain the injection of angiographic contrast media, angiographic studies of the major veins at risk must be done to delineate the anatomy of the disease and to estimate whether blood clots are present and whether or not they might be extractable or dissolvable ( Figs 2 1215, 3 1216). A liver biopsy specimen must be secured as the major indicator of the state of parenchymal integrity. In the assessment of blood flow, Doppler duplex ultrasound scanning and magnetic resonance data may eventually be as accurate as angiographic examinations, but without their attendant risks.

Pressure measurements within the veins are interesting, but are not essential, except to determine whether or not the pressure in the inferior vena cava is too high to prevent its use for decompression. Actually, measurements of portal venous and hepatic venous pressures tend to be specious because of artefacts induced by the volume of ascites and the mass of the swollen liver pressing on the inferior vena cava in a recumbent patient. One might have suspected that the narrowed inferior vena cava in Fig. 4 1217 would be associated with a high intraluminal pressure; in fact, the pressure was low.

Although the best time to see and treat a patient with Budd–Chiari syndrome is as soon as the diagnosis is entertained, a number of medical specialists generally consider the patient before a surgeon is called. The first judgement to be made is whether the disease is in an early stage, and, if so, whether it is likely to respond to thrombolytic therapy or to physical extraction of the clots (if that is the cause of the problem) by surgical or radiological means.

By the time the surgeon operates, the task is unenviable. For fear of immunological disaster, haematologists sometimes forbid the use of fresh frozen plasma, even if it is the only means of remedying intraoperative hypocoagulability, and massive doses of methylprednisolone are often given to counteract the immunological ravages of thrombogenic diseases (such as paroxysmal nocturnal haemoglobinaemia), making normal tissue fragile and diaphanous. On the other hand, high-protein ascites within the peritoneal cavity may be transformed into a thick peel adherent to the intestines and enveloping them in a shell of fibrous tissue called a small-bowel cocoon. Small-bowel obstruction is to be expected, and enteric leaks during surgery are common because of the inability to distinguish enveloping adhesions from the normal small-bowel wall.

Although a peritoneal–atrial prosthetic shunt of plastic tubing from the peritoneal cavity to the superior vena cava (LeVeen or Denver shunt) (Fig. 5) 1218 may relieve the massive ascites accompanying the Budd–Chiari syndrome, the proteinaceous fluid within the tubing has a natural propensity to clot, making long-term patency of the shunt problematic.

On occasion, surgical thrombectomy is successful. Lysis of clots with streptokinase or urokinase, by either systemic administration or direct instillation into the affected veins, may be successful but is unpredictable. Radiologically guided dilatation of the stenosed hepatic veins is sometimes worthwhile. There is no body of experience to gauge the comparative worth of lysis and radiological dilatation against that of the surgeon.

If these ‘lesser’ techniques are inappropriate or fail to relieve intravenous obstruction, three other approaches are possible:

1.Converting the portal vein or another splanchnic vein into an outflow tract for the liver;

2.Fracturing a membrane in the inferior vena cava, if one exists;

3.Transplanting a new liver.

Obstructive membranes in the inferior vena cava above the orifices of the hepatic veins are a genetic or acquired disease of Oriental populations; once a membrane is recognized, it is easily broken with a finger or a dilator inserted into the inferior vena cava (Fig. 6) 1219,1220. An expandable metal stent inserted into the vena cava under radiological guidance might help to assure patency of the fractured membrane.

But for the shortage of donor livers and the prudent caution exercised because of the problems that could face a transplant recipient, hepatic transplantation would be the best remedy for all cases of the Budd–Chiari syndrome, except those caused by membranous obstruction of the vena cava or those in which extraction of clot or neoplasm from around the orifices of the hepatic veins is feasible.

If the hepatic veins alone are occluded, but the portal vein and the inferior vena cava are patent, the simplest way of using the portal vein as an outflow tract is to make it one limb of a side-to-side portacaval shunt (Fig. 7) 1221. These circumstances, unhappily, are rare. In the Budd–Chiari syndrome the difficulty with any procedure requiring mobilization of the portal vein is that the nearby caudate ‘lobe’ is turgid; it encroaches on the operative site. Furthermore, in every instance the inferior vena cava and the portal vein have to be mobilized so that they can be joined on an obliquity (somewhat like a side-to-side choledochoduodenostomy).

For the inferior vena cava to reach the portal vein, it may be necessary to cut both the several lumbar veins behind the vena cava and those in the region of the right renal vein and cephalad to it, allowing one to free the inferior vena cava liberally. For the portal vein to reach the vena cava, mobilization from the bifurcation of the vein just before entry into the liver to a variable caudad distance behind the convexity of the pancreas is required. A good result can be expected from a side-to-side shunt in these conditions if the surgery goes well. Unfortunately, the presence of a portacaval shunt increases the difficulty of connecting blood vessels in a subsequent liver transplantation.

To avoid problems of mobilization, H-grafts are sometimes used. The advantage of an H-graft is that the portal vein and the inferior vena cava remain in situ, being connected with a segment of Dacron or Teflon vascular prosthesis, reinforced externally with plastic rings to keep it from collapsing (Fig. 8) 1222. Occasionally a segment of autogenous vein is used, but the result is unpredictable. The problem is that the bulk of the liver is often so great that even if the shunt looks perfect, shifting anatomy produces a kink.

Buckling of the prosthesis is sometimes a problem even when an externally supported (ringed) prosthesis is used, because the rings that one counts upon to support the otherwise floppy conduit can also serve as fulcrums, over which the fabric bends. Keeping the prosthesis short (perhaps only 3.5 cm long) and cutting the hoods of the prosthetic graft to ensure that they correspond perfectly to recipient orifices in the portal vein and in the inferior vena cava 90° away from each other are the keys to success.

In these operations, too, decompression of the liver is gained by allowing native structures to remain in place, but making them serve as an outflow conduit to the inferior vena cava or to the heart. These are the best answers to a difficult dilemma, excepting only hepatic transplantation. Originally, Fonkalsrud used a Dacron prosthesis as a bypass from the splenic vein to the inferior pulmonary vein. Cameron and others have extended the prosthesis to the right atrium (Fig. 6) 1219,1220, and have developed a number of practical variations.

The first was a mesocaval C-shunt, using the superior mesenteric vein as the outflow tract and a clot-free retroduodenal inferior vena cava as the decompression conduit, provided that a pressure gradient of more than 20 mmHg was not present between the right atrium and the inferior vena cava. Recently, Cameron's mesoatrial shunt (Fig. 6) 1219,1220 has proved more practical, provided the prosthesis is made of externally supported Gore-Tex and a more unyielding cylinder of Silastic or a sleeve of Dacron prosthesis is passed over the portion of the conduit at risk of compression in its substernal route to the heart.

If radiological incursions continue to be successful, the need for surgical splanchnic venous shunts will be reduced. A flexible, expandable, tube made of fine wire Wallstent® can sometimes be passed under radiological guidance through the superior vena cava, through the liver, and then into the portal vein, effectively creating a portasystemic venous shunt (Fig. 9) 1223.

Ahn SS, Yellin A, Sheng FC, Colonna JO, Goldstein LI, Busuttil RW. Selective surgical therapy of the Budd–Chiari syndrome provides superior survivor rates than conservative medical management. J Vasc Surg 1987; 5: 28–37.

Bismuth E, Hadengue A, Hammel P, Benhamou J–P. Hepatic vein thrombosis in Behçet's disease. Hepatology 1990; 11: 969–74.

Cambria RP, Shamberger RC. Small bowel obstruction caused by the abdominal cocoon syndrome: possible association with the LeVeen shunt. Surgery 1984; 95: 501–3.

Cameron JL, et al. The Budd-Chiari syndrome: Treatment by mesenteric–systemic venous shunts. Ann Surg 1983; 198: 335–46.

Halff G, Todo S, Tzakis AG, Gordon RD, Starzl TE. Liver transplantation for the Budd–Chiari syndrome. Ann Surg 1990; 211; 43–9.

Hay JM, et al. Syndrome de Budd–Chiari avec thrombose de la veine cave inférieure: traitement par une derivation mesenterico-innominé. Gastroenterol Clin Biol 1988; 12: 755–8.

Henderson JM, et al. Surgical options, hematologic evaluation, and pathologic changes in Budd–Chiari syndrome. Am J Surg 1990; 159: 41–8.

Klein AS, Cameron JL. Diagnosis and management of the Budd–Chiari syndrome. Am J Surg 1990; 160: 128–33.

Louis JJ, et al. Budd-Chiari syndrome: treatment with percutaneous transhepatic recanalization and dilation. Radiology 1989; 170: 791–3.

Ludwig J, Hashimoto E, McGill DB, van Heerden JA. Classification of hepatic venous outflow obstruction: ambiguous terminology of the Budd–Chiari syndrome. Mayo Clin Proc 1990; 65: 51–5.

Malt RA, Dalton JC, Johnson RE, Gurewich V. Side-to-side portacaval shunt versus nonsurgical treatment of Budd-Chiari syndrome. Am J Surg 1978; 136: 387–9.

Martin LG, Henderson JM, Millikan WJ Jr, Casarella WJ, Kaufman SL. Angioplasty for long-term treatment of patients with Budd-Chiari syndrome. Am J Roentgenol 1990; 514: 1007–10.

Nakao K, Adachi S, Kawashima Y, Okamoto E, Manabe H. A radical operation for the Budd–Chiari syndrome associated with obstruction of the inferior vena cava: A report of six patients. J Cardiovasc Surg 1984; 25: 216–21.

Oldhafer KJ, Ringe B, Wittekind C, Pichlmayr R. Budd–Chiari syndrome: Portacaval shunt and subsequent liver transplantation. Surgery 1990; 107: 471–4.

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

Rollins BJ. Hepatic veno-occlusive disease. Am J Med 1986; 81: 297–306.

Tullis JL. Hematologic problems of liver disease: coagulopathies and hypersplenism. In: McDermott WV Jr, ed. Surgery of the liver. Boston: Blackwell Scientific, 1989: 389–95.

Wang Z, et al. Recognition and management of Budd–Chiari syndrome: report of 100 cases. J Vasc Surg 1989; 10: 149–56.