Article Text

Encephalopathy without cirrhosis: hunt the shunt
  1. Department of Surgery
  2. King’s College Hospital
  3. Denmark Hill
  4. London SE5 9RS, UK

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

See article on page 442

The embryology of the portal venous system is extremely complex and yet major congenital malformations of the system have been rarely described. However, the increased use of sophisticated ultrasound techniques and the increase in liver transplantation have resulted in an increase in the number of reports of these abnormalities during the past 15 years. There is also an increased awareness of associated complications from these abnormalities including late onset encephalopathy.

The systemic venous system develops from the intra-embryonic anterior and posterior cardinal veins, whereas the portal venous system develops from the extra-embryonic vitelline and umbilical veins which provide drainage from the yolk sac and placenta, respectively.1 The vitelline veins are paired vessels which drain into the primitive sinus venosus and by the end of the fourth week of gestation, three cross communications have developed between them in subhepatic, retroduodenal and subdiaphragmatic positions.2

The portal vein in the embryo is derived from the inferior section of the left vitelline vein, the retroduodenal communication, and the superior section of the right vitelline vein, and so forms the S-shaped configuration of the mature portal vein. The inferior vena cava is also formed from the amalgamation of several venous channels in the embryo including the right end of the sinus venosus and the right end of the subdiaphragmatic anastomosis of the vitelline veins. Early in development blood from the left umbilical vein passes through the subhepatic anastomosis and the ductus venosus to reach the retrohepatic portion of the inferior vena cava, the common hepatic vein in the embryo. Anatomically, the ductus venosus therefore forms a direct connection between the umbilical vein and the inferior vena cava and provides a functional bypass through which part of the oxygenated umbilical blood from the placenta flows directly into the inferior vena cava.3 Human fetal studies clearly demonstrate that the umbilical vein gives branches to the left segments of the liver and that the opening of the ductus venosus lies opposite the junction of the umbilical vein with the left branch of the portal vein. Roughly 40 to 50% of umbilical vein blood passes through the ductus venosus whereas the remainder flows through the liver sinusoids.3

The wall of the ductus venosus contains only a small amount of smooth muscle and no muscular sphincter. Postnatal closure of the channel commences during the first minute after birth but is not completed until days 15 to 20. The mechanism of closure is not clear but it may be related to a decrease in pressure within the portal venous system after cessation of umbilical vein flow.3

Congenital malformations of the portal venous system are unusual. A preduodenal position of the portal vein is the result of a persistence of the most caudal of the inter-vitelline anastomoses and this abnormality is found most frequently in association with biliary atresia, duodenal atresia, and annular pancreas. Persistence of both vitelline veins may result in portal vein duplication, whereas more complex defects may include portal drainage into the superior vena cava, right atrium, or azygos system of veins.4

Abernethy (1793)5 gave the first description of a congenital extrahepatic mesenterico-caval anastomosis in association with an absence, or atresia, of the portal vein. The report was based on a postmortem examination of a 10 month old female infant who died of unknown cause. In addition to transposition of the great vessels and polysplenia, the portal vein terminated in the inferior vena cava at the level of the insertion of the renal veins. At least 13 examples of this type of congenital portosystemic anastomosis have now been reported, all in females. Complications in these cases included encephalopathy from the age of 44 years in one patient and a variety of liver tumours in five others which included focal nodular hyperplasia, adenoma, hepatoblastoma, and hepatoma. Recently it has been suggested that this type of congenital end-to-side portosystemic shunt should be designated as a “type 1” lesion.6

A second variety of side-to-side congenital extrahepatic portosystemic shunt (“type 2”) has also been recognised in two adult and four paediatric patients, all of whom were males.6 This type of communication between the portal vein and the retrohepatic vena cava is generally believed to represent a persistence of the ductus venosus and is amenable to surgical division and restoration of a normal intrahepatic flow of portal blood. Two adult patients, aged 49 and 67 years, presented with the clinical and biochemical features of hepatic encephalopathy. Surgical division of the shunt in the first patient resulted in severe portal hypertension secondary to intrahepatic portal vein hypoplasia and a portosystemic anastomosis was re-established as an emergency.7 The second patient responded well to surgical division of the shunt and there were no further episodes of encephalopathy.8 Surgical closures of “type 2” side-to-side shunts were also performed in two children without complication.6 Congenital portosystemic shunts, analogous to those described in humans, are well recognised in a variety of animals including dogs and cats.9 Encephalopathic ataxia in these animals is followed by coma and the clinical signs are accompanied by increasing blood ammonia concentrations. Urolithiasis was also found in 53% of the animals and was believed to be secondary to urinary excretion of ammonia and urate.

A familial incidence of patent ductus venosus (“type 2” shunt) was described in three young brothers who presented with encephalopathy. The symptoms and biochemical changes were successfully reversed in the two who underwent surgical closure of the shunt.10 Jacobet al, in this issue (see page 442), describe another family of three brothers in whom congenital “type 2” side-to-side shunts were demonstrated with ultrasound, computed tomography and angiographic imaging. Two of the patients had evidence of mild encephalopathy, one of whom had bilateral renal stones. Two liver masses of focal nodular hyperplasia were detected in the third patient. The clinical observations of encephalopathy, liver tumour and male dominance in the two reports of familial “type 2” shunts are therefore in accord with the previous single case reports. The authors of the current report suggest that their observations are in keeping with the transmission of an autosomal recessive trait.

In conclusion, reports of congenital portosystemic shunts have become more frequent during the past 15 years and they should be considered in the differential diagnosis of unexplained encephalopathy, particularly when symptoms appear in younger patients. The association of congenital shunts with primary liver tumours also suggests that a complete investigation of liver masses in young people should include a search for any congenital abnormality of the portal venous system. Surgical correction of “type 2” shunts prevents the onset of encephalopathy and can reverse established signs. Liver tumours can be resected safely, even in patients with “type 1” shunts,2 but severe encephalopathy in “type 1” lesions, which are associated with an absent intrahepatic portal vein, can only be relieved by liver transplantation.


Linked Articles