Article Text

Download PDFPDF
Congenital extrahepatic portosystemic shunt with progressive myelopathy and encephalopathy
  1. Danielle Lux1,
  2. Anna Naito1,
  3. Sreedharan Harikrishnan1,2
  1. 1 Neurology, East Kent Neurology Unit, East Kent University Hospitals NHS Foundation Trust, Canterbury, UK
  2. 2 Medical Education, Kings College London, London, UK
  1. Correspondence to Dr Danielle Lux, Neurology, East Kent Neurology Unit, East Kent University Hospitals NHS Foundation Trust, Canterbury CT9 4AN, UK; d.lux{at}


Portosystemic encephalopathy commonly occurs in patients with portal hypertension caused by end-stage liver disease or portal vein thrombosis. Congenital extrahepatic portosystemic shunt (CEPS) is an underdiagnosed and treatable condition that can cause encephalopathy and various neuropsychiatric symptoms. We report an unusual case of type 2 CEPS in a 29-year-old woman who presented with progressive myelopathy and fluctuating encephalopathy on a background of congenital cardiac disease. Investigations showed hyperammonaemia, and despite no evidence of portal hypertension on ultrasound imaging, CT scan of abdomen showed a shunt between the mesenteric and left internal iliac veins. Patients with unexplained fluctuating or progressive neuropsychiatric symptoms should have their serum ammonia checked. A raised serum ammonia concentration without known portal hypertension should prompt further investigations for extrahepatic shunts.

  • metabolic disease
  • neuropsychiatry
  • myelopathy
  • encephalopathy

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.


Congenital extrahepatic portosystemic shunt (CEPS) is a rare congenital anomaly that can present at any age with variable clinical manifestations. There are two types: type 1 has a completely absent portal vein; type 2 has a partial shunt with preserved hepatic portal flow.1 Type 2 (unlike type 1) is not usually associated with other congenital diseases, occurs equally in men and women and tends to present in adulthood.2 Most patients with type 2 CEPS are asymptomatic and may go undiagnosed for decades. However, several different neurological presentations are possible, including progressive parkinsonism; encephalopathy ranging from confusion to coma; seizures; organic psychosis; cerebellar syndromes and spastic paraparesis.3 Some develop multisystem complications such as hepatopulmonary syndrome, nephrotic syndrome and hepatic dysfunction.4 A low protein diet, lactulose and ligation or coiling of the shunt can significantly improve neuropsychiatric symptoms.3 We present a patient with type 2 CEPS who presented unusually with both myelopathy and encephalopathy, against a background of congenital cardiac disease. Neurologists should consider CEPS (including measuring serum ammonia) in patients with a range of neurological presentations, even when there is no cirrhosis or portal hypertension.

Case presentation

A 29-year-old woman presented with a 2-year history of fluctuating spastic paraparesis and cognitive impairment on a background of tetralogy of Fallot, surgically corrected at the age of 2 years. She had a previous history of antiphospholipid syndrome requiring long-term anticoagulation, migraine, chronic back pain and depression. She was under neuropsychological review for fluctuating cognitive impairment attributed to depression and psychotropic medications (tramadol 100 mg two times a day, gabapentin 900 mg three times a day).

On examination, she had symmetrical spastic paraparesis, hyper-reflexia and bilateral extensor plantar reflexes. Sensation was preserved. Cranial nerves and upper limbs were entirely normal. She scored 83/100 on Addenbrooke’s cognitive examination (ACE-R III) testing, with deficits in visual memory and fluency.

Routine bloods including liver function were normal. Serum angiotensin-converting enzyme (ACE) was 138 U/L (25–82); antinuclear antibody was mildly positive (1:80 titre) and speckled. Other serological tests were negative, including extractible nuclear antigen, antineutrophil cytoplasmic antibody, serum complement, copper studies, cryoglobulins, HIV, syphilis, Lyme and HTLV-1. Serum antibodies to N-methyl-D-aspartate receptor, voltage-gated potassium channels, glutamic acid decarboxylase and onconeuronal were negative. Skin biopsy was non-specific. MR scan of the spinal cord showed no signal change. MR scan of brain, however, showed T1-weighted bilateral basal ganglia hyperintensities (figure 1). MR cerebral venogram was normal.

Figure 1

MR scan of brain showing T1-weight bilateral basal ganglia hyperintensities.

Cerebrospinal fluid (CSF) identified an elevated CSF ACE at 1.97 (normal <1.2); oligoclonal bands were negative; other constituents were normal. CT scan of chest was normal, and further evaluation with fluorodeoxyglucose-positron emission tomography imaging showed no avid disease. She did not fulfil the Zajicek criteria for sarcoidosis.5

In light of the cognitive impairment and MR brain scan findings, we measured her serum ammonia, which was elevated at 86 umol/L (normal <50). Non-concurrent electroencephalography (EEG) was normal. Hepatitis serology and antimitochondrial antibodies were negative. Echocardiogram identified pulmonary regurgitation with a dilated right ventricle but no significant tricuspid regurgitation, in keeping with corrected tetralogy of Fallot. Despite evidence of right-sided cardiac impairment, liver ultrasound scan showed no hepatomegaly, cirrhosis or portal hypertension. In view of antiphospholipid syndrome, and despite targeted anticoagulation, we considered portal vein thrombosis and arranged further imaging with CT abdomen (figure 2). This showed a dilated inferior mesenteric vein with anastomosis with the left internal iliac vein constituting a portosystemic shunt. Triphasic liver CT scan confirmed portosystemic shunting and so we made a diagnosis of type 2 CEPS. Having identified a shunt, we did not perform a full screen for inherited metabolic disorders.

Figure 2

CT scan of abdomen (axial on left, coronal on right) showing the portosystemic shunt between the inferior mesenteric vein and the left internal iliac vein with narrow hepatic portal vein.

Serial reviews showed fluctuation with a correlation between hyperammonaemia and cognitive performance, with her ACE-R improving to 91/100 as her serum ammonia normalised. She was referred to hepatology and trialled on rifaximin. Her symptoms have since progressed and she has developed new onset extrapyramidal tremor. She is awaiting definitive treatment with endovascular embolisation.


Ammonia is a by-product of amino acid metabolism and intestinal urea-splitting bacteria. Glutamine and glutamate are metabolised generating free ammonium ions (NH4+) in the mitochondria, while free NH4+ is metabolised back to glutamine in the cytosol. The urea cycle, exclusive to liver mitochondria, converts NH4+, while the kidneys excrete variable amounts of ammonium dependent on acid–base balance. The brain cannot convert NH4+ to urea; however, the CNS concentration of ammonium is kept low by the astrocytic enzyme glutamine synthetase, which synthesises glutamine from glutamate and NH4+.

Failure to process nitrogen and maintain ammonia homeostasis—as occurs in urea cycle disorders and liver failure—results in hyperammonaemia and increases the circulating and tissue concentrations of glutamine. Additional metabolic and regulatory pathways impact on brain glutamine homeostasis, including alterations of high affinity astrocytic glutamine transporters responsible for glutamine clearance.6 Ammonia is toxic to the brain through several mechanisms: cellular glutamine accumulation with astrocyte swelling and cytotoxic oedema; glutamate release and neurotoxicity; impaired serotonergic and cholinergic neurotransmission; induced excitotoxicity via NMDA receptors; and nitric oxide release and oxidative stress.6 7 The brain glutamine concentration correlates positively with the severity of neurological symptoms.

In portosystemic shunts, the nitrogen load bypasses periportal hepatocyte conversion and directly enters the systemic circulation, leading to hyperammonaemia. The usual cause of portosystemic encephalopathy is portal hypertension from acquired causes such as liver cirrhosis, portal vein thrombosis and postgastric bypass.8 9 Congenital portosystemic shunts are an under-recognised and treatable cause. This case of CEPS is exceptional in that myelopathy and encephalopathy developed simultaneously, whereas in previous cases, each of these has occurred in isolation only. Our patient had tetralogy of Fallot (probably relevant) yet type 2 CEPS is not normally associated with such additional congenital anomalies.

Kerlan et al first reported CEPS in 1982.10 It manifests at any age11 and the clinical picture is varied. Those presenting with neuropsychiatric symptoms (such as ataxia and psychosis) in case series almost always have hyperammonaemia. Serum ammonia concentrations can fluctuate though, and patients may need repeat testing to support the diagnosis. CSF glutamine concentration may better indicate chronic hyperammonaemia but is not a routine test. A distinguishing feature of CEPS is that liver function tests are usually normal, and cirrhotic liver disease and portal hypertension are absent. Other toxic compounds usually removed by the liver, including manganese, may accumulate in the brain; patients with proven CEPS should have their serum manganese measured.3

Patients with hepatic impairment require care when using analgesics. Portosystemic shunting might decrease the first-pass metabolism of certain drugs and lead to their increased oral bioavailability.12 For example, the action of certain opioids such as codeine or tramadol relies on hepatic biotransformation to active metabolites. Thus, hepatic impairment would be expected to reduce their analgesic effect. However, their reduced hepatic clearance and increased bioavailability have prompted recommendations to prolong the dosing intervals in patients with hepatic impairment in order to prevent drug accumulation. Renally cleared drugs, notably gabapentin or low-dose tricyclic antidepressants, appear to be the safest options for managing neuropathic pain in patients with hepatic impairment.

Neuropsychological and neurophysiological tests are supportive but not specific. A minority of patients have triphasic waves on EEG but their presence can fluctuate and their absence may be falsely reassuring. MR scan of brain may show basal ganglia hyperintensity and MR spectroscopy can show increased cerebral glutamine resulting in astrocyte oedema; these changes are not specific to the underlying cause.3 These MRI modalities are more useful for monitoring patients’ progress with treatment rather than for diagnosis. However, monitoring blood serum ammonia or plasma glutamine concentrations may be of more practical relevance for evaluating response to treatment.

Imaging is the definitive method for the diagnosis of CEPS, and unexplained hyperammonaemia should prompt further investigations. In the absence of cirrhosis, patients need dedicated liver imaging using Doppler ultrasound, CT scan with portal system visualisation or MR angiography.13 The extent of the shunt can be measured with nuclear medicine studies such as transrectal portal scintigraphy with I123 iodoamphetamine.14 The portomesenteric blood bypasses the liver through the shunt into the systemic circulation. Hence, the shunt ratio can be calculated by how much isotope accumulates in the lung: >5% is abnormal and a>60% shunt ratio indicates high risk for developing spontaneous encephalopathy.15

It is important to establish the anatomy of the shunt as this may determine management, including surgery.16 In type 1 CEPS, where the portal vein is entirely absent, liver transplantation is the only treatment option,13 whereas in type 2 CEPS, symptoms can improve and even resolve completely with low-protein diet, lactulose and rifaximin, treatments that reduce ammoniagenesis (effective in recurrent or chronic encephalopathy).3 The definitive treatment is occlusion of the shunt by surgical ligation or interventional coiling.17 It may be difficult to classify the anatomy of type 2 CEPS further, looking at either the origin17 or caval ending16 of the shunt. The latter, in combination with portal venous pressure measurement, has been used successfully to dictate surgical strategy.18

It is essential to recognise CEPS early, since early intervention can improve or even completely resolve neuropsychiatric symptoms3 as well as other complications including liver nodules, pulmonary hypertension and hepatopulmonary syndrome.16 19

Key points

  • Congenital extrahepatic portosystemic shunt (CEPS) is a rare but treatable congenital anomaly that can cause a constellation of neuropsychiatric symptoms.

  • In patients with neuropsychiatric symptoms but without cirrhotic liver disease or portal hypertension, clinicians should consider type 2 CEPS and check the serum ammonia.

  • Patients with type 2 CEPS may have other congenital anomalies.

  • Type 2 CEPS can be treated by staged ligation or by minimally invasive endovascular closure, which improves or even resolves these symptoms.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.


  • Contributors AN and DL contributed equally to this paper and are joint first authors. DL and SH were responsible for the clinical care of the patient. All authors reviewed and approved the final manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed by Tom Britton, London, UK and Robin Lachmann, London, UK.

Other content recommended for you