Discussion

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