Statistics from Altmetric.com
In the absence of liver disease, hyperammonaemia is often not considered in the differential diagnosis of encephalopathy and, therefore, the diagnosis of hyperammonaemic encephalopathy may be delayed. We report a case of fulminant progression of hyperammonaemic encephalopathy after valproate treatment in a patient with ureterosigmoidostomy.
A 31 year old patient was admitted because of confusion and agitation. Ureterosigmoidostomy for congenital bladder exstrophy was performed when he was 9 years old. Over the past years, he was repeatedly hospitalised with episodes of abnormal behaviour, which were blamed on his abuse of various illegal drugs. At this admission, he was disoriented, with lapses into somnolence as well as agitation. Apart from that, his general and neurological status was unremarkable. C reactive protein was 80.5 mg/l, leucocytes 16160/μl, and blood urea 12.6 mmol/l. Venous ammonia concentration was mildly increased (63 μmol/l). Arterial blood gas analysis showed respiratory alkalosis. All other laboratory analyses including toxicological screening, microbiological tests, cerebrospinal fluid analysis, and cranial computed tomography were unremarkable. Plasma amino acids, organic acids, and orotic acid in urine were normal. After treatment with fluids and amoxicillin with clavulanate for urinary infection, C reactive protein, leucocytes, blood urea, and sodium concentrations normalised. Soon after admission, he developed generalised tonic-clonic seizures. Antiepileptic treatment was introduced with 900 mg valproate intravenously followed by continuous infusion (2000 mg/day). Because of repeated seizures, phenytoin was added two days later (1000 mg bolus followed by 1000 mg/day). Despite the antiepileptic treatment, the patient developed a status epilepticus and had to be intubated. The seizures finally stopped after barbiturate coma with thiopental (2 × 500 mg followed by 10 000 mg/day) was induced. Arterial blood ammonia concentration was now massively increased (2875 μmol/l). Emergency dialysis was started, and paromomycin and lactulose were administered. Under dialysis, the blood ammonia concentration decreased to 812 μmol/l. However, the patient developed dilated, areactive pupils. Computed tomography showed diffuse brain oedema. The patient died a few hours later, five days after admission. Total plasma carnitine (6 μmol/l, normal range 33–77 μmol/l) and free carnitine (3 μmol/l, normal range 25–54 μmol/l) concentrations were decreased. Acylcarnitine profile also showed decreased but no abnormal individual acylcarnitine species. Repeated analyses of amino acids in plasma and urine, urinary organic acids, and orotic acid showed no significant abnormalities. On postmortem examination, the liver was normal. In both kidneys, multiple abscesses were found. There was massive cerebral oedema and cerebellar herniation.
Ureterosigmoidostomy has been repeatedly associated with episodic hyperammonaemic encephalopathy.1 In this condition, the urine is excreted directly into the sigmoid colon and then excreted during defecation. Frequent complications include recurrent pyelonephritis, faecal incontinence, intestinal malabsorption, and hyperchloraemic acidosis.1 Hyperammonaemia develops as a result of the increased production of ammonia in the colon from bacterial ureolysis and subsequent absorption of ammonia in the colon. The bowel wall is much more permeable to ammonia than the bladder mucosa, allowing rapid absorption of the ammonia produced within the sigmoid colon. This exceeds the liver's capacity to detoxify ammonia by the urea cycle. In addition, a proportion of the reabsorbed ammonia bypasses the liver through oesophageal and haemorrhoidal veins. Thus, hyperammonaemic encephalopathy can develop even if liver function is normal.
In the present patient, the diagnosis of hyperammonaemic encephalopathy was not established until the ammonia concentration had risen greatly and the patient was already in critical condition. Urea cycle disorders and organic acidurias were excluded. The laboratory analysis and postmortem examination did not find any signs of hepatic dysfunction. There were no signs of inborn systemic carnitine deficiency. The ammonia concentration increased massively after the introduction of valproate. This can hardly be explained as a consequence of ureterosigmoidostomy alone: ammonia concentrations as high as these have never been observed in patients with hyperammonaemia following ureterosigmoidostomy. The patient had been treated a few weeks before for urosepsis following constipation, which theoretically can lead to very high ammonia concentrations. However, during the final episode, he was not constipated. Therefore, in our patient, pre-existing episodic hyperammonaemia with encephalopathy and seizures was obviously severely aggravated by valproate induced hyperammonaemia and depletion of carnitine. Patients with valproate induced encephalopathy have been repeatedly described, the pathophysiology of which seems to be heterogeneous.2 In some patients, previously subclinical urea cycle defects have become manifest after treatment with valproate.3 However, these disorders could be ruled out in our patient. Hyperammonaemia is a frequent side effect of valproate treatment and is often asymptomatic.2 It seems to occur more frequently in children but is also common in adults, particularly in the presence of other antiepileptic drugs, as was the case in our patient. The exact mechanism of valproate induced hyperammonaemia is unknown but it may appear independently of hepatotoxicity.4 Valproate has repeatedly been shown to reduce serum and liver carnitine concentrations, both with and without being associated with hyperammonaemia.5 While most of these patients were children, some cases in adults have been described.6 Valproate may reduce carnitine concentrations by forming an ester with carnitine, which is co-excreted with organic acids into the urine, or by altering renal reabsorption of acylcarnitine and free carnitine.7
In conclusion, our case shows that valproate may greatly aggravate pre-existing, mild hyperammonaemia. We suggest that valproate should be avoided in patients with even slight hyperammonaemia and normal liver function. Equally, we advise the close monitoring of ammonia and carnitine concentrations in patients with ureterosigmoidostomy, such as the one described here, if valproate cannot be avoided.
Competing interests: none declared
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.