To the Editor: We report the case of a 71-year-old woman who presented with a 3-week history of lethargy, subacute confusion and drowsiness. She was known to have a seizure disorder for which she had been taking lamotrigine 100 mg and sodium valproate 500 mg twice a day for 2 years.

On examination, the woman was disoriented with regard to person and time, and had constructional apraxia and asterixis. The rest of the physical examination was unremarkable.

A full blood count, electrolyte levels, coagulation parameters, arterial blood gas measurements and hepatitis serology were normal. Tests for immunological markers of autoimmune liver disease were negative. Liver function tests showed longstanding raised levels of alkaline phosphatase (158 U/L [reference range (RR), 30–110 U/L]) and γ-glutamyl transferase (434 U/L [RR, < 40 U/L]). Serum drug levels were sodium valproate 51.0 mg/L (therapeutic range, 50–100 mg/L) and lamotrigine 9.5 mg/L (therapeutic range, 3–14 mg/L). The venous blood ammonia level was 109 μmol/L (RR, < 50 μmol/L).

A liver ultrasound scan was normal. Computed tomography of the brain showed microvascular changes and an old cortical infarct. An electroencephalogram (EEG) showed diffuse slowing, with a predominance of rhythmical theta activity and some delta activity, suggestive of encephalopathy.

As hyperammonaemic encephalopathy secondary to sodium valproate therapy (VHE) was considered a possible diagnosis, sodium valproate treatment was discontinued. The patient’s confusion resolved completely and the asterixis disappeared within a week. At the same time, her blood ammonia level fell to 19 μmol/L and her EEG normalised.

Eight months after discontinuing sodium valproate treatment, the woman was still asymptomatic. A subsequent percutaneous liver biopsy, to investigate her persistently abnormal liver function, showed features consistent with primary biliary cirrhosis.

Sodium valproate is used not only for management of epileptic disorders but also for migraine prophylaxis and treatment of several psychiatric conditions. Although a generally well tolerated drug, it has a few well known side effects, including hyperammonaemia and, rarely, VHE.1-3 The possible pathophysiology of VHE has been described elsewhere.2 Gerstner et al reported on a series of 19 patients with VHE between 1994 and 2003.4 Review of the literature suggests that VHE is under-recognised, leading to considerable delay in the diagnosis of this potentially reversible condition.3,5

In our patient, it is reasonable to presume that sodium valproate precipitated the encephalopathy on a background of evolving unrecognised liver disease. The marked improvement in her clinical manifestations after discontinuation of valproate further supports this presumption. We have drawn attention to this case to highlight that VHE should be considered in patients presenting with confusion. Prompt measurement of the ammonia level and cessation of valproate treatment should be considered if clinically appropriate. Patients with previously unrecognised liver disease may be at particular risk.

Acknowledgement: We thank Professor Peter Roberts-Thomson, Director of the Department of Immunology at Flinders Medical Centre, for his expert opinion and advice.