A 35-year-old woman was admitted with a 5-day history of increasing drowsiness, fatigue, personality changes, generalised weakness and deteriorating mobility.
Her medical history included a 10-year history of schizo-affective disorder, lithium-induced diabetes insipidus and hypothyroidism.
Her medications included sodium valproate 1 g twice daily, carbamazepine, quetiapine and lithium.
On admission, the patient was drowsy, difficult to rouse (sleeping all night and in the afternoon).
Her Glasgow Coma Scale (GCS) was 14/15, her abbreviated mini-mental test score was 6/10 and she was noted to be slow in her responses.
Neurological examination showed reduced power in her lower limbs (3–4/5) in a pyramidal distribution, and brisk reflexes with intact sensation in all modalities.
Cardiovascular, respiratory and gastrointestinal examinations were unremarkable and there were no demonstrable features of hepatic disease.
Routine biochemistry was unremarkable.
The patient's alanine transaminase was 7 u/L (1–15), alkaline phosphatase 61 u/L (30–130), γ-glutamyl transferase 17 u/L (<45) and bilirubin was 6 μmol/L (<21).
She had normal serum B12 and folate levels, and normal thyroid function tests.
Serum lithium levels were within the therapeutic range.
Her serum ammonia was 47 μmol/L (11.2–35.4) and valproate concentrations were elevated 140 mg/L (50–100).
Brain imaging including CT and MRI did not show any haemorrhage or intracranial mass.
Lumbar puncture demonstrated clear cerebrospinal fluid (white cell count <1×106/L, protein 0.25 g/L (0.2–0.4)).
EEG demonstrated mild variable slowing and irregularity of background activity (likely due to medication).
The differential diagnosis included encephalitis, Guillain-Barré syndrome and acute disseminated encephalomyelitis (ADEM), which were excluded through our investigations.
In view of the raised valproate and ammonia levels, a diagnosis of valproate hyperammonaemic encephalopathy (VHE) was made at this point
The dose of sodium valproate was reduced by half rather than being completely withdrawn.
This was due to the complex and challenging nature of managing the patient's schizo-affective disorder.
Within 3 days, there was evident clinical improvement in her mental status.
The patient was able to sit and communicate with the team, and her abbreviated mental test score improved to 10/10.The serum ammonia normalised and valproate levels decreased to the therapeutic range.
After 4 days she was discharged home.
Results of investigations of her urea cycle were awaited.
Three weeks postdischarge, the patient was clinically stable in terms of mental functioning, having had no further episodes of confusion or drowsiness.
However, she continued to have fatigue and walking difficulties (due to residual leg weakness).
Results of her serum amino acids revealed that our patient had carnitine deficiency evident by low acyl-carnitine, low free serum carnitine (14.2 μmol/L, normal range: 23–52) and low total serum carnitine (19.9 μmol/L, normal range: 27–63).
She was started on oral carnitine 1 g twice daily.
Within 2 weeks, her symptoms resolved completely with normalisation of serum carnitine levels (free carnitine 25.2 μmol/L and total carnitine of 41.3 μmol/L).
The sodium valproate was initially stopped completely, but due to a serious relapse of the patient's schizo-affective disorder, it was restarted and she is now well on valproate 1 g once daily and long-term carnitine supplements.