To the Editor: It is well known that valproate acid can induce encephalopathy due to hyperammonemia.¹ However, it seems that interactions between antimanic medications including valproate acid may induce encephalopathy by mechanisms other than hyperammonemia or liver failure.

A 36-year-old woman was admitted to the hospital for a second manic episode. She was first treated with diazepam, 30 mg/day, to manage agitation, and divalproex sodium, 1,000 mg/day, was added as an antimanic medication. As her clinical state worsened over the next 3 days, divalproex sodium was increased to 1,500 mg/day and olanzapine, 10 mg/day, was added. Adding an antipsychotic while gradually increasing the dose of valproate acid is recommended in the absence of signs of improvement.² When this latest combination was started on day 4, the patient became drowsy, and therefore diazepam was stopped. However, her clinical state continued to worsen. The next morning, she gradually became comatose. All treatment was stopped, and a physical exam was performed. Electrolytes as well as liver and renal function tests were normal. Her serum valproate acid level remained in the therapeutic range according to clinical indications, at 120 mg/liter (plasma concentration for clinical response: 45–125 mg/liter).^3,4 The ammonia level was 26 μmol/L [laboratory range: 11–48 μmol/liter]. Previous findings included normal brain computed tomography. Her serum valproate acid dropped to 77 mg/L, and she was deeply comatose and in a nonreactive state some hours later. An EEG was performed and showed severe encephalopathy (Figure 1; panel A). She was transferred to the medical intensive care unit where she was intubated. Twenty-four hours later, after all the treatments had been discontinued, she completely recovered. Olanzapine alone was reintroduced within 48 hours and no side effects occurred. The EEG performed 12 days after the encephalopathy was normal (Figure 1; panel B). The patient was discharged taking 20 mg/day olanzapine alone.

This case provides evidence that an association of psychotropic drugs including valproate acid can lead to severe encephalopathy without toxic accumulations of ammonia. The main hypotheses that we put forward to explain the event are pharmacodynamic and pharmacokinetic interactions between the drugs. First, the psychotropic effects of these three drugs may be potentiated by their association and may induce deeper sedation. Second, valproate acid, olanzapine, and diazepam are strongly bound to plasma protein (80%–95%, 93%, and 98%–99%, respectively).⁵ This association of psychotropics may have saturated the binding sites of plasma proteins and thus increased the free fraction of these drugs, especially for valproate acid, whose free fraction increases with high dosages (10% at 40 mg/L to 18.5% at 130 mg/L).⁴ Since only the free fraction of drugs can penetrate the CSF, the interaction described may result in an overdose leading to neurologic toxicity. The serum valproate acid dosages (the total of free and bound fraction) at their appropriate levels may have hidden the overdose. Reducing doses of valproate acid would probably have prevented this neurologic adverse event. This case illustrates that associations of such antimanic medications must be used with caution.