Valproate-Induced Hyperammonemic Encephalopathy Treated by Hemodialysis

Abstract

Valproate-induced hyperammonemic encephalopathy is an unusual but serious complication that may occur in people with normal liver-associated enzyme levels, despite normal therapeutic doses and serum levels of valproate. Here, we describe an adolescent girl who had absence seizure and complained about progressive dizziness and general malaise several days after restarting valproate. Then, she presented vomiting and decreased consciousness three weeks after valproate use. Notably, her serum ammonia level was five times the upper limit of normal (184 μmmol/L), with normal liver-associated enzyme and supra-therapeutic valproate level. EEG showed continuous generalized slowing. The tandem mass analysis revealed carnitine deficiency. Consciousness improved after emergent hemodialysis. Ammonia level and EEG also returned to normal. Possible mechanisms, risk factors and the treatments of valproate-induced hyperammonemic encephalopathy are described. Physicians should consider this possibility when consciousness disturbance occurs in patients treated with valproate.

Keywords:

• hyperammonemia
• encephalopathy
• valproate

INTRODUCTION

Valproate is a branched-chain fatty acid that is widely used in the therapy of epilepsy. Most of the side effects are mild and transient. However, several serious adverse effects during valproate treatment can occur, such as hepatotoxicity, coagulation disorders, pancreatitis, bone marrow suppression, and hyperammonemia. Valproate occasionally induces stuporous or comatose states associated with an increase in frequent seizures and EEG changes. The encephalopathy frequently is accompanied by hyperammonemia without signs of hepatic failure.[1] Valproate-induced hyperammonemic encephalopathy (VHE) is a serious disease, but it can be reversed if a precocious diagnosis is made.[2] The diagnosis of hyperammonemia is often overlooked due to an unspecific clinical presentation even with normal liver-associated enzyme levels.[3] Here, we describe an adolescent girl who developed hyperammonemia encephalopathy while restarting valproate monotherapy for absence seizure. Rapid clinical symptoms and hyperammonemia were improved after discontinuing valproate and immediate hemodialysis.

CASE REPORT

A 14-year-old girl had an absence seizure via receiving regular valproate control for two years using the maintenance dose of 20–23 mg/kg/day (1000 mg/day). There was no seizure attack, and the drug level was within the normal limit during this period. Developmental milestones and school performance were normal. Four months before admittance, however, she discontinued the drug by herself. Poor appetite presented after stopping valproate, with weight loss of about 8 kg within three months. Unfortunately, myoclonic jerks of both hands were also complained about intermittently from two months before presentation. Thus, valproate was prescribed again by another doctor, using 25 mg/kg/day (1000 mg/day) three weeks before admittance. She complained of dizziness, general malaise, and vomiting several days after restarting valproate. She denied taking any other drugs in the most recent three months. Due to dizziness and general malaise deteriorating progressively, she visited our emergency department. On physical examination, body temperature was 35.7°C, pulse rate was 97 beats per minute, respiratory rate was 20 breaths per minute; and blood pressure was 103/72 mmHg. Her height was 152 cm (25–50^th percentile) and body weight was 42 kg (25^th percentile). The body mass index was 18.1 (normal range: 17.6–22.7). Neurologic examination was normal, including clear consciousness. However, she became lethargic gradually in the emergency room. Laboratory examination revealed ammonia was 184 μmmol/L (normal range < 33 μmmol/L), and valproate level was 182 μg/mL (therapeutic range 50–100 μg/mL). The other metabolic panel was as follows: aspartate aminotransferase 12 IU/L, alanine aminotransferase 5 IU/L. Blood sugar, blood urea nitrogen, creatinine, and electrolyte were normal. Blood gas analysis showed pH 7.455, PCO₂ 30.3 mm Hg, PO₂ 44.6 mm Hg, and HCO3⁻ 21.2 mEq/L. EEG showed continuous generalized slowing. Brain CT revealed no significant anomalies. Three hours later, she became irritable, with incoherent and dysarthric speech, delusion, and asterixis. The Glasgow Coma Scale was 14/15. Under the impression of VHE, we discontinued valproate therapy. Then emergent hemodialysis was performed once for three hours. Her consciousness recovered rapidly to normal 12 hours after hemodialysis. EEG also returned to a normal rhythm the next day (see Figure 1B). There was also a rapid correction of hyperammonemia when ammonia decreased to 1 μmmol/L on the next day and no further episode. The valproate level was 24 μg/mL detected two days after hemodialysis. The tandem mass spectrometry studies revealed carnitine deficiency (carnitine/acylcarnitine: C0 = 0.58, C2 = 0.52; normal range: C0: 3.18–1.35, C2: 5.56–1.5) before hemodialysis. After three months of follow-up, the patient had no cognitive impairment or neurologic sequelae. The serum carnitine was normal: free carnitine in serum was 5.67 mg/L (normal range: 4.3–8.5 mg/L), and total carnitine in serum was 11.09 mg/L (normal range: 6.3–11.6 mg/L). The anti-epileptic drug was changed to lamotrigine due to the absence of epilepsy recurring two months after discharge.

DISCUSSION

The reported incidence of asymptomatic hyperammonemia in children with valproate monotherapy is 19%.[4] However, the incidence of VHE is unknown. In contrast to hyperammonemia due to advanced liver disease, patients with VHE usually have normal liver-associated enzyme levels, suggesting a mechanism other than hepatic cell injury or death.[5]

The clinical manifestations of VHE include acute or subacute decreasing levels of consciousness that go from drowsiness to lethargy and coma, vomiting, dizziness, and focal neurological deficit. Low-grade fever and an increased in the frequency of seizures can also be found. Laboratory tests usually show a normal liver function with hyperammonemia.[2] VHE may occur in people with normal dose and serum levels of valproate.[6] The main findings of EEG are diffuse slowing with a predominance of rhythmical theta and delta activity, as seen in the initial EEG in our patient. Occasionally, triphasic waves can be found.[2] Our patient was compatible with all these diagnostic findings.

The pathogenesis of VHE is not completely understood. Valproate can lead to hyperammonemia because of several mechanisms. The most important one appears to be the inhibition of hepatic mitochondrial carbamoylphosphate synthetase-I, the enzyme that begins the urea cycle.[2] Hyperammonemia has been postulated as the main cause of encephalopathy. Ammonia toxicity is mediated by the excessive activation of the N-methyl-D-aspartic acid type of glutamate receptor, which can increase seizure frequency. On the other hand, excessive ammonia is conjugated in the brain with alpha-ketoglutarate to form glutamate and causes alpha-ketoglutarate depletion. Depletion of alpha-ketoglutarate in the brain produces a block in Krebs cycle and causes cell damage and neuronal death.[7] Some data suggest that VHE may be promoted either by a pre-existing carnitine deficiency or by deficiency induced by valproate per se.[8]

The risk factors of valproate-associated hyperammonemia include underlying urea cycle enzyme deficiencies, underlying liver disease, long-term valproate, concomitant anti-epileptic drug therapy, particular topiramate, or strict vegetarianism. VHE usually develops within days to weeks of initiation of treatment.[1] High initial dose and catabolic state also increase the risk.[5],[9] These three risk factors (initiation of treatment, high initial dose, and catabolic state) might have been associated with VHE in our patient while she restarted valproate therapy.

It has been known that there is no relationship between daily dosages and serum concentration of valproate, hyperammonemia, and the severity of VHE. In one report reviewing 15 VHE patients, the valproate drug level exceeded the therapeutic level in 40% of patients and was within the therapeutic level in the other 60% of patients.[3] Moreover, VHE may be evident even in patients who had previously received valproate without any clinical and laboratory problems,[10] as seen in our patient.

The primary treatment for VHE is the withdrawal of valproate. Reviewing the articles, complete consciousness recovery generally occurs over a period of 2 days to 14 days and has no mortality cases.[3],[11] Hydration, sodium phenyl acetate, and sodium benzoate may be of benefit.[12] On the other hand, L-carnitine may be helpful in the prevention, as well as treatment, of hyperammonia due to valproate.[5] Hemodialysis is a therapeutic option managing hyperammonemia. Most authors agree that ammonia levels greater than 680 μg/dL (400 μmol/L) under valproate therapy or significant clinical symptoms secondary to hyperammonemia require aggressive intervention.[13]

Valproic acid is a small molecule (1440 da) with a small volume of distribution (0.2 L/Kg). Ninety to ninety-five percent is protein bound at therapeutic concentrations. Metabolic acidosis, low albumin levels, and older age also correlate with decreased protein binding. Hemodialysis may thus be beneficial in valproic acid overdose because of these characteristics. In a review of 32 patients with toxic valproate levels, six patients received hemodialysis. The half-life of valproate could decrease from 2.2–4.3 hours during hemodialysis. The authors recommended hemodialysis as a component of the management of patients with necessitation for stability and deteriorating neurologic condition that necessitates intubation.[14] In our case, both hyperammonemia and toxic valproate level were seen with rapid deterioration in the level of consciousness. We regarded that it fit “significant clinical symptoms secondary to hyperammonemia” and decided to perform immediate hemodialysis. Hemodialys should be considered a component of the management in patients with hemodynamic instability and deteriorating neurological condition.

In conclusion, VHE is a potentially serious complication of valproate treatment. Physicians should consider this possibility while consciousness disturbance occurs in patients being treated with valproate, and should monitor ammonia and be aware of risk factors. Early intervention of VHE or withdrawal of valproate could be necessary. In our patient, hemodialysis rapidly improved the consciousness and hyperammonemia with the toxic valproate level.

DECLARATION OF INTEREST

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.