Status epilepticus secondary to hyperammonaemia: a late presentation of an undiagnosed urea cycle defect

Abstract

In this report, we describe the diagnosis, investigation and management of a patient presenting with refractory status epilepticus secondary to a previously unrecognised urea cycle defect, ornithine transcarbamylase deficiency, causing a hyperammonaemic encephalopathy. While metabolic disorders will be readily considered in a paediatric population presenting with difficult seizures, it is unusual for such cases to present in adulthood, and maintaining a broad differential in patients with status epilepticus is important. Early recognition and initiation of treatment are vital. Furthermore, the patient had been diagnosed with schizophrenia over a decade previously and more recently started on sodium valproate, a medication known to contribute to hyperammonaemia. This case also emphasises the importance of exclusion of underlying organic disease prior to diagnosis of psychiatric conditions.

Background

Constructing a differential diagnosis systematically evaluates and sequentially rules out potential causes for a patient’s constellation of symptoms. It traditionally begins with the most life-threatening conditions but considers the epidemiological probability of a given disease occurring in that particular patient. As such, rare but potentially life-altering conditions presenting with common and non-specific symptoms need to be highlighted so they are actively pursued in the elimination process of forming a differential diagnosis, given that they are otherwise likely to be overlooked on the basis of epidemiological improbability.

We report the case of a man presenting initially with behavioural change, who went on to develop generalised tonic–clonic seizures for the first time, and whose ultimate diagnosis was a late presentation of a rare metabolic disease: the urea cycle disorder, ornithine transcarbamylase (OTC) deficiency.

Although this patient’s diagnosis is rare, his initial symptoms of impaired mental status and first seizure account for an estimated 2% and 0.3% of all emergency department (ED) visits, respectively.^{1 2} According to a recent report, in the UK this would represent 470 000 visits per year due to behavioural change, and 70 500 visits due to first presentation seizure.³

This case highlights the need for maintaining a high index of suspicion for metabolic disorders when assessing patients who present to the ED with a first seizure at any age, as well as when it comes to investigating underlying organic causes for psychiatric presentations. For those with undiagnosed urea cycle disorders, this can be a matter of life and death.

Case presentation

A 46-year-old man with a longstanding diagnosis of schizophrenia presented to the ED having become acutely agitated overnight. His next of kin described him becoming restless, pacing around the house, followed by the onset of vomiting and progressive disorientation, failing to recognise family members.

He had been diagnosed with schizophrenia at the age of 15 and his medications on admission included aripiprazole and sodium valproate. He was also hypertensive and had a history of recurrent tonsillitis, with a recent episode 4 days prior to this admission. He was initially referred to the psychiatry team, who remarked on a reduced level of consciousness and recommended further evaluation by the medical team. He was admitted to the acute medical unit for assessment, where he was reported to be agitated and refusing medications.

Later the same day, he developed status epilepticus refractory to treatment with benzodiazepines and phenytoin, for which he was sedated, intubated, ventilated and admitted to intensive care. Initial CT scan (figure 1) of his brain did not show any abnormalities and his lumbar puncture results were not consistent with a central nervous system infection.

Figure 1.

Comparison of CT scans from initial presentation (09:24 hours, left) to a scan performed after non-convulsive status epilepticus was noted (23:40 hours, right). Left: Initial CT head (09:24 hours): no acute intracranial haemorrhage, territorial infarct or mass lesion. Ventricles and basal cisterns are symmetrical and within normal limits. Incidental cavum septum pellucidum. There is a small calcified meningioma associated with the left anterior falx cerebri. The mastoid air cells and paranasal air sinuses are clear. Skull vault and skull base are intact. Impression: No acute intracranial abnormality. Right: Repeat CT head (23:40 hours). There is generalised cerebral oedema with loss of grey-white matter differentiation and obliteration of the Cerebrospinal Fluid (CSF) spaces with no acute cerebral haemorrhage. There is effacement of the basal cisterns in keeping with impending cerebral herniation.

After 12 hours, discreet flickering of eyelids along with reduced pupillary reaction to light and rising lactate levels raised our suspicion of non-convulsive status epilepticus. We arranged a repeat CT (figure 1) of his brain which now showed severe cerebral oedema with sulcal effacement and signs of impending uncal herniation. He later developed generalised tonic–clonic seizures despite deep sedation, requiring burst-suppressive doses of thiopental for seizure control.

Non-hepatic hyperammonaemic encephalopathy was added to the differential diagnosis after discussion with the neurology team, who noted his sodium valproate was a possible cause of hyperammonaemia and an ammonia level was consequently sent. This was found to be raised at 557 μmol/L, over 10 times the upper limit of normal. Valproate levels were also requested, which were within range. Levocarnitine was started, with valproate-induced hyperammonaemia in mind as the cause of his presentation.

However, after discussion with a metabolic disorders specialist unit, a urea cycle disorder was felt to be most likely and specific tests were requested. These included plasma amino acid (table 1) and urine organic acid (table 2) measurements, as well as genetic testing specific to OTC deficiency. All protein intake was stopped, and calorie requirements were met with dextrose only. We started haemofiltration, in addition to the supportive measures already in place.

Table 1.

Plasma amino acids

Amino acid	Value (μmol/L)	Reference range
Taurine	79	19–173
Theonine	113	38–239
Serine	127	51–231
Glutamate	167	21–174
Glutamine	2292	307–768
Proline	203	45–452
Glycine	296	81–303
Alanine	724	112–686
Citrulline	40	8–57
Valine	382	96–566
Methionine	72	10–53
Isoleucine	175	26–159
Leucine	425	50–264
Tyrosine	120	26–154
Phenylalanine	105	34–110
Orithine	54	20–144
Lysine	1406	61–337
Histidine	165	40–143
Arginine	61	26–180

Laboratory comment: Gross increase in glutamine, consistent with severe hyperammonaemia. Marked increase in lysine, possibly also secondary to hyperammonaemia. Mild increases in a number of other amino acids (alanine, methionine, total isoleucine, histidine)—this is not a specific pattern associated with a particular inherited metabolic disease.

Table 2.

Urine organic acids

Urine organic acid	Value μmol/mmol creatinine
Methylmalonic acid	<3 (0–10.0)
Orotic acid	>25 (0.0–5.0)

Laboratory comment: Orotic acid >25 μmol/mmol creatinine (reference range <45). Gross increase in orotic acid with mildly increased uracil consistent with a diagnosis of urea cycle disorder.

Results of these tests for our patient were as follows:

Despite treatment, his ammonia remained unmeasurably high (>1140 μmol/L), and within days of initial presentation, he progressively lost brainstem reflexes, became increasingly cardiovascularly unstable and suffered a cardiac arrest due to uncal herniation.

Prior to our patient’s death, we sent samples for genetic testing, the results of which confirmed that our patient was hemizygous for the c.164A>G OTC gene variant. This is predicted to result in the substitution of tyrosine with cysteine at amino acid residue 55 if translated: p.(Tyr55Cys). A single patient with this variant has been reported in the literature.⁴ A different amino acid change has also been reported in a late-onset male patient affected by OTC deficiency.⁵ A diagnosis of OTC deficiency was confirmed based on these findings.

Differential diagnosis

Hyperammonaemia is a well-recognised feature of liver cirrhosis,⁶ in which the loss of hepatocytes, the only cells in which the urea cycle takes place, is the underlying cause. Failure of the urea cycle of any cause can also result in the accumulation of ammonia. Outside the setting of decompensated liver disease, it is a less commonly considered biochemical disturbance, particularly in the adult population.⁷

A differential for non-hepatic hyperammonaemia is summarised below⁸:

• Failure to clear ammonia: Inborn errors of metabolism resulting in decreased metabolism of ammonia, such as urea cycle defects, fatty acid oxidation defects and organic acidemias.

• Increased protein catabolism, resulting in overwhelming levels of ammonia with normal metabolic pathway capacity exceeded: for example, in mechanically ventilated patients, prolonged seizures, tumour lysis syndrome, following bariatric surgery and postpartum uterine involution.

• Drug-induced, for example, sodium valproate.

• A poorly collected or transported sample.

Inborn errors of metabolism should be considered when the hyperammonaemia is unexplained, progressive and overwhelming, despite supportive therapy. Patients may have a metabolic defect with a mild phenotype, providing partial enzyme activity. These patients may have a subclinical illness for many years, or present with mild symptoms, but are vulnerable to catastrophic episodes of hyperammonaemia if faced with another cause of metabolic stress.

Once hyperammonaemia is confirmed, investigations aim to narrow down the metabolic defect or specific enzyme responsible for the build-up of ammonia through the measurement of substrates or products of individual metabolic pathways. These will usually be directed by a metabolic medicine specialist but may include measurement of serum and urinary amino and organic acids which form the substrates, intermediates and products of different stages of the urea cycle. DNA testing provides definitive confirmation of the specific genetic defect responsible for the metabolic syndrome.

With regard to our patient, the high glutamine (table 1) is highly suggestive of a urea cycle defect: glutamine increases because a rising ammonia level results in metabolism by this alternative pathway (see figure 2). The low normal citrulline is indicative of a proximal defect—defect further around the cycle would cause a rise in the citrulline level as its metabolism slowed. Due to the deficiency of OTC, there will be a rise in carbamyl phosphate which can combine with aspartate to produce orotic acid, explaining the high urine orotic acid levels (table 2).

Figure 2.

The urea cycle and rationale for specific treatments (Illustrated by Dr Penny Beddoes). ARG, arginase; ASL, argininosuccinate lyase; ASS arginosuccinate sythetase; ATP, adenosine triphosphate; CPS1, carbamoyl phosphate synthetase 1; HCO₃, sodium bicarbonate; NAGS, N-acetylglutamate synthetase; NH₄, ammonia; OTC, ornithine transcarbamylase. *OTC is the enzyme that was deficient in the patient in this case. **Carbamyl phosphate metabolised via alternative pathway leading to increased concentrations of orotic acid in urine.

Other tests to consider in a patient presenting with this constellation of symptoms would include electroencephalogram (EEG) and MRI. The likely findings on an EEG would be those consistent with a hyperammonaemic encephalopathy: generalised, slowing of background activity with loss of reactivity and sharp triphasic waves. It may also assist in the detection of ongoing non-convulsive status epilepticus. The typical MRI findings described in a case series of patients with hyperammonaemia were symmetrical extensive cortical signal abnormalities, typically involving the insular and cingulate cortices, with restricted diffusion.⁹ Unfortunately, in our district general hospital setting, access to these tests requires interhospital transfer, and the clinical course of this patient overtook our ability to pursue these investigations.

Treatment

Guidelines for the treatment of decompensated urea cycle defects can be found on the British Inherited Metabolic Diseases Group (BIMDG) website under the page ‘Emergency guides’.¹⁰ Management focuses on maintaining a blood glucose of 6–10 mmol/L to ensure that calorific requirements are met by carbohydrate metabolism. In stable and relatively well patients, this may be achieved by the patient consuming two hourly oral high carbohydrate drinks. If the patient is unwell, intravenous 10% dextrose should be used at a rate of 2 mL/kg/hour. All protein intake should be discontinued, including Nasogastric (NG) feeding.

Guidelines for the makeup of specific intravenous infusions for the treatment of decompensated urea cycle defects are also available in the emergency guideline on the BIMDG website. For OTC deficiency, the infusion is made up of L-arginine, sodium benzoate and sodium phenylbutyrate. L-arginine acts to ‘drive’ the urea cycle, thereby increasing the metabolism of ammonia. Sodium benzoate and sodium phenylbutyrate act as ‘ammonia scavengers’. By reacting with glycine and glutamine, respectively, they facilitate the metabolism of ammonia through alternative pathways to the defective urea cycle.¹¹ The products of these alternative metabolic pathways are excreted in the urine. Figure 2 gives an illustration of the pathways in which these would work.

On the advice of our local metabolic diseases unit, we did not use ammonia scavengers or L-arginine for our patient as the severity of his hyperammonaemia necessitated haemofiltration as the primary method of treatment. Haemofiltration should be used if there are signs of acute encephalopathy as this suggests developing cerebral oedema.

Prior to discussion with the metabolic diseases unit, we had initiated L-carnitine supplementation for a presumed diagnosis of valproate-induced hyperammonaemia. L-carnitine deficiency is associated with more severe hyperammonaemia in the context of sodium valproate use and its supplementation is therefore recommended when valproate is suspected to be the primary cause.

Outcome and follow-up

Four days into his admission, despite seizure control and cessation of sedation, our patient had no brainstem reflexes on neurological examination. We were advised by the regional metabolic diseases unit that his prognosis was very grave—survival being extremely unlikely with an ammonia level of over 500 μmol/L. At this stage, our patient’s ammonia level was above the recordable range (>1140 μmol/L). Confirmation of brainstem death was not possible using brainstem testing as his ammonia level, as well as his recent treatment with midazolam and thiopental infusions to control his seizures, rendered these tests invalid. In accordance with Faculty of Intensive Care Medicine guidance, we performed CT brain angiography which demonstrated marked hypoperfusion of the brainstem, consistent with brainstem death. We initiated discussion with his family regarding cessation of supportive care. However, before this took place, the patient developed worsening haemodynamic compromise and suffered a cardiac arrest, likely secondary to uncal herniation.

Prior to his death, we sent blood samples from our patient for genetic testing to allow confirmation of the underlying metabolic disorder. This later confirmed a hemizygous sequence variant suggestive of OTC deficiency. The regional metabolic diseases team arranged follow-up with the patient’s family to provide genetic counselling and subsequent screening for those at risk. Of significance in this case, on initially being approached, the wife of our patient did not attend her appointment and did not engage with initial attempts for follow-up of the family. While seeking her consent for this article, we explained to her again the reasons for screening of the patient’s relatives: identification of other family members affected with the aim of preventing a potentially catastrophic decompensation. Our patient’s wife described her grief in the wake of her husband’s very sudden death, that had left her feeling unable to engage with genetic counselling but wished to pursue this having understood the reasons for it. We were able to put her back in contact with the metabolic diseases team so that this could go ahead.

Discussion

Urea cycle disorders are a group of diseases caused by inborn errors of metabolism, of which OTC deficiency is the most common, with a reported incidence between 1:8000 and 1:44 000.^{11 12} It is more common in countries with higher rates of consanguinity. It is X-linked and therefore more common and more severe in men, classically presenting with drowsiness, failure to thrive or seizures in the early neonatal period. Fewer than 10% of cases present in adulthood, where symptoms are commonly neuropsychiatric or gastrointestinal.¹³ Deficiency of OTC causes an accumulation of ammonia, a toxic product of protein metabolism via the urea cycle. This results in cerebral oedema with encephalopathy progressing to seizures, coma and death if not rapidly identified and treated. In women, it is a well-documented cause of postnatal psychosis^14–16 and despite being X-linked, women may be severely affected.

If presenting in adulthood, patients with a less severe phenotype may become symptomatic when the affected individuals are exposed to metabolic stress such as intercurrent illness—in this case, a recent episode of severe tonsillitis. Once seizures occur, the catabolic load of a prolonged generalised tonic–clonic seizure will contribute further to protein metabolism and hyperammonaemia in a cascade that is difficult to reverse.

Our patient had also been given a diagnosis of schizophrenia. While impossible to confirm, it may be that low-grade intermittent hyperammonaemia was responsible for his neuropsychiatric symptoms over the course of his life. If this were the case, sodium valproate, prescribed to him as a mood stabiliser, would likely have increased his vulnerability to decompensation as it is known to cause hyperammonaemia even in the absence of urea cycle defects. There are several case reports describing the unmasking of urea cycle defects after initiation of sodium valproate.^{17 18} The severity of the hyperammonaemia and the biochemical profile were in keeping with OTC deficiency as opposed to valproate-induced hyperammonaemia alone (as subsequently confirmed through genetic testing). It is sobering to consider that a patient labelled with a psychiatric disease had a possible underlying organic cause for his symptoms and was started on a course of treatment that, although intended to treat his neuropsychiatric symptoms, could have instead worsened the underlying pathology. The psychiatrists’ robust insistence on ‘medical clearance’ prior to psychiatric assessment or admission should not be undervalued. It is also important to consider an underlying metabolic defect in patient’s having severe reactions to valproate, even if the ammonia level should fall with cessation of therapy.

Finally, in the immediate time after her husband’s death, the wife of our patient felt profoundly depressed, resulting in a disengagement with genetic testing. Screening and identification of family members is of profound importance in the prevention of tragic outcomes such as occurred in this case. Metabolic teams arranging follow-up should be aware of the potential for disengagement as a result of grief and consider further attempts to contact families who do not initially engage but may feel able to do so with the passing of time.

Patient’s perspective.

The next of kin did not wish to write a specific account of their perspective. A copy of this report was provided, and consent was sought for publication. Specifically, our patient’s wife agreed to discussion of her reasons for not initially engaging with the metabolic diseases team for follow up screening. We have included this in the main text.

Learning points.

• In patients presenting with status epilepticus, it is important to consider a wider net of differential diagnoses, including late presentation of metabolic diseases.

• This case highlights the vital importance of excluding underlying organic disease prior to diagnosis of a psychiatric condition.

• When a genetic condition is suspected, diagnostic investigations must be readily initiated even if they will not change presumptive management, given that formal diagnosis will incur the need for genetic counselling for the family.

• When approaching families, services must remain mindful of the effects of grief among bereaved relatives and the impact this may have on their engagement with screening.