Carbamazepine-induced Hyperammonemia and Asterixis in Young Adults

Carbamazepine (CBZ) is a USFDA-approved antiepileptic and mood-stabilizer for treating epilepsy, trigeminal neuralgia, and bipolar disorder. Its common adverse effects are nausea, vomiting, ataxia, drowsiness, and diplopia, while rare ones include transient elevation of hepatic transaminases (5%–10%), persistent leucopenia (2%), and very rarely, asterixis, hyperammonemia, aplastic anemia (1:200,000), and dermatological reactions such as toxic epidermal necrosis and Stevens-Johnson syndrome. ¹

Antiepileptics such as valproate (VPA, most common), topiramate, phenytoin, gabapentin, and, very rarely, CBZ are known to cause hyperammonemia, ² a metabolic condition characterized by raised levels of ammonia, a nitrogen-containing compound, and potent neurotoxin. Clinical presentation of hyperammonemia includes neurological signs and symptoms that may be acute or chronic depending on the underlying abnormality. While many patients remain asymptomatic, early symptoms include anorexia, nausea, vomiting, somnolence or insomnia, lethargy or agitation, personality changes, increased frequency of seizures, ataxia, and asterixis, with clinical signs of hyperammonemia occurring at concentrations >60 µmol/L.^2,3 Less commonly, especially if undetected or untreated, it can develop into life-threatening complications such as cerebral edema and brain herniation. ⁴

Asterixis (flapping tremors) is a common presenting feature of hyperammonemia, which can also be found in toxic and metabolic encephalopathies, focal brain lesions (such as thalamic, parietal, or mesencephalic), poisoning, and overdose of antiepileptics. ¹ Very few case reports have found co-occurrence of asterixis and hyperammonemia associated with the use of CBZ at 400–1200 mg/d.^5-10 We report CBZ-induced hyperammonemia and asterixis at lower doses of CBZ (400-600 mg/d) in three young adults diagnosed with schizoaffective disorder and complex partial epilepsy, intellectual disability with behavioral problems, and depressive disorder and generalized epilepsy. Naranjo et al.’s algorithm for adverse drug reaction (ADR) probability score was utilized to assess the causal relation. ¹¹ Psychiatric disorders were diagnosed as per the ICD-11 criteria. ¹² The laboratory reference range for plasma ammonia was 7-40 µg/dL.

Case Series

Case 1

A 24-year-old married female, homemaker, post-graduate, with well-adjusted premorbid personality and no significant personal or family history, had a past history of first episode non-affective psychosis lasting for 4 months about 5 years ago, with prior good compliance and remission on antipsychotics for nearly 2 years. Now, she presented with features of (second episode) psychosis and mania for 3 weeks, along with 3-4 episodes of new-onset complex partial seizures (CPS) in this period. She was diagnosed with schizoaffective disorder, multiple episodes, currently symptomatic (6A21.10), and CPS. Complete blood count (CBC), liver function tests (LFT), renal function tests (RFT), blood glucose (BG) levels, magnetic resonance imaging (MRI) of the brain, and electroencephalogram (EEG) were normal. She was treated with tab. CBZ 200 mg/d initially for 5 days, which was later titrated to 400 mg/d, along with tab. risperidone (RSP) 8 mg/d and tab. trihexyphenidyl (THP) 6 mg/d. We preferred CBZ over VPA as the patient was in the reproductive age group and the couple were aspiring for a child. After another episode of CPS during the third week despite compliance, the dose of CBZ was hiked to 600 mg/d. As there was inadequate response to manic symptoms even at week 5, tab. lithium 900 mg/d was added with serum lithium within the therapeutic level (0.79 mmol/L). Due to poor response to medications, she was given six sessions of modified electroconvulsive therapy (MECT). She developed asterixis after 3 weeks of increase in the dose of CBZ to 600 mg/d, after 3 days of last MECT, and after 2 weeks of addition of lithium, without features of delirium. Her plasma ammonia level was 280 µg/dL. CBZ was substituted with oxcarbazepine (OXZ; 600 mg/d) for control of CPS, while other medications were continued at the same doses. After discontinuation of CBZ, plasma ammonia level reduced to 145 µg/dL in 1 week and complete cessation of asterixis was noted in 2 weeks. She was discharged and has maintained remission of both psychosis and seizures for the past 1 year on tab. OXZ 600 mg/d, tab. RSP 8 mg/d, tab. THP 6 mg/d, and tab. lithium 900 mg/d without recurrence of asterixis or hyperammonemia.

Case 2

A 31-year-old unmarried female, born out of a consanguineous marriage, had a history of developmental delay, poor scholastic performance, and severe behavioral problems (irritability, agitation, and aggression), and the latter was being managed with tab. CBZ 400 mg/d. She was brought with third-person auditory hallucinations, delusion of reference, anorexia, and unmanageable aggression for 4 months. She received a diagnosis of mild disorder of intellectual development (6A00.0; IQ = 66) comorbid with schizophrenia-first episode (6A20.0) with psychomotor symptoms in primary psychotic disorders (6A25.4). CBC, LFT, RFT, BG, and computed tomography of the brain were normal. She received tab. RSP 6 mg/d and tab. THP 6 mg/d, and the dose of CBZ was hiked to 600 mg/d based on body weight (38 kg). Within a week, she developed nausea, ataxia, and asterixis with a plasma ammonia level of 154.6 µg/dL and normal sensorium. CBZ was discontinued, leading to the cessation of asterixis and ataxia and the return of plasma ammonia levels to 41.2 µg/dL within 5 days. She was discharged and has maintained remission of psychosis and behavioral problems for the past 3 months on tab. RSP 6 mg/d and tab. THP 6 mg/d, without recurrence of asterixis or hyperammonemia.

Case 3

A 28-year-old male, illiterate, driver presented with a history of epilepsy for 2 years; the seizures were well-controlled on medications. However, he had discontinued the antiepileptics and presented with three episodes of generalized tonic–clonic seizures in a span of 5 days. He also received a comorbid psychiatric diagnosis of single episode depressive disorder-severe without psychotic symptoms (6A70.3). CBC, LFT, RFT, BG, MRI brain, and EEG were normal. The patient received tab. VPA 1 gm/d, tab. fluoxetine (FLX) 40 mg/d, clonazepam 1.5 mg in divided doses for 1 week, and five sessions of MECTs (for suicidality). As the patient developed fresh episodes of generalized seizures, tab. CBZ was added at 200 mg/d during the second week and gradually titrated to 400 mg/d. Subsequently, he developed asterixis and ataxia 1 week after the last MECT and 2 weeks after the last generalized seizure, with plasma ammonia levels of 88 µg/dL, without features of delirium. CBZ dose was reduced to 300 mg/d, leading to cessation of asterixis, and reduction of plasma ammonia level to 37.9 µg/dL after a week. The patient was discharged and maintained remission of both depression and seizures for the past 1 month on tab. FLX 40 mg/d, tab. VPA 1 gm/d, tab. CBZ 300 mg/d, and tab. clonazepam 0.5 mg/d, without recurrence of asterixis or hyperammonemia.

Discussion

Asterixis generally indicates serious underlying pathology such as encephalopathy secondary to severe hepatic dysfunction, metabolic derangements, toxic agents, or medications. ¹ Hyperammonemia associated with antiepileptics is often attributed to hepatic failure and/or drug-induced liver injury but can also be observed in patients without signs of hepatic failure/injury, ¹³ as in our cases where organic causes were absent. Isolated hyperammonemia in the absence of elevated liver enzymes can also occur due to either hepatocellular dysfunction without cellular damage or isolated mitochondrial dysfunction.^7,13 While VPA induces hyperammonemia through fulminant hepatic failure, by increasing the levels of propionic acid that inhibits carbamyl phosphate synthetase enzyme, disrupting urea cycle, and consequently decreasing the ammonia metabolism, by decreasing ammonia elimination, or by decreasing glutamate metabolism,^2,3 the exact mechanism by which CBZ causes hyperammonemia is not known.^2,4

Rittmannsberger and Leblhuber ⁸ reported four cases with asterixis as a presentation of neurotoxicity and elevated plasma ammonia levels of 78 µmol/L at therapeutic doses of CBZ (400-800 mg/d) either alone or with concomitant phenytoin (200 mg/d) and phenobarbital (100 mg/d) or in combination of psychotropics such as clozapine and lithium wherein asterixis developed after adding CBZ. Rivelli et al. ⁷ reported the emergence of flapping tremors and hyperammonemia (60 µmol/L) in a case of bipolar disorder on lithium, chlorpromazine, and CBZ (800 mg/d). Similarly, Ambrosetto et al. ⁶ reported two cases with asterixis and only laboratory abnormality of “slight” hyperammonemia at therapeutic doses of CBZ (800-1200 mg/d), wherein the plasma ammonia level decreased and asterixis ceased upon CBZ discontinuation or reduction in its dosage (600 mg/d), thereby highlighting hyperammonemia as a possible dose-related side-effect of CBZ.

Adams et al. ⁹ reported a case of CBZ-induced hyperammonemia (127 µg/dL) without asterixis 3 weeks after initiating CBZ (600 mg/d) in a young male with bipolar disorder, on concomitant topiramate, olanzapine, desmopressine, and quetiapine, at therapeutic serum levels of CBZ (3.9 µg/dL; reference range 4-12 µg/dL). In contrast, Terzano et al. ⁵ reported asterixis in patients with neurogenic painful syndrome treated with CBZ at therapeutic doses that attained toxic serum levels. However, none of these patients were evaluated for plasma ammonia levels, and most were on polypharmacy.

Singh et al. ¹⁰ reviewed that most patients presented with co-occurring hyperammonemia and asterixis after CBZ (400-1200 mg/d) therapy, while few reports had not evaluated for ammonia levels, and only one case had hyperammonemia without asterixis. Further, authors reported a case of asterixis, confusion, and hyperammonemia at therapeutic doses of CBZ (600 mg/d) in a case of bipolar disorder that resolved after discontinuation of CBZ.

In our cases, asterixis and hyperammonemia appeared within 1-3 weeks of either an increase in dose or initiation of CBZ (400-600 mg/d) in the presence of normal neuroimaging findings and hepato-renal parameters. In addition, cessation of asterixis, ataxia, and hyperammonemia occurred within a week of discontinuation (cases 1 and 2) or dose reduction (case 3) of CBZ, while other concomitant medications such as RSP, lithium, OXZ, and VPA were continued without change in doses. The probability score in Naranjo et al.’s ¹¹ algorithm for ADR was nine in all our cases, both for hyperammonemia and asterixis, highlighting their definite causal relationship with CBZ. We believe that the asterixis and hyperammonemia in our cases may be dose-related (≥400 mg/d), as none of our cases had cerebellar signs secondary to CBZ “toxicity” and in one of our cases, dose reduction led to cessation of these ADRs. Serum CBZ levels could have substantiated this possibility, but it was not performed due to non-affordability by patients/caregivers.

Although lithium also can very rarely induce asterixis, ¹⁴ it does not induce hyperammonemia. Instead, individuals on chronic lithium therapy excrete urinary ammonia through mechanisms independent of urinary pH and likely to involve increased collecting-duct ammonia secretion via the ammonia transporter (Rhcg). ¹⁵ Although RSP can induce extrapyramidal side effects, hyperammonemia and asterixis have not been independently implicated, but only with concomitant antiepileptics such as VPA and CBZ. Rodrigues-Silva et al. ¹⁶ reported a case of hyperammonemic encephalopathy during VPA therapy and upon initiation of RSP, possibly due to RSP’s interference with VPA’s binding to albumin, raising free VPA levels, which would impair the urea cycle and reduce ammonia conversion, leading to a hyperammonemic encephalopathy. Similarly, the possibility of RSP’s interference with CBZ’s binding to albumin and its consequences cannot be ruled out. In case -3, VPA was prescribed for epilepsy and FLX for severe depressive disorder, while CBZ was added 2 weeks later due to poor control of seizures. Although contrasting findings for FLX-CBZ related cytochrome-P450 3A4 interactions (moderate inhibition by main metabolite of FLX–norfluoxetine) leading to increased ¹⁷ or unaltered ¹⁸ serum concentrations of CBZ have been reported, the possibility of drug–drug interactions in susceptible individuals cannot be ruled out. Concomitant VPA–CBZ administration is frequently associated with clinical anticonvulsant toxicity at the beginning of bitherapy, ¹⁹ possibly due to raised plasma CBZ-epoxide concentrations secondary to VPA’s inhibitory action on the glucuronidation of CBZ-10,11-trans-diol, and probably also due to the inhibition of the conversion of CBZ-10,11-epoxide to the trans-diol derivative. ²⁰ However, this can be addressed with gradual titration of CBZ and adjustment of its dose to tolerable levels. ¹⁹

Extreme muscle contractions during seizures induce ammonia production by deaminating adenosine monophosphate in the purine nucleotide cycle. In addition, muscle hypoxia during seizures may also lead to lactic acidosis, causing ammonia production in the red blood cells. ²¹ However, generalized tonic–clonic seizures-induced transient hyperammonemia (THA) returns to normal range within 6-8 h of the cessation of seizure. ²² Two case reports have observed hyperammonemic encephalopathy as a cumulative effect after 12-15 ECT sessions.^23,24 Our two cases that received 5-6 MECT sessions with adequate muscle relaxation during the procedure, with the occurrence of hyperammonemia after 3-4 days of the last MECT session, without features of delirium and lack of temporal relation between hyperammonemia and generalized seizures, highlight that hyperammonemia is more likely due to CBZ than ECT. However, the latter’s role cannot be undermined in cases of post-ECT protracted delirium and as the cumulative effect of multiple sessions of ECTs.^23,24

Temporally related onset of hyperammonemia and asterixis during CBZ therapy, that ceased after discontinuation of CBZ while concomitant medications in prescribed doses were continued without reappearance of the ADRs, along with Naranjo et al.’s probability score for ADR being definite, highlight CBZ as the offending agent in our cases. However, the role of MECTs, generalized tonic-clonic seizures, and concomitant antipsychotics and mood stabilizers cannot be undermined as a risk for such adverse effects in susceptible individuals.

Conclusion

Our case series suggests a causal link between asterixis, hyperammonemia, and CBZ, although further studies will be necessary to confirm our findings. Nevertheless, clinicians need to exercise a high index of suspicion for such drug–drug interactions during concomitant administration of CBZ, VPA, and RSP. Although ECT-induced THA is self-limited, it is prudent to consider plasma ammonia during protracted or delayed-onset delirium after ECT. ²³