Acute metabolic encephalopathy in two patients treated with asparaginase and ondasetron

Abstract

L-asparaginase has been an important agent of acute lymphoblastic leukemia (ALL) therapy for more than 40 years. The major adverse events of L-asparaginase are immunoallergic reaction, pancreatitis, liver cytolysis (increased ASAT, ALAT, and bilirubin), and cerebral haemorrhage or thrombosis. Hyperammonemia has already been described with L-asparaginase treatment [1]. In the present article, we report acute metabolic encephalopathy in two young boys receiving asparaginase for B ALL relapse. We discuss the pathophysiology of this metabolic encephalopathy with hyperammonemia and its medical management. A 20-year-old man was diagnosed with bone marrow relapse of acute B lymphoblastic leukemia. He was previously treated according to the FRALLE 2000BT protocol (actual French Acute Lymphoblastic Leukaemia Group recommandations). The tolerance of this first chemotherapy was quite good; in particular, he had received asparaginase according to the protocol without adverse event. He was in complete remission for 3 years. For his relapse, he was first treated by vincristine, mitoxantrone, intrathecal methotrexate, and one injection of PEG-asparaginase at Day 9 (1000 IU/m2). One week after the PEG-asparaginase injection, despite intravenous ondasetron administration (Zofran® 24 mg/D, IVC), he presented vomiting and somnolence. The neurological examination showed brisk tendon reflexes. Ondasetron was continued because of vomiting attributed to prolonged side effects of chemotherapy. No antiemetic acting neuroleptic was added. Finally, he became lethargic in 3 days. CSF analyses and neuroimaging (CT scan and MRI) were normal. EEG showed a pattern of metabolic encephalopathy. Metabolic studies revealed moderate hyperammonemia (129 μmol/L [Nl < 50 μmol/L]) with normal liver function. Plasma aminoacids showed high glutamic acid: 284 μmol/L (Nl < 60 μmol/L), normal glutamine and total depletion in asparagine: 0 μmol/L. Hyperammonemia was treated with a low protein intake with IV sodium benzoate (350 mg/kg/D), and p.o. sodium phenylbutyrate (250 mg/kg/D) as alternative nitrogen excretion pathway. Moreover, we observed a huge increase of plasmatic phenylalanine: 1233 μmol/L (Nl < 90 μmol/L), probably related to the concomitant administration of ondasetron. As this drug contains aspartam (a precursor of phenylalanine), this drug was stopped. Within 2 days, the patient fully recovered from all neurological impairment while ammonia and EEG normalized. The treatment was stopped when asparagine was normalized, 35 days after the PEG-asparaginase injection. This management avoided an invasive procedure (hemodialysis) in this neutropenic patient. A few days after neurological recovery, chemotherapy was pursued and resulted in the second complete remission. Three months later, the patient underwent allogeneic stem cell transplantation. A 15-year-old boy was diagnosed with early acute T lymphoblastic leukemia bone marrow relapse, diagnosed during the maintenance treatment according to the FRALLE 2000BT protocol. In the first line of treatment, he had received all chemotherapy with a quite good overall tolerance. In particular, he had shown no side effects with asparaginase. For his relapse, he was included in the VANDEVOL protocol. This paediatric Phase I study, open in France since December 2009, includes clofarabine, vindesine, mitoxantrone, etoposide, intrathecal methotrexate, and L-asparaginase (four injections of kidrolase® 10000 IU/m2/dose at D7, 9, 11, 13). At D15, 2 days after the last asparaginase injection, despite ondasetron administration (Zofran® 24 mg/J, IVC), he presented vomiting. Zofran was continued and steroids added. Two days later, he continued with vomiting and lethargy. We considered metabolic encephalopathy. No brain exploration was performed. Effectively, we found the same biological pattern with hyperammonemia: 143 μmol/L (Nl < 50 μmol/L), normal liver function, high glutamic acid: 232 μmol/L (Nl < 60 μmol/L), normal glutamine and total depletion in asparagine: 0 μmol/L. Hyperammonemia was treated as in the first case with low protein intake, IV sodium benzoate (350 mg/kg/D) and p.o. sodium phenylbutyrate (250 mg/kg/D). We also observed a high phenylalanine level: 2242 μmol/L (Nl < 90 μmol/L). Ondasetron was stopped. As in the first case, clinical and biological abnormalities normalized within 2 days. Metabolic treatment was stopped when asparagine level rise up again. Chemotherapy was then continued. Neurological complications can occur from different origins in patients with chemotherapy (toxicity, infection, haemorrhage, etc.). Hyperammonemia must be systematically screened in patients presenting encephalopathy during chemotherapy, mostly in those with asparaginase treatment. This complication had already been described once, 24 years ago [1]. As leukemic cells do not express asparagine synthetase, they depend on exogenous asparagine uptake [2]. Asparaginase hydrolyzes asparagine to L-aspartic acid and ammonia, resulting in asparagine depletion and then acute lymphoblastic leukaemia cells apoptosis. The severe asparagine depletion leads to the diversion of glutamine metabolism toward the synthesis of asparagines. Concomitantly, asparagine is deaminated by asparaginase leading to a large accumulation of ammonia in the plasma. Moreover, Asparaginase also has a glutaminase activity which metabolizes glutamine into glutamic acid and ammonia. Both Asparaginase and glutaminase activity of the chemotherapy explain the accumulation of ammonemia (Fig. 1). In this situation, even if the glutamine level is normal, this aminoacid remains the ammonium donor responsible for the hyperammonemia. Then, we proposed a treatment similar to what is proposed in urea cycle disorders, aiming to decrease glutamine concentration and consequently ammonemia. The medical management included a decrease in protein intake and lactulose, which was given to reduce intestinal ammonia production. Moreover, Sodium Benzoate and Sodium Phenylbutyrate induced a waste of nitrogen by creating alternative pathways for nitrogen excretion [3], as there is a functional deficiency of the urea cycle pathway. Benzoate reacts with glycine by acetylation to produce hippurate, which is eliminated in urine (excretion of 1 mol of nitrogen per mole of benzoate). Phenylbutyrate reacts with glutamine via an acetylation process allowing the production of phenylacetylglutamine, which will also be excreted in urine (2 mol of nitrogen by mole of phenylbutyrate). This treatment (monitored on ammonia and glutamic acid levels) appeared to be effective in our two patients and was given as long as asparagine depletion related to asparaginase treatment persisted. In the two patients, the clinical state was worse than could be expected from ammonia levels. The efficiency of the management in the two cases proves that these symptoms were really related to the metabolic disturbance. This treatment, then, must be prescribed urgently in case of clinical worsening, because the only alternative treatment would be an extra renal dialysis, which is a much more invasive and dangerous procedure for these immunosuppressed patients. Physiopathology of hyperammonemia following asparaginase treatment. In normal situation (a), the nitrogen is excreted as urea through the urea cycle. The nitrogen coming from aminoacids catabolism is added to glutamate to synthetize glutamine. The glutaminase enzyme release ammonia from glutamine. This ammonia is normally transformed with glutamine in urea through the urea cycle and the nitrogen can be excreted as urea in urines. In case of L-asparaginase treatment (b), there are two cumulative effects which induce hyperammonemia. First, Asparagine is deaminated by asparaginase leading to an important accumulation of ammonia in the plasma. The glutamine metabolism is diverted to asparagine synthesis, which is immediately deaminated by asparaginase leading to a metabolic cycle of ammonia production. Second, asparaginase has also a glutaminase activity that metabolizes glutamine into glutamic acid and ammonia by the same deamination process. It induces a paradoxal pattern of high glutamic acid and low glutamine levels associated with hyperammonemia. The urea cycle is probably overloaded by the ammonia production and ammonemia rise up in blood. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] As we were confront with these patients, we systematically checked ammonia levels in eight paediatric patients (8-months to 14-year old) treated with L-asparaginase (kidrolase® 10000 IU/m2/dose or erwiniase 20000 IU/m2/dose) or PEG-asparaginase (1000 IU/m2). We measured the ammonia level after three or four L-asparaginase injections (48 hr after the last dose) or 1 week after one injection of Peg-asparaginase. None of these eight children was treated by Zofran IVC. We systematically found moderate hyperammonemia (75–210 μmol/L) without neurological signs. We also found high glutamic acid every time and low or normal glutamine plasma concentrations but never high phenylalanine level (Table I). It seems that hyperammonemia is almost a constant biological finding but neurological symptoms remain rare [4]. We can note that the two cases were previously treated with asparaginase without neurologic disorder. A prospective study of the prevalence of clinical manifestations due to hyperammonemia following asparaginase is required to determine the risk factors which can induce the clinical expression of these hyperammonemic episodes. As the hyperammonemia observed in our two patients was found to be insufficient to induce neurological disorders, we checked for other associated metabolic disorders. In both cases, we found very high phenylalanine levels because of ondasetron treatment, which was not observed in the other patients with hyperammonemia but without encephalopathy. This may have played a role because high levels of phenylalanine can increase the brain water content and can compete with the intake of essential aminoacids into the brain [5]. It is possible that the decrease in phenylalanine plasma concentration (induced by the withdrawal of ondasetron treatment) participated in the favorable outcome of the patients. As acute encephalopathy has never been seen in patients treated by ondasetron without asparaginase treatment, we can make the assessment that the neurological symptomatology is probably because of the combination of hyperammonemia and hyperphenylalaninemia, which could both be responsible for brain toxicity, even if this pathophysiology mechanism remains a hypothesis. To conclude, ammonia and plasma aminoacids must be measured in every patient under asparaginase chemotherapy presenting acute encephalopathy, and, in case of metabolic derangement, a specific management must be prescribed to avoid a more invasive procedure.