Keywords hyperammonemia - hyperammonemic encephalopathy - common acute lymphoblastic leukemia
Introduction
Hyperammonemia (HA) is defined as an elevation of plasma ammonia concentrations above
the normal range. There are several well-known risk factors in childhood, such as
chemotherapy, septic shock, gastrointestinal bleeding, renal diseases, severe muscle
exertion, heavy exercise, inborn errors of metabolism (e.g., urea cycle disorders),
acute liver failure due to hepatotropic infectious diseases or liver autoimmune diseases,
Reye syndrome, parenteral nutrition, urinary tract infection, and drug-related toxicity
(e.g., valproate or barbiturates).[1 ]
[2 ]
[3 ]
[4 ] The clinical presentation is heterogeneous ranging from mild symptoms like nausea
to acute life-threatening conditions with encephalopathy, cerebral edema, and coma.[1 ] Since the 1980s, there have been reports on HA following intensive chemotherapy
and bone marrow transplantation. The term idiopathic hyperammonemia (IHA) has been
suggested to describe this condition.[2 ]
[3 ] IHA is defined as elevated concentration of ammonia in serum with normal test results
for liver enzymes, and no evidence for inborn errors of metabolism or other identifiable
causes.[2 ]
[3 ] The underlying pathophysiological mechanism of this often fatal condition is not
yet fully understood and is very likely to be multifactorial.[3 ]
[4 ]
[5 ]
[6 ]
[7 ]
[8 ]
[9 ]
[10 ] There are several available treatment options for HA. Many of them were adopted
from the management of inborn errors of metabolism with HA, such as arginine (improvement/stimulation
of nitrogen excretion by the Krebs–Henseleit cycle). The main goals are the removal
of nitrogen by enhancing nitrogen excretion by hemodialysis and ammonia-trapping therapy
(alternate pathway therapy with sodium benzoate, sodium phenylacetate or sodium phenylbutyrate)
and the reduction of exogenous nitrogen load (optimization of enteral/parenteral nutrition).
To reduce the intestinal source of ammonium lactulosis and neomycin/rifaximin are
used. Treatment should be started early and be regularly reevaluated.[4 ]
We describe a unique clinical phenotype of IHA associated with marked tumor lysis
syndrome (TLS) after the administration of dexamethasone and mitoxantrone. The role
of mitoxantrone as a causative agent for IHA in children is elusive.
Case Presentation
Our patient, an 8-year-old male, presented with common lymphoblastic leukemia (central
nervous system [CNS] not involved) and received chemotherapy according to the CoALL-08/09
protocol (German Co-operative Study Group for Treatment of Acute Lymphoblastic Leukemia
08–09, Low-Risk group) from March 2017 to October 2017. From October 2017 to February
2018, he was on oral maintenance chemotherapy with methotrexate and 6-mercaptopurine.
In February 2018, our patient presented with progressive fatigue and reduced general
condition. Laboratory analysis revealed anemia (7.5 g/dL, reference range (RR): 11.8–14.8),
thrombocytopenia (7.0 G/L, RR: 195–464), and leukopenia (2.73 G/L, RR: 3.50–14.00).
The diagnostic workup revealed a relapse of common acute lymphoblastic leukemia. The
treatment was started according to the IntReALL-HR protocol (International Study for
Treatment of High Risk Childhood Relapsed Acute Lymphoblastic Leukemia 2010, High-Risk
group) with dexamethasone (20 mg/m2 , intravenously, twice a day) from therapy day 0 to 5 and mitoxantrone (10 mg/m2 , intravenously, once a day) on therapy day 0. Furthermore, he received a concomitant
intravenous hydration therapy (4,000 mL/m2 per day). He developed severe TLS with hyperuricemia (9.2 mg/dL, RR: 1.9–5.0), hyperphosphatemia
(7.4 mg/dL, RR: 3.0–5.4), hyperkalemia (6.6 mmol/L, RR: 3.3–4.7), and lactic acidosis
(pH = 7.22, RR: 7.35–7.43, lactate = 13.3 mmol/L, RR: 0.5–2.2, base excess: −14.6
mmol/L, RR: −1.5 to 3.0). The coagulation parameters were slightly abnormal (international
normalized ratio: 1.5, RR: 0.8–1.2, and partial thromboplastin time: 45 seconds, RR:
25–42). The liver function studies were normal (total bilirubin, alanine-aminotransferase,
and γ-glutamyltransferase). There was no evidence for acute kidney injury. Repeated
blood cultures remained sterile. On therapy day 1, he developed a qualitative and
quantitative disorder of consciousness in an undulating manner with confusion, agitation,
disorientation, anxiousness, and somnolence, respectively. An electroencephalography
revealed a generalized slowing, consistent with metabolic encephalopathy ([Fig. 1 ]). A computed tomography of the brain showed no apparent morphological involvement
of the CNS. A magnetic resonance imaging on the following day confirmed the results.
A sonography of the abdomen revealed a hepatosplenomegaly and diffuse leukemic infiltration
of liver, spleen, gut, and kidneys. There was no evidence for hepatic veno-occlusive
disease. With a lumbar puncture, malignant cells and a CNS infection could be excluded.
Strikingly, the concentration of ammonia in serum was significantly elevated (213
µmol/L, RR: 16–53). Sampling was performed by using a triple-lumen Hickman catheter.
The sample was transported refrigerated. The measurement was performed immediately
upon arrival at the laboratory. Additionally, analysis of organic acids in urine,
amino acids, and carnitine/acylcarnitine in plasma did not indicate an inborn error
of metabolism associated with HA. Moreover, a comprehensive workup for hepatotropic
infectious diseases and autoimmune liver diseases showed no significant findings.
Fig. 1 Electroencephalography with generalized slowing and periodic pattern well consistent
with metabolic encephalopathy.
Fig. 2 Plasmatic concentration of ammonia versus hours since onset of clinical symptoms.
The dashed line indicates the threshold (= 53 µmol/L) of pathological concentration
of ammonia in serum.
The TLS could be managed sufficiently with hydration, rasburicase, and forced diuresis.
The patient received crystalloid fluids only. Our management of HA comprised lactulosis
three times a day (3 × 3.3 g) and rifaximin twice a day (2 × 200 mg) via nasogastral
tube to reduce the intestinal source of ammonium and sodium benzoate (100 mg per kg
per day, continuously) intravenously for ammonia-trapping therapy. After a subsequent
plateau phase with slightly elevated and fluctuating results for the concentration
of ammonia in serum (with a range from 54 to 80 µmol/L), they returned to normal after
165 hours ([Fig. 2 ]).
Discussion
IHA is generally regarded as a diagnosis of exclusion. Since we do not have any identifiable
causes for HA in our patient, we suspect a case of IHA. IHA was previously described
in the literature in the context of high-dose chemotherapy for hematological malignancies,[3 ]
[4 ]
[5 ]
[6 ]
[7 ] chemotherapy in solid tumors,[11 ] bone marrow transplantation for malignant,[2 ]
[3 ]
[8 ]
[9 ]
[10 ] and nonmalignant diseases (Hurler syndrome) and solid organ transplants (lung transplant).[12 ]
[13 ] IHA is a rare, but frequently lethal complication, despite ammonia-trapping therapy
and hemofiltration.[5 ]
[6 ] Thus, we hypothesize a mild and unusually early-onset form of IHA in our patient,
who received dexamethasone and mitoxantrone on the first therapy day.
The occurrence of IHA during initiation of chemotherapy for leukemia is highly unusual
and was described in a pediatric population only twice: once on day 15 in a 15-year-old
male with newly diagnosed acute lymphoblastic leukemia and initiation of chemotherapy
with steroids only[14 ] and another time in a case of a 13-month-old male with a lethal IHA on day 35 after
induction therapy with seven different chemotherapeutic drugs.[6 ]
The only report of IHA in combination with mitoxantrone described the case of a 59-year-old
woman who received a polychemotherapy including mitoxantrone.[15 ] However, mitoxantrone was linked to mitochondrial cardio toxicity in a dose-dependent
manner during long-term use in a mouse model.[16 ] An acute adverse reaction to mitoxantrone cannot be ruled out in our patient. Yet,
a study in de novo diagnosed lymphoblastic leukemia could not show a higher toxicity
for mitoxantrone compared with daunorubicin.[17 ]
Many chemotherapeutic agents have been linked to IHA: vinciristine, busulfan, cyclophosphamid,
etoposide, methotrexate, cytarabain, daunomycin, and amsacrine.[4 ]
[5 ]
[18 ] However, some chemotherapeutic agents seem to cause less devastating outcomes when
it comes to HA, as for instance asparaginase and 5-Fluorouracil.[4 ] Some reports suggest specific adverse drug effects in these cases. For instance,
treatment with asparaginase for acute leukemia has been linked to transient HA.[19 ]
[20 ]
[21 ] Asparaginase hydrolyses asparagine in aspartic acid and ammonia as a mechanism of
action. Measuring plasma concentrations of ammonia was previously suggested to be
used as a parameter for drug efficacy and toxicity.[22 ]
[23 ]
[24 ] The intermediate product of 5- Fluorouracil, fluoracetate, is regarded as a direct
inhibitor of the urea cycle, thus leading to elevated plasmatic ammonia concentrations.[4 ]
[25 ]
For a functioning homeostasis of ammonia a functioning Krebs–Henseleit cycle and a
balanced ratio of protein anabolism and catabolism are crucial.[26 ] The capacity of the Krebs–Henseleit cycle varies in humans without clinical relevance.[19 ]
[27 ] However, when HA is combined with a previously unknown latent ornithine transcarbamylase
deficiency, a defect of the urea cycle, lethal outcomes have been reported.[23 ] Since we do not have evidence for underlying enzymatic defects in the Krebs-Henseleit
cycle in our patient and no chemotherapeutic agents with known transient HA as a side-effect
were used, we hypothesize that our patient had a mild form of IHA. A transient form
of HA due to massive acute protein breakdown in TLS should be considered as a differential
diagnosis. In addition, liver infiltration with leukemia blasts may have exaggerated
the metabolic decompensation.
Conclusion
Children undergoing chemotherapy are often temporarily in a poor general condition
and the clinical presentation of HA is heterogeneous. This could be one explanation
for HA being potentially under-recognized and thus under-diagnosed in TLS.
This case report should be a reminder for pediatric intensive care doctors to check
for HA in children with TLS and acute neurological deterioration. Routine measurements
of plasma ammonia concentrations in high-risk patients could lead to earlier diagnosis.
In addition, latent metabolic disorders and drug-induced HA should be taken into consideration.
To conclude, in children with acute alterations in consciousness of unknown origin
during chemotherapy and especially with TLS, HA should be considered and closely monitored.
