Neonatal Hyperammonemia Due to Ureaplasma Sepsis

We describe a baby with severe hyperammonemia who was initially suspected to have an inborn error of metabolism but instead had Ureaplasma sepsis. Hyperammonemia from Ureaplasma infection is well-documented in immunocompromised adults, but the phenomenon has not been described in neonates, in whom hyperammonemia is usually assumed to represent a hereditary metabolic disease. This 36-week gestation baby was transferred from another hospital because of metabolic acidosis, respiratory distress, and suspected seizures. His mother's pregnancy was complicated by maternal diabetes, premature rupture of membranes, and a 3-day history of vaginal bleeding. He was born limp, lethargic, and cyanotic, with Apgar scores of 3 and 7. On day 2 of life, he required intubation because of apnea and metabolic acidosis. Abnormal facial movements and posturing were initially suspected to represent seizures, so he was loaded with phenobarbital and levetiracetam. He also received empiric antibiotics and antiviral medications. Antiseizure medications were halted after continuous electroencephalography showed no epileptiform discharges during his abnormal movements. Blood cultures and cerebrospinal fluid analysis were unremarkable, aside from an elevated cerebrospinal fluid protein. A respiratory culture for Ureaplasma was negative, but next-generation DNA sequencing of serum confirmed evidence of Ureaplasma urealyticum, for which he received azithromycin. His initial serum ammonia level was dramatically elevated (1284 μg/dL). His ammonia level increased to 1374 μg/dL despite the infusion of sodium benzoate and sodium phenylacetate, and he began continuous kidney replacement therapy. Urine organic acids, plasma amino acids, serum pyruvate, and carnitine were normal. Genetic testing was not completed due to his improving clinical condition, his resolving hyperammonemia, and the Ureaplasma sepsis diagnosis. At 1 week of age, multifocal cerebellar hemorrhages were documented on ultrasound and computed tomography (Figure 1). The hemorrhages were also evident with magnetic resonance imaging (MRI). No hemorrhages were identified in other areas of the brain, nor did the MRI reveal abnormalities suggestive of inborn errors of metabolism. By 3 weeks of age, his condition had improved and his ammonia level had fallen to 63 μg/dL. At 13 months of age, he was starting to walk, playfully interacting, and saying several words. He has experienced no seizures or periods of lethargy. This child was transferred due to suspected seizures, but continuous electroencephalography showed no epileptiform discharges, even during the movements. His serum ammonia level was dramatically elevated, leading to the initial suspicion of an inborn error of metabolism. However, neither metabolic testing nor MRI showed evidence of hereditary metabolic disorders, and the subsequent resolution of his hyperammonemia and his normal outcome further support the absence of a genetic disease. The diagnosis of Ureaplasma urealyticum sepsis was based on next-generation DNA sequencing of serum to detect microbial cell-free DNA (cfDNA). This is a clinically validated diagnostic tool that can detect microbial cfDNA in up to 94% of the pathogens identified by blood culture in patients with sepsis [1]. Ureaplasma urealyticum is a tiny fastidious prokaryotic organism that is difficult to visualize or culture [2]. These organisms release substantial amounts of ammonia during urea hydrolysis, sometimes leading to clinical hyperammonemia [2, 3]. The diagnosis of hyperammonemia due to Ureaplasma sepsis is particularly challenging in neonates, whose hyperammonemia is likely to be mistakenly attributed to an inborn error of metabolism. Isolated cerebellar hemorrhages have been documented in infants with various organic acidemias [4-6]. Propionic, methylmalonic, and isovaleric acidemia typically present in babies as acute metabolic decompensation and encephalopathy, often associated with hyperammonemia. The occurrence of cerebellar hemorrhages in our patient with acquired hyperammonemia suggests that the cerebellum of neonates may be particularly vulnerable to hyperammonemia, whatever its origin. Catherine A. Kronfol: conceptualization, investigation, writing–review and editing. Aaron W. Hocher: conceptualization, writing–original draft, writing–review and editing. E. Steve Roach: conceptualization, writing–original draft, writing–review and editing, supervision. E. Steve Roach is the editor-in-chief of Annals of the Child Neurology Society. The other authors declare no conflicts of interest.