Secondary Logo

Journal Logo

Utility of Neuroimaging in Children Presenting to a Pediatric Emergency Department With Ataxia

Luetje, Maureen, DO*; Kannikeswaran, Nirupama, MD*; Arora, Rajan, MD*; Wang, Bo, MPH, PhD; Farooqi, Ahmad, MPhil, MSC, MA; Sivaswamy, Lalitha, MD*

doi: 10.1097/PEC.0000000000001823
Original Articles

Objectives The aim of this study was to evaluate the utility of neuroimaging in children who present to the pediatric emergency department with acute-/subacute-onset ataxia. Neuroimaging is performed in many children with ataxia to rule out serious intracranial pathology. There is, however, limited evidence to support such practice.

Methods This was a retrospective review of electronic medical records of children who presented to the emergency department with ataxia between 2007 and 2013. Patient demographics, historical features, physical examination findings, laboratory results, and neuroimaging results were collected. Neuroimaging studies that were classified as abnormal by a neuroradiologist were further reviewed and classified by the study neurologist as clinically significant or not.

Results The records of 141 subjects were analyzed. The most common causes of ataxia were infectious/postinfectious (36.2%) and ingestion (15.6%). Neuroimaging was performed in 104 children (73.8%). Neuroimaging was abnormal in 63 children (60.6%). However, these abnormalities were clinically significant in only 14 children (13.5%). Focal neurological findings were noted in 12 of 14 children (85.7%) with clinically significant neuroimaging.

Conclusions Clinically significant neuroimaging was noted in a minority of children who presented with acute/subacute ataxia. The majority of patients with clinically significant neuroimaging had focal neurological findings on examination. Neuroimaging may not be required in all children presenting to the ED with acute ataxia, but further large-scale studies are needed to validate these findings and identify a subset of patients with ataxia in whom imaging can be deferred.

From the *Children's Hospital of Michigan and

The Integrative Biosciences Center and

Department of Pediatrics, Wayne State University, Detroit, MI.

Disclosure: The authors declare no conflict of interest.

Reprints: Maureen Luetje, DO, Pediatric Emergency Department, Children's Hospital of Michigan, 3901 Beaubien Blvd, Detroit, MI 48201 (e-mail:

The prevalence of childhood ataxia is estimated to be 26 per 100,000 children.1 While the majority of causes of acute ataxia in pediatrics are benign, ataxia can be the presenting symptom of serious intracranial pathology or the initial manifestation of a chronic or progressive neurological disease.2

Neuroimaging is recommended and often performed in children who present to the pediatric emergency department (ED) with acute onset of ataxia.3 However, a systematic review of the literature, conducted by Whelan et al,4 indicated that the diagnostic yield of such neuroimaging studies may be low. Abnormalities of clinical significance were identified in only 2.5% of patients who underwent brain computed tomography (CT) and 5% of patients who received brain magnetic resonance imaging (MRI). The authors concluded that urgent neuroimaging is indicated for children with ataxia only in the presence of altered mental status, focal neurological signs, history of trauma, or when there is a clinical concern for a mass lesion.4 There is evidence to suggest that a detailed history and physical examination are far more likely to identify the etiology of acute ataxia than a battery of screening tests and that further investigations should be tailored to the clinical presentation at hand.5 Currently, evidence-based guidelines regarding neuroimaging for children who present with acute-onset ataxia to the ED are lacking.

The primary objective of this study was to evaluate the utility of neuroimaging in previously healthy children who presented to our pediatric ED with acute- or subacute-onset ataxia.

Back to Top | Article Outline


Study Setting

The study was conducted in an urban, academic children's hospital in Detroit, Mich, with an annual volume of 90,000 visits to the pediatric ED. This area is not endemic to any infectious organisms associated with cerebellitis. At this institution, pediatric neurologists are available for consultation in person or over the phone at all hours.

Back to Top | Article Outline

Study Design and Inclusion and Exclusion Criteria

A retrospective chart review was performed to identify patients evaluated for acute or subacute ataxia in our pediatric ED from 2007 to 2013. Participants included previously healthy children between the ages of 15 months and 18 years. Study patients were identified using the International Classification of Diseases, Ninth Revision, Clinical Modification selecting the diagnostic codes for gait abnormality, cerebellar symptoms, and ataxia.

Children younger than 15 months were excluded because of difficulty in accurately diagnosing ataxia in children of this age group. Children with ventriculoperiotoneal or ventriculoatrial shunts, developmental delay, and cerebral palsy were also excluded. Furthermore, children who had an established diagnosis of a condition where ataxia may be a primary or associated feature, such as multiple sclerosis, mitochondrial encephalopathy lactic acidosis and stroke-like symptoms, ataxia-telangiectasia, or Friedreich ataxia were excluded. However, if the child presented with ataxia and was diagnosed with multiple sclerosis after the completion of workup, they were included in the study. The study flowchart with reasons for exclusions is shown in Figure 1.



Back to Top | Article Outline

Study Definitions

For purposes of this study, acute ataxia was defined as symptom onset within 72 hours prior to presentation to our pediatric ED (consistent with the time frame used in the 2013 review of Whelan et al4), and subacute ataxia was defined as symptoms starting less than 1 week prior to initial evaluation, in a child with no underlying neurological disorder, as outlined above.

Neuroimaging of the study patients was reviewed and interpreted by our institution's pediatric neuroradiologists, who were not part of the study. The study's neurologist then reviewed all imaging deemed abnormal by the neuroradiologist and determined whether the abnormality was considered clinically significant. Clinically significant neuroimaging findings were defined as any radiological finding that resulted in medical or surgical treatment that would not have been otherwise initiated but for the results of the test. Furthermore, in order to avoid duplication of study results, patients with an abnormal CT and MRI were considered a single abnormal study.

Physical signs on neurological examination were classified as “focal” if they were findings that were not germane to a diagnosis of ataxia. Ataxia can be the result of pathology in the vestibular pathways, their connections to the cerebellum, the cerebellum itself, or in the central connections of the cerebellum to the brain stem or the sensory pathways. Because cerebellar signs such as past pointing, intention tremor, nystagmus, head titubation, and truncal instability would be considered within the spectrum of findings for a child with a self-remitting ataxia, they were excluded as focal findings. We used the term “focal” to indicate a disease process likely to explain the etiology of ataxia and could potentially be localized to a certain part of the central or peripheral nervous system. Focal findings were therefore defined as hemiparesis, paraparesis, cranial nerve signs other than nystagmus, papilledema, pathologically brisk or absent tendon reflexes, positive Babinski sign, aphasia, or dysphasia.

We used the following definitions for diagnostic categories of ataxia in our study. We defined “toxicity” as ataxia that was secondary to a medication that was prescribed to the child but was found to be above therapeutic levels (such as a child who was on an anticonvulsant for seizure disorder, which could cause ataxia when levels are supratherapeutic). Ingestion was defined as the intake of a medication or substance, which the patient was not prescribed and is known to cause ataxia, such as the ingestion of a benzodiazepine. Patients who reported a viral prodrome and/or whose examination revealed an infectious agent with otherwise benign workup were classified as postinfectious cerebellitis. Lower motor neuron conditions were defined as those causes of ataxia secondary to involvement of the anterior horn of the spinal cord or peripheral nerves and included Guillain-Barŕe syndrome and peripheral neuropathies. Finally, cases of those children who had a reported history of ataxia without any prodromal viral symptoms and had a relatively normal physical examination and whose ataxia resolved over the course of the ED visit were classified as ataxia not otherwise specified (NOS).

Back to Top | Article Outline

Data Collection

The data collected included patient demographics, pertinent historical data (history of ingestion, viral prodromal symptoms, recent immunization), presenting signs and symptoms, physical examination findings including neurological examination, laboratory studies performed, results of neuroimaging studies, and ED disposition. Both the emergency physician's notes and the neurology consultation note were reviewed for “neurological examination findings. Discrepancies between the neurological examination findings of the ED's clinician and neurology documentation were reconciled with preference to the neurologist's documentation. For the subset of patients who were discharged home from the ED, further review of the medical records was performed for up to the next 5 years. Return visits to the ED or any subsequent neurology clinic notes were reviewed for any further complaints of ataxia. This was done to ensure no serious diseases were missed during the initial evaluation of the patient.

Back to Top | Article Outline

Statistical Analysis

A standardized data collection instrument using Microsoft Excel (Microsoft Corp, Redmond, Wash) was created to record study variables. Double data entry with an Excel spreadsheet was performed to ensure accuracy of data collection. The principal investigator entered all study data. Descriptive statistics (means and proportions) were used to describe the study sample. All statistical analyses were performed using the SAS 9.4 statistical software package (SAS Institute Inc, Cary, NC).

Back to Top | Article Outline


A total of 141 patients were evaluated in our pediatric ED for acute/subacute ataxia during the study period. Of those, 70.2% (99/141) were subsequently evaluated by a pediatric neurologist. The mean age of the patients was 5.3 ± 4.5 years. The majority of patients were male (56%). Most children who were evaluated presented with acute ataxia (80.1%) and were admitted to the hospital (83.7%) from the ED. Of the 20 children who were discharged from the ED, 13 (65%) were seen in follow-up within our hospital system, and none had any further examination findings or symptoms of ataxia at a subsequent visit.

The most common causes of ataxia in our study cohort included postinfectious cerebellitis, which occurred in 51 patients (36.2%), and ingestions, which occurred in 22 patients (15.6%) (Fig. 2). Significant causes of ataxia requiring emergent medical or surgical treatment were noted in 16 patients (11.3%), including 8 central nervous system (CNS) tumors (5.6%), 4 cases of CNS vascular malformation/stroke (2.8%), 2 patients with encephalitis (1.4%), and 2 patients with Guillain-Barŕe syndrome (1.4%).



Neuroimaging (CT, MRI, or both) was performed in 73.8% (104/141) of the patients. Computed tomography of the brain was performed in 41.1% (58 /141) of patients, MRI of the brain was performed in 51.1% (72/141) of patients, and 18.4% (26/141) of patients had both a CT and MRI. Computed tomography was abnormal in 17 patients (29.3%), and MRI was abnormal in 46 patients (63.9%) (Fig. 3). Eight children had a normal CT scan but were found to have an abnormality on MRI, although none of these abnormalities were clinically significant. No cases occurred where a patient had a normal MRI but abnormal CT.



Neuroimaging abnormalities were classified as clinically significant by the study neurologist in 13.5% (14/104) of patients (Table 1). Two of the 14 patients (14.3%) with neuroimaging abnormalities had a nonfocal neurological examination on initial presentation in our ED. One of these patients was diagnosed with acute demyelinating syndrome, and the other one had an ependymoma. The child with the ependymoma presented with ataxia, head tilt, and worsening headache over a period of 2 weeks prior to presentation. The child with the demyelinating syndrome had worsening headaches for several weeks, along with projectile emesis.



Focal neurological deficits were found in 11.3% (16/141) of patients (Table 2). Six of the patients (37.5%) with focal neurological deficits had no abnormalities noted on neuroimaging; 2 of these patients were diagnosed with Guillain-Barŕe syndrome, and the other 4 were diagnosed with acute cerebellar ataxia, dystonia, vestibulitis, and idiopathic left lateral rectus nerve palsy.



Funduscopic examination was performed in 34 of 141 patients (24.1%) in the study cohort. Of the study patients who had “clinically significant” imaging, only 5 had a documented funduscopic examination. Papilledema was noted in 1 patient who was eventually diagnosed with a medulloblastoma.

Back to Top | Article Outline


This single-center study revealed that 13.5% of healthy children with acute- and subacute-onset ataxia have significant intracranial pathology on neuroimaging. All except 2 children with serious CNS pathology had focal neurological findings on presentation. The 2 children who did not present with focal neurological findings underwent neuroimaging on the basis of their prominent symptoms, as noted in RESULTS.

Neuroimaging has become ubiquitous in the pediatric ED setting for a variety of neurological symptoms ranging from dizziness6 to headache,7 first-time seizures,8 and concussion.9 However, most studies note the low yield of such routine practice, especially in situations wherein there are no concurrent historical factors or physical examination findings that suggest specific neurological pathology. In our cohort, most abnormalities noted on neuroimaging did not impact clinical decision making. These abnormalities included arachnoid cysts, mucus retention cysts, remote parenchymal injury causing encephalomalacia, mucosal thickening or opacification of sinuses, nonspecific T2 and fluid-attenuated inversion recovery signal abnormalities, chronic ventriculomegaly, and insignificant tonsillar ectopia.

A recent study was performed by Rudloe et al10 to determine the yield of neuroimaging in a cohort of 364 children who presented to the ED with acute ataxia (defined as <7 days' duration). The results from that study were remarkably similar to ours; that is, 13% of children with radiological findings required immediate intervention. The most common diagnoses classified as urgent in their cohort were intracranial tumors and acute demyelinating syndrome. Clinical features most suggestive of abnormalities on imaging were (1) age of the child, with older children noted to have more significant intracranial pathology; (2) duration of symptoms, with those that were present for more than 3 days being of significance; and (3) presence of additional neurological signs on examination, such as cranial nerve deficits or hemiparesis. The authors note that a predictive model using these 3 factors would have a 99% positive predictive value. The findings of our current study are in partial alignment with the study of Rudloe et al,10 in that the majority of patients with presence of clinically significant abnormalities on neuroimaging had focal neurological findings on examination. However, our small sample size precludes us from providing any meaningful recommendations regarding neuroimaging in children presenting to the ED with acute ataxia. Large-scale studies are needed to validate the findings of this study and to identify a subset of children with acute ataxia in whom imaging may not be necessary.

Cranial nerve deficits, in particular, appear to be associated with significant pathology on imaging including demyelinating diseases, strokes, and tumors. It is known that approximately 30% of children with a space-occupying lesion in the posterior fossa will have papilledema on funduscopic examination, and more than 80% will have other symptoms or signs of raised intracranial pressure, besides ataxia.11 Therefore, it is not individual symptoms, such as ataxia alone, but a combination of signs and symptoms that must alert a physician to a diagnosis of potentially serious intracranial pathology. Children with posterior fossa and brain stem tumors often have headache, emesis, cranial nerve involvement, and/or hemiparesis with or without a positive Babinski sign. A physician should actively seek to identify or exclude these findings in the setting of ataxia.

A significant limitation of several retrospective studies, including the current one, is the lack of documentation regarding funduscopic examination. The need for quick decision making in the ED setting and difficulty in performing a nondilated funduscopic examination in a younger child often lead to situations wherein cranial nerve examination is documented as “normal” but specifics regarding optic disc edema are not clearly noted. In this cohort, the 2-year-old patient who was eventually diagnosed with ependymoma did not have a funduscopic examination performed, but was found to have obstructive hydrocephalus on CT and was likely to have had papilledema on the funduscopic examination. The other child, who had significant findings on neuroimaging but a nonfocal examination, was diagnosed with acute demyelinating syndrome. On review of his documentation, he was noted to have symptoms suggestive of CNS pathology including new-onset headache, nausea, and projectile emesis that prompted the physician to perform neuroimaging. Conclusively, children presenting with ataxia and focal neurological signs should alert the physician to the necessity of neuroimaging. Furthermore, children with ataxia and no focal neurological signs, but other symptoms suggestive of increased intracranial pressure or optic disc edema, should also adhere to a neuroimaging evaluation.

Back to Top | Article Outline


This study was a retrospective chart review, which limited the amount of information that was available, due to lacunae in documentation. However, both the emergency physician dictation note and the neurology consultation documents were reviewed, if available, to minimize missing patient information. Diagnostic pathology could have been missed in those children (23/141 [16.3%]) who were evaluated and discharged from the ED, but we reviewed hospital records for those patients and found that the majority of patients (16/23) had a subsequent ED or clinic visit, but none involving ataxia. Of the 7 children who were discharged home and had no subsequent visits within the hospital system, 3 patients were diagnosed with ataxia due to ingestion and 1 patient each with migraine, transient ataxia, viral cerebellitis, and ataxia not otherwise specified. Computed tomography of the brain was performed on all of these patients, except for the known ingestions and the one patient with transient ataxia, and all of the CT scans were found to be negative for abnormal findings.

This study was conducted at a large academic center, with the ability to consult pediatric neurology at all hours of the day. A more thorough neurological examination, along with the ability to identify subtle abnormalities not identified by emergency medicine physicians, as well as specialist input and recommendations, could have influenced the decision to obtain neuroimaging. Hence, our study results cannot be generalized to settings where around-the-clock pediatric neurology support is not available.

The study patients were identified using International Classification of Diseases, Ninth Revision diagnostic codes for ataxia, cerebellar symptoms, and gait abnormality. It is possible that potential study patients could have been missed using this identification system.

Finally, this single-institution study is limited by a small sample size. A multicenter study with a larger sample size is necessary to formulate evidence-based decision (prediction) rules for obtaining neuroimaging in children presenting to the pediatric ED with acute ataxia.

Back to Top | Article Outline


Clinically significant neuroimaging was noted in a minority of children who presented to the pediatric ED with acute/subacute ataxia. The majority of patients with clinically significant neuroimaging had focal neurological findings on examination. A multicenter study with a large sample size would help formulate evidence-based prediction rules for identification of the subset of children with acute and subacute ataxia for whom neuroimaging may be unnecessary.

Back to Top | Article Outline


1. Musselman KE, Stoyanov CT, Marasigan R, et al. Prevalence of ataxia in children: a systematic review. Neurology. 2014;82:80–89.
2. Ryan MM, Engle EC. Acute ataxia in childhood. J Child Neurol. 2003;18:309–316.
3. Prasad M, Ong M, Setty G, et al. Fifteen-minute consultation: the child with acute ataxia. Arch Dis Child Educ Pract Ed. 2013;98:217–223.
4. Whelan HT, Verma S, Guo Y, et al. Evaluation of the child with acute ataxia: a systematic review. Pediatr Neurol. 2013;49:15–24.
5. Poretti A, Benson J, Huisman T, et al. Acute ataxia in children: approach to clinical presentation and role of additional investigations. Neuropediatrics. 2013;44:127–141.
6. Raucci U, Vanacore N, Paolino MC, et al. Vertigo/dizziness in pediatric emergency department: five years' experience. Cephalgia. 2015;36:593–598.
7. Sheridan DC, Meckler GD, Spiro DM, et al. Diagnostic testing and treatment of pediatric headache in the emergency department. J Pediatr. 2013;163:1634–1637.
8. Maytel J, Krauss JM, Novak G, et al. The role of brain computed tomography in evaluating children with new onset of seizures in the emergency department. Epilepsia. 2000;41:950–954.
9. Zonfrillo MR, Kim KH, Arbogast KB. Emergency department visits and head computed tomography utilization for concussion patients from 2006 to 2011. Acad Emerg Med. 2015;22:872–878.
10. Rudloe T, Prabhu SP, Gorman MP, et al. The yield of neuroimaging in children presenting to the emergency department with acute ataxia in the post–varicella vaccine era. J Child Neurol. 2015;30:1333–1339.
11. Wilne S, Collier J, Kennedy C, et al. Presentation of childhood CNS tumours: a systematic review and meta-analysis. Lancet Oncol. 2007;8:685–695.
Back to Top | Article Outline

From the Editors

Recently, the Editorial Board has conducted discussions about changing the focus of the journal to reflect the growing number of Original Research submissions that we receive each year.

We would like to accept and publish more of these articles and eliminate the large backlog of accepted articles that currently exists.

As part of this content shift, we will no longer be considering Illustrative Cases for publication, beginning July 1, 2019.

Over the journal’s 34+ years of publication, Illustrative Cases have been a reliable source of specific, case-based information, and we are grateful to all of the authors who sent them to us.


ataxia; intracranial pathology; neuroimaging

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.