lable at ScienceDirect

Pediatric Neurology 65 (2016) 14e30

Contents lists avai

Pediatric Neurology

journal homepage: www.elsevier .com/locate/pnu

Topical Review

Acute Ataxia in Children: A Review of the Differential Diagnosisand Evaluation in the Emergency Department

Mauro Caffarelli MD a,1, Amir A. Kimia MDb, Alcy R. Torres MDa,*

aDivision of Neurology, Boston Medical Center, Boston, MassachusettsbDivision of Emergency Medicine, Department of Medicine, Boston Children’s Hospital, Boston, Massachusetts

Article HistReceived O* Commu

NeurologyBoston, MA

E-mail a1Present aFrancisco,address: m

0887-8994/$http://dx.do

abstract

Acute ataxia in a pediatric patient poses a diagnostic dilemma for any physician. While the most common etiol-ogies are benign, occasional individuals require urgent intervention. Children with stroke, toxic ingestion,infection, and neuro-inflammatory disorders frequently exhibit ataxia as an essentialdif not the onlydpresentingfeature. The available retrospective research utilize inconsistent definitions of acute ataxia, precluding the abilityto pool data from these studies. No prospective data exist that report on patients presenting to the emergencydepartment with ataxia. This review examines the reported causes of ataxia and attempts to group them intodistinct categories: post-infectious and inflammatory central and peripheral phenomena, toxic ingestion, neuro-vascular, infectious and miscellaneous. From there, we synthesize the existing literature to understand whichaspects of the history, physical exam, and ancillary testing might aid in narrowing the differential diagnosis. MRI issuperior to CT in detecting inflammatory or vascular insults in the posterior fossa, though CT may be necessary inemergent situations. Lumbar puncture may be deferred until after admission in most instances, with suspicion formeningitis being the major exception. There is insufficient evidence to guide laboratory evaluation of serum,testing should be ordered based on clinical judgementdrecommended studies include metabolic profiles andscreening labs for metabolic disorders (lactate and ammonia). All patients should be reflexively screened for toxicingestions.

Keywords: acute ataxia, cerebellar ataxia, toxic ingestion, Guillain-Barré syndrome, differential diagnosis, children, magneticresonance, lumbar puncture

Pediatr Neurol 2016; 65: 14-30� 2016 Elsevier Inc. All rights reserved.

Introduction

Ataxia is a neurological disorder inwhich coordination ofmotor activity is impaired, preventing the execution of fluidmovements. Childrenwith ataxia classically present with aninability to ambulate as manifested by an ataxic gait, often

ory:ctober 31, 2015; Accepted in final form August 26, 2016nications should be addressed to: Dr. Torres; Division of; Boston Medical Center; One Boston Medical Center Place;02118.

ddress: artorres@bu.eduddress: Department of Pediatrics, University of California San550 16th Street, 4th Floor, San Francisco, CA 94143. Electronicauro.caffarelli@ucsf.edu.

e see front matter � 2016 Elsevier Inc. All rights reserved.i.org/10.1016/j.pediatrneurol.2016.08.025

described as a “wide-based gait” with “truncal instability.”Young children, especially, may manifest ataxia as a simpleinability or refusal to ambulate.1-4 The key to assessment ofacute ataxia in children is a thorough physical examination,which may reveal many possible associated findings andshed light on the location of the primary pathology. In thisreview, we present a broad differential diagnosis rangingfrom primary central causes to those secondary to infectionor toxic exposure. We hope this will serve as a guide forphysicians in the emergency department and consultingpediatric neurologists who are evaluating children pre-senting with acute-onset ataxia.

The causes of ataxia can be organized in an anatomicfashion. Although an “ataxic gait” on examination is typi-cally associated with cerebellar deficit, an abnormality in

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 15

gait can present as a consequence of deranged function inmany locations along the central nervous system (CNS).Motor weakness can originate from a lesion in the cerebralcortex, brainstem, spinal cord, peripheral nerves, orneuromuscular junction.1-4 Impaired balance often resultsfrom dysfunction of the vestibular apparatus in the innerear or the eighth cranial nerve.5

A thoughtful history and physical examination techniquewill aid in localizing the source of dysfunction. Childrenwhopresent with developmental delay or regression in mile-stones may suffer from a metabolic derangement such as anaminoaciduria (e.g., Hartnup andmaple syrupurine disease),pyruvate dehydrogenase deficiency, or biotinidase defi-ciency.6 A prior ataxic event in an otherwise healthy childmay suggest a paroxysmal disorder as is seen in genetic ionchannelmutations (i.e., the “episodic” ataxias) or in complexmigraines, inwhich the frequencyanddurationof attacks canvarywidely betweenpatients. Similarly, a previous historyofextremity weakness or vision loss may indicate a polyfocaldemyelinating disease such as multiple sclerosis (MS). Pri-mary “infectious” or secondary “postinfectious” cerebellitismay be implied, respectively, with an either concurrent orrecent viral syndrome, prompting the physician to elicit sucha history and examine the patient for signs of rash.

Most disorders discussed herein demonstrate depressedmentation or speech abnormalities such as fluctuations inclarity, tone, rhythm, and volume. The cranial nerve ex-amination may reveal a deficit, nystagmus, or palatalmyoclonus. These patients may have hypotonia and theirstrength may appear diminished. Finger-to-nose may showunder- or overshooting of the limbs. Patients may havedifficulty with rapid alternating movements and intentiontremor. Patients with peripheral neuropathies are usuallyhyporeflexive, conversely those with cerebellar insult mayhave pendular reflexes. Finally, gait testing may demon-strate poor maintenance of truncal position and titubation,or no ability to walk at all.

The temporal course of ataxia is diagnostically useful ifacute toxic, traumatic, infectious, or postinfectious causesare suspected. Chronic or episodic ataxias, on the otherhand, are generally the result of an inborn error of meta-bolism or hereditary channelopathies. Although an acuteonset of ataxia has historically been associated with abenign course,1 it can present a unique challenge to physi-cians as a few possible causes can pose a real concern forirreversible neurological damage.

Only a handful of studies have documented the di-agnoses one can encounter in children presenting withacute ataxia. Gieron-Korthals et al.7 in 1994 summarized 40children who had been symptomatic for less than 48 hours.They demonstrated that most instances of acute ataxia aredue to a postinfectious cerebellar process, toxic ingestion, orGuillaineBarré syndrome (GBS). Martínez-González et al.8

in 2006 similarly evaluated 39 children with symptomsfor less than 48 hours and found a similar array of etiologies,providing support to the findings of Gieron-Korthals et al.Rudloe et al.9 in 2014 demonstrated that clinically signifi-cant intracranial pathologydtumors, infarcts, and acutedisseminated encephalomyelitis (ADEM)dis frequentlydocumented in children with symptoms up to seven daysbut is far less likely in young children with symptoms forless than 72 hours. Finally, Thakkar et al.10 in 2016 described

120 children referred to a pediatric neurologist withsymptoms up to four weeks in which the leading diagnosisremained postinfectious in nature and the frequency oftoxic ingestion was much lower than preceding studies hadreported. The disparate conclusions that arise from thesestudies are rooted in their different designdall the afore-mentioned studies are retrospectivedand each uses adifferent definition for the duration of symptoms as well asdifferent inclusion and exclusion criteria.

Here we expand the differential diagnosis of acute ataxiainto broad categories, as one might do with a patientpresenting to care for the first time in the emergencydepartment. We then discuss the clinical, laboratory, andimaging features reported with each entity, so that a clini-cian may be able to begin prioritizing their evaluation basedon historical and examination features they may encounter.We give special attention to the evidence that supports theevaluation of acute ataxia in the emergency department tooptimize the initial diagnosis, keeping in mind that ongoingevaluation and follow-up of these patients may require adifferent set of clinical tools and practices. We then sum-marize the different aspects of evaluation (history, physicalexamination, laboratory tests, imaging, and lumbar punc-ture [LP]) and give our recommendations. Finally, we hopethis article will demonstrate the need for carefully designedprospective research to understand which circumstancesmight require close attention and swift action comparedwith those in which patients may be simply observed for aperiod of time.

Classification of etiologies

Central postinfectious and inflammatory causes

The most recognizable and generally most commoncauses of acute ataxia are postinfectious cerebellar processesthat are self-limited and resolve spontaneously.4,8 The term“acute cerebellar ataxia” has been firmly established in theliterature to imply a postinfectious process, but for the pur-pose of this review, we will use the term “acute post-infectious cerebellar ataxia” (APCA) as proposed by Porettiet al.11 for the sake of avoiding the less descriptive term.

In this section, we discuss APCA, cerebellitis, andADEM as the three major postinfectious syndromes thatfrequently present with ataxia. An important caveat inthe discussion of these postinfectious phenomena in-volves the introduction of the varicella vaccine in 1995and the understanding that many of the epidemiologicdata herein were collected in the period beforehand.Although neurological complications of varicella (i.e.,postcerebellar ataxia) occurs in nearly one of every 4000cases,12 estimates of the rate of cerebellar ataxia in thepostvaccine era approach 1.5 for every 1,000,000 vaccinedoses.13

Acute postinfectious cerebellar ataxiaThe observation of a sudden-onset ataxia in children after

an infectious illness was first documented for more than acentury ago.2 As other causes for cerebellar dysfunctionemerged, this remained a puzzling cause of ataxia for whichthe pathology remained unknown.1,3 Evidence for an auto-immune basis of APCA emerged with reports showing

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3016

oligoclonal bands in the cerebrospinal fluid (CSF), withseveral subsequent studies demonstrating immunoglobulinG and immunoglobulin M reactive to cerebellar Purkinjecells after varicella,14 EpsteineBarr virus (EBV),15,16 andmycoplasma17 infections. However, the mechanisms un-derlying both the pathologic specificity to and the physio-logic disruption of the cerebellum are unknown.

Our best understanding of the epidemiology andoutcome of APCA stems from work done by Connolly et al.,the only group to have prospectively studied the disorder,collecting data for more than a 23-year period. Of the 73children who satisfied entry criteria, 19 (26%) had a prioriinfectionwith varicella, two (3%) had EBV, and 36 (49%) hada nonspecific “viral” prodrome, whereas 14 (19.2%) had noidentifiable prodrome. The latency period from infection toataxia ranged from only a few days to three weeks with amean between nine and 11 days for all but EBV infections,which had a mean latency of 17.5 days. The mean age atpresentationwas 5.37 �4.00, although the data are skewedsuch that 60% of children are aged between two andfour years. Of 69 children who underwent LP, roughly halfhad leukocytes less than 5/mm3 and all bacterial and viralcultures were negative. Computed tomography (CT) wasunremarkable in all children imaged, whereas magneticresonance imaging (MRI) showed cerebellar T2 hyper-intensity in only one of nine children imaged. Of the in-dividuals who were followed, 70% returned to baselinewithin eight months of initial presentation.18

Since the study by Connelly et al., one other retrospectivestudy with essentially similar results has emerged.19 How-ever, without a good understanding of the pathophysiologyof the APCA, no tests are available to confirm the diagnosis.The work of Connelly et al. shows us that the mainstays inneurological diagnosis such as a thorough examination,imaging, and LP are not specific either.

Whelan et al.4 and others19-21 favor a combination of anormal mental status, a nonfocal neurological examination,and a clear history of recent infection as reasons to deferimaging or CSF studies for close neurological follow-up.However, clear criteria do not exist that reliably excludeconcerningdalthough perhaps less commonddiagnosesthat may require immediate intervention.

Postinfectious cerebellitisThe term “cerebellitis” appeared in the 1950s and 1960s

and was used interchangeably with cerebellar encephalitis,a cerebellar syndrome in the setting of a recent infectiousdisease such as infectious mononucleosis22 and varicella.23

These earlier descriptions of the term did not differentiate itfrom APCA, and thus it is unclear whether they weredescribing the same disease process.1,3 Horowitz et al. in1991 documented a patient with severe ataxia and en-cephalitis along with cerebellar edema on CT and MRI,24

invoking the subsequent use of cerebellitis to describe anataxia syndrome with the distinct feature of cerebellaredema on imaging.25,26

Sawaishi and Takada27 summarized patients with post-infectious ataxia who had cerebellar edema with mass ef-fect on imaging. Although these patients varied in severitywith a few mimicking APCA, many of them were lifethreatening or fatal, necessitating surgical decompres-sion.25,26 Sawaishi and Takada thus argued the diagnosis of

acute cerebellitis should be reserved for severe cases withpositive imaging findings, with the implication that itwould convey an urgent need for surgical evaluation.

Desai and Mitchell28 in 2012 argued that cerebellitis andAPCA should be thought of as two ends of a spectrum, withthe former diagnosis being applied largely to cases withcerebellar edema on imaging. However, Connelly et al.demonstrated that mild cerebellar edema may occur inchildren with ataxia who are otherwise well. In their study,Thakkar et al. report three (2.5%) of 120 patients diagnosedwith acute cerebellitis, which they defined as the presenceof imaging abnormalities in the setting of acute ataxiadoneof the patients developed fulminant brain swellingrequiring decompressive occipital craniectomy. Thiscompared with 71 (59.2%) patients with symptoms andnormal imaging findings diagnosed with APCA.10

Our view of cerebellitis aligns with Sawaishi and Takada,that it should denominate ataxia with the clinical distinc-tion of an abnormal mental status to suggest increasedintracranial pressure and thus the need for immediate im-aging. Without clear prospective data on the epidemiologic,clinical, laboratory, or imaging features of cerebellitis, it willbe crucial to have a low clinical threshold for imaging andsurgical evaluation should a child begin to show signs ofaltered mental status in the setting of a postinfectiouscerebellar ataxia. Although MRI may be more sensitive andspecific imaging compared with CT,29,30 CT may be neces-sary in more clinically urgent scenarios.

ADEM, the clinically isolated syndrome, and MSADEM, clinically isolated syndrome (CIS), and MS

describe three disease processes that have tremendousoverlap yet differ in their definition as delineated by specificclinical and imaging features.31 All three are characterizedby polyfocal immune-mediated demyelinating lesions ofthe CNS. The diagnosis of MS requires either evidence of aprior demyelinating event or new lesions on imagingfollow-up, whereas CIS and ADEM are differentiated by thepresence of encephalopathy (Table).31,33

When evaluating a patient with acute ataxia, a history ofprior episodes of neurological deficit is important formaking the diagnosis of MS. However, a patient presentingwith a first time demyelinating CNS event woulddby sakeof the definitiondbe diagnosed with ADEM or CIS unlessMRI were to reveal evidence of an older (i.e., nonenhancing)demyelinating lesion.31 However, the diagnosis of ADEM orCIS does not preclude a later diagnosis of MS. One study of40 children initially diagnosed with ADEM or CIS demon-strated that one of 15 children diagnosed with ADEM waslater reclassified as having MS, whereas 14 of 28 patientswith CIS had a relapse event, causing 13 of them to bediagnosed with MS.34 The presence of oligoclonal bands inthe CSF may indicate a higher likelihood of relapse andsubsequent reclassification to MS.35

Although we do not wish to diminish the importance ofdifferentiating between these entities, further discussion ofa patient with demyelinating disease would likely occurafter imaging reveals the presence of lesions. Because CIS isa relatively recent disease description that was first rigor-ously defined in 2007,32 prospective studies have not yetbeen published that identify the rate of ataxia as a pre-senting feature of the disease.

TABLE.Definitions of MS, ADEM, and CIS Based on Clinical Findings and Imaging Findings as Described in Krupp et al. 201332

Diagnosis Clinical Findings Imaging Findings

MS (Any of the following)� Two or more nonencephalopathic clinical CNS eventsseparated by more than 30 days

� One nonencephalopathic episode with MRI findings thatshow DIS and follow-up MRI that shows DIT

� One ADEM attack followed by a nonencephalopathicclinical event separated by �3 months with new MRIlesions

� First, single, acute event that does not meet ADEM criteriaand whose MRI findings demonstrate DIS and DIT

� Demonstration of DIT and DIS

ADEM � First time, polyfocal, clinical CNS event� Encephalopathy that is not explained by fever

� Diffuse, >1-2 cm lesions of white matter� Absence of DIT

CIS � Monofocal or polyfocal clinical CNS event� Absence of prior clinical evaluation of CNS demyelinatingdisease

� No encephalopathy

� Absence of DIT or DIS

Abbreviations:ADEM ¼ acute disseminated encephalomyelitisCIS ¼ clinically isolated syndromeCNS ¼ central nervous systemDIS ¼ dissemination in spaceDIT ¼ dissemination in timeMRI ¼ magnetic resonance imagingMS ¼ multiple sclerosisDefinition of DIT and DIS from Polman et al., 2010.33 DIT: a new T2 or contrast-enhancing lesion on follow-up MRI with reference to baseline or the simultaneous presence ofenhancing and nonenhancing lesions at any time. DIS: at least one T2 lesion in two of the following four areas of the CNSdperiventricular, juxtacortical, infratentorial, andspinal cord.

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 17

ADEM is an acute inflammatory demyelinating disorderthat is generally preceded by an infection and affects thewhite matter tracts of the brain, brainstem, spinal cord, andoptic nerves.32 The incidence is approximately 0.4/100,000among children and a mean age 6.5 years, with 64% of pa-tients aged between 2 and 10 years.36 The diagnosis re-quires polyfocal clinical CNS findings, encephalopathy notexplained by fever, and demonstration of characteristicmultifocal demyelinating lesions on MRI.31,37

Ataxia is a common presenting feature of ADEM. In aprospective study spanning 12 years, 42 (50%) of 84consecutive patients with ADEM presented with ataxia.Although it is unclear for howmany patients ataxia was theonly presenting feature, the study identified long tract signsin 71 (85%), hemiparesis in 64 (76%), and encephalopathy in58 (69%) patients, of whom 16 progressed into a coma.38

Encephalopathy can be defined broadly as any change inmental status, as patients can present with irritability,sleepiness, confusion, or obtundation.39 It is important tokeep inmind that because this studywas published in 2002,the operational difference between ADEM and CIS had notyet been formalized to differentiate between patients withand without encephalopathy.31,32

Of 82 patients prospectively included in the study byTenembaum et al., MRI showed only small (less than 5 mm)lesions in 52 (62%), large confluent lesions were observed in20 (24%), bithalamic involvement was seen in 10 (12%), andacute hemorrhagic lesions were seen in two patients. CSFstudies were abnormal in 24 children (28%), showing eitherlymphocytic pleocytosis (more than 180 cells/mm3) orincreased protein levels (more than 1 g/dL).38

Classic ADEM should be easy to clinically differentiatefrom APCA, as it will typically involve focal neurologicaldeficits along with a possible change in mental status

concurrent with ataxia, both of which are generally in-dications for rapid imaging. However, the relative incidenceof ADEM versus APCA is unknown, and no studies exist thatallow us to clinically differentiate ADEM from other causesof acute ataxia. AlthoughMartínez-Gonzáles et al. identifiedADEM as the cause of ataxia in one of 39 patients studied,8

the relative incidence of ADEM compared with other causesof ataxia is otherwise unknown. For the wary clinician, aperiod of observation may be useful in determining an“illness curve” of how symptoms are progressing.

Nearly all patients will eventually demonstrate completeneurological recovery,40 although patients should be closelymonitored for relapse.34,41 Although randomized controlledtreatment trials for ADEM have not been performed, rapidinitiation of high-dose intravenous methylprednisolone is amainstay in treatment.37 However, some cases may bemoresevere, and stabilization of the patient may be necessarybefore diagnostic evaluation can be initiated: a fraction ofcases will require intensive care unit admission as patientswith ADEM have been shown to progress to fulminant braininjury and even death.42 Furthermore, approximately 17% ofpatients were prospectively shown to present with lesionsin the brainstem that interfered with respiratory drive,necessitating prompt respiratory support.38 The adminis-tration of intravenous immunoglobulin in severe cases isnot unwarranted, although its efficacy in treatment has notbeen prospectively studied.43

Spinal and peripheral postinfectious and inflammatory causes

Sensory and motor deprivation of the lower extremitiesbecause of inflammation of the spine or peripheral nervescan often present as ataxia with no change in mental status.One of the more common classical causes of acute ataxia is

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3018

GBS, an idiopathic, postinfectious inflammatory disorder ofthe peripheral nerves.7 The spinal and peripheral post-infectious phenomena are a series of immune-mediateddisorders including GBS, Miller Fisher syndrome (MFS),and Bickerstaff brainstem encephalitis (BBE). Althoughthese syndromes are often thought to be distinct, theyfrequently overlap in their features with ataxia being amajor component of their clinical presentation. Thenonspecific finding of anti-GQ1b and anti-GM antibodies inthe serum of affected patients suggests GBS, MFS, and BBEare variants of a single disease process.44,45 The classicalassociation with antecedent campylobacter infection is alsobolstered by the association between the organism andanti-GM1 antibodies.43 Transverse myelitis, with its asso-ciated disorders, is also discussed in this section as a causeof acute ataxia.

GuillaineBarré syndromeGBS is a postinfectious polyneuropathy that can affect

individuals from 12 months into adulthood, with a peakincidence at five to six years.46 Patients generally presentwith symmetric lower extremity weakness and areflex-iadthe two major defining criteria for GBS diagnosis.47 Asmall percentage of patients (less than 5%) may presentwith normal or exaggerated reflexes.48 Autonomic insta-bility may also occur, resulting in heart-rate variability andcardiovascular collapse.49

Several minor criteria aid in differentiating GBS fromother causes of ataxia. For example, a prospective studyshowed that 85% of adult patients with GBS complained ofpain on admission.50 Another prospective study by Paradisoet al.51 showed 44 of 61 (72%) children with GBS com-plained of pain, although it is unclear whether painwas partof the initial presentation or if it appeared later on in thecourse of illness. Another retrospective study found thatchildren greater than five years of age with GBS were morelikely to initially complain of limb pain (53%) than theiryounger counterparts (24%).52

A characterizing feature of GBS is increased protein levelsin the CSF with albuminocytologic dissociation, meaningthat the cell counts are typically normal despite anincreased protein level.53 Although we intuitively realizethis is an important test for diagnosing GBS, CSF proteinlevels are commonly normal in the first week of presenta-tion. Early work by Ropper54 demonstrated that only 50% ofpatients initially presented with CSF protein greater than0.55 g/L. A more recent prospective study conducted byNishimoto et al.44 showed only 44% of patients with GBShave abnormal protein levels in the first week of illness.

Published European guidelines for the diagnosis of GBSargue that CSF studies are of very limited value, and thediagnosis should usually be made on clinical groundsalone.55 Our view is that, although CSF studies may beuseful in characterizing the course of disease and subse-quent response to treatment, they may be deferred untilafter the patient is admitted from the emergency depart-ment as they will rarely direct immediate management.

Imaging is not a part of diagnostic criteria in GBS and canbe deferred in cases with little clinical ambiguity.56 How-ever, as there can be overlap with MFS and BBE, which canrespectively involve the cranial nerves and brainstem, MRI

may be a diagnostic necessity. Zuccoli et al.57 gatheredretrospective neuroimaging data on 17 patients who pre-sented with clinical GBS, of whom 14 (82%) showedenhancement of the cauda equina.

Most patients who are adequately treated with immu-nomodulatory therapy will generally recover their neuro-logical function within the first year after the initialpresentation.58 However, as with most postinfectious dis-ease, GBS presents within a spectrum of varying severitywhere approximately 13% of children will progress to res-piratory failure requiring ventilator support.51

Miller Fisher syndromeAlthough the most common form of GBS involves

ascending paralysis in the extremities, several variant formshave been described, which often involve prominent cranialnerve involvement. MFS is one such variant, which alsodiffers from typical GBS, in that it tends to have amore rapidonset and greater severity.59 Although MFS is stronglyassociated with the clinical triad of ataxia, areflexia, andophthalmoplegia, it is important to note that any ascendingpolyneuropathy can present with dysfunction of any of thecranial nerves. One study prospectively found that of allchildren presenting with GBS-like symptoms such as lowerextremity weakness and areflexia, 27% also presented withcranial nerve dysfunction and another 46% developed somekind of cranial nerve dysfunction during the course ofhospitalization.46

Ancillary testing plays a similar role with MFS as it doeswith GBS. CSF shows increased protein levels in 25% ofpatients during the first week of presentation, again sug-gesting the limited role of LP in the initial evaluation ofMFS.44 As discussed previously, the presence of anti-GQ1bantibodies is highly suggestive of MFS but has also beenshown in polyneuropathies that do not have cranial nerveinvolvement.45

Bickerstaff brainstem encephalitisBBE is a relatively obscure medical condition described

as a clinical triad of ataxia, ophthalmoplegia, and a“disturbance of consciousness.”60 It is generally character-ized as occurring with an antecedent history of infection,albuminocytologic dissociation in the CSF, and anti-GQ1bantibodies in the serum.45 One group delineated criteriaas being “definite” when opthalmoplegia, ataxia, andimpaired consciousness are present in the setting of posi-tive serum anti-GQ1 antibodies and the absence of otherexcludable conditions.61

A nationwide survey of Japanese physicians was con-ducted to collect data on all patients with likely BBE be-tween 2006 and 2009. Thirty-seven patients were found tofulfill criteria for either definite (19) or probable (18) BBE. Ofthese, 54% presented with limb weakness and 67% withabsent or decreased reflexes. Ninety-two percent of patientspresented with ataxia, 100% with ophthalmoplegia, 41%with facial palsy, and 72% with oropharyngeal palsy. Labo-ratory investigation showed normal CSF cell counts (lessthan 5/mm3) in 56% and normal protein (less than 45 mg/dL) in 62% of patients. Brain MRI was abnormal, showing T2hyperintensity of the brainstem, thalamus, and cerebellumin eight (23%) of 35 patients. Patients were treated with

FIGURE 1.Relative rates of benzodiazepine, carbamazepine, and phenytoin ingestionin children aged less than six years. The comparison between 2009 to201369-71,76,77 and 1988 to199278-82 is made to highlight the relativereduction of carbamazepine and phenytoin ingestion juxtaposed with therelative increase in benzodiazepine ingestions in the time since Gieron-Korthals et al. published their study on the causes of ataxia. Patientsgreater than six years of age could not be compared, because the ageranges used to categorize the data had changed between the two periods oftime.

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 19

immunologic therapydintravenous immunoglobulin (92%),steroid (49%), or acyclovir (24%)dwith 91% of patients ableto walk independently at one-year follow-up.61

There is speculation that BBE is under-reported becauseit is not widely recognized as distinct from GBS or MFS.62

Although data on the disorder are scarce, one series docu-mented several patients with clinical brainstem lesions whohad simultaneous motor axonal neuropathy on electro-diagnostic studies suggesting a clinical overlap with GBS.60

The shared finding of serum anti-GQ1 antibodies providesfurther evidence of shared etiologic factors.45

BBE presents an interesting diagnostic problem, becauseimaging and LP appear to have little utility as with all otherpostinfectious phenomena discussed so far. The challengepresents itself with the presence of lesions that clearlylocalize to the brainstem, which would be a certain indi-cation for imaging. Our view is that BBE may serve as adiagnosis of exclusion in the face of cranial nerve abnor-malities with negative imaging findings.

Transverse myelitisAcute transverse myelitis (ATM) is a rare focal inflam-

matory disorder of the spine that affects one to four in everymillion persons per year, with a bimodal age distributionbetween the ages ten and 19 years and 30 and 39 years.63

The Transverse Myelitis Consortium Working Group out-lined a diagnostic algorithm in 2002, which included muscleweakness with upper motor neuron signs and sphincterdysfunction with a clear sensory level. With this clinicalsuspicion, the recommendation is to obtain a contrast-enhanced MRI within four hours to rule out compressivemyelopathy, whichmay necessitate a surgical evaluation or acourse of intravenous methylprednisolone.63

Idiopathic ATMdthe isolated and monophasic form ofthe diseasedwas retrospectively studied in children of aperiod spanning 23 years. Of 27 patients studied, ten (39%)had an antecedent infection and 20 (74%) presented withparaparesis. The mean age of onset was 9.5 � 5.7 years.Although ataxia is not explicitly stated as a clinical feature, asudden onset of paraparesis would present with an inabilityto walk, which in younger individuals might be difficult todistinguish from cerebellar ataxia. CSF pleocytosis was seenin ten (37%) and increased protein levels in 12 (44%) pa-tients. Imaging revealed T2 hyperintensity most often in thegray matter of the cervical and thoracic spine. Althoughsurrounding white matter was occasionally involved, iso-lated white matter lesions were not seen. Of 21 patientswho were imaged with MRI, seven (33%) had multifocallesions whereas six (28%) showed no abnormalities,although some of these data may have been collected witholder, less sensitive, imaging techniques.64

Although ATM can occur as a postinfectious process, it isalso known to be a rare feature in several connective tissuedisorders (sarcoidosis, Behçet’s disease, Sjögren’s syn-drome, systemic lupus erythematosus, and so forth) and isalso a common first presentation of MS.63 The presence ofoptic neuritis will suggest either neuromyelitis optica or MSas the underlying etiology. All cases of ATM eventuallyrequire CSF studies to examine for specific markers of dis-ease, although these tests do not need to be immediatelydone.65 A discussion on diagnosing the different causes ofATM is out of the scope of this review, as it would likely be

done after imaging is performed and the patient is admittedfrom the emergency department.

Toxic ingestion

Toxic ingestion is a common problem in the pediatricpopulation. In 2013, 1.3 million children aged less than19 years in the United States were reported to poison con-trol centers for accidental and intentional ingestion ofprescription and nonprescription medicationsd78% of vic-tims less than six years of age.66 Medication ingestion inchildren accounted for an estimated 9490 emergency hos-pitalizations between 2007 and 2011, 75% of which werechildren either one or two years of age. Benzodiazepineingestion accounts for approximately 10% of these patients,representing one of the most frequent classes pharmaceu-tical agents to be ingested.67 Although major efforts arecurrently being made to introduce safer packaging forcommonly ingested pharmaceuticals,68 toxic ingestion re-mains an important cause of ataxia.

Gieron-Korthals et al. showed that approximately onethird of acute ataxia in children resulted from toxic inges-tion. Although benzodiazepines were the most commonoffender, phenytoin, phenobarbital, and carbamazepinewere also important substances identified by the study.7

Martínez-Gonzáles et al.8 reported dextromethorphan andethanol to as two other toxic causes of acute ataxia. Newstudies are needed to evaluate which substances arecurrently implicated in acute ataxia, as the incidence ofpoisonings may have changed for more than the past20 years.

BenzodiazepinesBenzodiazepines are one of the most commonly ingested

substances by children and have been repeatedly published

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3020

as a frequent cause of acute ataxia.4,7,8,11 An average of10,000 ingestions per year occur in children.69-71 Clinicalfeatures include lethargy advancing to coma, with respira-tory depression as an important feature to be aware of.72

Although the sedative effects of many substances cancause pupillary constriction,73 benzodiazepines have beenshown to not affect the pupil size in the same way opiatesmight.74 Amnesia and diplopia are two other features thathave been described in the dental literature.75

The landmark 1994 study by Gieron-Korthals et al.demonstrated that only five (12.5%) of their 40 patientswith acute ataxia had ingested benzodiazepines, yet thatgroup represented the most cases of ataxia attributed to asingle drug class.7 Data collected by the American Associ-ation of Poison Control Centers show that, in children lessthan six years of age, benzodiazepine ingestion rates arehigher in recent years compared with the time period whenGieron-Korthals et al. collected their data (Fig 1).69,70,76,78-82

Despite the strong association between benzodiazepinesand acute ataxia in children, relatively few studies haveexamined this relationship. Wiley and Wiley in 1998 pub-lished a two-center retrospective analysis of childrenadmitted for benzodiazepine ingestion, reporting ataxia in40 (87%) of 45 patients of which eight (20%) presented withno other symptom. Altered mental status defined by aGlasgow coma scale (GCS) less than 15 was apparent in 32(70%) patients, of whom six (19%) had a GCS less than 6.Twenty-seven patients (60%) had a clearly identified time ofingestion.72

It is difficult to understand how these datadpublishednearly 20 years agodfit into the evaluation of acute ataxiatoday. However, it is clear that accidental ingestion ofbenzodiazepines in young children is still occurring at a rateunchanged, if not greater, than 20 years ago. The relativefrequency of ataxia caused by benzodiazepine ingestioncompared with other noningestion causes remains animportant question to tease out in future studies. Further-more, understanding what proportion of cases present withan a priori history of ingestionwill be helpful in determiningwhether to maintain a high index of suspicion in caseswhere such a history is absent. Toxicologic screening ofurine or serum is a rapid test that may be useful for childrenin clinically vague scenarios.

Antiepileptics: carbamazepine and phenytoinGieron-Korthals et al. reported that, apart from benzo-

diazepines, carbamazepine and phenytoin were the twoantiepileptic drugs (AEDs) most commonly implicated inacute ataxia. Three of their 40 patients had ingestedphenytoin and two had ingested carbamazepine.7 A retro-spective study by Lifshitz et al. reported ataxia in slightlyless than one third of children with carbamazepinepoisoning, with clinical improvement seen in all patients.Nystagmus is the most common presenting feature, andalthough ataxia may be seen as a more concerning symp-tom, the combination of the two may suggest carbamaze-pine poisoning.83

Carbamazepine ingestions have also decreased in thepast 20 years, most notably in children aged less thansix years (Fig 1).69-71,76-82 This finding coincides with agenerally reduced prescription rate of carbamazepine, andthe increasing use of newer AEDs.84,85 Oxcarbazepine, a

newer generation AED with a similar mechanism of action,has been shown to have more benign toxic profile in over-dosing, which does not seem to include ataxia, althoughdizziness and vertigo is a feature in about 1% of cases.86

Phenytoin toxicity is also classically associated withataxia, often in conjunction with vertical downbeatnystagmus and vomiting.87 Gieron-Korthals et al. describedthree patients with acute ataxia who were ultimately foundto have overingested phenytoin.7 However, averagephenytoin ingestions for more than the past five years areten-fold less than in the five years preceding their publi-cation,69-71,76-82 which may also be in part because of areduction in phenytoin prescriptions.85

Cough syrup: dextromethorphanDextromethorphan is the main ingredient in widely

available over-the-counter cough suppressants with CNSadverse effects that makes it a popular drug of abuse.88

Different effect “plateaus” have been described in whichincreasing concentrations of dextromethorphanwill lead todistinct symptoms: euphoria (2.5 mg/kg), imbalance andvisual sensations (7.5 mg/kg), altered consciousness anddelayed reaction times (15 mg/kg), hallucinations, ataxia,and complete dissociation (greater than 15 mg/kg).89

One 30-month-old child with an approximate 38 mg/kgingestion of dextromethorphan presented to the emergencydepartment with opisthotonos that resolved with admin-istration of diphenhydramine, although the patientdemonstrated ongoing ataxia and nystagmus until the nextmorning.90 It is important to recognize that although someover-the-counter formularies contain dextromethorphan asits only active ingredient, other products may also containpseudoephedrine, which may induce irritability and hy-peractivity in a child superimposed on ataxia andnystagmus.91 Fortunately, the rate of cough syrup ingestionshas been rapidly declining for more than the past five years(Fig 2).69-71,76 Data are needed to determine the incidence ofcough syrup ingestion as a cause of acute ataxia.

EthanolEthanol toxicity can present with a range of mental sta-

tus changes and respiratory depression, with children oftenat a risk of hypothermia and hypoglycemia.92 Ethanol iswell-documented cause of cerebellar dysfunction, althoughthe mechanism is not fully understood.93 Research in ani-mal models has shown that adolescents have a higherthreshold to motor dysfunction compared with adults,94

although a similar dose-dependent relationship has notbeen shown in humans. Other than the single patientretrospectively identified by Martínez-González et al.8 froma group of 39, the rate of ataxia in children with ethanolingestion has not been studied; therefore the blood con-centrations of ethanol needed to generate ataxia areunknown.

There appears to have been a large reduction in child-hood ethanol ingestion for more than the last several years.The American Association of Poison Control Centers’ Na-tional Poison Data System documented 22,137 ethanol in-gestions in 2009, 72% of which were children less than sixyears of age who ingested nonbeverage ethanol. Thisnumber dropped to 7,093 after two years, and in 2013 was5,839 (Fig 2).69-71,76 The reason for this sharp decline is

FIGURE 2.Rates of reported ingestions of different substances known to cause ataxia in children. Data gathered from the Poison Control Centers’ National Poison DataSystem.69-71,76,77

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 21

unclear, although one possible explanation is a trend awayfrom the sale and home-use of alcohol-basedmouthwash.95

Whether this trend is the reason for fewer childhoodalcohol ingestions or has had an appreciable effect on therate of ataxia in children is unknown.

MarijuanaAtaxia can be a result of marijuana use, which usually

occurs in adolescents participating in risky behavior. How-ever, young children presenting with ataxia, tachycardia,and vertical and horizontal nystagmus have been docu-mented to have cannabinoids in their serum.96,97 Marijuanaand the many synthetic cannabinoids are known to have awide range of toxidromic features: patients initially believea sense of euphoria, but can progress to paranoia, delusions,

hallucinations, panic attacks, nausea, vomiting, seizures,and dizziness.98

The rate of reported ingestions of marijuana andcannabinoid-containing edibles has steadily increasedin the past five years, especially in children less than sixyears of age (Fig 2),69-71,76 suggesting that they may beimportant considerations as the cause of acute ataxia inthe face of the increasing liberalization of these sub-stances. A thorough history should include the presenceof these substances in the household, including inhaledand edible forms. Marijuana is commonly baked intocookies or brownies, which has been shown to be a causeof accidental ingestion in children leading to coma.97

Cannabinoids may not be routinely tested as part oftoxicology screens; thus it is important to actively

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3022

consider marijuana toxicity in the differential diagnosisof acute ataxia.

Cerebrovascular

Childhood stroke is an extremely rare but devastatingevent, with cerebellar involvement occurring in aneven smaller subset of this group of patients.99 Thereare documented examples of both hemorrhagic and non-hemorrhagic cerebellar stroke in children, with listedoutcomes ranging from complete recovery to permanentneurological dysfunction and occasionally death.100,101 Insuch individuals, intervention must be initiated early to ach-ieve optimal outcomes. Ataxia can be a major presentingfeature for these children, and although rare, early diagnosis iscritical. Although we do include case reports and case seriesin this section, there is a knowledge gap in understandinghow acute ataxia may be the initial presentation of stroke.Further study is needed to help in differentiating strokefrom other causes of ataxia, as it is one of the etiologies thatcan be missed when there is much to gain from rapidintervention.

HemorrhagicIntracranial bleeding is a medical emergency, necessi-

tating prompt imaging and neurosurgical evaluation. Aretrospective study documented four children with cere-bellar hemorrhage that presented with acute-onset ataxia.For these patients, other neurological signs and symptomswere highly variable and occasionally absent. Headachewasa common accompanying symptom, where the timingranged from sudden onset to progressive for more than thecourse of three days. The GCS for these patients was normalin all except one. Cerebellar hemorrhage in children wasfrequently found to reveal underlying cavernous angiomaor an arteriovenous malformation in the posterior fossa.102

Although these conditions may require eventual surgicalcorrection, the need for immediate surgical intervention isdictated by the extent of fourth ventricle effacement and aGCS less than 13.103 Although most neurosurgical outcomedata are from studies done in adults, childhood does notappear to significantly influence indications for surgery.104

Headache appears to be the most frequent accompanyingsymptom, and is an indication for rapid imaging in childrenwho describe their headache as being “unusual,” “sudden,”or “worst.” AGCS less than 13 is also an important indicationfor imaging, with rapid imaging preferable over sensitiveimaging. Rapid MRI protocols allow a time in the scanner(“table time”) that is comparable with that of rapid CTscanning, although its use in children has only been studiedin its ability to assess for ventriculoperitoneal shunt mal-function.105,106 Furthermore, its use in emergency settingshas been limited to retrospective studies.65 As a result, thesensitivity of rapid MRI protocols in detecting intracranialpathology in the posterior fossa cannot reasonably beinferred and remains to be elucidated. Evidence of hemor-rhage in the posterior fossa in children less than three yearsof age may alert the physician to nonaccidental trauma.107

NonhemorrhagicAcute ataxia can be the result of an ischemic cerebellar

insult resulting from an occlusion in the posterior

circulation. Ischemic stroke in the brainstem is uncommon.One review documented 36 published examples of verte-bral artery dissection (VAD) in children formore than a spanof ten years.108 Another review of posterior circulationstroke in children indicated that slightly less than half ofcases were attributable to VAD.100 In the Gieron-Korthalset al. report, only one of the 40 patients identified in aten-year interval had a cerebellar infarction, and this indi-vidual was not described in detail.7

Posterior circulation ischemic stroke can occur at anyage.109 Hasan et al. suggested that about half of childrenwith VAD present with ataxia and that half of the patientshave no clear history of trauma to the head or neck region.Impairment of consciousness occurs in a third of the pa-tients. After ataxia, headache and vomiting were the twomost common presenting symptoms, each occurring in 38%and 34% of cases, respectively. Neurological examinationshowed a deficit in the visual system in 72% of cases,manifesting as either nystagmus, visual field deficits, ptosis,pupillary abnormality, or abducens (sixth) nerve palsy.Furthermore, 54% of cases occur with gross motor deficitmanifesting as either hemiparesis, hemiplegia, orquadriplegia.108

Hasan et al. reported abnormalities in CT angiography forall the cases they identified. Only ten of 63 children hadsigns “pathognomonic” for VAD, showing either an intimalflap (six) or eccentric filling defect (four) in the vertebralartery. The remaining 53 cases had signs “highly suggestive”of VAD, including occlusion, pseudoaneurysm, irregularity,stenosis, narrowing, thrombosis, or a “string of beads”appearance suggestive of fibromuscular dysplasia.108 Withthe low pretest probability of VAD in children, however, it isimportant to be cautious in interpreting abnormal results.Without data on the rate of these findings in asymptomaticchildren, the positive predictive value for such highly sug-gestive findings is unknown.

A retrospective cohort study at Boston Children’s iden-tified three of 364 children presenting to the emergencydepartment with ataxia duration of less than seven dayswho had a cerebellar infarct. Of the three patients identifiedin this study, MRI provided the standard of imaging,whereas CT failed to identify an abnormality in one of thecases.9

Venous sinus thrombosisCerebellar venous infarction can occur as a result of

isolated venous sinus thrombosis (VST) in the posteriorfossa, accounting for about 2% to 4% of all intracranialthromboses.110 One study by Ruiz-Sandoval et al. uncoverednine patients that presentedwith isolated cerebellar venousinfarction. Of these patients, all except one presented withataxia, with seven complaining of dizziness and seven witha depression in mental status (ranging from drowsiness tocoma). Three patients were aged �18 years (14, 15, and18 years), with all of them found to have an underlyingcoagulopathy (puerperium and protein C deficiency).110

CT is insensitive and nonspecific in diagnosing cerebellarVST, and MRI with magnetic resonance venography provesto be the diagnostic study of choice. Kulkarni et al.111

showed diagnostic findings in three of five patientsimaged, with MRI detecting the venous abnormality in theremaining two. Ruiz-Sandoval et al. reported CT as showing

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 23

abnormalities in all patients studied (hypodensities, masseffect, and hydrocephalus), although in none of the patientswas VST detected. MRI was abnormal in all patientsassessed, with the VST and cerebellar involvement detectedin all cases.110 These data included adolescents and adults,so their validity for younger individuals is unclear.

Infectious

MeningitisMeningitis is commonly included in the differential

diagnosis of acute ataxia,4,8,21 with one group advocatingfor LP in children who have altered mental status andnegative CT findings to rule out meningitis.19 Bacterialmeningitis is a serious life-threatening infection of themeningeal membranes that demands rapid diagnosis andsubsequent treatment with antibiotics.112 Thompsonet al.113 described the onset and time course of clinical signsand symptoms in 544 children, teasing out that fever, sepsisfeatures, hemorrhagic rash, impaired mental state, andmeningism and the main features that appear within thefirst 48 hours of disease. A systematic review of prospectivedata further characterized the sensitivity and specificity of26 different signs and symptoms seen with bacterial men-ingitis.114 Neither study mentions ataxia as an importantpresenting feature of disease.

A 1972 series described four children with bacterialmeningitis who presented with ataxia as a major feature ofillness. All had elevated temperatures and three exhibitedaltered mental status.115 Gieron-Korthals et al.7 includedone child with acute ataxia and viral meningitis; this pa-tient’s clinical details are not described in detail but therewas CSF pleocytosis. Thakkar et al.10 mentioned one pa-tient with ataxia and meningitis, but they did not specifywhether it was bacterial or viral and did not include theCSF findings.

Bacterial meningitis is an unlikely cause of acute ataxia,especially in a child without the classic features of fever ormeningismus. Furthermore, the introduction of the Hae-mophilus influenzae type-b and conjugate Pneumococcalvaccines has dramatically reduced the incidence of bacterialmeningitis in the United States to approximately 200 casesper year, at a mean age 0.6 (0.2 to 4.0) years.116 However,culture-negative (i.e., aseptic) meningitis remains a com-mon occurrence that is difficult to clinically distinguishfrom bacterial meningitis.117 A close examination of the rateof meningitis in the setting of acute ataxia is critical fordetermining the diagnostic utility of LP in the emergencydepartment.

RhombencephalitisRhombencephalitis is an extremely rare manifestation of

disseminated listeria or varicella infection, which has beenreported predominantly in immunosuppressed adults.118,119

In one case series of adults with Listeria rhombencephalitis,ataxia was shown in 27% of patients in whom fever wasuniversally present with meningismus in 64% of patients.All patients with CSF studies were shown to have pleocy-tosis, CT was positive in only eight of 21 patients.118 Onereport documented lesions in the brainstem and cerebellumon MRI.119

Enterovirus 71 is a cause of hand, foot, and mouth dis-ease that can have devastating neurological sequelae and isnotorious for a 1998 epidemic in Taiwan leading to thehospitalization of 320 childrenwith neurological disease, ofwhom 55 died. A total of 41 of these patients were studied,of whom 37 (90%) had rhombencephalitis, 20 (54%) ofwhich had grade I disease characterized by myoclonus witheither tremor or ataxia or both. Seven (19%) had grade IIIdisease, which entails fulminant neurogenic pulmonaryedema because of rapid cardiopulmonary failure. Despitemechanical ventilation and cardiopulmonary support, fiveof these patients died. Milder forms of rhombencephalitiswere varied in their clinical presentation, although nonepresented with changes in mental status. MRI showed le-sions restricted to the brainstem and not affecting anysupratentorial tissue.120 A treatment for enteroviral rhom-bencephalitis has not been proposed other than to provideappropriate supportive care.

LabyrinthitisLabyrinthitis refers to the neurological manifestations of

otitis media, usually as a result of bacterial toxins or hostinflammatory mediators migrating into the inner earwithout invasion of bacteria per se. Although older childrenwill usually report vertigo to their parents, younger childrenwho are unable to verbalize their complaint may appear tohave ataxic gait. Although it is a very unusual cause ofataxia, a clinical evaluation of otitis media should raisesuspicion for this condition. Children stereotypically pre-sent with nausea and vomiting, along with nystagmus.121

Gieron-Korthals et al.7 and Martínez-González et al.8 didnot attribute ataxia to labyrinthitis in any of their patients.MRI is the most sensitive means of detecting labyrinthitis,with the affected cochlea showing contrast enhance-ment.122 Symptoms typically resolve with antibiotics,although a child with multiple occurrences should beevaluated by otolaryngology.121

Miscellaneous

Opsoclonus-myoclonus syndromeOpsoclonus-myoclonus syndrome (OMS) is commonly

associated with occult neuroblastoma, although historicallythe failure to recognize the syndrome has led to a delay indiagnosing the tumor by more than a year after initialpresentation.123 Thakkar et al.10 noted that 10 (8.3%) of their120 patients had OMS, although the clinical details of thesecases are otherwise unreported. A prospective survey in theUnited Kingdom identified 19 individuals with OMS, two ofwhom presented with acute ataxia. In this study, mean ageof presentation was 18 months, and the children presentedwith a combination of ataxia, opsoclonus (chaotic conjugateeye movements often referred to as “dancing eyes”), andirritability. LP was abnormal in one of 10 patients, showingan unspecified pleocytosis.124

The prevalence of neuroblastoma in individuals withOMS has increased during the last several decades, probablydue to a combination of improvements in imaging and anincreasing recognition of the disorder. A retrospective studyof OMS showed a diagnostic prevalence of neuroblastomain 43% of patients evaluated in 2000. The diagnostic focus

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3024

with OMS has thus transitioned to excluding the presenceof occult neuroblastoma.125

Urine catecholamines have only a 24% sensitivity indiagnosing neuroblastoma in the presence of OMS.125

Although significantly increased numbers may help guidediagnosis, further investigation for neuroblastoma is betterdeferred for a time after the acute evaluation in the ED,either when the patient has been admitted or in clinicalfollow-up if the symptoms of ataxia do not subside.

Celiac diseaseCeliac disease is an autoimmune disease triggered by

gluten ingestion. Its onset can be at any age, and childrentypically present with vomiting, diarrhea, constipation, andabdominal pain with infants often demonstrating failure tothrive.126 Gluten sensitivity has been associated with ataxia,with one series demonstrating 16 of 48 children presentingat the onset of their celiac disease with one or moreneurological symptomsdtwo children presented withcerebellar ataxia and had cerebellar atrophy on CT, althoughit is unclear what the timing of symptoms was.127 Adifferent study, however, prospectively followed 835 chil-dren with gluten sensitivity, defined as any child withgluten sensitivity-associated antibodies in their sera andpositive intestinal biopsy. This study showed no childrenwith ataxia or other cerebellar manifestations, with only aslightly higher incidence of neurological manifestationscompared with control subjects.128 In one small series ofpatients with “gluten ataxia,” 12 of 13 patients showedantibodies in their sera that bound to Purkinje cells in rattissue, suggesting a pathologic basis for the association.129

The association of ataxia in childrenwith celiac disease iscontroversial and poorly studied. No individuals withgluten ataxia were documented in either the Gieron-Korthals et al.7 or Martínez-González et al.8 series of acuteataxia.

PseudoataxiaThe final diagnosis we will discuss is ataxia of a nonor-

ganic nature, or pseudoataxia. Gieron-Korthals et al.7

described one patient in their retrospective cohort as hav-ing had a “conversion reaction,” but detailed clinical fea-tures were not included. Many such terms are encounteredin the medical field that imply a nonorganic disorder,among them are “dissociative,” “functional,” “somatoform,”“psychogenic,” “supratentorial,” “hysterical,” and “medi-cally unexplained.” A gait disorder of this nature is oftenreferred to as an “idiopathic” gait disorder.

One study evaluated 103 children with acquired gaitdisorders, eight of whom had idiopathic ataxia amountingto an estimated incidence of 2.9 per 1,000,000 children. Ofthese eight children, five were female and three were male,and they ranged between ten and 15 years of age. All chil-dren with idiopathic gait disorder had significant schoolabsence and were scholastically above average before theirillness. Most of these children had many laboratory in-vestigations completed, including six patients who had MRIof either brain, spine, or both, and five who underwentelectroencephalography.130

Although laboratory and imaging investigations arenormal in this situation, there are no validatedmethodswithwhich a physician could diagnose pseudoataxia. One opinion

piece offers clinical methods for detecting functional gaitdisorders in children, such as an age-appropriate cognitivelychallenging task when walking.131 Although these methodsmay intuitively make sense, their sensitivity and specificityin detecting functional gait disorders have not been studiedand are thus unknown.

Evaluation of acute ataxia

Evaluating a child with acute ataxia is challenging,especially in the absence of clear guidelines to direct clinicalmanagement. Physicians in the emergency department arefaced with the dilemma of obtaining imaging or spinal fluidor deferring such testing for the wards. Until recently, theonly two studies that have examined all patients who pre-sentedwith acute ataxia are those by Gieron-Korthals et al.7

and Martínez-González et al.,8 both of which were retro-spective and only included approximately 40 patients.7,8

As initially discussed, one cannot undervalue the utilityof careful history taking and physical examination skills inevaluating a child with acute ataxia. More often than not, astraightforward cause of ataxia can be derived withoutancillary testing. However, the presence of lateralizingfindings on neurological examination poses a real concernfor acute intracranial pathology that must be aggressivelyinvestigated.

The current literature offers some guidelines on theevaluation of ataxia based on features of the clinical eval-uation and examination, but the positive and negativepredictive values of clinical findings rarely play a role inguiding testing. Furthermore, the sensitivity of certaintesting will play a major role in deciding which tests toorder, or whether to continue testing in the face of negativefindings. In this section of our review, we cover the differ-entiation between patients with isolated ataxia or thosewith concurrent altered mental status as well as the role ofancillary testing in the emergency department for a childpresenting with acute ataxia.

The history

The sudden onset of ataxia can be very distressing toparents and patients alike. It is important that the providerelicits much of the history by thorough questioning, as thepatient or their caretakers may not be able to associate priorevents with the current presentation. For the most acutecases in which there is concern for an evolving intracranialprocess or where the child is medically unstable, a briefhistory should include the presence of any other symptoms,known allergies, current medications, past medical history,last oral intake, and events leading up to the presentingsymptoms. In such scenarios, it may be appropriate to forgodetailed history taking for the sake of triage and timelyintervention.

The timing of symptoms can be helpful in narrowing thedifferential. Ataxia evolving for a period of less than72 hours is unlikely to be of neoplastic origin.7,8 The pres-ence of an infectious syndrome or vaccine administration asfar as 2 weeks before presentation may implicate one of themany postinfectious phenomena as the cause of ataxia. Anyassociated trauma to the head and neck is important to note

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 25

and merits thorough examination, close monitoring, andpotentially imaging.

Hemiplegic ataxia caused by weakness in the postictalstate has been described,132 although it has not been seen inany studies evaluating the different causes of acute ataxia inchildren. Nevertheless, a history focused on the possibilityof epilepsy is important in ambiguous cases as these pa-tients stand to benefit greatly from early diagnosis andintervention. A prior history of intermittent paroxysmalheadachesdwhich caregivers may believe is not relevant toa new onset of ataxia and thus may not readily offerdmayindicate complex migraine or hemiplegic migraine as thecausative entity.133 Recurrence of ataxia in an otherwiseundiagnosed patient should raise suspicion for migraineand seizures as the causative diagnosis. On the other hand, aheadache that is sudden in onset or described as the “worstheadache of my life” should trigger suspicion for hemor-rhage. Any unclear or inconsistent history that is concerningfor nonaccidental trauma should be taken seriously, ashemorrhage in the posterior fossa has been shown to have ahigh occurrence in such cases.107

An ingestion-focused history is important in this popu-lation, as toxic ingestion may mimic concerning pathology.Caretakers should be asked about any medications boththey and the patient may have been taking. Be sure to askspecifically if anyone who had been caring for the childsuffers from anxiety or seizures, as they may respectivelyindicate access to benzodiazepines or AEDs.

The birth history, other than usual questions surround-ing pregnancy, perinatal, and neonatal events, should focuson whether the child has had newborn screening for com-mon metabolic disorders. Ideally, these records would beavailable for review but if they are not, one should neverassume the results to be normal. Parents should be askedabout any history of spontaneous abortions or if any siblingshad abnormal results on their newborn screen. Althoughmetabolic disorders associated with ataxia tend to manifestgradually, the presentation is often “unmasked” by theonset of systemic illness or a discontinuation in their dailyvitamins.134

Physical and neurological examination

The initial examination should focus on signs of systemicillness, including an assessment of vital signs and an ex-amination of the skin for rash. Vital signs should be assessedfor abnormal temperature to assess for infection. The triadof hypertension, bradycardia, and hypopneamay occur withincreased intracranial pressure and has been shown tooccur with ischemia arising from the posteriorcirculation.135

A careful examination of the eyes is critical. Fundoscopicexamination should be performed to assess for signs ofincreased intracranial pressure, whereas retinal hemor-rhaging signals a high likelihood of nonaccidental trauma.Unilateral pupillary dilation can be the harbinger of tem-poral lobe herniation requiring immediate decompression.Bilateral nystagmus is often seen with toxic ingestion,whereas unilateral nystagmus implies a lateralizing defectand is therefore an indication for imaging. Any oph-thalmoplegia can result from a lesion in the correspondingcranial nerve or their nuclei in the brainstem, an assessment

of hearing and corneal sensation will help differentiatebetween the two possibilities. Sudden onset ocular findings,such as nystagmus or ophthalmoplegia, should be consid-ered warning signs of posterior circulation stroke andwarrant immediate imaging.

An important aspect of the physical examination in thesecases is deciding whether the ataxia is because of cerebellardysfunction. Speech can be affected in lesions of the cere-bellum ipsilateral to the dominant hand,136 and iscommonly described as a “scanning speech” in which pa-tients must carefully articulate individually spoken sylla-bles. Finger-to-nose and heel-to-shin can be used to detectdysmetria, a sign indicating ipsilateral cerebellar damage.Reflex testing of the knees may demonstrate a “pendular”response, in which the leg swings more than four timesafter the tendon is struck. Finally, gait testing may reveal astaggering, wide-based gait in which the patient appears tobe “drunk.” It is important to realize these tests may not bepossible to perform in younger children who may not yethave highly developed speech or motor skills and mayinstead present with a refusal to speak or walk.

An assessment of the child’s mental status is critical partof the neurological examination for these patients. Formany emergency department visits, the chief complaintmay often be a change in behavior or personality, withataxia revealed as a major feature only after taking thehistory or performing the physical examination. A formalassessment of the mental status should include the level ofalertness, attention span, speech and language, and short-term memory. As we have discussed, an altered mentalstatus is often the harbinger of urgent intracranial pathol-ogy such as ADEM, cerebellitis with mass effect, or stroke.Conversely, normal mentation is not normally seen withingestion of many toxic substances such as AEDs or ben-zodiazepines.72,83 A patient with ataxia and an alteredmental status should be assumed to have intracranial pa-thology until proven otherwise by MRI or unless there is anunequivocal history of toxic ingestion.

Laboratory testing

The utility of obtaining a complete blood count and acomprehensive metabolic profile has not been explicitlystudied in acute ataxia; however, these tests should be or-dered as a first pass screening tooldany derangement inelectrolytes (with attention to magnesium and phosphate)should be corrected accordingly and subsequently invokediagnostic algorithms themselves.137-139 Transaminitis mayhint toward an inborn error in metabolism.6 Vitamins B1,B12, and E may be implicated in ataxia and should beevaluated in a patient with a history of poor nutrition orcertain restrictive dietsdthese tests may be deferred untilafter admission.140

Any suspicion of underlying metabolic disorder as sug-gested by a positive family history or obstetric history ofspontaneous abortion, parental consanguinity, or that offailure or delay in achieving milestones preceding the onsetof ataxia should be preliminarily investigated by obtainingmetabolic screening laboratories: ammonia, lactate canscreen for mitochondrial disorders such as Leigh syndrome,pyruvate dehydrogenase deficiency,10,141 or errors incatabolism. Abnormalities in these values or a suspicion for

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3026

such a disorder may invoke follow-up levels of pyruvate,urinary amino acids, organic acids, carnitine, and an acyl-carnitine profile. However, these tests are better sent afteradmission and are best interpreted under the guidance of ageneticist or specialist in metabolic disorders.6

Imaging

The diagnostic yield of neuroimaging in children withacute ataxia has previously been estimated to be low, esti-mated by Whelan et al.4 in their review article to beapproximately 2.5% for CT and 5% for MRI. However, thedata they complied were largely from older retrospectivestudies published from 1994 to, most recently, 2006.7,8,18,20

Since then there have been advances in the diagnosticcapability of MRI that have allowed for a greater ability toresolve subtle areas of demyelination142,143 and cytotoxicedema.144 Advances in angiography and the use of contrasthave also improved the ability to detect tissue edema andneoplasms.144

Rudloe et al. reported clinically urgent neuroimagingfindings in 42 (12.5%) of 335 patients presenting with ataxiafor less than seven days. Twenty-two of these patients had atumor, and the remaining 20 had ADEM (12), infarct (three),encephalitis (one), transverse myelitis (one), hemorrhage(one), syrinx with hydrocephalus (one), and mastoiditiswith epidural abscess (one). CT was read as normal in 10(26%) of 38 patients that subsequently showed pathologyon MRI. Of 273 patients with symptoms for less thanthree days, 19 (7.0%) had significant imaging findings.9

Although Thakkar et al.10 did not explicitly report the totalrate of positive imaging results, they did report eight (6.7%)of 120 patients as having been diagnosed with imaging-dependent diagnoses: cerebellitis (three), cerebellar stroke(two), ADEM (two), and cerebral vein thrombosis (one). Thediagnostic yield of imaging cannot truly be assessed in theabsence of prospective research.

MRI is superior to CT in detecting intracranial pathology.Inflammatory diseases of the brain are unlikely to bedetected by CT. Rudloe et al.9 showed that of 10 patientsdiagnosed with ADEM, only two had lesions detected withCT, and the only patient with transverse myelitis had anormal CT. Transverse myelitis can only be detected withMRI and confirmed cases should be followed up with im-aging of the optic nerves to evaluate for neuromyelitisoptica.145

CT is useful for quickly evaluating acute stroke anddetermining if it is hemorrhagic or ischemic; however, MRImay be superior in detecting small or very acute hemor-rhagic lesions.29 Imaging of the posterior fossa with CT isfurther limited by beam hardening artifacts from sur-rounding bony structures.30 Although a study directlycomparing CT with MRI in imaging the posterior fossa hasnot been done, Rudloe et al.9 demonstrated four childrenwith cerebrovascular lesions in the posterior fossa, one ofwhich was not detected with CT.

Imaging the brain with cranial CT may be critical in itsrapid evaluation for neurosurgical issues such as the evo-lution of hemorrhage, mass effect, or obstruction of cerebralCSF outflow. In centers where rapid MRI is available,bypassing CT may save time, avoid unnecessary radiationand save resources as negative CT findings will not

necessarily rule out clinically urgent pathology.146 On theother hand, rapidMRI as previously discussed has only beenvalidated for assessing ventriculoperitoneal shunt mal-function and not for emergent headaches, depression inmental status, of focal neurological findings in the emer-gency department.65,105,106

Lumbar puncture

Abnormalities in the CSF, namely pleocytosis, are notunusual in the postinfectious ataxia disorders, APCA andcerebellitis.7,8,19,20 Cerebellitis is inconsistently associatedwith pleocytosis and cases with positive bacterial culturesare very rare.27,28 Protein levelsmay be increased in patientswith GBS and MFS, but the sensitivity is 44% and 27%,respectively, during the first week of presentation.44

Although increased protein levels in the setting of normalcell counts can help confirm the diagnosis, a presumptivediagnosis of the peripheral neuropathies is largely clinicaland has little dependence on LP.54,55 CSF protein levels aremostly important for prognosticating GBS and monitoringdisease progression,44 and therefore can be reserved untilafter admission to an inpatient unit.

Large studies on the presenting signs of meningitis focuson fever, changes in mental status, hemorrhagic rash, andseptic features to heighten the urgency of performing anLP.113,114,147 However, there is no evidence that ataxia withor without the presence of these features is more or lesssuggestive of bacterial meningitis. LP has been reported toprecipitate brain herniation in patients with bacterialmeningitis, and although CT may be helpful in findingcontraindications to LP, a normal CT scan does not precludethe risk of herniation, which is best assessed clinically(pupillary changes, irregular breathing, and so forth).148

LP can be diagnostic in the setting of subarachnoidhemorrhage, especially several days into the presentation asthe sensitivity of CT drops over time.149,150 However, thisphenomenon is exceedingly rare in children, and no suchpatients have presented with acute ataxia.9

The utility of obtaining CSF studies in the emergencydepartment when directing both diagnosis and manage-ment of acute ataxia is uncertain and remains largelyunstudied.

Toxicology screening

Toxic causes of ataxia can generally be identified fromthe clinical evaluation. Although the history is not alwaysafforded by caregivers, questioning on specific substancesor family members taking specific medications can beuseful. Gieron-Korthals et al.7 reported an a priori history ofingestion in 61% of patients with positive toxicologicfindings.

Screening for substances has long been considered animportant test in the evaluation of acute ataxia,4,11 espe-cially because many cases of ingestion will not necessarilypresent with a history of ingestion. Benzodiazepines remaina frequently ingested class of drugs, especially in toddlersand young children, that are known to cause ataxia. Broadscreening will also help in detecting cases of cannabinoidingestions, which may become more frequent in the face ofliberalizing laws against the sale of marijuana and edibles.

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 27

Ethanol levels should be obtained as part of standard toxi-cology panels, even in young children. Urine or bloodtesting should be considered routine in cases where thepatient is demonstrating decreased level of consciousness,acute confusional state, new onset of psychiatric symptoms,or a panic attack.151 As rapid confirmation of a toxic inges-tion will help to exclude more serious intracranial pathol-ogy from the differential. AEDs known to cause ataxia areless commonly prescribed as they were at the time Gieron-Korthals et al. published their study of ataxia, with in-gestions of these substances also less frequently reported.

Conclusions

The evaluation of a child with acute ataxia in the emer-gency department is a challenge for the clinician in deter-mining the extent and timing of initial screening tests. Inmost studies, abnormalities detected by screening testswere nondiagnostic except for in a minority of patients.Therefore, careful analysis of the medical history should becarried out as the differential is broad, including central andperipheral causes of which some are life threatening. Thespectrum of the etiologic factors in acute ataxia appears tohave changed in the years since Gieron-Korthals et al.published their landmark article in 1994. Some of the majorimplicated changes are the use of new AEDs instead ofphenytoin and carbamazepine and the use of new vaccinesthat may have reduced the incidence of classically post-varicellar cerebellar ataxia. Prospective data are needed todetermine the relative incidence of different causes of acuteataxia.

From the evaluations that could be performed in theemergency room, toxicology screening should be consid-ered in all children presenting with acute ataxia as it couldbe positive up to 49% of the cases or higher if the potentialsubstance is known. Brain MRI is preferred over cranial CTwhen structural lesions are suspected because it offers amore complete evaluation of the posterior fossa. Head CTshould be avoided to prevent radiation unless time is of theessence, as such is the case with hemorrhage, vasculardisorders, hydrocephalus, or impending herniation. Imag-ing can be deferred for observation and follow-up in theclinically improving child who is otherwise well appearing.LP should be considered in patients with fever, meningealfindings, or an abnormal or declining mental status, butotherwise may be deferred until after admission. Otherevaluations for less common causes should be considered inthe context of the presenting symptoms, but evaluation inthe ER is unlikely to change management. Urine catechol-amines in a child with OMS will not be reported until thepatient is already admitted and if it is negative will not ruleout this disorder. There is insufficient evidence at the pre-sent time to establish a role for other tests in less frequentconditions such as autoantibody testing or searching forinborn errors of metabolism in children presenting withacute ataxia. Testing for metabolitesdlactate, pyruvate,ammonia, amino acids, organic acids, carnitine, and acyl-carnitine profiledshould be done if there is any history ofdevelopmental regression or a family history of early deathsor pregnancy losses. Some tests could be normal in the firstfew days of a disease process, as is the case with CSF protein

in GBSda high index of suspicion is required in theappropriate clinical context.

Some institutions have protocols for the management ofacute ataxia in the emergency department and practitionersshould be familiar with these, but more research is neededto answer several fundamental questions with regards tothe initial evaluation. Acute ataxia is a relatively commonpresentation in children seen by pediatric acute servicesand child neurologists as a whole, but the frequency of theirspecific etiologies might be low. We are in need of multi-center prospective research that is well enough powered tounderstand the rates of different causes and how to reliablytest for them. Although a primary concern on initialassessment is to exclude serious causes of this clinicalsyndrome, one must also be able to quickly establish adiagnosis that might require rapid intervention. A goodclinician should be able to keep an open mind whenrecognizing the essentially benign nature of acute ataxia inmost children. The more confidence we have in evaluatingthese patients, the better we can provide guidance andreassurance to the families who may fear the worst.

References

1. Shanks RA. Notes on cerebellar ataxia in childhood. Arch Dis Child.1950;25:389-391.

2. Batten FE. Ataxia in childhood. Brain. 1906;28:484-505.3. Cottom DG. Acute cerebellar ataxia. Arch Dis Child. 1957;32:

181-188.4. 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. Philcox DV, Sellars SL, Pamplett R, Beighton P. Vestibular

dysfunction in hereditary ataxia. Brain. 1975;98:309-316.6. Hoffmann GF, Kölker S. Defects in amino acid catabolism and the

urea cycle. Handb Clin Neurol. 2013;113:1755-1773.7. Gieron-Korthals MA, Westberry KR, Emmanuel PJ. Acute childhood

ataxia: 10-year experience. J Child Neurol. 1994;9:381-384.8. Martínez-González MJ, Martínez-González S, García-Ribes A, Min-

tegi-Raso S, Benito-Fernández J, Prats-Viñas JM. Acute onset ataxiain infancy: its aetiology, treatment and follow-up. Rev Neurol.2006;42:321-324.

9. Rudloe T, Prabhu SP, Gorman MP, et al. The yield of neuroimagingin children presenting to the emergency department with acuteataxia in the post-varicella vaccine era. J Child Neurol. 2015;30:1333-1339.

10. Thakkar K, Maricich SM, Alper G. Acute ataxia in childhood: 11-year experience at a major pediatric neurology referral center.J Child Neurol. 2016;31:1156-1160.

11. Poretti A, Benson J, Huisman T, Boltshauser E. Acute ataxia inchildren: approach to clinical presentation and role of additionalinvestigations. Neuropediatrics. 2013;44:127-141.

12. Guess HA, Broughton DD, Melton LJ, Kurland LT. Population-basedstudies of varicella complications. Pediatrics. 1986;78:723-727.

13. Sharrar RG, LaRussa P, Galea SA, et al. The postmarketing safetyprofile of varicella vaccine. Vaccine. 2000;19:916-923.

14. Fritzler MJ, Zhang M, Stinton LM, Rattner JB. Spectrum of centro-some autoantibodies in childhood varicella and post-varicellaacute cerebellar ataxia. BMC Pediatr. 2003;3:11.

15. Ito H, Sayama S, Irie S, et al. Antineuronal antibodies in acutecerebellar ataxia following Epstein-Barr virus infection. Neurology.1994;44:1506-1507.

16. Uchibori A, Sakuta M, Kusunoki S, Chiba A. Autoantibodies in post-infectious acute cerebellar ataxia. Neurology. 2005;65:1114-1116.

17. Cimolai N, Mah D, Roland E. Anticentriolar autoantibodies inchildren with central nervous system manifestations of Myco-plasma pneumoniae infection. J Neurol Neurosurg Psychiatry. 1994;57:638-639.

18. Connolly AM, Dodson WE, Prensky AL, Rust RS. Course andoutcome of acute cerebellar ataxia. Ann Neurol. 1994;35:673-679.

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3028

19. Prasad M, Ong MT, Setty G, Whitehouse WP. Fifteen-minuteconsultation: the child with acute ataxia. Arch Dis Child Educ PractEd. 2013;98:217-223.

20. Nussinovitch M, Prais D, Volovitz B, Shapiro R, Amir J. Post-infec-tious acute cerebellar ataxia in children. Clin Pediatr. 2003;42:581-584.

21. Ryan MM, Engle EC. Topical review: acute ataxia in childhood.J Child Neurol. 2003;18:309-316.

22. Dowling MD, Van Slyck EJ. Cerebellar disease in infectious mono-nucleosis. Arch Neurol. 1966;15:270-274.

23. Levin S. Cerebellar ataxia following chickenpox. Lancet. 1960;275:1222-1223.

24. Horowitz MB, Pang D, Hirsch W. Acute cerebellitis: case report andreview. Pediatr Neurosurg. 1991;17:142-145.

25. Asenbauer B, McConachie NS, Allcutt D, Farrell MA, King MD. Acutenear-fatal parainfectious cerebellar swelling with favourableoutcome. Neuropediatrics. 1997;28:122-125.

26. Levy EI, Harris AE, Omalu BI, Hamilton RL, Branstetter BF,Pollack IF. Sudden death from fulminant acute cerebellitis. PediatrNeurosurg. 2001;35:24-28.

27. Sawaishi Y, Takada G. Acute cerebellitis. Cerebellum. 2002;1:223-228.

28. Desai J, Mitchell WG. Acute cerebellar ataxia, acute cerebellitis, andopsoclonus-myoclonus syndrome. J Child Neurol. 2012;27:1482-1488.

29. Kidwell CS, Chalela JA, Saver JL, et al. Comparison of MRI and CT fordetection of acute intracerebral hemorrhage. JAMA. 2004;292:1823-1830.

30. Alberico RA, Loud P, Pollina J, Greco W, Patel M, Klufas R. Thick-section reformatting of thinly collimated helical CT for reduction ofskull base-related artifacts. AJR Am J Roentgenol. 2000;175:1361-1366.

31. Krupp LB, Tardieu M, Amato MP, et al. International PediatricMultiple Sclerosis Study Group criteria for pediatric multiplesclerosis and immune-mediated central nervous system demye-linating disorders: revisions to the 2007 definitions. Mult Scler J.2013;19:1261-1267.

32. Krupp LB, Banwell B, Tenembaum S, International Pediatric MSStudy Group. Consensus definitions proposed for pediatric multi-ple sclerosis and related disorders. Neurology. 2007;68:S7-S12.

33. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria formultiple sclerosis: 2010 revisions to the McDonald criteria. AnnNeurol. 2011;69:292-302.

34. Dale RC, Pillai SC. Early relapse risk after a first CNS inflammatorydemyelination episode: examining international consensus defi-nitions. Dev Med Child Neurol. 2007;49:887-893.

35. Mikaeloff Y, Suissa S, Vallée L, et al. First episode of acute CNSinflammatory demyelination in childhood: prognostic factors formultiple sclerosis and disability. J Pediatr. 2004;144:246-252.

36. Leake JA, Albani S, Kao AS, et al. Acute disseminated encephalo-myelitis in childhood: epidemiologic, clinical and laboratory fea-tures. Pediatr Infect Dis J. 2004;23:756-764.

37. Alper G. Acute disseminated encephalomyelitis. J Child Neurol.2012;27:1408-1425.

38. Tenembaum S, Chamoles N, Fejerman N. Acute disseminatedencephalomyelitis: a long-term follow-up study of 84 pediatricpatients. Neurology. 2002;59:1224-1231.

39. Fridinger SE, Alper G. Defining encephalopathy in acute dissemi-nated encephalomyelitis. J Child Neurol. 2014;29:751-755.

40. Pavone P, Pettoello-Mantovano M, Le Pira A, et al. Acute dissemi-nated encephalomyelitis: a long-term prospective study and meta-analysis. Neuropediatrics. 2011;41:246-255.

41. Mikaeloff Y, Caridade G, Husson B, Suissa S, Tardieu M. Acutedisseminated encephalomyelitis cohort study: prognostic factorsfor relapse. Eur J Paediatric Neurol. 2007;11:90-95.

42. Absoud M, Parslow R, Wassmer E, et al. Severe acute disseminatedencephalomyelitis: a paediatric intensive care population-basedstudy. Mult Scler J. 2011;17:1258-1261.

43. Nishikawa M, Ichiyama T, Hayashi T, Ouchi K, Furukawa S. Intra-venous immunoglobulin therapy in acute disseminated encepha-lomyelitis. Pediatr Neurol. 1999;21:583-586.

44. Nishimoto Y, Odaka M, Hirata K, Yuki N. Usefulness of anti-GQ1bIgG antibody testing in Fisher syndrome compared with cerebro-spinal fluid examination. J Neuroimmunol. 2004;148:200-205.

45. Odaka M, Yuki N, Hirata K. Anti-GQ1b IgG antibody syndrome:clinical and immunological range. J Neurol Neurosurg Psychiatry.2001;70:50-55.

46. Korinthenberg R, Schessl J, Kirschner J. Clinical presentation andcourse of childhood Guillain-Barré syndrome: a prospective mul-ticentre study. Neuropediatrics. 2007;38:10-17.

47. van Doorn PA, Ruts L, Jacobs BC. Clinical features, pathogenesis,and treatment of Guillain-Barré syndrome. Lancet Neurol. 2008;7:939-950.

48. Yuki N, Kokubun N, Kuwabara S, et al. GuillaineBarré syndromeassociated with normal or exaggerated tendon reflexes. J Neurol.2011;259:1181-1190.

49. Samadi M, Kazemi B, Golzari Oskoui S, Barzegar M. Assessment ofautonomic dysfunction in childhood Guillain-Barré syndrome.J Cardiovasc Thorac Res. 2013;5:81-85.

50. Moulin DE, Hagen N, Feasby TE, Amireh R, Hahn A. Pain in Guillain-Barre syndrome. Neurology. 1997;48:328-331.

51. Paradiso G, Tripoli J, Galicchio S, Fejerman N. Epidemiological,clinical, and electrodiagnostic findings in childhood Guillain-Barrésyndrome: a reappraisal. Ann Neurol. 1999;46:701-707.

52. Sakakihara Y, Kamoshita S. Age-associated changes in the symp-tomatology of Guillain-Barré syndrome in children. Dev Med ChildNeurol. 1991;33:611-616.

53. Asbury AK. Guillain-Barré syndrome: historical aspects. Ann Neu-rol. 1990;27:S2-S6.

54. Ropper AH. The Guillain-Barré syndrome. N Engl J Med. 1992;326:1130-1136.

55. Van der Meché FG, van Doorn PA, Meulstee J, Jennekens FG, GBS-consensus group of the Dutch Neuromuscular Research SupportCentre. Diagnostic and classification criteria for the Guillain-Barrésyndrome. Eur Neurol. 2001;45:133-139.

56. van den Berg B, Walgaard C, Drenthen J, Fokke C, Jacobs BC, vanDoorn PA. GuillaineBarré syndrome: pathogenesis, diagnosis,treatment and prognosis. Nat Rev Neurol. 2014;10:469-482.

57. Zuccoli G, Panigrahy A, Bailey A, Fitz C. Redefining the Guillain-Barre spectrum in children: neuroimaging findings of cranialnerve involvement. Am J Neuroradiol. 2011;32:639-642.

58. Hughes RA, Swan AV, Raphaël JC, Annane D, van Koningsveld R,van Doorn PA. Immunotherapy for Guillain-Barré syndrome: asystematic review. Brain. 2007;130:2245-2257.

59. Lin JJ, Hsia SH, Wang HS, et al. Clinical variants of Guillain-Barrésyndrome in children. Pediatr Neurol. 2012;47:91-96.

60. Odaka M. Bickerstaff’s brainstem encephalitis: clinical features of62 cases and a subgroup associated with Guillain-Barre syndrome.Brain. 2003;126:2279-2290.

61. Koga M, Kusunoki S, Kaida K, et al. Nationwide survey of patientsin Japan with Bickerstaff brainstem encephalitis: epidemiologicaland clinical characteristics. J Neurol Neurosurg Psychiatry. 2012;83:1210-1215.

62. Sekiguchi Y, Mori M, Misawa S, et al. How often and when Fishersyndrome is overlapped by Guillain-Barré syndrome or Bickerstaffbrainstem encephalitis? Eur J Neurol. 2016;23:1058-1063.

63. Transverse Myelitis Consortium Working Group. Proposed diag-nostic criteria and nosology of acute transverse myelitis.Neurology. 2002;59:499-505.

64. Alper G, Petropoulou KA, Fitz CR, Kim Y. Idiopathic acute trans-verse myelitis in children: an analysis and discussion of MRIfindings. Mult Scler J. 2011;17:74-80.

65. Kim I, Torrey SB, Milla SS, Torch MC, Tunik MG, Foltin JC. Benefitsof brain magnetic resonance imaging over computed tomographyin children requiring emergency evaluation of ventriculoperitonealshunt malfunction: reducing lifetime attributable risk of cancer.Pediatr Emerg Care. 2015;31:239-242.

66. Mowry JB, Spyker DA, Brooks DE, McMillan N, Schauben JL. 2014Annual report of the American Association of Poison ControlCenters’ National Poison Data System (NPDS): 32nd annual report.Clin Toxicol. 2015;53:962-1147.

67. Lovegrove MC, Mathew J, Hampp C, Governale L, Wysowski DK,Budnitz DS. Emergency hospitalizations for unsupervised pre-scription medication ingestions by young children. Pediatrics.2014;134:e1009-e1016.

68. Budnitz DS, Salis S. Preventing medication overdoses in youngchildren: an opportunity for harm elimination. Pediatrics. 2011;127:e1597-e1599.

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e30 29

69. Mowry JB, Spyker DA, Cantilena Jr LR, Bailey JE, Ford M. 2012Annual report of the American Association of Poison ControlCenters’ National Poison Data System (NPDS): 30th annual report.Clin Toxicol. 2013;51:949-1229.

70. Bronstein AC, Spyker DA, Cantilena Jr LR, Rumack BH, Dart RC.2011 Annual report of the American Association of Poison ControlCenters’ National Poison Data System (NPDS): 29th annual report.Clin Toxicol. 2012;50:911-1164.

71. Bronstein AC, Spyker DA, Cantilena LR, Green JL, Rumack BH,Dart RC. 2010 Annual report of the American Association of PoisonControl Centers’ National Poison Data System (NPDS): 28th annualreport. Clin Toxicol. 2011;49:910-941.

72. Wiley CC, Wiley JF. Pediatric benzodiazepine ingestion resulting inhospitalization. J Toxicol Clin Toxicol. 1998;36:227-231.

73. Sigg EB, Sigg TD. The modification of the pupillary light reflex bychlorpromazine, diazepam, and pentobarbital. Brain Res. 1973;50:77-86.

74. Hou RH, Samuels ER, Langley RW, Szabadi E, Bradshaw CM.Arousal and the pupil: why diazepam-induced sedation is notaccompanied by miosis. Psychopharmacology. 2007;195:41-59.

75. Tweedy CM, Milgrom P, Kharasch ED, Kaakko T, Spieker M,Coldwell SE. Pharmacokinetics and clinical effects of sublingualtriazolam in pediatric dental patients. J Clin Psychopharmacol.2001;21:268-272.

76. Mowry JB, Spyker DA, Cantilena Jr LR, McMillan N, Ford M. 2013Annual report of the American Association of Poison ControlCenters’ National Poison Data System (NPDS): 31st annual report.Clin Toxicol. 2014;52:1032-1283.

77. Bronstein AC, Spyker DA, Cantilena JR LR, Green JL, Rumack BH,Giffin SL. 2009 Annual report of the American Association of PoisonControl Centers’ National Poison Data System (NPDS): 27th annualreport. Clin Toxicol. 2010;48:979-1178.

78. Litovitz TL, Schmitz BF, Holm KC. 1988 Annual report of theAmerican Association of Poison Control Centers National DataCollection System. Am J Emerg Med. 1989;7:495-545.

79. Litovitz TL, Schmitz BF, Bailey KM. 1989 Annual report of theAmerican Association of Poison Control Centers National DataCollection System. Am J Emerg Med. 1990;8:394-442.

80. Litovitz TL, Bailey KM, Schmitz BF, Holm KC, Klein-Schwartz W.1990 Annual report of the American Association of Poison ControlCenters National Data Collection System. Am J Emerg Med. 1991;9:461-509.

81. Litovitz TL, Holm KC, Bailey KM, Schmitz BF. 1991 Annual report ofthe American Association of Poison Control Centers National DataCollection System. Am J Emerg Med. 1992;10:452-505.

82. Litovitz TL, Holm KC, Clancy C, Schmitz BF, Clark LR, Oderda GM.1992 Annual report of the American Association of Poison ControlCenters Toxic Exposure Surveillance System. Am J Emerg Med.1993;11:494-555.

83. Lifshitz M, Gavrilov V, Sofer S. Signs and symptoms of carbamaz-epine overdose in young children. Pediatr Emerg Care. 2000;16:26-27.

84. Leong C, Mamdani MM, Gomes T, Juurlink DN, Macdonald EM,Yogendran M. Antiepileptic use for epilepsy and nonepilepsy dis-orders: a population-based study (1998-2013). Neurology. 2016;86:939-946.

85. Pickrell WO, Lacey AS, Thomas RH, Lyons RA, Smith PEM, Rees MI.Trends in the first antiepileptic drug prescribed for epilepsy be-tween 2000 and 2010. Seizure. 2014;23:77-80.

86. Spiller H, Strauch J, Essing-Spiller S, Burns G. Thirteen years ofoxcarbazepine exposures reported to US poison centers: 2000 to2012. Hum Exp Toxicol. 2016;35:1055-1059.

87. Mattson RH. Efficacy and adverse effects of established and newantiepileptic drugs. Epilepsia. 1995;36 Suppl 2:S13-S26.

88. Boyer EW. Dextromethorphan abuse. Pediatr Emerg Care. 2004;20:858-863.

89. Antoniou T, Juurlink DN. Dextromethorphan abuse. CMAJ. 2014;186:E631.

90. Warden CR, Diekema DS, Robertson WO. Dystonic reaction asso-ciated with dextromethorphan ingestion in a toddler. PediatrEmerg Care. 1997;13:214-215.

91. Roberge RJ, Hirani KH, Rowland PL, Berkeley R, Krenzelok EP.Dextromethorphan- and pseudoephedrine-induced agitated psy-chosis and ataxia: case report. J Emerg Med. 1999;17:285-288.

92. Lamminpää A. Acute alcohol intoxication among children andadolescents. Eur J Pediatr. 1994;153:868-872.

93. Dar MS. Ethanol-induced cerebellar ataxia: cellular and molecularmechanisms. Cerebellum. 2015;14:447-465.

94. Van Skike CE, Botta P, Chin VS, et al. Behavioral effects of ethanol incerebellum are age dependent: potential system and molecularmechanisms. Alcohol Clin Exp Res. 2010;34:2070-2080.

95. Moran JM. Home-use oral hygiene products: mouthrinses. Perio-dontol 2000. 2008;48:42-53.

96. Blackstone M, Callahan J. An unsteady walk in the park. PediatrEmerg Care. 2008;24:193-195.

97. Boros CA, Parsons DW, Zoanetti GD, Ketteridge D, Kennedy D.Cannabis cookies: a cause of coma. J Paediatr Child Health. 1996;32:194-195.

98. van Amsterdam J, Brunt T, van den Brink W. The adverse healtheffects of synthetic cannabinoids with emphasis on psychosis-likeeffects. J Psychopharmacol. 2015;29:254-263.

99. Jordan LC, Hillis AE. Hemorrhagic stroke in children. Pediatr Neurol.2007;36:73-80.

100. Ganesan V, Chong WK, Cox TC, Chawda SJ, Prengler M, Kirkham FJ.Posterior circulation stroke in childhood: risk factors and recur-rence. Neurology. 2002;59:1552-1556.

101. Feudale F, Liebelt E. Recognizing vertebral artery dissection inchildren: a case report. Pediatr Emerg Care. 2000;16:184-188.

102. de Ribaupierre S, Rilliet B, Cotting J, Regli L. A 10-year experiencein paediatric spontaneous cerebral hemorrhage: which childrenwith headache need more than a clinical examination? Swiss MedWkly. 2008;138:59-69.

103. Kirollos RW, Tyagi AK, Ross SA, van Hille PT, Marks PV. Manage-ment of spontaneous cerebellar hematomas: a prospective treat-ment protocol. Neurosurgery. 2001;49:1378-1386. discussion1386-7.

104. Puzzilli F, Salvati M, Bristot R, et al. Spontaneous cerebellar hae-morrhages in childhood. Childs Nerv Syst. 1997;13:601-603. dis-cussion 604.

105. Ashley WW, McKinstry RC, Leonard JR, Smyth MD, Lee BC, Park TS.Use of rapid-sequence magnetic resonance imaging for evaluationof hydrocephalus in children. J Neurosurg. 2005;103:124-130.

106. Wait SD, Lingo R, Boop FA, Einhaus SL. Eight-second MRI scan forevaluation of shunted hydrocephalus. Childs Nerv Syst. 2012;28:1237-1241.

107. Kadom N, Khademian Z, Vezina G, Shalaby-Rana E, Rice A, Hinds T.Usefulness of MRI detection of cervical spine and brain injuries inthe evaluation of abusive head trauma. Pediatr Radiol. 2014;44:839-848.

108. Hasan I, Wapnick S, Tenner MS, Couldwell WT. Vertebral arterydissection in children: a comprehensive review. Pediatr Neurosurg.2002;37:168-177.

109. Mackay MT, Prabhu SP, Coleman L. Childhood posterior circulationarterial ischemic stroke. Stroke. 2010;41:2201-2209.

110. Ruiz-Sandoval JL, Chiquete E, Navarro-Bonnet J, et al. Isolated veinthrombosis of the posterior fossa presenting as localized cere-bellar venous infarctions or hemorrhages. Stroke. 2010;41:2358-2361.

111. Kulkarni GB, Mustare V, Abbas MM. Profile of patients with cere-bral venous sinus thrombosis with cerebellar involvement. J StrokeCerebrovasc Dis. 2014;23:1106-1111.

112. Nigrovic LE, Kuppermann N, Macias CG, et al. Clinical predictionrule for identifying children with cerebrospinal fluid pleocytosis atvery low risk of bacterial meningitis. JAMA. 2007;297:52-60.

113. Thompson MJ, Ninis N, Perera R, et al. Clinical recognition ofmeningococcal disease in children and adolescents. Lancet. 2006;367:397-403.

114. Curtis S, Stobart K, Vandermeer B, Simel DL, Klassen T. Clinicalfeatures suggestive of meningitis in children: a systematic reviewof prospective data. Pediatrics. 2010;126:952-960.

115. Schwartz JF. Ataxia in bacterial meningitis. Neurology. 1972;22:1071-1074.

116. Nigrovic LE, Kuppermann N, Malley R, for the Bacterial MeningitisStudy Group of the Pediatric Emergency Medicine CollaborativeResearch Committee of the American Academy of Pediatrics.Children with bacterial meningitis presenting to the emergencydepartment during the pneumococcal conjugate vaccine era. AcadEmerg Med. 2008;15:522-528.

M. Caffarelli et al. / Pediatric Neurology 65 (2016) 14e3030

117. Dubos F, Martinot A, Gendrel D, Bréart G, Chalumeau M. Clinicaldecision rules for evaluating meningitis in children. Curr OpinNeurol. 2009;22:288-293.

118. Büla CJ, Bille J, Glauser MP. An epidemic of food-borne listeriosis inwestern Switzerland: description of 57 cases involving adults. ClinInfect Dis. 1995;20:66-72.

119. Alper G, Knepper L, Kanal E. MR findings in listerial rhombence-phalitis. Am J Neuroradiol. 1996;17:593-596.

120. Huang CC, Liu CC, Chang YC, Chen CY, Wang ST, Yeh TF. Neurologiccomplications in children with enterovirus 71 infection. N Engl JMed. 1999;341:936-942.

121. Kitsko DJ, Dohar JE. Inner ear and facial nerve complications ofacute otitis media, including vertigo. Curr Allergy Asthma Rep.2007;7:444-450.

122. Hegarty JL, Patel S, Fischbein N, Jackler RK, Lalwani AK. The valueof enhanced magnetic resonance imaging in the evaluation ofendocochlear disease. Laryngoscope. 2002;112:8-17.

123. Russo C, Cohn SL, Petruzzi MJ, de Alarcon PA. Long-term neuro-logic outcome in children with opsoclonus-myoclonus associatedwith neuroblastoma: a report from the Pediatric Oncology Group.Med Pediatr Oncol. 1997;28:284-288.

124. Pang KK, de Sousa C, Lang B, Pike MG. A prospective study of thepresentation and management of dancing eye syndrome/opsoclonus-myoclonus syndrome in the United Kingdom. Eur JPaediatric Neurol. 2010;14:156-161.

125. Brunklaus A, Pohl K, Zuberi SM, de Sousa C. Investigating neuro-blastoma in childhood opsoclonus-myoclonus syndrome. Arch DisChild. 2012;97:461-463.

126. Guandalini S, Assiri A. Celiac disease. JAMA Pediatr. 2014;168:272-277.

127. Diaconu G, Burlea M, Grigore I, Anton DT, Trandafir LM. Celiacdisease with neurologic manifestations in children. Rev Med ChirSoc Med Nat Iasi. 2013;117:88-94.

128. Ruggieri M, Incorpora G, Polizzi A, Parano E, Spina M,Pavone P. Low prevalence of neurologic and psychiatric man-ifestations in children with gluten sensitivity. J Pediatr. 2008;152:244-249.e1.

129. Hadjivassiliou M, Boscolo S, Davies-Jones GA, et al. The humoralresponse in the pathogenesis of gluten ataxia. Neurology. 2002;58:1221-1226.

130. Wassmer E, Wright E, Rideout S, Whitehouse WP. Idiopathic gaitdisorder among in-patients with acquired gait disorders admittedto a children’s hospital. Pediatr Rehabil. 2002;5:21-28.

131. Dooley JM, Gordon KE. Spurious neurological signs and symptomsin children. Paediatr Child Health. 2011;16:e51-e53.

132. Bansal SK, Chopra JS. Reversible postictal ataxic hemiparesis. Ital JNeurol Sci. 1991;12:75-79.

133. Gardiner AR, Jaffer F, Dale RC, et al. The clinical and geneticheterogeneity of paroxysmal dyskinesias. Brain. 2015;138:3567-3580.

134. Wolf B, Pomponio RJ, Norrgard KJ, et al. Delayed-onset profoundbiotinidase deficiency. J Pediatr. 1998;132:362-365.

135. Wan WH, Ang BT, Wang E. The cushing response: a case for areview of its role as a physiological reflex. J Clin Neurosci. 2008;15:223-228.

136. Baillieux H, De Smet HJ, Dobbeleir A, Paquier PF, De Deyn PP,MariEn P. Cognitive and affective disturbances following focalcerebellar damage in adults: a neuropsychological and SPECTstudy. Cortex. 2010;46:869-879.

137. Misra A, Kumar S, Venkatchalan U, Mukhopadhyay DK, Bhatla B.Cerebellar ataxia due to hyponatremia. Postgrad Med J. 1992;68:230-231.

138. Verbalis JG, Goldsmith SR, Greenberg A, et al. Diagnosis, evalua-tion, and treatment of hyponatremia: expert panel recommenda-tions. Am J Med. 2013;126:S1-S42.

139. Chang WT, Radin B, McCurdy MT. Calcium, magnesium, andphosphate abnormalities in the emergency department. EmergMed Clin North Am. 2014;32:349-366.

140. Anheim M, Tranchant C, Koenig M. The autosomal recessivecerebellar ataxias. N Engl J Med. 2012;366:636-646.

141. Menezes MP, Ouvrier RA. Peripheral neuropathy associated withmitochondrial disease in children. Dev Med Child Neurol. 2012;54:407-414.

142. Kober T, Granziera C, Ribes D, et al. MP2RAGE multiple sclerosismagnetic resonance imaging at 3 T. Invest Radiol. 2012;47:346-352.

143. Runge VM. Current technological advances in magnetic resonancewith critical impact for clinical diagnosis and therapy. InvestRadiol. 2013;48:869-877.

144. Runge VM, Aoki S, Bradley WG, et al. Magnetic resonance imagingand computed tomography of the braind50 years of innovation,with a focus on the future. Invest Radiol. 2015;50:551-556.

145. Goh C, Phal PM, Desmond PM. Neuroimaging in acute transversemyelitis. Neuroimaging Clin N Am. 2011;21:951-973. x.

146. Koral K, Blackburn T, Bailey AA, Koral KM, Anderson J. Strength-ening the argument for rapid brain MR imaging: estimation ofreduction in lifetime attributable risk of developing fatal cancer inchildren with shunted hydrocephalus by instituting a rapid brainMR imaging protocol in lieu of head CT. Am J Neuroradiol. 2012;33:1851-1854.

147. Brouwer MC, Tunkel AR, van de Beek D. Epidemiology, diagnosis,and antimicrobial treatment of acute bacterial meningitis. ClinMicrobiol Rev. 2010;23:467-492.

148. Joffe AR. Lumbar puncture and brain herniation in acute bacterialmeningitis: a review. J Intensive Care Med. 2007;22:194-207.

149. Horstman P, Linn FH, Voorbij HA, Rinkel GJ. Chance of aneurysm inpatients suspected of SAH who have a ‘negative’ CT scan but a‘positive’ lumbar puncture. J Neurol. 2011;259:649-652.

150. Edlow JA, Malek AM, Ogilvy CS. Aneurysmal subarachnoid hem-orrhage: update for emergency physicians. J Emerg Med. 2008;34:237-251.

151. Kozer E, Bar-Hamburger R, Rosenfeld N, et al. Strategy forincreasing detection rates of drug and alcohol abuse in paediatricemergency departments. Acta Paediatr. 2009;98:1637-1640.