Infectious and Autoantibody-AssociatedEncephalitis: Clinical Features andLong-term OutcomeSekhar C. Pillai, FRACPa,b, Yael Hacohen, MRCPCHc, Esther Tantsis, FRACP, PhDa,b, Kristina Prelog, FRANZCRd, Vera Merheb, BSca,Alison Kesson, PhDe, Elizabeth Barnes, PhDf,g, Deepak Gill, FRACPb, Richard Webster, FRACPb, Manoj Menezes, FRACPb,Simone Ardern-Holmes, FRACPb, Sachin Gupta, FRACPb, Peter Procopis, FRACPb, Christopher Troedson, FRACPb,Jayne Antony, FRACPb, Robert A. Ouvrier, FRACPb, Yann Polfrit, MBBSh, Nicholas W. S. Davies, PhDi, Patrick Waters, PhDc,Bethan Lang, PhDc, Ming J. Lim, PhDc,j, Fabienne Brilot, PhDa, Angela Vincent, FRCPathc, Russell C. Dale, PhDa,b

abstract BACKGROUND AND OBJECTIVES: Pediatric encephalitis has a wide range of etiologies, clinicalpresentations, and outcomes. This study seeks to classify and characterize infectious, immune-mediated/autoantibody-associated and unknown forms of encephalitis, including relativefrequencies, clinical and radiologic phenotypes, and long-term outcome.

METHODS: By using consensus definitions and a retrospective single-center cohort of 164 Australianchildren, we performed clinical and radiologic phenotyping blinded to etiology and outcomes, andwe tested archived acute sera for autoantibodies to N-methyl-D-aspartate receptor, voltage-gatedpotassium channel complex, and other neuronal antigens. Through telephone interviews, we definedoutcomes by using the Liverpool Outcome Score (for encephalitis).

RESULTS: An infectious encephalitis occurred in 30%, infection-associated encephalopathy in 8%,immune-mediated/autoantibody-associated encephalitis in 34%, and unknown encephalitis in 28%.In descending order of frequency, the larger subgroups were acute disseminated encephalomyelitis(21%), enterovirus (12%), Mycoplasma pneumoniae (7%), N-methyl-D-aspartate receptor antibody(6%), herpes simplex virus (5%), and voltage-gated potassium channel complex antibody (4%).Movement disorders, psychiatric symptoms, agitation, speech dysfunction, cerebrospinal fluidoligoclonal bands, MRI limbic encephalitis, and clinical relapse were more common in patients withautoantibodies. An abnormal outcome occurred in 49% of patients after a median follow-up of5.8 years. Herpes simplex virus and unknown forms had the worst outcomes. According to ourmultivariate analysis, an abnormal outcome was more common in patients with status epilepticus,magnetic resonance diffusion restriction, and ICU admission.

CONCLUSIONS: We have defined clinical and radiologic phenotypes of infectious and immune-mediated/autoantibody-associated encephalitis. In this resource-rich cohort, immune-mediated/autoantibody-associated etiologies are common, and the recognition and treatment of theseentities should be a clinical priority.

WHAT’S KNOWN ON THIS SUBJECT: Encephalitis isa serious and disabling condition. There areinfectious and immune-mediated causes ofencephalitis, but many cases remain undiagnosed.

WHAT THIS STUDY ADDS: This large single-centerstudy on childhood encephalitis provides insightinto the relative frequency and clinicoradiologicphenotypes of infectious, autoantibody-associated,and unknown encephalitis. Risk factors for anabnormal outcome are also defined.

aNeuroimmunology Group, Institute of Neuroscience and Muscle Research at the Kids Research Institute, Children’sHospital at Westmead, University of Sydney, Australia; bTY Nelson Department of Neurology and Neurosurgery andDepartments of dMedical Imaging, eInfectious Diseases and Microbiology, and fStatistics, the Children’s Hospital atWestmead, Sydney, Australia; cNuffield Department of Clinical Neurosciences, John Radcliffe Hospital, University of Oxford,Oxford, England; gNational Health Medical Research Council Clinical Trials Centre, University of Sydney, Camperdown, NewSouth Wales, Australia; hCentre Hospitalier Territorial Magenta, Service Pediatric, Nouméa, New Caledonia; iChelsea &Westminster Hospital, Department of Neurology, Imperial College Healthcare National Health Service Trust, London,England; and jEvelina Children’s Hospital, London, England

Dr Pillai acquired all clinical data, performed radiological phenotyping, retrieved investigationdata, performed outcome assessments, performed statistical analysis, wrote the first draft, andedited subsequent drafts; Dr Hacohen, Ms Merheb, and Drs Waters, Lang, Lim, and Brilotperformed and supervised all autoantibody testing and edited drafts; Dr Kesson performedinfectious investigation supervision, edited drafts, and

ARTICLE PEDIATRICS Volume 135, number 4, April 2015 by guest on January 23, 2020www.aappublications.org/newsDownloaded from

Encephalitis is inflammation of thebrain parenchyma and has manyinfectious and immune-mediatedetiologies.1 Acute encephalitis can belife threatening and results inpermanent neuropsychiatric andcognitive impairments in a significantproportion of survivors.

Inflammatory forms of encephalitissuch as acute disseminatedencephalomyelitis (ADEM) have beenknown for some time, but in the pastdecade other autoimmune forms ofencephalitis have been recognized,defined by the presence ofautoantibodies against cell surfaceantigens. Autoantibody detection isimportant as immune suppressionmay improve outcomes, particularlywhen given early in the disease.2,3 Arecent study suggested that N-methyl-D-aspartate receptor antibody(NMDAR-Ab) encephalitis, the best-known autoimmune form, is morecommon than herpes simplex virus(HSV) encephalitis in children andyoung adults,4 and a prospectivestudy of adults (n = 134) and children(n = 69) with encephalitis found thatinfectious encephalitis and immune-mediated forms were identified in42% and 21% of the cohort,respectively.5

However, most of the previous largerstudies of encephalitis in childrenwere undertaken before the advent ofneuronal autoantibody testing.6–8 Westudied clinical and radiologicfeatures, serum autoantibodies inacute archived samples, and long-term outcomes in a largeretrospective cohort of pediatricencephalitis from a single center.

METHODS

Patients, Definitions, and ClinicalData

Patients

We report patients with encephalitisseen between January 1998 andOctober 2010 at the Children’sHospital at Westmead, the largesttertiary children’s hospital in New

South Wales, Australia. Patients eitherpresented directly to the hospitalemergency department or werereferred from district hospitals inwestern Sydney and New South Wales(referral base ∼4.5 million people).Patients were identifiedretrospectively from InternationalClassification of Diseases, 10thRevision coding of encephalitis orencephalomyelitis at discharge, plusscreening of consultantcorrespondence from the Departmentof Neurology database and review ofcerebrospinal fluid (CSF) polymerasechain reaction testing from theDepartment of Infectious Diseases andMicrobiology. The following patientswere excluded: neonates (,1 monthold), patients .16 years old, known orsuspected cases of immunodeficiency,and viral or bacterial meningitiswithout encephalitis. Infection-associated encephalopathy (influenzaand rotavirus) cases were included. Atotal of 164 patients fulfilled thediagnostic criteria for acuteencephalitis, according to the followingdefinitions.

Definitions

We used the Granerod encephalitiscase definitions, with minormodifications.1 Encephalitis wasdefined as an acute encephalopathywith $2 of the following: fever $38°C,seizures or focal neurologic signs, CSFpleocytosis ($5 white blood cells/µL)or elevated CSF neopterin (.30 nmol/L), EEG consistent with encephalitis(presence of slowing in this study), orMRI findings suggestive ofencephalitis. CSF neopterin, which isproduced secondary to a- org-interferon stimulation, was includedas an alternative to CSF pleocytosisbecause it has been shown to bea useful marker of central nervoussystem (CNS) inflammation.9

Encephalopathy was defined as analtered or reduced level ofconsciousness and a change inpersonality or behavior or confusionlasting .24 hours. We did not uselethargy and irritability as sole features

of encephalopathy because of thenonspecific nature of these features insick children unless supported by EEGfeatures of encephalopathy.

Clinical Data During Acute Encephalitis

In total, 179 clinical or investigationvariables were recorded in everypatient (Supplemental Information).CSF results were available in 157(96%) patients, and EEG reports fromthe first week of admission wereavailable in 130 (79%) patients. Weobtained follow-up information bytelephone interview in 144 (88%)cases in 2011 and 2012 by using theLiverpool Outcome Score (LOS)questionnaire,10 supplemented byclinical records. All patients werefollowed up for a minimum of 12months after encephalitis, and the LOSwas recorded as follows: 1 = death,2 = severe, 3 = moderate, 4 = mild, and5 = normal, and specific domains ofabnormal outcome were also recorded.The outcome score describes thepatient’s abilities compared with peers,with mild describing impairmentswithout dependence on carers,whereas severe describes completedependence on carers. Clinical relapseswere recorded separately.

MRI

We had access to 1.5-Tesla MRI brainscans for review in 149 (91%)patients. MRI data were reviewed,blinded to clinical and diagnosticinformation, by pediatric neurologists(R.C.D., S.C.P.) and a pediatricneuroradiologist (K.P.), andconsensus was reached onabnormalities. The mean and medianduration of first MRI scan afteradmission were 4.8 days and 3 days,respectively (range 1–70 days).

Etiologic Investigation to DetermineEncephalitis Causation

Infectious Encephalitis and Infection-Associated Encephalopathy

The CSF, serological, and non-CNSspecimen testing for infectiousetiologies are summarized inSupplemental Table 3. Testing of up

PEDIATRICS Volume 135, number 4, April 2015 e975 by guest on January 23, 2020www.aappublications.org/newsDownloaded from

to 42 different microorganisms wascarried out as clinically indicated. Themean and median etiologic tests(including autoantibodies) perpatient were 14 and 13 tests,respectively (range 0–36).

CSF Testing

A CSF was obtained at a mean andmedian admission duration of 2.9days and 2 days, respectively (range1–48 days). Bacterial and viralcultures of the CSF were done on 155(95%) and 152 (93%) patients,respectively. CSF polymerase chainreaction testing was obtained in 131(80%) of the total cohort, withpatients with ADEM less frequentlytested (63%).

Serology

Serological testing for $1microorganisms was performed on139 (85%) patients. Serologicalstudies for microorganisms usedcomplement fixation tests andimmunoglobulin M (IgM) enzyme-linked immunosorbent assays(Table 1).

Non-CNS Specimens

These tests include blood culture (n =144, 87.8%), nasopharyngeal aspirate(n = 69, 42%), stool and rectal swab(n = 68, 41%), throat swab (n = 47,29%), and skin swab (n = 5, 3%)(Supplemental Table 3).

Immune-Mediated/Autoantibody-Associated Encephalitis

We used Granerod case definition forconfirmed ADEM. Retrospectiveautoantibody testing on stored acutesera was performed in 103 of the 129patients with non-ADEM encephalitis(80%). Patients with a preliminarydiagnosis of ADEM were excludedfrom autoantibody testing (n = 35).Antibodies to NMDAR (n = 103),voltage-gated potassium channelcomplex (VGKC-complex) (n = 102),leucine-rich glioma-inactivated 1(LGI1) (n = 99), contactin-associatedprotein-like 2 (CASPR2) (n = 99),glycine receptor (n = 102), and

glutamic acid decarboxylase (n = 102)were tested as previouslydescribed.11–14 Only 5 patients hadcombined testing of CSF and serumfor autoantibodies (NMDAR only).VGKC-complex antibody positivitywas based on titers .150 pM, assuggested in children.15 Dopamine-2receptor antibody testing (n = 103)was undertaken as described.16

Etiologic Causation

The final etiologic diagnosis wasdetermined by stringent applicationof the consensus guidelines proposedby Granerod et al.1 Cases wereclassified as confirmed, probable, orpossible. Confirmed cases haddetection of the organism orautoantibody in CSF or brain.1 Aprobable diagnosis was based onserological evidence of acute infectionor autoantibody, and a possible

diagnosis was based on detection ofthe organism from a specimen sampleoutside the CNS, such as stool ornasopharyngeal aspirate.1 Asaccepted in the literature, we usedthe term infection-associatedencephalopathy rather thanencephalitis to denote encephalitisrelated to influenza virus orrotavirus.17–19 In cases with multipleetiologies (Supplemental Table 4), theagent with the strongest hierarchicalassociation (confirmed . probable .possible) was designated as theetiology of encephalitis.

Ethics

The study was ethically approved(09/CHW/56), and written consentwas obtained from the families andpatients for a follow-up telephoneinterview and testing of stored acutesamples.

TABLE 1 Hierarchical Classification for the Etiology of Encephalitis

Etiology of Encephalitisa Confirmed Probable Possible Total N (%)

Infectious (N = 49, 30%)Enterovirus 17 — 3 20 (12)Mycoplasma pneumoniae — 11 — 11 (7)HSVb 4 1 4 9 (5)Cytomegalovirus — — 3 3 (2)Other infectionsc — 1 5 6 (4)

Infection-associated encephalopathy (N = 13, 8%)Influenza virus — — 5 5 (3)Rotavirus — — 5 5 (3)ANEd 3 — — 3 (2)

Immune-mediated or autoantibody-associated (N = 56, 34%)e

ADEM 35 — — 35 (21)NMDAR antibody 5 5 — 10 (6)VGKC-complex antibody — 7f — 7 (4)Dopamine D2 receptor antibody — 4 — 4 (2)

Unknown (N = 46, 28%) 46 (28)Total 64 29 25 164 (100)

ANE, acute necrotizing encephalopathy.Table adapted from Granerod et al. (2010).1 Serological testing for $1 microorganisms was performed on 139 (85%)patients. Serological studies for microorganisms used complement fixation tests (CFTs) and IgM enzyme-linked immu-nosorbent assays. Only 14 patients had paired acute and convalescent serology, and no patients had a diagnosis ofinfectious encephalitis based on a fourfold change in titer. A single raised anti-streptolysin O (group A Streptococcus, n =1) and a positive IgM and IgG (HSV [n = 3] and Epstein–Barr virus, parvovirus B19 [each, n = 1]) result were used todiagnose probable or possible infectious encephalitis. A positive IgM provided a probable infectious diagnosis forMycoplasma pneumoniae (n = 11).a 11 patients had $2 etiologic associations and are presented in Supplemental Table 4.b HSV-1 (n = 4, CSF PCR confirmed); HSV-2 (n = 1, skin swab polymerase chain reaction possible); 1 patient with probableHSV (relapsing) also had positive serum NMDAR and D2R antibodies.c Group A Streptococcus (n = 1, probable), adenovirus (n = 1, possible), Epstein–Barr virus (n = 1, possible), para-influenza 3 virus (n = 1, possible), parvovirus B19 (n = 1, possible), varicella zoster virus (n = 1, possible).d Associated infections (probable/possible) in 2 of 3 patients.e 22 patients tested positive for $1 of the following antibodies: NMDAR (n = 11), VGKC (n = 7), D2R (n = 6), and glycinereceptor (n = 1). LGI1, CASPR2, and glutamic acid decarboxylase antibodies were negative in all. One patient with a VGKC-complex antibody titer of 421 pM had positive glycine receptor antibody.f 5 of the 7 VGKC-positive patients had Mycoplasma pneumoniae IgM detected (see Supplemental Table 4).

e976 PILLAI et al by guest on January 23, 2020www.aappublications.org/newsDownloaded from

Statistical Analysis

For comparison of clinical variablesbetween the different major etiologicgroups, we used SAS version 9.3 (SASInstitute, Inc, Cary, NC). Categoricalvariables were compared with theexact x2 test and described in termsof proportions and Wilson confidenceintervals (CIs); continuous variableswere compared with theKruskal–Wallis test and described interms of median and range(Supplemental Tables 5–8). Weassessed the relationship of clinicalvariables to outcome by usingunivariate and multivariate logisticregression analyses. All variables witha univariate P ,.05 were included aspotential predictors in a multivariatemodel that used a stepwise backwardselection procedure in SPSS (IBMSPSS Statistics, IBM Corporation). Wedescribed these associations by usingodds ratios (ORs) and Wald CIs. Weassessed the association betweenclinical variables to autoantibodystatus by using the x2 test anddescribed it in terms of ORs and 95%CIs. We made no adjustment formultiple statistical testing.

RESULTS

All Encephalitis (n = 164)

Overview of Etiologies (Table 1)

A total of 164 children werediagnosed with acute encephalitis. Anetiology was proposed in 118 (72%)patients, of whom 64 (39%) wereconfirmed, 29 (18%) probable, and25 (15%) possible. In 46 (28%)patients, the cause was unknown. Theetiologic groups were infectiousencephalitis, n = 49 (30%); infection-associated encephalopathy, n = 13(8%); and immune-mediated/autoantibody-associated encephalitis,n = 56 (34%). Complete autoantibodyand some infectious serologicalfindings are presented in Table 1.Potential dual or multiple etiologyoccurred in 11 (7%) patients(Supplemental Table 4), including 8with Mycoplasma pneumoniae.

Demographics and Seasonality

The median age at presentation was5.5 years (range 5 weeks–15.1 years),and 94 (57%) patients were male.Figure 1 demonstrates that youngchildren are more frequently affected.The majority of encephalitis cases(n = 57, 35%) presented during theAustralian winter season(June–August, data not shown).

Clinical Features, CSF, EEG, and MRI

Figure 2 and Supplemental Table 5compare the clinical features, whichin descending order of frequencywere fever (n = 126, 77%), seizures(n = 88, 54%), headache (n = 71,43%), weakness or pyramidal signs(n = 61, 37%), any respiratorysymptoms (n = 57, 35%), andagitation (n = 56, 34%).

CSF pleocytosis was present in 110out of 155 (71%) patients andelevated CSF protein in 50 out of 154(33%) patients. CSF neopterin waselevated in 36/47 (77%) testedpatients. Oligoclonal bands weremeasured in 53 (32%), andintrathecal oligoclonal bands were

detected in only 2 patients (1 withNMDAR encephalitis, 1 unknown)and mirrored oligoclonal bands inanother 16 (5 infectious, 10 immune-mediated [ADEM 4, NMDAR 4,dopamine receptor D2 {D2R} 2], 1unknown). Five patients with positiveserum NMDAR antibody also hadCSF testing and were confirmed CSFNMDAR antibody positive. CSFfindings by subgroup are presented inSupplemental Table 6. The EEG wasabnormal in 115 out of 130 (88%)patients; EEG slowing was found in111 out of 130 (85%) andepileptiform discharges in 27 out of130 (21%) patients, respectively.Figure 3 and Supplemental Table 7describe the MRI features. The MRIbrain (T2-weighted fluid-attenuatedinversion recovery) was abnormal in121 out of 149 (81%) of patients.

Duration of Stay, Therapy, and Outcome

The mean and median length of stayfor patients were 27 and 13 days,respectively (range 1–618 days).Admission to the ICU was necessaryin 66 (40%) patients, with themajority needing management of

FIGURE 1Number of patients with acute encephalitis according to age at presentation and etiologic group.11% of patients were ,1 year of age, and 54% of patients were #5 years old, with the frequencydescending steadily at later ages.

PEDIATRICS Volume 135, number 4, April 2015 e977 by guest on January 23, 2020www.aappublications.org/newsDownloaded from

FIGURE 2Clinical features at any stage in all encephalitis and major etiologic subgroups (n $ 7). Data including statistical significance is in Supplemental Table 5.EV, enterovirus; GI, gastrointestinal; MycoP, Mycoplasma pneumoniae; Unk, unknown.

e978 PILLAI et al by guest on January 23, 2020www.aappublications.org/newsDownloaded from

FIGURE 3MRI features according to neuroanatomical location and sequence characteristics in all encephalitis and the major etiologic subgroups (n $ 7). Dataincluding statistical significance are in Supplemental Table 7. EV, enterovirus; FLAIR, fluid-attenuated inversion recovery; MycoP, Mycoplasma pneumoniae;T1-W, T1-weighted; T2-W, T2-weighted; Unk, unknown.

PEDIATRICS Volume 135, number 4, April 2015 e979 by guest on January 23, 2020www.aappublications.org/newsDownloaded from

encephalopathy, seizures, or both. Ofthe 164 patients, 147 (87%) receivedantibiotics or antiviral therapy. Themean and median duration of follow-up were 6.6 years and 5.8 years,respectively (range 1.1–14.4 years).There were 5 deaths (4 in thehospital).

Figure 4 compares the LOS measures.An abnormal outcome (LOS 1–4) waspresent in 71 out of 144 (49%)patients, including LOS 1 (death) (n =5, 4%), LOS 2 (severe) (n = 11, 8%),LOS 3 (moderate) (n = 29, 20%), andLOS 4 (mild) (n = 26, 18%). The mostcommon abnormal outcomes bycategory (Supplemental Table 8)were learning problems (n = 39,28%), behavioral problems (n = 33,24%), epilepsy (n = 25, 18%), andspeech problems (n = 24, 17%).Clinical relapse occurred in 9 (5%)patients: HSV encephalitis (n = 1),ADEM (n = 1), NMDAR-Abencephalitis (n = 3), D2R antibodyencephalitis (n = 2), and unknown(n = 2). Patients with NMDAR andD2R antibodies were more likely torelapse (P = .01).

Using univariate analysis, we foundthat an abnormal long-term outcomewas associated with status epilepticus(seizure.30 minutes) (OR 7.74; 95%CI, 2.51–23.89), ICU admission (OR3.83; 95% CI, 1.91–7.72), diffusionrestriction on MRI (OR 2.47; 95% CI,1.03–5.98), and a movement disorder(OR 2.44; 95% CI, 1.10–5.39).Patients with an infectious prodromewere less likely to have an abnormaloutcome (OR 0.28; 95% CI,0.12–0.63). On multivariate analysis,status epilepticus (OR 3.78; 95% CI,1.07–13.41, P = .04), diffusionrestriction (OR 2.47; 95% CI,0.95–6.46, P = .06), and ICUadmission (OR 2.27; 95% CI,0.92–5.59, P = .08) each predicted anabnormal outcome.

Subgroup Analysis

According to etiology, the followingclinical characteristics were morecommon (Fig 2): Respiratory

symptoms (Mycoplasma, VGKC-complex Ab), gastrointestinalsymptoms (enterovirus), seizures(HSV, VGKC-complex Ab), weaknessor pyramidal signs (ADEM), agitationand psychiatric symptoms(NMDAR), speech dysfunction (NMDAR),movement disorder (NMDAR),cerebellar (ADEM), sleep disturbance(NMDAR), brainstem (ADEM andenterovirus), and autonomic (NMDARand enterovirus).

According to etiology, the followingMRI features were more common(Fig 3): abnormalities of white matter(HSV, ADEM), cortical gray matter(HSV), basal ganglia (Mycoplasma andADEM), thalami (ADEM, HSV),cerebellum and brainstem (ADEM),and limbic encephalitis (NMDAR,VGKC-complex, Mycoplasma). TheMRI was least likely to be abnormal

(any T2-weighted fluid-attenuatedinversion recovery abnormality) inNMDAR encephalitis. Abnormalitiesdetected with diffusion-weightedsequencing, T1 sequences, contrastenhancement, hemorrhage, and cysticnecrosis were all more common inHSV (Fig 3 and SupplementalTable 7). HSV and unknownencephalitis had the worst outcomes,and ADEM had the best outcome(Fig 4).

Of the smaller subgroups (Table 1),all 4 patients with D2Rantibody–associated encephalitis haddystonia–Parkinsonism, and 3 out of4 had localized basal gangliainvolvement on MRI (basal gangliaencephalitis).16 Three patients hadacute necrotizing encephalopathywith the typical radiologicfeatures.18,20 There were no other

FIGURE 4The LOS (1–5) in all encephalitis (n = 144, 88%) and the major etiologic subgroups (n $ 7). Dataincluding statistical significance is in Supplemental Table 8.

e980 PILLAI et al by guest on January 23, 2020www.aappublications.org/newsDownloaded from

characteristic clinicoradiologicfeatures in the other small subgroups.

Clinicoradiologic Correlation WithAutoantibody Positivity

Using univariate analysis in patientswith non-ADEM encephalitis testedfor autoantibody (n = 103), we foundthat an autoantibody-associatedencephalitis was more likely (indescending order) with clinicalrelapse (OR 13.5; 95% CI, 2.46–74.04,P , .0001), movement disorder (OR11.2; 95% CI, 3.82–32.9, P , .0001),CSF intrathecal or mirroredoligoclonal bands (OR 7.89; 95% CI,1.53–40.5, P = .02), speechdysfunction (OR 5.92; 95% CI,2.03–17.2, P , .0001), psychiatricsymptoms (OR 4.79; 95% CI,1.73–13.2, P = .002), agitation (OR3.64; 95% CI, 1.34–9.94, P = .009),and MRI limbic encephalitisphenotype (OR 5.22; 95% CI,1.27–21.5, P = .03). Antibody-associated encephalitis was lesscommon with headache (OR 0.35;95% CI, 0.12–0.98, P = .04) or MRIthalamic involvement (OR 0.21; 95%CI, 0.05–0.98, P = .03).

Immunotherapy and Outcome

A total of 60 out of 164 patients(37%) received immunotherapy(corticosteroids, n = 59; intravenousimmunoglobulin, n = 16; rituximab,n = 2), most commonly for ADEM (27out of 35), NMDAR-Ab encephalitis (6out of 10), enterovirus (6 out of 20),and unknown (9 out of 46)subgroups. Abnormal outcome wasnot different in the patients whoreceived immunotherapy comparedwith those who did not (OR 1.17;95% CI, 0.60–2.31, P = .64). The meanlength of stay was longer in patientswho received immune therapy(41 days vs 19 days, P = .02).

DISCUSSION

This single-center cohort, followed upfor a median of 5.8 years, is the mostcomprehensive retrospective analysisof pediatric encephalitis to date. In

a resource-rich setting, we havedefined the relative frequency ofacute encephalitis syndromes byetiology, with emphasis on theemerging treatable autoantibody-associated encephalitides. Despitelong follow-up, the study confirms thehigh morbidity in survivors ofencephalitis, the significantproportion who lack a diagnosis, theneed to identify the autoantibody-associated syndromes, and the needto improve the acute managementand rehabilitation of these patientswith the hope of reducing long-termresidual neurocognitive impairments.

Immune-mediated/autoantibody-associated encephalitides were themost common etiologic group.Although by definition ADEM is anencephalomyelitis, it is notable thatmany previous encephalitis cohortsdo not include ADEM but generallyfocus on infectious encephalitisalone.6–8,21 NMDAR-Ab encephalitiswas diagnosed in 6% of patients, whohad typical clinical features of thedisease, with dominant movement,psychiatric, sleep, speech, andautonomic features. The numbercould be an underestimate, becauseserum testing was performed on only103 out of 129 (81%) of the patientswithout ADEM. Indeed, 4 of the 46patients with unknown encephalitishad a phenotype reminiscent ofNMDAR-Ab encephalitis but werenegative for serum antibody. CSFmight have been helpful,14,22

although that is not our experience.

VGKC-complex antibodies were firstreported in adults with limbicencephalitis.11 Subsequently, it wasfound that the VGKC-complexantibodies bound not to the VGKCsubunits themselves but to proteins,namely LGI1 and CASPR2,12,23 whichare complexed with the VGKCsubunits in vivo. In this report, 7(4%) of children had VGKC-complexantibodies, but only 2 had evidence oflimbic encephalitis on imaging, andthey were not positive for LGI1 orCASPR2 antibodies, as is often true

for pediatric cases with positiveVGKC-complex Ab.12,23–25 Although150 pM has been proposed asa positive result for VGKC-complex Abin children, levels .400 pM are morerelevant in adults,15,26 and only 2children had these levels.Nevertheless, VGKC-complexantibodies appear to be associatedwith inflammatory disorders(Hacohen et al in preparation). Theassociation seen here of positiveVGKC-complex Ab with Mycoplasmapneumoniae and with respiratorysymptoms (5 out of 7) suggests thatVGKC-complex Ab could be inducedas part of the immune response toMycoplasma pneumoniae. Mycoplasmapneumoniae–associated encephalitisoccurred in 7% of our cohort,although the diagnostic specificityof mycoplasma IgM has beenquestioned.27

Antibodies to D2R werefound in only 4 patients withdystonia–Parkinsonism and basalganglia lesions, representing a smallsubgroup (2%).16 There were nopatients with antibodies to glutamicacid decarboxylase and only 1 patientwith glycine R antibody (who alsohad VGKC-complex antibodies andMycoplasma), and reported pediatriccases with these antibodies are rare.Nine (20%) of the patients withunknown encephalitis did not haveserum available for autoantibodytesting, and comprehensive testingmay help to ensure recognition ofknown autoantibodies and to definesome of the “unknown” patients inthe future. For example,g-aminobutyric acid A receptorantibodies have recently been foundin patients with acute-onset severeepilepsy and encephalitis,28 and itwill be interesting to test pediatricpatients with encephalitis for thisautoantibody.

Infectious encephalitis was found in30%, and the most common infectionwas enterovirus (12%). Even withbroad testing for HSV (84% ofpatients), HSV encephalitis was

PEDIATRICS Volume 135, number 4, April 2015 e981 by guest on January 23, 2020www.aappublications.org/newsDownloaded from

diagnosed in only 5%, confirmingrecent reports that HSV encephalitisis less common in children thanadults with encephalitis.4,29,30

Despite early acyclovir therapy (notdescribed), the outcomes after HSVencephalitis were poor, as describedin recent pediatric cohorts.30,31

Interestingly, 1 patient with HSVencephalitis relapsed with chorea andhad NMDAR and D2R antibodies.32

The emerging theme of postviralautoimmunity is increasinglyrecognized, and it emphasizes theimportance of consideringimmunotherapies in post-HSVrelapses.32–34 There were somepatients with $2 etiologicassociations (Supplemental Table 4),and these patients providea challenge for diagnosis andillustrate the overlap betweeninfectious and autoantibody-associated syndromes.32,35–38

There were notable etiologicabsences in this cohort, such as noMycobacterium tuberculosisencephalitis. Unlike in the UnitedKingdom,5 Mycobacteriumtuberculosis is rare in Australianchildren. We also did not observe anyconfirmed varicella encephalitis,a result probably related partly to theintroduction of routine varicellavaccination during the study period.As is true for most resource-richcountries, we no longer observemeasles or mumps encephalitisbecause of vaccination. Other

etiologies not seen in our cohortinclude arboviruses and Bartonella.

There were small subgroups ofinfection-associated encephalopathy(8%) associated with influenza androtavirus, and these were kept asa separate group, by convention.18,39

Acute necrotizing encephalopathy,which is an infection-associatedencephalopathy, is observed morefrequently in Japanese and EastAsian children, presumably becauseof their differing geneticvulnerability.18

The main limitation of the study wasthe retrospective design andtherefore incomplete investigationand the potential incompleteascertainment of milder or atypicalcases. Despite these limitations, thestudy confirms that encephalitis isa serious disease with an abnormaloutcome in ∼50%, particularly incognitive and behavioral domains,with 3% mortality. In general, thepredictors of poor outcomes were ICUadmission and status epilepticus, asin previous studies.40–42 In addition,we found diffusion restriction on MRI,a marker of cytotoxic injury,correlated with abnormal outcomes.The use of immunotherapy was notassociated with better outcomes, butthe patients receiving immunetherapy had longer admissions,suggesting that these patients hada more complicated course.Immunotherapy has only recentlybecome established in the treatment

of autoantibody-associatedencephalitis (other than ADEM),3 andprospective studies are needed todetermine the benefit ofimmunotherapy in patients withencephalitis irrespective of whetheran autoantibody is detected. Thepercentage of unknown encephalitis(28%) in this study is the lowestamong substantial pediatriccohorts,6–8,40,41 but it represents animportant subgroup that iscomparable in size to the prospectiveUK cohort in adults and children.5

There were no distinguishing clinicalor radiologic characteristics of theunknown subgroup, mainly becauseof the likely heterogenous etiologies.The unknown encephalitis group maycontain both encephalitis of viralorigin (untested or unknown) andautoimmune etiologies. The pooroutcome in this group emphasizes theimportance of future research in thisarea.

ACKNOWLEDGMENTS

We thank the patients, their families,and all health professionals whocontributed to this study. S.C.P. hasfunding from the AustralianPostgraduate Award scheme. Y.H. hasfunding from Oxford UniversityClinical Academic Graduate Schooland Oxford National Institute forHealth Research Biomedical ResearchCentre. R.C.D. and F.B. have fundingfrom NHMRC (Australia) for thiswork.

was involved in clinical data acquisition and analysis; Dr Davies was involved in study design and edited drafts; Dr Prelog performed radiological phenotyping and

edited drafts; Drs Tantsis, Gill, Webster, Menzes, Ardern-Holmes, Gupta, Procopis, Troedson, Antony, Ouvrier, and Polfrit were involved in clinical data acquisition and

analysis and edited drafts; Dr Barnes performed statistical analysis and edited drafts; Dr Vincent supervised autoantibody testing and edited first and subsequent

drafts; Dr Dale conceived and designed the study, supervised all clinical aspects of the study, performed radiological phenotyping, wrote the first draft, and edited

subsequent drafts; and all authors approved the final manuscript as submitted.

www.pediatrics.org/cgi/doi/10.1542/peds.2014-2702

DOI: 10.1542/peds.2014-2702

Accepted for publication Jan 6, 2015

Address correspondence to Russell C. Dale, PhD, Clinical School, Children’s Hospital at Westmead, Locked Bag 4001, Westmead, NSW 2145, Australia. E-mail: russell.

dale@health.nsw.gov.au

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2015 by the American Academy of Pediatrics

e982 PILLAI et al by guest on January 23, 2020www.aappublications.org/newsDownloaded from

FINANCIAL DISCLOSURE: S.C.P., F.B., and R.C.D. are funded by Australian Postgraduate Award, Star Scientific Foundation, Petre Foundation, and National Health and

Medical Research Council. Y.H. has funding from Oxford University Clinical Academic Graduate School and Oxford National Institute for Health Research Biomedical

Research Centre. The other authors have indicated they have no financial relationships relevant to this article to disclose.

FUNDING: Dr Pillai has an Australian Postgraduate Award, and Drs Dale and Brilot have funding from Star Scientific Foundation, Petre Foundation, and National

Health and Medical Research Council. No other authors have funding relevant to this project.

POTENTIAL CONFLICT OF INTEREST: Dr Waters has received speaker honoraria from Euroimmun AG. The other authors have indicated they have no potential conflicts

of interest to disclose.

REFERENCES

1. Granerod J, Cunningham R, ZuckermanM, et al. Causality in acute encephalitis:defining aetiologies. Epidemiol Infect.2010;138(6):783–800

2. Dale RC, Brilot F, Duffy LV, et al. Utility andsafety of rituximab in pediatricautoimmune and inflammatory CNSdisease. Neurology. 2014;83(2):142–150

3. Titulaer MJ, McCracken L, Gabilondo I,et al. Treatment and prognostic factorsfor long-term outcome in patients withanti-NMDA receptor encephalitis: anobservational cohort study. LancetNeurol. 2013;12(2):157–165

4. Gable MS, Sheriff H, Dalmau J, Tilley DH,Glaser CA. The frequency of autoimmuneN-methyl-D-aspartate receptorencephalitis surpasses that of individualviral etiologies in young individualsenrolled in the California EncephalitisProject. Clin Infect Dis. 2012;54(7):899–904

5. Granerod J, Ambrose HE, Davies NWS,et al; UK Health Protection Agency (HPA)Aetiology of Encephalitis Study Group.Causes of encephalitis and differences intheir clinical presentations in England:a multicentre, population-basedprospective study. Lancet Infect Dis.2010;10(12):835–844

6. Koskiniemi M, Korppi M, Mustonen K,et al. Epidemiology of encephalitis inchildren. A prospective multicentrestudy. Eur J Pediatr. 1997;156(7):541–545

7. Fowler A, Stödberg T, Eriksson M,Wickström R. Childhood encephalitis inSweden: etiology, clinical presentationand outcome. Eur J Paediatr Neurol.2008;12(6):484–490

8. Kolski H, Ford-Jones EL, Richardson S,et al. Etiology of acute childhoodencephalitis at The Hospital for SickChildren, Toronto, 1994–1995. Clin InfectDis. 1998;26(2):398–409

9. Dale RC, Brilot F, Fagan E, Earl J.Cerebrospinal fluid neopterin in

paediatric neurology: a marker of activecentral nervous system inflammation.Dev Med Child Neurol. 2009;51(4):317–323

10. Lewthwaite P, Begum A, Ooi MH, et al.Disability after encephalitis:development and validation of a newoutcome score. Bull World Health Organ.2010;88(8):584–592

11. Vincent A, Buckley C, Schott JM, et al.Potassium channel antibody-associatedencephalopathy: a potentiallyimmunotherapy-responsive form oflimbic encephalitis. Brain. 2004;127(3):701–712

12. Irani SR, Alexander S, Waters P, et al.Antibodies to Kv1 potassiumchannel–complex proteins leucine-rich,glioma inactivated 1 protein andcontactin-associated protein-2 in limbicencephalitis, Morvan’s syndrome andacquired neuromyotonia. Brain. 2010;133(9):2734–2748

13. Hutchinson M, Waters P, McHugh J, et al.Progressive encephalomyelitis, rigidity,and myoclonus: a novel glycine receptorantibody. Neurology. 2008;71(16):1291–1292

14. Dalmau J, Lancaster E, Martinez-Hernandez E, Rosenfeld MR, Balice-Gordon R. Clinical experience andlaboratory investigations in patients withanti-NMDAR encephalitis. Lancet Neurol.2011;10(1):63–74

15. Haberlandt E, Bast T, Ebner A, et al.Limbic encephalitis in children andadolescents. Arch Dis Child. 2011;96(2):186–191

16. Dale RC, Merheb V, Pillai S, et al.Antibodies to surface dopamine-2receptor in autoimmune movement andpsychiatric disorders. Brain. 2012;135(pt11):3453–3468

17. Togashi T, Matsuzono Y, Narita M,Morishima T. Influenza-associated acuteencephalopathy in Japanese children in

1994–2002. Virus Res. 2004;103(1–2):75–78

18. Mizuguchi M, Yamanouchi H, Ichiyama T,Shiomi M. Acute encephalopathyassociated with influenza and other viralinfections. Acta Neurol Scand. 2007;115(4suppl):45–56

19. Lynch M, Lee B, Azimi P, et al. Rotavirusand central nervous system symptoms:cause or contaminant? Case reports andreview. Clin Infect Dis. 2001;33(7):932–938

20. Mizuguchi M. Acute necrotizingencephalopathy of childhood: a novelform of acute encephalopathy prevalentin Japan and Taiwan. Brain Dev. 1997;19(2):81–92

21. Glaser CA, Gilliam S, Schnurr D, et al;California Encephalitis Project,1998–2000. In search of encephalitisetiologies: diagnostic challenges in theCalifornia Encephalitis Project,1998–2000. Clin Infect Dis. 2003;36(6):731–742

22. Gresa-Arribas N, Titulaer MJ, Torrents A,et al. Antibody titres at diagnosis andduring follow-up of anti-NMDA receptorencephalitis: a retrospective study.Lancet Neurol. 2014;13(2):167–177

23. Lai M, Huijbers MGM, Lancaster E, et al.Investigation of LGI1 as the antigen inlimbic encephalitis previously attributedto potassium channels: a case series.Lancet Neurol. 2010;9(8):776–785

24. Hacohen Y, Wright S, Waters P, et al.Paediatric autoimmuneencephalopathies: clinical features,laboratory investigations and outcomesin patients with or without antibodies toknown central nervous systemautoantigens. J Neurol NeurosurgPsychiatry. 2013;84(7):748–755

25. Suleiman J, Brenner T, Gill D, et al. VGKCantibodies in pediatric encephalitispresenting with status epilepticus.Neurology. 2011;76(14):1252–1255

PEDIATRICS Volume 135, number 4, April 2015 e983 by guest on January 23, 2020www.aappublications.org/newsDownloaded from

26. Paterson RW, Zandi MS, Armstrong R,Vincent A, Schott JM. Clinical relevanceof positive voltage-gated potassiumchannel (VGKC)-complex antibodies:experience from a tertiary referralcentre. J Neurol Neurosurg Psychiatry.2014;85(6):625–630

27. Christie LJ, Honarmand S, Talkington DF,et al. Pediatric encephalitis: what is therole of Mycoplasma pneumoniae?Pediatrics. 2007;120(2):305–313

28. Petit-Pedrol M, Armangue T, Peng X, et al.Encephalitis with refractory seizures,status epilepticus, and antibodies to theGABAA receptor: a case series,characterisation of the antigen, andanalysis of the effects of antibodies.Lancet Neurol. 2014;13(3):276–286

29. Mailles A, Stahl JP; Steering Committeeand Investigators Group. Infectiousencephalitis in France in 2007: a nationalprospective study. Clin Infect Dis. 2009;49(12):1838–1847

30. Elbers JM, Bitnun A, Richardson SE, et al.A 12-year prospective study of childhoodherpes simplex encephalitis: is therea broader spectrum of disease?Pediatrics. 2007;119(2). Available at:www.pediatrics.org/cgi/content/full/119/2/e399

31. Schleede L, Bueter W, Baumgartner-SiglS, et al. Pediatric herpes simplex virus

encephalitis: a retrospective multicenterexperience. J Child Neurol. 2013;28(3):321–331

32. Mohammad SS, Sinclair K, Pillai S, et al.Herpes simplex encephalitis relapse withchorea is associated with autoantibodiesto N-methyl-D-aspartate receptor ordopamine-2 receptor. Mov Disord. 2014;29(1):117–122

33. Armangue T, Leypoldt F, Málaga I, et al.Herpes simplex virus encephalitis isa trigger of brain autoimmunity. AnnNeurol. 2014;75(2):317–323

34. Hacohen Y, Deiva K, Pettingill P, et al. N-methyl-D-aspartate receptor antibodiesin post–herpes simplex virusencephalitis neurological relapse. MovDisord. 2014;29(1):90–96

35. Prüss H, Finke C, Höltje M, et al. N-methyl-D-aspartate receptor antibodies inherpes simplex encephalitis. Ann Neurol.2012;72(6):902–911

36. Titulaer MJ, Höftberger R, Iizuka T, et al.Overlapping demyelinating syndromesand anti–N-methyl-D-aspartate receptorencephalitis. Ann Neurol. 2014;75(3):411–428

37. Hacohen Y, Absoud M, Woodhall M, et al;UK & Ireland Childhood CNSInflammatory Demyelination WorkingGroup. Autoantibody biomarkers in

childhood-acquired demyelinatingsyndromes: results from a nationalsurveillance cohort. J Neurol NeurosurgPsychiatry. 2014;85(4):456–461

38. Tsiodras S, Kelesidis I, Kelesidis T,Stamboulis E, Giamarellou H. Centralnervous system manifestations ofMycoplasma pneumoniae infections.J Infect. 2005;51(5):343–354

39. Amin R, Ford-Jones E, Richardson SE,et al. Acute childhood encephalitis andencephalopathy associated withinfluenza: a prospective 11-yearreview. Pediatr Infect Dis J. 2008;27(5):390–395

40. DuBray K, Anglemyer A, LaBeaud AD,et al. Epidemiology, outcomes andpredictors of recovery in childhoodencephalitis: a hospital-based study.Pediatr Infect Dis J. 2013;32(8):839–844

41. Wang IJ, Lee PI, Huang LM, Chen CJ, ChenCL, Lee WT. The correlation betweenneurological evaluations andneurological outcome in acuteencephalitis: a hospital-based study.Eur J Paediatr Neurol. 2007;11(2):63–69

42. Fowler A, Stödberg T, Eriksson M,Wickström R. Long-term outcomes ofacute encephalitis in childhood.Pediatrics. 2010;126(4). Available at:www.pediatrics.org/cgi/content/full/126/4/e828

e984 PILLAI et al by guest on January 23, 2020www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2014-2702 originally published online March 23, 2015; 2015;135;e974Pediatrics 

Ming J. Lim, Fabienne Brilot, Angela Vincent and Russell C. DaleLang,Robert A. Ouvrier, Yann Polfrit, Nicholas W. S. Davies, Patrick Waters, Bethan

Ardern-Holmes, Sachin Gupta, Peter Procopis, Christopher Troedson, Jayne Antony,Kesson, Elizabeth Barnes, Deepak Gill, Richard Webster, Manoj Menezes, Simone

Sekhar C. Pillai, Yael Hacohen, Esther Tantsis, Kristina Prelog, Vera Merheb, AlisonLong-term Outcome

Infectious and Autoantibody-Associated Encephalitis: Clinical Features and

ServicesUpdated Information &

http://pediatrics.aappublications.org/content/135/4/e974including high resolution figures, can be found at:

Referenceshttp://pediatrics.aappublications.org/content/135/4/e974#BIBLThis article cites 41 articles, 9 of which you can access for free at:

Subspecialty Collections

subhttp://www.aappublications.org/cgi/collection/neurologic_disorders_Neurologic Disordershttp://www.aappublications.org/cgi/collection/neurology_subNeurologyfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://www.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

by guest on January 23, 2020www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2014-2702 originally published online March 23, 2015; 2015;135;e974Pediatrics 

Ming J. Lim, Fabienne Brilot, Angela Vincent and Russell C. DaleLang,Robert A. Ouvrier, Yann Polfrit, Nicholas W. S. Davies, Patrick Waters, Bethan

Ardern-Holmes, Sachin Gupta, Peter Procopis, Christopher Troedson, Jayne Antony,Kesson, Elizabeth Barnes, Deepak Gill, Richard Webster, Manoj Menezes, Simone

Sekhar C. Pillai, Yael Hacohen, Esther Tantsis, Kristina Prelog, Vera Merheb, AlisonLong-term Outcome

Infectious and Autoantibody-Associated Encephalitis: Clinical Features and

http://pediatrics.aappublications.org/content/135/4/e974located on the World Wide Web at:

The online version of this article, along with updated information and services, is

http://pediatrics.aappublications.org/content/suppl/2015/03/17/peds.2014-2702.DCSupplementalData Supplement at:

ISSN: 1073-0397. 60007. Copyright © 2015 by the American Academy of Pediatrics. All rights reserved. Print the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on January 23, 2020www.aappublications.org/newsDownloaded from