Published Ahead of Print on May 4, 2016 as 10.1212/WNL.

0000000000002723

Acute disseminated encephalomyelitis in

228 patients
A retrospective, multicenter US study

Diederik L.H. Koelman, ABSTRACT

BSc Objective: To analyze the range of demographic, clinical, MRI, and CSF features of acute dissem-
Salim Chahin, MD, inated encephalomyelitis (ADEM), a rare, typically monophasic demyelinating disorder, and ana-
MSCE lyze long-term outcomes including time and risk factors for subsequent clinical events as well
Soe S. Mar, MD as competing diagnoses.
Arun Venkatesan, MD,
Methods: We performed a retrospective, multicenter study in 4 US academic medical centers of
PhD
all patients clinically diagnosed with ADEM. Initial presentation of pediatric and adult ADEM and
George M. Hoganson,
monophasic and multiphasic disease were compared. The Aalen-Johansen estimator was used to
MD
produce estimates of the probability of transitioning to a multiphasic diagnosis as a function of
Anusha K. Yeshokumar,
time since initial diagnosis, treating death and alternative diagnoses as competing risks.
MD
Paula Barreras, MD Results: Of 228 patients (122 children, age range 172 years, 106 male, median follow-up 24
Bittu Majmudar, MD months [25th75th percentile 667], 7 deaths), approximately one quarter (n 5 55, 24%) expe-
Joshua P. Klein, MD, rienced at least one relapse. Relapsing disease in children was more often diagnosed as multi-
PhD phasic ADEM than in adults (58% vs 21%, p 5 0.007), in whom MS was diagnosed more often.
Tanuja Chitnis, MD Encephalopathy at initial presentation (hazard ratio [HR] 0.383, p 5 0.001), male sex (HR 0.394,
David C. Benkeser, PhD p 5 0.002), and increasing age at onset (HR 0.984, p 5 0.035) were independently associated
Marco Carone, PhD with a longer time to a demyelinating disease relapse in a multivariable model. In 17 patients,
Farrah J. Mateen, MD, diagnoses other than demyelinating disease were concluded in long-term follow-up.
PhD Conclusions: Relapsing disease after ADEM is fairly common and associated with a few poten-
tially predictive features at initial presentation. Age-specific guidelines for ADEM diagnosis
and treatment may be valuable, and vigilance for other, mostly rare, diseases is imperative.
Correspondence to Neurology 2016;86:19
Dr. Mateen:
fmateen@partners.org
GLOSSARY
ADEM 5 acute disseminated encephalomyelitis; AQP4 5 aquaporin-4; CI 5 confidence interval; DMT 5 disease-modifying
therapy; HR 5 hazard ratio; ICD-9 5 International Classification of Diseases9; IPMSSG 5 International Pediatric MS Study
Group; IVIg 5 IV immunoglobulin G; MOG 5 myelin oligodendrocyte glycoprotein; mRS 5 modified Rankin Scale; MS 5
multiple sclerosis; NMOSD 5 neuromyelitis optica spectrum disorder; OR 5 odds ratio; PLEX 5 plasma exchange.

Acute disseminated encephalomyelitis (ADEM) is a rare inflammatory demyelinating disor-

der of the CNS, first described in 1724 in a patient after smallpox.1 A pathologic correlate
has been identified; nonetheless, the diagnosis remains clinical due to the lack of a distinctive
biological marker. The disease typically follows a self-limiting monophasic course, but mul-
tiphasic forms have been reported.2 Multiple sclerosis (MS) may be indistinguishable from
ADEM at first presentation, even if the diagnosis of ADEM is confined to patients
who present with encephalopathy. Accurate diagnosis at disease onset holds therapeutic
implications.
None of the 9 large (n $ 50) published cohorts of patients with an initial ADEM diagnosis is
United Statesbased.211 Most of these large cohorts have focused on the first demyelinating
episode in pediatric patients. Smaller studies have described ADEM in the United States, or have

From the Department of Neurology (D.L.H.K., T.C., F.J.M.), Massachusetts General Hospital, Boston; Department of Neurology (D.L.H.K.),
Academic Medical Center, Amsterdam, the Netherlands; Department of Neurology (S.C.), University of Pennsylvania, Philadelphia; Department of
Neurology (S.S.M., G.M.H., B.M.), Washington University School of Medicine, St. Louis, MO; Department of Neurology (A.V., A.K.Y., P.B.),
Johns Hopkins University School of Medicine, Baltimore, MD; Department of Neurology (J.P.K.), Brigham and Womens Hospital, Boston;
Harvard Medical School (J.P.K., F.J.M.), Boston, MA; and Department of Biostatistics (D.C.B., M.C.), University of Washington, Seattle.
Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

2016 American Academy of Neurology 1

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 focused on adult-onset ADEM.12,13 The clin- Definitions. Encephalopathy included altered consciousness or
altered behavior, confusion, or irritability, as proposed by the
ical and MRI criteria proposed to differentiate
IPMSSG.14 Severe fatigue and postictal confusion were not con-
ADEM from the likely first presentation of sidered to be encephalopathy. Presentation was polyfocal if clin-
MS do not provide a fully reliable prediction ical symptoms originated from different locations of the CNS.
algorithm, and not all have been tested in The emergence of either new or old neurologic symptoms or
lesions on MRI was defined as a flare if it occurred ,3 months
adult-onset ADEM.1416
after disease onset, and as a relapse if it occurred .3 months after
In this retrospective, multicenter study, we disease onset. A flare alone was not considered as multiphasic
describe the largest collected cohort of ADEM ADEM. A second relapse occurred .3 months after the first
to date. We compare the clinical characteris- relapse. Time of follow-up was defined as time from
presentation to last available follow-up. In case of death, date of
tics, MRI findings, CSF features, and long- death was defined as the last follow-up. Favorable outcome was
term outcomes including disease recurrence, designated as patients with an mRS score #2. Final diagnosis was
alternative diagnoses, and functional outcome left up to the discretion of the treating neurologist.
of both children and adults with ADEM. Standard protocol approvals, registrations, and patient
Finally, we assess prognostic factors for disease consents. The institutional ethics committee at each center
approved the study.
relapse and the clinical utility of the Interna-
tional Pediatric MS Study Group (IPMSSG) Statistical analysis. Categorical variables were compared using
Fisher exact test while continuous variables were compared using
criteria.14
the Mann-Whitney test. Initial presentation of pediatric and
adult ADEM and monophasic and multiphasic disease were
METHODS Study design and patient selection. We compared. Onset frequency by season was assessed using the
searched the hospitals billing systems for inpatient and outpa- Pearson x2 statistic. Treating death and alternative diagnosis as
tient medical records using ICD-9 codes for infectious ADEM competing risk, the Aalen-Johansen estimator was used to
(323.61) and noninfectious ADEM (323.81). Medical records of produce estimates of the probability of transitioning to
patients seen during first presentation or follow-up at the a multiphasic diagnosis as a function of time since initial
Massachusetts General Hospital, Boston, from January 2000 to diagnosis, and the Gray test was used to determine whether this
December 2014; St. Louis Childrens Hospital, Missouri, from function differed significantly between 2 subgroups.20,21 A
September 1985 to September 2014; Barnes-Jewish Hospital, multivariable cause-specific proportional hazards model was
St. Louis, from September 2007 to September 2014; the Johns used to assess risk factors for transition to multiphasic
Hopkins Hospital, Baltimore, Maryland, from January 1997 to ADEM.22 A multivariable logistic regression model was
January 2011; and Hospital of the University of Pennsylvania, constructed to identify risk factors for unfavorable outcome at
Philadelphia, from September 2000 to September 2014 last follow-up, adjusting for length of follow-up. Cause-specific
were reviewed by an author from each study site (D.L.H.K., hazard ratio (HR) and odds ratio (OR) estimates are reported
G.M.H., B.M., A.K.Y., P.B., and S.C.). We included patients along with corresponding 95% confidence intervals (95% CI).
with a clinical event in whom ADEM was diagnosed, based on All tests of hypotheses were 2-sided and performed at significance
the discretion of the treating neurologist. Patients without level a 5 0.05. Statistical analyses were implemented with the R
encephalopathy at clinical onset, but with an otherwise typical programming language.
presentation of ADEM, were included. Patients with both
a normal brain and spinal cord MRI, an initial presentation
RESULTS Characteristics at initial presentation. There
fitting criteria for aquaporin-4 (AQP4)positive neuromyelitis
optica spectrum disorder (NMOSD),17 or a change in diagnosis were 228 patients (122 children, 106 male, age 172
based on the first clinical presentationrather than on a flare, years) initially diagnosed with ADEM (table 1). The
a relapse, or deathwere excluded. treating neurologists diagnosis in patients with suffi-
Data collection. We systematically collected details of the first
cient information on initial presentation was in accor-
clinical event: sex, date of birth, preceding infection or vaccina- dance with IPMSSG criteria in 83 pediatric patients
tion, date of symptomatic presentation, medical history of neuro- (70%) but only 49 adult patients (47%). Absence of
logic concerns and other serious medical illnesses, clinical encephalopathy was the leading reason for patients
symptoms, CSF results, serum results, MRI findings adjudicated not to meet IPMSSG criteria (87 [96%]), followed
by radiology report, and treatment for the initial event. Addi-
by nonpolyfocal onset (6 [7%]) or normal brain MRI
tional characteristics were recorded if relevant. We assumed med-
ical assessments to be complete and clinical features normal if not
(3 [3%]).
explicitly reported as being remarkable. Spinal cord MRI, CSF, There were 26 patients (20 adults) with neurologic
and serum tests were considered missing or not done if not re- and 42 patients (30 adults) with non-neurologic ill-
ported. Subsequent events, including dates and alterations of nesses in their medical history. A majority of patients
diagnosis, were tracked. Final diagnosis and outcome were as- had an infection or vaccination ,4 weeks prior to
sessed at last follow-up. The treating neurologist rendered the
ADEM presentation (table 2). Season did not change
final diagnosis. An author at each study site retrospectively
onset frequency.
ascribed the modified Rankin Scale (mRS) scores based on the
documented examination if an mRS score was not already Imaging of the spinal cord was available in 179 pa-
noted in the medical record.18,19 Last follow-up was determined tients (79%), including the 3 patients who did not
by last visit or telephone encounter. have an abnormal brain MRI. In 2 patients, a lumbar

2 Neurology 86 May 31, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Table 1 Clinical characteristics of patients initially diagnosed with acute disseminated encephalomyelitis

Features All Pediatric Adult p Value

a
No. patients 228 122 106

Demographics

Age, y, median (25th75th percentile) 17 (736) 8 (512) 37 (2748)

Male, n (%) 106 (46) 61 (50) 45 (42) NS

Clinical presentation, n (%)

Preceding event 131 (58) 77 (63) 54 (52) NS

Polyfocal onset 220 (97) 120 (98) 100 (96) NS

Encephalopathy 137 (61) 85 (70) 52 (50) 0.003

Headache 120 (53) 70 (57) 50 (48) NS

Nausea/vomiting 86 (38) 52 (43) 34 (33) NS

Fever 90 (40) 68 (56) 22 (21) ,0.001

Seizures 36 (16) 21 (17) 15 (14) NS

Weakness 115 (51) 61 (50) 54 (52) NS

Ataxia 80 (35) 48 (39) 32 (31) NS

Gait abnormality 126 (56) 62 (51) 64 (62) NS

Optic neuritis 24 (11) 19 (16) 5 (5) 0.009

Other visual disturbances 52 (23) 26 (21) 26 (25) NS

Other cranial nerve palsies 82 (36) 43 (35) 39 (38) NS

Sensory abnormalities 60 (27) 17 (14) 43 (41) ,0.001

Meningismus 24 (11) 16 (13) 8 (8) NS

MRI findings, n (%)

Brain abnormal 225 (99) 121 (99) 104 (98) NS

Spinal cord abnormal 66 (37) 31 (36) 35 (38) NS

Infratentorial lesions 139 (63) 70 (59) 69 (66) NS

Brainstem involvement 116 (52) 62 (53) 54 (52) NS

Cerebellar involvement 90 (41) 40 (34) 50 (48) 0.040

Deep gray matter involvement 86 (39) 49 (42) 37 (36) NS

Periventricular involvement 86 (39) 32 (27) 54 (52) ,0.001

Other supratentorial lesions 184 (83) 104 (88) 80 (77) 0.032

Corpus callosum involvement 57 (26) 14 (12) 43 (41) ,0.001

CSF features, n (%)

Lumbar puncture performed 213 (93) 111 (91) 102 (96) NS

Inflammation 168 (81) 82 (77) 86 (86) NS

Pleocytosis (>5 cells/mL) 147 (71) 73 (68) 74 (74) NS

Elevated protein (>45 mg/dL) 97 (48) 36 (34) 61 (62) ,0.001

Oligoclonal bands present 35 (23) 11 (16) 24 (29) NS

Abbreviation: NS 5 not significant (p . 0.05).

Percentages for the MRI findings and CSF features are based on the number of available data.
a
Data on clinical presentation were missing for 2 adults.

puncture was not performed due to the concern for patients only in serum. All were seronegative during
cerebral herniation. CSF cell count and protein were the first clinical event. No patient was checked for mye-
available in 207 and 203 patients with a median of lin oligodendrocyte glycoprotein (MOG) antibody.
18 cells/mL (25th75th percentile 456) and 45 mg/dL
(25th75th percentile 2965), respectively. AQP4 anti- Follow-up. Follow-up was a median of 24 months
body was checked in 51 patients (22%): 3 patients both (25th75th percentile 667, range 0277, mean
in serum and CSF, 1 patient only in CSF, and 47 46) and longer in children compared to adults (30

Neurology 86 May 31, 2016 3

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 biopsy confirmation. Patients diagnosed with multi-
Table 2 Preceding systemic infections and
vaccinations
phasic ADEM presented more often with fever (50%
vs 13%, p 5 0.010) and encephalopathy (64% vs
Reported event No. (%) 38%, not significant); less often included sensory
Minor respiratory tract infection 62 (27) abnormalities (5% vs 42%, p 5 0.005); and more
Nonspecific infection 34 (15) often had a favorable outcome (96% vs 79%, not
significant) compared to patients diagnosed with
Infection with detected pathogen 18 (8)
MS. No patient with MS presented with meningeal
Gastrointestinal infection 11 (5)
signs. Three patients tested positive for AQP4 anti-
Pneumonia 6 (3)
body during follow-up, 2 of whom tested negative at
Ear, eye, or dental infection 5 (2) initial presentation. One patient was diagnosed with
Urinary tract infection 2 (1) NMOSD associated with Sjgren syndrome. The
Vaccinationa 10 (4) treating neurologist diagnosed 4 other patients with
NMOSD because of their clinical characteristics dur-
Preceding events occurred less than 4 weeks prior to initial
presentation.
ing the disease course.
a
Seven patients with a reported preceding vaccination In 17 patients (7 children), a final diagnosis out-
experienced a concomitant infection. side the demyelinating spectrum was made based on
a flare (n 5 3), a relapse (n 5 11), or death (n 5
[25th75th percentile 989] vs 21 [25th75th 3). Diagnosis was revised a median of 8 months
percentile 452] months, p 5 0.009). A monophasic (25th75th percentile 320) and as long afterwards
disease course, without any diagnostic revisions, as 75 months after the initial presentation. The diag-
occurred in 156 patients (84 male, 72 adults). Mono- noses of glioblastoma multiforme, astrocytoma, and
phasic ADEM was confirmed by histopathology in 2 CNS vasculitis were confirmed by brain biopsy.
adults. Differences between adult- and pediatric- The figure shows the cumulative incidence of
onset monophasic ADEM were identified (table 3). relapsing demyelinating disease after the initial diag-
At last follow-up, the diagnosis was revised in 72 nosis with stratification by several covariates. Most
patients (table 4). patients who experienced a relapse did so within the
The 55 patients (31 pediatric) who experienced first 2 years after initial presentation (85%). The pres-
a relapsing demyelinating disease (multiphasic ence of encephalopathy at initial presentation (HR
ADEM, MS, NMOSD) were followed longer than 0.383, 95% CI 0.2190.669, p 5 0.001), male sex
patients observed to remain monophasic (65 [25th (HR 0.394, 95% CI 0.2210.702, p 5 0.002), and
75th percentile 34127] vs 19 [25th75th percentile a higher age at onset (HR 0.984, 95% CI 0.969
346] months, p , 0.001). A competing demyelin- 0.999, p 5 0.035) were associated with a longer time
ating disease was diagnosed a median of 12 months to a demyelinating disease relapse in a multivariable
(25th75th percentile 923) after initial presenta- model.
tion. Four patients were diagnosed with MS based
on a flare. Patients who experienced a flare considered Treatment and functional outcome. All patients were
to be part of the initial ADEM event (n 5 42) were treated at the discretion of the treating neurologist.
significantly more likely to experience a subsequent Data on treatment for the first clinical event were
relapse more than 3 months after onset (64% vs 29%, available for 219 patients (96%). The reason not
p , 0.001). Patients with future relapsing demyelin- to treat (n 5 26) was documented for 8 patients,
ating disease were less often male (31% vs 52%, p 5 all of whom showed clinical improvement. In the
0.026) and presented more often with ataxia (50% vs other 193 patients, 188 patients were treated with
27%, p 5 0.013) and less often with encephalopathy steroids (185 IV), 37 patients with IV immuno-
(48% vs 73%, p 5 0.008), fever (28% vs 51%, p 5 globulin G (IVIg), and 17 patients with plasma
0.014), and meningeal signs (2% vs 17%, p 5 0.006) exchange (PLEX). IVIg was a first-line treatment
compared to patients who remained monophasic dur- in 5 patients; PLEX treatment was always preceded
ing a minimum follow-up time of 2 years (n 5 64). by treatment with steroids (n 5 8), IVIg (n 5 1), or
Rate of relapse was not significantly lower in patients both (n 5 8). MS disease-modifying therapy
who met the IPMSSG criteria (20% vs 33%) or when (DMT) for the first clinical event was prescribed
only adults (16% vs 32%) or children (23% vs 34%) for short-term use in 2 adults: 1 was subsequently
were included. diagnosed with AQP4 antibody-negative NMOSD
Pediatric patients with a relapsing demyelinating and the other remained monophasic more than 10
disease were more often diagnosed with multiphasic years later. Treatment with PLEX for the first
ADEM than adult patients (58% vs 21%, p 5 clinical event was more common in adults (13%
0.007). One child with multiphasic ADEM had vs 3%, p 5 0.012).

4 Neurology 86 May 31, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Table 3 Pediatric and adult patients with a final diagnosis of monophasic acute disseminated
encephalomyelitis

Feature All Children Adults p Value

a
No. patients 156 84 72

Demographics

Age at onset, y, median (25th75th percentile) 17 (736) 8 (513) 37 (2749)

Male, n (%) 83 (53) 49 (58) 34 (47) NS

Initial presentation,a n (%)

Preceding event 100 (65) 59 (70) 41 (58) NS

Polyfocal onset 150 (97) 82 (98) 68 (96) NS

Encephalopathy 102 (66) 62 (74) 40 (56) 0.027

Fever 69 (45) 51 (61) 18 (25) ,0.001

Headache 84 (54) 51 (61) 33 (46) NS

Nausea/vomiting 61 (39) 41 (49) 20 (28) 0.013

Seizures 22 (14) 13 (15) 9 (13) NS

Weakness 73 (47) 38 (45) 35 (49) NS

Ataxia 50 (32) 31 (37) 19 (27) NS

Gait abnormality 90 (58) 46 (55) 44 (62) NS

Optic neuritis 13 (8) 10 (12) 3 (4) NS

Visual problems 39 (25) 20 (24) 19 (27) NS

Other cranial nerve palsies 56 (36) 30 (36) 26 (37) NS

Sensory abnormalities 42 (27) 12 (14) 30 (42) ,0.001

Meningeal signs 22 (14) 15 (18) 7 (10) NS

MRI findings, n (%)

Spinal cord lesions 44 (37) 23 (40) 21 (34) NS

Brainstem lesions 81 (54) 44 (54) 37 (53) NS

Cerebellum/peduncle lesions 63 (42) 28 (35) 35 (50) NS

Infratentorial lesions 93 (62) 48 (59) 45 (64) NS

Deep gray matter lesions 64 (42) 38 (47) 26 (37) NS

Periventricular lesions 59 (39) 23 (28) 36 (51) 0.005

Other supratentorial lesions 128 (85) 76 (94) 52 (74) 0.001

Corpus callosum involvement 36 (24) 8 (10) 28 (40) ,0.001

CSF features

Inflammation, n (%) 121 (83) 63 (81) 58 (87) NS

Pleocytosis (5 cells/mL), n (%) 103 (71) 56 (72) 47 (70) NS

Elevated protein (>45 mg/dL), n (%) 75 (53) 29 (38) 46 (71) ,0.001

Cell count (cells/mL), median (25th75th percentile) 20 (462) 22 (454) 19 (475) NS

Protein (mg/dL), median (25th75th percentile) 46 (3065) 37 (2453) 54 (4484) ,0.001

Oligoclonal bands present, n (%) 24 (24) 8 (18) 16 (28) NS

Follow-up

Follow-up, mo, median (25th75th percentile) 19 (346) 20 (457) 17 (235) NS

Favorable outcome (mRS 02), n (%) 127 (84) 80 (95) 47 (70) ,0.001

Abbreviations: mRS 5 modified Rankin Scale score; NS 5 not significant (p . 0.05).

Percentages for the MRI findings and CSF features are based on the number of available data.
a
Data on clinical presentation were missing for one adult.

Seven patients were deceased at last known disease course (1 acute hemorrhagic leukoencephalitis,
follow-up. Three patients with ADEM died: 1 with 1 nonhemorrhagic ADEM). Other final diagnoses were
concomitant CNS lymphoma and 2 with a fulminant unspecified primary brain tumor, primary CNS

Neurology 86 May 31, 2016 5

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Relapsing disease after an initial ADEM diagnosis
Table 4 Final diagnosis of patients initially
diagnosed with ADEM
is reported in 10%31% of patients in prior co-
horts.2,6,9,10 In our study, of confirmed relapsing pa-
Final diagnoses No. (%) tients, the first relapse was reported within 2 years
Monophasic ADEM a
156 (68) after first presentation in 85% of patients, confirming
MS 24 (11) this cutoff, used in prior studies, as practical. In a large
natural history study of MS, the first interattack inter-
Multiphasic ADEM 23 (10)
val in patients was a median of 2 years.23 Yet 1 in 10
NMOSD 8 (4)
patients diagnosed with ADEM who remained
Susac syndrome 2 (1)
monophasic after 2 years follow-up subsequently
Lupus cerebritis 2 (1) experienced relapsing disease. Moreover, revision of
CNS lymphoma 2 (1) the diagnosis to a disease outside of the demyelinating
Astrocytoma grade 3 anaplastic 1 spectrum or to a relapsing disease occurred .5 and
Primary brain tumor 1
.10 years after initial presentation, respectively.
Therefore, a subgroup of patients who remained
CNS Lyme disease 1
monophasic during follow-up may progress to multi-
CLIPPERS 1
phasic disease in the future. Taken together, these
Mitochondrial disorder involving the CNS 1 findings highlight the importance of extended
Glioblastoma multiforme 1 follow-up and clinical vigilance for alternate diagnoses
PANS/autoimmune encephalitis 1 and subsequent relapsing diseases.
CNS vasculitis 1
Features helpful in predicting a subsequent relaps-
ing demyelinating disease included female sex and
Rabies myeloencephalitis 1
absence of encephalopathy at initial presentation.
Chronic relapsing inflammatory optic neuritis 1
Fever, ataxia, meningeal signs, and deep gray matter
Recurrent encephalitis 1
lesions on neuroimaging are also potentially useful
Abbreviations: ADEM 5 acute disseminated encephalomy- in the early prediction of relapsing disease.7,16,24 In
elitis; CLIPPERS 5 chronic lymphocytic inflammation with contrast with most prior studies, we did not find
pontine perivascular enhancement responsive to steroids; the presence of oligoclonal bands, presence of peri-
MS 5 multiple sclerosis; NMOSD 5 neuromyelitis optica
spectrum disorder; PANS 5 pediatric acute-onset neuro-
ventricular lesions, or corpus callosum involvement
psychiatric syndrome. on neuroimaging to be significantly associated with
a
Four patients with monophasic ADEM were diagnosed relapsing disease.7,15,16,2429 This can probably be ex-
with acute hemorrhagic leukoencephalitis.
plained by the selection of cases; patients with clinical
or diagnostic features well-known to be associated
lymphoma, astrocytoma grade 3, and rabies myeloence- with MS are less likely to be diagnosed with ADEM
phalitis. Only this last patient had a brain autopsy. initially.
Patients who required additional therapeutic man- In patients who remained monophasic, sensory
agement with PLEX or IVIg had a significantly lower abnormalities, elevated CSF protein, periventricular
chance of a favorable clinical outcome (66% vs 85%, lesions, and corpus callosum involvement were associ-
p 5 0.009). Older age (OR 1.050, 95% CI 1.029 ated with adult ADEM, while symptoms of men-
1.072, p , 0.001) and presence of encephalopathy ingoencephalitis, including encephalopathy, fever,
(OR 2.358, 95% CI 1.0215.445, p 5 0.044) but headache, and nausea or vomiting, were associated
not sex (p 5 0.980) were significantly associated with with pediatric ADEM. Pediatric and adult ADEM
an unfavorable outcome at the end of follow-up, after have been previously compared in 2 studies that re-
adjustment for length of follow-up. Of the 41 patients ported similar associations regarding some of these
who received IVIg or PLEX as secondary treatment, 28 features.9,30 Here, the functional outcome was less
patients (68%) had favorable clinical outcome. often favorable in adults. This may relate to the
reduced plasticity of the adult brain or differences
DISCUSSION In this multicentered retrospective in immune response, as has been hypothesized pre-
study of more than 200 patients initially diagnosed with viously.9 Although a favorable outcome was more
ADEM, the largest to date, we show several findings of common in patients who did not require additional
clinical relevance: (1) approximately one-quarter of therapeutic management, it is likely that IVIg and
patients experienced a relapse, (2) competing PLEX are each beneficial in some patients.
diagnoses were not uncommon, (3) some features of Children and adults experienced demyelinating re-
the initial presentation may distinguish monophasic lapses in similar proportions in our cohort. However,
from future relapsing disease, and (4) differences children with relapses were more often diagnosed
between pediatric and adult ADEM are identifiable. with multiphasic ADEM compared to adults, in

6 Neurology 86 May 31, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Figure Cumulative incidence of relapsing demyelinating disease over time after an initial diagnosis of acute disseminated encephalomyelitis

Estimate of the cumulative incidence of relapsing demyelinating disease after an initial diagnosis of acute disseminated encephalomyelitis throughout the
observed follow-up time in (A) the whole cohort; (B) pediatric vs adult-onset patients; (C) female vs male patients; and (D) patients with and without enceph-
alopathy. In all estimates, death and a nondemyelinating diagnosis were treated as competing risks. CI 5 confidence interval.

whom relapses were more often diagnosed as MS. remained monophasic, especially in adult onset, do
Differences between the initial presentation of MS not meet the proposed criteria. One retrospective
and multiphasic ADEM, such as the proportion of study31 of pathologically defined ADEM cases with
encephalopathy, fever, and sensory abnormalities, perivenous demyelination found that clinical criteria
were also age-dependent. The hesitancy to diagnose were 80% sensitive and 91% specific. The authors of
children with a disease that requires lifelong therapy that study found that ADEM was overdiagnosed
could explain the low proportion of children diag- and proposed depressed level of consciousness as
nosed with MS. a more specific clinical association with ADEM than
The IPMSSG criteria did not distinguish mono- encephalopathy.
phasic and relapsing demyelinating disease at first pre- The lack of uniform testing and consistent follow-
sentation. Moreover, a high proportion of cases that up is the major limitation of this study. Follow-up

Neurology 86 May 31, 2016 7

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 was significantly longer in patients diagnosed with data. Anusha K. Yeshokumar contributed to drafting and revising the
manuscript for content and acquisition of the data. Paula Barreras con-
a relapsing or alternative disease. It is assumed that
tributed to drafting and revising the manuscript for content and acquisi-
patients who did not seek care remained monophasic. tion of the data. Bittu Majmudar contributed to drafting and revising the
We analyzed data from 4 large hospitals, which likely manuscript for content and acquisition of the data. Joshua P. Klein con-
attract the more diagnostically and therapeutically tributed to drafting and revising the manuscript for content, acquisition
of the data, and interpretation of the data. Tanuja Chitnis contributed to
difficult cases and probably higher ratios of alternative drafting and revising the manuscript for content and interpretation of the
diagnoses. Our years of search were not uniform data. David C. Benkeser contributed to drafting and revising the manu-
among centers. MOG antibody testing has not been script for content, analysis and interpretation of the data, and statistical
analysis. Marco Carone contributed to drafting and revising the manu-
clinically available for patients in the United States,
script for content, analysis and interpretation of the data, and statistical
and AQP4 antibodies were not routinely assessed. analysis. Farrah J. Mateen contributed to drafting and revising the man-
Several studies have reported on AQP4 antibodies uscript for content, study design, analysis and interpretation of the data,
during or after an ADEM event, either as a more dis- and study supervision.

seminated form of NMOSD or a different pathogenic

ACKNOWLEDGMENT
variant. MOG antibodies may play a key role in the
D.L.H. Koelman thanks the American Academy of Neurology for the
pathogenesis in a subgroup of patients with an G. Milton Shy Medical Student Essay Award to present this paper at
ADEM phenotype, potentially distinguish ADEM the 2016 annual meeting in Vancouver, Canada.
from MS,32 and are associated with favorable out-
come after ADEM.33 Despite these limitations, the STUDY FUNDING
large size of the cohort with relatively equal distribu- No targeted funding reported.

tion among children and adults, the careful clinical

DISCLOSURE
phenotyping, diagnoses by subspecialized neurolo-
The authors report no disclosures relevant to the manuscript. Go to
gists, and thorough diagnostic testing allowed us to Neurology.org for full disclosures.
make useful inferences on predictors of relapsing dis-
ease and differences between adults and children. Received November 3, 2015. Accepted in final form February 24, 2016.
This study emphasizes the current difficulties in
the diagnosis of ADEM, exacerbated by the lack of REFERENCES
1. Clifton F. Dissertatio Medica Inauguralis: De Distinctis et
an available, discrete biomarker. Although the
Confluentibus Variolis [in Latin]. Leiden; 1724.
IPMSSG criteria have been partly successful in diag-
2. Tenembaum S, Chamoles N, Fejerman N. Acute dissem-
nosing pediatric ADEM, application to adults is less inated encephalomyelitis: a long-term follow-up study of
useful. The difference in diagnosis of relapsing disease 84 pediatric patients. Neurology 2002;59:12241231.
in children and adults is likely related to practice pref- 3. Hung K, Liao H, Tsai M. Postinfectious encephalomyeli-
erence rather than pathophysiology. Most patients tis: etiologic and diagnostic trends. J Child Neurol 1998;
with a relapsing disease after an initial ADEM diagno- 666670.
4. Idrissova ZR, Boldyreva MN, Dekonenko EP, et al.
sis are probably representations of MS. The initiation
Acute disseminated encephalomyelitis in children: clini-
of MS DMT for this subgroup of patients could be cal features and HLA-DR linkage. Eur J Neurol 2003;10:
considered. The number of alternative diagnoses in 537546.
this cohort stresses the need for vigilance to exclude 5. Singhi PD, Ray M, Singhi S, Khandelwal NK. Acute dis-
other, most often rare disease entities, especially seminated encephalomyelitis in North Indian Children:
because of their possible treatment options and fatal clinical profile and follow-up. J Child Neurol 2006;21:
851857.
outcomes. While this study identified several factors
6. Mikaeloff Y, Caridade G, Husson B, et al. Acute dissem-
at initial presentation to predict a relapsing disease, inated encephalomyelitis cohort study: prognostic factors
no fully reliable prognosticators are known. A large, for relapse. Eur J Paediatr Neurol 2007;11:9095.
perhaps national or international, registry for patients 7. de Seze J, Debouverie M, Zephir H. Acute fulminant
with ADEM, including biomarker collection, appears demyelinating disease. Arch Neurol 2007;64:14261432.
imperative to understand ADEM, compared to first 8. Panicker JN, Nagaraja D, Kovoor JME, Subbakrishna DK.
Descriptive study of acute disseminated encephalomyelitis and
presentations of MS and other related diseases within
evaluation of functional outcome predictors. J Postgrad Med
the CNS demyelinating spectrum. 2010;56:1216.
9. Ketelslegers IA, Visser IER, Neuteboom RF, et al. Disease
AUTHOR CONTRIBUTIONS course and outcome of acute disseminated encephalomy-
Diederik L.H. Koelman contributed to drafting and revising the manu- elitis is more severe in adults than in children. Mult Scler
script for content, study design, analysis and interpretation of the data,
2011;17:441448.
acquisition of data, statistical analysis, and study coordination. Salim
10. Marchioni E, Ravaglia S, Montomoli C. Postinfectious neu-
Chahin contributed to drafting and revising the manuscript for content
and acquisition of the data. Soe S. Mar contributed to drafting and revis-
rologic syndromes: a prospective cohort study. Neurology
ing the manuscript for content and interpretation of the data. Arun Ven- 2013;80:882889.
katesan contributed to drafting and revising the manuscript for content 11. Koelman DLH, Mateen FJ. Acute disseminated encepha-
and interpretation of the data. George M. Hoganson contributed to lomyelitis: current controversies in diagnosis and outcome.
drafting and revising the manuscript for content and acquisition of the J Neurol 2015;262:20132024.

8 Neurology 86 May 31, 2016

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 12. Leake JAD, Albani S, Kao AS, et al. Acute disseminated 10: relapses and long-term disability. Brain 2010;133:
encephalomyelitis in childhood: epidemiologic, clinical 19141929.
and laboratory features. Pediatr Infect Dis J 2004;23: 24. Alper G, Heyman R, Wang L. Multiple sclerosis and acute
756764. disseminated encephalomyelitis diagnosed in children after
13. Schwarz S, Mohr A, Knauth M, et al. Acute disseminated long-term follow-up: comparison of presenting features.
encephalomyelitis: a follow-up study of 40 adult patients. Dev Med Child Neurol 2009;51:480486.
Neurology 2001;56:13131318. 25. Absoud M, Lim MJ, Chong WK, et al. Paediatric acquired
14. Krupp LB, Tardieu M, Amato MP, et al. International demyelinating syndromes: incidence, clinical and magnetic
Pediatric Multiple Sclerosis Study Group criteria for pedi- resonance imaging features. Mult Scler 2013;19:7686.
atric multiple sclerosis and immune-mediated central ner- 26. Dale RC, de Sousa C, Chong WK, et al. Acute dissemi-
vous system demyelinating disorders: revisions to the 2007 nated encephalomyelitis, multiphasic disseminated
definitions. Mult Scler 2013;19:12611267. encephalomyelitis and multiple sclerosis in children. Brain
15. Callen DJA, Shroff MM, Branson HM, et al. Role of MRI 2000;123:24072422.
in the differentiation of ADEM from MS in children. 27. Neuteboom RF, Boon M, Catsman Berrevoets CE, et al.
Neurology 2009;72:968973. Prognostic factors after a first attack of inflammatory CNS
16. Verhey LH, Branson HM, Shroff MM, et al. MRI param- demyelination in children. Neurology 2008;71:967973.
eters for prediction of multiple sclerosis diagnosis in chil- 28. Mikaeloff Y, Adamsbaum C, Husson B, et al. MRI prog-
dren with acute CNS demyelination: a prospective nostic factors for relapse after acute CNS inflammatory
national cohort study. Lancet Neurol 2011;10:1065 demyelination in childhood. Brain 2004;127:19421947.
1073. 29. Neuteboom RF, Ketelslegers IA, Boon M, et al. Barkhof
17. Wingerchuk DM, Banwell B, Bennett JL, et al. Interna- magnetic resonance imaging criteria predict early relapse
tional consensus diagnostic criteria for neuromyelitis opti- in pediatric multiple sclerosis. Pediatr Neurol 2010;42:
ca spectrum disorders. Neurology 2015;85:177189. 5355.
18. Rankin J. Cerebral vascular accidents in patients over the 30. Lin CH, Jeng JS, Hsieh ST, et al. Acute disseminated
age of 60: II: prognosis. Scott Med J 1957;2:200215. encephalomyelitis: a follow-up study in Taiwan. J Neurol
19. Farrel B, Godwin J, Richards S, Warlow C. The United Neurosurg Psychiatry 2007;78:162167.
Kingdom transient ischaemic attack (UK-TIA) aspirin 31. Young NP, Weinshenker BG, Parisi JE, et al. Perivenous
trial: final results. J Neurol Neurosurg Psychiatry 1991; demyelination: association with clinically defined acute
54:10441054. disseminated encephalomyelitis and comparison with
20. Aalen O, Johansen S. An empirical transition matrix for pathologically confirmed multiple sclerosis. Brain 2010;
non-homogeneous Markov chains based on censored ob- 133:333348.
servations. Scand J Stat 1978;5:141150. 32. Ketelslegers IA, Van Pelt DE, Bryde S, et al. Anti-MOG
21. Gray R. A class of K-sample tests for comparing the cumu- antibodies plead against MS diagnosis in an acquired
lative incidence of a competing risk. Ann Stat 1988;16: demyelinating syndromes cohort. Mult Scler J 2015;4:
11411154. 18.
22. Prentice R, Kalbfleish J, Peterson A, et al. The analysis of 33. Baumann M, Sahin K, Lechner C, et al. Clinical and
failure times in the presence of competing risks. Biometrics neuroradiological differences of paediatric acute dissemi-
1978;34:541554. nating encephalomyelitis with and without antibodies to
23. Scalfari A, Neuhaus A, Degenhardt A, et al. The natural the myelin oligodendrocyte glycoprotein. J Neurol Neuro-
history of multiple sclerosis, a geographically based study surg Psychiatry 2014:265272.

Neurology 86 May 31, 2016 9

2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Acute disseminated encephalomyelitis in 228 patients: A retrospective, multicenter US
study
Diederik L.H. Koelman, Salim Chahin, Soe S. Mar, et al.
Neurology published online May 4, 2016
DOI 10.1212/WNL.0000000000002723

This information is current as of May 4, 2016

Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/early/2016/05/04/WNL.0000000000
002723.full.html
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Acute disseminated encephalomyelitis
http://www.neurology.org//cgi/collection/acute_disseminated_encephal
omyelitis
All Demyelinating disease (CNS)
http://www.neurology.org//cgi/collection/all_demyelinating_disease_c
ns
Cohort studies
http://www.neurology.org//cgi/collection/cohort_studies
Multiple sclerosis
http://www.neurology.org//cgi/collection/multiple_sclerosis
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright 2016 American Academy of Neurology. All
rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.