ORIGINAL RESEARCH

published: 22 April 2021

doi: 10.3389/fped.2021.605042

Clinical Characteristics and

Prognostic Factors of Children With
Anti-N-Methyl-D-Aspartate Receptor
Encephalitis
Sai Yang, Liming Yang, Hongmei Liao, Mei Chen, Mei Feng, Shulei Liu and Lihong Tan*

Department of Neurology, Hunan Children’s Hospital, Changsha, China

Objective: Anti-N-methyl-D-aspartate (anti-NMDA) receptor encephalitis is the most

common autoimmune encephalitis in pediatric patients. The study aimed to investigate
the clinical characteristics and prognostic factors of anti-NMDA receptor encephalitis in
children in South China.
Edited by:
Carl E. Stafstrom, Methods: This was a retrospective study of children diagnosed with anti-NMDA
Johns Hopkins Medicine,
United States
receptor encephalitis between 01/2014 and 12/2017 at Hunan Children’s Hospital.
Reviewed by:
Laboratory, brain magnetic resonance imaging (MRI), and electroencephalography data
Kevin Rostasy, were collected. The short-term (6-month) outcomes were assessed using the Liverpool
Vestische Kinder- und Jugendklinik score by the same pediatric neurologist. The children were divided into good (scores 4–5)
Datteln, Germany
Thea Giacomini, and poor (score <3) clinical outcomes.
Istituto Giannina Gaslini (IRCCS), Italy
Results: Among the 51 patients, 21 (41.2%) were male. The most common clinical
*Correspondence:
symptoms were dyskinesia (88.2%), personality change (84.3%), seizure (82.4%), and
Lihong Tan
[email protected] cognitive disorder (31.4%). Two were transferred to another hospital, 45 (91.8%) received
intravenous immunoglobulins, 41 (83.7%) received methylprednisolone, and 8 (16.3%)
Specialty section:
received plasma exchange. Eight (16.3%) received rituximab for second-line treatment,
This article was submitted to
Pediatric Neurology, six after intravenous immunoglobulin and methylprednisolone treatment, and two after
a section of the journal plasma exchange therapy failed. Seven were lost to follow-up. The short-term outcome
Frontiers in Pediatrics
was good in 23 patients. Cognitive disorder [odds ratio (OR): 23.97, 95% confidence
Received: 11 September 2020
Accepted: 23 March 2021
interval (CI): 1.12–513.30, P = 0.042) and abnormal brain MRI (OR: 14.29, 95% CI:
Published: 22 April 2021 1.36–150.10, P = 0.027] were independently associated with a poor short-term outcome
Citation: after adjustment for age, GCS, and rituximab use.
Yang S, Yang L, Liao H, Chen M,
Feng M, Liu S and Tan L (2021) Conclusions: MRI abnormalities and cognitive disorders are independently associated
Clinical Characteristics and Prognostic with poor short-term outcomes in children with anti-NMDA receptor encephalitis. The
Factors of Children With
use of rituximab is not associated with the 6-month outcomes.
Anti-N-Methyl-D-Aspartate Receptor
Encephalitis. Front. Pediatr. 9:605042. Keywords: anti-N-methyl-D-aspartate receptor encephalitis, pediatrics, neurology, prognosis, treatment,
doi: 10.3389/fped.2021.605042 rituximab

Frontiers in Pediatrics | www.frontiersin.org 1 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

INTRODUCTION received second-line treatment (12%) (14). On the other hand,

Titulaer et al. (5) reported that second-line treatment was a factor
N-methyl-D-aspartate (NMDA)-receptor encephalitis is an acute associated with good outcomes.
form of encephalitis caused by an autoimmune reaction to the The first case of anti-NMDA receptor encephalitis in China
GluN1 subunit of the NMDA neuronal receptor (1, 2). It usually was reported in 2010, followed by many adult and some pediatric
affects women of reproductive age but can occur in men and cases (18, 19). Still, very few dedicated studies have investigated
patients of all ages (2, 3). Although the prevalence and incidence anti-NMDA receptor encephalitis in pediatric patients or a South
of anti-NMDA receptor encephalitis remain unclear (4), pediatric China population. Therefore, this study aimed to describe the
patients accounted for 37% of the patients in a previous clinical characteristics of children with anti-NMDA receptor
study (5). Anti-NMDA receptor encephalitis is identified more encephalitis in South China and explore the prognostic factors
frequently than any specific viral etiology in children with associated with short-term outcomes.
unknown encephalitis (6). In addition, anti-NMDA receptor
encephalitis is a leading cause of autoimmune encephalitis (7, METHODS
8). Compared with adults, anti-NMDA receptor encephalitis in
children is characterized mainly by seizures, dyskinesia, and Study Design
localized neurological symptoms (5, 9). Recovery is typically This retrospective observational cohort study was approved by
slow and erratic and may take >3 years (2). Relapse occurs in the institutional review board of Hunan Children’s Hospital,
20–25% of the patients (2, 10). The disease is fatal in 4% of Changsha, China. The need for individual consent was waived
the cases (3). by the committee.
The first-line treatment for anti-NMDA receptor encephalitis
is immunotherapy, including methylprednisolone, intravenous Subjects
immunoglobulins, and plasma exchange. The second-line The children diagnosed with anti-NMDA receptor encephalitis
treatment is immunosuppression, typically with rituximab or between January 1, 2014, and December 1, 2017, at the
cyclophosphamide. Nevertheless, standardized treatments are neurology department of Hunan Children’s Hospital of China
not well-defined (5, 8, 9). In addition, who should receive were included. The diagnostic criteria of pediatric anti-NMDA
second-line treatment and when second-line treatment should be receptor encephalitis were (1) age 0–16 years, (2) with at
initiated remain unclear, with wide variations among hospitals. least one sign or symptom among the following: abnormal
Hence, the use rate of second-line immunosuppression is 14– behavior (psychiatric symptoms) or cognitive dysfunction,
81% (5, 9, 11–14). According to a recent survey of pediatric language dysfunction (continuous, uninterrupted mandatory
neurologists, the time of initiating second-line therapy varies language, language reduction, silence), seizures, dyskinesia
widely, from started at diagnosis (0 months) or at 6 months in or myotonia, abnormal posture, decreased consciousness,
the absence of improvement to the primary therapy (8). autonomic dysfunction or central hypoventilation, and positive
The reason for the variability in second-line treatment use for anti-NMDA receptor (GluN1 subunit) IgG antibody
is the unclear effect reported in previous observational studies [antibody detection included the cerebrospinal fluid (CSF)] (20),
(15–17). In a study by Zekeridou et al. (9), although more and (3) exclusion of other possible causes of encephalitis (1). All
patients received second-line treatment, the relapse rate (8%) was children included in this study had positive results for CSF or
similar to that reported in other studies in which fewer patients serum anti-NMDA receptor (GluN1 subunit) IgG antibody.

FIGURE 1 | Magnetic resonance imaging (MRI) of a representative pediatric patient admitted for convulsions for 1 week and diagnosed with anti-NMDA receptor
encephalitis (NMDA receptor antibody in CSF was 1:10) secondary to HSV encephalitis. The patient was first diagnosed with HSV encephalitis in another hospital.
Then, acyclovir was given, but the treatment effect was not good. Bilateral frontal and temporal lobes and insula showed a wide lamellar abnormal signal. The
boundary of the lesions was not clear. T1WI showed a low signal (A,B), and T2WI showed a high signal (C). FLAIR sequence was dominated by a high signal (D). The
lesion-adjacent brain showed scattered and patchy low signal ditch, crack broadening and deepening, and right ventricle temporal horns. The third ventricle was
slightly expanded, with no obvious shift of the midline structure.

Frontiers in Pediatrics | www.frontiersin.org 2 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

Clinical Data Collection

The demographic information, including age, sex, and parents’
educational level, was routinely collected by nurses. The clinical
symptoms were recorded and categorized by experienced
physicians: fever, seizure, consciousness, movement disorder
(e.g., orofacial dyskinesias and choreoathetosis), cognitive
disorders, personality change, autonomic dysfunction (e.g.,
ataxia), and upper respiratory tract infections. Consciousness was
assessed using the Glasgow Coma Score (GCS) and classified as
severe (GCS: 3–8), moderate (GCS: 9–12), and mild (GCS: 13–
15) (21). The Wechsler Intelligence Scale for Children-Fourth
edition-Chinese version (WISC-IV-Chinese) (22) was used to
assess the cognitive capacities of children of 6–16 years of age
when they were conscious, able to communicate, and could
take a 15-min test. Considering that the children might not
be able to complete all the WISC tests, only two index scores
representing speech and memory abilities in the WISC, i.e., the
Verbal Comprehension Index (VCI) and the Working Memory
Index (WMI), were used in the cohort. These two index scores
ranged from 40 to 160. Cognitive disorder was defined as having
at least one index score <70. The cognitive capacities of children
aged <6 years were assessed by physicians asking questions (e.g.,
what did you eat for breakfast) based on some items of the Gesell
Developmental Schedules. If a child exhibited speech difficulties
or memory deficits during these conversations, the physician
diagnosed the child with a cognitive disorder (23). Personality
change assessment was performed according to the Mental
Health Law of the People’s Republic of China and referred to the
diagnostic criteria in ICD-10 (24), i.e., a significant and persistent
change in behavioral patterns and interpersonal relationships for
at least 2 months. The results of the measurement of the Health
of the Nation Outcome Scale for Children and Adolescents
were evaluated (25). FIGURE 2 | Electroencephalogram (EEG) of three patients. (A) A 6-years old
The results from the first magnetic resonance imaging (MRI), female patient with anti-NMDA receptor encephalitis, in a noisy,
electroencephalography (EEG), CSF, and blood after admission non-cooperative, awakened state (EEG parameters: SENS: 10 HF, high pass:
70 Tc, low pass: 0.1 Cal, notch value: 50, sampling rate: 1,000, monitoring
were collected. For this study, an abnormal MRI was defined as
duration: 2 h). The delta rhythm (1.5–3 Hz, 120–380 UV) was used as the main
signs of brain parenchyma damage (any MRI signal abnormality background, with a small amount of low amplitude beta activity and bilateral
indicating acute injury, infection, or old injury) (Figure 1), symmetry. The background rhythm was significantly moderated. (B) A 2 years
and an abnormal EEG was defined as the presence of slow- and 10 months female patient anti-NMDA receptor encephalitis, in an
wave activity and spikes (Figure 2). The laboratory tests for awakened state (EEG parameters: SENS: 10 HF, high pass: 70 Tc, low pass:
0.1 Cal, notch value: 50, sampling rate: 1,000, monitoring duration: 4 h). The
CSF included glucose, protein, and white blood cells (WBC). delta rhythm (2–3 Hz, 80–300 UV) was used as the main background, with
Herpes simplex virus (HSV) and Epstein-Barr virus (EBV) slightly more θ rhythms (4–6 Hz, 50–120 UV) and a small amount of low
infection were tested from the blood. Tumor screening was amplitude beta activity. (C) A 5-years old female patient with anti-NMDA
performed for all children. Abdominal and pelvic ultrasound receptor encephalitis, in an awakened state (EEG parameters: SENS: 10 HF,
was performed, followed by CSF and serum paraneoplastic high pass: 70 Tc, low pass: 0.1 Cal, notch value: 50, sampling rate: 1,000,
monitoring duration: 4 h). The α rhythm (8–9 Hz, 50–90 UV) was the main
antibody testing, including antineuronal nuclear autoantibody background in the two occipital regions. The δ and θ waves (2–5 Hz, 40–120
1 (ANNA 1), ANNA 2, and ANNA 3. If teratomas were UV) were abundant in each region. The low amplitude β activity was observed
detected by ultrasound, the patients underwent surgery. If in each region, which was symmetrical on both sides. Rhythm regulation and
paraneoplastic antibody testing was positive but no teratomas amplitude modulation in the occipital area were poor. The superiority of the
occipital area is obvious.
were detected by ultrasound, a chest computed tomography (CT)
scan was performed.

Treatment
The first-line treatment, including methylprednisolone,
intravenous immunoglobulins, and plasma exchange (alone improvements after 7–14 days of first-line treatment (26, 27). No
or combined), was administered at the time of diagnosis. standard treatment protocol regarding dose, time of initiation,
Second-line rituximab was used if the patients showed no and length of each treatment were followed.

Frontiers in Pediatrics | www.frontiersin.org 3 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

Follow-Up and Short-Term Outcome numbers (percentages) and analyzed using the chi-square test or
Assessment Fisher’s exact test. The variables associated with the outcome (P <
The patients were routinely followed at 6 months after discharge. 0.10) in the univariable analyses were entered into a multivariable
The children were re-evaluated by the same pediatric neurologist. logistic regression analysis. Age is a common prognostic factor
The Liverpool Outcome Score system was used by the physicians and was also included in the multivariable analysis regardless
to assess the short-term outcomes of the children at follow-up. of the univariable analysis results. The results were presented as
The Liverpool Outcome Score system is a commonly used tool odds ratios (OR) with the 95% confidence intervals (CI). P < 0.05
to determine a child’s level of disability and the likelihood of (two-sided) was considered significant.
leading an independent life after Japanese encephalitis (28, 29).
It reflects a child’s actual neurological status and is specific for
children with encephalitis. In addition, based on two field tests in RESULTS
Southeast Asia, including one in India and another in Malaysia, Characteristics of the Patients
the Liverpool Score System has good inter-observer and intra- Figure 3 presents the patient flowchart. Of the 51 patients
observer agreement and good agreement compared with those recruited, 2 were transferred to other hospitals after the diagnosis
of clinical assessments (κ = 0.906) (30). The children were was confirmed, 7 were lost to follow-up, and 42 completed the
divided into two groups. A good clinical outcome was defined 6-month follow-up after discharge. Of the 51 patients, 21 were
as Liverpool Score 5 (full recovery) or 4 (minor sequelae), while male (41.2%), and 30 were female (58.8%), for a male-to-female
a poor clinical outcome was defined as Liverpool Score <3 ratio of 1:1.42 (Table 1). Mean age was 7.4 ± 3.2 years. There were
(moderate or severe sequelae or death). 16 children younger than 6 years. Among them, 50 were positive
for serum anti-NMDA receptor, 48 were positive for CSF anti-
Statistical Analysis NMDA receptor, and 47 were positive for both serum and CSF
SAS 9.4 (SAS Institute Inc., Cary, NY, USA) was used for receptors. The most common clinical symptoms were dyskinesia
statistical analysis. Continuous data were tested for normal (n = 45, 88.2%), personality change (n = 43, 84.3%), epilepsy (n
distribution using the Kolmogorov-Smirnov test. They were = 42, 82.4%), and cognitive impairment (n = 16, 31.4%). GCS
presented as means ± standard deviations and analyzed using was >13 points in 33 children (64.7%), 9–12 in 13 (25.5%), and
Student’s t-test. The categorical variables were presented as <8 points in five (9.8%). Ten patients (19.6%) had HSV infection

FIGURE 3 | Patient flowchart.

Frontiers in Pediatrics | www.frontiersin.org 4 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

and seven (13.7%) had EBV infection. Brain MRI showed brain TABLE 1 | Demographics, clinical symptoms, examination results, treatment,
parenchymal damage in 27 patients (52.9%); among them, the and outcome.

abnormalities were observed in the whole brain (n = 1), temporal Variables Values (%)
lobe (n = 3), temporal lobe and other parts (n = 12), frontal
lobe (n = 4), frontal lobe, and other parts (n = 5), and atypical Demographics
pathological changes (n = 2). EEG revealed slow-wave activity in Age (years) 7.4 ± 3.2
42 patients (82.4%). A 12.5-year-old female patient with ovarian Sex
teratoma underwent tumor resection. Female 30/51 (58.8%)
Of the 49 patients who received treatment, 45 (91.8%) Male 21/51 (41.2%)
received intravenous gamma globulin, 41 (83.7%) received Education level of parents
methylprednisolone, and 8 (16.3%) received plasma exchange, of Elementary 8/51 (15.7%)
whom 4 were severe cases and hence received plasma exchange Middle school 15/51 (29.4%)
directly in the intensive care unit. The other 4 patients switched High school 15/51 (29.4%)
to plasma exchange after intravenous administration of gamma College or higher 13/51 (25.5%)
globulin and methylprednisolone. Eight patients (16.3%) were Clinical symptoms
treated with rituximab, 6 of whom received intravenous gamma Seizure 42/51 (82.4%)
globulin and methylprednisolone without improvement for 7–14 Fever 14/51 (27.5%)
days and received rituximab directly, and 2 received rituximab Upper respiratory infection 2/51 (3.9%)
after plasma exchange therapy failed. The specific treatment plan Autonomic dysfunction 12/51 (23.5%)
was unknown. Supplementary Table 1 presents the treatments of Personality change 43/51 (84.3%)
the patients. Movement disorder 45/51 (88.2%)
Cognitive disorder 16/51 (31.4%)
Clinical Factors Influencing the Outcome Glasgow coma scale
The results showed that cognitive disorder (P = 0.002) and brain
13–15 33/51 (64.7%)
abnormal MRI (P = 0.025) were associated with a poor short-
9–12 13/51 (25.5%)
term outcome. The use of rituximab (P = 0.015) was positively
3–8 5/51 (9.8%)
associated with a poor short-term outcome. There was a trend
Test results
toward children with low GCS scores to have poor short-term
WBC in CSF 24/51 (47.1%)
outcomes (P = 0.051) (Table 2).
Protein in CSF 32/51 (62.8%)

Multivariable Analysis Glucose in CSF 14/51 (27.5%)

In the multivariable logistic regression model, cognitive disorder HSV in blood 10/51 (19.6%)

(OR: 23.97, 95% CI: 1.12–513.30, P = 0.042) and abnormal EBV in blood 7/51 (13.7%)

brain MRI (OR: 14.29, 95% CI: 1.36–150.10, P = 0.027) were EEG: slow-wave activity 29/35 (82.9%)
independently associated with a poor short-term outcome after EEG: spike 7/35 (20.0%)
adjustment for age, GCS, and rituximab use. The use of rituximab MRI abnormality location
was not related to the short-term outcomes (P = 0.540) (Table 3). Whole brain 1 (2.0%)
Temporal lobe 3 (6.0%)
Temporal lobe and other parts 12 (24.0%)
DISCUSSION
Frontal lobe 4 (8.0%)

Anti-N-methyl-D-aspartate (anti-NMDA) receptor encephalitis Frontal lobe and other parts 5 (10.0%)
is the most common form of autoimmune encephalitis in Atypical pathological changes 2 (4.0%)
pediatric patients. This study aimed to investigate the clinical Treatment
characteristics and prognostic factors of anti-NMDA receptor Immunoglobulin 45/49 (91.8%)
encephalitis in children in South China. The results showed Methylprednisolone 41/49 (83.7%)
that the most common clinical features of pediatric anti- Plasma exchange 8/49 (16.3%)
NMDA receptor encephalitis were dyskinesia, personality Rituximab 8/49 (16.3%)
change, seizure, and cognitive disorders. MRI abnormalities and Outcome in the follow-up (poor) 19/42 (45.2%)
cognitive disorders were independently associated with poor
WBC, white blood cells; CSF, cerebrospinal fluid; HSV, herpes simplex virus; EBV,
short-term outcomes in children with anti-NMDA receptor Epstein-Barr virus; MRI, magnetic resonance imaging; EEG, electroencephalogram.
encephalitis. The use of rituximab was not independently
associated with the 6-month outcomes.
The most common clinical features in children with
anti-NMDA receptor encephalitis were movement disorder, frequently exhibited seizures and movement disorders (both
personality change, and seizure. Children with cognitive disorder of which >80%) compared with the largest adult case series
and brain parenchyma damage demonstrated by MRI were more reported by Titulaer et al. (5). These findings were consistent with
likely to have a poor short-term outcome. The children more Zekeridou et al., who compared 36 pediatric patients and 71 adult

Frontiers in Pediatrics | www.frontiersin.org 5 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

TABLE 2 | Comparisons of clinical factors between the two groups.

Variables Poor outcome (n = 19) Good outcome (n = 23) P

Demographics
Age (years) 8.0 ± 1.3 6.4 ± 2.1 0.114
Sex 0.976
Female 10 (52.6%) 12 (52.2%)
Male 9 (47.4%) 11 (47.8%)
Education level of the parents 0.288
Elementary 4 (21.1%) 2 (8.7%)
Middle school 4 (21.1%) 6 (26.1%)
High school 8 (42.1%) 6 (26.1%)
College or higher 3 (15.8%) 9 (39.1%)
Clinical symptoms
Seizure 15 (79.0%) 21 (91.3%) 0.384
Fever 5 (26.3%) 5 (21.7%) >0.999
Upper respiratory infection 2 (10.5%) 2 (8.7%) 0.199
Autonomic dysfunction 3 (15.8%) 10 (43.5%) 0.305
Personality change 15 (79.0%) 21 (91.3%) 0.384
Movement disorder 17 (89.5%) 21 (91.3%) >0.999
Cognitive disorder 11 (57.9%) 3 (13.0%) 0.002
Glasgow coma scale 0.051
13–15 9 (47.4%) 19 (82.6%)
9–12 7 (36.8%) 3 (13.0%)
3–8 3 (15.8%) 1 (4.4%)
Test results
WBC in CSF 8 (42.1%) 12 (52.2%) 0.516
Protein in CSF 11 (57.9%) 17 (73.9%) 0.273
Glucose in CSF 3 (15.8%) 7 (30.4%) 0.305
HSV in blood 4 (21.1%) 1 (4.4%) 0.158
EBV in blood 4 (21.1%) 1 (4.4%) 0.158
MRI (abnormal) 14 (73.7%) 9 (39.1%) 0.025
EEG: slow-wave activity 8 (42.1%) 14 (60.9%) >0.999
EEG: spike 2 (10.5%) 4 (17.4%) >0.999
Treatment
Immunoglobulin 17 (89.5%) 23 (100.0%) 0.199
Methylprednisolone 15 (79.0%) 21 (91.3%) 0.384
Plasma exchange 6 (31.6%) 8 (34.8%) 0.112
Rituximab 7 (36.8%) 1 (4.4%) 0.015

WBC, white blood cells; CSF, cerebrospinal fluid; HSV, herpes simplex virus; EBV, Epstein-Barr virus; MRI, magnetic resonance imaging; EEG, electroencephalogram.

patients from the same center (seizure: 50 vs. 23%, movement et al. (32) reported abnormal MRI in 31% of their patients.
disorder: 83 vs. 55%) (9). On the other hand, fewer tumors were Bacchi et al. (33) showed that in children with head MRI
detected in pediatric patients than in adult patients. Only 1.9% of abnormalities, the incidence of disturbance of consciousness,
all pediatric patients (1/51) had tumors in the present study. In probability of recurrence, and Glasgow severity score was higher
contrast, 34% of all the adult patients had tumors in the study by than in those displaying a normal head MRI. Bartels et al. (34)
Zekeridou et al. (9), and 52% of all the female patients aged >12 reported that 39.5% of the children with anti-NMDA receptor
years had tumors in the study by Titulaer et al. (5). The younger encephalitis had abnormal MRI findings, mainly white matter
the patient, the less likely the tumor was to appear, suggesting that hyperintensities at T2/fluid-attenuated inversion recovery. MRI
the pathogenesis in children and adults might be different (31). showed an abnormality in 52.9% of the children in this
In 2011, Dalmau et al. (12) pointed out that about 55% of study, consistent with Dalmau et al. (12). Zekeridou et al. (9)
patients with anti-NMDA receptor encephalitis had an abnormal reported that cognitive impairment reached 92%, but cognitive
MRI. In 2013, Titulaer et al. (5) reported that 33% of their impairment was only 3% in children with prodromal symptoms.
patients had an abnormal MRI. Zekeridou et al. (9) and Jones Jones et al. (32) reported 56% of cognitive impairment. The

Frontiers in Pediatrics | www.frontiersin.org 6 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

TABLE 3 | Association between clinical factors and poor short-term outcome in the treatment selection bias in a multivariable logistic regression
multivariable analysis. model was adjusted for age, cognitive disorder, abnormal MRI,
Variable P OR 95% CI
and GCS, which were variables significantly associated with the
outcome in the univariable analyses. After adjustment for the
Age 0.938 1.02 0.68–1.51 treatment selection bias in the multivariable analysis, rituximab
Cognitive disorder 0.042 23.97 1.12–513.30 use was no longer significantly related to the outcome.
GCS (9–12 vs. 13–15) 0.556 1.38 0.12–16.04 A significant effect of rituximab on anti-NMDA receptor
(3–8 vs. 13–15) 0.212 11.02 0.39-310.65 encephalitis treatment was not observed in this study.
MRI abnormal 0.027 14.29 1.36–150.10 Nevertheless, Titulaer et al. (5) reported that second-line
Rituximab 0.540 2.477 0.14–45.01 treatment resulted in improved outcomes (OR = 2.69). This
inconsistency might be due to different follow-up durations; a
OR, odds ratio; CI, confidence interval; GCS, Glasgow Coma Scale; MRI, magnetic
short-term outcome at 6 months was reported in the present
resonance imaging.
study, whereas the study by Titulaer et al. included an outcome
assessment at 4–24 months. In addition, the small sample size
of the present study limited the power to detect a significant
present study reported only 31.4% of cognitive impairment, difference. The results of the present study indicated that brain
which was significantly lower than in the two studies mentioned parenchyma damage shown by MRI and cognitive disorders
above. It was possibly because the children in this study were were predictors of a poor short-term outcome. For patients with
divided according to <6 and >6 years old, and the WISC-IV abnormal MRI findings and cognitive disorders, the second-line
and VCI + WMI evaluation systems were adopted to determine treatment application at the time of diagnosis can be explored
the cognitive impairment in the children. The assessment was rather than waiting until no response is observed from the
more complete and detailed. MRI findings in children with first-line treatment. However, no standard treatment protocol
anti-NMDA receptor encephalitis showed brain parenchymal regarding dose, time of initiation, and length of each treatment
damage, which is a factor for poor prognosis. Zekeridou et al. were followed. After admission, the patients were first evaluated
(9) showed that age (>12 years), admission to the ICU, MRI clinically according to the diagnostic criteria of the Chinese
abnormalities, and cognitive impairment were factors for poor Guidelines for the Diagnosis and Treatment of Autoimmune
prognosis. This was consistent with the findings of the present Encephalitis (2017 Edition). Immunotherapy was initiated
study. Jones et al. (32) investigated MRI abnormalities associated in the presence of EEG changes and after excluding other
with poor prognosis, but did not mention cognitive impairment, etiologies such as infection, structural changes, etc. Still, in some
and only highlighted the level of consciousness. Wang et al. patients with atypical clinical manifestations, immunotherapy
(35) suggested that MRI abnormalities and prognosis were not was delayed until the blood and cerebrospinal fluid antibodies
directly related. Mo et al. (36) showed that age, disturbance of were positive. Therefore, there was no uniform timing for
consciousness, and slow waves on EEG were associated with the use of immunotherapy among the patients. Studies are
poor prognosis. The clinical characteristics associated with a poor necessary to determine the optimal timing of immunotherapy in
prognosis still need further exploration. such patients.
The first-line immunotherapy for anti-NMDA receptor This study has several limitations. The sample size was small,
encephalitis includes intravenous gamma globulin, steroids, resulting in wide CIs and poor power. In addition, this study
and plasmapheresis. The second-line immunotherapy includes was retrospective, and the percentage of missing data was 17.7%,
rituximab and cyclophosphamide (12). For refractory patients, leading to bias. Therefore, the conclusions from this study need to
second-line immunotherapy is likely to improve the clinical be replicated in a larger sample size. Moreover, the patients were
outcomes (5). Kahn et al. (8) asked 19 questions to 151 pediatric followed for only 6 months, and hence the relapse information
neurologists in 70 different institutions to identify similarities and long-term outcomes could not be assessed. Because of
and differences in the treatment of anti-NMDA receptor different cognitive functions and neurodevelopment, different
encephalitis. The start time of the second-line treatment varied assessment tools had to be used in children <6 years of age (n
greatly, and the effect of the treatment was not clear. The patient’s = 16) and >6 years of age, probably leading to bias. Finally,
symptoms in the present study did not improve when rituximab as per the routine practice at this hospital, the outcomes were
was used after 7–14 days of a lack of response to the primary only evaluated using the Liverpool score. Future studies could
therapy. Only one patient had a good prognosis (1/8, 12.5%). use other tools used in previous studies of anti-NMDA receptor
An overall response was seen in seven children (7/8, 87.5%), encephalitis, e.g., the NEOS score (37). The use of such tools
which was similar to a previous study (5/6, 83.3%) (19). In the could allow a more robust and quantitative assessment of the
univariable analyses, second-line rituximab was associated with disease outcomes.
the short-term outcome but not in the multivariable analysis. It In conclusion, the most common clinical features of pediatric
could be explained by treatment selection bias, a common type anti-NMDA receptor encephalitis are dyskinesia, personality
of bias in observational studies. In such studies, patients with change, seizure, and cognitive disorders. MRI abnormalities
poor clinical characteristics are more likely to receive second-line and cognitive disorders are independently associated with poor
treatment and have poor outcomes because of their initial clinical outcomes at 6 months in children with anti-NMDA receptor
characteristics. The actual treatment effect cannot be estimated encephalitis. The use of rituximab is not independently associated
without controlling for the treatment selection bias. In this study, with 6-month outcomes. Additional studies are needed to

Frontiers in Pediatrics | www.frontiersin.org 7 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

determine the clinical characteristics and prognosis of pediatric MC participated in data collection and statistical analysis. SY
anti-NMDA receptor encephalitis. wrote the manuscript. All authors read and approved the
final manuscript.
DATA AVAILABILITY STATEMENT
FUNDING
The raw data supporting the conclusions of this article will be
made available by the authors, without undue reservation. This work was supported by Clinical medical technology
innovation guidance project in Hunan Province (Grant Number:
ETHICS STATEMENT 2020SK505).

The studies involving human participants were reviewed and ACKNOWLEDGMENTS

approved by Hunan Children’s Hospital. Written informed
consent from the participants’ legal guardian/next of kin was Xiao Zhang, Peng Wei, and Li Ming for their guidance on EEG
not required to participate in this study in accordance with the and brain imaging in this article.
national legislation and the institutional requirements.
SUPPLEMENTARY MATERIAL
AUTHOR CONTRIBUTIONS
The Supplementary Material for this article can be found
SY analyzed and interpreted data regarding anti-NMDA receptor online at: https://www.frontiersin.org/articles/10.3389/fped.
encephalitis. SY and MF performed the statistical analysis. 2021.605042/full#supplementary-material

REFERENCES and paraclinical observations in a predominantly non-paraneoplastic disorder

of both sexes. Brain. (2010) 133:1655–67. doi: 10.1093/brain/awq113
1. Graus F, Titulaer MJ, Balu R, Benseler S, Bien CG, Cellucci T, et al. A clinical 12. Dalmau J, Lancaster E, Martinez-Hernandez E, Rosenfeld MR,
approach to diagnosis of autoimmune encephalitis. Lancet Neurol. (2016) Balice-Gordon R. Clinical experience and laboratory investigations
15:391–404. doi: 10.1016/S1474-4422(15)00401-9 in patients with anti-NMDAR encephalitis. Lancet Neurol. (2011)
2. Peery HE, Day GS, Dunn S, Fritzler MJ, Prüss H, De Souza 10:63–74. doi: 10.1016/S1474-4422(10)70253-2
C, et al. Anti-NMDA receptor encephalitis. The disorder, the 13. Armangue T, Titulaer MJ, Málaga I, Bataller L, Gabilondo I, Graus F, et al.
diagnosis and the immunobiology. Autoimmun Rev. (2012) Pediatric anti-N-methyl-D-aspartate receptor encephalitis-clinical analysis
11:863–72. doi: 10.1016/j.autrev.2012.03.001 and novel findings in a series of 20 patients. J Pediatr. (2013) 162:850–
3. Rosenfeld MR, Dalmau JO. Paraneoplastic disorders of the 6.e2. doi: 10.1016/j.jpeds.2012.10.011
CNS and autoimmune synaptic encephalitis. Continuum. (2012) 14. Wright S, Hacohen Y, Jacobson L, Agrawal S, Gupta R, Philip S, et al. N-
18:366–83. doi: 10.1212/01.CON.0000413664.42798.aa methyl-D-aspartate receptor antibody-mediated neurological disease: results
4. Prüss H, Dalmau J, Harms L, Höltje M, Ahnert-Hilger G, of a UK-based surveillance study in children. Arch Dis Child. (2015) 100:521–
Borowski K, et al. Retrospective analysis of NMDA receptor 6. doi: 10.1136/archdischild-2014-306795
antibodies in encephalitis of unknown origin. Neurology. (2010) 15. Dale RC, Gorman MP, Lim M. Autoimmune encephalitis in children: clinical
75:1735–9. doi: 10.1212/WNL.0b013e3181fc2a06 phenomenology, therapeutics, and emerging challenges. Curr Opin Neurol.
5. Titulaer MJ, McCracken L, Gabilondo I, Armangué T, Glaser C, Iizuka T, et al. (2017) 30:334–44. doi: 10.1097/WCO.0000000000000443
Treatment and prognostic factors for long-term outcome in patients with anti- 16. Nosadini M, Mohammad SS, Ramanathan S, Brilot F, Dale RC. Immune
NMDA receptor encephalitis: an observational cohort study. Lancet Neurol. therapy in autoimmune encephalitis: a systematic review. Expert Rev
(2013) 12:157–65. doi: 10.1016/S1474-4422(12)70310-1 Neurother. (2015) 15:1391–419. doi: 10.1586/14737175.2015.1115720
6. Gable MS, Sheriff H, Dalmau J, Tilley DH, Glaser CA. The frequency 17. Staley EM, Jamy R, Phan AQ, Figge DA, Pham HP. N-methyl-D-
of autoimmune N-methyl-D-aspartate receptor encephalitis surpasses aspartate receptor antibody encephalitis: a concise review of the
that of individual viral etiologies in young individuals enrolled disorder, diagnosis, and management. ACS Chem Neurosci. (2019)
in the California Encephalitis Project. Clin Infect Dis. (2012) 10:132–42. doi: 10.1021/acschemneuro.8b00304
54:899–904. doi: 10.1093/cid/cir1038 18. Huang X, Fan C, Wu J, Ye J, Zhan S, Song H, et al. Clinical analysis on anti-N-
7. Granerod J, Ambrose HE, Davies NW, Clewley JP, Walsh AL, Morgan D, methyl-D-aspartate receptor encephalitis cases: Chinese experience. Int J Clin
et al. Causes of encephalitis and differences in their clinical presentations in Exp Med. (2015) 8:18927–35.
England: a multicentre, population-based prospective study. Lancet Infect Dis. 19. Sai Y, Zhang X, Feng M, Tang J, Liao H, Tan L. Clinical diagnosis
(2010) 10:835–44. doi: 10.1016/S1473-3099(10)70222-X and treatment of pediatric anti-N-methyl-D-aspartate receptor encephalitis:
8. Kahn I, Helman G, Vanderver A, Wells E. Anti–N-methyl-D- a single center retrospective study. Exp Ther Med. (2018) 16:1442–
aspartate (NMDA) receptor encephalitis:a survey of treatment progress 8. doi: 10.3892/etm.2018.6329
and prospects from pediatric neurologists. J Child Neurol. (2017) 20. Dalmau J, Gleichman AJ, Hughes EG, Rossi JE, Peng X, Lai
32:243–5. doi: 10.1177/0883073816675557 M, et al. Anti-NMDA-receptor encephalitis: case series and
9. Zekeridou A, Karantoni E, Viaccoz A, Ducray F, Gitiaux C, Villega analysis of the effects of antibodies. Lancet Neurol. (2008)
F, et al. Treatment and outcome of children and adolescents with 7:1091–8. doi: 10.1016/S1474-4422(08)70224-2
N-methyl-D-aspartate receptor encephalitis. J Neurol. (2015) 262:1859– 21. Sternbach GL. The Glasgow coma scale. J Emerg Med. (2000) 19:67–
66. doi: 10.1007/s00415-015-7781-9 71. doi: 10.1016/S0736-4679(00)00182-7
10. Irani SR, Vincent A. NMDA receptor antibody encephalitis. Curr Neurol 22. Wasserman GA, Liu X, Parvez F, Factor-Litvak P, Kline J, Siddique AB,
Neurosci Rep. (2011) 11:298–304. doi: 10.1007/s11910-011-0186-y et al. Child intelligence and reductions in water arsenic and manganese:
11. Irani SR, Bera K, Waters P, Zuliani L, Maxwell S, Zandi MS, et al. N- a two-year follow-up study in Bangladesh. Environ Health Perspect. (2016)
methyl-D-aspartate antibody encephalitis: temporal progression of clinical 124:1114–20. doi: 10.1289/ehp.1509974

Frontiers in Pediatrics | www.frontiersin.org 8 April 2021 | Volume 9 | Article 605042

Yang et al. Children With Anti-NMDA Receptor Encephalitis

23. Gläscher J, Tranel D, Paul LK, Rudrauf D, Rorden C, Hornaday A, et al. Island children in New Zealand. Dev Med Child Neurol. (2017) 59:719–
Lesion mapping of cognitive abilities linked to intelligence. Neuron. (2009) 24. doi: 10.1111/dmcn.13420
61:681–91. doi: 10.1016/j.neuron.2009.01.026 33. Bacchi S, Franke K, Wewegama D, Needham E, Patel S, Menon D.
24. Wing TL. ICD-10 medical coding: the role of perioperative services Magnetic resonance imaging and positron emission tomography in anti-
in addressing implementation challenges. AORN J. (2016) 103:177– NMDA receptor encephalitis: a systematic review. J Clin Neurosci. (2018)
88. doi: 10.1016/j.aorn.2015.12.008 52:54–9. doi: 10.1016/j.jocn.2018.03.026
25. Eggleston MJ, Watkins WG. Mental health services for children and 34. Bartels F, Krohn S, Nikolaus M, Johannsen J, Wickstrom R, Schimmel M, et al.
adolescents in New Zealand, outcomes, and the Health of the Nation Clinical and magnetic resonance imaging outcome predictors in pediatric
Outcome Scale for Children and Adolescents (HoNOSCA). N Z Med J. (2008) anti-N-methyl-D-aspartate receptor encephalitis. Ann Neurol. (2020) 88:148–
121:83–91. Available online at: http://www.nzma.org.nz/journal/121-1271/ 59. doi: 10.1002/ana.25754
2983/ 35. Wang Y, Zhang W, Yin J, Lu Q, Yin F, He F, et al. Anti-N-methyl-D-
26. Dale RC, Brilot F, Duffy LV, Twilt M, Waldman AT, Narula S, et al. Utility and aspartate receptor encephalitis in children of Central South China: clinical
safety of rituximab in pediatric autoimmune and inflammatory CNS disease. features, treatment, influencing factors, and outcomes. J Neuroimmunol.
Neurology. (2014) 83:142–50. doi: 10.1212/WNL.0000000000000570 (2017) 312:59–65. doi: 10.1016/j.jneuroim.2017.09.005
27. Binstadt BA, Caldas AM, Turvey SE, Stone KD, Weinstein HJ, Jackson J, et al. 36. Mo Y, Wang L, Zhu L, Li F, Yu G, Luo Y, et al. Analysis of risk factors
Rituximab therapy for multisystem autoimmune diseases in pediatric patients. for a poor prognosis in patients with anti-N-methyl-D-aspartate receptor
J Pediatr. (2003) 143:598–604. doi: 10.1067/S0022-3476(03)00382-2 encephalitis and construction of a prognostic composite score. J Clin Neurol.
28. Ma J, Jiang L. Outcome of children with Japanese encephalitis and predictors (2020) 16:438–47. doi: 10.3988/jcn.2020.16.3.438
of outcome in southwestern China. Trans R Soc Trop Med Hyg. (2013) 37. Dalmau J, Armangue T, Planaguma J, Radosevic M, Mannara F, Leypoldt
107:660–5. doi: 10.1093/trstmh/trt064 F, et al. An update on anti-NMDA receptor encephalitis for neurologists
29. Griffiths MJ, Lemon JV, Rayamajhi A, Poudel P, Shrestha P, Srivastav V, and psychiatrists: mechanisms and models. Lancet Neurol. (2019) 18:1045–
et al. The functional, social and economic impact of acute encephalitis 57. doi: 10.1016/S1474-4422(19)30244-3
syndrome in Nepal–a longitudinal follow-up study. PLoS Negl Trop Dis.
(2013) 7:e2383. doi: 10.1371/journal.pntd.0002383 Conflict of Interest: The authors declare that the research was conducted in the
30. Lewthwaite P, Begum A, Ooi MH, Faragher B, Lai BF, Sandaradura absence of any commercial or financial relationships that could be construed as a
I, et al. Disability after encephalitis: development and validation potential conflict of interest.
of a new outcome score. Bull World Health Organ. (2010)
88:584–92. doi: 10.2471/BLT.09.071357 Copyright © 2021 Yang, Yang, Liao, Chen, Feng, Liu and Tan. This is an open-access
31. Luca N, Daengsuwan T, Dalmau J, Jones K, deVeber G, Kobayashi J, article distributed under the terms of the Creative Commons Attribution License (CC
et al. Anti-N-methyl-D-aspartate receptor encephalitis: a newly recognized BY). The use, distribution or reproduction in other forums is permitted, provided
inflammatory brain disease in children. Arthritis Rheum. (2011) 63:2516– the original author(s) and the copyright owner(s) are credited and that the original
22. doi: 10.1002/art.30437 publication in this journal is cited, in accordance with accepted academic practice.
32. Jones HF, Mohammad SS, Reed PW, Dunn PPJ, Steele RH, Dale RC, No use, distribution or reproduction is permitted which does not comply with these
et al. Anti-N-methyl-D-aspartate receptor encephalitis in Māori and Pacific terms.

Frontiers in Pediatrics | www.frontiersin.org 9 April 2021 | Volume 9 | Article 605042