Articles

Treatment and prognostic factors for long-term outcome in

patients with anti-NMDA receptor encephalitis:
an observational cohort study
Maarten J Titulaer, Lindsey McCracken, Iñigo Gabilondo, Thaís Armangué, Carol Glaser, Takahiro Iizuka, Lawrence S Honig, Susanne M Benseler,
Izumi Kawachi, Eugenia Martinez-Hernandez, Esther Aguilar, Núria Gresa-Arribas, Nicole Ryan-Florance, Abiguei Torrents, Albert Saiz,
Myrna R Rosenfeld, Rita Balice-Gordon, Francesc Graus, Josep Dalmau

Summary
Background Anti-NMDA receptor (NMDAR) encephalitis is an autoimmune disorder in which the use of immuno- Lancet Neurol 2013; 12: 157–65
therapy and the long-term outcome have not been defined. We aimed to assess the presentation of the disease, the Published Online
spectrum of symptoms, immunotherapies used, timing of improvement, and long-term outcome. January 3, 2013
http://dx.doi.org/10.1016/
S1474-4422(12)70310-1
Methods In this multi-institutional observational study, we tested for the presence of NMDAR antibodies in serum or
See Comment page 123
CSF samples of patients with encephalitis between Jan 1, 2007, and Jan 1, 2012. All patients who tested positive for
Department of Neurology and
NMDAR antibodies were included in the study; patients were assessed at symptom onset and at months 4, 8, 12, 18, Neurosciences, Hospital of the
and 24, by use of the modified Rankin scale (mRS). Treatment included first-line immunotherapy (steroids, University of Pennsylvania,
intravenous immunoglobulin, plasmapheresis), second-line immunotherapy (rituximab, cyclophosphamide), and Perelman School of Medicine,
tumour removal. Predictors of outcome were determined at the Universities of Pennsylvania (PA, USA) and Barcelona Philadelphia, PA, USA
(M J Titulaer MD,
(Spain) by use of a generalised linear mixed model with binary distribution. L McCracken BS,
E Martinez-Hernandez MD,
Results We enrolled 577 patients (median age 21 years, range 8 months to 85 years), 211 of whom were children Prof M R Rosenfeld MD,
(<18 years). Treatment effects and outcome were assessable in 501 (median follow-up 24 months, range 4–186): Prof R Balice-Gordon PhD,
Prof J Dalmau MD);
472 (94%) underwent first-line immunotherapy or tumour removal, resulting in improvement within 4 weeks in Department of Neurology,
251 (53%). Of 221 patients who did not improve with first-line treatment, 125 (57%) received second-line Hospital Clinic, The University
immunotherapy that resulted in a better outcome (mRS 0–2) than those who did not (odds ratio [OR] 2·69, CI of Barcelona (d’Investigacions
Biomèdiques August Pi i
1·24–5·80; p=0·012). During the first 24 months, 394 of 501 patients achieved a good outcome (mRS 0–2; median
Sunyer [IDIBAPS]), Spain
6 months, IQR 2–12) and 30 died. At 24 months’ follow-up, 203 (81%) of 252 patients had good outcome. Outcomes (M J Titulaer, I Gabilondo MD,
continued to improve for up to 18 months after symptom onset. Predictors of good outcome were early treatment T Armangué MD, E Aguilar BS,
(0·62, 0·50–0·76; p<0·0001) and no admission to an intensive care unit (0·12, 0·06–0·22; p<0·0001). 45 patients had N Gresa-Arribas PhD, A Saiz MD,
Prof M R Rosenfeld, Prof
one or multiple relapses (representing a 12% risk within 2 years); 46 (67%) of 69 relapses were less severe than initial
F Graus MD, Prof J Dalmau);
episodes (p<0·0001). In 177 children, predictors of good outcome and the magnitude of effect of second-line California Department of
immunotherapy were similar to those of the entire cohort. Public Health, Richmond, CA,
USA (C Glaser MD);
Department of Neurology,
Interpretation Most patients with anti-NMDAR encephalitis respond to immunotherapy. Second-line immunotherapy
Kitasato University School of
is usually effective when first-line treatments fail. In this cohort, the recovery of some patients took up to 18 months. Medicine, Sagamihara, Japan
(T Iizuka MD); Department of
Funding The Dutch Cancer Society, the National Institutes of Health, the McKnight Neuroscience of Brain Disorders Neurology, Columbia
University, New York, NY, USA
award, The Fondo de Investigaciones Sanitarias, and Fundació la Marató de TV3.
(Prof L S Honig MD);
Department of Rheumatology,
Introduction Encephalitis Project, CA, USA), the frequency of anti- Hospital of Sick Children,
In 2005, a syndrome with prominent psychiatric symp- NMDAR encephalitis surpassed that of any specific viral University of Toronto, ON,
Canada (S M Benseler MD);
toms, memory loss, decrease of level of consciousness, cause in young individuals.6 Patients usually present Department of Neurology,
and central hypoventilation was described in four young with acute behavioural change, psychosis, and catatonia Brain Research Institute,
women with an ovarian teratoma and antibodies against that evolve to include seizures, memory deficit, Niigata University, Japan
an antigen highly expressed in the hippocampus.1 Soon dyskinesias, speech problems, and autonomic and (I Kawachi MD); Department of
Neurology, Hospital Sant Pau,
thereafter, the target antigen was identified as the NMDA breathing dysregulation.3 The syndrome resembles the Universitat Autònoma de
receptor (NMDAR).2 The disorder, named anti-NMDAR phenotypes obtained with genetic and pharmacological Barcelona, Spain
encephalitis, has since been seen in patients of all ages, decreases in expression and function of NMDAR.7,8 (E Martinez-Hernandez);
but most often in young adults and children with or Findings from experiments in which patients’ antibodies Department of Neurology,
Children’s Hospital of
without teratomas.3,4 Findings from epidemiological were added to cultures of hippocampal neurons or Philadelphia, University of
studies suggest that this disorder is the most common infused into the hippocampus of rodents showed specific Pennsylvania, Philadelphia,
cause of autoimmune encephalitis after acute antibody-mediated internalisation of NMDAR and PA, USA (N Ryan-Florance MD);
demyelinating encephalomyelitis.5 In a centre focused on alteration of the mechanisms of synaptic plasticity.9,10 Biostatistics and Data
Management Platform,
the study of encephalitis of unclear cause (California Despite the severity of the disease, patients often improve

www.thelancet.com/neurology Vol 12 February 2013 157

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 Articles
IDIBAPS, Hospital Clinic, after intensive care support, immunotherapy, and
Barcelona, Spain lengthy stays in hospital with multidisciplinary care.3,11
(A Torrents BS); and Institució
Catalana de Recerca i Estudis
However, the effectiveness of immunotherapy and its
Avançats (ICREA, Barcelona, long-term effect have not been established. Many patients
Spain (Prof J Dalmau) do not respond to first-line immunotherapy, such as
Correspondence to: steroids, plasmapheresis, and intravenous immuno-
Prof Josep Dalmau, ICREA- globulins, and for these patients the treatment strategy
IDIBAPS-Hospital Clínic,
and outcome are unknown. Moreover, available evidence
Universitat de Barcelona,
Department of Neurology, c/ suggests that the syndrome and response to treatment
Villarroel 170, Barcelona 08036, can differ between children and adults.4 To address these
Spain issues, we assess the presentation of the disease, the
josep.dalmau@uphs.upenn.
edu
spectrum of symptoms, immunotherapies used, timing
of improvement, and long-term outcome.
Methods
Study design and participants
In this observational cohort study, we tested for the
presence of NMDAR antibodies in serum or CSF
samples from patients at the hospitals of the University
of Pennsylvania (Philadelphia, PA, USA) and the
University of Barcelona (Spain), and from patients whose
samples were sent to these institutions (from 200 centres
See Online for appendix in 35 countries; appendix) for study between Jan 1, 2007,
and Jan 1, 2012. Samples were recorded as positive if they
fulfilled previously reported criteria,3 including a
characteristic pattern of immunostaining of the neuropil
of rat brain, and specific reactivity with HEK293 cells
expressing GluN1 (also known as NR1) subunits of the
NMDAR.
All patients who tested positive for NMDAR antibodies
were included in the study without selection criteria
other than availability of clinical information. Clinical
information was obtained by the study investigators or
referring physicians at the acute stage of the disease. On
the basis of the reported manifestations of this disorder,3
we categorised symptoms into eight groups: behaviour
and cognition, memory, speech, seizures, movement
disorder, loss of consciousness, autonomic dysfunction,
501 patients with NMDAR encephalitis
29 had no treatment
472 had first-line immunotherapy and
tumour removal (when applicable)
251 had improvement* 221 had little or no response*
96 had no additional treatment or 125 had second-line treatment
repetition of first-line treatment
Figure 1: Study profile
*The occurrence of improvement or absence of improvement was assessed at 4 weeks from initiation of
treatment.
158
Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
and central hypoventilation. We defined children as
patients younger than 18 years. Previous reports that
suggested differences in symptom presentation and
tumour association in children and adolescents4 led us
to examine these two features in three groups of patients:
pre-pubertal (<12 years), post-pubertal (12–17 years), and
adults (≥18 years). We obtained follow-up information at
regular intervals after symptom onset (months 4, 8, 12,
18, and 24); we assessed neurological status with the
modified Rankin scale (mRS).12 We assessed results
from only the first MRI, electroencephalogram (EEG),
and CSF examinations. All patients were screened at
least once for systemic tumours. First-line immuno-
therapy was defined as the use of steroids, intravenous
immunoglobulins, or plasma exchange alone or com-
bined; second-line immunotherapy included rituximab
or cyclophosphamide alone or combined. There was no
predefined protocol establishing the order or combin-
ation of treatments within each line of treatments. We
recorded initial treatment as a failure if no sustained
improvement occurred within 4 weeks after initiation of
immunotherapy or tumour removal, and if the mRS
score remained at 4 or higher. We excluded patients with
a follow-up shorter than 4 months from analysis of
effects of treatment and outcome. We defined relapse of
encephalitis as the new onset or worsening of symptoms
occurring after at least 2 months of improvement or
stabilisation. Written consent for studies was obtained
from families and patients’ representatives; studies were
approved by the institutional review boards of the Uni-
versity of Pennsylvania and the University of Barcelona.
Statistical analysis
We analysed demographic information and symptoms
with Fisher exact, Fisher-Freeman-Halton test (an
extension of the Fisher exact test for r × c contingency
tables), or Mann-Whitney U test, as appropriate. Because
of a skewed distribution, we used a log-transformation
for age at symptom onset, duration of follow-up, and
delay until initiation of treatment. For untreated patients,
the last date of follow-up (maximum 24 months) was
considered as the delay in initiation of treatment. Factors
affecting outcome were assessed by univariable binary
logistic regression (good outcome defined as mRS 0–2)
independently of the treatments given. Factors associated
with a good outcome (p<0·1) were included in generalised
linear mixed models with binary distribution, declaring
the subject as random and the other factors as fixed
effects.13 All mRS scores at defined timepoints were used
(months 4, 8, 12, 18, and 24; each patient’s follow-up
could provide a maximum of five scores). We included
time from onset of disease as an additional variable,
allowing us to analyse patients with different durations
of follow-up. After creation of this model, we established
the effect of second-line immunotherapy by adding
second-line immunotherapy to the same model applied
to only those patients who had failed first-line
www.thelancet.com/neurology Vol 12 February 2013
 Articles

A
70 Female patients (non-tumour)
Female patients (tumour)
Male patients (non-tumour)
Male patients (tumour)
60

50
Number of patients

40

30

20

10

0
0–5 6–11 12–17 18–23 24–29 30–35 36–41 42–47 48–53 54–59 60–65 66–71 72–77 78–83 84–89
Age at onset (years)
B Age <12 years (n=111) C Age 12–17 years (n=99) D Age ≥18 years (n=364)
Behaviour*
Cognition
Memory deficit
Speech disorder*
Loss of consciousness
Movement disorder
Seizures*
Other

E
100 p<0·0001

90 p=0·024
p=0·002
80
Proportion of patients with symptoms (%)

70

60

50
p<0·0001 Figure 2: Demographic
information, distribution by
40
age of initial symptom, and
cumulative symptoms
30
during the first month of the
p<0·0001
disease
20
p<0·0001 Patients’ age at disease onset
(A) and first symptom at
10
disease onset according to
patients’ age (B–D; first
0
symptom of three patients
rs

8 s
<1 rs

–1 rs

8 s
<1 rs

–1 rs

8 s
<1 rs

–1 rs

8 s
<1 rs

–1 rs

8 s
<1 rs

–1 rs

8 s
s

<1 rs
–1 rs

8 s
<1 rs

–1 rs

–1 rs
8 s
<1 rs

8 s

–1 rs

8 s
s
≥1 ar

≥1 ear

≥1 ear

≥1 ear

≥1 ear

≥1 ear
ar

≥1 ear

≥1 ear

≥1 ear

≥1 ear
ar
ea

a
12 ea

12 ea

a
12 ea

a
12 ea

a
12 ea

a
12 ea

a
12 yea

12 ea

12 ea

unknown; *p<0·0001
ye
ye

ye

ye

ye

ye

ye

ye

ye

ye

ye
2y
2y

2y

2y

2y

2y

2y

2y

2y

y
8

2
8

8
<1

<1
–1

between the three age

12

groups). Distribution by age

Behaviour/ Memory Speech Seizures Movement Loss of Autonomic Central Cerebellar Hemiparesis of cumulative symptoms
cognition deficit disorder disorder consciousness symptoms hypoventilation ataxia during the first month of the
disease (E).

www.thelancet.com/neurology Vol 12 February 2013 159

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 Articles

immunotherapy. We analysed relapse frequency with 12 years and 45 years (205 had tumours); only four girls
Kaplan-Meier curves using log-rank statistics. We used younger than 12 years (6% [ four of 68 girls]; p<0·0001),
SAS (version 9.3) for all multivariable analyses and SPSS and seven men (6% [seven of 109 men]; p<0·0001) had a
(version 20) for all other analyses. tumour (p<0·0005 for tumour risk in male patients
versus female patients; appendix). 207 tumours (94% of
Role of the funding source all tumours) were ovarian teratomas, four tumours (2%)
The study sponsors had no role in the study design, data were extraovarian teratomas, and the remaining nine
collection, data analysis, data interpretation, or writing of tumours (4%) were as follows: two of each of lung, breast,
the report. The corresponding author had full access to and testicular tumours; one ovarian carcinoma; one
all the data in the study and had final responsibility for thymic carcinoma; and one pancreatic cancer. Six of the
the decision to submit for publication. nine patients with tumours other than teratomas were
older than 45 years. Asian and black patients were more
Results likely to have a teratoma (81 [45%] Asian patients and
Overall, we included 577 patients in the demographic 38 [48%] black patients) than were white patients
analyses (135 of whom seen by study investigators, all (58 [31%]) or Hispanic patients (23 [27%]; p=0·007;
others seen by referring clinicians), of whom 501 had appendix).
been followed-up for at least 4 months (median 238 adults (65%) presented with behavioural
24 months, range 4–186; figure 1). Median age at disease problems, and 55 children (50%) younger than 12 years
onset was 21 years (range 8 months to 85 years). presented with seizures or movement disorders
468 patients (81%) were female (figure 2 and appendix). (figure 2). In adolescents, the frequencies of the
211 (37%) of all 577 patients were younger than 18 years, indicated first symptoms were in between those of
and 28 (5%) were 45 years or older (figure 2). Many young children and adults. The occurrence of abnormal
patients younger than 12 years were boys (43 patients behaviour, movement disorders, and seizures was
[39%]), and many patients older than 45 years were men different between the three age groups (p<0·0001;
(12 patients [43%]; p<0·0001 for both). 220 (38%) patients figure 2) Within the first 4 weeks of symptom onset,
had an underlying neoplasm, of whom 213 (97%) were most patients developed a similar spectrum of
female (46% of all girls and women). The presence of a symptoms irrespective of their age. Although movement
tumour predominated in female patients aged between disorders were more common in children, memory
deficits and central hypoventilation occurred more often
in adults (figure 2). Atypical symptoms such as
Non-tumour Tumour All (n=577) p value
cerebellar ataxia or hemiparesis were more common in
(n=357) (n=220)
children than in adults. During the first month of the
Number of symptoms disease, 498 (87%) of 571 patients developed four or
8 28 (8%) 24 (11%) 52 (9%) 0·37* more of the eight categories of symptom; only six
7 64 (18%) 48 (22%) 112 (20%) patients (1%) had only one symptom (table 1). The
6 72 (20%) 52 (24%) 124 (22%) assessment of disease severity showed that most
5 70 (20%) 37 (17%) 107 (19%) patients had a maximum mRS score of 5 and were
4 72 (20%) 31 (14%) 103 (18%) admitted to the intensive care unit (table 1).
3 34 (10%) 13 (6%) 47 (8%) MRI, EEG, and CSF examinations were abnormal
2 10 (3%) 10 (5%) 20 (3%) in 180 (33%) of 540 patients, 432 (90%) of 482 patients, and
1 5 (1%) 1 (<0·5%) 6 (1%) 418 (79%) of 532 patients, respectively. Detection of
Unknown 2 (1%) 4 (2%) 6 (1%) NMDAR antibodies was compared in 250 paired serum
Maximum mRS samples and CSF samples randomly selected by use of an
For the random integer 5 296 (83%) 199 (90%) 495 (86%) 0·086† online random integer generator, showing more sensitivity
generator see http://www. 4 49 (14%) 19 (9%) 68 (12%) in CSF (100%) than serum (85%; p<0·0001; appendix).
random.org/integers/
3 6 (2%) 1 (<0·5%) 7 (1%) In the 501 patients with at least 4 months follow-up,
Unknown 6 (2%) 7 (1%) 1 (<0·5%) 462 (92%) were treated with first-line immunotherapy,
Stay in intensive care unit 134 (27%) with second-line immunotherapy (figure 1 and
Yes 248 (69%) 187 (85%) 435 (75%) <0·0001‡ table 2), and 10 (2%) with tumour removal without
No 100 (28%) 32 (15%) 132 (23%) immunotherapy, and 29 (6%) were not treated. The
Unknown 9 (3%) 1 (<0·5%) 10 (2%) combination of first-line immunotherapy most often
used was steroids and intravenous immunoglobulins
Data are n (%). mRS=modified Rankin scale. *Fisher exact test: 1–3 vs 4–8 symptoms,
(202 [44%] patients). Of 197 patients with teratoma,
comparing patients with and without tumours. †Fisher-Freeman-Halton test.
‡Fisher exact test. 189 (96%) had tumour resection, 20 (11%) of them at
neurological relapse or after recovery.
Table 1: Number and severity of symptoms during the first month of
During the first 24 months, 394 of 501 patients who
the disease
had at least 4 months’ follow-up had an mRS score of 0–2

160 www.thelancet.com/neurology Vol 12 February 2013

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 Articles

(median 6 months, IQR 2–12), and 30 died. At 24 months’

Non-tumour (n=304) Tumour (n=197) All (N=501) p value*
follow-up, 203 (81%) of 252 patients had a favourable
outcome and 24 (10%) died (figure 3 and appendix). In Median time from symptom onset 21 (14–49) 21 (14–42) 21 (14–46) 0·090
until treatment in days (IQR)
univariable analysis, the factors associated with good
First-line immunotherapy 283 (93%) 179 (91%) 462 (92%) 0·40
outcome included no need for admission to an intensive
Steroids 265 (87%) 156 (79%) 421 (84%) 0·024
care unit, early treatment, and low severity of disease
Intravenous immunoglobulins 221 (73%) 125 (63%) 346 (69%) 0·030
within 4 weeks of onset (maximum mRS; table 3). A
Plasmapheresis 80 (26%) 83 (42%) 163 (33%) 0·0003
longer follow-up was associated with a better outcome
Second-line immunotherapy† 93 (31%) 41 (21%) 134 (27%) 0·017
(p<0·0001; table 3). In multivariable analysis the factors
associated with good outcome included early treatment Rituximab 71 (23%) 30 (15%) 101 (20%) 0·030

and no need for admission to an intensive care unit; the Cyclophosphamide 50 (16%) 31 (16%) 81 (16%) 0·90
proportion of patients with good outcomes improved up Other immunotherapy‡ 23 (8%) 8 (4%) 31 (6%) 0·13
until 18 months (table 3). Median time from symptom onset ·· 1·4 (0·7–2·6, ··
until tumour removal in months –13 to 177)
Of the 472 patients who were treated with first-line (IQR, range)
immunotherapy or tumour removal, about half had Surgery§ 14 (5%) 189 (96%) ·· <0·0001
symptom improvement within 4 weeks of treatment During initial episode 14 (5%) 169 (86%) ··
(figure 1). During the first 24 months, 241 of these At relapse 0 7 (4%) ··
patients had an mRS score of 0–2 (median 3 months,
After recovery 0 13 (7%) ··
IQR 2–5); at 24 months’ follow-up, 111 (97%) of
Failure of first-line immunotherapy¶
115 patients had a good outcome (figure 3). Of the
Yes 145 (48%) 76 (39%) 221 (44%) 0·069
221 (47%) patients who had no improvement with these
No 138 (45%) 103 (52%) 241 (48%)
treatments, 125 (57%) received second-line immuno-
Surgery with no immunotherapy 1 (<0·5%) 9 (5%) 10 (2%)
therapy and 96 (43%) had no further immunotherapy or
No treatment 20 (7%) 9 (5%) 29 (6%)
continued receiving first-line immunotherapy (figure 1).
Of the 125 patients who received second-line immuno- Data are n (%), unless otherwise stated. *Fisher’s exact test, except for time from symptom onset, which is
therapy, 84 had an mRS score of 0–2 during the first Mann-Whitney U test. †84% (n=85) of patients treated with weekly rituximab received four treatments of 375 mg/m²,
8% (n=8) received five to eight treatments; and 8% (n=8) received fewer than four treatments (in most cases two
24 months (median 10 months, IQR 6–21); in the treatments of 500 mg/m² given 2 weeks apart); 73% (n=59) of patients treated with monthly cyclophosphamide
96 patients who did not receive second-line received three to six treatment cycles of 750 mg/m², 15% (n=12) received seven to 12 cycles, and 12% (n=10) received
immunotherapy, 49 had an mRS score of 0–2 (median one or two cycles. Four serious adverse events were reported including, one anaphylactic reaction and one infection
due to rituximab, both causing a delay in treatment, and one infection and one severe lymphopenia due to
15 months, IQR 7 to not achieved). At 24 month follow-up
cyclophosphamide causing discontinuation of this drug—no treatment-related deaths or irreversible complications
43 (78%) of 55 patients who failed first-line treatment and were identified. ‡Azathioprine, mycophenolate mofetil, tacrolimus, or methotrexate. §Eight patients with a tumour
received second-line treatment and 32 (55%) of 58 patients did not have surgery: six had teratomas and five of them died of encephalitis; one was treated for an adenocarcinoma
who failed first-line and not receive second-line treatment of the breast and the encephalitis improved; and one had a small-cell lung cancer and died. ¶12 patients failed first-line
therapies at relapse: seven without tumour had responded to first-line therapies at initial episode, the other five (one
had a good outcome (figure 3). In multivariable analysis, of them with a teratoma) did not have immunotherapy at their first episode.
we identified the use of second-line immunotherapy as
an additional factor for good outcome (table 3). Table 2: Overview of treatments (n=501)
29 patients were not treated with immunotherapy or
surgery. The reasons for no treatment included first with fewer relapses (p=0·038; figure 4) Only 12 patients
diagnosis of the disease during relapses in 14 patients with teratoma identified at symptom presentation or
(initial episodes not treated), retrospective diagnosis in thereafter had neurological relapses (appendix).
nine patients (four post mortem and five with archived The use of immunotherapy in the initial episode of
serum or CSF samples), and spontaneous improvement encephalitis was associated with a lower frequency of
within a few weeks in six patients. Overall, in this group, relapses (p=0·038; figure 4 and appendix). Second-line
29% had a poor outcome (mRS 3–5 or dead). immunotherapy was associated with fewer relapses in
During the 24 months’ follow-up, 45 patients (a 12% patients without tumour (p=0·007; appendix). Similarly,
risk) had clinical relapses, 15 (33%) of whom had multiple the introduction of second-line immunotherapy during
relapses (appendix). Compared with the initial episode, relapses was associated with a decreased occurrence of
46 (67%) of 69 relapses were less severe (as shown by a subsequent relapses (p=0·024; appendix).
lower mRS score measured at the stage of maximum The time between symptom onset and initiation of
severity), more often mono-symptomatic (24 [35%] treatment was shorter in children (all individuals younger
relapses), and resulted in fewer admissions (12 [17%] than 18 years) than it was in adults (21 days [IQR 14–35] vs
relapses) to the intensive care unit (p<0·0001 for all). In 28 days [14–49]; p=0·007; appendix). Overall, the outcome
16 (23%) relapses, the severity of symptoms was similar to in children was much the same as that in adults (p=0·92;
that of the patient’s initial episode, but for seven (10%) table 3 and appendix). In children, multivariable analysis
relapses it was worse. Patients without a tumour had a identified similar risk factors as those of adults
higher frequency of relapses than did those with a tumour (appendix). 81 (46%) of 177 children failed first-line
(p=0·0007), and the use of immunotherapy was associated immunotherapy—in these patients, the magnitude of

www.thelancet.com/neurology Vol 12 February 2013 161

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 Articles

A B
100 mRS score
6
5
80 4
3
Proportion of patients (%)

2
1
60
0

40

20

0
4 8 12 18 24 4 8 12 18 24
Number of patients 501 450 382 308 252 251 219 181 144 115

C D
100

80
Proportion of patients (%)

60

40

20

0
4 8 12 18 24 4 8 12 18 24
Follow-up (months) Follow-up (months)
Number of patients 96 88 78 66 58 125 116 96 72 55

Figure 3: Clinical outcome

Clinical outcome in all patients (A; 29 patients did not have immunotherapy or tumour removal), patients who responded to first-line immunotherapy (steroids,
immunoglobulins, plasmapheresis; B), patients who failed first-line immunotherapy and did not receive second-line therapy (C), patients who failed first-line
immunotherapy and received second-line therapy (rituximab, cyclophosphamide, or both; D).

the beneficial effects of second-line immunotherapy
(used in 53 [65%] patients; appendix) was similar to that
of the entire cohort, although because of the smaller
sample size was not significant (OR 3·35, CI 0·86–12·98;
p=0·081; appendix).
Discussion
Our findings show that immunotherapy and tumour
removal, if applicable, result in substantial neurological
improvement in 81% of patients with anti-NMDAR
encephalitis after a median follow-up of 24 months
Moreover, second-line immunotherapy with rituximab,
cyclophosphamide, or both, improved the outcome of
patients who did not respond to first-line treatment and
decreased the occurrence of relapses.
Symptom presentation varied between children and
adults (symptoms were more often neurological in
children and more often psychiatric in adults), but in
most cases the progression of symptoms evolved towards
a similar syndrome. 1 month after disease onset, 87% of
patients had four or more of the indicated groups of
symptoms. Recognition of the combinations of these
symptoms, which usually includes alteration of
cognition and behaviour (delusions, psychosis, catatonia)
and abnormal movements (orofacial dyskinesias,
choreoathetosis), should prompt testing for antibodies
against the GluN1 subunit of the NMDAR, especially in
young people.
Identification of NMDAR antibodies confirms the
diagnosis of the disorder and should lead to the search for
a tumour, which, if present, is almost always an ovarian
teratoma that contains nervous tissue and expresses
NMDAR.14,15 Teratomas located elsewhere and other
tumours are uncommon. In this study, the occurrence of
an underlying teratoma was greater in female patients
aged 12 years or older than it was in young children and
male patients (52% vs 6%). Therefore, if a tumour is not
detected, the extent and frequency of tumour screening
should take into consideration a patient’s age and sex. In
female patients aged 12 years or older, a screening

162 www.thelancet.com/neurology Vol 12 February 2013

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 Articles

approach similar to that of paraneoplastic syndromes is

p Odds ratio (95% CI) Events
appropriate (eg, MRI of the abdomen and pelvis every
6 months for 4 years),16 but in young children (<12 years) Univariable analysis
and male patients the need for repeat screening is unclear. Stay in intensive care unit <0·0001 0·22 (0·10–0·46) ··
If a tumour is detected, removal is important because it Time until treatment initiation (loge) 0·009 0·79 (0·66–0·94) ··
speeds up improvement and decreases relapses.3,17 Time of follow-up (loge) <0·0001 1·09 (1·07–1·11) ··
Two independent predictors of good outcome were the Maximum mRS 0·011 ·· ··
lower severity of symptoms, assessed as no need for 3 (n=7) 0·569 1·86 (0·22–15·59) ··
admission to an intensive care unit, and the prompt 4 (n=58) 0·003 8·66 (2·08–36·11) ··
initiation of immunotherapy and tumour removal, if 5 (n=436)* 1·00 ·· ··
appropriate. A notable finding was the protracted phase Age (loge) 0·87 1·02 (0·77–1·36) ··
of recovery that continued until the 18 months’ follow-up. ≥18 (vs <18) years 0·92 0·98 (0·62–1·53) ··
By establishing the last follow-up assessment at Tumour 0·55 1·15 (0·74–1·78) ··
24 months, our study probably underestimates the Sex 0·93 1·02 (0·59–1·77) ··
frequency and level of recovery because some patients Multivariable analysis
had a shorter follow-up, others continued to improve Stay in intensive care unit <0·0001 0·12 (0·06–0·22) 394
after 24 months, and those with substantial improvement Time until start of treatment (loge) <0·0001 0·62 (0·50–0·76) 394
tended to discontinue follow-up (appendix). For the same Follow-up <0·0001 ·· 394
reasons, our findings have probably overestimated the 4 months† ·· 0·04 (0·02–0·06) 224
occurrence of mortality; imputation analysis estimated 8 months† <0·0001 0·20 (0·12–0·35) 110
7% mortality at 24 months (appendix). 12 months† 0·0044 0·37 (0·21–0·66) 32
Decisions about the type and duration of immuno- 18 months† 0·0066 0·76 (0·42–1·37) 22
therapy were made by referring physicians and were 24 months† 0·36 1·00 6
based on clinical symptoms, not antibody titres. 92% of Maximum mRS 0·51 ·· 394
patients underwent first-line immunotherapy including Multivariable analysis of first-line failure
steroids, intravenous immunoglobulins, and plasma- Stay in intensive care unit <0·0001 0·07 (0·02–0·25) 133
pheresis alone or combined. This treatment and tumour Time until start of treatment (loge) 0·005 0·55 (0·36–0·83) 133
removal, if applicable, resulted in improvement within the Follow-up <0·0001 ·· ··
first 4 weeks of treatment in 53% of patients, and 97% of
4 months† ·· 0·01 (0·00–0·02) 33
them showed a good outcome at 24 months’ follow-up
8 months† <0·0001 0·11 (0·05–0·23) 57
(mRS score of 0–2). The main challenge was posed by the
12 months† 0·004 0·26 (0·12–0·54) 22
47% of patients who did not improve with first-line
18 months† 0·004 0·70 (0·33–1·49) 16
treatment within the first 4 weeks of starting this
24 months† 0·35 1·00 5
treatment. These patients continued to have a poor mRS
Second-line treatment 0·012 2·69 (1·24–5·80) 133
score (median score of 5). In this group, patients who
received rituximab, cyclophosphamide, or both, had better loge=natural logarithm. mRS=modified Rankin scale.*An mRS score of 5 was the reference category. †To assess
outcomes than did those who continued with first-line continuous improvement over time, the outcome at each individual timepoint was compared with that of the previous
timepoint, with the p values indicated as well as the amount of events in those specific months after the previous
immunotherapy or who received no further immuno- timepoint. After 8 months the number of patients achieving good outcome (or events) becomes small (as reflected in
therapy. Similar results were obtained for paediatric the large confidence intervals). Patients improving from mRS 2 to 0 or 1 (or from 5 to 4 or 3) were not counted as
patients; in this smaller group the effect of second-line events; these are all considered good outcome (or poor outcome), respectively. The 24-month follow-up is the reference
value for the odds ratios (and the associated lower and higher range of the confidence interval) of individual timepoints.
immunotherapy was not statistically significant but the
estimated benefit was similar to that of the entire cohort Table 3: Factors associated with good outcome (mRS 0–2)
(ORs 3·35 vs 2·69). Second-line immunotherapy was well
tolerated—only four patients had serious adverse effects,
all of which were reversible (table 2). Indeed, after failing first-line immunotherapy, all these
In this study, the effect of individual treatments patients remained in a similar poor neurological status,
(eg, plasma exchange vs intravenous immunoglobulins enabling the comparison between those who
or rituximab) could not be compared side-by-side subsequently received second-line immunotherapy with
statistically. This is because the effect or lack of effect those who did not. The choice of adding second-line
of a specific treatment directly affected physicians’ immunotherapy was based on the treating physician’s
decisions about additional treatment, leading to a preference, family acceptance of potential side-effects,
paradox in which the more treatment a patient receives, and availability of the drugs. Although selection bias
the worse their outcome, but in fact the worse a patient cannot be ruled out in observational studies like this
is doing, the more treatment they receive. However, we one, we could not identify obvious bias between patients
could statistically assess the effect of the addition of who after failing first-line treatments were treated with
second-line immunotherapy to patients whose second-line treatment and those who did not: the
symptoms were refractory to first-line treatments. severity of symptoms (92% vs 88%) and duration of stay

www.thelancet.com/neurology Vol 12 February 2013 163

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 Articles

A Panel: Research in context

25 Non-tumour
Tumour
All
Systematic review
We searched Medline and Embase up until Oct 01, 2012, for
articles published in any language with the search terms
20 “Anti-N-Methyl-D-Aspartate Receptor Encephalitis”,
“Receptors, N-Methyl-D-Aspartate”, and “Encephalitis”. We
restricted searches to studies in human beings. We also
reviewed the reference lists of the papers identified by this
15 search. We identified ten studies and one extended review
describing at least three patients; five of the studies included
Relapse (%)

at least 25 patients.
Interpretation
10
Our study extends epidemiological and clinical knowledge on
anti-NMDA receptor encephalitis. It lends support to the
female preponderance, the high frequency of the disease in
children, and a higher risk for an underlying teratoma in black
5
people. The higher frequency of teratomas in Asian patients
is a novel finding. The findings suggest that the clinical
presentation in children is often different from that in adults;
this finding has been previously suggested in small series but
0
0 2 4 6 8 10 12 14 16 18 20 22 24 not substantiated until now. In this study, the occurence of
Number at risk
Non-tumour 296 243 194 151 114
relapses was less frequent than in previous reports.
Tumour 193 170 144 116 96 Additionally, the less severe nature of the symptoms
All 489 413 338 267 210
compared with the initial episodes was noted in most
B relapses. To our knowledge, this is the first study to provide
25 No treatment extensive and structured follow-up and prognostic factors
Only first-line immunotherapy
Second-line immunotherapy that affect outcome. The suggested better outcome of
patients with teratoma compared with those without
teratoma was not confirmed in this study; however, we
20 identified stay in an intensive care unit and the time from
symptom onset to initiation of treatment as important
prognostic factors. First-line immunotherapy was often
insufficient to control the disease; in this group of patients,
15 the addition of second-line immunotherapy was associated
Relapse (%)

with a better outcome compared with those who did not

receive second-line immunotherapy.

10
in an intensive care unit (88% vs 87%) were much the
same, and we recorded no differences in age, sex,
ethnicity, and delay of initial immunotherapy.
The lower frequency of neurological relapses (12% vs
5
20–24% reported in previous studies11,18) was probably
caused by better recognition of the disorder, earlier
treatment, and increasing use of second-line immuno-
therapy. Indeed, patients who received second-line
0
0 2 4 6 8 10 12 14 16 18 20 22 24 immunotherapy during the initial episode of encephalitis
Follow-up (months) had fewer relapses. Although this finding should be
Number at risk
No treatment 41 37 34 28 26 considered with caution because of the small number of
Only first-line 315 264 213 163 133 patients, a similar effect was seen in patients who had
immunotherapy
Second-line 134 111 92 64 50 previous relapses—in these patients the introduction of
immunotherapy second-line immunotherapy reduced the occurrence of
subsequent relapses. The presence of a teratoma not
Figure 4: Frequency of relapses, by tumour status (A) and line of immunotherapy (B)
In panels A and B, the total number of patients at baseline is 513 instead of 577 because for some patients no
detected or untreated during the initial episode, or the
follow-up was available. The number at risk at 4 months’ follow-up is 489 and not 501 because some patients had possibility of tumour recurrence, should be considered in
already had a relapse (event). patients with a relapse.

164 www.thelancet.com/neurology Vol 12 February 2013

Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.

The discovery of anti-NMDAR encephalitis has changed
the diagnostic and treatment approach to several
neurological and psychiatric disorders, and has led to the
identification of an expanding group of autoimmune
encephalitis related to antibodies against synaptic
proteins.19 Rapidly progressive alterations of memory,
behaviour, cognition, and psychosis previously thought to
be idiopathic or virally induced are now known to be
autoimmune (panel).20–23 To our knowledge, a systematic
analysis of the effects of sequential (first-line, second-line)
immunotherapy and their effect on long-term outcome
has not been previously done in any of these disorders or
other antibody-mediated neurological diseases.
Our study has the limitations of not being randomised,
but it is precursory to future trials to establish the efficacy
of each individual treatment (eg, steroids, intravenous
immunoglobulins, plasmapheresis, rituximab, and
cyclophosphamide) and duration of immunotherapy.
Moreover, the large number of events (patients achieving
a good outcome) and prospective follow-up of many
patients reported in this study, provide level II-2
evidence24 of the usefulness of immunotherapy and
tumour removal in anti-NMDAR encephalitis.
Contributors
MJT and JD designed the study and created the figures. MJT did the
literature search. MJT, FG, and JD arranged the funding. All authors
contributed to data collection and analysis and critical appraised the final
version of the paper.
Conflicts of interest
MJT is supported by a KWF fellowship 2009-4451 of the Dutch Cancer
Society. Some authors are supported in part by NIH RO1NS077851 (JD),
RO1CA89054 (JD), and RC1NS068204 (RB-G, JD); a McKnight
Neuroscience of Brain Disorders award (RB-G and JD), Fondo de
Investigaciones Sanitarias (FIS, Spain, 11/01780, JD; PS09/0193, FG; and
08/00285, EM-H), Fundació la Marató de TV3 (JD), and National
Institute on Aging (P50AG08702), the Alzheimer’s Association, the
Alzheimer’s Disease Drug Discovery Foundation, and the Taub Institute
(LSH). JD receives royalties from Athena Diagnostics for a patent for the
use of Ma2 as an autoantibody test, and licensing fees from Euroimmun
for a patent for the use of NMDAR as an autoantibody test.
Acknowledgments
We would like to thank all physicians, patients, and families of patients
that provided clinical information.
References
1 Vitaliani R, Mason W, Ances B, Zwerdling T, Jiang Z, Dalmau J.
Paraneoplastic encephalitis, psychiatric symptoms, and
hypoventilation in ovarian teratoma. Ann Neurol 2005; 58: 594–604.
2 Dalmau J, Tuzun E, Wu HY, et al. Paraneoplastic anti-N-methyl-D-
aspartate receptor encephalitis associated with ovarian teratoma.
Ann Neurol 2007; 61: 25–36.
3 Dalmau J, Gleichman AJ, Hughes EG, et al. Anti-NMDA-receptor
encephalitis: case series and analysis of the effects of antibodies.
Lancet Neurol 2008; 7: 1091–98.
4 Florance NR, Davis RL, Lam C, et al. Anti-N-methyl-D-aspartate
receptor (NMDAR) encephalitis in children and adolescents.
Ann Neurol 2009; 66: 11–18.
www.thelancet.com/neurology Vol 12 February 2013
Downloaded for Anonymous User (n/a) at Costa Rica University from ClinicalKey.com by Elsevier on October 11,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
Articles
5 Granerod J, Ambrose HE, Davies NW, et al. Causes of encephalitis
and differences in their clinical presentations in England:
a multicentre, population-based prospective study. Lancet Infect Dis
2010; 10: 835–44.
6 Gable MS, Sheriff H, Dalmau J, Tilley DH, Glaser CA. The
Frequency of autoimmune N-Methyl-D-Aspartate receptor
encephalitis surpasses that of individual viral etiologies in young
individuals enrolled in the California Encephalitis Project.
Clin Infect Dis 2012; 54: 899–904.
7 Krystal JH, Karper LP, Seibyl JP, et al. Subanesthetic effects of the
noncompetitive NMDA antagonist, ketamine, in humans.
Psychotomimetic, perceptual, cognitive, and neuroendocrine
responses. Arch Gen Psychiatry 1994; 51: 199–214.
8 Mohn AR, Gainetdinov RR, Caron MG, Koller BH. Mice with
reduced NMDA receptor expression display behaviors related to
schizophrenia. Cell 1999; 98: 427–36.
9 Hughes EG, Peng X, Gleichman AJ, et al. Cellular and synaptic
mechanisms of anti-NMDA receptor encephalitis. J Neurosci 2010;
30: 5866–75.
10 Mikasova L, De Rossi P, Bouchet D, et al. Disrupted surface cross-
talk between NMDA and Ephrin-B2 receptors in anti-NMDA
encephalitis. Brain 2012; 135: 1606–21.
11 Irani SR, Bera K, Waters P, et al. N-methyl-D-aspartate antibody
encephalitis: temporal progression of clinical and paraclinical
observations in a predominantly non-paraneoplastic disorder of
both sexes. Brain 2010; 133: 1655–67.
12 van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J.
Interobserver agreement for the assessment of handicap in stroke
patients. Stroke 1988; 19: 604–7.
13 Verbeke G, Molenberghs G. Models for discrete longitudinal data.
New York: Springer-Verlag, 2005.
14 Tuzun E, Zhou L, Baehring JM, Bannykh S, Rosenfeld MR,
Dalmau J. Evidence for antibody-mediated pathogenesis in
anti-NMDAR encephalitis associated with ovarian teratoma.
Acta Neuropathol 2009; 118: 737–43.
15 Dabner M, McCluggage WG, Bundell C, et al. Ovarian teratoma
associated with Anti-N-Methyl D-Aspartate receptor encephalitis:
a report of 5 cases documenting prominent intratumoral lymphoid
infiltrates. Int J Gynecol Pathol 2012; 31: 429–37.
16 Titulaer MJ, Soffietti R, Dalmau J. et al. Screening for tumours in
paraneoplastic syndromes: report of an EFNS Task Force.
Eur J Neurol 2011; 18: 19–e3.
17 Sabin TD, Jednacz JA, Staats PN. Case records of the Massachusetts
General Hospital. Case 26-2008. A 26-year-old woman with
headache and behavioral changes. N Engl J Med 2008; 359: 842–53.
18 Gabilondo I, Saiz A, Galan L, et al. Analysis of relapses in
anti-NMDAR encephalitis. Neurology 2011; 77: 996–99.
19 Lancaster E, Dalmau J. Neuronal autoantigens—pathogenesis,
associated disorders and antibody testing. Nat Rev Neurol 2012;
8: 380–90.
20 Chapman MR, Vause HE. Anti-NMDA receptor encephalitis:
diagnosis, psychiatric presentation, and treatment. Am J Psychiatry
2011; 168: 245–51.
21 Kayser MS, Dalmau J. The emerging link between autoimmune
disorders and neuropsychiatric disease. J Neuropsychiatry Clin Neurosci
2011; 23: 90–97.
22 Pruss H, Holtje M, Maier N, et al. IgA NMDA receptor antibodies
are markers of synaptic immunity in slow cognitive impairment.
Neurology 2012; 78: 1743–53.
23 Pruss H, Kinke C, Holtje M, et al. NMDA receptor antibodies in
herpes simplex encephalitis. Ann Neurol 2012; published online
July 16. DOI:10.1002/ana.23689.
24 US Preventive ServicesTask Force (August 1989). Guide to clinical
preventive services: report of the US Preventive Services Task Force.
DIANE Publishing, 1989.
165