Annals of Clinical Epidemiology
Online ISSN : 2434-4338
Recent overview of patients with anti-N-methyl-D-aspartate receptor encephalitis using a national inpatient database in Japan
Daisuke ShigemiHiroki MatsuiKiyohide FushimiHideo Yasunaga
ジャーナル オープンアクセス HTML

2019 年 1 巻 1 号 p. 11-17



Anti-N-methyl-D-aspartate receptor (anti-NMDAR) encephalitis is an autoimmune disease with multiple neurologic symptoms with or without a tumor, including ovarian teratoma. In the present study, a national inpatient database in Japan was used to investigate the characteristics, treatment, and outcomes of hospitalized patients with anti-NMDAR encephalitis who received initial treatment.


Using the Diagnosis Procedure Combination database, we identified all patients who were diagnosed with anti-NMDAR encephalitis and received initial first-line treatments (methylprednisolone, intravenous immunoglobulin, plasma exchange, and tumor removal) and second-line treatments (cyclophosphamide and rituximab) from July 2010 to March 2017. We excluded patients who received no immunotherapy or surgical treatment and those for whom data were missing. We investigated the characteristics, treatment, and outcomes of eligible patients.


In total, 163 eligible patients were identified. Of these patients, 116 (71%) were female and 44 (28%) were ≤19 years of age. Among the female patients, 44 (38%) had a tumor. Thirty-nine patients (24%) were admitted to the intensive care unit during their hospitalization. Methylprednisolone, intravenous immunoglobulin, and plasma exchange were used as initial therapy in 82%, 56%, and 34% of patients, respectively. Second-line treatments were rarely used. Most patients were alert or had a slight disturbance of consciousness upon discharge. The proportion of patients discharged to home was 55%.


The results of this Japanese study on anti-NMDAR encephalitis suggest that patients’ characteristics and outcomes including the male-to-female ratio, proportion of associated tumors, treatment options, and consciousness disturbance may differ from those in previous reports from other countries.


Encephalitis is an inflammatory condition of the brain. The etiologies of encephalitis include paraneoplastic, autoimmune, and infectious conditions [13].

Anti-N-methyl-D-aspartate receptor (anti-NMDAR) encephalitis is caused by specific antibodies to NMDAR. It is an autoimmune encephalitis that was initially reported in 2007 [4]. Diagnostic criteria for probable and definite anti-NMDAR encephalitis were proposed in 2016. In these criteria, both cell-based assays and tissue-based assays using rat brain immunohistochemistry are recommended in the diagnosis of anti-NMDAR encephalitis [5].

After publication of several case series from Japan, the USA, and the UK [2, 4, 68], a large cohort study of 557 patients with anti-NMDAR encephalitis in 2013 revealed the treatment and prognostic factors for long-term outcomes in these patients [9]. Briefly, young women were more likely to develop anti-NMDAR encephalitis, and ovarian teratoma was frequently associated with anti-NMDAR encephalitis. However, several studies showed that anti-NMDAR encephalitis affected male patients and/or infants and children [10, 11].

Several treatment options for anti-NMDAR encephalitis have been recommended since 2010, including methylprednisolone, intravenous immunoglobulin, and plasma exchange as first-line therapy and rituximab and cyclophosphamide as second-line therapy [10]. Surgical tumor removal is also recommended for patients with tumors, especially teratomas. A large cohort study in 2013 showed that 82% of patients recover substantially at 24 months after disease onset [9]. Indicators of poor neurologic outcomes include the need for intensive care support and a delay in immunotherapy initiation [9].

Although clinical characteristics and therapeutic evidence have been accumulated through studies of anti-NMDAR encephalitis, these studies included relatively small proportions of Japanese patients. Several studies of anti-NMDAR encephalitis have originated from Japan. For example, a retrospective case series study on autoimmune neurological disorders among Japanese patients showed that the NMDAR was the most frequently identified among neuronal cell surface antigens (34/190 patients) [12]. However, the study did not assess clinical characteristics or outcomes. No previous reports have investigated more than 100 patients with anti-NMDAR encephalitis from Japan. A previous report from Japan written in English included the largest number of patients with anti-NMDAR encephalitis (n = 15) [13].

Using a national inpatient database in Japan, the present study was performed to evaluate the characteristics of patients with anti-NMDAR encephalitis who received immunotherapy or underwent tumor removal during hospitalization.


In this retrospective cohort study, we used the Diagnosis Procedure Combination database, a national inpatient database for acute-care inpatients in Japan. The details of this database have been described elsewhere [14]. Briefly, about 1,200 hospitals, including all 82 academic hospitals, participate in the database and provide data for approximately 8 million admissions annually; these admissions represent about 50% of all acute-care inpatients in Japan. Academic hospitals are obliged to participate in the database, whereas participation by community hospitals is voluntary. The database contains discharge abstracts and administrative claims data together with the following data: unique identifiers of hospitals; dates of admission and discharge; patient details (age, sex, body height and weight); smoking status (nonsmoker or current/past smoker); type of admission (planned or emergency); primary and secondary diagnoses; pre-existing comorbidities on admission and complications after admission; medical procedures, including types of surgery; medications and devices used; in-hospital mortality; pregnancy status (pregnant or not); gestational age on admission; and delivery during hospitalization. Diagnoses, comorbidities, and complications are recorded using the International Classification of Diseases, Tenth Revision (ICD-10) codes and text data in Japanese. The database contains no laboratory data or gynecologic examination findings (including ultrasound findings, magnetic resonance imaging findings, and neurological test results). The attending physicians are encouraged to accurately record the diagnoses by linkage of data entry with reimbursement for health care costs. A previous study showed that the validity of the diagnostic records in the database is generally high and that the sensitivity and specificity of the primary diagnoses are 50%–80% and 96%, respectively. The specificity and sensitivity of procedures were found to exceed 90% [15].

We identified all patients who were diagnosed with anti-NMDAR encephalitis (ICD-10 code G048 or G049 with Japanese text of diagnosis) and received initial treatment in a participating hospital from July 2010 to March 2017. We excluded patients who received neither immunotherapy nor surgical treatment and those for whom data were missing. We investigated the eligible patients’ characteristics, treatments, and outcomes.

We categorized the patients’ age as follows: ≤12, 13–19, 20–29, 30–39, 40–49, 50–59, and ≥60 years. The database also provides several clinical scores, including the Japan Coma Scale (JCS) score. The JCS correlates well with the Glasgow Coma Scale; consciousness is scored at 100 points on the JCS, and this is equivalent to a score of 6–9 on the Glasgow Coma Scale. The JCS scores are defined as follows: 0, alert consciousness; 1–3, wakefulness without any stimuli; 10–30, arousal by some stimuli; and 100–300, coma [16, 17].

The outcomes assessed in the present study were the length of hospital stay, consciousness on discharge, discharge to home, and in-hospital death.

Categorical variables are shown as number and percentage and were compared using Fisher’s exact test. Continuous variables are shown as mean and standard deviation or median and interquartile range (IQR) and were compared using Student’s t test or the Mann–Whitney U test.

All statistical analyses were performed using Stata software version 15.0 (StataCorp LP, College Station, TX, USA). All tests were two-tailed, and the threshold for significance was P < 0.05.

The current study was approved by the Institutional Review Board of The University of Tokyo, which waived the requirement for informed patient consent because of the anonymous nature of the data.


During the study period, we identified 163 eligible patients with anti-NMDAR encephalitis who underwent initial treatment in 105 hospitals. Eighty patients had a suspected diagnosis of anti-NMDAR encephalitis without any medical or surgical treatment during the study period, but we did not include these patients. The recorded diagnoses of anti-NMDAR encephalitis for these untreated patients may have had less validity because it is unlikely that patients with true anti-NMDAR encephalitis do not receive any treatment in the first admission.

Table 1 shows the baseline characteristics of the eligible patients by sex. Of these patients, 116 (71%) were female. The median age was 28 years (IQR, 19–37 years), and 44 (28%) patients were ≤19 years old. The proportions of patients with a benign or malignant tumor were 2/47 (4.3%) among male patients and 44/116 (38%) among female patients. Assessment of the JCS score on admission showed that 73 of 117 patients (62%) without a tumor and 36 of 46 patients (78%) with a tumor had disturbance of consciousness (JCS score = 1–300). Thirty-nine patients (24%) were admitted to the intensive care unit (ICU) during hospitalization. Female patients were less likely than male patients to be older, to have alert consciousness upon admission, and to have a smoking habit.

Table 1 Baseline patient characteristics by sex
Female Male P value
n 116 47
Age, years <0.001
 ≤12 3 (2.6) 5 (10.6)
 13–19 29 (25.0) 7 (14.9)
 20–29 42 (36.2) 9 (19.2)
 30–39 28 (24.1) 6 (12.8)
 40–49 4 (3.5) 12 (25.5)
 50–59 4 (3.5) 4 (8.5)
 ≥60 6 (5.2) 4 (8.5)
Body mass index, kg/m2 0.66
 <18.5 27 (23.3) 15 (31.9)
 18.5–24.9 66 (56.9) 23 (48.9)
 25.0–29.9 17 (14.7) 6 (12.8)
 ≥30.0 6 (5.1) 3 (6.4)
Japan Coma Scale score on admission 0.005
 0 (alert) 29 (25.0) 25 (53.2)
 1–3 (wakefulness without any stimuli) 59 (50.9) 14 (29.8)
 10–30 (arousal by some stimuli) 11 (9.5) 5 (10.6)
 100–300 (coma) 17 (14.7) 3 (6.4)
Smoking 9 (7.8) 13 (27.7) 0.002
Use of ambulance 41 (35.3) 13 (27.7) 0.37
Complication of any tumor 44 (37.9) 2 (4.3) <0.001
Pregnancy 1 (0.9)

Data are presented as n (%).

Of 46 patients with diagnosis of a tumor, 41 (89%) underwent tumor removal. Ten patients underwent total or partial ovarian resection without apparent evidence of an ovarian tumor.

Table 2 shows an overview of treatments by sex. The proportion of laparoscopic surgery was 65% among all surgical procedures. Female patients were more likely than male patients to undergo management in the ICU within 2 days of admission, mechanical ventilation, and tracheotomy. Methylprednisolone, intravenous immunoglobulin, and plasma exchange were used in 82%, 56%, and 34% of patients, respectively. Cyclophosphamide was used only in 13 patients, and no patients received rituximab.

Table 2 Overview of treatments
n = 116
n = 47
P value
Intensive care unit admission
 within 2 days of admission 17 (14.7) 1 (2.1) 0.025
 during hospitalization 32 (27.6) 7 (14.9) 0.106
Tracheal intubation 31 (26.7) 7 (14.9) 0.15
Mechanical ventilation 53 (45.7) 11 (23.4) 0.008
Tracheotomy 36 (31.0) 3 (6.4) 0.001
Tumor removal 51 (44.0) 0 (0.0) <0.001
First-line therapy
 Methylprednisolone 94 (81.0) 39 (83.0) 0.83
 Intravenous immunoglobulin 66 (56.9) 25 (53.2) 0.73
 Plasma exchange 45 (38.8) 10 (21.3) 0.04
Second-line therapy
 Cyclophosphamide 7 (6.0) 6 (12.8) 0.20
 Rituximab 0 (0.0) 0 (0.0)

Data are presented as n (%).

Table 3 shows the outcomes during hospitalization by sex. The length of hospital stay was not significantly different between patients with and without a tumor (median, 61 [IQR, 42–161] days vs. 58 [30–99] days, respectively; P = 0.35). Almost all patients had alert consciousness or slight disturbance of consciousness on discharge. The proportions of patients who were discharged to home and transferred to another hospital were 55% and 36%, respectively. Two in-hospital deaths occurred.

Table 3 Outcomes during hospitalization
n = 116
n = 47
P value
Length of hospital stay 64.5 (33–145.5) 44 (21–83) 0.04
 with tumor 61 (36–166.5)
 without tumor 65 (33–130)
Japan Coma Scale score on discharge 0.55
 0 (alert) 89 (78.1) 40 (85.1)
 1–3 (wakefulness without any stimuli) 22 (19.3) 7 (14.9)
 10–30 (arousal by some stimuli) 3 (2.6) 0 (0.0)
Discharge status 0.89
 Discharge to home 61 (52.6) 28 (59.6)
 Transfer to other hospital 43 (37.1) 15 (31.9)
In-hospital death 2 (1.7) 0 (0.0) 1.00

Data are presented as median (interquartile range) or n (%).


In the present study of patients treated for anti-NMDAR encephalitis in Japan, female patients accounted for 71% and the proportion of teenage patients (<20 years) was 28%. The proportions of patients with a tumor were 4% of male patients and 38% of female patients. Methylprednisolone, intravenous immunoglobulin, and plasma exchange were used as first-line therapy in 82%, 56%, and 34% of patients, respectively. Only 8% of patients received cyclophosphamide as second-line therapy. Most patients had almost clear consciousness on discharge; however, only approximately half of them were discharged to home.

The current study revealed characteristics of Japanese patients with anti-NMDAR encephalitis that partially differ from the characteristics found in previous studies. An earlier study using clinical data of 557 patients reported that the proportions of female patients, patients with associated tumors, and ICU-admitted patients were 81%, 38% (6% of male and 46% of female patients), and 77%, respectively [9]. Conversely, the current study showed lower proportions of female patients, patients with associated tumors, and ICU-admitted patients compared with the previous study. The proportion of ICU admissions is relatively low in Japan (24%). This finding may be explained by earlier initiation of treatment due to better access to high-level medical services compared with other countries. A previous study on cross-cultural comparison of critical care between Japan and the United States reported that Japanese ICUs were more likely to be organized to care for older people than ICUs in the United States [18]. This demographic difference could also be related to the lower proportion of ICU admissions among patients with anti-NMDAR encephalitis in Japan.

Neither first-line treatment (intravenous immunoglobulin and plasma exchange) nor second-line treatment (rituximab and cyclophosphamide) for anti-NMDAR encephalitis is included in the universal healthcare coverage in Japan. The first-line therapeutic regimens and surgical procedures in the current study were similar to the recommendations in 2010 [10]. However, the second-line treatments greatly differed between the 2010 recommendations and the real-world clinical practice in Japan. Japanese physicians are usually reluctant to use these medications because of the limitations established by the universal healthcare coverage. A previous study showed better outcomes among patients who did than did not receive second-line treatments [9]. Based on the low proportion of second-line immunotherapy in the current study, improvements in therapeutic options could be important for patients with anti-NMDAR encephalitis.

Most patients in the current study obtained favorable short-term outcomes in terms of lower mortality and smaller proportions of consciousness disturbance on discharge. The reason for this remains unclear, but it may be explained by the high proportion of improvement among patients who received first-line immunotherapy and/or surgical treatment in Japan.

Ten patients underwent surgery without an apparent diagnosis of a tumor. Prompt removal of ovarian tumors was reported as an independent predictor of a positive outcome [9]. Because tumor removal can lead to a decrease in ovarian function, the decision regarding surgical treatment for premenopausal women should be discussed with prudent clinical assessment. In addition, this disorder can improve without tumor removal even in patients with ovarian teratoma-associated anti-NMDAR encephalitis, and approximately 80% of patients recover at 24 months of disease onset [9]. However, some clinical reports have described cases of teratoma-associated anti-NMDAR encephalitis in which the teratoma was identified only microscopically [1921]. The current study and previous reports suggest that even in the absence of imaging evidence of ovarian tumors, a surgical approach and ovarian pathology may be considered for patients who do not experience improvement by nonsurgical treatment. Some studies have shown that prompt laparoscopic tumor removal prior to a definitive diagnosis could be beneficial for anti-NMDAR encephalitis [22, 23].

Several limitations of this study should be acknowledged. First, because the current study was retrospective and observational in nature, a causal relationship between the patients’ characteristics and outcomes could not be confirmed. Second, because we used an administrative claims database, the recorded diagnoses may have been less accurate than those in a planned cohort study. In particular, we identified the diagnosis of anti-NMDAR encephalitis using ICD-10 codes G048 or G049 along with Japanese text “anti-NMDAR encephalitis.” This approach does not ensure validity of the diagnosis, and there may be misdiagnosis in our study population. Third, the definition of a benign or malignant tumor was based on the recorded ICD-10 diagnoses along with text data in Japanese or surgical procedures because no information on pathological findings was available. Moreover, we could not calculate the exact proportion of teratoma among ovarian tumors. Finally, we were unable to investigate long-term outcomes because the database does not provide post-discharge outcomes.

In Japan, the backgrounds of patients with anti-NMDAR encephalitis may differ from those of patients in other countries. The current study suggests that the overall short-term outcomes of patients with anti-NMDAR encephalitis may be favorable in Japan, although there are some limitations with respect to treatment options.


This work was supported by grants from the Ministry of Health, Labour and Welfare, Japan (H29-Policy-Designated-009 and H29-ICT-General-004); and the Ministry of Education, Culture, Sports, Science and Technology, Japan (17H04141).


The authors report no potential conflicts of interest.

© 2019 Society for Clinical Epidemiology