Anti-N-methyl-ᴅ-aspartate receptor (NMDAR) encephalitis is an autoimmune encephalopathy that was first diagnosed in 2007 [1]. This disorder is characterized by a rapid, less than 3 months, onset of several symptoms that include abnormal behaviors, speech dysfunction, epileptic seizures, dyskinesia or rigidity, decreased level of consciousness, or autonomic dysfunction [2]. More than 85% of patients with this disease exhibit psychiatric symptoms [3]. NMDAR diagnostic confirmation requires detection of anti-NMDAR immunoglobulin G antibodies targeting the GluN1 subunit of the NMDAR, a neuronal surface synaptic receptor, in the cerebrospinal fluid (CSF) [4]. However, diagnosis and treatment are often delayed when no obvious abnormalities are identified in the CSF profile or results of brain magnetic resonance imaging (MRI) and electroencephalography (EEG). This report is of a case in which the patient was diagnosed with anti-NMDAR encephalitis even though the initial workup identified no abnormalities.

Written informed consent for publishing this case report and its associated laboratory results was obtained from the patient.

A 40-year-old man presented to our outpatient department with a 3-week history of abnormal behavior. His caregiver reported that he had been speaking abusively and demonstrated short-term memory loss that manifested as a failure to recall his recent actions and to only intermittently remember procedures for performing simple tasks like opening a door in his home. On the day of clinical presentation, he was observed shoplifting, forgetting to pay for store items before leaving. He remembered the incident but could not recall his reasoning associated with the event. He also verbalized a feeling of not being in control of his actions. He reported that he had not consumed any unusual foods or alcohol and that he had not taken any medications.

Approximately 2 and a half years before this event, he had received immunotherapy at another hospital for a presumptive diagnosis of autoimmune encephalitis. At that time, he was experiencing headache, disorientation, and aphasia. He was treated with intravenous acyclovir, methylprednisolone, immunoglobulin, rituximab, and tocilizumab for approximately 1 month. A fluid-attenuated inversion recovery image from brain MRI performed at the beginning of hospitalization demonstrated diffuse high signal intensity in the left fronto-parieto-temporal cortex and insula region. A CSF examination revealed a white blood cell (WBC) count of 80/mm³, predominantly mononuclear cells. A brain positron emission tomography (PET) scan performed during the third week of treatment indicated diffuse hypermetabolism in the left fronto-parieto-temporal cortex.

During hospitalization, he experienced seizures, and multiple EEGs were performed; however, no epileptiform discharges were observed on these EEGs. He was prescribed levetiracetam 2,000 mg/day and lacosamide 300 mg/day. These were gradually tapered until the cessation of therapy. CSF autoimmune encephalitis-associated antibodies were not identified; the final diagnosis was seronegative autoimmune encephalitis. He was discharged in the sixth week of hospitalization without any obvious sequelae. Six months after discharge, he received rituximab 500 mg; and at the 8-month follow-up examination, he began a 4-month azathioprine of 100 mg/day, maintenance therapy regimen.

We extended azathioprine maintenance for 2 additional months. His Mini-Mental State Examination (MMSE) score was 30 at the time of azathioprine discontinuation. He had no neurological symptoms after immunosuppressant therapy cessation; however, one year after discontinuation of the azathioprine therapy, he experienced an unprovoked seizure during sleep. An EEG revealed a sharp wave in the right frontal region; we, therefore, re-initiated levetiracetam, 1,000 mg/day. Two weeks later, he was prescribed escitalopram, 10 mg/day, and alprazolam, 0.5 mg/day, for the onset of depression and anxiety. Approximately 3 weeks after this therapeutic reinitiation, he began exhibiting the same abnormal behaviors he exhibited on the initial presentation. He was hospitalized 3 weeks later.

On admission, his vital signs were within the normal range; headache and fever were absent. Neurologic examination was unremarkable except for decreased language fluency and mild dysarthria. An EEG performed on the day of admission demonstrated no epileptiform discharges. A brain MRI and a CSF study that included cytological evaluation and polymerase chain reaction screening for central nervous system infectious disease were conducted the following day. No abnormalities were noted on these examinations, but his MMSE score was 21. Chest and abdominopelvic computed tomography (CT) performed the next day had no findings suggestive of a tumor or infection. Serum tumor markers of prostate-specific antigen, alpha-fetoprotein, carcinoembryonic antigen, carbohydrate antigen 19-9, squamous cell carcinoma antigen, and beta-human chorionic gonadotropin were not observed. Paraneoplastic antibodies, such as anti-Hu, anti-Yo, anti-Ri, anti-amphiphysin, anti-CV2, anti-Ma2/Ta, anti-recoverin, anti-Sry-like high mobility group box 1, and anti-titin antibodies were also absent on blood testing. However, antibodies against myelin oligodendrocyte glycoprotein were present. Therefore, to rule out the possibility of adverse effects from recently prescribed medications, therapy with escitalopram (10 mg/day) and alprazolam (0.5 mg/day) was discontinued. This was followed by close patient monitoring.

On his fifth day of hospitalization, he exhibited a blank stare for approximately 5 minutes during a walk and told his caregiver that he was already dead. Epileptiform discharges were absent on a follow-up EEG performed on the sixth hospitalization day, but intravenous immunoglobulin (IVIG; 0.4 g/kg/day) was initiated due to the possibility of autoimmune encephalitis recurrence. The next day, the second day of IVIG administration, he intermittently experienced generalized sweating and became unresponsive to voice commands. However, a follow-up brain MRI failed to demonstrate any abnormalities. On the eighth day of hospitalization, the patient had a consistently tense posture with his left arm half-bent, of which he had no awareness. He reported sudden transient episodes of profuse sweating in his hands and feet. A testicular sonographic examination performed on the following day showed no abnormalities. Composite Autonomic Scoring Scale performance on the 10th day of hospitalization suggested severe sudomotor dysfunction. At that time, he completed his 5-day IVIG treatment regimen.

The next day, he exhibited aggressive behavior, including screaming and kicking objects. As a result, rituximab 375 mg/m² was administered on hospital day 13. On hospital day 14, his aggressive behavior and catatonia resolved, and his language fluency and dysarthria improved significantly. A whole-body PET-CT performed on the same day showed no abnormalities. A second dose of rituximab at the previous dosage was administered on the 20th day of hospitalization; and, having returned to health, he was discharged the next day.

Test results of a CSF specimen collected on the second day of hospitalization confirmed the presence of anti-NMDAR antibodies. For this, he was administered two additional rituximab treatments at the same dosage: one at 6 days and one at 13 days post-discharge. An MMSE conducted approximately 4 months after discharge showed a score of 30, and the patient remained asymptomatic. Figure 1 summarizes the patient’s previous and current clinical courses.

The patient in this case had anti-NMDAR encephalitis, the prompt diagnosis of which was challenging because all initial standard encephalitis diagnostic tests had negative results. The presence of positive findings suggestive of encephalitis in cases of anti-NMDAR encephalitis is relatively low. In 15% to 96% of anti-NMDAR encephalitis cases, the initial CSF profile is devoid of signs of an inflammatory condition [5], and 50% to 77% of cases display no obvious MRI abnormalities suggestive of the condition [6]. Kwon et al. [7] reported a case of anti-NMDAR encephalitis in which a patient presenting with abnormal behavior for 5 days before hospital admission had unremarkable initial CSF and MRI findings. However, a CSF study performed 1 week later demonstrated a WBC count of 18/mm³, emphasizing the importance of follow-up testing. In our case, however, the patient visited our hospital approximately 3 weeks after symptom onset. Therefore, difficulties arose with interpreting the initial CSF findings as false negative because the testing was not conducted early enough in the disease process. Espinola-Nadurille et al. [5] analyzed CSF samples from 29 patients with confirmed anti-NMDAR encephalitis and found pleocytosis in only 13 (44.8%). The CSF was collected an average of 12.6 ± 8 days after symptom onset. Dürr et al. [8] reported that pleocytosis in anti-NMDAR encephalitis cases is reduced in CSF testing on samples obtained more than 3 weeks after symptom onset. In our case, the patient’s CSF was initially collected approximately 3 weeks after symptom onset; this might explain the absence of abnormal findings.

The diagnostic criteria for probable anti-NMDAR encephalitis proposed by Graus et al. [4] emphasize the need to weigh clinical manifestations more heavily than laboratory findings. In our case, psychosis was the initial and predominant symptom. This was followed by speech dysfunction, memory impairment, blank staring episodes indicative of epileptic seizures, sudomotor dysfunction, and left arm catatonia. Although CSF testing and MRI and EEG had no obvious abnormal findings, the clinical presentation suggested the possibility of autoimmune encephalitis. Therefore, we initiated immunotherapy despite these negative laboratory results. The patient’s symptoms improved dramatically after receiving rituximab. Cases of anti-NMDAR encephalitis have favorable prognoses when appropriate treatment is initiated early in the disease process [9]. Anti-NMDAR encephalitis is believed to be more common in women, but the patient in this case was a man. The symptoms presented were very typical of anti-NMDAR encephalitis, suggesting a lack of sex-related differences in symptoms. Altintas et al. reported no apparent sex difference in the presenting symptoms of anti-NMDAR encephalitis [10].

In conclusion, autoimmune encephalitis should be considered as a differential diagnosis even in the absence of paraclinical abnormalities. This case highlights the importance of prioritizing clinical presentation in the diagnostic approach to autoimmune encephalitis.