Relapsing Anti-NMDAR encephalitis with later detected MOG-IgG-Positive cortical encephalitis: a case report

Abstract

Background

Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis is the most common autoimmune encephalitis. Neuroimaging findings are frequently normal or nonspecific, and predominant cortical involvement at disease onset is uncommon in anti-NMDAR encephalitis. Overlap with myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) has been increasingly recognized; therefore, the value of individual reports lies in defining their longitudinal phenotypic evolution and diagnostic limitations.

Case presentation

We report a 28-year-old woman with relapsing anti-NMDAR encephalitis followed for nearly seven years across three episodes (the index attack and two relapses). The initial presentation in 2018 was typical anti-NMDAR encephalitis, with positive cerebrospinal fluid (CSF) anti-NMDAR antibodies and complete clinical remission after the first episode. The 2022 relapse manifested as focal cortical encephalitis with progressive unilateral cortical swelling, sulcal enhancement, and seizures; MOG-IgG was not tested. In 2025, she developed recurrent headache, intracranial hypertension, and extensive unilateral cortical involvement, with later detected serum MOG-IgG at 1:32. CSF anti-NMDAR and MOG-IgG were not tested at this admission. The episode was therefore adjudicated as MOG-IgG-positive cortical encephalitis within a longitudinal anti-NMDAR/MOGAD overlap course rather than a proven simultaneous dual-antibody relapse. Oral methylprednisolone was followed by clinical and radiological improvement.

Conclusions

Across nearly seven years, this case documents phenotypic evolution from CSF-confirmed anti-NMDAR encephalitis to recurrent unilateral cortical encephalitis and later serum MOG-IgG detection. A changing cortical phenotype should prompt renewed antibody assessment at each recurrence rather than reliance on the index antibody profile. Because MOG-IgG was not tested in earlier episodes and CSF anti-NMDAR was not tested in 2025, the timing and mechanistic contribution of each antibody cannot be determined.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12883-026-05217-5.

Background

Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis is the most prevalent autoimmune encephalitis, accounting for approximately 80% of reported cases [1]. Clinically, it is characterized by a broad spectrum of neuropsychiatric manifestations, including psychiatric and behavioral disturbances, cognitive impairment, language dysfunction, seizures, movement disorders, altered consciousness, autonomic instability, and central hypoventilation [2]. Brain magnetic resonance imaging (MRI) is often normal or shows only subtle, scattered cortical or subcortical hyperintensities on fluid-attenuated inversion recovery (FLAIR) sequences. In some patients, bilateral mesial temporal signal abnormalities consistent with limbic encephalitis may be observed, whereas in rare cases lesions extend beyond the limbic system. Occasional reports have also described imaging features suggestive of inflammatory demyelination within the central nervous system, including in the cohort reported by Titulaer et al. [3]. However, anti-NMDAR encephalitis initially presenting with predominant cortical involvement remains distinctly uncommon [4, 5].

Here, we report a woman followed for nearly seven years across three episodes: an index episode of CSF-confirmed anti-NMDAR encephalitis in 2018 and relapses in 2022 and 2025 with recurrent unilateral cortical involvement. Serum MOG-IgG was first tested and detected during the third episode, prompting classification of the longitudinal course as anti-NMDAR encephalitis with later MOG-IgG-positive cortical encephalitis [6]. Because anti-NMDAR/MOGAD overlap is now well described, the specific contribution of this report is the extended follow-up, serial clinical and MRI evolution, repeated cortex-predominant phenotype with intracranial hypertension, and recurrence after prior rituximab exposure.

Case presentation

Figure 1 summarizes the longitudinal clinical course.

Fig. 1.

Longitudinal clinical course. The timeline summarizes the patient’s remote seizure history from 2011 to 2016, the first anti-NMDAR encephalitis episode in December 2018, relapse in July 2022 with day 8 clinical and radiological worsening, rituximab/B-cell and antibody follow-up during 2023–2024, relapse in July 2025 with serum MOG-IgG positivity and cortical edema, and marked MRI improvement in October 2025 after oral methylprednisolone OP, opening pressure; IVIG, intravenous immunoglobulin; IVMP, intravenous methylprednisolone; OCBs, oligoclonal bands

First hospitalization (December 2018)

A 28-year-old woman was first admitted in December 2018 with a 10-day history of intermittent right-sided distending headache and visual distortion, followed by one week of episodic loss of consciousness with generalized convulsions two to three times daily. She had experienced infrequent convulsive episodes since 2011, with previously normal brain MRI findings, and had been treated with lamotrigine until 2016, after which the seizures ceased. No autoimmune or infectious evaluation had been performed at that time.

On admission, vital signs and general and neurological examinations were normal. Routine laboratory tests, rheumatologic and infectious serology, tumor markers, and endocrine profiles were unremarkable. Brain MRI with and without contrast, whole-body positron emission tomography-computed tomography (PET-CT), and contrast-enhanced CT of the chest, abdomen, and pelvis showed no abnormalities. CSF analysis demonstrated an opening pressure of 200 mmH₂O, a white blood cell (WBC) count of 151/µL, and a protein concentration of 0.48 g/L; CSF pathogen next-generation sequencing (NGS) was negative. CSF anti-NMDAR antibodies were positive at a titer of 1:100, confirming the diagnosis of anti-NMDAR encephalitis. The patient received intravenous immunoglobulin (IVIG; 0.4 g/kg/day for 5 days) and intravenous methylprednisolone (1,000 mg/day for 3 days), followed by a six-month oral corticosteroid taper, resulting in complete clinical remission. MOG-IgG was not tested in serum or CSF during this episode.

Second hospitalization (July 2022)

In July 2022, after 3.5 years of clinical stability, the patient was readmitted with a 10-day history of left-sided headache and a 3-day history of speech disturbance. The headache worsened with exertion and was accompanied by nausea and vomiting, but not fever. She developed mixed aphasia but did not experience seizures. Neurological examination revealed neck stiffness with preserved limb strength.

Serum anti-NMDAR antibodies were detected at a titer of 1:10. Lumbar puncture showed an opening pressure exceeding 330 mmH₂O, a WBC count of 166/µL (75% mononuclear cells), and a protein concentration of 0.53 g/L. CSF anti-NMDAR antibodies were positive at a titer of 1:32, and CSF NGS remained negative. Brain MRI demonstrated cortical and sulcal FLAIR hyperintensity in the left temporo-parieto-occipital region (Fig. 2). Pelvic ultrasonography revealed small bilateral ovarian cysts. MOG-IgG was not tested in either serum or CSF during this relapse.

Fig. 2.

Brain MRI in July 2022 showed cortical and sulcal FLAIR hyperintensity in the left temporo-parieto-occipital region

Relapsing anti-NMDAR encephalitis was diagnosed. IVIG (0.4 g/kg/day for 5 days) and intravenous methylprednisolone (1,000 mg/day for 3 days) were administered, resulting in initial headache relief and a reduction in CSF inflammatory indices (opening pressure 200 mmH₂O; WBC 50/µL). However, on day 8 of hospitalization, the patient developed worsening aphasia, agraphia, behavioral agitation, memory impairment, impaired consciousness, and recurrent seizures. Repeat MRI revealed marked cortical swelling involving the left frontal, temporal, parietal, occipital, and insular lobes, with sulcal enhancement (Fig. 3). A second course of IVIG combined with a single intravenous infusion of rituximab (300 mg) was administered, after which the patient gradually improved clinically.

Fig. 3.

Repeat brain MRI on day 8 of the 2022 relapse demonstrated marked swelling of the left frontal, temporal, parietal, occipital, and insular cortices, with sulcal enhancement

Six months later, follow-up MRI showed substantial lesion resolution (Fig. 4). The patient continued oral corticosteroids and mycophenolate mofetil for one year, after which both were discontinued. In September 2023, the peripheral B-cell count was 221/µL (8.1%); CSF and serum autoimmune and paraneoplastic antibody panels and oligoclonal bands were negative, and brain MRI was normal. Rituximab was administered (100 mg and 500 mg on days 1 and 2, respectively). In March 2024, when the B-cell count had decreased to 81/µL (2.7%), an additional 500 mg of rituximab was given. Follow-up in July and October 2024 showed B-cell counts of 0 and 116/µL (3.7%), respectively, with persistently negative serum autoimmune antibodies. After two years of stability, rituximab was discontinued. In February 2025, the B-cell count increased to 264/µL (11.6%) without clinical relapse.

Fig. 4.

Follow-up brain MRI approximately six months after treatment showed near-complete resolution of the cortical lesions

Third hospitalization (July 2025)

In July 2025, the patient presented with a one-week history of intermittent vertex headache and no focal neurological deficits. The peripheral B-cell count was 451/µL (18.5%), and serum anti-NMDAR antibodies were weakly positive at a titer of 1:1. Rituximab was administered (100 mg and 400 mg on days 1 and 2, respectively). On day 3, after cold exposure, she developed low-grade fever (37.6 °C) and worsening headache.

CSF examination revealed an opening pressure of 300 mmH₂O, a WBC count of 224/µL (72% mononuclear cells), and a protein concentration of 0.59 g/L; CSF NGS was negative. At this admission, CSF anti-NMDAR antibodies and CSF MOG-IgG were not tested; the reason was not documented in the retrospective medical record. Chest CT was unremarkable, and pelvic ultrasonography again showed a functional ovarian cyst. Brain MRI demonstrated extensive swelling of the left frontal, temporal, parietal, and insular cortices, with linear sulcal enhancement (Fig. 5). Serum MOG-IgG was detected at a titer of 1:32. Ophthalmological examination and optical coherence tomography were normal. These findings were adjudicated as MOG-IgG-positive cortical encephalitis within a longitudinal anti-NMDAR/MOGAD overlap course. The weakly positive serum anti-NMDAR result at 1:1 alone was insufficient to establish a concurrent anti-NMDAR relapse, and the absence of CSF anti-NMDAR testing precluded confirmation of simultaneous dual-antibody activity. Oral methylprednisolone (48 mg/day) resulted in significant symptomatic improvement. One week later, CSF opening pressure decreased to 240 mmH₂O and the WBC count decreased to 50/µL. Follow-up MRI in October 2025 showed marked resolution of cortical edema.

Fig. 5.

Brain MRI during the July 2025 relapse revealed extensive swelling of the left frontal, temporal, parietal, and insular cortices, with linear sulcal enhancement; follow-up MRI in October 2025 showed marked radiological improvement after oral methylprednisolone

Discussion and conclusions

This case documents three episodes over nearly seven years: an index episode of CSF-confirmed anti-NMDAR encephalitis followed by two relapses with a changing cortex-predominant phenotype. The main contribution is not the existence of anti-NMDAR/MOGAD overlap itself, which is increasingly recognized, but the extended follow-up showing recurrent unilateral cortical encephalitis, intracranial hypertension, later serum MOG-IgG detection, and recurrence after prior rituximab exposure. These observations raise diagnostic and pathophysiological questions but do not establish a causal mechanism.

Tumor-related mechanisms are a recognized contributor to relapse in anti-NMDAR encephalitis, particularly in patients who do not undergo tumor removal or tumor-directed therapy [7, 8]. Occult neoplasms may ectopically express NMDAR and sustain antibody production until eradicated. However, extensive and repeated tumor screening in the present case failed to identify a neoplastic source, suggesting that alternative mechanisms were involved.

Immunological factors may have contributed to this disease course. Interactions among peripheral immunity, brain borders, and CNS compartments may sustain humoral immune activity [9, 10]. Mature plasma cells do not express CD20 and are not directly depleted by rituximab [11]. During the 2025 episode, CSF pleocytosis and elevated protein confirmed CNS inflammation but did not identify which antibody system was active [12]. Peripheral B cells had repopulated to 451/µL at that relapse, so the available count documented repopulation rather than continued peripheral depletion. Epitope spreading is one plausible model linking recurrent anti-NMDAR encephalitis and later MOG-IgG detection, because blood-brain barrier disruption and neuronal injury may expose previously sequestered antigens and facilitate secondary autoimmune responses [9, 13, 14]. However, this interpretation remains inferential and does not prove when MOG-IgG first became detectable or that it caused the phenotypic evolution [15, 16].

Treatment-related factors should also be considered. Observational data associate delayed or less intensive immunotherapy with relapse and suggest that early rituximab during the index episode may reduce subsequent relapse risk, but they do not establish a universal maintenance regimen [17]. In this patient, the 2025 episode occurred after prior corticosteroids, IVIG, rituximab, and B-cell-guided retreatment, but after peripheral B-cell repopulation. The available count therefore documents peripheral repopulation rather than continued depletion at relapse, and the clinical course cannot by itself establish treatment-refractory intrathecal antibody production [11].

The detection of serum MOG-IgG during the 2025 clinical and radiological deterioration was clinically relevant in the context of the accompanying cortical phenotype, although causality cannot be established from a single case. The response to methylprednisolone and resolution of cortical swelling are consistent with the steroid responsiveness described in MOGAD [18, 19]. MOGAD can include dynamic cortical or meningocortical abnormalities with prominent cortical edema, features that became evident later in this patient’s course [20, 21]. Because serum anti-NMDAR antibodies were only weakly positive at 1:1 and CSF anti-NMDAR antibodies were not tested in 2025, the episode cannot be classified as a proven simultaneous dual-antibody relapse.

Proposed explanations for coexisting NMDAR and MOG antibodies include parallel or secondary autoimmune responses after CNS inflammation, immunotherapy-related immune reconstitution, infection-related immune activation, and blood-brain barrier disruption [3, 15, 22–24]. Molecular mimicry has also been proposed, but none of these mechanisms can be established in this patient.

The missed MOG-IgG testing during the 2022 relapse is a discrete diagnostic lesson from this case. Antibody results from the index episode should not be assumed to define every recurrence when the clinical or MRI phenotype changes. In patients with recurrent cortex-predominant encephalitis, particularly with cortical swelling, sulcal or leptomeningeal enhancement, or intracranial hypertension, the diagnostic assessment should be renewed at each episode. Testing should be guided by the new phenotype and may include serum MOG-IgG and CSF anti-NMDAR antibodies, together with evaluation for infectious and other inflammatory causes [8, 25–27].

From a therapeutic perspective, this single case does not justify uniformly more aggressive immunotherapy for all patients with first-episode anti-NMDAR encephalitis. Current guidance supports prompt first-line immunotherapy, escalation to second-line treatment in severe or refractory disease, and consideration of maintenance therapy for recurrent disease [28–31]. The present course instead supports a risk-adapted strategy, with repeated clinical, MRI, and antibody reassessment when the phenotype evolves or recurrence occurs [32].

Previous reports provide important context for this clinical course (Table 1). MOG-IgG may precede, coincide with, or be detected after anti-NMDAR encephalitis, and reported phenotypes include demyelinating events, cortical or subcortical lesions, seizures, headache, altered consciousness, and leptomeningeal enhancement [3, 13–16, 33, 34]. Titulaer et al. highlighted demyelinating manifestations among patients with anti-NMDAR encephalitis, while subsequent reviews and cohorts showed temporal heterogeneity, frequent relapse, and variable outcomes [3, 15, 16, 33, 34]. Against this literature, the distinguishing features of the present case are approximately 82 months of observation, evolution across an index episode and two relapses, recurrent unilateral cortical encephalitis with intracranial hypertension, and a 2025 recurrence after prior rituximab exposure. Because MOG-IgG was not tested in 2018 or 2022, the interval to its first detection is known, but the interval to its biological appearance is not.

Table 1.

Previous cases of anti-NMDAR encephalitis and MOGAD overlap.

Study	Cases or setting	Timing of MOG-IgG	Main phenotype	Treatment and outcome	Follow-up duration	Relevance to this case
Titulaer et al. (3)	Large anti-NMDAR encephalitis cohort; 23 of 691 patients had prominent demyelinating manifestations.	MOG antibodies were detected in patients with brainstem or multifocal demyelinating syndromes and in some simultaneous presentations.	Anti-NMDAR encephalitis with demyelinating or multifocal CNS features.	Outcomes varied across phenotypes; the report highlighted reassessment when demyelinating features occur.	Variable across patients; not uniformly reported.	Established that anti-NMDAR encephalitis may coexist with demyelinating antibody syndromes.
Ding et al. (15)	Systematic review and meta-analysis of 58 double-positive cases.	MOG-IgG could precede, coincide with, or be detected after anti-NMDAR encephalitis.	Frequent relapse with heterogeneous encephalitic and demyelinating manifestations, including cortical, subcortical, brainstem, and spinal cord involvement.	Relapse was common and treatment courses varied between reports.	Variable across cases; not uniformly reported.	Provides comparative frequencies for relapse and cortical involvement.
Berek et al. (33)	Case report with long-term follow-up and systematic review of 31 published cases.	MOG-IgG may precede, coincide with, or follow anti-NMDAR encephalitis.	MOG positivity was usually linked to a demyelinating event.	Long-term follow-up emphasized recurrent or multiphasic disease.	66 months for the index case; variable across reviewed cases.	Supports longitudinal antibody profiling across relapses.
Duan et al. (16)	Long-term retrospective study of 10 relapsing overlap patients.	MOG positivity occurred before, simultaneously with, or after anti-NMDAR encephalitis.	Relapsing MOGAD overlapping with anti-NMDAR encephalitis.	Repeated relapses were described; outcomes were heterogeneous.	Mean 25.4 months.	Shows temporal heterogeneity similar to the present patient.
Liu et al. (13)	Series of 13 overlap cases.	Timing varied across patients.	Headache, fever, altered consciousness, neuropsychiatric symptoms, seizures, cortical or subcortical lesions, and leptomeningeal enhancement.	Immunotherapy was generally used; clinical courses were heterogeneous.	6–48 months.	Highlights cortical/subcortical lesions and meningeal enhancement as overlap clues.
Kondo et al. (34)	Literature review of 13 acute overlap cases treated with corticosteroids.	Acute overlap presentations.	Anti-NMDAR/MOG overlap encephalitis with steroid-responsive features.	Steroid therapy was associated with improvement, although relapses occurred during tapering or after cessation.	Variable; some reports provided less than 6 months.	Supports the clinical relevance of steroid responsiveness in overlap cases.
Zhong et al. (38)	Case report and literature review of anti-NMDAR/FLAMES overlap.	Concurrent serum and CSF MOG-IgG and anti-NMDAR positivity in the index case.	Seizures with cortical or meningeal abnormalities; the overlap review included headache, raised intracranial pressure, CSF inflammation, and cortical FLAIR lesions.	High-dose corticosteroids; clinical and radiological improvement.	168 days for the index case.	Direct comparison for the FLAMES-spectrum phenotype in anti-NMDAR/MOGAD overlap.
Present case	Single case followed from 2018 to 2025 across an index episode and two relapses.	MOG-IgG was first tested and detected in serum in 2025; it was not tested in 2018 or 2022.	Recurrent unilateral cortical encephalitis with intracranial hypertension and later serum MOG-IgG positivity.	Improvement after methylprednisolone; the 2025 episode occurred after prior rituximab exposure and after B-cell repopulation.	Approximately 82 months (December 2018-October 2025).	Notable for extended follow-up, phenotypic evolution across three episodes, and repeated cortex-predominant disease.

Abbreviations: CNS Central nervous system MOGAD, FLAMES FLAIR-hyperintense lesions in anti-MOG-associated encephalitis with seizures MOG Antibody-associated disease, MRI Magnetic resonance imaging, NMDAR N-methyl-D-aspartate receptor

FLAIR-hyperintense lesions in anti-MOG-associated encephalitis with seizures (FLAMES) was first recognized as a MOG-IgG-positive unilateral cortical encephalitis phenotype and was subsequently characterized as a distinct clinicoradiographic syndrome [35, 36]. Typical features include seizures, headache, focal cortical symptoms, and unilateral cortical FLAIR hyperintensity with swelling; sulcal FLAIR abnormality and leptomeningeal enhancement occur within the broader MOGAD meningocortical spectrum [37]. Published anti-NMDAR/FLAMES overlap cases likewise describe seizures, headache, CSF inflammation, raised intracranial pressure, and cortical or meningeal MRI abnormalities [38]. The 2022 relapse in our patient closely resembled seizure-defined FLAMES, but MOG-IgG was not tested and MOGAD cannot be assigned retrospectively. The 2025 episode reproduced the unilateral cortical and meningeal pattern with headache, intracranial hypertension, and serum MOG-IgG positivity, but without seizures; it is therefore better framed within the broader MOGAD meningocortical spectrum rather than strict seizure-defined FLAMES.

Rituximab depletes CD20-positive B cells but does not directly eliminate mature CD20-negative plasma cells. In an international cohort of 121 patients with MOGAD, some patients relapsed despite apparent suppression of circulating B cells, indicating that peripheral B-cell counts do not fully predict disease control [39]. Neuropathological studies of anti-NMDAR encephalitis have demonstrated CD138-positive plasma cells and plasmablasts in perivascular and interstitial CNS compartments, providing a biological basis for persistent intrathecal antibody production [40]. In this patient, however, the 2025 relapse occurred after peripheral B-cell repopulation to 451/µL; both repopulation and CD20-negative antibody-secreting cells are plausible considerations, but neither mechanism can be established. Plasma cell-directed bortezomib has been reported in a small series of severe, standard-treatment-refractory anti-NMDAR encephalitis [41]. It was not used in this patient, and this case does not support routine escalation to plasma cell-directed therapy; such treatment remains a potential option for selected severe refractory cases after multidisciplinary assessment.

Abbreviations: CNS, central nervous system; FLAMES, FLAIR-hyperintense lesions in anti-MOG-associated encephalitis with seizures; MOGAD, MOG antibody-associated disease; MRI, magnetic resonance imaging; NMDAR, N-methyl-D-aspartate receptor.

Several limitations constrain interpretation. MOG-IgG was not tested in serum or CSF in 2018 or 2022, and CSF MOG-IgG was not tested in any episode. CSF anti-NMDAR antibodies were not tested in 2025, and the retrospective record did not document why the test was omitted. Consequently, the time at which MOG-IgG first became detectable cannot be determined, and the third episode cannot be classified as a proven simultaneous anti-NMDAR/MOGAD relapse. These limitations also prevent causal attribution of the evolving phenotype to epitope spreading or a specific intrathecal antibody response.

In summary, this case documents a nearly seven-year evolution from CSF-confirmed anti-NMDAR encephalitis to recurrent unilateral cortical encephalitis and later serum MOG-IgG detection. The key clinical lesson is to reassess the antibody profile at each recurrence when the phenotype changes. The 2022 episode was FLAMES-like but cannot be retrospectively assigned to MOGAD, and the 2025 episode is best interpreted as MOG-IgG-positive cortical encephalitis within a longitudinal overlap course rather than proven simultaneous dual-antibody activity. Mechanistic explanations, including epitope spreading and compartmentalized humoral immunity, remain hypotheses.

Abbreviations

AE

Autoimmune encephalitis

AQP4

Aquaporin-4

BBB

Blood-brain barrier

CNS

Central nervous system

CSF

Cerebrospinal fluid

CT

Computed tomography

FLAIR

Fluid-attenuated inversion recovery

IVIG

Intravenous immunoglobulin

MOG

Myelin oligodendrocyte glycoprotein

MOGAD

MOG antibody-associated disease

MRI

Magnetic resonance imaging

NGS

Next-generation sequencing

NMDAR

N-methyl-D-aspartate receptor

OCBs

Oligoclonal bands

OCT

Optical coherence tomography

PET-CT

Positron emission tomography-computed tomography

WBC

White blood cell

Authors' contributions

YL and ZZ contributed equally to data collection and manuscript drafting. ZS, ZX, and BL contributed to clinical data interpretation and imaging analysis. ZLZ supervised the project and critically revised the manuscript. All authors approved the final version of the manuscript.

Funding

This work was supported by the Tianjin Key Medical Discipline Construction Project (Grant No. TJYXZDXK-3-014B).

Data availability

All data generated or analysed during this study are included in this published article. Further inquiries can be directed to the corresponding author.

Declarations

Ethics approval and consent to participate

Ethical approval was waived in accordance with local institutional policies, because this manuscript reports a single retrospective clinical case without experimental intervention. Written informed consent for publication was obtained from the patient.

Consent for publication

Written informed consent was obtained from the patient for publication of this case report and any accompanying images or potentially identifiable data.

Competing interests

The authors declare no competing interests.