Anti-NMDA Receptor Encephalitis Following SARS-CoV-2 Infection: A Systematic Review and Pooled Case Series

1. Introduction

The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), first identified in Wuhan, China, in December 2019, is responsible for the global outbreak of Coronavirus Disease 2019 (COVID-19), which has since posed major public health challenges worldwide [1,2].

While the acute infection has been the primary focus of clinical and research attention, increasing evidence highlights that 10–30% of patients experience persistent, long-term effects collectively known as post-COVID syndrome (PCS) [3]. The National Institute for Health and Care Excellence (NICE) defines PCS as symptoms that develop during or after a COVID-19 infection that last for more than 12 weeks, and cannot be explained by an alternative diagnosis [3]. PCS encompasses a wide range of manifestations, including chronic fatigue, musculoskeletal pain, respiratory difficulties, cardiovascular irregularities, and, notably, neurological and neuropsychiatric conditions [3,4].

Among these neurological sequelae, encephalitis has emerged as a significant but poorly understood complication of COVID-19. The proposed pathogenesis of COVID-19-related encephalitis is multifactorial, with three main mechanisms suggested: (i) direct invasion of the central nervous system (CNS), (ii) systemic inflammation, and (iii) molecular mimicry. Direct viral invasion may occur through either transsynaptic propagation or hematogenous spread, facilitated by SARS-CoV-2 binding to angiotensin-converting enzyme 2 (ACE2) receptors on neuronal and endothelial cells [5,6]. However, as cerebrospinal fluid (CSF) polymerase chain reaction tests are frequently negative in COVID-19 encephalitis, direct invasion is considered less likely [7,8,9,10]. More plausibly, systemic inflammation and cytokine dysregulation, particularly involving IL-6 and Th17 cell pathways, can disrupt the blood–brain barrier and contribute to neurological damage [11,12,13,14,15,16]. Additionally, molecular mimicry may trigger autoimmune responses, with antibodies generated against SARS-CoV-2 antigens cross-reacting with neuronal epitopes, potentially leading to autoimmune encephalitis such as anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis [6,17,18,19,20,21].

Anti-NMDAR encephalitis is an autoimmune disorder characterized by IgG antibodies against the NR1 subunit of the NMDA receptor, resulting in neuropsychiatric and neurological symptoms [22]. Although relatively rare—approximately one case per 1.5 million people annually—it is one of the most recognized subtypes of autoimmune encephalitis, affecting predominantly young women [23].

The disease typically evolves in stages, beginning with viral-like prodromes, followed by acute psychiatric symptoms (e.g., psychosis, agitation), then progressing to seizures, movement disorders, dysautonomia, and, in severe cases, coma [24]. Recovery is often prolonged, with persistent cognitive and behavioral impairments even after inflammation resolves [24].

The diagnostic criteria for anti-NMDAR encephalitis are categorized as probable or definite [24]. A probable diagnosis involves the rapid onset (within three months) of at least four out of six major symptom groups: abnormal psychiatric or cognitive behavior, speech dysfunction (e.g., pressured speech, verbal reduction, or mutism), seizures, movement disorders (such as dyskinesias or rigidity), decreased consciousness, and autonomic dysfunction or central hypoventilation [24]. Supporting findings may include an abnormal EEG, showing focal or diffuse slow or disorganized activity, epileptic activity, or an extreme delta brush pattern, along with CSF abnormalities like pleocytosis or oligoclonal bands [24]. Alternatively, three major symptoms plus a systemic teratoma can support a probable diagnosis [24]. For a definite diagnosis, one major symptom group along with IgG GluN1 antibodies is required; if only serum is available, additional confirmatory tests, such as immunohistochemistry or cell-based assays, are recommended [24].

Before the COVID-19 pandemic, viral infections (e.g., herpes simplex, varicella zoster) and paraneoplastic conditions such as ovarian teratoma were known triggers of anti-NMDAR encephalitis [25,26]. However, the potential association between SARS-CoV-2 and the development of this autoimmune encephalitis remains underexplored. Given the global burden of COVID-19 and growing case reports of post-COVID encephalitis, there is an urgent need to synthesize the emerging evidence.

The objective of this systematic review is to systematically map and analyze the literature on anti-NMDAR encephalitis following COVID-19 infection. Specifically, this review aims to describe the clinical manifestations and diagnostic challenges, summarize current therapeutic approaches, enhance early recognition to improve patient outcomes, and highlight gaps in knowledge to guide future research. Additionally, we present eight post-COVID-19 anti-NMDAR encephalitis cases to contribute to the growing body of evidence.

2. Methods

Design: This study consisted of two components, a sysytematic review of published literature and a case series of eight new patients diagnosed with post-COVID-19 anti-NMDAR encephalitis at Loghman Hakim Hospital, Tehran, Iran. The systenatic review was conducted following the PRISMA guidelines. We additionally followed the methodological framework established by Arksey and O’Malley, which is commonly recommended for systematic reviews.

2.1. Search Strategy

We searched five databases of PubMed, MEDLINE, Scopus, Embase, and Web of Science to extract case report and case series articles on autoimmune encephalitis associated with SARS-COV-2 published from the inception to September 23, 2024. The search strategy consisted of keywords including: ([SARS-CoV-2] OR COVID OR COVID-19 OR coronavirus) AND (Anti-NMDR Encephalitis OR Non paraneoplastic Anti-NMDR Encephalitis) OR ([Anti-N-Methyl-D-Aspartate Receptor Encephalitis] OR [NMDA] OR [Autoimmune encephalitis] OR [anti-NMDA receptor autoantibody] OR [Receptors, N-Methyl-D-Aspartate]). The search field was limited to ‘title/abstract’. No language restriction was applied. complete search strategy table presented in the supplementary file (Table S1).

Eligibility criteria: We included studies that reported cases of anti-NMDA receptor encephalitis associated with COVID-19 infection. We excluded studies involving other types of encephalitis, as well as cases of autoimmune encephalitis without confirmed anti-NMDA receptor antibodies in CSF or serum, and cases involving individuals under 18 years old. Excluding patients under 18 avoids heterogeneity due to distinct pathophysiology, syndromes, and clinical data, ensuring a focused and consistent analysis of adult post-COVID encephalitis.

Study Selection: The selection of articles followed a three-step process. First, duplicate articles were omitted. The titles and abstracts of the remaining articles were screened by two reviewers independently. Next, the full texts of the selected studies were reviewed in accordance with the eligibility criteria. Any disagreements were resolved by a third reviewer. Data extraction was then performed independently by two reviewers. A PRISMA flow chart of the search process is shown in Figure 1.

• Data Extraction: The following data points were systematically collected from each eligible study: Study ID, patient demographics (age, sex, past medical history), COVID-19 test results, interval between COVID-19 presentation and onset of encephalitis, neurological and psychiatric symptoms, neuroimaging and EEG findings, serum and CSF analysis, treatment modalities, and patient outcomes. Two reviewers independently extracted the data using a standardized form. Discrepancies were resolved through discussion or consultation with a third reviewer. The data were then compiled into a comprehensive dataset for further analysis.

Data Synthesis and Analysis: We synthesized the extracted data to categorize qualitative and quantitative data points for comprehensive analysis. Using IBM SPSS statistics 27 (IBM Corp., Armonk, NY, USA), we applied the following methodologies:

Qualitative Data: Summarized as percentages to identify trends and patterns.

Quantitative Data: Summarized as mean ± standard deviation (SD) to understand data distribution.

This combined approach provided an overview of the clinical features, diagnostic approaches, treatments, and outcomes of anti-NMDAR encephalitis following SARS-CoV-2 infection.

2.2. Case Series

Setting and Patients: Eight patients with post-COVID-19 anti-¹NMDAR encephalitis were identified and treated at Loghman Hakim Hospital, Tehran, Iran. All patients were ≥18 years old with confirmed anti-NMDAR antibodies in CSF and/or serum.

Data Collection: For each patient, the following clinical data were collected from medical records: Demographics and past medical history, COVID-19 diagnostic confirmation, Interval between COVID-19 presentation and onset of encephalitis, Neurological and psychiatric manifestations, neuroimaging and EEG findings, CSF and serum findings, treatment modalities and patient outcomes.

Ethical Considerations: The case series component was conducted in compliance with institutional ethical standards. Ethical approval was obtained from the Shahid Beheshti University of Medical Sciences Research Ethics Committee (approval number: IR.SBMU.RETECH.REC.1403.864). Written informed consent was obtained from patients or their legal guardians for inclusion in the study and publication of anonymized clinical data.

3. Results

3.1. Study Selection

A systematic search of PubMed, Scopus, Web of Science, and Embase initially identified 817 potentially relevant records. After removing 261 duplicate records, 556 articles were screened based on titles and abstracts. Of these, 489 records were excluded for not meeting the eligibility criteria, leaving 67 reports sought for full-text retrieval. One report could not be retrieved, resulting in 66 full-text articles assessed for eligibility. Following detailed evaluation against the inclusion and exclusion criteria, 51 reports were excluded for reasons including: pediatric population (under 18 years), absence of confirmed anti-NMDAR antibodies in CSF or serum, non-SARS-CoV-2-associated encephalitis, or insufficient clinical data for extraction. Ultimately, 15 studies met all eligibility criteria and were included in the final scoping review and pooled analysis. A PRISMA flow diagram summarizing the study selection process is presented in Figure 1.

3.2. Demographic Characteristics

The study included 23 adult patients with anti-NMDAR encephalitis temporally associated with SARS-CoV-2 infection, comprising 15 cases identified from the literature (65.2%) and eight newly reported cases from Loghman Hospital, Tehran, Iran (34.8%). Ages ranged from 18 to 76 years, with a mean of 35.6 ± 16.4 years and a median of 30 years (IQR: 22.5–47.0), with a female predominance (60.9%, n = 14) compared with males (39.1%, n = 9).

Thirteen patients (56.5%) had at least one comorbidity, with the remainder having no significant past medical history. The most frequent comorbid conditions were psychiatric or substance-use disorders (5/23, 21.7%), cardiovascular disease such as hypertension or prior myocardial infarction (3/23, 13.0%), autoimmune diseases such as systemic lupus erythematosus and p-ANCA vasculitis (2/23, 8.7%), and prior neurologic conditions such as epilepsy or stroke (2/23, 8.7%). Other reported conditions included oncologic/paraneoplastic history (e.g., ovarian teratoma, breast cancer in remission) (2/23, 8.7%), chronic infections (e.g., hepatitis C, prior varicella zoster virus) (8.7%), metabolic disorders such as diabetes mellitus (1/23, 4.3%), renal disease such as nephrotic syndrome (1/23, 4.3%), and pregnancy-related complications (1/23, 4.3%). A specific paraneoplastic trigger, ovarian teratoma, was identified in one case (4.3%) (See Table 3).

Table 1. Characteristics of the Cases sourced from the literature.

Study ID	Age, Sex	Relevant PMH	COVID test	Onset	Key neuro/psychiatric features	Imaging	EEG	Serum and CSF analysis	Treatment	outcome
Allahyari, F. et al. [27]	18, F	—	RT-PCR +	Simultaneous	AMS, GTC seizures, meningismus	CT/MRI: generalized edema	-	CSF + anti-NMDAR; SARS-CoV-2 PCR +	IVIG, IVMP, oral steroids	Full recovery
Álvarez Bravo et al. [28]	30, F	Left ovarian teratoma	RT-PCR +	3 days	Psychomotor agitation, hallucinations, focal → GTC seizures,	MRI: Lt hippocampal hyperintensity	Lt frontotemporal epileptic discharges; delta brush	CSF + anti-NMDAR	IVIG, IVMP, AEDs, rituximab	Partial recovery
O.V. Ulyanova et al. [29]	22, F	Hepatitis C	RT-PCR +	9 days	Psychosis, catatonia → GTC seizures	CT/MRI: normal	Diffuse abnormal potentials	Serum & CSF + anti-NMDAR	IVMP, plasmapheresis, IVIG, AEDs	Death (cerebral edema, pneumonia)
Gabriele Melegari et al. [30]	31, F	—	COVID IgG +	20 days	Anosmia and Ageusia	MRI: olfactory nucleus hyperintensity	—	Serum + MAG and NMDA Abs	Supportive, vitamin C	Partial recovery
T. Hainmueller et al. [31]	18, M	—	Rapid antigen +	5 weeks	Seizure, confusion, psychosis, catatonia	CT/MRI: normal	Diffuse polymorphic slowing	CSF + anti-NMDAR	AEDs, IVMP, IVIG, rituximab	Full recovery
Caitlin N et al. [32]	27, M	Mood/psychotic disorder	RT-PCR +	simultaneous	Racing thoughts, hallucinations, suicidal ideation, seizures	MRI: normal	Normal	CSF + anti-NMDAR	IVMP, plasmapheresis, rituximab	Full recovery
Lee H. et al. [33]	21, F	—	RT-PCR +	3 days	Short-term memory loss, abnormal behavior	MRI: cerebellar & hippocampal lesions	Diffuse beta ± sharp waves temporal	CSF & serum + anti-NMDAR; lymphocytic pleocytosis	Acyclovir, IVMP, IVIG	Partial recovery
McHattie et al. [34]	53, F	Breast Ca (remission), depression	RT-PCR +	2 weeks	Palilalia, echolalia, dysautonomia, focal seizures	MRI: Lt amygdala & putamen hyperintensity	Slow activity, no epileptiform discharges	CSF + anti-NMDAR, lymphocytic pleocytosis	HCQ, IVIG, tocilizumab, AEDs	Partial recovery
Monti et al. [35]	50, M	HTN	RT-PCR +	3 days	Refractory status epilepticus, orofacial dyskinesia	MRI: Normal	Delta brush; periodic theta	CSF: pleocytosis, elevated IL-6, oligoclonal bands; + anti-NMDAR	IVMP, IVIG, plasmapheresis, AEDs	Full recovery
Naidu k. et al. [36]	50, F	—	RT-PCR +	2 weeks	Progressive weakness, confusion, seizures (two admissions)	MRI: multifocal white-matter & corticospinal hyperintensities	—	CSF + anti-NMDAR; elevated CSF IgG	IVMP, IVIG, plasmapheresis, oral steroids	Partial recovery
Panariello et al. [37]	23, M	Drug abuse (THC, cocaine, PCP)	RT-PCR +	simultaneous	Agitation, psychosis, dyskinesia, autonomic failure	CT: normal	Theta activity 6 Hz	CSF: pleocytosis, IL-6 ↑; + anti-NMDAR	IVMP, IVIG	Full recovery
Sanchez-Larsen [38]	22, F	Focal epilepsy (non-lesional)	RT-PCR +	5 days	Seizure → aphasia, psychosis, insomnia	MRI: normal	Frontal intermittent rhythmic delta activity	CSF & serum + anti-NMDAR	Benzodiazepines, antipsychotics, IVMP, IVIG, rituximab	Full recovery
João Moura et al. [39]	76, M	CVA, MI, pANCA vasculitis, prior VZV	RT-PCR +	simultaneous	Altered mental status, receptive aphasia, meningismus	MRI: old ischemic lesions + recent infarct	—	CSF & serum + anti-NMDAR	IVIG, IVMP, acyclovir	Death (resp. failure)
Mestre Fusco et al. [40]	30, F	—	RT-PCR +	3 days	Agitation, dysarthria, hallucinations	MRI: left hemispheric/hippocampal hyperintensity	Lt frontotemporal epileptiform discharges, delta brush	CSF & serum + anti-NMDAR	IVMP, IVIG, rituximab, AEDs	partial recovery
Valadez-Calderon et al. [41]	28, M	—	RT-PCR +	2 weeks	Catatonia, status epilepticus, hallucinations	MRI: bilateral anterior cingulate cortex & temporal hyperintensities	Subcortical dysfunction on EEG	CSF + anti-NMDAR & GAD65	IVIG, IVMP	partial recovery

Table notes: *Onset = interval from COVID diagnosis to encephalitis presentation (as reported). Abbreviations: AMS = altered mental state; GTC = generalized tonic-clonic; AEDs = anti-epileptic drugs; IVMP = intravenous methylprednisolone; IVIG = intravenous immunoglobulin; CT = computed tomography; MRI = magnetic resonance imaging.

Table 1. Characteristics of the Cases sourced from the literature.

Study ID	Age, Sex	Relevant PMH	COVID test	Onset	Key neuro/psychiatric features	Imaging	EEG	Serum and CSF analysis	Treatment	outcome
Allahyari, F. et al. [27]	18, F	—	RT-PCR +	Simultaneous	AMS, GTC seizures, meningismus	CT/MRI: generalized edema	-	CSF + anti-NMDAR; SARS-CoV-2 PCR +	IVIG, IVMP, oral steroids	Full recovery
Álvarez Bravo et al. [28]	30, F	Left ovarian teratoma	RT-PCR +	3 days	Psychomotor agitation, hallucinations, focal → GTC seizures,	MRI: Lt hippocampal hyperintensity	Lt frontotemporal epileptic discharges; delta brush	CSF + anti-NMDAR	IVIG, IVMP, AEDs, rituximab	Partial recovery
O.V. Ulyanova et al. [29]	22, F	Hepatitis C	RT-PCR +	9 days	Psychosis, catatonia → GTC seizures	CT/MRI: normal	Diffuse abnormal potentials	Serum & CSF + anti-NMDAR	IVMP, plasmapheresis, IVIG, AEDs	Death (cerebral edema, pneumonia)
Gabriele Melegari et al. [30]	31, F	—	COVID IgG +	20 days	Anosmia and Ageusia	MRI: olfactory nucleus hyperintensity	—	Serum + MAG and NMDA Abs	Supportive, vitamin C	Partial recovery
T. Hainmueller et al. [31]	18, M	—	Rapid antigen +	5 weeks	Seizure, confusion, psychosis, catatonia	CT/MRI: normal	Diffuse polymorphic slowing	CSF + anti-NMDAR	AEDs, IVMP, IVIG, rituximab	Full recovery
Caitlin N et al. [32]	27, M	Mood/psychotic disorder	RT-PCR +	simultaneous	Racing thoughts, hallucinations, suicidal ideation, seizures	MRI: normal	Normal	CSF + anti-NMDAR	IVMP, plasmapheresis, rituximab	Full recovery
Lee H. et al. [33]	21, F	—	RT-PCR +	3 days	Short-term memory loss, abnormal behavior	MRI: cerebellar & hippocampal lesions	Diffuse beta ± sharp waves temporal	CSF & serum + anti-NMDAR; lymphocytic pleocytosis	Acyclovir, IVMP, IVIG	Partial recovery
McHattie et al. [34]	53, F	Breast Ca (remission), depression	RT-PCR +	2 weeks	Palilalia, echolalia, dysautonomia, focal seizures	MRI: Lt amygdala & putamen hyperintensity	Slow activity, no epileptiform discharges	CSF + anti-NMDAR, lymphocytic pleocytosis	HCQ, IVIG, tocilizumab, AEDs	Partial recovery
Monti et al. [35]	50, M	HTN	RT-PCR +	3 days	Refractory status epilepticus, orofacial dyskinesia	MRI: Normal	Delta brush; periodic theta	CSF: pleocytosis, elevated IL-6, oligoclonal bands; + anti-NMDAR	IVMP, IVIG, plasmapheresis, AEDs	Full recovery
Naidu k. et al. [36]	50, F	—	RT-PCR +	2 weeks	Progressive weakness, confusion, seizures (two admissions)	MRI: multifocal white-matter & corticospinal hyperintensities	—	CSF + anti-NMDAR; elevated CSF IgG	IVMP, IVIG, plasmapheresis, oral steroids	Partial recovery
Panariello et al. [37]	23, M	Drug abuse (THC, cocaine, PCP)	RT-PCR +	simultaneous	Agitation, psychosis, dyskinesia, autonomic failure	CT: normal	Theta activity 6 Hz	CSF: pleocytosis, IL-6 ↑; + anti-NMDAR	IVMP, IVIG	Full recovery
Sanchez-Larsen [38]	22, F	Focal epilepsy (non-lesional)	RT-PCR +	5 days	Seizure → aphasia, psychosis, insomnia	MRI: normal	Frontal intermittent rhythmic delta activity	CSF & serum + anti-NMDAR	Benzodiazepines, antipsychotics, IVMP, IVIG, rituximab	Full recovery
João Moura et al. [39]	76, M	CVA, MI, pANCA vasculitis, prior VZV	RT-PCR +	simultaneous	Altered mental status, receptive aphasia, meningismus	MRI: old ischemic lesions + recent infarct	—	CSF & serum + anti-NMDAR	IVIG, IVMP, acyclovir	Death (resp. failure)
Mestre Fusco et al. [40]	30, F	—	RT-PCR +	3 days	Agitation, dysarthria, hallucinations	MRI: left hemispheric/hippocampal hyperintensity	Lt frontotemporal epileptiform discharges, delta brush	CSF & serum + anti-NMDAR	IVMP, IVIG, rituximab, AEDs	partial recovery
Valadez-Calderon et al. [41]	28, M	—	RT-PCR +	2 weeks	Catatonia, status epilepticus, hallucinations	MRI: bilateral anterior cingulate cortex & temporal hyperintensities	Subcortical dysfunction on EEG	CSF + anti-NMDAR & GAD65	IVIG, IVMP	partial recovery

Table notes: *Onset = interval from COVID diagnosis to encephalitis presentation (as reported). Abbreviations: AMS = altered mental state; GTC = generalized tonic-clonic; AEDs = anti-epileptic drugs; IVMP = intravenous methylprednisolone; IVIG = intravenous immunoglobulin; CT = computed tomography; MRI = magnetic resonance imaging.

Table 2. Characteristics of the eight new cases from Loghman hospital, Tehran, Iran.

Age and Sex	Past Medical History	COVID-19 Test	Onset of Encephalitis from COVID-19 Presentation	Neurological and Psychiatric Symptoms	Imaging	EEG	Serum and CSF Analysis	Treatment	Outcome
58, F	DM; bipolar	RT-PCR +	1 week	Fluctuating consciousness, delirium	MRI: normal	Focal Lt epileptic discharge; moderate encephalopathy	Serum + NMDAR Ab	IVMP, plasmapheresis, rituximab	Partial recovery
66, F	HTN	RT-PCR +	10 days	Recurrent transient Rt hemiparesis, Aphasia	MRI: nonspecific periventricular foci; MRA: left fetal PCA	Focal Lt epileptic discharge	Serum + NMDAR Ab	IVMP; levetiracetam	Full recovery
32, F	—	RT-PCR +	3 weeks	Disorientation, delirium, frontal headache	MRI: bilateral hippocampal signal changes; limbic encephalitis pattern	Focal Rt temporal discharge	Serum + NMDAR Ab	Acyclovir, IVMP, levetiracetam	Full recovery
23, M	Nephrotic syndrome; SLE	RT-PCR +	4 weeks	Ataxia, dysarthria, vertigo	MRI/MRA: normal	Normal	Serum + NMDAR Ab	IVMP, rituximab	Full recovery
42, F	Pregnant; molar pregnancy	RT-PCR +	4 weeks	Parkinsonism signs, oromandibular dyskinesia, mood fluctuation	MRI: bilateral sphenoid sinusitis	Mild diffuse encephalopathy	Serum + NMDAR Ab	IVMP, levodopa	Partial recovery
18, F	Bipolar disorder	RT-PCR +	2 weeks	GTC seizures, myoclonus, mood disorder	MRI: Rt hippocampus atrophy	Focal Rt epileptic discharge	Serum + NMDAR Ab (2x)	IVMP, valproate, levetiracetam, quetiapine, lithium	Full recovery
37, M	—	RT-PCR +	2 months	Progressive rigidity, spasticity	MRI: normal	Mild diffuse encephalopathy	Serum + NMDAR Ab and GAD65	Clonazepam	Partial recovery
44, M	—	RT-PCR +	3 weeks	Ataxia, choreiform movements (Rt UL)	MRI: multiple high T2/FLAIR WM lesions (supratentorial & infratentorial)	Mild diffused encephalopathy	Serum + NMDAR Ab	IVMP	Partial recovery

Table 2. Characteristics of the eight new cases from Loghman hospital, Tehran, Iran.

Age and Sex	Past Medical History	COVID-19 Test	Onset of Encephalitis from COVID-19 Presentation	Neurological and Psychiatric Symptoms	Imaging	EEG	Serum and CSF Analysis	Treatment	Outcome
58, F	DM; bipolar	RT-PCR +	1 week	Fluctuating consciousness, delirium	MRI: normal	Focal Lt epileptic discharge; moderate encephalopathy	Serum + NMDAR Ab	IVMP, plasmapheresis, rituximab	Partial recovery
66, F	HTN	RT-PCR +	10 days	Recurrent transient Rt hemiparesis, Aphasia	MRI: nonspecific periventricular foci; MRA: left fetal PCA	Focal Lt epileptic discharge	Serum + NMDAR Ab	IVMP; levetiracetam	Full recovery
32, F	—	RT-PCR +	3 weeks	Disorientation, delirium, frontal headache	MRI: bilateral hippocampal signal changes; limbic encephalitis pattern	Focal Rt temporal discharge	Serum + NMDAR Ab	Acyclovir, IVMP, levetiracetam	Full recovery
23, M	Nephrotic syndrome; SLE	RT-PCR +	4 weeks	Ataxia, dysarthria, vertigo	MRI/MRA: normal	Normal	Serum + NMDAR Ab	IVMP, rituximab	Full recovery
42, F	Pregnant; molar pregnancy	RT-PCR +	4 weeks	Parkinsonism signs, oromandibular dyskinesia, mood fluctuation	MRI: bilateral sphenoid sinusitis	Mild diffuse encephalopathy	Serum + NMDAR Ab	IVMP, levodopa	Partial recovery
18, F	Bipolar disorder	RT-PCR +	2 weeks	GTC seizures, myoclonus, mood disorder	MRI: Rt hippocampus atrophy	Focal Rt epileptic discharge	Serum + NMDAR Ab (2x)	IVMP, valproate, levetiracetam, quetiapine, lithium	Full recovery
37, M	—	RT-PCR +	2 months	Progressive rigidity, spasticity	MRI: normal	Mild diffuse encephalopathy	Serum + NMDAR Ab and GAD65	Clonazepam	Partial recovery
44, M	—	RT-PCR +	3 weeks	Ataxia, choreiform movements (Rt UL)	MRI: multiple high T2/FLAIR WM lesions (supratentorial & infratentorial)	Mild diffused encephalopathy	Serum + NMDAR Ab	IVMP	Partial recovery

Table 3. Demographic and baseline characteristics of included anti-NMDAR encephalitis cases following COVID-19 (n = 23).

Characteristic	Summary
Age, years	Mean ± SD: 35.6 ± 16.4; Median (IQR): 30 (22.5–47.0); Range: 18–76
Sex	Female: 14/23 (60.9%); Male: 9/23 (39.1%)
Source of cases	Literature cases: 15/23 (65.2%); New institutional cases (Loghman): 8/23 (34.8%)
Any comorbidity*	13/23 (56.5%)
• Psychiatric or substance-use disorder	5/23 (21.7%)
• Cardiovascular (e.g., hypertension, prior MI)	3/23 (13.0%)
• Autoimmune disease (e.g., SLE, p-ANCA vasculitis)**	2/23 (8.7%)
• Neurologic history (e.g., epilepsy, prior stroke)**	2/23 (8.7%)
• Oncologic/paraneoplastic history (e.g., ovarian teratoma, breast cancer remission)**	2/23 (8.7%)
• Metabolic (e.g., diabetes)	1/23 (4.3%)
• Renal (e.g., nephrotic syndrome)	1/23 (4.3%)
• Pregnancy-related (e.g., molar pregnancy)	1/23 (4.3%)
• Chronic infectious (e.g., hepatitis C, prior VZV)	2/23 (8.7%)
Specific paraneoplastic trigger (e.g, Ovarian teratoma)	1/23 (4.3%)

Notes: Percentages use n = 23 as the denominator. Categories are not mutually exclusive; one case may contribute to multiple rows, so subtotals exceed 100%. *Examples shown in parentheses; full details appear in the individual case tables.

Table 3. Demographic and baseline characteristics of included anti-NMDAR encephalitis cases following COVID-19 (n = 23).

Characteristic	Summary
Age, years	Mean ± SD: 35.6 ± 16.4; Median (IQR): 30 (22.5–47.0); Range: 18–76
Sex	Female: 14/23 (60.9%); Male: 9/23 (39.1%)
Source of cases	Literature cases: 15/23 (65.2%); New institutional cases (Loghman): 8/23 (34.8%)
Any comorbidity*	13/23 (56.5%)
• Psychiatric or substance-use disorder	5/23 (21.7%)
• Cardiovascular (e.g., hypertension, prior MI)	3/23 (13.0%)
• Autoimmune disease (e.g., SLE, p-ANCA vasculitis)**	2/23 (8.7%)
• Neurologic history (e.g., epilepsy, prior stroke)**	2/23 (8.7%)
• Oncologic/paraneoplastic history (e.g., ovarian teratoma, breast cancer remission)**	2/23 (8.7%)
• Metabolic (e.g., diabetes)	1/23 (4.3%)
• Renal (e.g., nephrotic syndrome)	1/23 (4.3%)
• Pregnancy-related (e.g., molar pregnancy)	1/23 (4.3%)
• Chronic infectious (e.g., hepatitis C, prior VZV)	2/23 (8.7%)
Specific paraneoplastic trigger (e.g, Ovarian teratoma)	1/23 (4.3%)

Notes: Percentages use n = 23 as the denominator. Categories are not mutually exclusive; one case may contribute to multiple rows, so subtotals exceed 100%. *Examples shown in parentheses; full details appear in the individual case tables.

3.3. COVID-19 Diagnosis and Onset of Neurological Symptoms

Regarding COVID-19 diagnostic testing, the majority of patients (91.3%, n = 21) were confirmed positive by RT-PCR, while one patient (4.3%) was diagnosed via COVID-19 IgG serology and another (4.3%) by rapid antigen testing (RAT). The interval from COVID-19 diagnosis to encephalitis onset had a median of 10 days (IQR 3–21 days; range 0–60 days). Four patients (17.4%) presented with encephalitic symptoms simultaneous with their COVID-19 presentation. Twenty-one patients (91.3%) developed encephalitis within the first month after COVID-19 diagnosis; two patients (8.7%) had onset beyond 1 month (Figure 2; Table 4).

3.4. Neurological and Psychiatric Presentations

Neurological and neuropsychiatric manifestations were highly prevalent among the patients. Overall, 91.3% (n=21/23) experienced at least one neuropsychiatric symptom. Psychosis and hallucinations were reported in 56.5% (n=13), while agitation or catatonia occurred in 43.5% (n=10). Behavioral or cognitive disturbances, including memory and executive function impairment, were present in 52.2% (n=12). Seizures were documented in 52.2% (n=12), most commonly generalized tonic–clonic events, although focal seizures were also observed. Movement disorders, such as dyskinesias, rigidity, chorea, or parkinsonism, were identified in 34.8% (n=8). Speech dysfunction, including mutism, dysarthria, or broader language impairments, was noted in 26.1% (n=6). Autonomic dysfunction, primarily dysautonomia and hypoventilation, was less frequent 17.4% ( n=4). Altered levels of consciousness, ranging from confusion and delirium to coma, were observed in 43.5% (n=10). Additional neurological features, including ataxia, headache, weakness, and sensory changes, were present in 30.4% (n=7) (Table 5).

3.5. Neuroimaging and EEG Findings

Neuroimaging (brain MRI or CT) was performed in 22 patients (92%), of whom 59.1% (n=13) demonstrated abnormalities, while 40.9% (n=9) had normal or nonspecific findings. The most common abnormalities were temporal/hippocampal involvement in 27.3% (n=6), followed by multifocal white-matter lesions in 9.1% (n=2), with other findings including generalized edema, hippocampal atrophy, and atypical changes such as sinusitis or amygdala involvement. In contrast, 40.9% (n=9) of cases had normal or nonspecific findings (Table 6). Electroencephalography (EEG) was performed in 21 patients, revealing abnormalities in 76.2% (n=16), while 23.8% (n=5) had normal recordings. The most frequent abnormal pattern was diffuse slowing/encephalopathy in 42.9% (n=9), followed by epileptiform discharges in 38.1% (n=8), followed by a smaller subset (14.3%, n=3) exhibited the characteristic delta brush pattern, while polymorphic or focal slowing was less commonly observed.

3.6. CSF and Serum Analysis

Anti-NMDAR antibodies were detected in CSF in 14/23 patients (60.9%) and in serum in 14/23 patients (60.9%); both CSF and serum were positive in 5/23 (21.7%). Among additional CSF parameters, pleocytosis was observed in 34.8% (n=8), typically with a lymphocytic predominance. Elevated protein levels were found in 26.1% (n=6), and oligoclonal bands (OCBs) were present in 13.0% (n=3). Glucose abnormalities were not consistently detected. (See Table 6). Two patients showed concurrent presence of anti-NMDA-R and anti-GAD65 antibodies: one in cerebrospinal fluid (CSF) and one in serum ([41] and one of our new cases). Additionally, one patient exhibited co-occurrence of anti-NMDA-R and anti-MAG antibodies in CSF [30].

3.7. Treatment Approaches

Most patients received first-line immunotherapy, most commonly intravenous methylprednisolone (IVMP) in 87.0% (n=20/23), followed by intravenous immunoglobulin (IVIG) in 56.5% (n=13/23), and plasmapheresis (PLEX) in 34.8% (n=8/23). Among those receiving IVMP, the average dosage and duration were approximately 770.83 ± 254.49 mg/day over 4.08 ± 1.02 days, while IVIG was administered at roughly 2 g/kg over 4.54 ± 0.51 days. PLEX sessions averaged 2.46 ± 0.51 liters per session over 4.46 ± 0.51 sessions (reported where available).

Second-line therapy with rituximab was used in 30.4% (n=7/23) of patients. Cyclophosphamide was not used in any cases, and tocilizumab was administered in 4.3% (n=1/23).

Adjunctive therapies included antiepileptic drugs (AEDs) in 47.8% (n=11/23), psychotropic medications in 26.1% (n=6/23), antiviral or antibiotic therapy (e.g., acyclovir, hydroxychloroquine) in 17.4% (n=4/23), and tumor removal (teratoma) in 4.3% (n=1/23) (See Table 7).

3.8. Clinical Outcomes

Among the patients, 52.2% (n=12/23) achieved full recovery, defined as complete resolution of neurological and psychiatric symptoms with no residual deficits. Partial recovery was observed in 39.1% (n=9/23), characterized by substantial improvement but with lingering symptoms such as mild cognitive impairment, mood disturbances, or ongoing rehabilitation needs. Death occurred in 8.7% (n=2/23) of patients, resulting from cerebral edema and respiratory failure in the context of viral pneumonia [29,39]. The median follow-up duration was 3 months, ranging from hospital discharge to 12 months. Most patients with partial recovery required ongoing neurorehabilitation or medication adjustments to manage residual neurological or psychiatric symptoms (See Table 8).

4. Discussion

4.1. Summary of Key Findings

In this systematic review plus case series of 23 adult patients with anti-NMDAR encephalitis temporally associated with SARS-CoV-2 infection (15 literature cases + 8 new cases from Loghman our center), we identified a consistent clinical phenotype and several reproducible epidemiologic and diagnostic patterns. Our study had a mean age of 35.6 ± 16.4 years (median 30 years) and a female predominance (60.9%). SARS-CoV-2 infection was confirmed by RT-PCR in the large majority (91.3%). The median interval between COVID-19 diagnosis and encephalitis onset was 10 days (IQR 3–21), and 91.3% of patients developed encephalitis within one month of infection. Neuropsychiatric features were nearly universal (91.3%), with psychosis/hallucinations (56.5%), behavioral/cognitive changes (52.2%), seizures (52.2%), movement disorders (34.8%) and altered consciousness (43.5%) commonly observed. MRI was abnormal in 59.1% of cases (most commonly temporal/hippocampal changes), while EEG showed abnormalities in 76.2% (diffuse slowing, epileptiform activity and, less commonly, delta brush). Anti-NMDAR antibodies were detected in CSF and serum in 60.9% each (both positive in 21.7%); pleocytosis and elevated protein were present in a minority. First-line immunotherapies (IV methylprednisolone 87.0%, IVIG 56.5%, plasmapheresis 34.8%) were commonly used and rituximab was employed in 30.4% as second-line therapy. At a median follow-up of 3 months, full recovery occurred in 52.2% of patients, partial recovery in 39.1%, and death in 8.7%.

4.2. Pathophysiological Mechanisms

Several pathophysiological mechanisms have been implicated in the development of COVID-19-associated anti-NMDAR encephalitis, including molecular mimicry, systemic inflammation leading to cytokine storm, blood-brain barrier (BBB) disruption, and dysregulation of microRNAs (miRNAs) [42]. In molecular mimicry, SARS-CoV-2 non-structural proteins such as NSP8 and NSP9 exhibit structural similarities to NMDA receptor (NMDAR) subunits GluN1 (NR1) and GluN2A (NR2A), respectively [12,42]. This resemblance can trigger an immune response where T cells recognize these viral proteins, activating B cells to produce cross-reactive immunoglobulin G (IgG) antibodies that bind to NMDARs. Consequently, these antibodies induce receptor crosslinking and internalization, reducing synaptic NMDAR density and impairing glutamatergic neurotransmission, which manifests as neuropsychiatric symptoms such as altered mental status, seizures, and cognitive deficits [12,42]. Additionally, this process involves bystander T cell activation, particularly IL-6-dependent Th17 cell differentiation, which secretes IL-17 to further promote inflammation and BBB permeability [42].

Systemic inflammation in severe COVID-19 often escalates into a cytokine storm, characterized by elevated levels of pro-inflammatory mediators such as IL-6, IL-1β, tumor necrosis factor-alpha (TNF-α), and interferon-gamma (IFN-γ) [42]. These cytokines can cross the BBB and initiate a positive feedback loop, activating resident microglia and astrocytes, leading to reactive gliosis, neuronal damage, and axonal degeneration [42]. IL-6, in particular, plays a central role by altering neuronal and glial function, correlating with disease severity, and exacerbating BBB disruption through endothelial activation, tight junction degradation, and increased vascular permeability [42]. This compromised BBB integrity facilitates the translocation of peripherally generated autoantibodies and immune cells into the central nervous system (CNS), perpetuating autoimmune encephalitis [42].

Furthermore, alterations in miRNA expression provide additional insights into the overlapping pathophysiology [33,43]. Several miRNAs upregulated in COVID-19, such as miR-107 (linked to neuroendocrine pathways), miR-26b (associated with immune regulation in ovarian teratoma-related encephalitis), and miR-155 (involved in inflammatory JAK/STAT and RAS signaling), have also been implicated in classical anti-NMDAR encephalitis, where they modulate gene expression, immune responses, and neuronal homeostasis [43]. Other shared miRNAs include miR-29b, let-7a, let-7f, and miR-21, which may contribute to autoimmune triggers by influencing pathways like molecular mimicry or cytokine production [43]. However, the notable absence of let-7b—a primary downregulated biomarker in idiopathic or paraneoplastic anti-NMDAR encephalitis, known for its role in neuronal differentiation and immune suppression—suggests that COVID-associated cases may involve distinct or incomplete mechanisms [43]. This discrepancy could arise from SARS-CoV-2’s unique ability to hijack host miRNAs or encode viral miRNA-like molecules (e.g., CoV2-miR-O7a), leading to atypical immune dysregulation and a potentially lower incidence of full-blown encephalitis compared to non-COVID forms [43].

4.3. Comparison with Prior (Non-COVID and COVID) Anti-NMDAR Literature

4.3.1. Demographics

Our study’s mean age (35.6 years) and female predominance (60.9%) align broadly with adult anti-NMDAR literature, which generally reports a young adult skew and strong female predominance [44,45]. However, heterogeneity exists: some COVID-focused collections reported older median ages [46], while others documented younger mean ages and higher female percentages [47,48]. These differences likely reflect variable case ascertainment (case reports vs. institutional series), geographic sampling, and inclusion criteria (adult vs. mixed age cohorts).

4.3.2. COVID-19 Diagnosis and Timing of Encephalitis Onset

The short median latency in our series (10 days; 91.3% within 1 month) is consistent with other COVID-associated case reports and small series, which describe neurologic symptom onset typically within days to a few weeks after SARS-CoV-2 infection [21,48]. This relatively tight temporal clustering supports — but does not prove — a biologically plausible post-infectious or para-infectious relationship that differs from paraneoplastic or delayed post-infectious presentations seen in non-COVID studies.

4.3.3. Neurological and Psychiatric Presentations

The clinical phenotype in our sample (psychiatric disturbance, seizures, movement disorder, altered consciousness) mirrors the classical anti-NMDAR syndrome reported across pre-pandemic and pandemic-era series [44,47,49]. The prevalence of psychosis and behavioral disturbance in our cohort (56.5% and 52.2%, respectively) is comparable to many series, though some COVID case collections documented even higher rates of isolated psychiatric presentations. Seizure frequency in our study (52.2%) is within the range seen in larger meta-analyses [44,50,51], although seizure rates vary across studies depending on age structure and case severity.

4.3.4. EEG Findings

EEG abnormalities were common (76.2%) and were dominated by diffuse slowing/encephalopathy and epileptiform discharges — a pattern reported repeatedly [49,52]. EDB, proposed as a relatively specific marker of anti-NMDAR encephalitis, was identified in 14.3% of our cases; prior estimates of EDB prevalence range widely (≈11–31% in different series — [44,49,50], reflecting differences in disease severity, EEG timing, and interpretative criteria. Importantly, several groups have identified quantitative EEG markers (e.g., beta:delta ratio, BDR) that may enhance specificity [53]. Taken together, our data confirm EEG’s high sensitivity as an early objective test, but they also highlight EDB’s variable occurrence and the need for standardized EEG timing and analytic approaches in COVID-associated cohorts.

4.3.5. Neuroimaging

Approximately 59% of our patients had MRI abnormalities, most frequently involving the limbic system (temporal/hippocampal regions). Additional findings included generalized brain edema, hippocampal atrophy, sinusitis, and multifocal white-matter lesions. This frequency is consistent with prior reports showing a substantial proportion of normal MRIs in anti-NMDAR encephalitis [44,54], and that when present, focal medial temporal and frontal T2/FLAIR hyperintensities are common. The takeaway for clinicians is clear: a normal brain MRI is common and should not delay serologic/CSF testing or empiric therapy when the clinical picture is suggestive.

4.3.6. CSF and Serologic Testing

CSF and serum anti-NMDAR detection rates in our pooled sample (~61% each) appear lower than some larger aggregated datasets reporting higher CSF positivity [44,49]. This discrepancy likely reflects methodological heterogeneity (different cell-based assays or laboratory platforms), timing of sampling relative to disease onset, and incomplete reporting. Pleocytosis and elevated protein were present in roughly a third and a quarter of cases, respectively — again reflecting substantial inter-case variability. A few reports documented concurrent autoantibodies (anti-GAD65, anti-MAG), consistent with episodic broader autoimmune activation in a minority of patients. Notably, a small number of COVID cases had detectable SARS-CoV-2 in CSF [21], but this was uncommon; the significance of CSF viral detection for pathogenesis is unresolved.

4.3.7. Treatment Approaches

Treatment strategies in our cohort paralleled standard non-COVID anti-NMDAR practice: high rates of first-line immunotherapy (IVMP, IVIG, PLEX) and use of rituximab for refractory cases. This resembles patterns reported in [44,48,55], though differences in second-line use and access (e.g., cyclophosphamide, tocilizumab, bortezomib) reflect temporal trends and regional practice variation. Available data do not yet clearly indicate that COVID-associated cases respond differently to standard immunotherapies.

4.3.8. Clinical Outcomes

Outcomes in our cohort (52.2% full recovery, 39.1% partial recovery, 8.7% mortality) fall within the wide outcome range reported pre-pandemic and during COVID reporting; mortality (8.7%) is notable and may be influenced by small numbers and case selection (publication bias toward severe/complicated cases). Several prior analyses have identified predictors of poorer outcome — older age, central hypoventilation, ICU support, and delayed immunotherapy — which remain relevant in interpreting our results [44,49,55]. The short median follow-up (3 months) in our series limits conclusions about long-term recovery and relapse risk.

4.4. Emerging Trends Specific to COVID-Associated Anti-NMDAR Encephalitis

From synthesis of the literature plus our cases, several trends emerge that may help clinicians recognize and manage this entity:

• Tight temporal clustering. Most COVID-associated cases occur within days to weeks of infection (median 10 days; 91.3% within 1 month), supporting at least a temporal link that is plausibly biologically meaningful.
• Preservation of classic phenotype. The hallmark neuropsychiatric syndrome (early psychiatric disturbance → seizures → movement disorder/autonomic features) is preserved, meaning clinicians should apply the same diagnostic vigilance in COVID-exposed patients.
• Diagnostics. CSF and serum anti-NMDAR testing remains central; however, sensitivity varied across reports. EEG abnormalities are common and sometimes the earliest objective clue; MRI is frequently normal or nonspecific. Co-existence of other autoantibodies (e.g., GAD65, MAG) was documented in a minority, suggesting broader autoimmune activation in some patients.
• Therapeutics. First-line immunotherapies (IVMP, IVIG, PLEX) remain the cornerstone; rituximab is used for refractory disease. Treatment patterns broadly mirror non-COVID practice but resource access and timing may influence outcomes.
• Prognosis. While many patients improve, a substantial proportion have residual deficits and a measurable early mortality. Short median follow-up in most reports limits conclusions about long-term cognitive and functional recovery.

4.5. Strengths and Limitations

This work integrates a systematic review with a contemporaneous institutional case series, offering both breadth (literature) and depth (detailed clinical data from eight new cases). The methods followed PRISMA-ScR guidance and used standardized data extraction, allowing quantitative summaries of timing, phenotype, investigations and management.

Several limitations affect interpretation, first, the overall sample is small (n=23), limiting statistical inference and generalizability. Second, the dataset pools heterogeneous sources (case reports/series and a single-center cohort), introducing reporting and selection biases: severe or atypical cases are more likely to be published. Third, diagnostic approaches varied (CSF vs serum testing, assay types) and not all reports included comprehensive inflammatory biomarkers (e.g., IL-6), limiting mechanistic insight. Fourth, follow-up was short (median 3 months), preventing robust conclusions about long-term cognitive and functional outcomes. Finally, confounding by alternative triggers (occult teratoma, other infections) and the possibility of coincidental temporal association cannot be completely excluded in individual reports.

4.6. Alternative Interpretations and Potential Confounders

A cautious interpretation recognizes that temporal association does not prove causation. Anti-NMDAR encephalitis may arise coincidentally in patients who happen to acquire SARS-CoV-2, particularly when community prevalence is high. Conversely, severe systemic illness, ICU admission and hypoxia could non-specifically predispose to neuropsychiatric complications. Variability in antibody assay sensitivity and specificity also raises the possibility of false positives or cross-reactivity. Publication bias (over-reporting of dramatic cases) likely inflates estimates of severity and unusual features.

4.7. Clinical and Public-Health Implications, and Future Directions

Clinically, our synthesis supports maintaining a low threshold to evaluate for anti-NMDAR encephalitis in adults who develop new psychiatric symptoms, seizures, movement disorders or unexplained altered consciousness in close temporal proximity to SARS-CoV-2 infection. Early EEG and CSF analysis (including cell-based antibody assays) and prompt initiation of first-line immunotherapy when clinical suspicion is high may improve outcomes. Tumor screening should still be performed according to standard protocols.

For research and public health, priority next steps include: (1) prospective, multicenter registries to estimate incidence and natural history after COVID-19; (2) standardized case definitions and assay harmonization to reduce measurement heterogeneity; (3) longitudinal follow-up studies to quantify cognitive and psychiatric sequelae; and (4) mechanistic studies (cytokine profiles, molecular mimicry, BBB integrity, miRNA signatures) to clarify causality and identify biomarker-guided treatment strategies. Finally, comparative studies contrasting COVID-associated versus non-COVID anti-NMDAR encephalitis with matched controls would help isolate the unique contribution of SARS-CoV-2.

5. Conclusions

Anti-NMDAR encephalitis occurring after SARS-CoV-2 infection appears to present with the classical neuropsychiatric syndrome and to occur most often within the first month after infection. EEG and CSF antibody testing remain central to diagnosis, although sensitivity varies and MRI may be normal. Most patients respond to established immunotherapies, but a substantial minority have residual deficits or life-threatening complications. Larger, prospective and mechanistic studies are needed to confirm causality, define risk factors, refine diagnostic algorithms (including EEG biomarkers), and optimize treatment strategies for this emerging post-infectious neurologic complication.