Autoimmune encephalitis (AE) is a group of inflammatory brain disorders in which the immune system mistakenly attacks healthy brain tissue, leading to diverse neurological and psychiatric manifestations. Once considered rare, AE has emerged as one of the most common causes of encephalitis worldwide, with an incidence estimated at 10-13 per 100,000 person-years[1]. The condition encompasses a broad spectrum of clinical phenotypes, ranging from limbic encephalitis with prominent memory disturbances to diffuse encephalopathy with seizures, movement disorders, and autonomic dysfunction[2].
The recognition of AE has revolutionized our understanding of previously enigmatic encephalitic syndromes. Since the discovery of anti-NMDA receptor (NMDAR) encephalitis in 2007[3], over 20 distinct autoantibody targets have been identified, each associated with characteristic clinical features, treatment responses, and prognoses. These antibodies target either neuronal surface antigens (NSAbs) or intracellular antigens, with significant implications for pathogenesis and therapeutic approaches[4].
Autoimmune encephalitis affects individuals across all age groups, though certain subtypes show characteristic age distributions. Anti-NMDAR encephalitis predominantly affects young adults, with a median age of 21 years and a strong female predominance (approximately 4:1 ratio), particularly in patients with ovarian teratomas[5]. In contrast, LGI1-associated limbic encephalitis more commonly affects older males, with a median age of 64 years[6].
The overall incidence of AE has increased substantially over the past two decades, partly due to improved diagnostic awareness and testing capabilities. Population-based studies from Europe and North America indicate that AE now accounts for 20-30% of all encephalitis cases, surpassing infectious etiologies in many regions[7]. Approximately 70-80% of AE cases occur in individuals without known predisposing factors, while the remaining cases are associated with underlying neoplasms (paraneoplastic AE), infections (post-infectious AE), or other autoimmune disorders[8].
While AE occurs worldwide, some studies suggest regional variations in incidence and antibody prevalence. The peak incidence appears in winter and spring months for some subtypes, potentially reflecting seasonal infection patterns that trigger post-infectious autoimmunity[1:1]. However, more research is needed to establish definitive seasonal patterns.
Certain human leukocyte antigen (HLA) alleles confer increased susceptibility to specific AE subtypes. HLA-DRB110:01 is strongly associated with anti-IgLON5 disease[9], while HLA-DQB106:02 shows association with anti-NMDAR encephalitis in some populations. These genetic associations suggest underlying immune dysregulation as a predisposing factor[10].
Family history of autoimmune disease is reported in 15-20% of AE patients, with higher rates of autoimmune thyroid disease, type 1 diabetes, and systemic lupus erythematosus. This suggests a shared genetic susceptibility to immune dysregulation[11].
Gender distribution varies by AE subtype:
These gender differences likely reflect hormonal influences on immune function and differential tumor associations.
The pathogenesis of autoimmune encephalitis involves distinct mechanisms depending on the antibody target. Antibodies against neuronal surface antigens typically cause reversible neuronal dysfunction through several mechanisms:
Internalization and modulation: Anti-NMDAR antibodies bind to the GluN1 subunit, causing receptor internalization and reduction of surface NMDAR density, leading to impaired synaptic transmission and cognitive deficits[12]
Functional blockade: Some antibodies, such as those targeting LGI1 or GABA_B receptor, interfere with ligand binding or channel function without causing receptor internalization[13]
Complement activation: Antibodies against intracellular antigens (e.g., Hu, Ma2) are thought to mediate cytotoxicity through T-cell mechanisms rather than direct complement-mediated damage[14]
The NMDAR is composed of GluN1 and GluN2 (A-D) subunits. Anti-NMDAR antibodies primarily target the GluN1 subunit, which is essential for receptor function. The binding site localizes to the extracellular domain, and antibody binding can be displaced by exogenous agonists, suggesting competitive inhibition[12:1].
LGI1, a ligand for ADAM22, modulates AMPA receptor function. Anti-LGI1 antibodies disrupt this interaction, leading to impaired synaptic transmission and hyperexcitability. The characteristic faciobrachial dystonic seizures likely arise from disrupted temporal lobe circuitry[15].
Beyond humoral immunity, T-cell-mediated responses contribute significantly to tissue injury in AE. CD8+ cytotoxic T lymphocytes target neurons expressing cognate antigens, particularly in paraneoplastic encephalitis associated with intracellular antigen antibodies[16]. Microglial activation and inflammatory cytokine production create a pro-inflammatory microenvironment that perpetuates neuronal injury[17].
The blood-brain barrier (BBB) plays a critical role in AE pathogenesis. Disruption of the BBB allows peripheral antibodies and immune cells to enter the central nervous system, while also enabling CNS antigens to trigger peripheral immune responses. Studies using dynamic contrast-enhanced MRI have demonstrated BBB leakage in many AE patients, particularly in limbic structures[18].
The inflammatory response in AE involves multiple cell types and signaling pathways:
Tumors expressing neuronal antigens trigger immune responses that cross-react with central nervous system tissues. The immune response involves both cellular and humoral mechanisms, with antibodies serving as markers of the underlying immune dysregulation rather than direct mediators of damage in many paraneoplastic syndromes[20].
The most well-established paraneoplastic association is between ovarian teratomas and anti-NMDAR encephalitis. Teratoma cells contain neuronal tissue that expresses NMDAR, triggering an immune response that generates cross-reactive antibodies. Tumor resection often leads to clinical improvement, supporting this pathogenic mechanism[5:1].
Other paraneoplastic associations include:
Anti-NMDAR encephalitis represents the most common and well-characterized AE subtype. The classic presentation involves a prodromal phase with headache, fever, and viral-like symptoms, followed by psychiatric features including anxiety, agitation, behavioral changes, and psychosis[21]. The psychiatric manifestations can be severe enough to lead to initial psychiatric admission before neurological diagnosis.
Seizures occur in over 80% of patients, often progressing to refractory status epilepticus. Movement disorders are common, particularly orofacial dyskinesias, choreoathetosis, and dystonia. Autonomic dysfunction manifests as tachycardia, hypertension, hyperthermia, and can progress to severe dysautonomia requiring intensive care support[22].
The disease course typically progresses through characteristic stages:
Limbic encephalitis presents with the triad of seizures, memory dysfunction, and psychiatric symptoms. LGI1 encephalitis typically presents with faciobrachial dystonic seizures (FBDS), characterized by brief, unilateral facial and arm contractions, often preceding the development of limbic encephalitis by days to weeks[15:1]. These seizures are highly characteristic and should prompt immediate antibody testing.
Cognitive impairment often includes prominent anterograde amnesia with relative preservation of other cognitive domains. Patients may develop emotional lability and personality changes, reflecting limbic system involvement.
Anti-IgLON5 disease presents as a unique syndrome of progressive insomnia, dysarthria, dysphagia, gait disturbance, and cognitive decline. Polysomnography reveals abnormal sleep architecture with REM sleep behavior disorder. The disease has strong associations with HLA-DRB1*10:01 and may have a neurodegenerative component[9:1].
Beyond the core syndromes, AE can present with:
Seizures in AE are often refractory and may present with various semiologies:
EEG findings correlate poorly with seizure burden, and subclinical seizures may persist despite suppressive therapy.
The 2016 diagnostic criteria for AE proposed by Graus et al. provide a structured approach to diagnosis[18:1]. The criteria require:
A definite diagnosis requires the combination of clinical features with confirmatory antibody testing. A probable diagnosis can be made when clinical features are characteristic but antibody testing is negative or unavailable.
AE must be differentiated from:
Key distinguishing features include subacute onset, CSF inflammation, and characteristic MRI findings.
CSF findings in AE are variable but commonly include:
MRI findings depend on the AE subtype:
FDG-PET can reveal hypometabolism in temporal lobes or diffuse cortical hypermetabolism in anti-NMDAR encephalitis[23:1]. PET imaging may be more sensitive than MRI for detecting early changes.
EEG is abnormal in 90% of AE patients, showing:
Serum and CSF testing should be performed simultaneously, as some antibodies (particularly anti-NMDAR) have higher sensitivity in CSF[24]. Cell-based assays (CBA) are the gold standard for surface antibody detection, with immunohistochemistry and Western blot serving as supplementary techniques.
Antibody categories:
The antibody target influences prognosis and treatment response, making accurate identification crucial.
First-line therapies should be initiated as soon as AE is suspected:
These treatments can be used alone or in combination, with early intervention associated with improved outcomes[11:1]. Combination therapy (corticosteroids + IVIG or plasma exchange) is often recommended for severe cases.
For patients with inadequate response to first-line therapy (persistent symptoms >2-4 weeks):
Second-line therapy is required in approximately 30-50% of anti-NMDAR encephalitis cases[19:1].
In paraneoplastic AE, tumor identification and removal is crucial:
Comprehensive cancer screening should be performed in all AE patients, including:
Aggressive supportive management is often required:
Outcomes vary substantially based on AE subtype and treatment timing:
Anti-NMDAR encephalitis: Approximately 80% achieve good recovery (mRS 0-2) with aggressive immunotherapy, though recovery may take months to years. Relapse occurs in approximately 10-20% of patients[19:2]. Persistent cognitive deficits occur in up to 25% of patients.
Limbic encephalitis (LGI1, GABA_BR): Generally favorable prognosis with early immunotherapy, though cognitive deficits often persist. Approximately 70% achieve functional independence.
Anti-IgLON5 disease: More progressive course, with approximately 30% mortality at 5 years despite immunotherapy. Most patients require ongoing immunosuppressive therapy[9:2].
Paraneoplastic encephalitis with intracellular antigens: Generally poorer outcomes, with permanent neurological deficits common even with aggressive treatment.
Positive prognostic factors include:
Negative prognostic factors include:
Emerging evidence links autoimmune encephalitis to neurodegenerative processes. Some patients with AE develop persistent cognitive deficits resembling frontotemporal dementia or Alzheimer's disease[25]. The long-term cognitive outcomes of AE survivors are an area of active investigation.
Conversely, patients with neurodegenerative diseases may develop AE-like autoimmunity, complicating diagnostic differentiation[26]. The presence of autoantibodies in neurodegenerative diseases may represent:
Recent studies have identified autoantibodies in patients with Alzheimer's disease, Parkinson's disease, and frontotemporal dementia, raising questions about whether autoimmunity contributes to or is a consequence of neurodegeneration.
Anti-IgLON5 disease represents an intriguing intersection between autoimmunity and neurodegeneration. The presence of tau pathology in some patients, combined with HLA-DRB1*10:01 association and progressive course, suggests underlying neurodegeneration triggered or exacerbated by autoimmune mechanisms[9:3].
Autoimmune encephalitis in children has distinct features:
Pediatric AE requires specialized management including developmental support and school reintegration planning.
Animal models have provided crucial insights into AE pathogenesis:
These models enable testing of novel therapeutic strategies before clinical translation.
AE imposes significant healthcare burden:
Early diagnosis and treatment reduce overall costs by shortening hospital stay and improving outcomes.
AE survivors face significant quality of life challenges:
Comprehensive follow-up programs addressing physical, cognitive, and psychological domains are essential.
Current research focuses on:
Incidence of autoimmune encephalitis: a population-based study. JAMA Neurology. 2019. ↩︎ ↩︎
Antibody-mediated encephalitis: a treatable cause of dementia. Current Alzheimer Research. 2017. ↩︎
NMDAR encephalitis: clinical features, diagnosis and therapy. New England Journal of Medicine. 2007. ↩︎
Autoimmune encephalitis: expanding the spectrum of neurological manifestations. Lancet Neurology. 2020. ↩︎
Anti-NMDA receptor encephalitis: clinical features and long-term outcomes. Lancet Neurology. 2016. ↩︎ ↩︎
LGI1 encephalitis: clinical features and treatment outcomes. Brain. 2015. ↩︎
Changing patterns in autoimmune encephalitis: a population-based study. Neurology. 2020. ↩︎
Paraneoplastic autoimmune encephalitis: clinical features and tumor associations. Neurology Neuroimmunology Neuroinflammation. 2021. ↩︎
Anti-IgLON5 disease: a novel tauopathy. Brain. 2015. ↩︎ ↩︎ ↩︎ ↩︎
HLA associations in autoimmune encephalitis. Journal of Neurology Neurosurgery and Psychiatry. 2019. ↩︎
Treatment outcomes in autoimmune encephalitis. Lancet Neurology. 2016. ↩︎ ↩︎
Anti-NMDA receptor antibodies cause receptor internalization. Neuron. 2008. ↩︎ ↩︎ ↩︎
LGI1 antibodies cause limbic dysfunction without receptor internalization. Brain. 2012. ↩︎
T-cell mediated pathogenesis in paraneoplastic encephalitis. Annals of Neurology. 2015. ↩︎
Faciobrachial dystonic seizures precede LGI1 encephalitis. Brain. 2015. ↩︎ ↩︎
CD8+ T-cell cytotoxicity in autoimmune encephalitis. Brain Pathology. 2017. ↩︎
Microglial activation in autoimmune encephalitis. Acta Neuropathologica. 2020. ↩︎ ↩︎
Diagnostic criteria for autoimmune encephalitis: Graus criteria. Lancet Neurology. 2016. ↩︎ ↩︎
Relapse in anti-NMDA receptor encephalitis. Lancet Neurology. 2016. ↩︎ ↩︎ ↩︎
Paraneoplastic autoimmunity: mechanisms and clinical implications. Nature Reviews Neurology. 2015. ↩︎
Psychiatric manifestations of anti-NMDA receptor encephalitis. Lancet Psychiatry. 2014. ↩︎
Autonomic dysfunction in anti-NMDA receptor encephalitis. Lancet Neurology. 2016. ↩︎
FDG-PET and CSF biomarkers in autoimmune encephalitis. Journal of Nuclear Medicine. 2017. ↩︎ ↩︎
CSF vs serum antibody testing in autoimmune encephalitis. Neurology. 2018. ↩︎
Cognitive outcomes in autoimmune encephalitis. JAMA Neurology. 2019. ↩︎
Autoimmunity in neurodegenerative diseases. Nature Reviews Neurology. 2022. ↩︎