[Skip to Navigation]
September 13, 2021

Discerning the Role of Autoimmunity and Autoantibodies in Epilepsy: A Review

Author Affiliations
  • 1NYU Comprehensive Epilepsy Center, New York University, New York
  • 2Mater Advanced Epilepsy Unit, Brisbane, Australia
  • 3The University of Queensland, Brisbane, Australia
  • 4Department of Statistics, QIMR Berghofer Institute, Brisbane, Australia
  • 5Jefferson University, Philadelphia, Pennsylvania
  • 6Mayo Clinic, Rochester, Minnesota
JAMA Neurol. 2021;78(11):1383-1390. doi:10.1001/jamaneurol.2021.3113

Importance  The literature on neural autoantibody positivity in epilepsy has expanded over the last decade, with an increased interest among clinicians in identifying potentially treatable causes of otherwise refractory seizures.

Observations  Prior studies have reported a wide range of neural autoantibody positivity rates among various epilepsy populations, with the highest frequency reported in individuals with focal epilepsy of unknown cause and new-onset seizures. The antibodies in some cases are of uncertain significance, and their presence can cause conundrums regarding therapy.

Conclusions and Relevance  There is likely some role for neural autoantibody assessment in patients with unexplained epilepsy who lack clear evidence of autoimmune encephalitis, but the clinical implications of such testing remain unclear owing to limitations in previous published studies. A framework for study design to bridge the current gaps in knowledge on autoimmune-associated epilepsy is proposed.

Add or change institution
2 Comments for this article
Which patients should be considered non-suspicious of autoimmune encephalitis in autoimmune-associated epilepsy studies?
Pablo Cabezudo-Garcia, MD. | Instituto de Investigación Biomédica de Málaga (IBIMA). UGC Neurociencias; Servicio de Neurología; Unidad de Epilepsia Hospital Regional Universitario de Málaga. Spain.
We agree with the pertinent recommendations made by Steriade et al.¹ for future studies evaluating patients with suspected autoimmune-associated epilepsy (AAE). Some of these concerns had already been pointed out by our group in a systematic review and meta-analysis to determine the pooled prevalence of neural autoantibodies in patients with epilepsy of unknown etiology² where we excluded studies with autoimmune encephalitis (AIE) or only high suspicion of AIE patients.
Future consensus regarding whether seizures are secondary to AIE or high suspicion of AIE exists, and when a patient has possible AAE without encephalitis is advisable. The authors propose to use
a cut-off in the APE2 score as the first step to classify patients as acute symptomatic seizures secondary to AIE or possible AAE without encephalitis. Aware that these scores are “imperfect”, the authors clarify that predictive scores should be modified depending on the population of epilepsy under study, especially in those epilepsy groups with some characteristic associated with AIE such as drug resistance, new onset-seizures or autonomic seizures, and symptoms. Also, to remark that other characteristics associated with the presence of antineuronal antibodies in patients with epilepsy as cognitive symptoms, temporal lobe involvement, or previous history of autoimmune disease³, are not uncommon in epilepsies of any etiology and also appear frequently in AIE. It would be necessary to agree to what extent these characteristics individually or in combination make a patient highly suspect of suffering from AIE.
Another recommendation made by the authors is that cerebrospinal fluid (CSF) determinations should always be carried out if available. Although we fully agree with this assertion, the authors specify that this determination should be made in patients with positive results in serum and in "all individuals at high clinical suspicion". This statement is still contradictory to the previous recommendation of only including patients without high suspicion of AIE, and we believe that carrying out a systematic CSF determination in all patients in a study with prospective recruitment is the most appropriate method, even taking into account the difficulties involved. We consider that patients with epilepsy in which CSF was obtained in routine clinical practice should be considered as patients with clinical suspected AIE.
Finally, the term autoimmune-associated epilepsy was proposed in 2020 by some of the authors of this review as chronic seizures (despite immunotherapy) determined to be secondary to enduring immune factors, post-encephalitic structural injury, or a combination (i.e., anti-GAD65 encephalitis)⁴. Regarding this proposed definition, a different term may be more appropriate for patients with epilepsy of unknown etiology with the presence of antineuronal antibodies and without “high suspicion” of encephalitis until future studies determine if these autoantibodies play a role or are indicative of an immunologic process.

1.Steriade C, et al. Discerning the Role of Autoimmunity and Autoantibodies in Epilepsy: A Review.
2.Cabezudo-García P, et al. Prevalence of Neural Autoantibodies in Epilepsy of Unknown Etiology: Systematic Review and Meta-Analysis.
3.de Bruijn MAAM, et al. Antibodies Contributing to Focal Epilepsy Signs and Symptoms Score.
4. Steriade C, et al. Acute symptomatic seizures secondary to autoimmune encephalitis and autoimmune-associated epilepsy: Conceptual definitions.
Neuronal Antibody Testing in Epilepsy: How to Preselect Patients?
Marienke de Bruijn, MD PhD | Erasmus University Medical Center, Rotterdam, the Netherlands
With many interest we have read this very nice review.¹ Identifying seizures with an autoimmune etiology is an important topic, because rapid identification of patients leads to earlier initiation of immunotherapy, the most effective treatment when compared to anti-seizure medication.²
Several studies reported, due to variable study designs and different selection criteria, widely variable antibody prevalence (3.4%-16.7%), with lowest frequencies in unselected patient cohorts without clear encephalitis and highest prevalence in selected patients with signs of encephalitis.
The authors review 27 publications containing 4067 individuals with seizures, all tested for neuronal antibodies in serum. In only 3.4%
CSF was tested, an expected finding as only patients without overt encephalitis were included in this review. Neuronal antibodies were detected in serum in 405 patients (10%). Neuronal antibodies without clinical significance (low concentration anti-GAD65, double negative anti-VGKC) were excluded (n=167). The majority of the remaining 238 patients (5.9% of the entire cohort) had high-concentration anti-GAD65, followed by anti-GlyR, anti-NMDAR, anti-CASPR2, and anti-LGI1. Caution is however warranted in positive patients lacking confirmation in CSF (NMDAR, GAD65).The authors note that CSF confirmation occurred in only 7% of the positive cases. Luckily, this is probably higher: at least the 20 patients in the ACES study³ were all confirmed, but scored erroneously as ‘not applicable’ according to Table 1. Similarly, critical consideration is warranted in antibodies without alternative test for confirmation available (GlyR) or without additional confirmation performed (GAD65, NMDAR, CASPR2; immunohistochemistry or live hippocampal neurons).
The authors combine two recently created and validated clinical scores (APE2⁴ and ACES³) to preselect patients requiring antibody testing. As the scores were developed independently, and serve different purposes, combination seems the worst of two worlds: the new flowchart is not validated and difficult in its use. The APE2 score was created using patients data including many with overt encephalitis signs, whereas the ACES score was designed using data from patients with focal epilepsy without clear signs of encephalitis. The ACES score is better suited to use in chronic patients (>1year, especially GAD65), while both APE2 and ACES score proved useful in patients with new-onset seizures. Meeting the cut-off in any of both scores is sufficient to recommend screening, while high scores in both should increase suspicion and lead to more extensive testing.

1. Steriade C, Gillinder L, Rickett Ket al. Discerning the Role of Autoimmunity and Autoantibodies in Epilepsy: A Review. JAMA Neurol. 2021 Sep 13.
2. Bruijn MA, Sonderen van A, Coevorden-Hameete van MH, et al. Evaluation of seizure treatment in anti-LGI1, anti-NMDAR, and anti-GABABR encephalitis. Neurology. 2019 May 7;92(19):e2185-e2196.
4. Dubey D, Pittock SJ, McKeon A. Antibody Prevalence in Epilepsy and Encephalopathy score: Increased specificity and applicability. Epilepsia. 2019 Feb;60(2):367-369.
4. Bruijn de MAAM, Bastiaansen AEM, Mojzisova H et al. Antibodies Contributing to Focal Epilepsy Signs and Symptoms Score. Ann Neurol. 2021 Apr;89(4):698-710.

Authors: MAAM de Bruijn, MJ Titulaer.