Autoimmune-associated epilepsy or acute symptomatic seizures? A case series of recurrent seizures in patients with myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD)

Abstract

Acute symptomatic seizures are a well recognized symptom of myelin oligodendrocyte antibody-associated disease (MOGAD), but long-term seizure outcomes and risk of epilepsy are understudied. In a retrospective cohort of 135 consecutive patients meeting consensus criteria for MOGAD without a prior diagnosis of epilepsy or concurrent NMDA-R Ab positivity, 19 developed seizures after MOGAD onset (16 %). Of these 19 patients, 7 patients (37 % of those with seizures, and 5 % of the total cohort). experienced one or more seizure recurrence during MOGAD remission (i.e. outside of an acute attack). Five of the 7 patients with recurrent seizures remained on antiseizure medication (ASM) (four on monotherapy and one on two ASMs) at last follow-up (median duration of follow up: 92 months). We discuss phenotypes of recurrent seizures in patients with MOGAD in the context of the conceptual framework of acute symptomatic seizures versus autoimmune encephalitis associated epilepsy (AEAE).

1. Introduction

Seizures in autoimmune encephalitis (AE) have been conceptualized as either “acute symptomatic seizures” or “autoimmune associated epilepsy”, (Steriade et al., 2020) the latter refined to “autoimmune encephalitis associated epilepsy” (AEAE) (Steriade et al., 2025). While this conceptual distinction has been adopted in the field of AE, (Steriade et al., 2025) its operational definition remains tentative and its relevance to seizures in demyelinating diseases remains unexamined.

Seizures occur in ~20 % of cases (Li et al., 2022) meeting 2023 international consensus criteria for myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), (Banwell et al., 2023) more common than with other demyelinating syndromes such as multiple sclerosis or neuromyelitis optica. Diagnosis requires a core clinical demyelinating event (criterion A), a positive MOG-IgG test (in some cases supported by clinicoradiographic features) (criterion B), and exclusion of better alternative diagnoses (criterion C) (Banwell et al., 2023). Core clinical syndromes fulfilling criterion A include optic neuritis, myelitis, acute disseminated encephalomyelitis (ADEM), cerebral monofocal or polyfocal deficits, brainstem or cerebellar deficits, or cerebral cortical encephalitis (CCE) often with seizures (Banwell et al., 2023). Seizures in MOGAD have often been described at presentation or during relapse, most commonly during ADEM or CCE, with recurrent seizures usually occurring at the time of a subsequent relapse (Montalvo et al., 2022). The significance of seizures occurring during remission, i.e. between MOGAD relapses, and in the absence of explanatory MRI findings, is poorly understood. Closer examination of clinical characteristics of seizures in MOGAD within the “acute symptomatic seizures versus AEAE” conceptual framework may lend insights into the need for long term anti-seizure medication (ASM) and clinical spectrum of signs of active CNS disease prompting changes in immunotherapy (Montalvo et al., 2022; Vaingankar et al., 2025; Gutman et al., 2018).

In this exploratory study, we leveraged a large, deeply phenotyped MOGAD cohort with long term follow-up to characterize the clinical phenotypes of seizures in MOGAD, with attention to temporal relation to core clinical syndromes, radiologic and electroclinical correlations and MOG antibody titer. We focused on phenotypic relationships with duration of ASM, immunotherapy and seizure outcome.

2. Methods

2.1. Study population

We comprehensively reviewed electronic medical records for all patients who were diagnosed with MOGAD by a neuroimmunologist at the NYU MS Center, New York, NY based on consensus criteria (Banwell et al., 2023). Out of 135 patients with MOGAD (mean age 25.5 years +/− sd 17, median age 22 years, 51/135 or 38 % under 18, 61.5 % female), 19 patients had seizures (16 %) and were included in our study. Patients were excluded if they had a prior diagnosis of epilepsy (5 patients), a positive antibody test associated with non-MOGAD autoimmune encephalitis (2 patients) or inadequate documentation of seizure (7 patients).

2.2. Clinical data collection

Treatment course, laboratory, imaging, electrodiagnostic features and outcome data were collected with special emphasis of seizure semiology, frequency; their relation to MOGAD relapses; and EEG data and relevant MRI features. Specifically, attention was paid to timing of the MOGAD relapse in relation to recurrent seizures. A MOGAD relapse was defined clinically according to Criterion A of the most recent proposed consensus criteria (Banwell et al., 2023). ‘Active relapse’ was defined as the period of 3 months from symptom onset or any ongoing clinical and/or radiologic activity.

3. Results

3.1. Overall characteristics of MOGAD-associated seizure cohort

Of the 19 individuals with seizures and MOGAD, 16 were children and 3 were adults (see Table 1, median age 7 years); 13/19 were female (68 %). 5/19 (26 %) had monophasic MOGAD and 14/19 (74 %) experienced at least one more relapse. Seizure onset occurred during the initial clinical attack in 13/19 (68 %) of patients. The presenting core clinical syndrome fulfilling Criterion A of current proposed diagnostic criteria included cerebral lesions in 10 of 19 cases (53 %). Isolated ON occurred at onset in 4 of 19 cases (21 %), and at least once in 11 of 19 cases (58 %). 7 out of 19 patients (37 %) experienced a seizure at a separate time from a MOGAD clinical attack or relapse (Fig. 1). Of these, onset syndromes included meningoencephalitis in four patients, and three of these four had another core syndrome during a relapse. The other three phenotypes with recurrent seizures were recurrent ADEM (n = 1), relapsing ON (n = 1) and monophasic brainstem relapse (n = 1). While seizure onset outside the initial clinical event of MOGAD was less common than seizure during initial MOGAD attack, a similar proportion of such patients experienced recurrent seizures (3/6, 50 %) when compared to the rate of recurrence among patients whose first seizure occurred during the initial event (5/13, 38 %). Eight of 11 patients (73 %) with both ON and seizure during their disease course experienced an ON relapse.

Table 1.

Demographics and electroclinical features.

	All	Seizure/s outside relapse
number of patients	19	7
median age at onset	7.5	6
female sex	12	4
focal onset semiology	14	6
impaired cons	5	2
focal to BTC	9	5
BTC only	5	1
subclinical ever	2	1
initial EEG done/avail	12	7
normal	3	3
epileptiform	3	2
focal slow present	4	2
gen slow only	2	0
LT ASM use > 3 m	11	6
continued at last f/u.	5	5
monotherapy	8	3
2 ASMs	3	3
refractory	0	0
Followup duration (mo)
median	90	80
range	1–480	26–109

Fig. 1.

The clinical course of 7 patients who experienced seizures without clinical or radiographic concurrent evidence of MOGAD relapse are illustrated. In some, seizures occurred outside of a MOGAD syndrome, and in others, the index seizure occurred concurrently with a MOGAD syndrome, representing acute symptomatic seizures, but later recurred without definite clinical or radiographic evidence of MOGAD. Abbreviations: ADEM - acute disseminated encephalomyelitis, CCE – cortical cerebral encephalitis, MDEM – multiphasic disseminated encephalomyelitis, ME – meningoencephalitis, ON – optic neuritis.

Of the 19 patients with seizures, 6 (32 %) never received an ASM, and 4 (21 %) were successfully weaned off ASMs prior to discharge from hospital with acute MOGAD presentation. ASM wean occurred at various intervals thereafter (2 between 3 and 12 months, 2 after 12 months), with 5 (26 %) remaining on ASMs at last follow-up.

3.2. Electroclinical characteristics of MOGAD-associated seizures

14 patients experienced focal semiology, with focal impaired consciousness in 5 patients (26 % of cohort), with or without observable components such as gaze deviation, hand shaking, arm posturing, or facial movements. Of those 14 patients, nine evolved to a focal to bilateral tonic clonic seizure at least once. The remainder of patients (5/19 – 26 %) solely experienced a bilateral tonic clonic activity, of presumed focal onset. Acutely, electroclinical findings were often concordant with acute MRI lesions, for example in the case of a patient with right hemispheric, maximum frontotemporal interictal epileptiform discharges (IEDs), corresponding to MRI FLAIR hyperintensities including the right mesial temporal regions (Fig. 2A). Follow-up EEGs, including prolonged ambulatory or inpatient video-EEGs, were normal in 6/7 patients with recurrent seizures. In one patient, an area of encephalomalacia in the right posterior quadrant was associated with right frontal IEDs, possibly related to network hyperexcitability (Fig. 2B).

Fig. 2.

A. A patient with acute disseminated encephalomyelitis (ADEM) and multifocal lesions including in the right amygdala (white arrows) had right temporal spike wave complexes and acute symptomatic seizures. B. A patient with cortical cerebral encephalitis involving the right parietotemporal region (white arrow) evolved to have volume loss in that region (arrowhead) and remote seizures without MOGAD relapse, with corresponding sharp waves in the right frontal region, suspected due to spread and network parieto-frontal excitability.

3.3. Recurrent seizures as acute symptomatic seizures vs autoimmune associated epilepsy

Fig. 1 outlines the timeline of recurrent seizures in the seven patients in whom recurrent seizures occurred during MOGAD remission; such seizures could not be readily classified as ‘acute symptomatic’ and may therefore represent epilepsy. Three of those patients failed ASM weans early in their disease course, and were continued on ASMs, with rare nondisabling seizures. Two patients experienced breakthrough seizures in the setting of steroid taper, with seizures heralding an optic neuritis attack in two cases. In three patients, MOG titers increased at the time of recurrent seizures, without evidence of a MOGAD syndrome on MRI. Follow-up MRIs were obtained in 5 of the 7 patients with recurrent seizures and were normal in all but one, who harbored gliosis (Fig. 2B). Repeat CSF was not sampled when recurrent seizures occurred without clinical or radiographic evidence of a MOGAD syndrome, precluding confident exclusion of aseptic meningoencephalitis in those patients.

4. Discussion

In our longitudinal cohort of MOGAD patients with seizures, seizures recurred in over one third of patients, a higher frequency of recurrent seizures than previously reported (Montalvo et al., 2022). While a significant proportion of patients will experience seizures outside of an attack and will require ongoing ASMs, their course, while consistent with epilepsy, is more benign than that of autoimmune associated epilepsy after neural autoantibody associated AE, which is frequently drug-resistant (Steriade et al., 2025). We also identified a small subgroup in whom recurrent seizures coincided with a rise in MOG titers with or without a subsequent typical relapse. Thus, recurrent seizures even without a typical clinical presentation may represent acute symptomatic seizures related to active autoimmunity. MRI-negative aseptic meningoencephalitis is increasingly recognized in MOGAD, (Aboseif et al., 2025) though not yet part of consensus criteria, and may present initially with seizures and be underrecognized without appropriate antibody testing without other clear clinical or radiographic evidence of MOGAD relapse.

Since a proportion of patients experienced seizures outside of a core syndrome or new MRI lesions, we posit that a proportion of patients with acute symptomatic seizures and MOGAD may progress to a MOG-related autoimmune associated epilepsy (MOG-AAE). In contrast with neural autoantibody mediated AE and subsequent epilepsy, breakthrough seizures outside of acute relapses in our cohort were isolated and sporadic, in keeping with a post-acute injury pattern seen outside of autoimmunity (Misra et al., 2023).

As reported previously, (Gutman et al., 2018) seizures may be the initial manifestation of MOGAD, particularly meningocortical presentations which we hypothesize lie along a spectrum with cortical encephalitis and may be symptomatic of cortical involvement (Shivarthi et al., 2025; Budhram et al., 2022; Vega et al., 2023). Seizures may precede overt parenchymal changes such as those seen in ADEM, or even optic neuritis. Two of our cases of MOGAD without encephalitic syndromes had index seizures outside an apparent syndrome, suggesting radiographic lag as a potential explanation for MOGAD patients with relapsing ON and seizures separated in time. In other cases of relapsing ON, seizures could immediately precede a frank clinical relapse of ON. This provides a particularly interesting subgroup of patients with relapsing optic neuritis without overt cortical lesions and raises questions on possible subradiographic (by standard 3 T MRI) cortical parenchymal involvement which could lead to seizures without overt MRI changes. In one patient with seizure during apparent remission, CSF analysis may have been appropriate to assess for the possibility of aseptic meningitis.

The main limitation of our study is the small number of patients with remission-phase seizures, constraining subgroup analysis and limiting conclusions regarding features that differentiate relapse-related seizures from MOG-AAE. This exploratory study demonstrates the need for a prospective, multicenter study documenting seizures, especially those outside of apparent relapses, to estimate their frequency. The secondary objective would be to identify early predictors of subsequent seizures and epilepsy in patients who experience seizure during relapse. Standardized intervals for serial contrast-enhanced imaging, electrodiagnostic testing and CSF sampling could better characterize the temporal relationship between radiographic evidence of meningocortical involvement, EEG abnormalities and seizures. We hypothesize that while some seizures, such as those during a steroid taper without concurrent clinical evidence of MOGAD, may represent acute symptomatic seizures heralding a relapse; the remainder of patients with recurrent seizures after acute symptomatic seizures related to MOGAD likely represent MOG-AAE, with a relatively benign course as compared to neural antibody AEAE.