Abstract
Liu X, Guo K, Lin J, Gong X, Li A, Zhou D, Hong Z. Epilepsia. 2022;63(7):1812-1821. doi:10.1111/epi.17245
This study was undertaken to update and evaluate long-term seizure outcomes in patients with autoimmune encephalitis (AE) based on a large cohort study with long follow-up. In this prospective observational registry study, we analyzed data from patients with AE mediated by common types of neuronal surface antibodies (anti-N-methyl-d-aspartate receptor [NMDAR], anti-leucine-rich glioma-inactivated 1 [LGI1]/contactin-associated protein-like 2 [Caspr2], anti-γ-aminobutyric acid type B receptor [GABAB R]). All patients were recruited from the Department of Neurology at West China Hospital between October 2011 and June 2019, and data were collected prospectively on their demographic and clinical characteristics, treatment strategy, and seizure outcomes, with a median follow-up of 42 months (range = 6-93 months). Potential risk factors associated with seizure recurrence were also assessed. Of 320 AE patients, 75.9% had acute seizures, among whom >90% of patients had their last seizure within 12 months of disease onset. During our follow-up, 21 (9.3%) patients experienced seizure recurrence. Patients with anti-GABAB R encephalitis had a higher cumulative incidence of seizure recurrence than those with anti-NMDAR (log-rank p = .03) or anti-LGI1/Caspr2 encephalitis (log-rank p = .04). Among patients with anti-NMDAR encephalitis, women had a significantly higher cumulative incidence of seizure recurrence than men (log-rank p = .01). Interictal epileptiform discharges (IEDs) or seizures captured on continuous electroencephalogram (EEG) in the acute phase were identified as potential risk factors for seizure recurrence (p = .04, p = .007). Among 163 patients with ≥24 months of follow-up, five (3.1%) showed persistent seizures and required ongoing antiseizure medications despite aggressive immunotherapy. Seizure recurrence occurred in a small number of patients and chronic epilepsy occurred in 3.1% of patients during prolonged follow-up. Across all types of AE, risk factors for seizure recurrence were IEDs or seizures captured on EEG in the acute phase; for anti-NMDAR encephalitis, female sex was also a risk factor.Objective:
Methods:
Results:
Significance:
Commentary
Seizures commonly occur in the acute setting of autoimmune encephalitis (AE), sometimes as the first presenting sign, so epileptologists can be an integral part of the diagnostic team of neurologists in cases of suspected AE. Now that the diagnosis and recognition of AE has improved, there has been increasing interest in understanding the complexity of the subacute and chronic recovery phase of AE beyond functional outcomes, including cognitive status 1 and treatment-resistant seizures. 2 -4 Identifying the frequency and predictors of specific outcomes of AE can enable more nuanced prognostication and counseling in the acute phase, but also inform future interventional studies. Recently, the ILAE Autoimmunity Taskforce recommended a conceptual distinction between seizures occurring acutely in the setting of AE (acute symptomatic seizures) versus chronically in a recurrent unprovoked manner (autoimmune-associated epilepsy). 5 The latest publication by Liu et al 6 makes an explicit attempt at identifying which patients with NMDA-R, LGI1, CASPR2, and GABA-B antibody-mediated encephalitis go on to develop autoimmune-associated epilepsy.
This large observational cohort study stems from an initiative entitled Outcomes of Autoimmune Encephalitis Study in Western China (ONE-WC), which started in 2011 and now includes a substantial number of patients with long-term follow-up. This cohort is heavily weighted toward NMDA-R encephalitis (about 75% of those reported in this study), which impacts the significance of the findings to a broader group of AE. The cohort has been well characterized, with periodic visits or telephone follow-ups conducted by neurologists to ascertain outcomes, including the presence of seizures. This study focused on cell-surface antibody-mediated encephalitis and included predominantly NMDA-R (n = 234), but also LGI1 (n = 37), CASPR2 (n = 12), and GABA-B (n = 37). A small proportion of patients (about 5%) received second-line immunotherapy (e.g., rituximab or cyclophosphamide), which may reflect differences in regional clinical practice, or inclusion of an earlier patient population when aggressive immunotherapy may not have been the standard approach in cases with incomplete response to first-line immunotherapy. Meanwhile, the majority (two-thirds) were maintained on chronic immunosuppression after discharge, although it is not stated on how this treatment was maintained.
Overall, this cohort reports favorable seizure outcomes and very low rates of autoimmune-associated epilepsy at long-term follow-up (3.7%). About two-thirds stopped having seizures within 4 weeks of initiation of immunotherapy, and the majority stopped thereafter. A proportion of patients experienced seizure recurrence, which typically responded to reinitiation of immunotherapy, suggesting this was a relapse of encephalitis rather than the development of chronic epilepsy. Outcomes did not differ significantly between those who continued on anti-seizure medications (ASMs) beyond the first 3 months of illness versus those discontinued ASMs. Of the 163 patients with greater than 2 years of follow-up, only 12 still took ASMs, and 7 of these had drug-resistant epilepsy. These findings are similar to that reported by a Dutch cohort 2 which had reported very low (1%) rates of persistent seizures at last follow-up in a group of 110 patients with NMDA-R, LGI1, and GABA-B antibody-mediated encephalitis. It should be noted that a recent long-term follow-up study of 49 patients with LGI1 encephalitis reported that about one-fifth of the patients continued to have seizures at the 2 year mark and beyond, 7 a significantly higher proportion than that suggested by this recent Chinese cohort 6 and prior Dutch cohort. 2 These discrepancies in outcomes may be due to genetic contributions to outcomes, differences in treatment practices across various centers, but also differences in outcome ascertainment. Indeed, the latest LGI1 encephalitis cohort with higher rates of epilepsy reports focal aware semiology (e.g., “rush of adrenaline,” “nausea, dazed, and flushing sensation”), which may go underreported unless specifically queried by the investigators. 7 Therefore, the true range of autoimmune-associated epilepsy may be higher than the one reported in this study.
Although the group that eventually suffers from autoimmune-associated epilepsy is likely a small one in the case of cell surface antibody-mediated encephalitis, it illustrates a core challenge in the management of seizures associated with autoimmunity (with or without cell surface antibodies)—how does one determine the role of autoimmunity in continuing to generate seizures in the chronic phase? Does an absence of response to first-line immunotherapy mean these epilepsies are structural in nature (e.g., related to hippocampal sclerosis or gliosis secondary to an encephalitic insult)? Or does our approach to immunotherapy need to differ in the chronic phase of disease and an incomplete immunotherapy response simply represents a lack of rationally targeted immunotherapy? The absence of reliable biomarkers to determine whether seizures continue to be caused by an active neuroinflammatory process makes this determination challenging, and the clinician is often left with difficult decisions in a patient with chronic drug-resistant epilepsy who suffered an initial encephalitic illness—proceeding to an epilepsy surgery workup versus retrial of aggressive immunotherapies such as second-line immunotherapies or biologics in a patient with no response to first-line immunotherapy. This challenging decision-making process uncommonly happens in patients with a history of cell surface antibody-mediated encephalitis, but it is the norm for those with antibodies targeting intracellular antigens (e.g., GAD65 antibodies). In those who continue to experience seizures after AE, is there still a spark? A better understanding of the biology of the disease through longitudinal studies investigating candidate biomarkers of neuroinflammation and autoimmunity will likely help us improve our decision-making and design better interventions.
