Abstract
Lee WJ, Lee HS, Kim DY, Lee HS, Moon J, Park KI, Lee SK, Chu K, Lee ST. Brain. 2022;145(10):3509-3521. doi:10.1093/brain/awac166
Seronegative autoimmune encephalitis is autoimmune encephalitis without any identifiable pathogenic antibody. Although it is a major subtype of autoimmune encephalitis, many unmet clinical needs exist in terms of clinical characteristics, treatments and prognosis. In this institutional cohort study, patients diagnosed with seronegative autoimmune encephalitis with available 2-year outcomes were analysed for the disease course, 2-year outcome prediction system, effect of immunotherapy, necessity of further immunotherapy at 6 or 12 months and pattern of brain atrophy. Seronegative autoimmune encephalitis was subcategorized into antibody-negative probable autoimmune encephalitis, autoimmune limbic encephalitis and acute disseminated encephalomyelitis. Poor 2-year outcome was defined by modified Rankin scale scores 3–6, and the 2-year serial data of Clinical Assessment Scales in Autoimmune Encephalitis score was used for longitudinal data analyses. A total of 147 patients were included. The frequency of achieving a good 2-year outcome (modified Rankin scale 0–2) was 56.5%. The antibody-negative probable autoimmune encephalitis subtype exhibited the poorest outcomes, although the baseline severity was similar among the subtypes. The RAPID score, consisting of five early usable clinical factors, refractory status epilepticus, age of onset ≥60 years, probable autoimmune encephalitis (antibody-negative probable autoimmune encephalitis subtype), infratentorial involvement and delay of immunotherapy ≥1 month, was associated with poorer 2-year outcomes. Any immunotherapy was associated with clinical improvement in the patients with low risk for poor 2-year outcomes (RAPID scores 0–1), and the combination immunotherapy of steroid, immunoglobulin, rituximab and tocilizumab was associated with better outcomes in the patients with high risk for poor 2-year outcomes (RAPID scores 2–5). In patients with persistent disease at 6 months, continuing immunotherapy was associated with more improvement, while the effect of continuing immunotherapy for more than 12 months was unclear. In the longitudinal analysis of MRI, the development of cerebellar atrophy indicated poor outcomes, while the absence of diffuse cerebral atrophy or medial temporal atrophy indicated the possibility of a good outcome. This study provides information about the clinical characteristics and courses, the effect of immunotherapy and its duration, and prognostic factors in seronegative autoimmune encephalitis.
Commentary
Autoimmune encephalitides are a group of immune-mediated disorders that can cause epilepsy and status epilepticus. It is a diverse group that is further classified by anatomical, serological, and etiological criteria. 1 Classical presentations such as limbic encephalitis (LE), the N-methyl-D-aspartate receptor (NMDAR) or leucine-rich, glioma inactivated-1 (LGI1)-antibody syndromes, and typical paraneoplastic syndromes in the setting of a known cancer are well rooted in the medical literature. The identification of a pathogenic antibody targeting an onconeuronal or cell-surface antigen helps confirm the diagnosis. In these situations, clinical management is well established; relying on immune therapy and the screening and treatment of an underlying cancer, if appropriate. Early treatment is usually associated with better outcomes and is very rewarding for the treating physician to see seizures fade away and the patient progressively recovers with appropriate treatment.
Yet, these only represent a fraction of all the patients presenting with a clinical syndrome compatible with autoimmune encephalitis. Not unfrequently, a patient might present with a constellation of signs and symptoms suggestive of encephalitis, such as seizures, altered mental status, or cognitive and behavioral issues, but no firm ancillary evidence of an infectious or autoimmune disorder can be uncovered. 2,3 The cerebrospinal fluid (CSF) is perhaps mildly abnormal, the magnetic resonance imaging (MRI) may show nonspecific diffusion-weighted imaging (DWI) or T2/fluid-attenuated inversion-recovery (FLAIR) abnormalities, but no pathogen or antibody is identified. This clinical picture is sometimes referred to as antibody-negative probable autoimmune encephalitis (ANPRA). The uncertainty regarding the cause of the patient’s affliction often leads to difficult challenges in clinical management. Is this autoimmune or infectious encephalitis? Which treatment should be given? Is it safe to administer immune therapies if an infectious process has not been formally ruled out? While the clinician ponders these questions, precious time might be lost, perhaps at the expense of patients’ outcome. Fortunately, a recent study proposed a clinical score that helps differentiate viral from autoimmune encephalitis with excellent accuracy. 4
In this large prospective case series 5 from a national referral center for autoimmune encephalitis, the authors report the clinical picture, treatment, and evolution of 147 patients fulfilling criteria of autoimmune encephalitis, including ANPRA (117), LE (n = 23), and acute disseminated encephalomyelitis (n = 7), but in which no antibody was identified. Patients all underwent a standardized and comprehensive assessment for autoimmune and infectious encephalitis. Immunohistochemistry (IHC) on rodent brain sections followed by specific antibody detection by commercially available kits was used. The sensitivity is not 100%, which might have led to an underestimation of seropositive cases. Also, it is unclear if IHC-positive but specific antibody-negative cases, which might correspond to yet unknown or recently described antibodies, were considered as seronegative. The ANPRA group thus likely is a heterogeneous group of disorders with varied immunological mechanisms. Some may correspond to antibody-mediated encephalitis with a yet-to-be discovered antibody, while others might not. For instance, Case #2 presented in Supplemental Figure 4 shows most of the features of the recently redefined cryptogenic febrile-illness related epilepsy syndrome (FIRES), 6 including refractory status epilepticus (RSE) and T2/FLAIR hyperintense abnormalities in the claustrum, 6 which is thought to result from a non-antibody-mediated autoinflammatory mechanism. 7
Patients also underwent a very careful and close clinical follow-up with repeated assessments at weekly interval for the first 3 months then monthly up to 1 year then every 3 months up to 2 years. The investigators used both the broad modified Rankin Scale (mRS) and the more specific clinical assessment scale in autoimmune encephalitis (CASE), a 9-item scale encompassing consciousness, cognition, behavior, seizure activity, movement disorders, and brainstem and cerebellar function. Seizures occurred in 119 (81%) cases and were slightly, albeit not significantly, more frequent in the ANPRA group. Refractory SE occurred in 44 (30%) cases. The clinical presentation of ANPRA slightly differed from LE, with more signs of brainstem and cerebellar involvement. Seizures and cognitive and behavioral manifestations were similarly frequent in both groups. The MRI findings in ANPRA cases showed more extensive abnormalities involving the neocortex, subcortical white matter, basal grey matter and brainstem; while by definition abnormalities were almost restricted to the mesial temporal lobes in LE cases. This is further evidence that ANPRA is a heterogeneous group of disorders.
Most patients received steroids (80%), intravenous immunoglobulins (97%), and rituximab (79%); 40% received tocilizumab, an anti-interleukin-6 receptor antibody, or cyclophosphamide (10%). The latter were almost exclusively used in ANPRA and LE.
Patients were followed for a median duration of 29 months. At baseline, more than 50% were severely disabled and bedridden (mRS 5; CASE > 12). Clinical improvement mostly occurred during the first year, with close to 50% achieving functional autonomy (mRS
In multivariate logistic regression, the authors identified 5 independent factors associated with worse functional outcome: RSE, age
Some of these findings are expected: age is associated with unfavorable outcome in almost every acute neurological disorder and a few studies 2 have found SE to be associated with worse outcome in encephalitis, as it is in many other acute conditions.
More importantly, immune therapy appeared to be beneficial in seronegative autoimmune encephalitis as its earlier use is associated with better outcomes. Further, the early addition of tocilizumab or rituximab to first-line immune therapies was associated with faster recovery, regardless of other identified prognostic variables. These findings are similar to what has been shown in seropositive autoimmune encephalitis, such as the NMDAR 8 and LGI1 9 antibody syndromes, and in non-antibody-mediated autoinflammatory conditions, such as FIRES. 10 They also strongly suggest that even in the absence of an identified antibody or infectious agent, most cases of suspected encephalitis result from an autoimmune or at least autoinflammatory mechanism. Further research is required to unravel their pathophysiological pathway and identify early diagnostic biomarkers, focusing for instance on cytokine profiling 11 or single cell transcriptomics 12 in the CSF. In the meantime, this study already has an important clinical implication: we should not be reluctant to use early immune therapy in encephalitis that remain cryptogenic once usual causes have been ruled out.
Based on the variables associated with outcome, the authors propose the RAPID prognostic score, which can be easily applied in practice to help inform patients and families, if not to make clinical decisions. For instance, patients with RAPID score <3 had a faster recovery and a much higher (>80%) chance of low disability (mRS < 3) at 2 years.
Finally, while approximately half of the cohort received rituximab in the first month and tocilizumab in the first 2 months of the disease, these second-line therapies were also used later in the course of the disease. This was notably the case in patients with persistent disease 6 months after onset. In these patients, their use was also associated with clinical improvement.
Obviously, all the limitations of an observational study apply. For instance, treatment choices may have depended on factors that were not accounted for. Still, short of randomized controlled trials, this large observational study provides important and useful information on the clinical course, prognosis, and response to immune treatment of these still poorly understood conditions.
