A clinical approach to new-onset psychosis associated with immune dysregulation: the concept of autoimmune psychosis

Psychosis associated with autoantibodies targeting well-characterized synaptic and neuronal cell surface proteins [3‐5]. Indeed, findings from several case reports and clinical case series have linked autoantibodies targeting NMDAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR), GABA_ßR and VGKC-protein complex, to new-onset psychosis among other psychiatric disturbances (e.g., mania, anxiety, aggression, emotional lability, and personality changes) as the dominant presenting symptoms [17‐23, 27, 38‐42]. Although the majority of these patients eventually exhibit a wide constellation of neurological symptoms throughout the course of the illness, isolated neuropsychiatric symptoms with dominant psychosis may rarely remain the key symptoms; we suggest that the latter pattern might constitute the presentation of milder form or forme fruste of autoimmune encephalitis. Thus, we recommend screening for these antibodies in serum and CSF of individuals presenting with new-onset atypical drug-refractory neuropsychiatric symptoms including new-onset psychosis, particularly in those exhibiting any of the above highlighted clinical accompaniments. This is important, as structural neuroimaging abnormalities are commonly not noted on conventional MRI in patients with autoimmune encephalitis associated with these synaptic and neuronal surface autoantibodies [43, 44]. Indeed, recent studies of anti-NMDAR encephalitis showed that structural neuroimaging abnormalities are absent on initial presentation and follow-up brain MRIs in about 89 and 79% of cases, respectively, despite associated profound cognitive, neurological, and neuropsychiatric disturbances [43]. Furthermore, identification of specific autoantibodies does not only establish the diagnosis of autoimmune psychosis, but also can clarify its subtype and guide treatment.

Psychosis associated with classical systemic inflammatory and autoimmune disorders (e.g., neuropsychiatric systemic lupus erythematosus), neuroinflammatory disorders (e.g., inflammatory demyelinating disorders such as multiple sclerosis and acute disseminated encephalomyelitis, and neurosarcoidosis), Hashimoto’s encephalopathy, or autoantibodies targeting intracellular antigens (i.e., glutamic acid decarboxylase and onconeuronal antigens).

Neuropsychiatric systemic lupus erythematous (NPSLE) occurs in 25 to 75% of SLE patients [1]. NPSLE includes cognitive deficit, headache, psychiatric symptoms (e.g., anxiety, depression, and psychosis), seizures, movement disorders, non-specific demyelination, and cerebrovascular events [32, 45, 46]. Psychiatric symptoms often occur within the first 2 years of disease onset [1] and are rarely limited to isolated psychosis [32]. The pathophysiology of NPSLE is multifactorial and includes inflammation-mediated microangiopathy and autoantibodies such as anti-GluN2 (NR2) subunits of NMDAR, anti-ribosomal P (positive in 90% of psychotic SLE patients), anti-endothelial cell, anti-ganglioside, and anti-dsDNA, as well as anti-phospholipid antibodies [1]. Other systemic inflammatory disorders such as Sjögren’s syndrome, vasculitis, and Behçet’s disease can also present with a variety of neuropsychiatric features including psychosis, in addition to a wide constellation of neurological disturbances such as cognitive changes [18]. Please refer to the comprehensive review published on this topic [18].

Central nervous system (CNS) inflammatory demyelinating disorders such as acute disseminated encephalomyelitis [47] and multiple sclerosis [48] can present with diverse psychotic features, suggesting that these disorders may share some common immune and inflammatory mechanisms with primary psychotic disorders [49]. Neurosarcoidosis is another neuroinflammatory disorder that can have neuropsychiatric comorbidities with psychotic features. Indeed, psychotic manifestations can be seen in up to 17% of individuals with neurosarcoidosis [50].

Hashimoto’s encephalopathy is a relatively uncommon steroid-responsive encephalopathy. It is characterized by rapidly progressive non-specific encephalopathy, seizures, myoclonus, stroke-like episodes, and diverse neuropsychiatric features that include mood disorders as well as psychosis [33]. This disorder is associated with thyroid antibodies, which can are also present in about 13% of healthy individuals [33]. Brain MRI and CSF investigations may be normal or revealing non-specific abnormalities [33].

Glutamic acid decarboxylase antibody seropositivity has been associated with a variety of neurological disorders including limbic encephalitis, which typically present with neurological (e.g., cognitive, short-term memory impairment, confusion, seizures) and psychiatric symptoms (mood alteration, anxiety, psychosis) [1, 33]. However, studies that are more recent indicated that other concomitant antibodies, such as those targeting GABA_ßR, are more likely to be the culprit [51, 52].

Paraneoplastic limbic systems can also present with a variety of neuropsychiatric including psychosis and neurological symptoms such as seizures and cognitive impairment [33]. Although onconeuronal antibodies are associated with paraneoplastic CNS syndromes, direct cytotoxic effects of T cell immunity appear to be more etiologically relevant to pathophysiology of paraneoplastic limbic encephalitis [53].

Seronegative but probable autoimmune psychosis (SPAP). The evidence supporting the relevance of SPAP as a distinct diagnostic category in clinical practice is indirect, limited, and largely consists of case reports that document responsiveness of secondary psychosis to immune therapies despite negative comprehensive serological assays for systemic and central nervous system autoimmune diseases, as well as lack of any other discernable organic etiology [54, 55]. The molecular and cellular changes of the pathoimmunological processes underlying SPAP remain largely mysterious. However, some of these cases may represent a subgroup of the increasingly recognized spectrum of seronegative autoimmune encephalitides [33, 50, 54‐61] presenting primarily with acute or subacute onset encephalopathy, short-term memory impairment, and neuropsychiatric features with dominant psychotic symptoms [54, 55]. Its etiological connection to the underlying immunological dysregulation can be inferred only after excluding other alternative organic causes such as infectious, toxic-metabolic, substance abuse, neurodegenerative, or genetic disorders. We suggest that seronegativity is likely attributed to autoantibodies that either have not yet been discovered or present exclusively in the CSF that is often not analyzed in this subgroup of individuals; (I) many cases of paraneoplastic limbic encephalitis previously characterized as seronegative were subsequently found to be GABA_ßR antibody seropositive [52], and (II) NMDAR antibodies were detected only in CSF of about 13% of individuals with confirmed anti-NMDAR encephalitis through research-based combined testing of immunohistochemistry and CBAs [62]. The strong immunoprecipitation property of the brain can also result in high binding of autoantibodies to their respective receptors, thus limiting their free circulation and detection in the serum and CSF [63]. Furthermore, methodological heterogeneity might occasionally contribute to false seronegativity; although the current CBA testing using commercially available kits is very specific, it appears to be less sensitive for detecting antibodies targeting synaptic and neuronal cell surface antigens (e.g., NMDAR, GABA_ßR, AMPAR) in patients with suspected synaptic autoimmune encephalitis, as compared to research-based combined testing of both rodent brain immunohistochemical staining as well as CBAs [64]. Indeed, about 12% of all confirmed autoimmune encephalitis cases were identified only after demonstrating positive rodent brain reactivity on immunohistochemical staining followed by research-based CBAs [64]. Alternatively, CNS autoimmunity may be mediated by not-yet-defined innate- or T cell autoimmunity, instead of circulating autoantibodies [1]. Thus, diagnosing SPAP requires high index of clinical suspicion together with lower threshold for referring patients with suspected SPAP for basic diagnostic investigation such as neuroimaging, EEG testing, and CSF analysis. However, given the relatively low diagnostic yield of conventional brain MRI [43, 44] and the limited use of CSF analysis, it is conceivable that SPAP frequently escapes clinical detection. Indeed, findings from recent studies assessing the sensitivity of various diagnostic tests in autoimmune encephalitis indicate that brain FDG-PET offers a higher diagnostic yield, with cerebral cortical hypometabolism being the most frequently observed abnormality, compared with early EEG, conventional brain MRI, and even CSF analysis [65]. A clinically meaningful response to immune therapy trial, similar to the approach used in seronegative rheumatologic disorders [66], might provide indirect evidence of the underlying CNS autoimmunity, particularly when associated with restoration of functional neuroimaging abnormalities of glucose metabolism or blood flow following immune therapy [36, 67, 68]. The yield of brain biopsy, similar to that of autoimmune encephalitis, is likely low [41]. However, brain biopsy might be considered in carefully selected cases of severe encephalopathy associated with refractory psychotic features of undefined organic causes to confirm their etiological connection to autoimmunity before initiating more aggressive immunosuppressive therapies. Indeed, we have reported cases and reviewed literature of brain biopsy-proven seronegative autoimmune encephalitis presenting with isolated neuropsychiatric disorders including psychosis [54, 56]. Although the typical neuroinflammatory findings of perivascular and parenchymal mononuclear inflammatory infiltrates and/or microglial nodules are typically not etiologically specific, but can, in the proper clinical context and after excluding other organic etiologies such as infections, provide indirect evidence of probable neuroimmune process.