Autoimmune Disorders and Psychiatric Manifestations: Lupus, MS, Thyroid, PANDAS, Anti-NMDA Receptor Encephalitis, and the Neuroinflammation Paradigm

The recognition that autoimmune processes can produce psychiatric syndromes indistinguishable from primary mental illness represents one of the most consequential developments in neuropsychiatry over the past two decades. From the landmark identification of anti-NMDA receptor encephalitis by Josep Dalmau and colleagues in 2007 to the expanding literature on neuroinflammatory biomarkers in major depression and schizophrenia, the field has moved well beyond the classical separation of "organic" and "functional" psychiatric disorders.

Autoimmune conditions affect approximately 5–8% of the general population, with women disproportionately affected at a ratio approaching 2:1 across most conditions. Psychiatric manifestations occur in a substantial proportion of these patients — estimates range from 20% to over 90% depending on the specific autoimmune disorder and how psychiatric symptoms are assessed. Critically, psychiatric symptoms may be the presenting feature in a significant minority of cases, preceding any recognized systemic autoimmune disease by months or years. This temporal dissociation creates a diagnostic trap: patients receive psychiatric diagnoses and treatments that, at best, partially address symptoms while the underlying autoimmune pathology progresses untreated.

The neurobiological mechanisms linking autoimmunity to psychiatric symptoms are diverse and often disease-specific, encompassing direct antibody-mediated neuronal dysfunction, cytokine-driven neurotransmitter perturbation, microglial activation, blood-brain barrier (BBB) disruption, demyelination, and vasculitis-mediated ischemia. Understanding these mechanisms is not merely academic — it dictates whether a patient requires immunotherapy, psychotropic medication, or both, and it fundamentally shapes prognosis.

This article provides a detailed clinical review of psychiatric manifestations across five major autoimmune conditions — systemic lupus erythematosus (SLE), multiple sclerosis (MS), autoimmune thyroid disease, PANDAS/PANS, and anti-NMDA receptor encephalitis — before examining the broader neuroinflammation hypothesis and its implications for psychiatric practice.

Systemic lupus erythematosus is a chronic, multi-system autoimmune disorder with a prevalence of approximately 20–150 per 100,000 persons, varying substantially by ethnicity and geography. Women of childbearing age are affected at a 9:1 female-to-male ratio. The American College of Rheumatology (ACR) has defined 19 neuropsychiatric syndromes attributable to SLE (NPSLE), which occur in an estimated 12–95% of patients depending on case definitions and ascertainment methods. When rigorous attribution criteria are applied, approximately 30–40% of SLE patients develop clinically significant neuropsychiatric manifestations.

Psychiatric Manifestations

The most prevalent psychiatric syndromes in NPSLE include:

• Cognitive dysfunction: Present in 20–80% of SLE patients across studies. Deficits typically involve attention, working memory, processing speed, and executive function. The Hopkins Lupus Cohort has documented cognitive impairment in approximately 33% of patients using standardized neuropsychological batteries.
• Major depression: Lifetime prevalence of 17–75%, with most rigorous estimates centering around 25–40% — substantially higher than the 7% point prevalence in the general population.
• Anxiety disorders: Present in 24–57% of NPSLE patients.
• Psychosis: Occurs in 2–11% of SLE patients, often with paranoid delusions, auditory hallucinations, and disorganized thinking. Lupus psychosis is a medical emergency requiring immediate immunosuppressive treatment.
• Acute confusional state (delirium): Reported in 4–7% and associated with high mortality without treatment.

Neurobiological Mechanisms

Multiple, often concurrent, pathophysiological mechanisms drive NPSLE:

• Anti-neuronal antibodies: Anti-NMDA receptor subunit NR2 antibodies (anti-NR2) are found in approximately 25–35% of SLE patients and have been associated with cognitive dysfunction and depression. These antibodies, when they penetrate the BBB, can cause excitotoxic neuronal injury through NMDA receptor activation, followed by receptor internalization and subsequent hypofunction — a mechanism paralleling hypotheses in schizophrenia.
• Anti-ribosomal P antibodies: Found in 12–26% of SLE patients, these are strongly associated with lupus psychosis (sensitivity ~75%, specificity ~75% in some series), though their pathogenic mechanism remains debated. Some evidence suggests they cross-react with neuronal surface proteins.
• Cerebral vasculitis and thrombosis: Antiphospholipid antibodies (present in 30–40% of SLE patients) cause a prothrombotic state contributing to cerebral ischemia, particularly in deep white matter and hippocampal regions.
• Cytokine-mediated neuroinflammation: Elevated intrathecal IL-6, IL-8, and interferon-α levels in NPSLE patients correlate with psychiatric symptom severity. Type I interferons, central to SLE pathogenesis, directly affect tryptophan metabolism via indoleamine 2,3-dioxygenase (IDO) activation, shunting tryptophan from serotonin synthesis toward the kynurenine pathway. Downstream kynurenine metabolites include the NMDA receptor agonist quinolinic acid (neurotoxic) and the neuroprotective kynurenic acid. The balance between these metabolites — the "kynurenine neurotoxic ratio" — is shifted toward neurotoxicity in active NPSLE.
• BBB disruption: Anti-endothelial cell antibodies and complement-mediated vascular injury compromise BBB integrity, allowing peripheral autoantibodies and inflammatory mediators access to CNS parenchyma.

Treatment

Management of NPSLE depends on the attributed mechanism. Inflammatory/autoimmune-mediated manifestations (psychosis, acute confusional states, myelopathy) are treated with high-dose corticosteroids (typically methylprednisolone 1 g/day IV for 3–5 days) followed by oral prednisone taper, with or without cyclophosphamide. The Euro-Lupus Nephritis Trial protocol for cyclophosphamide (500 mg IV every 2 weeks × 6 doses) has been adapted for neuropsychiatric lupus, though randomized trial data specifically in NPSLE are limited. Observational studies report response rates of 50–80% for lupus psychosis with combined corticosteroid-cyclophosphamide therapy. Rituximab has been used in refractory cases with reported response rates of 60–90% in case series, though no RCTs exist for NPSLE specifically. For thrombotic manifestations, anticoagulation is the primary treatment. Psychotropic medications are used adjunctively — SSRIs for depression, low-dose atypical antipsychotics for psychosis — but should not replace immunotherapy when there is evidence of active autoimmune CNS disease.

Prognostic Factors

Poor prognostic indicators in NPSLE include diffuse (vs. focal) CNS involvement, antiphospholipid antibody positivity (associated with recurrent events), persistent hypocomplementemia, high anti-dsDNA titers, and delayed initiation of immunotherapy. Cognitive dysfunction, once established, tends to be persistent: the Hopkins Lupus Cohort found that approximately 50% of patients with cognitive impairment showed no improvement over 2 years despite systemic disease control.

Multiple sclerosis is a chronic demyelinating autoimmune disorder of the central nervous system affecting approximately 2.8 million people worldwide (2020 Atlas of MS), with prevalence rates of 50–300 per 100,000 in high-incidence regions. The disease is characterized by inflammatory demyelination and axonal injury mediated by autoreactive T cells and B cells targeting myelin antigens. Psychiatric manifestations in MS are among the most prevalent and disabling comorbidities, yet they remain systematically underdiagnosed and undertreated.

Prevalence of Psychiatric Disorders in MS

• Major depressive disorder: Lifetime prevalence of approximately 50%, with 12-month prevalence of 25–30% — roughly 2–3 times that of the general population. The landmark Canadian Community Health Survey (Patten et al., 2003) established a 12-month MDD prevalence of 25.7% in MS versus 8.9% in matched controls.
• Anxiety disorders: 12-month prevalence of 22–36%, with generalized anxiety disorder and adjustment disorder being most common.
• Bipolar disorder: Prevalence of approximately 5–10%, roughly 2–3 times expected population rates.
• Pseudobulbar affect (PBA): Characterized by involuntary, contextually inappropriate laughing or crying, PBA affects approximately 10–46% of MS patients depending on assessment methods. It is distinct from depression and reflects disruption of cortico-pontine-cerebellar circuits.
• Psychosis: Rare but occurs at 2–3 times the general population rate (~2–4%).
• Cognitive impairment: Present in 40–70% of MS patients, affecting processing speed, episodic memory, and executive function. Cognitive impairment in MS is the strongest predictor of unemployment, above physical disability.

Neurobiological Mechanisms

The psychiatric manifestations of MS are not simply "reactive" to disability. While psychosocial factors contribute, substantial evidence supports direct neurobiological causation:

• Lesion location effects: Demyelinating plaques in specific locations correlate with psychiatric presentations. Left frontal and temporal lesions are associated with depression. Frontal and temporal lesions, particularly involving the arcuate fasciculus and prefrontal regions, increase psychosis risk. Hippocampal demyelination, now recognized as common in MS, correlates with memory deficits and depression.
• Gray matter atrophy: Independent of white matter lesion burden, gray matter atrophy — particularly of the hippocampus, amygdala, and prefrontal cortex — predicts depressive symptoms. This gray matter pathology is driven by meningeal inflammation and subpial cortical demyelination, increasingly recognized as a key disease process in progressive MS.
• Neuroinflammatory mechanisms: Activated microglia in normal-appearing white and gray matter produce pro-inflammatory cytokines (TNF-α, IL-1β, IL-6) that disrupt serotonergic, glutamatergic, and dopaminergic neurotransmission. The kynurenine pathway is upregulated in MS, with elevated quinolinic acid levels in the CSF correlating with both neurological disability and depressive symptom severity.
• Hypothalamic-pituitary-adrenal (HPA) axis dysregulation: Demyelination of hypothalamic pathways and chronic inflammatory activation produce HPA axis hyperactivity, evidenced by enlarged adrenal glands and elevated cortisol levels in MS patients, paralleling findings in primary major depression.
• Iatrogenic factors: Interferon-beta therapy, a mainstay of MS disease-modifying therapy, is associated with depression onset or exacerbation in approximately 20–40% of patients in some series, likely through IDO activation and serotonin depletion. This has contributed to the shift toward newer disease-modifying therapies with more favorable psychiatric profiles.

Treatment Considerations

Depression in MS responds to standard antidepressant pharmacotherapy, though the evidence base is thinner than for primary depression. A 2011 Cochrane review (Koch et al.) found limited but supportive evidence for SSRIs, with desipramine and paroxetine showing efficacy over placebo in small trials. Cognitive behavioral therapy (CBT) has moderate evidence in MS-related depression, with effect sizes (Cohen's d) of approximately 0.5–0.7 in meta-analyses. PBA is treated with dextromethorphan/quinidine (Nuedexta), which demonstrated a large effect size in the STAR trial (NNT ≈ 3 for ≥50% reduction in episodes). Importantly, optimal management of MS-related psychiatric symptoms requires concurrent disease-modifying therapy — untreated inflammatory disease activity predicts psychiatric symptom persistence.

Autoimmune thyroid diseases — primarily Hashimoto's thyroiditis and Graves' disease — are the most common autoimmune conditions overall, affecting approximately 5% of the general population (with anti-thyroid antibody positivity present in up to 10–12% of women). Their psychiatric manifestations occur through two distinct but overlapping pathways: thyroid hormone dysregulation and direct autoimmune CNS effects.

Thyroid Hormone–Mediated Psychiatric Effects

Thyroid hormones (T3 and T4) are essential neuromodulators that regulate serotonergic, noradrenergic, and GABAergic neurotransmission, neuronal myelination, and synaptic plasticity. Psychiatric manifestations vary by direction of thyroid dysfunction:

• Hypothyroidism (Hashimoto's thyroiditis): Depression occurs in approximately 40–60% of clinically hypothyroid patients. Cognitive dysfunction ("brain fog"), psychomotor retardation, and anhedonia are characteristic. Subclinical hypothyroidism (elevated TSH, normal free T4) is associated with a 1.5–3-fold increased risk of depression in several meta-analyses, including a 2018 meta-analysis by Tang et al. (OR = 2.35, 95% CI 1.84–3.02). Hypothyroid depression is thought to result from reduced serotonin synthesis (thyroid hormones upregulate tryptophan hydroxylase), decreased cortical 5-HT2 receptor sensitivity, and reduced hippocampal BDNF expression.
• Hyperthyroidism (Graves' disease): Anxiety is the most prominent psychiatric feature (60–80%), along with irritability, insomnia, emotional lability, and impaired concentration. Frank psychosis occurs in approximately 1–2% ("thyroid storm psychosis"). Mania or hypomania occurs in 2–5% of cases. The mechanism involves enhanced beta-adrenergic sensitivity, increased central dopaminergic and noradrenergic activity, and direct T3 effects on amygdala excitability.
• "Myxedema madness": A rare but dramatic presentation of severe hypothyroidism featuring paranoid psychosis, hallucinations, and delirium. Classically described in the pre-levothyroxine era, it still occurs when hypothyroidism is unrecognized.

Hashimoto's Encephalopathy (Steroid-Responsive Encephalopathy Associated with Autoimmune Thyroiditis — SREAT)

Hashimoto's encephalopathy (HE), now often termed SREAT to reflect its defining feature, is a rare autoimmune encephalopathy (estimated prevalence 2.1/100,000) characterized by cognitive dysfunction, psychiatric symptoms (psychosis, depression, personality change), seizures, myoclonus, and/or stroke-like episodes in the presence of elevated anti-thyroid peroxidase (anti-TPO) or anti-thyroglobulin antibodies and a dramatic response to corticosteroids. Crucially, HE can occur in euthyroid patients — the encephalopathy is not caused by thyroid hormone levels but rather by an autoimmune process affecting the brain directly.

The pathophysiology of HE remains incompletely understood. Anti-TPO antibodies are likely an epiphenomenon rather than directly pathogenic; the actual CNS-targeting antibodies are unknown in most cases. Brain biopsy studies (rare) have shown perivascular lymphocytic infiltration, suggesting a T cell–mediated vasculitis. Some patients have anti-neuronal antibodies on research assays, and there is overlap with other autoimmune encephalitides.

Diagnosis is challenging and is essentially one of exclusion: encephalopathy + elevated anti-thyroid antibodies + exclusion of other causes + steroid response. EEG is abnormal in ~90% of cases (typically generalized slowing). CSF may show elevated protein in 75% and pleocytosis in 25%. MRI is abnormal in approximately 50% but nonspecific.

Treatment and Outcomes

The defining feature of HE/SREAT is its responsiveness to glucocorticoids — approximately 90–98% of patients show improvement with high-dose corticosteroid therapy. However, relapse occurs in approximately 25–50% during steroid tapering, necessitating steroid-sparing agents (azathioprine, mycophenolate, rituximab) in a substantial subset. Long-term prognosis is generally favorable with appropriate immunotherapy, but delayed diagnosis — average time to diagnosis exceeds 2 years in some series — is associated with persistent cognitive deficits. The most critical diagnostic pitfall is the failure to check anti-thyroid antibodies in patients presenting with unexplained encephalopathy or treatment-refractory psychiatric symptoms.

For thyroid hormone–mediated psychiatric symptoms, hormone replacement (levothyroxine for hypothyroidism) or suppression (methimazole/propylthiouracil for hyperthyroidism) is the primary treatment. Depression associated with hypothyroidism resolves in the majority of patients with adequate thyroid hormone replacement, though residual depressive symptoms persist in approximately 10–20% despite normalized TSH, suggesting concurrent primary depression or permanent structural CNS effects. Augmentation of antidepressants with T3 (liothyronine 25–50 μg/day) has demonstrated efficacy in treatment-resistant depression, most notably in the STAR*D trial, where T3 augmentation produced remission rates of approximately 24.7% — comparable to lithium augmentation (15.9%) with a significantly more favorable side effect profile.

Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal Infections (PANDAS) and the broader category of Pediatric Acute-onset Neuropsychiatric Syndrome (PANS) represent a distinct clinical entity in which autoimmune or inflammatory mechanisms trigger abrupt-onset OCD and/or tic disorders in children, typically following infection. While PANDAS specifically implicates Group A streptococcal (GAS) infection as the trigger, PANS encompasses cases triggered by other infections, metabolic disturbances, or unknown precipitants.

Diagnostic Criteria and Epidemiology

PANDAS criteria (Swedo et al., 1998) require: (1) presence of OCD and/or a tic disorder, (2) prepubertal onset, (3) abrupt onset of symptoms or dramatic exacerbations, (4) association with GAS infection, and (5) neurological abnormalities (choreiform movements, motor hyperactivity). PANS criteria (2012) are broader: (1) abrupt onset of OCD or severely restricted food intake, (2) concurrent presence of additional neuropsychiatric symptoms (≥2 of: anxiety, emotional lability, irritability/aggression, behavioral regression, deterioration in school performance, sensory or motor abnormalities, somatic symptoms including sleep disturbance and urinary symptoms), and (3) symptoms not better explained by a known neurological or medical disorder.

Prevalence is difficult to establish and remains debated. Estimates suggest that PANDAS/PANS may account for approximately 5–10% of childhood-onset OCD cases, though some investigators consider this an overestimate while others argue for underrecognition. The mean age of onset is approximately 6–8 years, with a male preponderance (2–3:1) mirroring childhood tic disorders.

Pathophysiology: Molecular Mimicry and the Basal Ganglia

The leading pathophysiological model invokes molecular mimicry, wherein antibodies generated against GAS antigens (particularly the M protein and N-acetyl-beta-D-glucosamine, or GlcNAc) cross-react with neuronal targets in the basal ganglia — specifically, proteins on the surface of cholinergic interneurons in the caudate and putamen. This mechanism is analogous to Sydenham's chorea, the neuropsychiatric manifestation of rheumatic fever, which shares striking clinical overlap with PANDAS (OCD symptoms occur in approximately 70% of Sydenham's chorea patients).

Specific proposed targets include:

• Dopamine D1 and D2 receptors: Anti-neuronal antibodies from PANDAS patients have been shown to activate CaM kinase II signaling in human neuronal cell lines, a pathway that increases dopamine release. This hyperdopaminergic state in the basal ganglia could drive both tics and OCD-like repetitive behaviors.
• Lysoganglioside GM1 and tubulin: Cross-reactive antibodies against these targets have been detected at higher titers in PANDAS patients compared to strep-infected controls without neuropsychiatric symptoms (Cunningham panel).
• Blood-brain barrier disruption: Inflammatory cytokines released during streptococcal infection may transiently increase BBB permeability, allowing circulating autoantibodies access to the CNS.

Neuroimaging evidence supports basal ganglia involvement: the landmark study by Giedd et al. (2000) using volumetric MRI found that the average volume of the caudate, putamen, and globus pallidus was significantly larger in PANDAS patients during acute exacerbations compared to healthy children, a pattern also seen in Sydenham's chorea. Basal ganglia inflammation has been confirmed in a small PET study using the microglial activation marker TSPO.

Treatment Approaches

Treatment of PANDAS/PANS is multimodal and remains an area of active debate:

• Antibiotic therapy: Acute GAS infection requires standard antibiotic treatment (penicillin, amoxicillin). The role of prophylactic antibiotics is more controversial. A randomized controlled trial by Snider et al. (2005) found that penicillin and azithromycin prophylaxis reduced streptococcal infections and neuropsychiatric exacerbations compared to placebo over 12 months.
• Standard psychiatric treatments: CBT (particularly exposure and response prevention, ERP) and SSRIs are first-line for OCD symptoms in PANDAS/PANS, as in primary OCD. Clinical experience suggests that PANDAS patients may be more SSRI-sensitive, requiring lower starting doses. Response rates to combined ERP + SSRI are similar to primary childhood OCD (approximately 60–70% showing clinically significant improvement).
• Immunomodulatory therapies: For moderate-to-severe cases or those refractory to standard treatment, immunotherapy is considered. A pivotal double-blind, placebo-controlled trial (Perlmutter et al., 1999) demonstrated that both plasma exchange and IVIG produced significant symptom improvement in PANDAS (58% and 45% improvement in OCD severity, respectively, vs. no significant change in placebo). However, a subsequent larger trial of IVIG alone (Williams et al., 2016) showed more modest benefits. NSAID therapy (ibuprofen) is sometimes used for milder cases, with anecdotal reports of benefit but no RCT data.
• Corticosteroids: Short courses of oral corticosteroids (e.g., prednisone 1–2 mg/kg/day for 5 days) have been used with reported benefit in acute flares, though no RCT evidence exists.

Controversy and Diagnostic Challenges

PANDAS remains one of the most contentious diagnoses in child psychiatry and neurology. Critics raise several valid concerns: GAS infections are extremely common in the target age group, making temporal associations potentially coincidental; the Cunningham panel and other commercial autoantibody assays lack adequate sensitivity and specificity for clinical diagnosis; and a definitive biomarker does not exist. The 2017 PANS/PANDAS Consortium Guidelines (Swedo et al.) emphasized that diagnosis remains clinical and that immunotherapy should be reserved for well-characterized cases with clear infection-triggered exacerbations, not applied broadly to childhood OCD.

Anti-N-methyl-D-aspartate receptor (anti-NMDAR) encephalitis, first described by Dalmau and colleagues in 2007, has become the paradigmatic autoimmune neuropsychiatric disorder. It is now recognized as the most common autoimmune encephalitis, surpassing all individual viral etiologies of encephalitis in young populations. The disorder is caused by IgG autoantibodies targeting the GluN1 (NR1) subunit of the NMDA receptor, leading to receptor internalization, NMDA receptor hypofunction, and a characteristic neuropsychiatric syndrome.

Epidemiology

The California Encephalitis Project found that anti-NMDAR encephalitis accounted for approximately 4% of all encephalitis cases and was more common than any single viral etiology in patients under 30. Estimated incidence is approximately 1.5 per million per year. Approximately 80% of patients are female, and median age of onset is 21 years. In approximately 40–50% of female patients, the disorder is associated with an ovarian teratoma containing neural tissue that likely serves as the immunogenic trigger. In male patients and children, tumors are found in less than 5–10% of cases, and a post-infectious trigger (often HSV encephalitis) is increasingly recognized.

Clinical Presentation and Stages

The disorder follows a characteristic staged progression:

• Prodromal phase (days 1–14): Headache, fever, and upper respiratory or gastrointestinal symptoms in approximately 70% of cases, often leading to initial misdiagnosis as a viral illness.
• Psychiatric phase (weeks 1–3): Prominent psychiatric symptoms dominate early presentation and lead to initial psychiatric hospitalization in approximately 77% of adult cases. Common symptoms include anxiety, insomnia, paranoid delusions, auditory and visual hallucinations, bizarre behavior, catatonia, mania, and agitation. Presentation in children is somewhat different, with more prominent behavioral change, tantrums, irritability, and new-onset seizures.
• Neurological phase (weeks 2–8): Progressive movement disorders (orofacial dyskinesias — highly characteristic — dystonia, choreoathetosis), seizures, autonomic instability (tachycardia, blood pressure lability, central hypoventilation), and decreased level of consciousness. Catatonia with mutism is common and can be a clinical clue differentiating the condition from primary psychosis.
• Recovery phase: Gradual improvement over weeks to months in the reverse order of symptom development, with psychiatric symptoms often the last to fully resolve.

Pathophysiology

The mechanism is remarkably specific and well-characterized. Anti-GluN1 antibodies cause selective, antibody-concentration-dependent internalization of NMDA receptors from the synaptic surface. This process is reversible upon antibody removal, explaining the recovery potential with immunotherapy. The resulting NMDA receptor hypofunction closely parallels the glutamate hypothesis of schizophrenia — indeed, NMDA receptor antagonists (ketamine, phencyclidine) produce a clinical syndrome strikingly similar to anti-NMDAR encephalitis, including psychosis, catatonia, movement abnormalities, and autonomic instability.

The NMDA receptor hypofunction disproportionately affects GABAergic interneurons, which have the highest NMDA receptor density. Loss of GABAergic interneuron function results in disinhibition of glutamatergic pyramidal neurons, producing a paradoxical state of extracellular glutamate excess alongside NMDA receptor hypofunction. This model explains the excitatory symptoms (seizures, dyskinesias) coexisting with the "negative" schizophrenia-like symptoms (catatonia, mutism, cognitive dysfunction). The frontostrial circuitry is particularly vulnerable, consistent with the prominent catatonic, psychotic, and dyskinetic features.

Diagnosis

Definitive diagnosis requires detection of anti-GluN1 IgG antibodies in CSF, which has sensitivity exceeding 98% and specificity approaching 100% when tested in reference laboratories using cell-based assays. Serum testing alone has a false-negative rate of approximately 14–17% and should not be used to exclude the diagnosis if clinical suspicion is high. CSF pleocytosis (lymphocytic) is present in approximately 80%, and EEG shows diffuse slowing or the "extreme delta brush" pattern (present in approximately 30%, but highly specific when found). MRI is normal in approximately 50% of patients at presentation — a normal MRI does not rule out the diagnosis.

Treatment and Outcomes

Treatment is staged immunotherapy with concurrent tumor removal when applicable:

• First-line immunotherapy: High-dose IV methylprednisolone (1 g/day × 5 days), IVIG (0.4 g/kg/day × 5 days), and/or plasma exchange. Approximately 53% of patients improve with first-line therapy alone (Titulaer et al., 2013, the largest outcome study with 577 patients).
• Second-line immunotherapy: Rituximab (375 mg/m² weekly × 4 weeks) and/or cyclophosphamide (750 mg/m² monthly). Among those who failed first-line therapy, approximately 67% improved with second-line treatment.
• Tumor removal: In teratoma-associated cases, oophorectomy is critical and improves outcomes. Patients with tumor removal had significantly better 2-year outcomes than those without (in the Titulaer cohort).

Overall, approximately 81% of patients achieve good outcomes (modified Rankin Scale 0–2) at 24 months, though recovery is often slow and incomplete for cognitive and psychiatric symptoms. Approximately 12% of patients relapse, usually within the first 2 years. Mortality has decreased to approximately 3–7% with aggressive immunotherapy, down from early case series mortality of 25%.

Prognostic Factors

Favorable prognosis is associated with: tumor identification and removal, earlier initiation of immunotherapy (within 4 weeks of onset), less severe initial presentation, and no ICU admission. The strongest predictor of poor outcome is delayed treatment — mean time from symptom onset to immunotherapy initiation exceeds 2 months in many centers, often because initial psychiatric presentation leads to prolonged psychiatric management before the autoimmune etiology is recognized. Persistent cognitive deficits (particularly executive dysfunction and memory impairment) are present in approximately 20–30% of recovered patients on detailed neuropsychological testing.

Beyond the specific autoimmune disorders discussed above, a growing body of evidence suggests that neuroinflammatory processes contribute to a subset of patients with diagnoses that have traditionally been considered "primary" psychiatric disorders — particularly schizophrenia, bipolar disorder, and major depressive disorder.

Evidence for Neuroinflammation in Primary Psychiatric Disorders

• Inflammatory biomarkers: Meta-analyses consistently demonstrate elevated peripheral inflammatory markers (CRP, IL-6, TNF-α, IL-1β) in active episodes of schizophrenia (Goldsmith et al., 2016), bipolar disorder, and major depression. In the MDD literature, a large meta-analysis by Dowlati et al. (2010) confirmed significantly elevated IL-6 and TNF-α levels relative to healthy controls. CRP levels above 3 mg/L predict poorer response to SSRIs and better response to anti-inflammatory augmentation strategies.
• Microglial activation: Positron emission tomography (PET) studies using TSPO radioligands have demonstrated increased microglial activation in the prefrontal cortex and anterior cingulate of patients with schizophrenia, depression, and bipolar disorder during acute episodes. A meta-analysis by Plitman et al. (2021) confirmed elevated TSPO binding in schizophrenia.
• Genetic overlap: Genome-wide association studies (GWAS) have identified shared genetic loci between autoimmune disorders and psychiatric conditions. The MHC region on chromosome 6, particularly HLA-B and complement component C4 variants, confers risk for both schizophrenia and autoimmune diseases. The landmark C4 study by Sekar et al. (2016) demonstrated that structurally distinct C4 alleles associated with increased C4A expression were associated with greater schizophrenia risk, linking excessive complement-mediated synaptic pruning to psychosis vulnerability.
• Epidemiological associations: Large Danish registry studies (Benros et al., 2011, 2013) found that having any autoimmune disorder increased schizophrenia risk by approximately 29% (IRR = 1.29), and having a prior serious infection increased risk by approximately 60%. The combination of autoimmune disease and infection was associated with a 2.25-fold increased risk.
• Autoantibody prevalence in first-episode psychosis: Studies screening first-episode psychosis patients for anti-neuronal antibodies have yielded variable but intriguing results. A systematic review by Pollak et al. (2019) estimated that approximately 5–8% of first-episode psychosis cases may have clinically relevant anti-neuronal antibodies, though methodological heterogeneity across studies limits precision. The SINAPPS2 study, a UK-based trial, is investigating immunotherapy in antibody-positive psychosis patients.

Anti-Inflammatory Treatments in Psychiatric Disorders

The neuroinflammation hypothesis has generated therapeutic trials:

• Celecoxib augmentation in schizophrenia: A meta-analysis by Zheng et al. (2017) found that adjunctive celecoxib improved total PANSS scores in early-phase schizophrenia (SMD = −0.41, 95% CI −0.65 to −0.17) but not in chronic schizophrenia, suggesting an anti-inflammatory treatment window.
• Anti-inflammatory augmentation in MDD: A comprehensive meta-analysis by Köhler-Forsberg et al. (2019) found that anti-inflammatory agents (NSAIDs, cytokine inhibitors, statins, pioglitazone) as adjuncts to antidepressants produced a pooled effect size of approximately SMD = −0.55 for depressive symptoms. The NNT for antidepressant response with anti-inflammatory augmentation was approximately 6–8 in available trials.
• Minocycline in schizophrenia: Minocycline, a tetracycline antibiotic with anti-inflammatory and neuroprotective properties, has been studied in multiple RCTs. A meta-analysis found modest benefits for negative symptoms (SMD ≈ −0.40), but the BeneMin trial (2018), the largest individual RCT (n = 207), was negative for its primary outcome.

The Emerging Concept of Autoimmune Psychosis

Recognizing that some patients presenting with psychosis have an autoimmune etiology, a consensus position paper (Pollak et al., 2020) proposed criteria for "autoimmune psychosis." Red flags suggesting autoimmune etiology include: rapid onset (days to weeks rather than months), new-onset in patients over 40, presence of movement disorders or catatonia, seizures, autonomic instability, treatment resistance to conventional antipsychotics, and concurrent autoimmune disease. Screening recommendations include anti-neuronal antibody panels, inflammatory markers, thyroid antibodies, ANA, and in appropriate clinical contexts, lumbar puncture.

The clinical challenge in autoimmune neuropsychiatry is bidirectional: autoimmune conditions are misdiagnosed as primary psychiatric disorders, but equally, primary psychiatric disorders can be erroneously attributed to autoimmune processes in the era of heightened awareness. Rigorous diagnostic reasoning requires attention to several principles.

Red Flags Suggesting Autoimmune Etiology in Psychiatric Presentation

• Acute or subacute onset (days to weeks) rather than the insidious onset typical of schizophrenia or mood disorders
• Rapid progression of symptoms
• Prominent cognitive dysfunction (especially encephalopathic features: disorientation, fluctuating consciousness) disproportionate to the apparent psychiatric diagnosis
• Movement abnormalities: orofacial dyskinesias (before antipsychotic exposure), catatonia, chorea, myoclonus
• Seizures, especially new-onset in a psychiatric setting
• Autonomic instability: unexplained tachycardia, blood pressure lability, diaphoresis
• Treatment resistance to appropriate psychotropic medications
• Age-atypical presentation (e.g., first psychotic episode after age 40)
• Known autoimmune disease or strong family history of autoimmunity
• Recent infection preceding psychiatric onset

Diagnostic Pitfalls

Several specific pitfalls warrant emphasis:

• Normal MRI does not exclude autoimmune encephalitis. MRI is normal in approximately 50% of anti-NMDAR encephalitis cases at presentation and may be normal in early NPSLE and HE.
• Negative serum antibody testing does not exclude autoimmune encephalitis. CSF testing is required for definitive exclusion, particularly for anti-NMDAR antibodies.
• Anti-thyroid antibodies are common in the general population. Elevated anti-TPO antibodies alone do not establish Hashimoto's encephalopathy; the diagnosis requires encephalopathy plus exclusion of other causes plus steroid responsiveness.
• Psychiatric symptoms in autoimmune disease may be reactive rather than autoimmune. Not all depression in SLE is NPSLE; adjustment to chronic illness, corticosteroid effects, and comorbid primary depression all contribute. Clinical context, temporal relationships, and disease activity indices help distinguish mechanisms.
• Corticosteroid effects on mood are common and dose-dependent. Psychiatric symptoms (mania, depression, psychosis) occur in approximately 5–18% of patients receiving ≥40 mg/day prednisone equivalent, complicating assessment in autoimmune patients receiving steroid therapy.

Recommended Workup for Suspected Autoimmune Psychiatric Presentation

A staged approach is clinically practical:

• First-tier (all patients with atypical or treatment-resistant psychiatric presentations): CBC with differential, CMP, TSH, free T4, anti-TPO antibodies, ANA, ESR, CRP, urinalysis
• Second-tier (if first-tier abnormal or high clinical suspicion): Complement levels (C3, C4), anti-dsDNA, antiphospholipid antibodies, anti-neuronal antibody panel (serum), brain MRI with contrast, EEG
• Third-tier (high suspicion for autoimmune encephalitis): Lumbar puncture with CSF cell count, protein, glucose, oligoclonal bands, anti-neuronal antibodies in CSF, cytology. Pelvic ultrasound or CT/MRI in female patients to evaluate for ovarian teratoma.

While specific treatments vary by condition (as detailed above), several overarching principles guide management:

Immunotherapy as Primary Treatment

When psychiatric symptoms are attributable to autoimmune mechanisms, immunotherapy is the primary treatment, not psychotropic medication. This represents a fundamental departure from standard psychiatric practice and requires interdisciplinary collaboration between psychiatry, neurology, and rheumatology or immunology. The evidence base for immunotherapy in specific conditions is summarized:

• Anti-NMDAR encephalitis: Strong evidence (multiple large cohort studies). First-line immunotherapy response rate ~53%; second-line response rate ~67% of first-line failures. Overall good outcome at 24 months: ~81%.
• NPSLE psychosis/delirium: Moderate evidence (observational studies, guidelines). Response to corticosteroids + cyclophosphamide: 50–80%.
• Hashimoto's encephalopathy: Moderate evidence (case series). Steroid response rate: ~90–98%.
• PANDAS/PANS: Limited evidence (2 small RCTs for IVIG/plasma exchange). IVIG/plasma exchange improvement: 45–58% in the pivotal Perlmutter trial.

Psychotropic Medications as Adjunctive Therapy

Psychotropic medications remain important for symptom management during acute illness and recovery. However, several caveats apply:

• Antipsychotics should be used cautiously in anti-NMDAR encephalitis, as these patients have heightened susceptibility to neuroleptic malignant syndrome (NMS) and dystonic reactions. Low-potency agents or quetiapine are generally preferred when antipsychotics are necessary.
• SSRIs are appropriate for depression and OCD symptoms across autoimmune conditions but may be less effective without concurrent treatment of the underlying autoimmune process.
• Benzodiazepines (particularly lorazepam) are first-line for catatonia in autoimmune encephalitis, as in other settings. Response rates for lorazepam in catatonia are approximately 60–80%, though the response may be incomplete or transient if the underlying autoimmune process continues.

Long-Term Monitoring and Relapse Prevention

Most autoimmune neuropsychiatric conditions carry relapse risk. Anti-NMDAR encephalitis relapses in approximately 12% of patients within 2 years, often heralded by return of psychiatric symptoms. NPSLE psychiatric manifestations recur in approximately 15–35% of patients over 5 years. Monitoring protocols should include regular clinical assessment for early psychiatric symptom recurrence, relevant serological markers (anti-dsDNA, complement levels for SLE; anti-NMDAR antibodies for encephalitis), and periodic neuropsychological testing in patients with prior cognitive deficits.

Understanding prognostic factors is critical for clinical counseling and treatment planning. Across autoimmune neuropsychiatric conditions, several themes emerge:

Timing of Treatment

Across virtually all autoimmune neuropsychiatric conditions, earlier treatment initiation predicts better outcomes. In anti-NMDAR encephalitis, the Titulaer et al. (2013) cohort demonstrated that immunotherapy initiated within 4 weeks of symptom onset was associated with significantly better functional outcomes at 24 months. In NPSLE, delayed recognition and treatment of psychosis or delirium is associated with higher mortality and persistent cognitive deficits. This underscores the clinical urgency of accurate early diagnosis.

Cognitive Outcomes

Cognitive deficits are the most persistent sequelae across conditions. Even in "good outcome" anti-NMDAR encephalitis patients, detailed neuropsychological testing reveals executive dysfunction and memory impairment in 20–30%. In MS, cognitive decline is progressive in many patients and correlates more with gray matter atrophy than white matter lesion load. In NPSLE, cognitive impairment is frequently persistent despite systemic disease control. These observations suggest that while immunotherapy addresses the inflammatory/autoimmune insult, secondary neuronal injury (excitotoxicity, axonal loss, synaptic damage) may be irreversible once established.

Predictors of Good vs. Poor Outcome

• Good prognosis: Early treatment, identifiable and removable trigger (e.g., teratoma), younger age at onset (for anti-NMDAR encephalitis), robust initial response to immunotherapy, absence of severe autonomic complications, and low comorbid cardiovascular risk (for NPSLE).
• Poor prognosis: Delayed treatment, ICU admission, status epilepticus, severe autonomic dysfunction, older age at onset, persistent intrathecal antibody production, high cumulative inflammatory burden (disease activity over time), and antiphospholipid antibody positivity in SLE.

Impact of Psychiatric Comorbidity on Autoimmune Disease Outcomes

The relationship between psychiatric symptoms and autoimmune disease outcomes is bidirectional. Depression in MS is associated with greater inflammatory disease activity, faster disability progression, reduced adherence to disease-modifying therapy, and increased mortality. In SLE, comorbid depression is associated with higher disease activity, more flares, and greater organ damage accrual over time. This creates a vicious cycle in which autoimmune inflammation drives psychiatric symptoms, which in turn worsen autoimmune outcomes through behavioral (non-adherence, unhealthy behaviors), neuroendocrine (HPA axis activation, autonomic dysregulation), and immunological (stress-related immune activation) pathways.

The field of autoimmune neuropsychiatry is advancing rapidly, but significant limitations and unanswered questions remain:

Current Limitations

• Lack of randomized controlled trials: With the partial exception of PANDAS, there are no RCTs of immunotherapy specifically for psychiatric manifestations of SLE, MS-related depression, or Hashimoto's encephalopathy. Treatment recommendations are based primarily on cohort studies, case series, and extrapolation from non-psychiatric outcome data.
• Diagnostic biomarker gaps: For NPSLE, there is no single biomarker with adequate sensitivity and specificity to confirm CNS autoimmune involvement. Anti-ribosomal P and anti-NR2 antibodies are associated but not diagnostic. For PANS/PANDAS, the lack of a validated diagnostic biomarker remains a fundamental obstacle.
• Heterogeneity of "neuroinflammation": The term is used broadly to encompass microglial activation, cytokine elevation, autoantibody-mediated processes, and complement activation. These represent distinct mechanisms requiring different treatments. Lumping them together risks overgeneralization.

Emerging Research Areas

• Expanded autoantibody discovery: The number of identified pathogenic anti-neuronal antibodies continues to grow. Beyond anti-NMDAR, antibodies targeting LGI1, CASPR2, AMPAR, GABA-B, GABA-A, DPPX, and IgLON5 have been identified, each with distinct clinical syndromes. Many patients with suspected autoimmune encephalitis remain "seronegative" on currently available panels, suggesting undiscovered antibodies.
• Precision immunotherapy: Research is moving toward tailored immunotherapy based on the specific autoimmune mechanism. For antibody-mediated disorders, rituximab (B cell depletion) is increasingly used as early second-line therapy. For T cell–mediated processes, different agents may be needed. Biomarker-guided immunotherapy selection is an active area.
• The gut-brain-immune axis: Emerging evidence links gut microbiome composition to both autoimmune disease activity and psychiatric symptoms. Altered microbiota have been documented in MS, SLE, and major depression, with shared patterns of reduced diversity and increased proinflammatory species. Clinical trials of probiotics and fecal microbiota transplantation in autoimmune and psychiatric conditions are underway.
• Neuroimaging biomarkers: Advanced neuroimaging techniques — PET with microglial activation tracers, magnetic resonance spectroscopy measuring glutamate and glutathione, and novel MRI sequences detecting cortical and meningeal inflammation — offer the prospect of identifying neuroinflammation in vivo before irreversible neuronal damage occurs.
• The SINAPPS2 trial: This ongoing UK-based randomized trial is testing whether immunotherapy (IVIG + rituximab) improves outcomes in patients with psychosis who test positive for anti-neuronal antibodies. Results will be pivotal for establishing whether routine antibody screening in first-episode psychosis is warranted and whether a distinct "autoimmune psychosis" entity can be treated with immunotherapy.

The autoimmune neuropsychiatric disorders reviewed here — NPSLE, MS, autoimmune thyroid disease, PANDAS/PANS, and anti-NMDA receptor encephalitis — collectively demonstrate that the boundary between "medical" and "psychiatric" illness is artifactual and clinically dangerous. Psychiatric symptoms in these conditions are not ancillary; they are often the presenting and most disabling features. Failing to recognize the autoimmune etiology results in ineffective treatment, disease progression, and in some cases, preventable death.

For practicing clinicians, the key implications are practical: maintain a high index of suspicion for autoimmune etiologies in atypical, acute-onset, or treatment-resistant psychiatric presentations; pursue appropriate laboratory and neuroimaging workup in a staged fashion; collaborate across disciplines; and recognize that immunotherapy, not psychotropic medication, is the primary treatment when autoimmune mechanisms are driving psychiatric symptoms.

The broader neuroinflammation paradigm, while still maturing, suggests that even among patients who do not meet criteria for a specific autoimmune disorder, inflammatory processes may contribute to psychiatric symptom burden in a meaningful subset. As biomarkers improve and clinical trials mature, the integration of immunological thinking into psychiatric practice will likely expand, offering new treatment options for patients who currently derive insufficient benefit from conventional approaches.

This article is intended for educational purposes only and does not constitute medical advice. Autoimmune neuropsychiatric conditions require multidisciplinary evaluation and individualized treatment planning by qualified specialists.