The Evolving Standard of Care for Autoimmune Neuropsychiatric Illness

Autoimmune Encephalitis and Psychosis

Reports of psychiatric illness associated with autoimmune encephalitis (AE) have increased in the medical literature over the last two decades. A variety of autoimmune diseases, including multiple sclerosis, systemic lupus erythematosus, thyrotoxicosis, Sjögren syndrome, and vasculitis, can cause psychiatric symptoms ranging from subtle personality changes to psychosis and catatonia.⁴ In 2001 and 2006, Diamond and colleagues published research implicating autoantibodies against the N-methyl-D-aspartate (NMDA) receptor in the neuropsychiatric manifestations of lupus.^5,6 In 2005, Dalmau and colleagues⁷ described four cases of young women patients with subacute or acute psychiatric symptoms, including behavioral and personality changes, rapid neurological deterioration including seizures, and altered mental status. In addition to abnormalities seen on brain magnetic resonance (MRI) and positron emission tomography imaging, the patients’ spinal fluid contained antibodies against an unknown protein present in hippocampal neurons. Notably, all four patients were eventually diagnosed with ovarian teratoma. In 2007, Dalmau et al. identified the antigens as subunits of the N-methyl-D-aspartate receptor (NMDAR) and characterized the clinical syndrome as paraneoplastic autoimmune NMDAR-related encephalitis,⁸ now known as anti-NMDAR encephalitis. Of the 12 patients described in the 2007 report, three presented initially with short-term memory loss followed by onset of psychiatric symptoms or altered mental status. Nine of the twelve initially had psychiatric symptoms such as personality or behavioral change, agitation, or paranoia; six of the twelve had initially been evaluated by psychiatrists; and five of the twelve had been admitted to psychiatric units.⁸

Since 2007, numerous additional etiologies for AE have been characterized, with pathogenic antibodies against a variety of neuronal intracellular antigens (including Hu and glutamic acid decarboxylase (GAD)), synaptic receptors (including NMDA and AMPA), or cell-surface proteins (including leucine-rich glioma-inactivated 1 (LGI-1), myelin oligodendrocyte glycoprotein antibody (MOG), and Aquaporin 4).⁹ In light of these findings, clinicians must now consider AE as part of the differential diagnosis for acute psychiatric illness, particularly in cases of psychosis that are atypical in onset or associated with other general medical concerns (e.g., emerging after a neoplasm or viral infection).¹⁰ Of note, while anti-NMDAR encephalitis has become known in part for its prominent psychiatric phenotypes,¹¹ including psychosis, neuropsychiatric symptoms (particularly memory deficits or altered mental status) are common in other forms of AE, including limbic encephalitis.^9,11

As summarized by Graus and Dalmau, the term “autoimmune psychosis” was originally coined “to designate patients with schizophrenia or [first episode psychosis] suspected to be autoimmune in origin on the basis of the detection of serum […] neuronal autoantibodies” (Ref. 10, p 112). These antibodies (such as NMDAR antibodies), however, have been found in one to five percent of both patient groups and healthy controls, and without corresponding cerebrospinal fluid (CSF) antibody findings, or MRI or electroencephalogram (EEG) abnormalities, the presence of serum antibodies is not a specific finding for AE¹² and is not sufficient to warrant immunotherapy.¹³

Table 1 summarizes diagnostic approaches for AE and autoimmune psychosis. In 2016, Graus et al. provided a set of criteria for “possible” AE, based on presence of working memory deficits, altered mental status, or psychiatric symptoms, as well as new focal central nervous system (CNS) findings, new or unexplained seizures, CSF pleocytosis, or MRI features.⁹ The same paper also proposed criteria for “definite” AE subtypes, including anti-NMDAR encephalitis and autoimmune limbic encephalitis.⁹ In 2018, Dubey et al. proposed the Antibody Prevalence in Epilepsy and Encephalopathy (APE2) score.¹⁴ An APE2 score of four or more was 99 percent sensitive and 93 percent specific for the presence of neural-specific antibodies.¹⁴ This score is routinely used in clinical practice to determine the risk of possible or definite AE.¹⁶ In 2020, Pollak et al. proposed criteria, intended for use in psychiatric contexts, for the diagnosis of “possible,” “probable,” and “definite” psychosis of autoimmune origin.¹⁵ A diagnosis of “probable” autoimmune psychosis requires CSF, MRI, or EEG findings; and “definite” autoimmune psychosis requires a finding of IgG class antineuronal antibodies in CSF. Autoimmune psychosis is arguably a variant of AE,¹⁵ and Graus and Dalmau stated that “outside of the context of AE, psychosis [of] autoimmune origin has not been shown to exist” (Ref. 10, p 113). Terminological debate aside, a diagnosis of AE or autoimmune psychosis may have significant implications for a given patient’s treatment and prognosis.

AE may initially present as another neurological or psychiatric condition. A 2018 study retrospectively examined the initial presentations of 50 patients who were eventually diagnosed with AE. In 34 of 50, the admission diagnoses were different: ten patients (20%) had been diagnosed with epilepsy, eight (16%) had been diagnosed with psychiatric disorders, and two (4%) had been diagnosed with dementia.¹⁷ A 2020 case report¹⁸ described a patient who received a diagnosis of functional neurological disorder, and later was diagnosed with unspecified tic disorder, adjustment disorder, and mild cognitive impairment, before eventually being diagnosed with anti-LGI-1 limbic encephalitis. The authors stated that the delay in obtaining a lumbar puncture and CSF analysis delayed the diagnosis of limbic encephalitis.¹⁸

A 2018 comparative study estimated the prevalence of probable or definite AE (based on the 2016 criteria⁹) as 13.7/100,000 person-years, similar to the prevalence of all infectious encephalitis (11.6/100,000) in the study population.¹⁹ But, based on subsequent publications, the prevalence of definitive AE appears to be lower than previously thought, including in patients with psychosis (with or without neurological symptoms). In a 2021 prospective observational study of 105 patients with first-episode psychosis,¹² 20 percent of the patients fulfilled criteria for “possible” or “probable” autoimmune psychosis based on the 2020 criteria.¹⁵ Despite this, none of these 105 patients had neuronal autoantibodies in their serum or CSF, and none were ultimately diagnosed with anti-NMDAR or another AE.¹² At a six-month follow-up, 101 of the patients had a primary psychiatric diagnosis (such as schizophrenia or bipolar I disorder).¹² The remaining four patients had diagnoses of frontotemporal dementia, HIV-associated encephalopathy, hyperthyroid encephalopathy, and encephalopathy secondary to suspected metabolic disease.¹²

A 2023 retrospective study examined 393 cases of patients who had been diagnosed with AE but subsequently received a different diagnosis at an autoimmune neurology clinic.²⁰ Of the 393, 107 patients were found to have been misdiagnosed, and 77 did not fulfill the diagnostic criteria for “possible” AE.²⁰ The updated diagnoses included functional neurological disorder (25% of the 107), neurodegenerative dementia (21%), psychiatric disease (18%), cognitive deficits from comorbidities (10%), cerebral neoplasm (9.5%), and other (17%).²⁰ These results are consistent with AE being rare. The authors of this study noted that the estimated cumulative incidence of AE is three to nine per million person-years, and it accounts for fewer than one percent of first-episode psychosis cases.²⁰

While AE is far less common than primary psychiatric diseases, either a missed diagnosis or a misdiagnosis could delay appropriate treatment and result in harm to the patient. AE has high morbidity and mortality and can involve weeks or months of hospitalization, relapse, and persistent symptoms even after resolution of an acute episode.¹¹ The 2020 consensus criteria state that immunotherapy should be considered in cases of “probable” or “definite” autoimmune psychosis.¹⁵ Treatment for AE is multimodal and involves plasma exchange and immunosuppression, such as high-dose steroids or agents such as methotrexate or rituximab, removal of tumor (if present), and seizure control.¹⁵ Immunotherapy initiated within weeks after symptom onset is recommended to reduce relapses and cognitive impairment and improve functional outcomes.^11,15 A diagnosis of AE, however, may require a protracted period of symptom evolution, evaluation, and testing. Immunotherapy also has substantial side effect burden and cost. Management of psychiatric symptoms associated with AE also requires psychiatric treatment, such as antipsychotic medications, to manage psychosis symptoms, even though patients with AE may be more susceptible to adverse reactions such as neuroleptic malignant syndrome (NMS).¹⁵ Treatment for catatonia in AE may include benzodiazepines or electroconvulsive therapy.¹⁵ Additional treatments include supportive care and rehabilitation: physical therapy, occupational therapy, and psychotherapy for emotional sequelae. Given both the rarity and morbidity of AE, combined with the risks of misdiagnosis and provision of inappropriate treatments, the careful evaluation of a patient with suspected AE carries great significance.

The Existing Standard of Care

A psychiatrist’s duty to exercise due care in diagnosis includes a comprehensive interview and review of the patient’s history, medical records, diagnostic tests, and collateral information.²¹ The psychiatrist is also responsible for performing or referring a patient to a competent physician for a physical and neurological examination and clinically appropriate tests to rule out general medical conditions that could be causing the patient’s mental illness.²¹ As Noffsinger and Magalotti noted in 2022, a psychiatrist may be liable for a failure to evaluate for or diagnose a medical condition, for mistaking a psychiatric illness for another medical condition, or for a failure to manage a medical condition.²²

Claims related to misdiagnosis or delayed diagnosis in psychiatric care may be resolved through settlement. A 1994 article reviewed 34 liability claims filed between 1978 and 1991 against the University of Texas System in its care of adult psychiatric patients.²³ One of these cases, which resulted in an out-of-court settlement with the largest monetary award of the claims surveyed, involved an allegation of negligent failure to diagnose Cushing’s disease. The patient had developed acute psychosis, received three years of psychiatric treatment, including electroconvulsive therapy, and was subsequently diagnosed with Cushing’s disease. The authors noted, “[I]t is vital for psychiatrists to consider organic disorders within a differential diagnosis to provide appropriate medical treatment and avoid this type of malpractice litigation” (Ref. 23, p 467). These findings suggest that for patients exhibiting atypical physical symptoms associated with psychiatric illness, psychiatrists may wish to seek consultation from primary care or other specialty physicians, with documentation of this consultation.

Judicial precedent in this area of misdiagnosis, delayed diagnosis, or inadequate management of medical conditions in patients receiving psychiatric care includes cases where the plaintiff had an endocrine condition,^24,–,26 genetic disorder,²⁷ tumor,^28,–,30 stroke,^30,31 brain infection,^32,33 traumatic injury,^34,35 or complex medical history.³⁶ The cases highlighted below suggest courts will continue to face challenges in adjudicating malpractice claims as the field of psychiatry becomes more complex and intertwined with other medical specialties, as in the evaluation and management of AE.

A finding of malpractice typically requires finding that a physician’s actions deviated from the accepted standard of care. This was reinforced in Anderson v. House of Good Samaritan Hospital in 2007,³³ wherein the plaintiff had received diagnoses of labyrinthitis and depression and was hospitalized involuntarily in a psychiatric ward. According to court records, the plaintiff became catatonic and was eventually transferred to a medical hospital, where he was diagnosed with acute disseminated encephalomyelitis secondary to a viral infection. The plaintiff alleged malpractice against the hospitals and treating psychiatrist.³³ The trial court jury was instructed to consider whether the physician had made an error in judgment, and the court found that none of the defendants were negligent. The Supreme Court of New York ruled that the trial court had erred, and that the “issue of fact” was instead whether the defendant’s assessment and treatment fell short of the medically accepted standard of care.

The standard of care as determined by expert testimony may depend on whether the concern is diagnosis or management of a general medical condition. The 1997 ruling in Vilcinkas v. Johnson³² considered the care of a patient who was hospitalized in a psychiatric unit, transferred to a medical intensive care unit, and diagnosed with herpes simplex encephalitis. The Supreme Court of Nebraska ruled that a family practice physician was able to testify regarding the standard of care for a medical doctor, including the defendant psychiatrist, who was treating a medical condition. In contrast, in another 1997 ruling, that of Whaling v. Joyce,²⁶ the plaintiff alleged a psychiatrist who had been treating her depression and anxiety was negligent in failing to diagnose her Graves disease. The plaintiff’s expert witness was an endocrinologist. The Circuit Court of the City of Fredericksburg, Virginia, granted the defendant psychiatrist’s motion for summary judgment on the basis that the proposed expert testifying regarding the standard of care should have recent clinical experience and knowledge of standards in the defendant’s specialty.

Physicians may be held liable for failing to use appropriate diagnostic procedures, as emphasized by the 1983 ruling in Snow v. State.³⁷ The courts have also recognized that the standard of care may depend on the time period during which a patient receives care. In the 1990 case of East v. United States,²⁴ the plaintiff’s decedent had received treatment for severe depression. Court records stated he was eventually diagnosed with hypothyroidism, but despite treatment and normalization of his thyroid function, his depression persisted, worsened, and he eventually died by suicide. The plaintiff alleged that the decedent’s physicians (including psychiatrists) failed to diagnose and treat his hypothyroidism in a timely manner, which proximately caused his severe depression and suicide. The U.S. District Court of Maryland ruled that “the standard of care between 1983-1985 [when the decedent was being treated] did not require a psychiatrist to perform a thyroid function test on all outpatients complaining of depression” unless they had “signs and symptoms suggestive of an organic illness involving the thyroid” (Ref. 24, p 17).

In order for a delayed diagnosis of a general medical condition to be ruled as malpractice, plaintiffs have had to demonstrate that the delayed diagnosis was a proximate cause of harm,²⁹ that an earlier diagnosis and treatment would have been beneficial,²⁸ or that physicians should have identified a diagnosis earlier based on the patient’s signs and symptoms.³⁵ An adequate psychiatric evaluation may require a screening physical examination, the lack of which was relevant in Zavalas v. State in 1993³⁸ and was deemed a failure constituting inadequate outpatient care in O’Sullivan v. Presbyterian Hospital in 1995.³⁹ Addressing abnormal physical exam findings may be difficult, but no less crucial, in patients with chronic medical conditions and complex histories. In the 1995 case of Deasy v. United States,³⁶ physicians were found to have fallen below the standard of care when they treated the plaintiff patient for bipolar mania, disregarded his medical history including Ormond’s disease (idiopathic retroperitoneal fibrosis) and the adverse effects of medications (including steroids), and failed to provide timely and appropriate treatment of his edema.

The psychiatrist’s scope of care may not necessarily include direct care of general medical conditions, but typically includes knowing whether the patient is under the care of other practitioners, and making appropriate referrals if the patient is not receiving necessary general medical care.^30,40 In the 1985 case of Witt v. Agin,³⁰ a plaintiff alleged malpractice in multiple physicians’ failing to diagnose a growing nonmalignant brain tumor over two years. Court records showed the defendant psychiatrist’s motion for summary judgment was granted on the basis that he did not commit malpractice, because he provided short-term care for the plaintiff’s emotional problems and knew that she was under the care of other doctors for her other medical conditions. The 1988 case of Wozniak v. Lipoff²⁵ emphasized the importance of referring a patient for competent specialty care in a timely manner. According to court records, the defendant internist had been treating the decedent patient’s Graves disease. Over a period of months the patient developed severe anxiety, paranoia, and depression.²⁵ The patient was eventually referred to an endocrinologist, and died by suicide a day after the endocrinology appointment.²⁵ An expert witness psychiatrist testified that the patient’s depression was most likely related to the Graves disease, and the Supreme Court of Kansas affirmed in the ruling that negligence had occurred.

Psychiatric facilities routinely rely on a general medical hospital or emergency department (ED) to provide “medical clearance” prior to admission or may transfer a patient to an ED or medicine service for care of acute general medical conditions. But, depending on a psychiatric facility’s own scope and resources, courts may consider if both the facility and clinicians working in the facility have a duty of directly evaluating patients’ general medical conditions. In Wilburn v. Cleveland Psychiatric Institute in 1998,⁴¹ the plaintiff alleged that the psychiatric facility’s inadequate evaluation of the patient’s stroke-like symptoms, which had been provisionally diagnosed as conversion disorder, resulted in the patient’s second stroke. The trial court ruled that the facility was not negligent because it should not be held to the same standard of care as a general medical hospital in the diagnosis and treatment of stroke. But, the Court of Appeals of Ohio reversed the judgment on the basis that the facility had claimed to have the capability to address patients’ medical needs and did not provide the necessary evaluation to rule out conversion disorder. In the similar case of Brodowski v. Ryave in 2005,³¹ a plaintiff with initial diagnosis of stroke versus conversion reaction was briefly admitted to a psychiatric unit and subsequently developed worsening stroke symptoms, with imaging findings showing an infarction and carotid artery clot. The trial court jury returned a verdict of “no negligence” against multiple defendant physicians and the Superior Court of Pennsylvania affirmed. But, a dissenting opinion stated there was evidence that the inpatient psychiatrist did not adequately screen the referral to the psychiatric unit.³¹

Litigation in Autoimmune Encephalitis

There are multiple cases of malpractice litigation regarding patients with infectious meningoencephalitis who had a missed diagnosis, delayed diagnosis, or delayed treatment.^{32,33,42,–,44} Cases involving nonpsychiatric patients with infectious meningitis have resulted in multimillion dollar verdicts or settlements for the plaintiffs.^42,–,44 To our knowledge, however, there is sparse judicial precedent regarding medical negligence in failing to diagnose or treat an AE in psychiatric contexts.

In the 2022 case of Schultz v. Ercole,⁴⁵ inpatient pediatricians were defendants in a suit alleging deliberate indifference toward a child patient who experienced neuropsychiatric symptoms including pain, facial paralysis, panic attacks, and hallucinations in 2019. The child was initially diagnosed with anxiety, and the differential diagnosis eventually included Pediatric Autoimmune Neuropsychiatric Syndrome (PANS), Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal Infections (PANDAS), AE, conversion disorder, and a stress or adjustment disorder.⁴⁵ Court records stated that the child was treated with intravenous immunoglobulin (IVIG). During a readmission, a consultant neurologist diagnosed the child with PANS. The lawsuit claimed that the defendant pediatricians were deliberately indifferent to the child’s medical needs by disregarding the PANS/PANDAS diagnosis. The U.S. District Court for the Western District of Texas dismissed the claim and ruled that the defendants had considered multiple possible diagnoses, had considered the consultant’s opinion, had consulted literature on PANS/PANDAS, and had ordered appropriate treatments. The court noted that the consultant who diagnosed PANS had not directly examined the child or performed diagnostic testing. The court also noted that negligence or malpractice do not constitute deliberate indifference, which is a particularly high standard to meet. The court stated that physicians do have a duty to explore “every possible explanation” (Ref. 45, p 8) for their patient’s illness.

An Uncertain Standard of Care

Because of the rarity and variable presentations of AE, which many clinicians have never encountered or treated, patients with AE may experience progressive illness over weeks or months before they are able to receive appropriate treatment. In considering the possibility of AE as part of a differential diagnosis, one approach is to utilize the APE2 risk stratification score,¹⁴ as well as the presence or absence of clinical “red flag” signs or symptoms as described by the authors of the 2020 Consensus Criteria.¹⁵ Several “red flags” (recent tumor diagnosis, movement disorder, severe or disproportionate cognitive dysfunction, decreased consciousness, unexplained seizures, or significant autonomic dysfunction) would increase the APE2 score and also form part of the proposed diagnostic criteria for “possible” autoimmune psychosis. Other red flags (recent infection, new-onset or changed headache, focal neurological signs, unexplained hyponatremia, or a history of autoimmune disorders such as lupus or autoimmune thyroid disease) are less specific but would still raise suspicion and warrant consideration of medical contributors to the patient’s psychiatric symptoms.¹⁵

But, an early presentation of AE may be limited to psychiatric symptoms, without clear red flags. Some patients may not even show characteristic MRI or EEG findings until later in the disease course.¹⁸ A 2013 observational cohort study of 571 patients diagnosed with anti-NMDAR encephalitis found that 23 patients (4%) had isolated psychiatric symptoms (most commonly psychotic symptoms with a mood component) without neurological symptoms.⁴⁶ For five patients with isolated psychiatric symptoms who were eventually diagnosed with anti-NMDAR encephalitis, the time from symptom onset to treatment ranged from two to 60 weeks, with a median of nine weeks.⁴⁶ A significant limitation of the 2020 Consensus Criteria is that they may be less helpful in cases with isolated¹⁰ or more chronic psychiatric symptoms.

If a delay in pursuing AE evaluation would be considered an error in judgment or a deviation from the acceptable standard of care would depend on the specific clinical circumstances. Substantial clinical judgment is involved in deciding whether to refer a patient for additional specialty care or additional testing. Additionally, factors outside the control of the treating physician can also affect the standard of care for what a “reasonable practitioner” would or could do in similar circumstances. Depending on the patient’s geographic location, socioeconomic status, insurance status, and other characteristics, access to a neurologist and to diagnostic testing may be limited and may require referral to a facility at a significant distance from the original point of care. Disparities in access may be further heightened by socioeconomic disparities.

Even after a CSF sample is obtained, testing for neuronal autoantibodies in CSF may take weeks to obtain. According to the Mayo Clinic Laboratory’s website, the Autoimmune/Paraneoplastic Encephalopathy panel takes eight to 12 days to obtain the results.⁴⁷ The Clinical Laboratories Test Directory at the University of California, San Francisco states the result time for this test is four to 10 days.⁴⁸ Whether a patient has access to a tertiary medical center with experience evaluating for AE, including familiarity with expediting the necessary send-out tests, could be an additional factor shaping a patient’s care. Evaluation for AE is extensive, time-intensive, and costly,⁴⁹ as it involves specialists, invasive procedures, neuroimaging, and specialty laboratory tests, all before immunotherapy. Physicians must reconcile these diagnostic and care access challenges with ongoing systems pressure to practice cost-conscious care,^50,51 including not ordering unnecessary or low-yield interventions,⁵² as well as pressure to shorten the lengths of hospitalizations.^53,–,55 Patients may also encounter difficulty with affording treatment. In 2015, a patient with AE filed suit against her health insurers and the New York State Department of Financial Services after they denied coverage for her rituximab and IVIG immunotherapy.⁵⁶ According to 2016 court records, the case was discontinued (additional information was not provided).