Case

A woman in her mid-20s, with no personal or family history of mental illness, presented to her general practitioner with anxiety and low mood. Sick leave from her stressful job provided no relief. Over 5 weeks, her mood, appetite and sleep rapidly worsened. She experienced severe panic attacks and later, an auditory hallucination on at least one occasion. Mirtazapine and zopiclone provided no immediate benefit, and an urgent referral to the community mental health team was made.

Her behaviour became acutely disturbed at home. She rapidly alternated between being angrily agitated and mutely staring. She spoke in strange voices, and the content could be nonsensical at times. Occasionally, she randomly burst into song. She repeatedly squatted and bit her lip. She followed her parents around the house, fearing harm could come to them. At different points in time, she complained of having a ‘sore head’ and ‘floppy’ legs.

Given the severe deterioration in her mental health, an emergency psychiatric admission using mental health legislation was arranged. On admission, routine physical observations (including temperature), examination and blood tests (including white cell count and C-reactive protein) were normal.

On the ward, she appeared profoundly depressed, with extremely impoverished and latent speech. At times, she also appeared fearfully hypervigilant or perplexed. She exhibited mutism, staring, stereotypy and mannerisms – features suggestive of catatonia. Her social judgement was impaired and she lacked insight. She required intense one-to-one nursing care and direction with washing and dressing, implying severe cognitive dysfunction (no formal testing was done).

The initial working diagnosis was a severe depressive episode with psychotic symptoms (ICD-10 code F32.3), given the prominence of her severely depressed mood, which preceded the disorganised behaviour, impoverished speech and psychomotor retardation.

Over 4 weeks on the ward, her mood partially lifted in response to treatment with venlafaxine, olanzapine (low dose) and regular lorazepam (for catatonia). After some uneventful passes home, early in-patient discharge was facilitated by her very supportive parents.

However, upon review in clinic a few days later, concerns were raised regarding an apparent worsening in her cognition (e.g. gross disorientation and poor memory) and motor skills, both fine (e.g. button opening) and gross (e.g. descending stairs). Furthermore, abnormal involuntary movements were observed in one hand.

Given the onset of new focal neurological deficits, it was decided to take blood tests to screen for autoimmune encephalitis. Fortuitously, results were returned within 24 hours (by the local neuroimmunology laboratory), and revealed the presence of serum N-methyl-D-aspartate receptor (NMDAR) antibodies.

Urgent admission to the neurology department was arranged, where work-up and treatment for probable anti-NMDAR encephalitis began. Cerebrospinal fluid (CSF) analysis also detected NMDAR antibodies, thereby confirming the diagnosis. Notably, magnetic resonance imaging (MRI) head and electroencephalogram (EEG) results were normal.

The encephalitis progressed despite treatment with steroids and intravenous immunoglobulins, resulting in fluctuating consciousness and autonomic instability for 1–2 weeks. Rituximab was then commenced, to which there was a gradual, marked and sustained response. Ovarian teratoma and other tumours were excluded through imaging. At 1 year after disease onset, the patient's recovery continued at home.

In Box 1, the patient in this case describes her experience.

Learning objectives

1. (a) What is autoimmune encephalitis?

2. (b) What is anti-NMDAR encephalitis?

3. (c) When should anti-NMDAR encephalitis be suspected?

4. (d) How should suspected anti-NMDAR encephalitis be investigated?

5. (e) How should the practical management and treatment of anti-NMDAR encephalitis be approached?

6. (f) What is the prognosis for anti-NMDAR encephalitis?

Autoimmune encephalitis

Autoimmune encephalitis refers to autoantibody-mediated inflammation of the brain. It manifests with a pattern of psychiatric and/or neurological symptoms, dependent on the specific autoantibody involved. Red flags for autoimmune encephalitis in patients with psychosis are shown in Box 2; these are useful when considering differential diagnoses, the indication for immunological testing and guiding discussion on the most suitable initial admitting speciality.

The number of known autoantibodies pathologically implicated in cases of autoimmune encephalitis expands every year, as new discoveries are made. Well-established ones include autoantibodies to the voltage-gated potassium channel (VGKC)-complexed proteins LGI1 and CASPR2, AMPA, and GABA_B.^{Reference Endres, Maier, Leypoldt, Wandinger, Lennox and Pollak4} However, by far the most common type of autoimmune encephalitis, and the one most frequently associated with psychiatric symptoms at onset, is anti-NMDAR encephalitis.^{Reference Rogers, Pollak, Blackman and David5}

Anti-NMDAR encephalitis

NMDAR antibody testing

Once the differential diagnosis of anti-NMDAR encephalitis is considered, the priority investigation should be to test for the presence of NMDAR antibodies in the (blood) serum. Practically, this is easily achievable in mental health settings by means of a simple blood draw in a yellow-/gold-topped (serum separating tube; SST) bottle. For gold-standard testing, a paired CSF sample should also be obtained (and tested) at the same time; however, it is acknowledged lumbar puncture mandates acute hospital admission, and for this reason, in reality, it is often tested at a later stage . Regarding the prospect of routine screening for autoantibodies in patients presenting with psychiatric symptoms, see Box 4.

Importantly, rates of seropositivity of anti-NMDAR antibodies (of all Ig classes targeting GluN1, including IgG) in healthy individuals are surprisingly high, at over 10%.^{Reference Hammer, Stepniak, Schneider, Papiol, Tantra and Begemann25} Moreover, rates of seropositivity in healthy individuals are not significantly different from those in affected individuals; blood brain barrier permeability and activation of intrathecal antibody production are thought to be the major determining factors in whether disease develops.^{Reference Hammer, Stepniak, Schneider, Papiol, Tantra and Begemann25–Reference Yu, Wu, Cao, Li, Liao and Wei28} Therefore, in practice, a positive serum result – even in a patient with psychosis – does not automatically equate to a diagnosis of anti-NMDAR encephalitis.^{Reference Martinez-Martinez and Titulaer29} However, it does mean further investigation is warranted – at a minimum with CSF analysis, EEG and MRI of the head – and expert neurology evaluation should be sought.^{Reference Pollak and Lennox30}

In the majority of anti-NMDAR encephalitis cases, both the serum and CSF have detectable NMDAR antibodies. However, at least one in ten cases are seronegative but CSF-positive.^{Reference Masdeu, Dalmau and Berman31,Reference Gresa-Arribas, Titulaer, Torrents, Aguilar, McCracken and Leypoldt32} Therefore, irrespective of the serum result in cases of high clinical suspicion, CSF testing is indicated to either support or exclude the diagnosis.

Other CSF findings in support of anti-NMDAR encephalitis are a lymphocytic pleocytosis, mildly raised protein concentration and presence of oligoclonal bands restricted to CSF (indicating intrathecal IgG synthesis).^{Reference Dalmau, Lancaster, Martinez-Hernandez, Rosenfeld and Balice-Gordon9,Reference Blinder and Lewerenz33}

In conjunction with medical/neurology colleagues, careful consideration needs to be given to the practical challenges of safely conducting lumbar puncture (and MRI) in a patient who may be agitated and experiencing psychosis. Heavy sedation or general anaesthesia is sometimes required. Psychiatrists can help by ensuring the legal framework pertaining to detention, capacity and consent issues is satisfactorily addressed. Also, the provision of trained mental health nursing staff is conducive to optimising care.

Importantly, these challenges must be overcome to ensure equitable access to crucial investigations for all patients. The overwhelming benefits of early diagnosis and correct treatment with immunotherapy must be prioritised throughout, regardless of the patient's initial treating speciality or hospital location.

Currently, antibody testing via highly specialised national reference laboratories is the only option in many contexts, and can necessitate a turnaround time of several weeks. In the future, accessibility of testing will grow as local laboratories develop their own testing capabilities for the neuronal antibodies most commonly implicated in autoimmune encephalitis.^{Reference Goodfellow and Mackay34}

If delay in obtaining results is encountered in the context of a patient with a worsening presentation (e.g. new neurological signs) and/or the suspicion of an organic aetiology is high, do not wait for the result.^{Reference Rössling and Prüss35} Arrange a medical admission in liaison with neurology, so the patient can benefit from specialist evaluation (including other investigations) and monitoring, and preliminary treatment can be commenced.^{Reference Venkatesan and Geocadin36}

Other investigations

Routine blood test results (including C-reactive protein and white cell count) are most often within normal limits in cases of anti-NMDAR encephalitis.^{Reference Liu, Liang, Liu, Luo, Liu and Fan37} Neuroimaging results are also most often normal. For example, MRI of the head was found to be normal in 67–89% of cases.^{Reference Irani, Bera, Waters, Zuliani, Maxwell and Zandi7,Reference Titulaer, McCracken, Gabilondo, Armangué, Glaser and Iizuka15} When abnormalities are found, they are non-specific, e.g. mild or transient white matter changes in any region of the brain.^{Reference Lee and Lee38} Compared with MRI, computed tomography imaging of the head provides inferior detail for detection of any central nervous system inflammatory changes that may be seen in autoimmune encephalitis, and therefore MRI is much preferred.^{Reference Granerod, Davies, Mukonoweshuro, Mehta, Das and Lim39} Notwithstanding, as a widely and rapidly accessible modality, computed tomography remains useful for urgent exclusion of gross central nervous system pathology, such as space-occupying lesions and intracranial haemorrhage. EEG findings are abnormal in 90% of anti-NMDAR encephalitis cases,^{Reference Titulaer, McCracken, Gabilondo, Armangué, Glaser and Iizuka15} but are generally non-specific (focal or diffuse slowing, disorganised or epileptic activity). Nevertheless, it is suggested that an ‘extreme delta brush’ pattern is a unique feature of anti-NMDAR encephalitis.^{Reference Schmitt, Pargeon, Frechette, Hirsch, Dalmau and Friedman40}

Crucially, therefore, normal results from any, or even all, of these investigations (routine blood tests, computed tomography, MRI and EEG) do not rule out a possible diagnosis of anti-NMDAR encephalitis. This reinforces the diagnostic importance of antibody testing in serum and CSF in suspected cases.

In confirmed anti-NMDAR encephalitis cases, whole-body imaging for tumours should be conducted.^{Reference Dalmau, Lancaster, Martinez-Hernandez, Rosenfeld and Balice-Gordon9,Reference Lee and Lee38} Specifically, ovarian teratoma in female adolescents and young adults should be searched for by means of ultrasound scan and MRI of the pelvis.

Diagnostic criteria

Clinical criteria for probable and definite anti-NMDAR encephalitis were developed by expert consensus in 2016, with the aim of advancing earlier diagnosis and immunotherapy (Box 5).^{Reference Graus, Titulaer, Balu, Benseler, Bien and Cellucci41}

Box 5

Diagnostic criteria for anti-NMDAR encephalitis (reproduced with permission from Graus et al).^{Reference Graus, Titulaer, Balu, Benseler, Bien and Cellucci41}

Notably, these criteria adopt a neurology focus, giving rise to questioning of their practical utility for psychiatrists in a mental health setting.^{Reference Pollak, Lennox, Müller, Benros, Prüss and van Elst1} Of specific concern is the fact that they do not support diagnosis in those presenting with primarily or exclusively psychiatric symptoms, despite evidence that such individuals exist.^{Reference Kayser, Titulaer, Gresa-Arribas and Dalmau42}

Treatment of anti-NMDAR encephalitis

Prognosis

A validated predictive prognosis tool exists, called the anti-NMDAR Encephalitis One-Year Functional Status (NEOS) score.^{Reference Balu, McCracken, Lancaster, Graus, Dalmau and Titulaer50} It is scored on five categorical variables, with the presence of each worsening prognosis: intensive care unit admission, treatment delay of >4 weeks, lack of clinical improvement within 4 weeks, abnormal MRI and CSF white blood cell count of >20 cells/μL.

Following acute recovery from anti-NMDAR encephalitis, transfer to a specialist neuro-rehabilitation facility is often appropriate. Baseline and repeated neurocognitive assessments assist in comprehensive monitoring of the patient's functional recovery. Over 18 to 24 months, with treatment, eight out of ten cases make a gradual, but significant recovery.^{Reference Titulaer, McCracken, Gabilondo, Armangué, Glaser and Iizuka15} Prolonged deficits can include executive dysfunction, impulsivity, disinhibition and sleep abnormalities. Patients often have little memory of their illness.

Relapses occur (12% risk within 2 years), more often in cases not associated with a tumour at presentation, and the majority are of a less severe nature compared with the index illness.^{Reference Titulaer, McCracken, Gabilondo, Armangué, Glaser and Iizuka15} Relapses require further treatment with immunotherapy. Notably, a quarter of relapses present with psychiatric symptoms only.^{Reference Kayser, Titulaer, Gresa-Arribas and Dalmau42} Therefore, psychiatrists need to be alert to this fact in someone with a past medical history of anti-NMDAR encephalitis.

Case reflections

In retrospect, the case described many of the red flags for autoimmune encephalitis, and nearly all of the ‘SEARCH For NMDAR-A’ diagnostic clues were indeed present. This emphasises the importance of raising awareness of features suggestive of anti-NMDAR encephalitis among psychiatrists, to enable its early recognition.

Furthermore, this case demonstrates the importance of a fully comprehensive and systematic specialist work-up, including serum/CSF antibody testing, in suspected cases of anti-NMDAR encephalitis. In this instance, the patient's MRI head and EEG results were normal, despite the relatively advanced stage of the encephalitis at the time of these investigations. The true diagnosis was only revealed by antibody testing of serum and CSF.

Finally, doctors must be mindful of the risk of diagnostic overshadowing in such cases where psychiatric and physical elements coexist, and be prepared to advocate for their patients to ensure they receive appropriate medical care.