NMDA Receptor Encephalitis

By Dr Ed Needham

To me, the autoimmune encephalitides epitomise the advent of neurology as an acute speciality. Advances over the last decade have seen the delineation of previously undefined disorders into semiologically distinct illnesses with correlating autoantibodies. This has re-invigorated the relationship between neurology and intensive care medicine because of the requirement for critical care during aggressive immunosuppression which often leads to remarkably good outcomes.

Illustrative Case: Miss A is a 30 year old previously well accountant, who 4 weeks previously developed a flu-like illness with fever and mild headache. After 3 days, her partner noticed a marked behaviour change with agitation, persecutory delusions and hallucinosis. Her GP referred for an urgent psychiatric review but before this could happen she began to exhibit repetitive involuntary facial movements and chorea, suggesting a more organic pathology. She was referred to the acute medical unit, where she became progressively comatose and began seizing.

MRI brain revealed right mesial temporal hyperintensity on FLAIR sequencing.

This presentation is typical of NMDA receptor encephalitis, the most common of the antibody-mediated encephalitides. First described in 2007, the typical patient is a young woman (8:2 female:male), who exhibits a viral-like prodrome, before developing psychiatric disturbance, movement disorders, encephalopathy and seizures. Dysautonomia is very common (mandating haemodynamic monitoring), and can mimic sepsis by disordered thermoregulation.

Routine investigations should include blood and CSF antibody titres to the NMDA receptor (NMDAr) (~15% have positive CSF, but negative serum), MRI brain (often normal, but may show non-specific abnormalities), routine CSF (often mildly lymphocytic, sometimes with unmatched CSF oligoclonal bands) and EEG (which may show the pathognomonic “delta-brush” pattern – delta activity with superimposed beta). In 10-50% of patients, an ovarian teratoma is responsible for driving the autoimmune process, and an exhaustive search is essential, as removal of the tumour can be of profound therapeutic benefit. A comprehensive workup should include CT chest/abdo/pelvis, transvaginal ultrasound, MRI ovaries and PET-CT.

Management principally revolves around immunosuppression and seizure control. Seizure control can be incredibly difficult. Firstly, deciphering true seizures from non-epileptic movement disorders is hard clinically, and even with EEG correlation can be ambiguous given the markedly abnormal EEG traces these patients have. Secondly, the seizures are very resistant to anti-epileptic medications, and often only come under control when the underlying immune process is quelled. Immunotherapy should be aggressive, and typically starts with high dose steroids alongside either plasma exchange or intravenous immunoglobulins. Second line therapy is with either cyclophosphamide, rituximab, or both. Other therapies including alemtuzumab and intrathecal methotrexate have been tried, but remain in the realm of case reports. Clearly, such profound immunosuppression in an intubated, critically ill patient is perilous, and pedantic surveillance for infection is important.

Recovery can be painfully slow, with coma persisting for weeks, but good outcome is common even after protracted illness. Nearly 80% of those who are severe enough to necessitate second line treatment have a favourable outcome at 24 months.

The second most common autoimmune encephalitis is driven by antibodies to the voltage-gated potassium channel. This typically appears in older patients, often men, who present with dystonic faciobrachial spasms, hyponatraemia, seizures and coma. Associated tumours are much less common (<10%), but still CT body and PET-CT are prudent. The work-up and treatment are the same, with outcome data paralleling that of NMDAr encephalitis. Encephalitis has also been described with antibodies directed at AMPA, GABA(b), Glycine, mGluR5 and DPPX receptors amongst others, and many antibody-negative variants almost certainly have an as yet unrecognised antibody association.

Comparison should be drawn with the “traditional” paraneoplastic encephalitides associated with anti-Hu/Ma/Ri etc. antibodies. These disorders are invariably associated with underlying malignancy (often occult), and are associated with a much bleaker prognosis. Although antibodies are detectable, these are thought to be the product of underlying cell-mediated autoimmunity, rather than directly pathogenic, and the response to immunotherapy is often very poor.