FOSUN HEALTH Logo

Autoimmune Encephalitis Overview

catalogue

Autoimmune Encephalitis Overview

Autoimmune Encephalitis Overview

Autoimmune encephalitis is inflammation of the brain caused by the immune system rather than by an infection. It usually begins over days to weeks, and it very often begins as a change in behaviour, mood or thinking rather than as something that looks neurological — which is why it is so frequently mistaken at first for a psychiatric illness, for delirium, or for an infection. That mistake is expensive, because this is one of the few serious brain diseases where starting treatment earlier measurably changes how much of the person comes back. This guide follows the order families actually need: what the first change looks like and why it is missed, which antibody is behind it and why that changes the plan, how severe it can become and what recovery really looks like, and the two timing rules that should never wait for a test result.

Autoimmune Encephalitis Overview
  • written opinion once complete records arrive

  • neurology, neurocritical care, neuropsychology and rehabilitation planning together

  • serum and cerebrospinal fluid cell-based assay alongside 3.0T MRI and long-term video EEG

  • countries and regions our international patients come from

What autoimmune encephalitis is — and what it is not

Encephalitis means inflammation of the brain. Autoimmune encephalitis means the inflammation is caused by the person's own immune system rather than by a virus or a bacterium. Antibodies, or the immune cells behind them, target proteins that nerve cells need in order to signal. The result is a brain that works, but works wrongly — and, crucially, a brain being damaged by something that can usually be switched off.

The most important thing to understand is that the immune attack is the disease, and time spent letting it run is brain spent. This is not a degenerative condition unfolding on its own schedule. It is an active, treatable process, and the evidence points one way: the earlier the attack is interrupted, the more complete the recovery tends to be.

The illness is subacute. It builds over days to a few weeks and, by convention, within about three months — much faster than a dementia, much slower than a stroke. And the opening symptom is often not what people expect: new fear, agitation, hallucinations, delusions or a personality change in someone with no psychiatric history is one of the recognised ways this begins, which is why the first clinician to see the person is often a psychiatrist or an emergency physician rather than a neurologist.

What it is not: not an infection, although infection must be excluded first and fastest; not a primary psychiatric illness, although one can look almost identical and can also coexist; not a degenerative dementia and not a stroke; not caused by stress, by parenting, or by anything the person did. In most cases no external trigger is ever found, and the absence of one is not evidence against the diagnosis.

Two sentences decide almost everything that follows. The antibody family decides the plan: antibodies against proteins on the cell surface usually respond to immunotherapy, while antibodies against proteins inside the cell usually point to a tumour and respond less well. And a negative antibody panel does not exclude the illness — the panel is finite, assays differ, and antibody-negative autoimmune encephalitis is a recognised diagnosis when the criteria are met. The diagnosis page sets out how that judgement is made.

The antibodies, and what each one looks like

The antibody is not a label for its own sake. It names what is being attacked, and that decides two things: whether a tumour has to be hunted for and how hard, and which immunotherapy is likely to work. The table is ordered by that distinction rather than by frequency.

Antibody How it presents Look for alongside What the result changes
NMDA receptor Psychiatric symptoms first in many — fear, agitation, hallucinations, catatonia — then memory failure, seizures, reduced consciousness, mouth and jaw movements, autonomic instability Ovarian teratoma, most often in younger women; a viral-like prodrome; new psychiatric or movement symptoms weeks after herpes encephalitis Early immunotherapy; tumour removal where a teratoma is found; intensive care support is often needed
LGI1 Brief stereotyped jerking of one arm and the same side of the face; temporal lobe seizures; short-term memory failure; low sodium Onset most often in men over fifty; the movement often precedes the memory change — the window in which treatment prevents the decline Immunotherapy treats the movements too; antiseizure medicine alone controls them poorly; sodium is monitored
CASPR2 Confusion and memory change with nerve over-excitability: cramps, twitching, sweating, weight loss, severe insomnia, neuropathic pain Thymoma; insomnia with autonomic overactivity and confusion is sometimes called Morvan syndrome Immunotherapy plus a thymic tumour search; symptomatic treatment of pain, sweating and sleep
AMPAR Limbic encephalitis — memory failure, confusion, seizures — sometimes with prominent psychiatric symptoms Thymus, lung or breast tumour more often than in other limbic presentations; relapse is described Immunotherapy with a thorough tumour search, repeated if the first screen is negative
GABAA receptor Frequent or hard-to-stop seizures, sometimes status epilepticus, with confusion and behaviour change Thymoma in a proportion; imaging may show scattered abnormalities rather than one lesion Immunotherapy alongside antiseizure treatment; the seizures often need immunotherapy to settle
DPPX Diarrhoea and weight loss often precede the illness, followed by confusion, tremor, unsteadiness and an exaggerated startle A gastrointestinal prodrome lasting months, easily attributed elsewhere; the part patients rarely volunteer Immunotherapy; ask about the prodrome specifically, because it is the clue most often missed
IgLON5 A sleep disorder dominates — dream enactment, breathing pauses, insomnia — with unsteadiness, swallowing difficulty and later cognitive change Later onset; a slow course that can look neurodegenerative; a sleep study is often the first useful test Response is less predictable than in the rest of this group; sleep, airway and swallowing support are central
Intracellular group (Hu, Yo, Ri, Ma2, CRMP5, amphiphysin, GAD65) Limbic encephalitis, sensory loss, cerebellar unsteadiness, brainstem problems or stiffness; onset more gradual, course more fixed An underlying tumour in most cases — small cell lung, thymoma, breast, testicular, ovarian; GAD65 often without one Finding and treating the tumour is the dominant step; immunotherapy response is poorer; the search is repeated over following years

Two points follow. The same antibody can present differently at different ages, and the absence of the textbook picture is not evidence against it. More than one antibody can be present, and then the clinical picture — not the laboratory list — decides what is treated first.

How severe it becomes, and how recovery is measured

This illness has a wider range than almost anything else in neurology. One person is irritable and forgetful on a general ward; another is ventilated with seizures that will not stop and blood pressure that swings without warning. Both are the same disease, and neither tells you the final outcome. What families need is two separate measurements: how severe the acute illness is now, and what the recovery curve looks like from here.

Level or phase What it looks like What it means for care What it does not tell you
Awake on a ward Confused, forgetful, behaviourally changed, but alert, swallowing and breathing normally Immunotherapy on a ward; safe environment, seizure observation, swallowing assessment, early rehabilitation referral It does not predict the outcome. Mild presentations can worsen quickly
Fluctuating consciousness Alertness rising and falling over hours, sometimes with hallucinations between clear periods Continuous observation and a low threshold for escalation; the fluctuation is the encephalitis, not a psychiatric illness It is not deliberate, and it is not the ward environment
Seizures or status epilepticus Repeated fits, or a fit that does not stop between doses of medicine Intensive care, EEG monitoring, anaesthesia where needed; antiseizure medicine escalated while immunotherapy starts in parallel Poor seizure control on medicine alone is not treatment failure — in several subtypes the seizure is the immune disease
Autonomic instability Blood pressure and heart rate swinging widely, temperature episodes without infection, sweating Monitored setting; certain antipsychotics used cautiously because they can worsen this It is not anxiety and not caused by the medication alone; it settles as the inflammation settles
Reduced breathing drive Shallow or irregular breathing, sometimes needing ventilation, most often in NMDA receptor disease Intensive care with ventilatory support. This is a phase rather than a verdict Time on a ventilator is not a measure of brain damage; the person is often aware of far more than assumed
The recovery months Improvement that is uneven: weeks of plateau, then a step forward; fatigue, sleep and mood lagging behind Rehabilitation throughout, reviewed rather than restarted at each plateau A plateau, or a worsening during a steroid reduction, is common and is not the same as failure
Relapse Return of symptoms after clear improvement, most often during or after a reduction of immunosuppression Re-escalation of immunotherapy; the maintenance plan exists to catch this early A relapse does not mean the original treatment failed. Most relapses respond

The order of recovery is usually the reverse of the order of onset. Consciousness and autonomic stability return first, then movement, then behaviour, and memory, attention and executive function last — which means a family may see the person "return" while the person is still struggling with exactly the functions that independence depends on. Fatigue, sleep and mood trail behind everything else. Nobody can promise a full return to the previous level; what can be said honestly is that improvement continues over months and that rehabilitation is what converts neurological recovery into a life. The recovery page sets out how that phase is managed.

How the diagnosis is assembled

No single test makes this diagnosis. It is assembled from a timeline, an examination, spinal fluid, imaging, an antibody panel and a tumour search — and the order matters, because two of these steps have to happen before their results are available.

Step What it involves What it establishes What it does not rule out
The timeline A structured account from someone who knew the person before: what changed first, in what order, from roughly which date Whether the onset is subacute — an entry criterion — and whether behaviour, memory or seizures came first Nothing on its own. It is the frame every other result is read inside
Examination and bedside cognition Neurological examination plus short cognitive testing, repeated over days rather than once Focal findings, abnormal movements, or a memory pattern that fits a limbic process A normal early examination does not exclude it. Findings accumulate
Blood tests Inflammatory markers, thyroid function, vitamin B12, autoimmune serology, and the metabolic and toxic causes that imitate this picture Whether something reversible is contributing; a baseline before immunosuppression Normal blood tests do not exclude encephalitis. This is diagnosed in the brain and the spinal fluid
Lumbar puncture Cell count and differential, protein, glucose, oligoclonal bands, cytology where indicated, and microbiology including herpes simplex PCR Whether there is inflammation; whether an infection is present; whether bands indicate a local immune response A normal first tap does not exclude it. Cell counts can be normal early or after steroids
Contrast-enhanced MRI at 3.0T Brain imaging read by someone looking for the patterns of encephalitis, not only for a lesion Signal change in the medial temporal lobes or elsewhere; whether a structural alternative exists A normal scan is common, particularly in NMDA receptor disease, and must not delay treatment
EEG, including long-term video EEG Routine recording, extended where seizures or fluctuating consciousness are in question Slowing consistent with encephalitis, or seizures — including subtle ones that look like behaviour An abnormal EEG is supportive, not specific; a normal one excludes neither
Neuronal antibody panel Cell-based assay on both serum and cerebrospinal fluid, with tissue-based assay as an adjunct Which protein is targeted, and therefore which family — cell-surface or intracellular — the illness belongs to A negative panel does not exclude antibody-negative disease; a low-titre serum-only result needs clinical reading
Tumour screening CT or PET-CT of chest, abdomen and pelvis, pelvic imaging, testicular ultrasound where relevant, mammography; whole-body PET-CT if the first screen is negative and the antibody is intracellular Whether a tumour is driving the immune response — which, in the intracellular group, changes everything A negative first screen is not the end. In that group, screening is repeated over following years

Two rules govern the timing, and both are about acting before a result arrives. First, aciclovir is started on suspicion — while infection is excluded, not after it is confirmed — because herpes simplex encephalitis is treated before it is proven. Second, immunotherapy is not held for the antibody result, which takes days to weeks, because earlier treatment is associated with a more complete recovery. The clinical criteria authorise treatment: a subacute change in memory, mental state or behaviour, plus at least one of new seizures, new focal findings, spinal fluid inflammation or MRI changes, after reasonable exclusion of alternatives. The diagnosis page carries the full sequence.

Why patients travel to China for an encephalitis assessment

Patients come for three reasons: an assessment that has not converged, a subtype that has not been named, or a plan that is not being executed at the pace the illness requires. The value of a second opinion here is not a different drug — it is a team that runs the whole sequence in one pathway and says plainly which criteria are met and what would change the answer.

What the pathway looks like. Neurology, neurocritical care, neuropsychology, neuroradiology and rehabilitation plan together. Contrast-enhanced 3.0T MRI, dual-source CT, PET-CT, 3D DSA and long-term video EEG sit in the same programme, as do lumbar puncture with spinal fluid studies and the serum and spinal fluid antibody panel. Where a tumour is suspected, screening is organised as one block. A written opinion is returned within 48 hours of complete records arriving. Fully treated costs typically run at 30–50% of comparable care in Europe or the United States, more than 30 commercial insurers are billed directly, and our international patients come from 124 countries and regions.

What a records-only opinion can and cannot do. It can say whether the criteria are met, whether the exclusion work is complete, which antibody is likely and which family it belongs to, whether the tumour screen was adequate, and whether the immunotherapy given was the right intensity. It cannot examine the person, repeat the tap or run a scan — and those are sometimes exactly what the case turns on. We say so rather than implying otherwise, and where a visit would add nothing we say that too, before any payment is discussed. What must be said about timing: this is not a destination for an acutely unstable person. Someone in status epilepticus, with an unstable airway or unpredictable blood pressure, must be stabilised where they are. Send the timeline first — it is the most valuable document in the file.

Frequently Asked Questions

Can autoimmune encephalitis really look like a psychiatric illness?

Yes, and this is the commonest reason the diagnosis is delayed. New fear, agitation, hallucinations, delusions or a personality change in a person with no psychiatric history is one of the recognised ways this illness starts, and the person is frequently seen first in a psychiatric or emergency setting. The features that should make anyone look further are a change that is new rather than lifelong, an onset measured in days to weeks, memory failure alongside the psychiatric symptoms, any seizure, any abnormal movement, any fluctuation in consciousness, and a poor response or an unusual sensitivity to antipsychotic medicine. A primary psychiatric illness remains possible and often coexists, but it should be a conclusion reached after the neurological assessment, not before it.

How fast does it start — would we have noticed?

Almost always yes, and the speed is one of the diagnostic clues. Autoimmune encephalitis is subacute: it develops over days to a few weeks and, by convention, within about three months. That is much faster than a degenerative disease and much slower than a stroke, which arrives in minutes. The practical consequence is that a written timeline is one of the most useful things a family can bring: what the person was like before, what changed first, and on roughly which date. That timeline often does more to establish the diagnosis on the first day than any single test.

How much of the person comes back, and how long does it take?

Recovery is measured in months, not weeks, and it is usually uneven. Many people make a substantial recovery over the first year or two, and a proportion are left with residual difficulty in memory, attention, mood, sleep or stamina. Nobody can promise a return to exactly the previous level, and the honest answer is a trajectory rather than a guarantee: the severity of the acute phase, how early treatment started, whether a tumour was found and removed, and whether relapse occurred all influence it. The order of recovery is often the reverse of the order of onset — consciousness and autonomic stability return first, then movement, then behaviour, and memory and executive function last, with fatigue, sleep and mood trailing behind everything else.

Do we have to wait for the antibody result before starting treatment?

No, and waiting is the commonest avoidable mistake. The antibody panel can take days to weeks to return, and treatment started earlier is associated with a more complete recovery. What has to happen first is not the antibody result but the exclusion of infection: aciclovir is started on suspicion while cerebrospinal fluid results are awaited, because herpes simplex encephalitis is the one alternative that must not be missed and is treated before it is confirmed. Once the clinical criteria are met — subacute change in memory, mental state or behaviour, plus new seizures, focal findings, cerebrospinal fluid inflammation or MRI changes, with alternatives reasonably excluded — immunotherapy begins. The antibody result then names the subtype and settles the plan, including whether a tumour search needs to be repeated.

Get the timeline read, and the immune attack interrupted

Send a written timeline of what changed and when, the discharge summaries you have, any lumbar puncture results with the cell count and protein, any antibody reports, MRI and EEG reports, a list of current medicines, and — if the episodes were captured — a short video of one of them. Our team will return a written opinion on whether the diagnostic criteria are met, whether the exclusion work is complete, which antibody is likely and what that means for the plan, and what can be established from your records before any travel is arranged.

Want to see the whole process first? Your full medical journey, step by step →

Book an Appointment Request a Video Consultation

Follow Fosun Health

Apply for an Appointment

Do you have test results*