Clinical Cases in Pharmacology Clinical Cases  ·  Neurology IV  ·  Neuroimmunologic/Paraneoplastic  ·  Anti-NMDA Receptor Encephalitis: Immunotherapy Escalation Timing
Neurology IV, Case NeuroImmun-0001 — Neuroimmunologic/Paraneoplastic

Anti-NMDA Receptor Encephalitis: Immunotherapy Escalation Timing

Twelve days into first-line immunotherapy for anti-NMDA receptor encephalitis, a patient's dyskinesias haven't budged — and the consensus threshold for adding a second immunosuppressive agent is landing in the same week as her teratoma resection.

Abbreviations, terms, and other agents mentioned in this case NMDAR — N-methyl-D-aspartate receptor, the synaptic target of the pathogenic antibody in this disease  ·  IVIG — intravenous immunoglobulin  ·  PE — plasma exchange (plasmapheresis)
Presentation

Maya T., 24, is two years into veterinary school and plays cello in a community orchestra most Sunday evenings — the kind of double life her friends used to tease her about, since half her waking hours were split between exam rooms and practice rooms. Three weeks ago her roommate noticed she had stopped sleeping and started accusing the mail carrier of following her; within a week she was disorganized enough in speech that a friend drove her to the emergency department instead of her usual clinic. She has no psychiatric history of her own, no substance use, and was, by every account of everyone who knows her, entirely well until this began. Within days of admission she developed orofacial dyskinesias — a chewing, grimacing movement she cannot suppress — followed by stretches of unresponsiveness alternating with agitation, and then blood pressure swings severe enough to move her to the ICU. Cerebrospinal fluid showed a mild lymphocytic pleocytosis; both serum and CSF tested positive for NMDAR antibodies. A pelvic ultrasound, ordered on that result alone before any other workup pointed there, found a 4cm ovarian teratoma, resected five days later.

She is now twelve days into high-dose methylprednisolone and IVIG, started the same day the antibody result came back, and the orofacial dyskinesias have not meaningfully changed; she remains minimally interactive, tracking but not consistently following commands. Titulaer and colleagues’ 577-patient cohort found roughly half of first-line-treated patients show real improvement by four weeks — she is not there yet, but she is not past that window either, which is exactly what makes the next decision genuinely unsettled rather than overdue. The tumor is out, which matters, since ongoing antigen exposure from an unresected teratoma is one of the clearest drivers of a slow first-line response — but antibody-mediated synaptic dysfunction does not necessarily clear on the same timeline the tumor does, and nothing about a clean resection guarantees the next week looks better than this one did.

Maya T. · 24 Day 12, Post-Resection
Presentation
Subacute psychosis → orofacial dyskinesias → autonomic instability over 3 weeks
CSF / serology
Lymphocytic pleocytosis; NMDAR antibody positive, serum and CSF
Current exam
Orofacial dyskinesias persist; minimally interactive, inconsistent command-following
Autonomic status
Improved from ICU nadir; intermittent BP swings continue
Teratoma
4cm, resected postoperative day 5, margins clear
First-line therapy
Methylprednisolone + IVIG, day 12, no clear clinical response yet

ICU, day twelve

Autoimmune Neurology Specialist Opening

Twelve days of first-line therapy with no real movement on the dyskinesias is inside the window international consensus treats as the trigger to add rituximab, not before it. Titulaer and colleagues’ 577-patient cohort found roughly half of first-line-treated patients improve by four weeks, and Nosadini and colleagues’ international consensus recommendations treat non-improvement at around two weeks as the point to start a second-line agent rather than wait for the full four to elapse. That consensus document was written for pediatric NMDAR encephalitis, not for a twenty-four-year-old — but the same two-week mark is what the adult autoimmune-encephalitis literature uses to define first-line treatment failure, so I don’t think the number moves much for her. Rituximab’s own mechanism — depleting the B cells still producing antibody — takes real time to show effect; starting it once she is obviously failing costs us exactly the days we cannot spare with her still in the ICU on this trajectory.

Critical Care Physician Response

You’re right that the two-week threshold is real and that we’re sitting right on it — but the number that threshold is built on is measured from symptom onset or first-line initiation broadly, not from five days post-teratoma-resection specifically, and removing four centimeters of antigen-producing tissue from a twenty-four-year-old is not nothing.

“Costs us exactly the days we can’t spare” assumes the days between now and next week are otherwise empty — they’re not; they’re the exact window in which resection-driven improvement, when it happens, tends to show up first. I’d rather watch that window close before adding a second immunosuppressive agent to a patient who is already fighting autonomic swings and whatever an ICU stay does to infection risk.

Clinical Pharmacologist Final

There’s a version of this where neither of you has to be wrong yet. Rituximab and ‘wait’ aren’t the only two options on the table — a second course of plasma exchange pulls circulating antibody out directly, on a timescale rituximab’s own depletion mechanism can’t match, and it isn’t a new drug class, just more of one we’ve already used. Pham and colleagues’ long-term follow-up found patients who received IVIG after plasma exchange did better than those who got it in the other order — nine patients, and the authors say themselves the series is too small to conclude much from, so I’d hold it loosely. It is still a reason to think we haven’t finished trying what’s already in front of us. A second PE course buys real days of observation on the resection question without foreclosing rituximab if she’s still not moving by day sixteen or eighteen.

Regimen selected
Methylprednisolone (continuing)
Corticosteroid · High-dose IV, day 12
First-line agent since diagnosis; continued alongside the plan below, not withdrawn for lack of response.
IVIG (continuing)
Immunoglobulin · First-line, day 12
Continued for the same reason as the corticosteroid course — no clinical evidence yet that either has failed outright, only that response is incomplete.
Plasma Exchange (repeat course)
Antibody Removal · Started today
Chosen over immediate second-line escalation; removes circulating NMDAR antibody directly and faster than rituximab’s depletion mechanism; sequencing signal from Pham and colleagues’ nine-patient series, which the authors themselves call too small to be conclusive.
Rituximab — Held in Reserve
Anti-CD20 Monoclonal Antibody · Contingent
Not started today; explicit trigger set for day 16–18 if repeat PE and continued first-line therapy have not produced clear improvement.
Cyclophosphamide — Not Adopted
Alkylating Agent · Second-line alternative
Available as a second-line option per consensus guidance but not chosen today; rituximab is preferred as the first second-line agent if escalation becomes necessary.
Where this was left

Agreed: a second course of plasma exchange starts today, continuing alongside the ongoing methylprednisolone and IVIG, with an explicit reassessment at day sixteen to eighteen — if there is no clear improvement in the dyskinesias or her level of interaction by then, rituximab starts without further debate.

Not agreed: how much weight the post-resection recency argument should carry the next time this decision comes up. The critical care physician remains unconvinced the two-week consensus threshold should apply as rigidly in the days immediately following tumor resection; the autoimmune neurology specialist thinks the threshold should hold regardless of resection timing, since the consensus data was never stratified that finely to begin with. Both agreed the disagreement was real enough to leave stated rather than smoothed over into today’s plan.

Educational content only — a composite teaching case, not a real patient encounter or a substitute for clinical guidance. About These Cases →