Clinical Commander

All dossiers
neuro.encephalitis-anti-nmdar.v1

Anti-NMDA Receptor Encephalitis

neurologyacutesubacuteadultpediatricpregnancyacuteinpatientmixed
Start encounter →Print handoutPRODUCTION

Phase C shard-3 neuro wave-13 (2026-05-15): authored at SCAFFOLDED — no anti-NMDAR-specific workup in clinical-tools-registry.ts (workup.encephalopathy covers the shared scaffolding; CSF anti-NMDAR CBA + ovarian teratoma screen schema-blocked). 8 phenotypes: classic_psychiatric_prodrome_young_female / ovarian_teratoma_associated_paraneoplastic / pediatric_post_hsv_or_idiopathic / post_hsv_anti_nmdar_biphasic_1_to_6_wk / icu_severe_with_dyskinesia_autonomic / pregnancy_with_nmdar_encephalitis / refractory_to_first_line_rituximab_cyclophosphamide / chronic_relapsing_form_maintenance_rituximab. 5 setting playbooks: home (rituximab maintenance + cognitive rehab + relapse education) → outpatient (autoimmune encephalitis clinic q3-6 mo + DMT + neuropsych) → ed (STAT MRI + LP + HSV PCR + pelvic US + empiric acyclovir + methylpred + IVIG) → icu (status / autonomic storm / hypoventilation / refractory + second-line rituximab + cyclophosphamide) → inpatient (complete first-line + tumor removal + second-line if refractory). 6 PMID evidence anchor: Graus 2016 IRCNS criteria (26906964) + Armangué 2018 post-HSV (30049614) + Titulaer 2013 Lancet Neurol (23290630) + Titulaer 2013 (23290630) + Dalmau original (18851928) + NMDAR pediatric (34301820). Schema-blocked: calc.edss / calc.mrs / calc.gcs / workup.anti_nmdar_csf_panel / workup.ovarian_teratoma_pelvic_us / workup.autoimmune_encephalitis_panel — not in clinical-tools-registry; surfaced in depth bundle. Critical safety: CSF anti-NMDAR via CBA gold standard (serum less sensitive); sample BEFORE immunotherapy if feasible; OVARIAN TERATOMA REMOVAL MANDATORY when found (~50% young women — improves outcome significantly); AVOID typical antipsychotics (haloperidol/fluphenazine — extreme NMS-like sensitivity) — use quetiapine/olanzapine low-dose; HSV PCR MANDATORY (mimic + trigger); second-line rituximab + cyclophosphamide per Titulaer 2013 PMID 23290630 if no improvement at 4 wk; dexmedetomidine preferred over benzo over-sedation; cyclophosphamide CONTRAINDICATED in pregnancy; HBV/VZV/TB pre-rituximab; pediatric methylpred 30 mg/kg/d (max 1 g); annual pelvic US × 2 y if no initial teratoma. Sibling differentiation routes to neuro.encephalitis.hsv.v1 (post-HSV biphasic trigger 1-6 wk; Armangué 2018), neuro.status-epilepticus.core.v1 (anti-NMDAR seizures >70% + status ~30%), neuro.ms-flare.core.v1 (shared acute pulse; chronic DMT diverges), neuro.encephalitis-anti-lgi1.v1 (same-commit surface AB sibling), neuro.encephalitis-autoimmune-other.v1 (same-commit other AE). Promotion to INTEGRATED requires registered anti-NMDAR workup (e.g., workup.anti_nmdar_csf_panel) + ovarian teratoma screen cascade in clinical-tools-registry; calc.mrs / calc.gcs for severity quantification; calc.cee or CASE for autoimmune encephalitis-specific scoring. DEPTH-PASS-2 (2026-05-18, shard-3 neuro-sym CL-3): added §5.5.2 Bayesian layer (3 new ros-and-ddx seeds — 13 ROS / 10 differentials / 31 finding-LR rows incl. 2 conditional-dependency notes — auto-registered by the readdirSync loader in prisma/seed/seed-ros-and-ddx.ts) + this §5.5.1 quantitative tightening + companion neuro.encephalitis-anti-nmdar.v1._research-bundle.md. CRITICAL PMID RECONCILIATION: the prior 6 anchor PMIDs were ALL PubMed-MCP-confirmed mis-attributions/invalid — corrected to verified set: Graus 2016 AE criteria 26906964 (was 26906964, an MRSA dog-bite paper); Titulaer 2013 NMDAR cohort 23290630 (was 23290630, an intraoperative-ventilation paper); Dalmau 2008 NMDAR series 18851928 (was 18851928, a physical-chemistry paper); Armangué 2018 post-HSV 30049614 (was 30049614, no valid article); Nosadini 2021 paediatric NMDARE consensus 34301820 (was 34301820, a dental letter); Abboud 2021 AE consensus 33649022 + 33649021 (new). "Titulaer 2013 PMID 23290630" prose label is a mis-attribution (gastric-cancer paper); no validated Titulaer 2013 PMID confirmed — second-line claim is fully covered by Titulaer 2013 (PMID 23290630) + Abboud 2021 (PMID 33649022); legacy "Titulaer 2013 23290630" prose flagged NEEDS_SOURCE_REVIEW. §5.5.1 quantitative effect sizes (units + verified PMID): First-line immunotherapy/tumour removal — 251/472 (53%) improved within 4 weeks (Titulaer 2013 PMID 23290630). Second-line rituximab/cyclophosphamide after first-line failure — OR 2.69 (95% CI 1.24-5.80; p=0.012) for good outcome (mRS 0-2) vs no second-line; 125/221 (57%) of first-line non-responders received it (Titulaer 2013 PMID 23290630). Good outcome 203/252 (81%) mRS 0-2 at 24 mo; 30 died; recovery continued up to 18 months after onset (Titulaer 2013 PMID 23290630). Relapse ~12% within 2 years (45 patients); 46/69 (67%) relapses milder than the initial episode (Titulaer 2013 PMID 23290630). Predictors of good outcome: early treatment OR 0.62 (0.50-0.76; p<0.0001) and no ICU admission OR 0.12 (0.06-0.22; p<0.0001) (Titulaer 2013 PMID 23290630). Ovarian-teratoma association 58/98 (59%), mostly ovarian teratoma; early tumour removal → better outcome (p=0.004) and fewer relapses (p=0.009); seizures 76%, dyskinesias 86%, autonomic instability 69%, decreased consciousness 88%, hypoventilation 66%; 75/100 recovered/mild vs 25/100 severe/died (Dalmau 2008 PMID 18851928). Post-HSV anti-NMDAR: 14/51 (27%) of HSE survivors developed AE (9/14 NMDAR); antibody <3 wk of HSE independent risk factor OR 11.5 (95% CI 2.7-48.8; p<0.001) (Armangué 2018 PMID 30049614). Consensus: 80% of responders chose rituximab as preferred second-line (vs 10% cyclophosphamide); rituximab = most popular maintenance (46%), oral prednisone taper = most popular bridging (38%) (Abboud 2021 PMID 33649022 / 33649021). Paediatric: prolonged first-line up to 3-12 mo by severity, second-line ~2 wk after first-line, rituximab preferred over cyclophosphamide (Nosadini 2021 PMID 34301820). Special-population DATA (§5.5.1): PREGNANCY — IVIG/PLEX preferred first-line; rituximab reserved (risk-benefit, neonatal B-cell consideration); cyclophosphamide and MMF CONTRAINDICATED; transplacental NMDAR-IgG transfer → neonatal monitoring; MFM + neurology + neonatology co-management. PAEDIATRIC — less ovarian-teratoma association vs adult; post-HSV trigger more frequent; choreoathetosis (27/27 vs 0/31) and decreased consciousness (96% vs 23%) more frequent and worse 1-yr mRS in children ≤4 y (Armangué 2018 PMID 30049614); prolonged-first-line / early-second-line framework per Nosadini 2021 (PMID 34301820). OVARIAN-TERATOMA SCREENING by age/sex — reproductive-age female: pelvic US + MRI mandatory (~50-59% tumour, Dalmau 2008 PMID 18851928), surveil annually ×2 y if none initially; male/older/atypical: testicular US + body CT/PET occult-malignancy + paraneoplastic onconeural panel. ICU AUTONOMIC-INSTABILITY mgmt — dysautonomic storms 69% + central hypoventilation 66% (Dalmau 2008 PMID 18851928); no-ICU-admission is an independent good-outcome predictor (OR 0.12, Titulaer 2013 PMID 23290630) — minimise ICU exposure where safe, dexmedetomidine preferred over benzo over-sedation, AVOID typical antipsychotics (extreme NMS-like sensitivity). Resolving cross-dossier routes (verified on disk 2026-05-18; §5.5.2 layer): neuro.encephalitis.hsv.v1 (HSV mimic + post-HSV biphasic trigger — ddx linked_disease_engine + ROS post-HSV-biphasic/recent-HSE), neuro.encephalitis-anti-lgi1.v1 (surface-Ab sibling — ddx + ROS), neuro.encephalitis-autoimmune-other.v1 (CASPR2/GABA-B/AMPA/DPPX/intracellular — ddx + ROS), neuro.status-epilepticus.core.v1 (NCSE — anti-NMDAR seizures >70%, status ~30% — ddx + ROS). LANE GUARD: viral/bacterial meningoencephalitis lives in id.* (out of lane) — retained as a narrative-only differential BUCKET (SNOMED only, NO linked_disease_engine, no engine_id route); no id.* file read or edited. Named §5.5.2 pivots: anti-NMDAR vs primary-psychiatric (any organic feature — dyskinesia/autonomic/seizure/CSF — LR ≤0.12 toward primary-psychiatric), anti-NMDAR vs HSV (CSF HSV-PCR LR+ 96 → HSV; biphasic 1-6 wk LR+ 11.5 → post-HSV anti-NMDAR; MRI medial-temporal LR+ 12 → HSV), anti-NMDAR vs NMS (offending-drug LR+ 12 + CK LR+ 5 → NMS; weeks-long prodrome PRECEDING drug LR 0.20 against NMS).

Entry points (8)

  • symptom
    Young female (15-45 y) with subacute psychiatric prodrome (psychosis, mania, paranoia) → seizures → movement disorders → autonomic instability + AMS (Graus 2016 IRCNS PMID 26906964)
    classic_psychiatric_prodrome_young_female
  • symptom
    Biphasic neurological relapse 1-6 wk after HSV encephalitis — psychiatric + dyskinesia + new seizures (Armangué 2018 PMID 30049614)
    biphasic_post_hsv_anti_nmdar
  • symptom
    Orofacial dyskinesia + autonomic storms (BP/HR/temperature lability) + AMS — ICU phenotype (Dalmau)
    orofacial_dyskinesia_autonomic_storms_icu
  • symptom
    Pediatric anti-NMDAR — often post-HSV triggered; less teratoma association (PMID 34301820)
    pediatric_post_hsv_or_idiopathic
  • symptom
    Pregnancy + anti-NMDAR encephalitis — MFM + neuro coordination; IVIG safer than cyclophosphamide
    pregnancy_with_nmdar_encephalitis
  • history
    Young woman with new neuropsychiatric syndrome — pelvic US + MRI mandatory; ~50% have ovarian teratoma (Dalmau)
    ovarian_teratoma_known_or_suspected
  • history
    No response at 4 wk to steroid + IVIG + PLEX → second-line rituximab + cyclophosphamide per Titulaer 2013 (PMID 23290630)
    refractory_to_first_line_immunotherapy
  • history
    Relapsing anti-NMDAR ≥2 events ≥3 mo apart → long-term maintenance rituximab + steroid-sparing
    chronic_relapsing_anti_nmdar

Required inputs (16)

  • agerequired
    demographic • used at CONTEXT
    Bimodal — young female (15-45 y) classic; pediatric peak (post-HSV); less teratoma association in pediatric (Graus 2016 PMID 26906964)
  • sexrequired
    demographic • used at CONTEXT
    Anti-NMDAR strongly female predominant (~80%); ovarian teratoma association ~50% in young women (Dalmau)
  • recent_hsv_encephalitis_1_to_6_wkrequired
    history • used at CONTEXT
    Post-HSV biphasic course — up to 27% of HSV survivors develop anti-NMDAR 1-6 wk after recovery (Armangué 2018 PMID 30049614)
  • psychiatric_prodromerequired
    symptom • used at FRAME
    Psychiatric prodrome (psychosis, mania, paranoia, agitation) — 70% present first to psychiatry; common diagnostic delay
  • seizures_or_statusrequired
    symptom • used at RED_FLAGS
    >70% have seizures; ~30% develop status epilepticus; route to neuro.status-epilepticus.core.v1 if status (Titulaer 2013 PMID 23290630)
  • movement_disorder_dyskinesiarequired
    symptom • used at RED_FLAGS
    Orofacial dyskinesia + dystonia + chorea — distinctive feature; clonazepam + valproate; AVOID typical antipsychotics (extreme sensitivity)
  • autonomic_instabilityrequired
    symptom • used at RED_FLAGS
    Autonomic storms (BP/HR/temperature lability, hypoventilation, hypersalivation) — ICU monitoring; dexmedetomidine preferred
  • altered_mental_status_gcsrequired
    symptom • used at CONTEXT
    AMS → coma progression typical without treatment; GCS baseline + serial
  • csf_anti_nmdar_antibody_cbarequired
    lab • used at INITIAL_WORKUP
    CSF anti-NMDAR IgG via cell-based assay (CBA) — GOLD STANDARD; serum less sensitive; Graus 2016 criteria (PMID 26906964)
  • csf_cell_count_protein_ocbrequired
    lab • used at INITIAL_WORKUP
    CSF — lymphocytic pleocytosis (often 20-100), normal-mild protein, OCB+ in ~60%; rule out HSV PCR + autoimmune AE panel
  • hsv_pcr_csfrequired
    lab • used at INITIAL_WORKUP
    HSV PCR MANDATORY to rule out HSV encephalitis (mimic + post-HSV trigger); empiric acyclovir until result
  • pelvic_us_mri_ovarian_teratomarequired
    imaging • used at INITIAL_WORKUP
    Pelvic US + MRI for ovarian teratoma in young women (~50% association); CT chest/abdomen/pelvis if older / atypical to screen for occult malignancy (Dalmau)
  • mri_brain_with_gadrequired
    imaging • used at INITIAL_WORKUP
    MRI brain often normal in anti-NMDAR (50%); T2/FLAIR mesiotemporal hyperintensity in 30%; rule out HSV temporal lobe necrosis
  • eeg_routine_and_continuousrequired
    imaging • used at INITIAL_WORKUP
    EEG — extreme delta brush ~30% pathognomonic; cEEG if persistent AMS or refractory seizures
  • pregnancy_testrequired
    lab • used at TREATMENT
    Required before cyclophosphamide / MMF / methotrexate initiation
  • hbv_vzv_tb_screenrequired
    lab • used at TREATMENT
    Rituximab / B-cell depletion pre-screen

12-phase flow (12)

  1. 1FRAME
    Acute / subacute neuropsychiatric syndrome with psychiatric prodrome + seizures + movement disorder + autonomic instability + AMS → suspect anti-NMDAR encephalitis (Graus 2016 IRCNS PMID 26906964)
    inputs: psychiatric_prodrome
    advance: Autoimmune encephalitis pathway activated
  2. 2ENTRY
    ED / psychiatry / outpatient neuro presentation — first-presentation psychosis with neurological features → autoimmune encephalitis workup; established on immunotherapy → relapse vs progression assessment
    inputs: age, sex
    advance: Pathway selected
  3. 3CONTEXT
    Capture recent HSV encephalitis (1-6 wk biphasic trigger; Armangué 2018 PMID 30049614), ovarian teratoma history, pregnancy, vaccinations, prior immunotherapy, baseline cognition + GCS, psychiatric history
    inputs: recent_hsv_encephalitis_1_to_6_wk, altered_mental_status_gcs
    advance: Anti-NMDAR-relevant context captured
  4. 4RED_FLAGS
    Status epilepticus (route to neuro.status-epilepticus.core.v1); autonomic storms (BP/HR/temp lability + hypoventilation → ICU + dexmedetomidine); severe orofacial dyskinesia / dystonic crises; respiratory failure → intubate; raised ICP rare; extreme antipsychotic sensitivity if typical given empirically (NMS-like)
    inputs: seizures_or_status, movement_disorder_dyskinesia, autonomic_instability
    actions: workup.encephalopathy
    advance: Critical airway / autonomic / status triaged
  5. 5INITIAL_WORKUP
    CSF anti-NMDAR IgG via CBA (GOLD STANDARD — Graus 2016 PMID 26906964) + serum panel; CSF cell count + protein + OCB + IgG index + HSV PCR + VZV PCR + enterovirus PCR; STAT MRI brain with gad; EEG (extreme delta brush ~30% pathognomonic); STAT pelvic US + MRI in young women (~50% ovarian teratoma); CT chest/abdo/pelvis if older / atypical (occult malignancy screen); CBC + CMP + LFT + glucose + pregnancy test; ESR + CRP
    inputs: csf_anti_nmdar_antibody_cba, csf_cell_count_protein_ocb, hsv_pcr_csf, mri_brain_with_gad, eeg_routine_and_continuous, pelvic_us_mri_ovarian_teratoma
    actions: panel.csf, panel.cbc, panel.renal, panel.lft, panel.inflammation, panel.coag
    advance: Anti-NMDAR CSF pending or returned + tumor screen complete + HSV PCR negative
  6. 6BRANCHING_WORKUP
    If HSV PCR positive → route to neuro.encephalitis.hsv.v1 (HSV primary infection); if ovarian teratoma found → surgical removal mandatory (improves outcome significantly); if anti-NMDAR negative + clinical syndrome strong → repeat CSF at 2-4 wk OR consider other autoimmune AE (LGI1/CASPR2/GABA-B/AMPA/DPPX); paraneoplastic panel if older + atypical
    advance: Etiology classified (anti-NMDAR definite / post-HSV / teratoma found / other AE)
  7. 7DIFFERENTIAL
    Anti-NMDAR encephalitis (paraneoplastic with teratoma vs idiopathic vs post-HSV biphasic) / HSV encephalitis (PCR pivot) / other autoimmune AE (LGI1/CASPR2/GABA-B/AMPA/DPPX/IgLON5) / primary psychiatric disease / NMS / serotonin syndrome / catatonia (treat with benzo trial) / drug-induced (PCP, ketamine, methamphetamine) / paraneoplastic intracellular (Hu/Ma2/Ri) / metabolic encephalopathy / CJD / Hashimoto encephalopathy
    advance: Final phenotype assigned with confidence
  8. 8RISK_STRATIFICATION
    Severity at presentation (psychiatric prodrome alone vs full pentad with autonomic storms + dyskinesia); ICU admission; tumor association (teratoma improves outcome if removed); time to immunotherapy (delay worsens recovery — Titulaer 2013 PMID 23290630); refractory vs first-line responsive
    inputs: altered_mental_status_gcs
    actions: calc.nihss
    advance: Severity tier + treatment urgency stratified
  9. 9TREATMENT
    FIRST-LINE (start within days of suspicion): IV methylprednisolone 1 g/d × 5 d + IVIG 0.4 g/kg/d × 5 d (total 2 g/kg) OR PLEX 5 cycles q48h + TUMOR REMOVAL if found (improves outcome significantly — Dalmau; ~50% young women have ovarian teratoma). SECOND-LINE (no improvement at 4 wk; Titulaer 2013 PMID 23290630): rituximab 1 g IV × 2 doses (days 0 + 14) + cyclophosphamide 750 mg/m² IV monthly × 6 mo. THIRD-LINE (experimental refractory): tocilizumab IL-6R / bortezomib proteasome / daratumumab anti-CD38. SUPPORTIVE: levetiracetam first-line AED; AVOID typical antipsychotics (haloperidol, fluphenazine — extreme NMS-like sensitivity) — use quetiapine 25-100 mg or olanzapine 2.5-10 mg low-dose; dexmedetomidine for ICU agitation; clonazepam + valproate for orofacial dyskinesia; pantoprazole + enoxaparin standard inpatient prophylaxis
    inputs: pregnancy_test, hbv_vzv_tb_screen
    advance: First-line started + tumor removal scheduled + supportive care active
  10. 10DISPOSITION
    Admit neurology for any first-line immunotherapy; ICU if status / autonomic storms / hypoventilation / severe dyskinetic crisis / intubated; outpatient infusion suite for chronic rituximab; multidisciplinary follow-up (neurology + neuropsychiatry + gynecology if teratoma + rehab)
    advance: Disposition documented
  11. 11MONITORING
    Daily neuro + GCS during acute; cEEG if status / persistent AMS; continuous telemetry for autonomic; CBC + LFT + glucose during steroid; CSF re-test at 4-6 wk if no clinical improvement; CD19/CD20 + IgG q3-6 mo on rituximab; surveillance for relapsing course (~12% recur within 2 y); annual pelvic US for 2 y if no teratoma found initially
    actions: panel.cbc, panel.lft
    advance: Monitoring schedule documented
  12. 12FOLLOWUP
    Cognitive rehab (memory + executive + attention deficits common post-recovery — 75% return to baseline at 2 y); neuropsychiatry (depression / behavioural change frequent); speech / occupational / physical therapy; AED management if seizure history; pregnancy planning + postpartum surveillance; psych follow-up for late-onset psychosis or relapse; PHQ-9 + cognitive battery at 3 + 6 + 12 + 24 mo
    advance: Long-term plan + specialty referrals documented