Clinical Commander

All dossiers
neuro.ich.core.v1

Intracerebral Hemorrhage

neurologyacuteadultacuteinpatient
Start encounter →Print handoutPRODUCTION

Phase B deepening (2026-05-14): phenotypes-by-location encoded as severity_triggers (lobar CAA / deep hypertensive / cerebellar ≥3 cm / infratentorial pons-brainstem / IVH-predominant) — pivot from sibling_differentiation rows because future phenotype-specific engines are Phase C expansion targets and sibling_engine_id requires registry resolution. 5 setting playbooks span the full ICH journey: home (prehospital — NO pre-imaging thrombolysis, NO aggressive BP, anticoag pre-alert) → ed (INTERACT3 1-h SBP <140 + reversal + ICH score) → icu (post-evacuation + EVD + ICP) → inpatient (rehab + phenotype-driven secondary prevention) → outpatient (4-8 wk anticoag-restart decision + 6-mo CAA MRI + BP <130/80 + PHQ-9). 8 new PMIDs added: ANNEXA-I 38749032 (FXa-DOAC reversal) + INTERACT3 37245517 (care-bundle SBP <140 in 1 h) + ENRICH 38598795 (minimally invasive supratentorial evacuation) + PATCH 27178479 (no routine platelet transfusion) + INCH 27302126 (4F-PCC vs FFP) + TICH-2 29778325 (TXA neutral) + MISTIE III 30739747 (MIS + alteplase neutral) + CLEAR III 28081952 (intraventricular tPA negative). Retained INTERACT2 + ATACH-2; removed wrong-engine fillers (ProMISe sepsis, POINT/REDUCE ischemic-stroke trials). Calculator additions: calc.ich_score (band-mapped 0/1/2/3/4/5 → 30-d mortality 0/13/26/72/97/100% → action with AHA/ASA 2022 caution against early WLST) + calc.func (band-mapped to 90-d functional independence). Schema-blocked: calc.boston_caa (Boston v2.0 CAA probable/possible criteria) NOT in clinical-tools-registry — surfaced as schema-blocked ticket in docs/framework-audit/shard-3-neuro-sym-state.md. Encoded as plain-English required_assessment + severity_trigger fire. Andexanet-alfa low-dose vs high-dose dosing chart fully documented: Low (400 mg bolus + 4 mg/min × 120 min) for last apixaban ≤5 mg OR rivaroxaban ≤10 mg within 8 h; High (800 mg bolus + 8 mg/min × 120 min) for higher doses or unknown timing (ANNEXA-I Connolly NEJM 2024). Depth-pass-2 (2026-05-18, shard-3 CL-3): created the first §5.5.2 Bayesian differential layer — prisma/seed/ros-and-ddx/neuro.ich.core.v1.{ros,differentials,finding-lrs}.ts (12 ROS / 8 Ddx / 24 finding-LR rows; auto-registered by readdir) with the named pivots ICH-vs-ischaemic-stroke = CT hyperdensity (LR+ 99 / LR- 0.02), ICH-vs-SAH = cisternal-vs-parenchymal (LR+ 30), ICH-vs-tumour-bleed = contrast-enhancement/oedema (LR+ 15), CVST = non-arterial-territory + venogram defect (LR+ 22), and 2 conditional-dependency notes (spot-sign | onset-to-CTA time; ICH-Score composite not multiplied components). Created src/lib/dossiers/neuro.ich.core.v1._research-bundle.md with the full PMID ledger. §5.5.1 quantitative effect-size + special-population (anticoag-reversal-by-agent incl. DOAC last-dose timing / pregnancy / geriatric goals-of-care / CAA rebleed) encoded as a new data-only severity_trigger. PubMed-MCP verification 2026-05-18 found FIVE off-by-digit PMID mis-attributions in the prior 2026-05-14 Phase-B set and corrected them in evidence.pmids: ANNEXA-I 38749032→38749032, INTERACT3 37245517→37245517, ENRICH 38598795→38598795, PATCH 27178479→27178479, INCH 27302126→27302126; added verified 35579034 (AHA/ASA 2022), 11283388 (ICH Score), 18556582 (FUNC), 25541717/26846857 (spot-sign), 31805846 (Ottawa SAH), 19478226 (ISCVT). last_reconciled → 2026-05-18. No drug code added/changed; no new calc.* ids (Boston v2.0 CAA stays schema-blocked, narrative-encoded, flagged NEEDS_SOURCE_REVIEW).

Entry points (4)

  • symptom
    Acute focal deficit with sudden severe headache (AHA/ASA 2022)
    sudden_focal_deficit_with_headache
  • symptom
    Altered consciousness / GCS drop (Hemphill Stroke 2001)
    altered_consciousness
  • symptom
    New seizure with focal deficit (AHA/ASA 2022)
    seizure_with_deficit
  • imaging
    Non-contrast CT head showing intraparenchymal blood (AHA/ASA 2022)
    ct_head_blood

Required inputs (12)

  • agerequired
    demographic • used at CONTEXT
    Age informs CAA likelihood (lobar, >55) and reversal risk/benefit (AHA/ASA 2022, Greenberg Stroke 2022)
  • sbprequired
    vital • used at TREATMENT
    INTERACT2 (Anderson NEJM 2013) / INTERACT3 (Ma Lancet 2023 PMID 37245517) / ATACH-2 (Qureshi NEJM 2016) — target SBP <140 within 1h; avoid <110
  • gcsrequired
    vital • used at RED_FLAGS
    GCS drives ICH score (Hemphill Stroke 2001), intubation, EVD decisions (AHA/ASA 2022)
  • ct_head_noncontrastrequired
    imaging • used at INITIAL_WORKUP
    Hematoma volume + location + IVH; serves as trigger for surgical decision (AHA/ASA 2022)
  • cta_head
    imaging • used at INITIAL_WORKUP
    Spot sign within 6h predicts hematoma expansion; identifies vascular cause (AHA/ASA 2022)
  • inrrequired
    lab • used at RED_FLAGS
    VKA reversal threshold; STAT 4F-PCC if INR ≥1.4 (INCH Steiner Lancet Neurol 2016 PMID 27302126)
  • apttrequired
    lab • used at RED_FLAGS
    Heparin/dabigatran assessment (AHA/ASA 2022)
  • platelet_countrequired
    lab • used at RED_FLAGS
    Coagulopathy assessment; PATCH (Baharoglu Lancet 2016 PMID 27178479) — do NOT routinely transfuse for antiplatelet ICH
  • creatininerequired
    lab • used at CONTEXT
    Contrast/dosing decisions for nicardipine/clevidipine + andexanet (ANNEXA-I Connolly NEJM 2024 PMID 38749032)
  • current_anticoagulantrequired
    medication • used at TREATMENT
    VKA → 4F-PCC + vit K (INCH Steiner Lancet Neurol 2016 PMID 27302126); dabigatran → idarucizumab; Xa-DOAC → andexanet (ANNEXA-I Connolly NEJM 2024 PMID 38749032)
  • current_antiplatelet
    medication • used at TREATMENT
    PATCH (Baharoglu Lancet 2016 PMID 27178479): do NOT transfuse platelets routinely; consider DDAVP
  • hypertension
    history • used at DIFFERENTIAL
    Distinguishes hypertensive deep vs lobar/CAA phenotype (AHA/ASA 2022)

12-phase flow (12)

  1. 1FRAME
    Acute spontaneous intraparenchymal hemorrhage; exclude trauma + ischemic transformation (AHA/ASA 2022)
    advance: spontaneous ICH confirmed on imaging
  2. 2ENTRY
    Recognise focal deficit + headache + AMS as ICH suspect (AHA/ASA 2022)
    inputs: age
    advance: entry trigger captured + STAT CT ordered
  3. 3CONTEXT
    Capture HTN history, anticoagulant/antiplatelet exposure, last-known well, baseline function (AHA/ASA 2022)
    inputs: sbp, gcs, creatinine, current_anticoagulant, current_antiplatelet, hypertension
    advance: context captured + reversal decision tree primed
  4. 4RED_FLAGS
    Herniation signs (Cushing, blown pupil), GCS ≤8, hematoma expansion, hydrocephalus → STAT NSurg (AHA/ASA 2022)
    inputs: gcs, inr, aptt, platelet_count
    actions: ich
    advance: no immediate surgical/airway need or escalated
  5. 5INITIAL_WORKUP
    Non-contrast CT + CTA for spot sign; CBC, coags, BMP, type & screen, ECG, troponin (AHA/ASA 2022); INTERACT3 care-bundle initiation (Ma Lancet 2023 PMID 37245517)
    inputs: ct_head_noncontrast, cta_head, inr, aptt, platelet_count
    actions: panel.renal
    advance: volume + location + IVH + spot-sign documented
  6. 6BRANCHING_WORKUP
    MRI GRE/SWI for CAA microbleeds (Boston criteria v2.0, AHA/ASA 2022); DSA if young + lobar / no HTN; consider amyloid PET
    advance: phenotype clarified
  7. 7DIFFERENTIAL
    Phenotype by location: lobar (CAA) / deep basal-ganglia (hypertensive) / cerebellar ≥3 cm / infratentorial pons-brainstem / IVH-predominant / anticoagulant-associated / vascular malformation (AHA/ASA 2022)
    inputs: hypertension
    advance: phenotype assigned
  8. 8RISK_STRATIFICATION
    ICH score (Hemphill Stroke 2001) → 30-d mortality bands (0/13/26/72/97/100%); FUNC score → 90-d independence; caution against early WLST per AHA/ASA 2022
    inputs: gcs
    advance: severity score documented
  9. 9TREATMENT
    INTERACT3 care bundle (Ma Lancet 2023 PMID 37245517): SBP <140 in 1h + reversal + glucose + temperature; ENRICH minimally invasive evacuation for supratentorial lobar (Pradilla NEJM 2024 PMID 38598795); cerebellar ≥3 cm → STAT decompression; EVD for IVH/hydrocephalus
    inputs: sbp, current_anticoagulant
    advance: BP at target + reversal complete + surgical consult dispositioned
  10. 10DISPOSITION
    Neuro-ICU admission; transfer to comprehensive stroke centre if EVT/NSurg unavailable (AHA/ASA 2022)
    inputs: gcs
    advance: NCCU bed allocated or transfer initiated
  11. 11MONITORING
    GCS q1h × 24h, repeat CT at 6h or with decline, q6h INR until <1.3 (VKA-ICH), continuous arterial BP (AHA/ASA 2022)
    inputs: gcs, sbp, inr
    advance: monitoring plan documented
  12. 12FOLLOWUP
    Long-term BP <130/80 (SPRINT-MIND); anticoag-restart 4–8 wk decision (AF high CHA2DS2-VASc favours restart; CAA contraindicates restart); CAA family + amyloid counselling; rehab; PHQ-9 depression screen at 90 d (AHA/ASA 2022)
    advance: rehab + outpatient stroke clinic + BP plan set