Skip to content

cluster-headache

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

cluster-headache

README.md

Cluster Headache (CH) — QSP Model

Hypothalamic circadian generator → trigeminovascular CGRP/PACAP release → parasympathetic outflow (SPG) and partial Horner's → 15–180 min attacks of unilateral periorbital VAS 9–10/10 pain occurring in stereotyped circadian/circannual bouts. Episodic in 80–90 %, chronic in 10–20 %. Male:female ≈ 3:1.

Files

File Description
ch_qsp_model.dot Graphviz source — 12 clusters · 104 nodes
ch_qsp_model.svg · ch_qsp_model.png Rendered mechanistic map
ch_mrgsolve_model.R 25-compartment ODE QSP, 7 drugs + O2 + GON block
ch_shiny_app.R 8-tab Shiny dashboard
ch_references.md 72 PubMed references grouped by topic

Mechanistic map — 12 clusters

  1. Genetic & chronobiologic susceptibility — HCRTR2, ADCYAP1, MTNR1A, CLOCK, BMAL1, ADH4, family Hx ×5–18.
  2. Hypothalamic generator — posterior hypothalamic gray (May 1998 Lancet), SCN / PVN / orexin / pineal melatonin, blunted cortisol & testosterone, in-bout vs remission gate.
  3. Trigeminovascular activation — V1 trigeminal ganglion → dura · ICA · pial vessels; CGRP (jugular ↑ in attack, Goadsby 1994), PACAP-38, VIP, substance P, NO.
  4. Parasympathetic limb (trigeminal-autonomic reflex) — SSN → GSPN → SPG → lacrimation, rhinorrhea, conjunctival injection, forehead sweating.
  5. Sympathetic disruption — pericarotid plexus → SCG → ipsilateral ptosis + miosis (partial Horner's).
  6. Central pain processing — TCC (NMDA/AMPA) → ipsilateral thalamic VPM/Po → S1·insula·ACC; restlessness / agitation distinguishes CH from migraine.
  7. Receptor pharmacology targets — 5-HT 1B/1D/1F, L-type Ca²⁺, GSK-3β/IMPase (Li), OX2R, SSTR2/5, TRPV1, CGRP/CGRP-R mAbs.
  8. Drugs — acute (O2, sumatriptan SC, zolmitriptan IN, lidocaine IN, octreotide), transitional (prednisone, GON block), preventive (verapamil, lithium, topiramate, melatonin, galcanezumab CGRP-mAb, erenumab off-label, civamide IN), devices (SPG-stim, nVNS, ONS, posterior hypothalamic DBS), research (psilocybin / LSA).
  9. Pharmacokinetics — sumatriptan SC (CL 18 L/h, t½ ≈ 2 h), zolmitriptan IN, verapamil PR (CYP3A4, active norverapamil), lithium (renal CL ≈25 mL/min), topiramate (CL 1.2 L/h, t½ 21 h), galcanezumab (CL 0.008 L/h, t½ 27 d), prednisolone.
  10. Clinical endpoints — attacks/week (primary), VAS pain, pain-free 15 min, ≥50 % responder rate, serum CGRP / VIP, AE bands (PR/QT, lithium tox, topiramate cognition, mAb injection-site).
  11. Triggers / lifestyle — alcohol (bout-on only), histamine SC, nitroglycerin, nitrites/MSG, altitude/flights, odors, heat, daytime naps.
  12. Comorbidity & burden — depression 55–65 %, suicide ideation 20–25 % (highest among pain disorders), CV risk on verapamil, lost work, HIT-6/CH-NDI.

ODE model — 25 compartments

  • Drug PK (16): sumatriptan SC (1-cmt), zolmitriptan IN (2-cmt + peripheral), verapamil PR (2-cmt), lithium (1-cmt, renal-CL adjusted by CrCL), topiramate (1-cmt), galcanezumab SC (1-cmt with first-order SC absorption, linear FcRn-recycled), prednisolone.
  • Disease PD (9): hypothalamic drive (circadian + bout gate), CGRP tone, PACAP tone, pial effect site, attack hazard (smoothed), cumulative attacks, bout timer, O2 effect compartment, GON-block effect compartment (decays t½ ≈ 14 d).
  • Composite preventive effect is a multiplicative escape: 1 − ∏(1 − Eᵢ) over verapamil, lithium, topiramate, galcanezumab, prednisolone, GON.
  • Acute abortive is a similar product over sumatriptan, zolmitriptan and O2 effect.
  • Calibrations: Cohen 2009 JAMA (O2 78 % pain-free@15 min), Ekbom 1991 NEJM (SC sumatriptan 74 %), Leone 2000 Neurology (verapamil 240 mg/d), Steiner 1997 Cephalalgia (lithium), Goadsby 2019 NEJM / Dodick 2020 Cephalalgia (galcanezumab ECH/CCH), Obermann 2021 Lancet Neurol (prednisone bridge), Leroux 2011 Lancet Neurol (GON block).

Six treatment scenarios (12-week simulation horizon)

Scenario Description
S0 No treatment (natural bout)
S1 O2 + sumatriptan 6 mg SC for an indexed attack
S2 Verapamil 240 mg PR BID
S3 Verapamil + lithium 300 mg BID (chronic CH)
S4 Galcanezumab 300 mg SC q4w
S5 Prednisone 60 → 40 → 20 mg taper + verapamil
S6 GON block (one-shot 0.65 effect) + verapamil

Shiny dashboard — 8 tabs

  1. Patient & disease profile (chronic vs episodic, hypothalamic drive)
  2. Drug PK (acute / preventive / mAb)
  3. Pathway PD (hypothalamic drive · CGRP/PACAP · pial)
  4. Clinical endpoints (attacks/week, hazard, trial-anchored table)
  5. Scenario comparison (S0–S6)
  6. Biomarkers (CGRP, PACAP, pial)
  7. Safety (verapamil PR/QT, lithium therapeutic band, galcanezumab exposure)
  8. References (renders ch_references.md)

Run

# Render map
dot -Tsvg ch_qsp_model.dot -o ch_qsp_model.svg
dot -Tpng -Gdpi=150 ch_qsp_model.dot -o ch_qsp_model.png

# Simulate (R)
Rscript -e 'source("ch_mrgsolve_model.R"); print(simulate_all())'

# Dashboard
Rscript -e 'shiny::runApp("ch_shiny_app.R")'

Built by Claude Code Routine — 2026-06-30.