Compounds with a graded row

Ordered by rung, strongest first. The rung rates the study design; the chip beside it says what the study measured. Both are needed.

1. Oxytocin

   human RCT Direct outcome In registered trials

   Who was studied

   adults 18-65 with chronic migraine (with or without aura) by ICHD-3, n=88 randomised, 12 weeks; intranasal oxytocin 30 IU once daily or 30 IU twice daily versus placebo nasal spray, quadruple-blinded (TNX-1900, NCT05679908)

   This is the one compound here that was given as a candidate treatment, and it was measured directly: mean change in monthly migraine headache days over 12 weeks. Posted registry results show no separation from placebo - placebo -8.17 days, oxytocin 30 IU once daily -7.78 days, oxytocin 30 IU twice daily -5.77 days. No regulator has approved oxytocin for migraine.

   Source ClinicalTrials.gov 1 primary source read for this row

2. PACAP (Pituitary Adenylate Cyclase-Activating Peptide)

   human RCT Direct outcome Not approved for this condition

   Who was studied

   adults with migraine without aura and headache-free healthy controls; PACAP-38 10 pmol/kg/min IV over 20 min, n=12 migraine + 12 healthy (Schytz 2009); PACAP-27 10 pmol/kg/min IV over 20 min, n=20 migraine without aura (Ghanizada 2020). People with migraine with aura were not the studied group in either trial.

   In both randomised, double-blind, placebo-controlled crossover trials PACAP was infused in order to TRIGGER attacks, not to treat them: PACAP-38 provoked migraine-like attacks in 7 of 12 migraine patients and none on placebo, and the shorter isoform PACAP-27 provoked attacks in 11 of 20 and 2 of 20 on placebo. This is evidence that PACAP causes migraine attacks in susceptible people; no study has administered PACAP-38 or PACAP-27 as a migraine treatment, and the drug programmes built on this work develop antibodies and receptor antagonists that BLOCK PACAP - those are different molecules and their evidence does not belong to PACAP itself.

   Source DOI 2 primary sources read for this row

3. Substance P

   human RCT Mechanistic only Not approved for this condition

   Who was studied

   healthy adults with no headache disorder, n=21, double-blind placebo-controlled two-way crossover; substance P 1.5 pmol/kg/min IV over 20 min. People with migraine were not enrolled.

   Recorded, and not evidence for this condition

   Substance P was infused in order to PROVOKE headache, and it did: headache within 12 hours in 15 of 21 healthy participants versus 2 of 21 on placebo, with dilation of the superficial temporal artery. No study measured any migraine outcome - the participants did not have migraine, and migraine attack frequency, severity and duration were not measured, so the link from this result to migraine is an argument from mechanism rather than a result. A provocation trial in people with migraine without aura is registered (NCT06959004) and has not reported.

   Source DOI 1 primary source read for this row

4. Nociceptin

   human observational Mechanistic only Not approved for this condition

   Who was studied

   adults with migraine without aura, n=18, headache-free at sampling, versus n=24 controls; a further 6 sampled within the first 3 hours of an attack. Migraine with aura was not studied.

   Recorded, and not evidence for this condition

   This study MEASURED circulating nociceptin; it did not administer it to anyone. Interictal plasma nociceptin was lower in migraineurs than controls (5.79 vs 9.74 pg/mL) and lower still during attacks, which is an association between a blood marker and having migraine. No study has given nociceptin to people with migraine, so no migraine outcome has ever been measured for this compound, and nothing here shows that raising or lowering nociceptin changes attacks.

   Source DOI 1 primary source read for this row

Checked, and nothing recorded

1 compound was checked against migraine and produced no gradeable row.

• Bradykinin — checked, no row on file

A compound is here because someone looked and the sources did not support a row — not because nobody looked. Under the rubric, uncertainty resolves to no row, and a row resting on a vendor page or a clinic blog is worse than no row at all.