Rats that received fremanezumab fifteen minutes before a brain bleed had higher cerebral blood flow five minutes into the bleed than rats that received saline. By forty-eight hours they were also doing better on a motor coordination test. By fourteen days the gap had closed: survival, body weight, well-being, and a composite neurological score came out roughly even between the two groups.
Fremanezumab is a once-monthly injection that blocks CGRP ↗, a small neuropeptide that dilates blood vessels in the head. Migraine patients take it to stop attacks. It works because CGRP is what fires during a migraine to open vessels and trigger pain. The open clinical question, ever since these antibodies were approved, has been the cerebrovascular flip side. If a patient on an anti-CGRP antibody suffers a brain bleed, does the vasodilator they no longer have access to make the bleed worse? Or does the blunted vasodilation actually help by limiting how much blood pours through the rupture in the first minutes?
A preclinical group reported its first direct test of that question online May 29 in the Journal of Headache and Pain ↗. They induced a pre-chiasmatic subarachnoid hemorrhage in rats, the kind of brain bleed produced by a ruptured aneurysm, fifteen minutes after intravenous fremanezumab at 30 mg/kg or vehicle. They tracked cerebral blood flow in the first minutes, motor performance on a rotating pole at forty-eight hours, neurological score over fourteen days, body weight, well-being, and survival.
The five-minute window came out the opposite of the worried direction. Fremanezumab-treated animals had higher cerebral blood flow than vehicle-treated ones in the first five minutes after the bleed. At forty-eight hours, the antibody group performed better on the rotating pole, a standard motor coordination test in rat stroke work. From day three onward, the two arms converged, and by day fourteen there was no significant difference in any of the four composite endpoints.
Why this matters
CGRP-blocking monoclonal antibodies (erenumab, fremanezumab, galcanezumab, eptinezumab) are first-line preventives at the chronic and severe end of the migraine population. The clinical assumption when they were approved was that a single brain bleed in a patient who happens to be on one of them should not be different from a brain bleed in anyone else. Nobody had a model to test it. This is the first that tries.
The headline pattern is early benefit, late wash. The first minutes of a bleed are the window when global cerebral perfusion is fighting against vasoconstriction triggered by the blood itself. If a drug tilts the vasoreactivity toward more flow, the brain's first reaction is to do better. By day three, autoregulation and inflammation and rebleeding and a dozen other processes have all caught up, and the early input is no longer detectable in survival or composite neurology. The early signal is real. It just does not propagate to a population-scale outcome.
The same study landed a separate pathophysiology finding that has nothing to do with fremanezumab. The authors used capsaicin (the chili-pepper compound, a standard way to provoke neuropeptide release at sensory nerve endings) on rat dura mater and on isolated trigeminal ganglia. After SAH, capsaicin still released the same amount of CGRP from the ganglion, but the dura mater released noticeably less. The signal that is supposed to come out of the perivascular nerve endings into the meningeal space was the one that broke. The signal coming from the cell bodies that generate it was unchanged. That is consistent with peripheral release exceeding what the cell body can resupply on a short timescale, and it says trigeminovascular CGRP biology in cerebrovascular disease is not the same as trigeminovascular CGRP biology in migraine.
What this is not
This is one rat model. The authors used 30 mg/kg fremanezumab intravenously. Clinical migraine dosing for an adult is 225 mg or 675 mg every one to three months, which is roughly 3 to 9 mg/kg in a 75 kg human, dosed at a steady-state interval. The rats had antibody on board for fifteen minutes when the bleed happened. A clinical patient has antibody on board for weeks. The pharmacology does not translate directly. What the paper supports is a narrow claim: pretreatment at acute dose did not worsen the outcome of a single experimental bleed in this strain of rat, and showed an early benefit signal in the first hours.
The open avenues are obvious. Whether the delayed cerebral ischemia phenotype (the late vasospasm complication that drives most SAH morbidity in humans) is affected. Whether the dura release deficit recovers, and on what timescale. Whether the early cerebral blood flow benefit holds in a milder bleed model where the vasoconstrictive response is more nuanced. None of these are answered in this paper, and the authors say so.
On peptidemodel, α-CGRP ↗ is the canonical endogenous ligand here, a 37-residue neuropeptide that docks the CALCR/RAMP1 complex on vascular and trigeminal cells. Fremanezumab itself is not a peptide and has no card; it is a humanized IgG2 antibody that sequesters the free peptide before it can reach the receptor. The receptor stays in the platform's CALCR/RAMP ↗ target page, and that is the molecular shape this work circles around.