Fremanezumab in Pediatric Chronic Migraine: An Episodic-Migraine Approval and a Negative Chronic Cohort
The first CGRP monoclonal antibody approved for children was approved for episodic migraine. This patient has chronic migraine, and the same trial program tested that separately — which turns a general question about placebo response into a concrete one about which cohort she actually belongs to.
A.K. missed nineteen days of ninth grade this term, not from any one dramatic absence but in the ordinary erosion of a headache diary nobody wants to keep: a morning here, a half-day there, a study hall spent lying down in the nurse’s office instead of finishing a worksheet. She is 14, weighs 52kg, plays club volleyball when her head allows it, and has had migraine since she was 10, now averaging sixteen headache days a month, twelve of them meeting criteria for migraine — enough to qualify as chronic. Her family has done what pediatric headache clinics ask first: a consistent sleep schedule, hydration, a headache diary, six months of biofeedback-based behavioral therapy with a psychologist who specializes in pediatric chronic pain. Improvement has been real but incomplete — headache days down from twenty to sixteen, not enough to keep her off the school’s attendance-review list.
Pediatric migraine prevention carries an evidence problem that is neither subtle nor historical. The 2017 CHAMP trial (Powers et al.) randomized children and adolescents to amitriptyline, topiramate, or placebo and found no significant difference between either active drug and placebo — a five-decade assumption about what worked in kids collapsing against a roughly fifty percent response rate that showed up regardless of which pill they actually took. Fremanezumab’s August 2025 pediatric approval rests on SPACE (Hershey et al., published in the New England Journal of Medicine), the first CGRP monoclonal antibody study built for children rather than extrapolated downward from adults by weight, and SPACE did separate from placebo: 2.5 fewer monthly migraine days against placebo’s 1.4.
That she now needs something pharmacologic is not what the room is arguing about — six months of behavioral therapy at a plateau settled that. SPACE enrolled children with no more than fourteen headache days a month. She has sixteen. Her 52kg clears the label’s 45kg floor and her age sits inside the approved 6-to-17 band, but her headache frequency does not: the approval reads for the preventive treatment of episodic migraine, and by her own diary she has chronic migraine. The program did test that population — a separate pediatric chronic migraine cohort of 289 children — and there fremanezumab and placebo were indistinguishable, 3.8 fewer monthly migraine days against 3.7. The trial that made fremanezumab available to children is not the trial that enrolled children like her, and the one that did found nothing.
Six months into behavioral therapy, weighing a first pharmacologic option
I still want to start fremanezumab, and I want to be honest about what I’m arguing against. SPACE’s chronic migraine cohort was flatly negative — 3.8 monthly migraine days against placebo’s 3.7. But look at what the placebo arm did. It dropped 3.7 days on its own. A trial whose placebo arm improves that much has almost no room left to show a drug effect, and a null produced that way tells us the study couldn’t measure the question, not that the answer is no.
A.K. also isn’t where that cohort started. She has already been through six months of behavioral therapy and plateaued, which is not the same population as children who might have improved on attention alone.
I want to separate two things you’ve combined. The first is factual and doesn’t depend on how we read the null: the pediatric approval is for episodic migraine, capped at fourteen headache days a month. She has sixteen. Prescribing here is off-label, and the family should hear that word from us rather than discover it from a denial letter.
The second is the argument itself. “The placebo response swamped the effect” is exactly what was said after CHAMP, and it is available after every negative pediatric trial that will ever be run — there is no result it couldn’t absorb. I’m not saying the drug doesn’t work in these children. I’m saying that reasoning can’t be what tells us it does, and right now it’s the only thing on offer.
Something neither of you has said out loud yet, and I think it reframes the whole conversation: all three drugs are in the same evidential position. Topiramate and amitriptyline failed against placebo in CHAMP. Fremanezumab failed against placebo in the chronic cohort of its own pivotal program. There is no option on this list with a positive pediatric trial in a child who looks like her.
So the question isn’t which drug the evidence favors, because it favors none of them. It’s whether we try anything, and on what terms. I’d try fremanezumab — the tolerability profile in SPACE was genuinely clean and she is missing school — but with the family told plainly that this is off-label for her pattern and that the matching cohort was negative, and with a stop rule set now: twelve weeks, and if her school attendance hasn’t moved, we stop rather than titrate hope. A reassessment without a stopping condition is how a child ends up on a drug for two years because nobody wanted to be the one to call it.
Agreed: fremanezumab 225mg monthly started as an explicitly off-label trial for chronic migraine, with the family told directly that the pediatric approval covers episodic migraine and that the chronic-migraine cohort of the approving trial did not separate from placebo. Behavioral therapy continued. Twelve-week assessment set against both her headache diary and her school attendance record.
If school attendance has not improved at twelve weeks, fremanezumab stops. All three voices wanted the stopping condition fixed before any result existed to argue about.
The Pediatric Headache Neurologist reads the chronic cohort's null as an underpowered trial swamped by its own placebo arm; the Clinical Pharmacologist reads that explanation as unfalsifiable and declines to weigh it. Neither moved, and the plan was built so it doesn't depend on which is right.