A patient came in last month holding her phone, scrolling through six weeks of a headache diary. She’d started Nurtec for acute attacks, then switched to Qulipta daily for prevention, and she still had twelve headache days a month. “I thought this was supposed to be the good stuff,” she said. She wasn’t wrong to expect more — gepants are effective for a lot of people. But “not working” almost never means the drug class failed outright. Usually it means the dose, the timing, the diagnosis, or something running underneath the migraine wasn’t addressed, and that’s fixable once we find it.
What gepants actually do — and don’t do
Gepants (rimegepant/Nurtec, ubrogepant/Ubrelvy, atogepant/Qulipta, zavegepant/Zavzpret) block the CGRP receptor directly. That’s different from triptans, which act on serotonin receptors to constrict blood vessels, and different from the injectable CGRP antibodies like Aimovig or Emgality, which mop up CGRP in the bloodstream over weeks. Gepants work faster to leave the system and, for most patients, cause less of the chest tightness or jaw pressure some people get from triptans. Rimegepant and atogepant also have the unusual property of working for both acute treatment and prevention, which is genuinely useful — but it also means people sometimes take an acute-dosed gepant and expect preventive-level results, or vice versa, and the mismatch looks like failure.
The reasons gepants fail — in the order I actually see them
Taken too late. This is the single most common reason an acute gepant “doesn’t work.” CGRP receptor blockade is most effective before central sensitization sets in — roughly the first 60 minutes of pain. Patients who wait to see if the headache will “just pass” and then take Nurtec two hours in are treating a fundamentally harder problem. I ask people to treat at the first clearly recognizable migraine symptom, not once the pain is severe.
Medication overuse underneath it. If someone is using an acute gepant, a triptan, or an NSAID more than 10 days a month, the brain can develop rebound sensitization that no preventive — gepant or otherwise — will out-run. I’ve had patients on Qulipta who kept using Ubrelvy four times a week for breakthrough pain; the Qulipta wasn’t failing, the overuse pattern was working against it. This has to be untangled before we can judge whether the preventive is doing its job.
Not enough time on the preventive dose. Atogepant’s preventive effect builds over the first four to twelve weeks, similar to a CGRP antibody, not overnight like an acute drug. Judging Qulipta at two weeks is judging it before it’s had a fair chance. I ask for a minimum of eight weeks at a stable dose before we call it ineffective.
Wrong migraine subtype, or not migraine at all. Gepants are built for migraine biology specifically. If the underlying pattern is actually chronic tension-type headache, cervicogenic pain, or a secondary cause, a CGRP-targeted drug has little to work with no matter the dose. This is worth revisiting honestly, especially if the pain doesn’t have migraine features like throbbing quality, nausea, or light and sound sensitivity.
Genuine non-response. Roughly a third of patients don’t respond meaningfully to any single CGRP-pathway drug, gepant or antibody. CGRP is one driver of migraine, not the only one, and some people’s attacks are propagated through other pathways we target differently.
When “gepants aren’t working” is really a diagnosis problem
I’ll be direct about this one because it comes up more than people expect: if a well-dosed, well-timed gepant trial of eight-plus weeks hasn’t moved the needle at all, and medication overuse has been ruled out, that’s a reason to revisit whether “migraine” is even the right label. I’ve relabeled patients into hemicrania continua, new daily persistent headache, and medication-overuse patterns that were driving the whole picture — each of which responds to something other than CGRP blockade. A drug that targets the wrong mechanism will look like a failed drug every time.
What I actually change next
First, I fix timing and dosing before touching the drug itself — most “failures” resolve here. Second, if overuse is present, we work through a structured taper; you can’t accurately test a preventive while the brain is still rebound-sensitized. Third, if a fair trial genuinely hasn’t worked, I’ll switch within the class (rimegepant and atogepant have different receptor kinetics, and one working where the other didn’t isn’t unusual) or move to a CGRP monoclonal antibody, which hits the same pathway through a different mechanism and timeline. Fourth, for patients who’ve now failed multiple CGRP-pathway drugs, Botox or a non-CGRP preventive like a beta-blocker or topiramate often gets better results — CGRP isn’t the only lever. And throughout, I’m reassessing whether the diagnosis itself still holds up, because the drug can only be as accurate as the label we’re treating.
If you’ve been on Nurtec, Ubrelvy, Qulipta, or Zavzpret for a full, honest trial and you’re still counting more bad days than good ones, that’s not a reason to give up on treatment — it’s a reason to get the plan re-examined by someone who will look past the prescription pad. For more on what a stalled treatment plan usually means, see Headache Not Getting Better? What Doctors Miss. If you were on a triptan before this and it also fell short, Triptans Not Working? covers that side of it, and if you’ve already tried an injectable CGRP antibody, CGRP Injections Not Working? walks through those next steps. For background on how gepants work as a class, Gepants for Migraine: How the New Pills Work is a fuller guide.
Next steps
Before changing anything on your own, run through our free headache red-flags checklist to make sure nothing about your pattern needs urgent attention. If your treatment plan has stalled despite honest effort, you can talk to a specialist and get the plan reassessed directly.
This article is for general education and isn’t a substitute for a medical evaluation. If you’re having a sudden, severe, or unusual headache, seek emergency care right away.
From the textbook. Chapter 14, “Acute Treatment of Migraine,” of my clinical textbook Headache Medicine covers this topic in full for doctors, with the evidence and decision pathways behind it. About the book →