Migraine Preventive Not Working? How Long It Takes — and When to Switch

Written and medically reviewed by Tae Ho Guak, MD, board-certified neurologist · Reviewed July 2026 · Editorial policy

The short answer, before anything else: a migraine preventive needs a fair trial before you can honestly call it a failure. For the oral drugs — topiramate, propranolol, amitriptyline, candesartan and the rest — that means at least eight weeks at the full target dose, not eight weeks of slowly working up to it. For the monthly CGRP injections, give it about three months, and sometimes longer, because a real share of people who look like non-responders at twelve weeks turn the corner by month six. Judging a preventive too early is the single most common reason I see patients abandon a drug that would have worked.

A man came to my office last spring convinced he had “tried everything.” When we went through it, the picture was different. He had been given topiramate, taken 25 mg for about ten days, felt foggy, and stopped. That is not a failed preventive. That is a starter dose, taken for a fraction of the time it needs, abandoned during the exact window when side effects are loudest and benefit hasn’t arrived yet. He hadn’t tried topiramate at all. He had sampled it.

How long a preventive really needs

Preventive medications are not painkillers. They don’t act on the headache in front of you; they slowly lower how often and how hard the attacks come. That work takes weeks, and most of the oral options have to be titrated — started low and raised in steps — specifically so your body tolerates them. So the clock that matters is not “how long since I started the pill.” It is “how long have I been at the full, intended dose.”

The American Headache Society’s guidance is clear on this: with oral preventives, you should stay at a target therapeutic dose for a minimum of eight weeks before deciding it isn’t working. Add the weeks of titration on the front end, and the honest timeline to know whether a drug like topiramate or propranolol is helping is often two to three months. Quitting at three weeks tells you nothing except that the side effects were unpleasant early — which, for most of these drugs, they are, and then usually settle.

The CGRP monoclonal antibodies — erenumab, galcanezumab, fremanezumab, eptinezumab — are faster and gentler, but they still ask for patience. The standard is to reassess at three months. What’s become clearer in the last few years is that three months isn’t the finish line. In real-world studies, a meaningful group of people who hadn’t hit the response threshold by twelve weeks did reach it by six months. So a “no” at three months on a CGRP injection is often really a “not yet.”

“I tried that and it didn’t work” — usually one of these

When someone tells me a preventive failed, four explanations cover the large majority of cases, and only one of them is the drug actually not working for them.

The first is the dose was too low. A preventive at half its effective dose is not a gentler version of the treatment — it is often no treatment at all. The second is the trial was too short, the topiramate-for-ten-days problem. The third is that it was stopped during titration, before the target dose was ever reached, usually because of early side effects that would have faded. The fourth — and this one is quietly common — is that something else in the background was blocking any preventive from working, no matter how good the drug was. I’ll come back to that, because it’s the one people miss.

None of these mean you’re out of options. They mean the medication never got a real chance, and a real chance is a completely reasonable thing to give it before moving on.

What “working” actually means

Here is where expectations quietly sabotage good treatment. A preventive is considered successful if it cuts your migraine days by about half. Not to zero — by half. If you were having twelve headache days a month and you’re now having six, that is a drug that is working, and working well, even though you still get headaches.

Patients often measure success as “I still got migraines, so it failed.” By that standard almost every preventive fails, because none of them promise a headache-free life. When I ask people to actually count — which is why I’m relentless about keeping a diary — the drug they were about to throw out has frequently taken them from ten bad days to five. That’s not failure. That’s the target being hit.

The flip side matters too: if you’re a partial responder, don’t assume the benefit has topped out at eight weeks. Continued use at the same dose can keep adding benefit over the following six to twelve months. Sometimes the right move isn’t switching — it’s staying the course a little longer.

When it genuinely is time to switch

So when is it a real failure? When you’ve been at the full target dose for at least eight weeks (or three-plus months on an injection), you’ve counted your days honestly, and you’re still not near a fifty-percent improvement — or when the side effects are genuinely intolerable at the dose you need. Those are the two legitimate reasons to change: it doesn’t work despite a fair trial, or it works but you can’t live with how it makes you feel.

At that point, switching is not defeat, and it is not the end of the road. There are several drug classes for prevention, they work through different mechanisms, and failing one says very little about whether the next will help. This is the same logic I lay out for the acute drugs in why triptans stop working and what a neurologist tries next, and for the injectables in what to do when CGRP injections don’t work. Not responding to one is information, not a verdict.

Before you blame the drug

The failure that gets missed most often isn’t the medication at all. It’s a background problem the preventive can’t overcome. The biggest one is medication-overuse headache: if you’re taking acute painkillers or triptans more than about ten to fifteen days a month, they can lock you into a daily background headache that no preventive will lift until the overuse is unwound. I see people cycle through three or four preventives, all “failing,” when the real issue was sitting in the medicine cabinet the whole time — more on that in how everyday painkillers cause daily headache.

The other quiet saboteur is the wrong diagnosis. A preventive aimed at migraine won’t do much if the problem was never migraine — if it’s actually a headache type that responds to something else entirely. When several sensible treatments all seem to fail, that’s not usually bad luck. It’s a signal that the diagnosis, not just the drug, deserves a second look — which is exactly what I walk through in why your headache isn’t getting better.

If you’re on your first preventive and tempted to quit, my honest advice is usually the same: check the dose, check the calendar, count your days, and make sure nothing in the background is blocking the road. Give a good drug a fair trial. And if it truly fails that fair trial, switch without guilt — there’s plenty left to try.

---

Take the next step:

📋 Not sure your headache is being taken seriously? Start with my free red-flag checklist — the warning signs that mean a headache needs urgent attention: Get the free headache red-flags checklist.

💬 Stuck on a preventive that isn’t working? You can talk to a specialist about whether it’s the dose, the drug, or the diagnosis.

This article is for education and is not a substitute for a personal medical evaluation. Preventive medications and their dosing should be managed with your own physician. If you have a sudden, severe “worst headache of my life,” a headache with fever and a stiff neck, weakness, confusion, or loss of vision, treat it as an emergency and go to the nearest emergency room.

— Dr. Tae-Ho Guak, MD, board-certified neurologist