What This Guideline Is

The American Academy of Neurology and the American Headache Society have published a joint practice guideline on pharmacologic migraine prevention in adults. It appears in Neurology and is endorsed by the American Academy of Family Physicians.

The panel convened in January 2018. The American Headache Society joined the project in March 2022. The systematic review underpinning it covers studies published through June 2024 and is published separately as a companion paper. The result is 17 recommendations broken into 52 individual statements.

Guidelines are worth reading differently from single studies. A trial tells you what happened in one group of people. A guideline tells you what a panel of specialists concluded after weighing all of the trials together, including the ones that did not work out, and what level of confidence they are willing to attach to each conclusion.

When to Start a Preventive

Three statements cover this, all Level B.

Clinicians should tell every patient with migraine that effective preventive treatments exist for frequent attacks. Preventive treatment should be offered to anyone with four or more migraine days per month, or four or more moderate to severe headache days per month. It should also be offered to anyone with substantial disability from migraine, regardless of how the day count falls.

The disability clause matters. Someone with three severe attacks a month that each wipe out two days of work is a candidate under this guideline even though they sit below the four-day threshold.

How the Choice Is Framed

The guideline states plainly that no single medication has been shown to be clearly superior for reducing attack frequency, and that head to head trials have generally not been of sufficient quality to establish one drug over another.

Instead of ranking drugs, the panel compares them on four properties and then asks what the individual patient is optimizing for.

Three Level B statements sit underneath this. Clinicians should inform patients of appropriate choices given their medical and psychiatric history, current medications and contraindications. They should discuss and understand the patient's preferences about potential adverse effects. They should do the same for treatment modality, meaning oral against injectable.

The Four Lists

Recommendation 3 sets out named drug lists. Rather than one list of best options, there are separate lists depending on what the patient is prioritizing. All four are Level B.

The pattern across the first three lists is worth noting. The CGRP-targeting drugs appear in both the efficacy list and the tolerability list, but not in the long-term safety list, because they have not been in use long enough for that question to be settled. Topiramate is the only drug that appears in the efficacy list and the long-term safety list for both migraine types, and it is absent from the tolerability list. Cost is handled separately and depends entirely on formulary and insurance.

How Long to Give It

Recommendation 14 covers the question patients ask most often once they have started something.

The underlying reason given is that superiority to placebo was typically demonstrated between 8 and 12 weeks for oral medications in the systematic review, with 12 weeks the most common timepoint. Studies of onabotulinumtoxinA evaluated response at 24 weeks. Adverse events, by contrast, were reported within days to weeks, and some led to discontinuation before any therapeutic response could appear.

Recommendation 15 covers how efficacy gets measured. Clinicians should document attack frequency, and may use a consistent tool such as a headache diary, HIT-6 or MIDAS across visits. One statement here is Level A: the assessment of treatment efficacy must be patient-centric and based on individual goals. Frequency, severity, associated symptoms, quality of life and acute medication use should all feed into the judgment.

Pregnancy, Lactation and Age

Pregnancy carries four Level A statements, the densest concentration of strong obligations in the guideline.

The onabotulinumtoxinA data amounts to one Class III study of 397 women, mostly exposed before conception or in the first trimester, with an overall fetal defect prevalence of 2.6% that the authors describe as consistent with general population rates. The guideline is explicit that this is extremely limited.

For lactation there is a single Level A statement: clinicians must counsel patients that preventives pass into breast milk in varying amounts and can affect breastfed infants, using evidence-based sources such as LactMed. Information on breast milk passage is described as scarce for the newer treatments.

The topiramate and contraception point comes with a useful nuance. Doses of 200 mg per day or lower produce minimal or no reduction in estrogen and norethindrone levels, and low-dose topiramate as used in migraine prevention has not been shown to increase unintended pregnancy compared with propranolol, amitriptyline or other oral preventives. The concern applies to the higher end of the dose range.

Comorbidities and Medication Overuse

Where one drug can treat migraine and a coexisting condition, the guideline frames that as a reasonable option that reduces pill burden and cost. It also says that if either condition fails to respond, the approach should be abandoned and each condition treated on its own terms.

The body mass recommendation is framed against the drugs that push the other way. Amitriptyline, propranolol, pizotifen and flunarizine all promote weight gain in some patients, and cross-sectional data links obesity to migraine chronification.

Three Level B statements follow. Preventive medication should be offered to those meeting criteria for medication overuse, and to those with medication-overuse headache. Preventives with evidence in this population, named as CGRP monoclonal antibodies, atogepant, onabotulinumtoxinA and topiramate, should be considered before options lacking that evidence.

One finding here changes older practice. Available evidence suggests that starting a preventive on its own is as effective as starting a preventive alongside acute medication withdrawal. The guideline flags an important limit: those studies included only small numbers of people overusing opioids or barbiturate-containing medications, so the conclusion does not extend to those drugs.

Stopping a Preventive

Only two studies of preventive cessation were of sufficient design to inform recommendations, which is a small base for a question almost every long-term patient eventually asks.

After six months of topiramate at 50 to 200 mg, randomized discontinuation produced a small increase in migraine attack days of 1.19 days per four weeks (95% CI 0.71 to 1.66), more days of acute medication use, and a decrease in migraine-related quality of life on MIDAS. After six months of galcanezumab, more than half of the month-six responders had lost their response five months after stopping.

Both statements are about having the conversation rather than about a default action. The guideline notes that patients vary in how much they want to taper an effective treatment, and that migraine is a lifelong condition with fluctuations in frequency, severity and duration.

Key Takeaways

The document does not name a best drug, and its structure explains why. Once no option is clearly superior on efficacy, the useful question becomes which property a given patient most needs optimized, and the four lists answer that directly.

For anyone already on a preventive, the timing recommendations are the most immediately usable part. A preventive that has been at a tolerated dose for less than 8 weeks has not yet had a fair trial, and one that has been at maximum tolerated dose for 12 to 24 weeks without adequate response has had one.