Migraine Prevention: The Evidence-Based Supplement Protocol

← Back to articles

Migraine prevention is one of the few conditions where several supplements carry guideline-level support. The 2012 American Academy of Neurology / American Headache Society evidence-based update rated magnesium, riboflavin, and MIG-99 (feverfew) as Level B (probably effective), and coenzyme Q10 as Level C (possibly effective). Butterbur received the strongest grade (Level A) at the time, but is no longer recommended for routine use because of pyrrolizidine-alkaloid hepatotoxicity (covered below). Importantly, the trials behind these grades are mostly small, and later meta-analyses have been more equivocal than the early single RCTs suggested. Supplements are an adjunct to — not a replacement for — sleep, hydration, trigger management, and, where needed, prescription preventives such as CGRP-targeting therapy.

Magnesium — 400–600 mg elemental daily (moderate evidence)

Magnesium is the best-studied option. The Peikert RCT (81 adults) found oral magnesium 600 mg/day cut attack frequency by 41.6% versus 15.8% on placebo over weeks 9–12. A 2024 dose-response meta-analysis of 22 trials reported a mean reduction of roughly 2.5 attacks/month and about 1.7 fewer monthly migraine days versus control, and a 2016 meta-analysis of 10 oral-magnesium prophylaxis trials likewise found reduced frequency and intensity. Not all reviews agree — a 2019 systematic review rated the magnesium severity evidence as "low strength" and non-significant — so treat the effect as real but modest. Magnesium citrate or glycinate is better tolerated than oxide; loose stools are dose-limiting. The main caution is reduced clearance in advanced kidney disease.

Riboflavin (vitamin B2) — 400 mg daily (moderate evidence)

Schoenen's 1998 RCT (55 adults) established the 400 mg/day dose: 59% of riboflavin patients were responders (≥50% fewer attacks) versus 15% on placebo, with a number-needed-to-treat of about 2.3. The 2024 dose-response meta-analysis found riboflavin reduced attack frequency by roughly 1.3 attacks/month. The proposed mechanism is support of mitochondrial flavin-dependent energy metabolism. Riboflavin is inexpensive and very well tolerated; the only common effect is harmless bright-yellow urine. Note that most positive data is in adults — pediatric trials have been less consistent.

Coenzyme Q10 — 100 mg three times daily (limited / mixed evidence)

The Sándor RCT (42 adults) found CoQ10 300 mg/day raised the 50%-responder rate to 48% versus 14% on placebo. But the meta-analytic picture is genuinely mixed: a 2021 meta-analysis (6 trials, 371 participants) found CoQ10 reduced attack frequency and duration but not severity, whereas a 2019 meta-analysis found no statistically significant effect on frequency, duration, or severity and graded the evidence only "moderate strength." CoQ10 is safe and well tolerated, so it is a reasonable add-on, but the honest summary is "possibly effective," not proven.

Feverfew (Tanacetum parthenium) — standardized extract (limited evidence)

The MIG-99 RCT (170 adults) of a standardized CO₂ feverfew extract (6.25 mg three times daily) found a modest reduction of about 1.9 versus 1.3 attacks/month over placebo, with an odds ratio of 3.4 favoring feverfew on responder analysis. However, a Cochrane review of 5 trials (343 patients) concluded results were mixed and did not convincingly establish efficacy over placebo, though feverfew appeared safe. Use a standardized extract; taper rather than stopping abruptly (post-feverfew rebound headache has been described). Avoid in pregnancy and in those allergic to the daisy (Asteraceae) family.

What doesn't work, or is overhyped

Butterbur carried the AAN's only Level A rating, but raw and inadequately purified Petasites contains hepatotoxic pyrrolizidine alkaloids; cases of liver injury led the AAN to retract its recommendation pending safer sourcing, and the American Headache Society no longer endorses it for routine use. Avoid 5-HTP alongside triptans or SSRIs/SNRIs (serotonin-syndrome risk). Melatonin has some signal for cyclical and cluster-type headache but is not established for routine migraine prevention. Omega-3 fatty acids were rated by the AAN as having conflicting or inadequate evidence. CBD has thin migraine-specific data and frequent label-accuracy problems.

How to run the protocol

Keep a 1-month baseline headache diary, then add components one at a time roughly 4 weeks apart so you can attribute any effect. Magnesium and riboflavin are the most evidence-supported and lowest-risk starting points; CoQ10 and feverfew are reasonable add-ons. Give any preventive a full 8–12 weeks before judging it, and use a ≥50% reduction in monthly migraine days as the responder benchmark. Non-responders should discuss prescription prophylaxis (including CGRP monoclonal antibodies or gepants) with a clinician. See the condition page and the related menstrual migraine protocol.