Migraine Prevention: What the Evidence Actually Shows for Riboflavin, Magnesium, CoQ10, and Feverfew

This article is for educational purposes intended for healthcare practitioners and informed readers. It does not constitute medical advice.

Why Migraine Prevention Draws So Much Nutraceutical Interest

Migraine is a common, disabling neurovascular disorder, and the standard pharmacological preventives (beta-blockers, anticonvulsants, tricyclics, and the newer CGRP-targeted therapies) are effective for many patients but come with side-effect burdens that lead a substantial proportion to discontinue treatment or seek adjunctive options. This has made migraine one of the more heavily researched areas in nutraceutical medicine, with an unusual amount of randomised, placebo-controlled trial data for a field that is often short on rigorous evidence.

That trial base is exactly what makes migraine prevention a useful case study in honest evidence reporting. Some of the commonly recommended naturopathic interventions (high-dose riboflavin, coenzyme Q10) rest on real, if small, randomised trials with clear effect sizes. Others (magnesium) have genuinely mixed results depending on which trial is read. And one, butterbur, is a cautionary story: it carried the highest evidence rating of any complementary treatment in the 2012 American Academy of Neurology and American Headache Society (AAN/AHS) guideline, before hepatotoxicity case reports made its use inadvisable outside tightly controlled, purified formulations. A clinician who recommends nutraceuticals for migraine needs to know which of these stories applies to which supplement.

This review covers the primary trial evidence for riboflavin, magnesium, coenzyme Q10, and feverfew, and explains why butterbur is no longer recommended by mainstream neurology bodies despite once holding their strongest evidence grade.

The 2012 AAN/AHS Evidence Framework

Before looking at individual nutrients, it helps to see how mainstream neurology graded them. In 2012, the Quality Standards Subcommittee of the AAN and the AHS published an evidence-based guideline update on NSAIDs and complementary treatments for episodic migraine prevention, reviewing 49 Class I or Class II trials published between 1999 and 2009.

Their ratings for the nutraceuticals covered in this article:

This is the framework most clinicians still cite. It is also, on the butterbur point, now out of date, which is precisely why understanding the underlying trials (rather than just the letter grade) matters.

Riboflavin (Vitamin B2): The Best-Replicated Single Trial

The rationale for riboflavin in migraine prevention comes from the mitochondrial energy hypothesis: some evidence points to impaired mitochondrial oxidative phosphorylation as a contributing factor in migraine pathophysiology, and riboflavin is a precursor for flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), essential cofactors in the mitochondrial electron transport chain.

The pivotal trial is Schoenen, Jacquy, and Lenaerts (1998), a randomised, placebo-controlled trial of 55 migraine patients comparing riboflavin 400 mg daily against placebo over three months. Using intention-to-treat analysis, riboflavin was superior to placebo for reducing attack frequency (p = 0.005) and headache days (p = 0.012). The 50% responder rate (patients whose headache days improved by at least half) was 59% for riboflavin versus 15% for placebo (p = 0.002), giving a number-needed-to-treat of 2.3, an unusually strong figure for a nutraceutical trial. Adverse events were minor: one case each of diarrhoea and polyuria in the riboflavin group.

This is a single trial with a modest sample size, and it has not been replicated at the same scale since 1998, a limitation worth stating plainly rather than glossing over. A 2017 systematic review by Thompson and colleagues in the Journal of Clinical Pharmacy and Therapeutics found the broader riboflavin literature thinner and more heterogeneous than the strength of the original trial might suggest, though it did not contradict the core finding. What riboflavin has going for it clinically is an excellent safety profile, negligible cost, and a plausible mechanism, which is why it remains a reasonable first-line nutraceutical option even though the evidence base is narrower than its reputation implies.

Typical dosing: 400 mg/day, taken as a single dose. Benefit is not usually apparent for 6 to 12 weeks, and patients should be counselled accordingly. Riboflavin causes harmless bright yellow discolouration of urine, which is worth mentioning in advance so it is not mistaken for an adverse effect.

Magnesium: Genuinely Mixed Evidence, Not Settled Science

Magnesium is frequently presented in naturopathic sources as straightforwardly effective for migraine prevention. The trial record does not support that level of confidence, and this is the clearest example in this review of why reading beyond the marketing summary matters.

The positive trial: Peikert, Wilimzig, and Köhne-Volland (1996), published in Cephalalgia, randomised 81 patients to oral trimagnesium dicitrate 600 mg (24 mmol) daily or placebo for 12 weeks after a 4-week baseline. In weeks 9 to 12, attack frequency fell 41.6% in the magnesium group versus 15.8% with placebo (p < 0.05), with migraine days and symptomatic medication use also significantly reduced. Diarrhoea (18.6%) and gastric irritation (4.7%) were the main adverse events, consistent with the known GI tolerability profile of citrate-based magnesium salts at this dose.

The negative trial: Pfaffenrath and colleagues (1996), also in Cephalalgia and published the same year, randomised patients to 10 mmol magnesium twice daily or placebo. This trial was stopped early after an interim analysis of 69 patients found no meaningful difference in responder rate (28.6% magnesium versus 29.4% placebo) and no benefit on migraine days or attack frequency. Adverse events were more common with magnesium (45.7% versus 23.5%, mostly soft stool and diarrhoea) without a corresponding efficacy benefit.

Two well-designed, similarly sized, double-blind trials published in the same journal in the same year reached opposite conclusions. A plausible explanation is magnesium salt and dosing form (the positive trial used trimagnesium dicitrate; formulation affects both bioavailability and GI tolerability, and poor tolerability can undermine treatment adherence in trials that do not report per-protocol as well as intention-to-treat results). Later reviews, including a 2018 systematic review by von Luckner and Riederer in Headache, have generally concluded that the evidence supports a plausible but not firmly established benefit, consistent with the AAN/AHS Level B ("probably effective") rating rather than anything stronger.

Typical dosing in trials showing benefit: 600 mg elemental magnesium daily (as trimagnesium dicitrate in the positive trial), in divided doses to improve GI tolerability. Diarrhoea is the main limiting side effect and is dose-related; magnesium glycinate or magnesium bisglycinate is commonly substituted in clinical practice for better tolerability, though this specific salt form has less direct trial data in migraine than trimagnesium dicitrate.

Coenzyme Q10: Promising but Thin

CoQ10 shares the mitochondrial rationale with riboflavin: as a component of the electron transport chain and an antioxidant, CoQ10 has been proposed to support the same energy metabolism pathway implicated in migraine pathophysiology.

The main trial is Sándor and colleagues (2005) in Neurology: a double-blind, randomised, placebo-controlled trial of CoQ10 (100 mg three times daily, 300 mg total) versus placebo in 42 migraine patients over three months. CoQ10 was superior to placebo for attack frequency, headache days, and days with nausea by the third treatment month, and was well tolerated. The 50% responder rate for attack frequency was 47.6% for CoQ10 versus 14.4% for placebo, giving a number-needed-to-treat of 3.

This is a smaller trial than the riboflavin study (42 versus 55 participants) and has not been substantially extended by larger confirmatory RCTs since, which is reflected in its more cautious Level C ("possibly effective") rating from the AAN/AHS, one grade below riboflavin and magnesium. The mechanistic overlap between CoQ10 and other mitochondrial-support strategies used in naturopathic practice is discussed further in our mitochondrial dysfunction and functional medicine guide, and practitioners weighing which CoQ10 formulation to use should see our CoQ10 and ubiquinol comparison for the bioavailability evidence between ubiquinone and ubiquinol forms.

Typical dosing: 100 mg three times daily (300 mg/day total), the dose used in the positive trial. CoQ10 is fat-soluble and better absorbed taken with meals containing some fat.

Feverfew: Low-Quality Evidence, Modest Effect at Best

Feverfew (Tanacetum parthenium) has one of the longer traditional-use histories of any migraine botanical, but the Cochrane systematic review by Wider, Pittler, and Ernst (2015) is a useful corrective to overstated claims about it.

The review identified six eligible double-blind, placebo-controlled trials totalling 561 participants. The results were inconsistent: three earlier trials with small samples (17 to 60 participants) reported positive effects, while two more rigorous trials (50 and 147 participants) found no significant difference from placebo. The largest and most recent trial included in the update (218 participants, using a standardised extract dosed from a prior dose-finding study) reported a reduction in migraine frequency of 1.9 attacks per month versus 1.3 for placebo, a between-group difference of 0.6 attacks per month. No significant differences were found for attack intensity, duration, or nausea.

The review's own conclusion is unusually candid for a supplement with this much market presence: "this constitutes low quality evidence, which needs to be confirmed in larger rigorous trials with stable feverfew extracts and clearly defined migraine populations before firm conclusions can be drawn." Feverfew was not associated with significant safety concerns beyond mild gastrointestinal complaints and mouth ulcers, which is the main point in its favour: even a modest, uncertain benefit is a reasonable trade against a low-risk profile, but it should be presented to patients as exactly that, modest and uncertain, rather than as an established preventive.

Butterbur: Why a Level A Recommendation Was Withdrawn

Butterbur (Petasites hybridus) is the most important cautionary example in this review, and the reason it deserves more space here than its current clinical relevance might suggest.

In the 2012 AAN/AHS guideline, Petasites was the only complementary treatment to receive a Level A rating ("effective, should be offered"), based on trials using a specific proprietary extract. That guideline is still findable and still cited in some naturopathic and integrative sources without qualification.

The problem is hepatotoxicity. Butterbur is a member of the Asteraceae family and, unless carefully processed, contains pyrrolizidine alkaloids (PAs), compounds with well-established hepatotoxic, carcinogenic, and mutagenic potential once metabolically activated in the liver. A pharmacovigilance case series by Anderson and Borlak (2019), published in the Journal of Clinical Medicine, reviewed spontaneous reports of herb-induced liver injury (HILI) associated with a PA-reduced butterbur extract (Petadolex). Most reported cases involved mild, transient liver enzyme elevation, but nine patients developed severe HILI, with a median time to onset of roughly 103 days, and two required liver transplantation. Causality was assessed as likely in 3 cases, possible in 13, and unlikely or unclassifiable in the remainder, an important nuance, since not every reported case is confidently attributable to the product. Even accounting for that uncertainty, the signal was strong enough that mainstream neurology guidance has moved away from recommending butterbur, and it has been withdrawn from sale in some markets, including the UK, over these concerns.

The practical lesson is not that butterbur definitely causes liver injury in every user, nor that the original trials showing efficacy were flawed. It is that an herbal extract's safety profile can look acceptable in short trials and still carry a rare but serious signal that only emerges once a much larger population has used it over longer periods, and that guideline recommendations, even Level A ones, need to be checked against the current literature rather than cited from a single source year after year. Any practitioner still recommending butterbur for migraine should, at minimum, ensure the product is a certified PA-free extract, use it for a defined trial period rather than indefinitely, and monitor liver function.

Practical Summary

None of these interventions replace an accurate migraine diagnosis or exclusion of secondary headache causes. Sudden, severe ("thunderclap") headache, new headache with fever or neurological deficit, or a marked change from a patient's usual headache pattern warrants urgent medical assessment before any preventive protocol, natural or pharmaceutical, is considered.

Frequently Asked Questions

Is riboflavin or magnesium better for migraine prevention?

They are not directly compared in a head-to-head trial, so a definitive answer is not available. Riboflavin has the single strongest trial result (NNT 2.3) but rests on one unreplicated 1998 study. Magnesium has genuinely mixed trial evidence, with formulation and dose likely explaining some of the discrepancy between positive and negative results. Both carry a Level B "probably effective" rating and a favourable safety profile, and in practice many naturopathic protocols combine them rather than choosing one.

Why is butterbur no longer recommended if it once had the best evidence rating?

The 2012 Level A rating reflected the migraine-prevention trial data available at the time. Since then, pharmacovigilance case series have documented herb-induced liver injury, including severe cases requiring transplantation, associated with butterbur extracts. The efficacy trials were not necessarily wrong, but efficacy and long-term safety are separate questions, and the safety signal that emerged after wider use changed the risk-benefit calculation. This is a useful example of why any single guideline citation needs to be checked against more recent literature.

Is coenzyme Q10 worth trying if the evidence is weaker than riboflavin?

It is a reasonable adjunct given its favourable safety profile and plausible shared mechanism with riboflavin, but patients should be told honestly that the trial base is a single 42-person study, smaller and less replicated than the riboflavin evidence. Framing it as "possibly effective" rather than "proven" is both accurate and consistent with how the AAN/AHS graded it.

References

• Holland S, Silberstein SD, Freitag F, Dodick DW, Argoff C, Ashman E. Evidence-based guideline update: NSAIDs and other complementary treatments for episodic migraine prevention in adults. Neurology. 2012;78(17):1346-1353.
• Schoenen J, Jacquy J, Lenaerts M. Effectiveness of high-dose riboflavin in migraine prophylaxis: a randomized controlled trial. Neurology. 1998;50(2):466-470.
• Sándor PS, Di Clemente L, Coppola G, et al. Efficacy of coenzyme Q10 in migraine prophylaxis: a randomized controlled trial. Neurology. 2005;64(4):713-715.
• Peikert A, Wilimzig C, Köhne-Volland R. Prophylaxis of migraine with oral magnesium: results from a prospective, multi-center, placebo-controlled and double-blind randomized study. Cephalalgia. 1996;16(4):257-263.
• Pfaffenrath V, Wessely P, Meyer C, et al. Magnesium in the prophylaxis of migraine: a double-blind placebo-controlled study. Cephalalgia. 1996;16(6):436-440.
• Wider B, Pittler MH, Ernst E. Feverfew for preventing migraine. Cochrane Database of Systematic Reviews. 2015;4:CD002286.
• Anderson N, Borlak J. Hepatobiliary events in migraine therapy with herbs: the case of Petadolex, a Petasites hybridus extract. Journal of Clinical Medicine. 2019;8(5):652.

This article is intended for educational purposes and professional practice reference. It does not constitute individual medical advice. Clinical decisions should be made in the context of a full patient assessment by a qualified practitioner.