Feverfew

Tanacetum parthenium (L.) Schultz Bip. is a member of Asteraceae (Compositae), the daisy family; NCBI Taxonomy files it as taxon 127999. Older names include Chrysanthemum parthenium, Matricaria parthenium, and Pyrethrum parthenium, and NCCIH lists bachelor’s buttons and featherfew among its common names. NCCIH describes it as a perennial native to parts of western Asia and the Balkans that now grows throughout the world, and EMA’s assessment report notes that it grows prolifically in gardens and other open spaces. It forms a bushy, aromatic plant with a camphor-like smell, yellowish-green leaves deeply divided like a chrysanthemum’s, and many small daisy-like flower heads, 12–22 mm across, with white rays around a yellow centre. NCCIH points out that American feverfew (Parthenium integrifolium) is a different plant.

The medicinal part is the leaf or the dried aerial parts. Its signature constituents are sesquiterpene lactones stored in glandular hairs on the leaves, flowers, and seeds. The main one is parthenolide (PubChem CID 6473881, C15H20O3), a germacranolide with a reactive α-methylene-γ-lactone ring and an epoxide. EMA’s assessment report says parthenolide makes up as much as 85% of the sesquiterpene lactone content and up to about 1% of dried leaf, is found in leaves and flower heads but not stems, and varies with season and growth stage; the European Pharmacopoeia requires at least 0.20% in the dried herb. Feverfew also contains flavonoids such as apigenin and luteolin glycosides, and an essential oil in which camphor and trans-chrysanthenyl acetate together can make up about 70% (EMA).

Product quality varies enormously. EMA cites an analysis by Nelson and colleagues of US capsules containing 25–500 mg of feverfew leaf: leaf content matched the labels, but parthenolide per capsule varied 150-fold, from 0.02 to 3.0 mg, so following label directions would deliver anywhere from 0.06 to 9.7 mg of parthenolide a day. Other analyses found commercial extracts that did not contain the labelled parthenolide, and batches from the same manufacturer that differed significantly. Parthenolide degrades with heat and moisture, and EMA advises storing feverfew cool, dry, well closed, and protected from light. Because it is not clear that parthenolide is the active ingredient, a parthenolide figure on a label is a quality marker, not a guarantee of effect.

EMA’s assessment report traces medicinal feverfew to the Materia Medica of Dioscorides and through the herbals of Dodoens, Gerard, and Culpeper in the seventeenth century. The English name is usually derived from the Latin febrifugia, a fever-reducer, and the herb was used for intermittent fevers, headache, menstrual problems and difficult labour, toothache, stomach ache, and insect bites. The name parthenium has been linked to a story told by Plutarch of a man saved with the herb after falling from the Parthenon during its construction, and alternatively to the Greek parthenios, virgin, reflecting its reputation for women’s complaints; the German name Mutterkraut, mother herb, carries the same idea. In 1772 John Hill wrote that “in the worst headache this herb exceeds whatever else is known”.

The modern revival began in Britain. EMA notes that feverfew liquid extracts and tinctures held UK product licences from the early 1970s, and that interest surged in the late 1970s when the press reported that a group of migraine sufferers in Wales had found relief after taking the leaves for some time. Johnson surveyed 300 feverfew users in 1983, and in 1985 Heptinstall and colleagues reported in The Lancet that feverfew extracts inhibited secretion from blood platelets and white blood cells. The same year Johnson and colleagues published a small placebo-controlled trial in the BMJ at the City of London Migraine Clinic, and in 1988 a Nottingham team including Heptinstall published a larger crossover trial in The Lancet, putting feverfew on the map as a migraine preventive.

Germany produced the next generation of trials, using a stable extract made with supercritical carbon dioxide, MIG-99, tested first in a dose-finding trial (Pfaffenrath, 2002) and then in the larger Diener trial (2005). Powdered feverfew capsules were registered as traditional medicines in Spain and France from 1991, and traditional-use registrations followed in the UK (2007), Austria (2012), Germany (2015), and Sweden (2018). The EU’s Committee on Herbal Medicinal Products adopted its first feverfew monograph on 25 November 2010 and a revision on 8 July 2020, accepting traditional use for migraine prophylaxis but rejecting well-established use because the clinical evidence was judged low quality. EMA opened a call for data for a periodic review of the monograph in 2026.

How feverfew might prevent migraine is not settled. Early work focused on platelets. Heptinstall and colleagues (Lancet 1985) found that feverfew extracts inhibited the release of serotonin from platelets triggered by a range of stimuli and consistently inhibited platelet aggregation, without blocking thromboxane synthesis, a pattern different from other antiplatelet agents; the extracts also inhibited secretion from white blood cells more strongly than very high concentrations of NSAIDs. Kwok and colleagues (Chemistry and Biology 2001) later showed that parthenolide binds directly to and inhibits IκB kinase beta, a key switch in the NF-κB inflammatory pathway, through its α-methylene-γ-lactone group. EMA’s assessment report notes that the same reactive group, which binds sulphur-containing sites on proteins, is thought to underlie the ability of sesquiterpene lactones to cause allergic contact dermatitis.

A more migraine-specific mechanism came from Materazzi and colleagues (Pain 2013). Migraine is triggered by release of calcitonin gene-related peptide (CGRP) from trigeminal nerves. In cell and rodent studies, parthenolide acted on TRPA1, an irritant-sensing channel on these nerves, first stimulating it as a partial agonist and then desensitizing it, which reduced CGRP release and CGRP-driven widening of blood vessels in the meninges. This is laboratory evidence only. There is a complication: in a phase I study in cancer patients, cited by EMA and MSKCC, parthenolide could not be detected in plasma at doses up to 4 mg a day, and a trial of an alcoholic extract standardized to parthenolide (De Weerdt, 1996) found no benefit, so other constituents may matter, or the effective preparation may be the whole leaf.

On drug metabolism, MSKCC cites laboratory evidence that feverfew inhibits CYP1A2, 2C8, 2C9, 2C19, 2D6, and 3A4, with unknown clinical relevance, and EMA’s assessment report describes a suspected case of propranolol toxicity in which inhibition of drug metabolism was proposed as the mechanism. No clinical drug interaction studies have been carried out.

Migraine prevention rests on six double-blind trials. Johnson and colleagues (BMJ 1985) studied 17 people who were already eating fresh feverfew leaves daily: eight continued on freeze-dried feverfew capsules (25 mg twice daily) and nine were switched to placebo. Those on placebo had a significant increase in the frequency and severity of headache, nausea, and vomiting, while the feverfew group stayed stable. Murphy, Heptinstall, and Mitchell (Lancet 1988) randomized 72 people to one capsule of dried feverfew leaf daily or placebo, for four months each in a crossover design; 59 provided full data. Feverfew was associated with fewer and less severe attacks (3.6 versus 4.7 attacks per two months, as summarized by EMA) and less vomiting, with no change in attack duration and no serious side effects. EMA’s assessors note the small size, the lack of a washout period, and unclear randomization methods.

Later trials were more rigorous and less flattering. De Weerdt and colleagues (Phytomedicine 1996) found no effect on headache number or severity in 50 people given an alcoholic extract containing 0.5 mg of parthenolide daily. Pfaffenrath and colleagues (Cephalalgia 2002) tested three doses of the CO2 extract MIG-99 in 147 people for 12 weeks and found no significant difference from placebo overall; only a predefined subgroup of 49 people with at least four attacks a month improved, most clearly with 6.25 mg three times daily. Diener and colleagues (Cephalalgia 2005) then tested 6.25 mg three times daily for 16 weeks, with 170 people in the intention-to-treat analysis. Attacks fell from 4.76 a month by 1.9 with MIG-99 and by 1.3 with placebo, a small but statistically significant difference, and adverse events possibly related to treatment occurred in 8.4% versus 10.2%. EMA notes that use of acute painkillers during the trial was allowed but not reported.

Wider, Pittler, and Ernst (Cochrane 2015) included those six trials, with 561 participants. The trials were too different to pool. The reviewers concluded that the Diener trial adds some positive evidence, a difference of 0.6 attacks per month, but that this is low quality evidence that needs confirmation in larger rigorous trials with stable extracts; they found no major safety concerns. NCCIH cites a 2020 review of seven studies with 634 participants that found the studies small, varied in product and dose, and inconsistent. The American Academy of Neurology and American Headache Society guideline (Holland, Neurology 2012) rated MIG-99 as probably effective for preventing episodic migraine (Level B), alongside magnesium, riboflavin, and several NSAIDs. EMA judged the totality of evidence too weak for well-established use and accepts feverfew only as a traditional medicine.

Other uses have little support. The only controlled trial in rheumatoid arthritis (Pattrick and colleagues, Annals of the Rheumatic Diseases 1989) gave 41 women 70–86 mg of dried chopped feverfew or placebo daily for six weeks and found no important differences in any clinical or laboratory measure. NCCIH says there is little or no evidence for other conditions, and MSKCC lists no scientific support for painful or heavy periods and no human data for psoriasis. Parthenolide has anticancer activity in laboratory studies, but because it could not be detected in patients’ blood, researchers turned to a more soluble synthetic analogue (MSKCC); feverfew is not a cancer treatment.

EU traditional use (EMA/HMPC/48715/2017, revision 1, adopted 8 July 2020): powdered feverfew herb in solid oral dosage forms for prophylaxis of migraine headaches in adults, after serious conditions have been excluded by a doctor. Single dose 100 mg once daily or 200 mg three times daily; daily dose 100–600 mg. EMA says the 100 mg daily dose may be increased gradually until an effect is obtained, not exceeding 600 mg a day. Two months is the usual period of treatment needed to see an effect; if migraines still recur after two months, consult a doctor. Not recommended under 18.

Trial doses were mostly lower than EMA’s maximum. Johnson’s patients took 50 mg of freeze-dried leaf daily, close to the roughly 60 mg they had been getting from about two and a half fresh leaves; Murphy’s capsules contained on average 82 mg of dried leaf; and the MIG-99 trials used 6.25 mg of a CO2 extract three times daily. Results with one product cannot be assumed for another, especially given the wide variation in parthenolide content between brands. Feverfew is a preventive: take it daily, keep a headache diary, and judge the effect after two to three months, the period EMA’s assessors suggest for evaluating it. It is not designed to stop an attack once it has started.

Chewing fresh leaves, the folk method, is the route most associated with mouth ulcers and sore, swollen lips and tongue; EMA’s assessment report, citing the WHO monograph, says the local reaction may be avoided by using encapsulated products. Choose a product that states the plant part, the amount of herb or extract, and ideally its parthenolide content, from a manufacturer with independent quality testing. If you decide to stop after months or years of use, consider reducing the dose gradually and be aware that headaches may temporarily rebound; tapering is a reasonable precaution, although it has not been tested.