Life With What
There's a whole family waiting for you →

Asteraceae

Tanacetum parthenium

Feverfew

Mixed Caution
Part used
leafaerial
Native range
The Balkans and Anatolia; naturalised widely
Morphology
A bushy aromatic perennial to about sixty centimetres with yellow-green pinnate leaves and small daisy flowers with short white rays.
Also known as
Chrysanthemum parthenium

What is in it

sesquiterpene lactones parthenolide
flavonoids
volatile oil

Evidence

Trials for migraine prevention are inconsistent. Cochrane found the evidence insufficient, though a well-conducted trial of a stable CO2 extract was positive. Parthenolide content in commercial products varies enormously and several trials likely used inert material.

Safety

Mouth ulceration and loss of taste occur with chewed fresh leaf. A post-feverfew syndrome of rebound headache, anxiety and stiffness is described on abrupt cessation. Allergy in those sensitive to Asteraceae. Contraindicated in pregnancy.

Interactions: Additive bleeding risk with anticoagulants

Regulatory: Permitted as a listed medicine ingredient.

How the traditions classify it

Galenic Hot and dry. Used for fevers — the common name is a corruption of febrifuge — and for gynaecological complaints.

Three systems, rating the same plant. Where they disagree, none of them is correcting the others — they are separate frameworks that arrived at their categories independently.

Historical use

Recorded from Dioscorides onward and used across Europe for fever, headache and childbirth. Its modern reputation for migraine dates from the 1970s, when British migraine sufferers began reporting benefit and researchers followed it up.

Recorded as history and attributed where possible. Long use establishes that a plant was widely used, not that it worked, and not that it was safe.

Worth knowing

Feverfew is the clearest case in this collection of a herb whose trials probably failed because the material was inert.

Parthenolide is unstable and degrades in storage. Surveys of commercial feverfew products have repeatedly found parthenolide content ranging from a reasonable level down to essentially nothing, including products that met their label claim for dried leaf weight while containing no detectable active compound.

So a body of inconsistent trials could reflect a herb that does not work, or a herb tested using material that had degraded on a shelf. This is the preparation-variability problem in its purest form, and it is why the one clearly positive trial used a stabilised CO2 extract with measured content.

Nomenclature verified against GBIF backbone taxonomy.