Butterbur

Petasites hybridus (L.) G.Gaertn., B.Mey. & Scherb. belongs to Asteraceae, the daisy family, and is a close relative of coltsfoot (Tussilago farfara); Grieve notes that Linnaeus placed it in the coltsfoot genus, and NCCIH lists Petasites officinalis and Tussilago hybrida among its older names. The Botanical Society of Britain and Ireland (BSBI) describes it as a European temperate species found from Spain and Portugal north to southern Sweden and Norway, to about 63° N, and east into neighboring parts of Asia, and introduced in parts of Scandinavia, North America, and New Zealand. It is a very robust perennial of moist, fertile, often alluvial soils by rivers and streams, in wet meadows, marshes, flood plains, copses, and roadside ditches, recorded up to 433 m in County Durham. It spreads mainly by a thick, brittle, creeping rhizome, blackish outside and white inside, with a bitter taste (Grieve). The leaves die down in autumn, leaving buds buried in the soil over winter, and BSBI notes that once its dense canopy expands in April it shades out competing plants, so that it often grows in stands with almost nothing beneath.

The flowers come first. From late February or early March into April or May, the rhizome sends up thick, scaly, hairy flowering stems up to 40 cm tall, each carrying from 50 to more than 100 small flower heads in a dense spike, variously described as pink, purple, or pale lilac (BSBI). Every floret is tubular; there are no petal-like rays. Plantlife notes that the densely packed flowers are popular with bees and a valuable source of nectar early in the year, when little else is in bloom. As the flowers fade the leaves expand: all basal, rounded, and deeply heart-shaped at the base, often more than 30 cm across and occasionally up to 90 cm, on stout, hollow stalks. They are cottony at first and stay grey and downy underneath. Grieve’s A Modern Herbal (1931) calls them the largest leaves of any plant in Great Britain, and the BSBI account notes that the plant could almost be mistaken for rhubarb; wild rhubarb and bog rhubarb are among its folk names.

Butterbur is dioecious, with functionally male and female plants. Male heads hold nectar-rich florets that shed pollen; female heads hold around 100 to 150 female florets, and after flowering their stems keep growing, often past a meter, to release seeds on white, wind-borne plumes. BSBI records that female plants are frequent only in northern and central England, while male clones are widespread, many perhaps planted as early food for hive bees; Grieve notes that Swedish farmers planted butterbur near their beehives. Chemically, the plant is known for petasins, sesquiterpene esters that include petasin (PubChem CID 5281526, C20H28O3) and its isomer isopetasin (CID 5318627), and for pyrrolizidine alkaloids such as senecionine and senkirkine (Kulinowski, Journal of Ethnopharmacology 2022). In Swiss wild populations, Wildi and colleagues (Planta Medica 1998) found 7.4–15.3 mg of petasin per gram of dried rhizome and 3.3–11.4 mg/g in leaves, but 5–90 ppm of alkaloids in rhizomes against 0.02–1.50 ppm in leaves. The rhizome used for migraine extracts is the part richest in toxic alkaloids.

Butterbur’s names describe its leaves. Gerard’s Herbal explains that the Greek name comes from the hugeness of the leaf, like a petasos or hat, big enough to keep a man’s head from rain and sun, and Grieve links petasos to the felt hats worn by shepherds and by Mercury in classical images. Kulinowski and colleagues quote the first-century Greek physician Dioscorides describing a plant with large, hat-shaped leaves, “as if they were mushrooms,” good for malignant ulcers. The English name is usually explained by the use of the leaves to wrap butter in hot weather, and BSBI counts at least 29 local English names, most referring to the leaf as a shelter or a butter wrapper. Plantlife lists devil’s hat and umbrella plant, and NCCIH notes names in other languages meaning rain hat or hat plant. Lyte’s herbal of 1578, quoted by Grieve, says a single leaf is large enough to cover a small table, as with a carpet, and Grieve adds that farmyard poultry sheltered under the leaves from rain and the noonday sun.

The other strand of its reputation is plague. German names such as Pestwurz, plague root, date from the Middle Ages, when the smell and smoke of burning butterbur roots were thought to fight the plague (NCCIH; Kulinowski), and Plantlife records the old English name pestilence wort. Grieve explains that the name marked the plant as a remedy in time of pestilence, not a cause of it, and quotes Lyte calling it “a soveraigne medicine against the plague.” Gerard recommended the dried root, powdered and drunk in wine, against plague and “pestilent fevers” because it provoked sweating, and also against worms and “naughty filthy ulcers”; Culpeper called it “a great strengthener of the heart” and good for those who “wheeze much or are shortwinded.” By the early twentieth century Grieve described the rhizome as a heart tonic and diuretic, given as a warm decoction for fevers, asthma, colds, and urinary complaints. She also recorded a piece of love divination in which a young woman sowing butterbur seed before sunrise on a Friday would see her future husband mowing nearby.

The modern story is one of a promising medicine overtaken by safety questions. Anderson and Borlak (Journal of Clinical Medicine 2019) report that the maker of Petadolex switched its extraction solvent from methylene chloride to carbon dioxide in 1988 to improve removal of pyrrolizidine alkaloids. By early 2004 the independent German drug bulletin arznei-telegramm counted six German reports of severe liver damage linked to the extract, and Swissmedic, the Swiss regulator, revoked authorization of the products sold in Switzerland, judging the link possible to probable in every case and the balance of benefit and harm negative. In Germany the regulator, BfArM, refused to renew Petadolex’s authorization on formal grounds, treating the carbon dioxide extract as a different product from the one originally approved, and it was no longer available there from mid-July 2009 (Ärzte Zeitung; Kulinowski). In the UK, the MHRA declined a traditional herbal registration in 2011 and in January 2012 advised consumers not to take unlicensed butterbur products, advice it repeated in 2021. Meanwhile US and Canadian headache guidelines had endorsed butterbur in 2012, before the AAN retired its guideline in 2015.

Most research points to the petasins, concentrated in the rhizome, as the active constituents, though MSKCC notes that a butterbur extract also inhibited the inflammatory enzyme COX-2 in the laboratory in a way that did not track petasin content. In human white blood cells, Thomet and colleagues (Biochemical Pharmacology 2001) found that the leaf extract Ze 339 and isolated petasin blocked production of leukotrienes, inflammatory messengers involved in allergy, and also blocked the rise in calcium inside the cells that precedes it; zileuton, a prescription leukotriene blocker, inhibited leukotrienes but not the calcium signal. That is the main laboratory rationale for butterbur in hay fever. For migraine, Horak and colleagues (Journal of Pharmacology and Experimental Therapeutics 2009) found that petasins block Cav2.1 calcium channels, which control the release of nerve transmitters and contribute to signaling in the trigeminal pain system involved in migraine. The sulfur-containing S-petasin and iso-S-petasin were the most potent, and the block grew stronger with rapid, repeated stimulation.

A second migraine mechanism involves calcitonin gene-related peptide (CGRP), a messenger released by sensory nerves around the brain during migraine attacks. Benemei and colleagues (British Journal of Pharmacology 2017) showed in rodent tissues and in cells carrying the human channel that isopetasin first activates TRPA1, an irritant-sensing channel on pain nerves, and then leaves those nerves desensitized, reducing CGRP release; repeated doses by mouth reduced pain behavior in mice and the widening of meningeal arteries triggered by irritants in rats. A 2022 narrative review by Borlak, Diener, and colleagues (Frontiers in Neurology) adds laboratory evidence that butterbur inhibits cyclooxygenase, lipoxygenase, and phospholipase A2 enzymes and desensitizes TRPA1 and TRPV1 channels, concluding that it works by blocking CGRP signaling. These are animal and cell studies, and how much of each constituent reaches human nerves at trial doses has not been established.

The toxicity of pyrrolizidine alkaloids is well understood. The European Medicines Agency explains that liver enzymes, chiefly CYP3A4 and CYP3A5, convert unsaturated PAs into highly reactive pyrroles that bind proteins and DNA and damage the cells lining the liver’s smallest blood vessels. The vessel walls thicken until blood flow is blocked, causing hepatic veno-occlusive disease (also called sinusoidal obstruction syndrome) and, when more stable metabolites reach the lungs, pulmonary arterial hypertension; several PAs cause cancer in animals, and their adducts can persist in tissues for months or years. Because CYP3A activity varies about 30-fold between people, susceptibility varies too. The liver injuries reported with the PA-free migraine extract looked different: Anderson and Borlak note that none was diagnosed as sinusoidal obstruction, and biopsies showed an inconsistent mix of necrosis, fatty change, inflammation, and cholestasis. In rats, six months of dosing caused bile duct overgrowth only at 45 to 90 times the maximum clinical dose. When Swissmedic acted in 2004, the mechanism and the responsible constituents were unknown, and the cases are still described as idiosyncratic.

Migraine prevention is the best-studied use. Lipton and colleagues (Neurology 2004) randomized 245 adults aged 18 to 65, who had two to six migraine attacks a month, to a standardized root extract at 75 mg twice daily, 50 mg twice daily, or placebo for four months. In the per-protocol analysis, attack frequency fell by 48% with the higher dose, 36% with the lower dose, and 26% with placebo; only the higher dose was significantly better than placebo. Sixty-eight percent of people on the higher dose at least halved their attacks, against 49% on placebo, and the higher dose also beat placebo on this measure at one, two, and three months. The most frequent side effects possibly related to treatment were mild digestive complaints, mainly burping. The result is clinically meaningful, but it comes from one product, and the 50 mg dose did not separate from placebo on the main measures.

The earlier trial was smaller and messier. Grossmann and Schmidramsl (International Journal of Clinical Pharmacology and Therapeutics 2000) gave 60 people two 25 mg capsules of Petadolex twice daily, or placebo, for 12 weeks and reported that attacks fell by up to 60% from baseline, significantly more than with placebo, with no adverse events. Because the original protocol and analysis had major shortcomings, Diener and colleagues (European Neurology 2004) reanalyzed the data independently: attacks fell from 3.4 to 1.8 a month with butterbur (33 people) and from 2.9 to 2.6 with placebo (27 people), and 45% versus 15% at least halved their attack frequency. The reanalysis concluded only that butterbur “may be effective.” Agosti and colleagues (Phytomedicine 2006), reviewing the two trials together, found moderate evidence for 150 mg a day, a higher dose than was then recommended, and said more rigorous studies of long-term effectiveness and safety were needed before it could be recommended.

Guideline panels in 2012 read this evidence generously. Holland and colleagues, writing for the American Academy of Neurology and American Headache Society (Neurology 2012), concluded that Petasites is effective and should be offered to reduce the frequency and severity of migraine attacks (Level A), above the probably effective rating (Level B) given to the feverfew extract MIG-99, magnesium, and riboflavin. The Canadian Headache Society (Pringsheim, Canadian Journal of Neurological Sciences 2012) listed butterbur among 11 preventives with a strong recommendation. On September 16, 2015, the AAN Board of Directors retired the 2012 guideline “due to serious safety concerns” with butterbur, stating that recommendations in retired guidelines are no longer valid or supported (quoted by Malone and Tsai, Journal of Family Practice 2018). The trials did not change; the judgment about whether they justified the liver risk did. Both positive adult trials tested one proprietary extract, so they do not vouch for other products: Avula and colleagues (Journal of Pharmaceutical and Biomedical Analysis 2012) found no detectable petasins in six of 21 butterbur supplements, and pyrrolizidine alkaloids in seven.

Children have been studied, but not enough to recommend butterbur for them. Pothmann and Danesch (Headache 2005) gave 108 children and adolescents aged 6 to 17 with migraine 50 to 150 mg of the root extract a day, depending on age, for four months in an open study with no comparison group; 77% reported at least halving their attack frequency, and 7.4% had side effects, mostly burping. In a small randomized trial, Oelkers-Ax and colleagues (European Journal of Pain 2008) assigned 58 primary school children to butterbur, music therapy, or placebo for 12 weeks, on top of headache education. Right after treatment only music therapy beat placebo; six months later both music therapy and butterbur did, and every group had improved substantially even during the baseline period. NCCIH notes that studies including children and adolescents found PA-free products seemed safe for up to 16 weeks, but the liver reports in adults and the EMA’s lower alkaloid allowance for smaller bodies argue for caution.

Hay fever is the second use, based on the leaf extract Ze 339. Schapowal (BMJ 2002) randomized 125 people with seasonal allergic rhinitis to Ze 339 four times daily or cetirizine for two weeks; quality-of-life and global scores improved similarly. In 186 people with intermittent allergic rhinitis, two doses of Ze 339 beat placebo, the higher dose more so (Archives of Otolaryngology–Head and Neck Surgery 2004), and in a 330-person trial Ze 339 three times daily and fexofenadine were each better than placebo and similar to each other (Phytotherapy Research 2005). But in a placebo-controlled crossover trial of 35 people in Scotland, Gray and colleagues (Annals of Allergy, Asthma and Immunology 2004) found that butterbur 50 mg twice daily had no effect on nasal airflow, symptoms, or quality of life. NCCIH says leaf extract studies suggest it may help; Malone and Tsai call the data not convincing. MSKCC notes a few studies in asthma, and NCCIH says no conclusions can be reached for conditions beyond migraine and hay fever.

The clinical evidence belongs to two standardized carbon dioxide extracts, not to the plant. For migraine, the root extract was tested at 50 mg twice daily, given as two 25 mg capsules, in the early trial, and at 75 mg twice daily, 150 mg a day, in the Lipton trial, the only dose that clearly beat placebo; children in the open study took 50 to 150 mg a day depending on age. The trials lasted three to four months, and NCCIH describes safety data on PA-free products for up to 16 weeks. In Anderson and Borlak’s case series, most liver reports involved doses between 50 and 225 mg a day, so staying within trial doses does not rule out the risk. For hay fever, Ze 339 tablets each contain 8 mg of total petasins (Schapowal 2005; MSKCC) and were taken three or four times a day for two weeks. Results from these extracts cannot be transferred to other brands, which differ in plant part, extraction method, petasin content, and alkaloid removal.

PA-free labeling is the minimum, not a guarantee. NCCIH says only butterbur products processed to remove pyrrolizidine alkaloids and labeled or certified PA-free should be considered. The European Medicines Agency’s limit for these alkaloids in herbal medicines is less than 1.0 µg a day for an adult, reduced by body weight for children, so that a 20 kg child’s limit is 0.5 µg. Raw rhizome can contain 5–90 ppm of PAs (Wildi), which means a single gram could hold up to 90 µg, many times that limit. That rules out homemade teas, tinctures, and root powders, along with the old decoctions in Grieve’s herbal. Supplement quality is also uneven: in Avula’s analysis, six of 21 butterbur supplements contained no measurable petasins and seven contained pyrrolizidine alkaloids. If you use butterbur at all, choose a product with certified PA testing and a stated petasin content, and avoid blends that state neither.

Practical points: butterbur is a preventive, taken daily and judged in trials after three to four months; it does nothing for an attack in progress. Talk to a doctor before starting, especially if you take other daily medicines or use painkillers often, and ask whether liver tests make sense. Malone and Tsai recommend liver function monitoring for anyone using butterbur, although Swissmedic warned in 2004 that because the liver injuries were unpredictable, regular testing could not guarantee safety. Stop at once and seek care for yellow skin or eyes, dark urine, pale stools, itching, nausea, or unusual tiredness. Do not put PA-containing butterbur preparations on broken skin, since NCCIH warns that the alkaloids could be absorbed through cuts or scrapes. A headache diary kept before and during use is the simplest way for you and your doctor to tell whether it is helping enough to justify the risk.