ButterburBotanical

• Migraine prevention in adultsLikely helps

  The pivotal trial randomized 245 adults: 75 mg twice daily cut migraine frequency 48% against 26% on placebo over 4 months, p=0.0012 [4].
  
  A 50 mg twice-daily arm managed only a non-significant 36% reduction (p=0.127), so the dose matters [4].
  
  An earlier 60-patient trial saw monthly attacks fall from 3.5 to 1.7 on butterbur, against 2.9 to 2.6 on placebo [3].
  
  In 2012 the American Academy of Neurology graded it Level A, 'established as effective', and the Canadian Headache Society still recommends it strongly [1].
  
  The AAN withdrew that endorsement in 2015-2016 - over liver safety, not over whether it works [1]. ^[4]

• Migraine in children and adolescentsPossibly helps

  An open-label study of 108 patients aged 6-17 reported 77% had at least a 50% drop in attack frequency after 4 months, with overall frequency down 63% [8].
  
  That study had no control group. A controlled trial of 58 patients comparing butterbur, music therapy, and placebo found butterbur's responder rate did not differ significantly from placebo [11].
  
  Music therapy outperformed both, and by 6 months there was no significant difference across all three arms - weight the controlled result more heavily [11]. ^[8]

• Allergic rhinitis (hay fever)Possibly helps

  Two sizeable randomized trials came out well: 125 patients found the leaf extract Ze 339 comparable to the antihistamine cetirizine, and 330 patients found butterbur comparable to fexofenadine [6].
  
  That second trial also beat placebo, and a third confirmed a dose-dependent effect [7].
  
  At least one smaller placebo-controlled trial, cited in an academic otolaryngology review, found no significant difference from placebo. Mostly positive, then, but not uniformly [7]. ^[6]

• AsthmaPossibly helps

  NIH's StatPearls describes only a limited body of RCT evidence for asthma, well short of what exists for migraine [2].
  
  The archived NCCAM government fact sheet is blunter: there is not enough evidence to establish efficacy and safety for asthma [9].
  
  One small trial found complementary anti-inflammatory activity when butterbur was added to inhaled corticosteroids [2]. ^[2]

• Alzheimer's diseasePossibly helps

  Preliminary in vitro and animal research points to neuroprotective and anti-inflammatory activity relevant to Alzheimer's [2].
  
  No human clinical trial has tested it, so this sits entirely in the preclinical column [2]. ^[2]

• PA-free root and rhizome extract, e.g. PetadolexMigraine prevention. This is the form behind essentially all the adult migraine trial data

  Elemental content

  Standardized to a minimum 15% petasin and isopetasin - 7.5 mg per 50 mg capsule

  Absorption

  Made by supercritical CO2 extraction, which is what strips the pyrrolizidine alkaloids out

  Best for

  Migraine prevention. This is the form behind essentially all the adult migraine trial data [4]

• PA-free leaf extract, Ze 339Allergic rhinitis, where it was compared head-to-head against cetirizine and fexofenadine

  Elemental content

  A CO2-extracted leaf-based tablet, a different preparation from the migraine root extract

  Absorption

  Dosed several times daily in trials rather than twice

  Best for

  Allergic rhinitis, where it was compared head-to-head against cetirizine and fexofenadine [6]

• Raw plant, teas, tinctures, powdersNothing. NCCIH and LiverTox both say only products processed and certified PA-free should be considered

  Elemental content

  Unprocessed, with pyrrolizidine alkaloids intact

  Absorption

  Nothing here has removed the hepatotoxic alkaloids

  Best for

  Nothing. NCCIH and LiverTox both say only products processed and certified PA-free should be considered [1]

• Topical preparationsNCCIH specifically warns that PA-containing topical products may not be safe

  Elemental content

  PA content varies by product and is often unstated

  Absorption

  Alkaloids can be absorbed through cuts or broken skin

  Best for

  NCCIH specifically warns that PA-containing topical products may not be safe [1]

• Common side effects

  Belching is the most common complaint, alongside stomach upset, diarrhea, headache, itchy eyes, drowsiness, and fatigue [1].
  
  In the head-to-head trial against cetirizine, butterbur produced markedly fewer sedating adverse events - two-thirds of the antihistamine group's adverse events were sedation-related [6]. ^[1]

• Pyrrolizidine alkaloids and the liver

  This is the highest-consequence fact on the page. Unrefined butterbur contains pyrrolizidine alkaloids, plant toxins the liver converts into reactive compounds that damage liver cells [15].
  
  The classic mechanism is veno-occlusive disease, in which small liver blood vessels become blocked [15].
  
  A WHO Vigibase analysis covering 1992-2006 linked P. hybridus extracts to 40 hepatotoxicity cases, including 9 acute hepatitis cases and 2 liver failures or transplants [2].
  
  Estimated incidence of reversible cholestatic hepatitis with long-term use is around 1 in 175,000 [2]. ^[2]

• The Petadolex counter-analysis

  A 2018 reanalysis of the Petadolex brand specifically, using the formal RUCAM causality method, found no probable relationship between the product and the serious liver-injury reports [5].
  
  Only 2 non-serious, reversible liver-enzyme elevations were rated probably related [5].
  
  It was manufacturer-funded but methodologically formal, and the wider signal may reflect poorly purified competitor products. Two datasets, two stories, not resolved here [5]. ^[5]

• 'PA-free' is a label claim, not a verified fact

  Testing cited by NIH's StatPearls found detectable toxic alkaloids in 7 of 21 commercial butterbur products [2].
  
  The FDA does not independently verify PA-free labelling, so the phrase describes a manufacturing intent rather than a checked result [1]. ^[2]

• Which plant part, and which species

  NIH's StatPearls describes the migraine extract as coming from Petasites hybridus leaves [2].
  
  The primary migraine RCTs, LiverTox, and a 2022 Frontiers review all describe a root and rhizome extract instead - the weight of the primary literature runs against StatPearls here [4].
  
  Separately, a product labelled simply butterbur may be P. japonicus or P. frigidus, neither of which is the species that was studied [11]. ^[4]

• Alcohol and existing liver disease

  WebMD states there are no known food or drink interactions, with one exception: alcohol may raise the risk of liver damage alongside butterbur [13].
  
  Existing liver disease is a reason to avoid butterbur outright, since it might make that disease worse [13]. ^[13]

• Ragweed and daisy-family allergy

  Butterbur belongs to the Asteraceae (Compositae) family, alongside ragweed, chrysanthemums, marigolds, and daisies [1].
  
  NCCIH, StatPearls, and WebMD all warn that people allergic to those plants may react to butterbur too [2]. ^[2]

• Pregnancy and breastfeeding

  NCCIH says PA-containing butterbur should not be used in pregnancy or while breastfeeding, since the alkaloids may cause birth defects or liver damage [1].
  
  For PA-free products little is known either way, and Drugs.com advises avoiding it entirely, citing the carcinogenic and mutagenic potential of any residual alkaloids [11]. ^[1]

• Children

  PA-free products have been studied in children aged 6-17 for up to 16 weeks and were reasonably well tolerated in that migraine research [13].
  
  StatPearls notes safety criteria have not been established under age 6, and the NCCAM fact sheet says any use in children should be supervised by a qualified practitioner [9]. ^[13]

• Anticholinergic and cardiovascular medications

  NIH's StatPearls lists butterbur as contraindicated with anticholinergic medications [2].
  
  Drugs.com's professional monograph flags theoretical risk with calcium channel antagonists and inotropic agents, while stating no interaction is well documented clinically [11]. ^[2]

• Liver-enzyme inducers

  CYP3A4 inducers - St John's wort, carbamazepine, phenytoin, rifampin - carry a theoretical risk [11].
  
  RxList adds phenobarbital and rifabutin to that list [12].
  
  The concern is more toxic alkaloid metabolite production by the liver, not a change in butterbur's own effect [12]. ^[11]

Petasites hybridus is not a food. It has essentially no USDA- or EFSA-style nutrient composition data, and the raw plant's pyrrolizidine alkaloids make it unsafe to eat without specialized processing [14].

An academic review describes it as gaining popularity as a food supplement - meaning supplement products, not culinary use [14].

The confusion comes from a cousin. Petasites japonicus, Japanese butterbur or fuki, has stalks that are commonly grown and eaten as a vegetable in Japan [11].

That is a different species from the one in the migraine and allergy trials, so 'butterbur is eaten in Japan' does not make the supplement plant a food [11].

No study has looked at time of day for butterbur, and no source reviewed specifies taking it with or without food [3].

What the trials show instead is a split-dose pattern. Migraine prevention used 75 mg twice daily, morning and evening [4].

Allergic rhinitis trials went further, dosing three to four times daily [6].

NIH's LiverTox describes the general regimen the same way: 100-150 mg/day in 2-3 divided doses [3].