Butterbur (Petadolex) for migraine prevention and the pyrrolizidine question

Butterbur (Petasites hybridus) is one of the few botanicals that ever earned a Level A recommendation from the American Academy of Neurology for migraine prevention — and then lost it. The story illustrates how clinical efficacy in well-conducted trials can collide with toxicology that the supplement market is structurally unprepared to manage.

The evidence base that earned the recommendation

Two double-blind RCTs in adults established the efficacy signal. The 2000 trial by Grossmann and Schmidramsl randomised 60 patients to 50 mg butterbur extract twice daily and reported a 46 percent reduction in attack frequency at 12 weeks (PMID: 11075019).¹ The pivotal 2004 trial by Lipton and colleagues randomised 245 patients to 75 mg twice daily, 50 mg twice daily, or placebo and showed a dose-dependent decrease in migraine frequency, with the 75 mg arm achieving a 48 percent reduction versus 26 percent on placebo (PMID: 15623680).² A pediatric/adolescent open-label study by Pothmann reported a 63 percent responder rate (PMID: 16236023).³

The pyrrolizidine alkaloid problem

Petasites hybridus naturally contains pyrrolizidine alkaloids (PAs) — hepatotoxic and potentially carcinogenic plant compounds. The trials used a CO2-extracted preparation marketed as Petadolex in which PAs are below 0.08 ppm. Crude butterbur powder or unfractionated extracts can contain much higher PA levels, and PA toxicity has been documented in hepatic veno-occlusive disease cohorts (PMID: 21111003).⁴ Even within the purified-extract pathway, post-marketing cases of hepatotoxicity emerged. The 2012 European spontaneous-reporting analysis identified at least 40 cases of liver injury associated with butterbur products, several of which involved purified extracts (PMID: 22585594).⁵

Why the AAN withdrew its recommendation

In 2012, German regulator BfArM revoked the marketing authorization of Petadolex over hepatotoxicity signals. In 2015, the American Academy of Neurology updated its evidence-based migraine prevention guideline, stating that butterbur could no longer be recommended owing to safety concerns even though the efficacy data remained Level A on technical grounds (PMID: 26464511).⁶ The Canadian Headache Society took a similar position. The Cochrane review on feverfew and other herbal preventives notes that butterbur efficacy data are higher quality than most herbal headache evidence but flagged the same safety issue (PMID: 25892430).⁷

What the 2024-2025 picture looks like

Several reformulated products remain on the US dietary-supplement market, where labeling cannot make a migraine claim. A 2023 European Medicines Agency assessment concluded that current PA-free analytic methods could in principle support a future regulated product, but no manufacturer has so far reopened the German authorization (EMA herbal monograph reassessment). A 2024 narrative review in Current Pain and Headache Reports noted that with CGRP antagonists now available, the risk-benefit calculation for butterbur has shifted decisively against it for most patients (PMID: 38381258).⁸

If a patient is still considering butterbur

The only defensible use pattern is a clearly labeled PA-free, CO2-extracted product with batch-level certificates of analysis and baseline plus periodic liver-function monitoring under clinician supervision. Pregnant patients should not use butterbur in any form because the PA exposure cannot be guaranteed and trace PAs are particularly concerning to the developing liver. The 2018 update to the German neurology society guideline still lists butterbur as effective but contraindicated in routine prevention pending regulatory clarification.