Feverfew and migraine
What 3 graded rows on feverfew agree on for migraine, and where they differ. Counted side by side, never averaged.
Consensus finder Rows from the Feverfew page, reviewed Oct 21, 2026 How it decides
What the rows agree on
Each block and row is one graded row from the Feverfew page, with its PubMed record.
Your question
Feverfew and migraine
3 rows · 878 peopleSplit. 1 row found an improvement, 1 found no difference and 1 had mixed results.
- CGrade C× 3
- Rows
- 3
- Trials
- 8
- People
- 878
- Pooled reviews may count some twice
Where they differ
From the rowsThe best-graded row, grade C, is a systematic review of 561 people (2015) and reported mixed results.
The row that found an improvement ran longer (4 months) than the row that found no difference (3 months).
The row that found an improvement is newer (2005) than the row that found no difference (2002).
1 row with mixed results is left out of this comparison.
The rows
1 improved · 1 no difference · 1 mixed- CGrade CMigraine attacks per month · Improved
RCT · n 170 · 2005 · Adults with migraine (International Headache Society criteria)
6.25 mg three times daily · 16 weeks
Attacks fell by 1.9 a month versus 1.3 on placebo (P=0.046); responders were more likely on feverfew (odds ratio 3.4), and possibly related adverse events were similar (8.4% vs 10.2%).
- CGrade CMigraine attacks per month (dose-finding trial) · No difference
RCT · n 147 · 2002 · Adults with migraine with or without aura (International Headache Society criteria)
2.08, 6.25 or 18.75 mg three times daily · 12 weeks
No significant difference from placebo in the whole group and no dose-response. In a predefined subgroup of 49 people with at least 4 attacks a month, attacks fell by 1.8 per 28 days on 6.25 mg three times daily vs 0.3 on placebo. 35% of all patients reported at least one adverse event, at similar rates on feverfew and placebo.
- CGrade CMigraine frequency · Mixed
Systematic review, 6 studies · n 561 · 2015 · People with migraine in 6 double-blind trials
Varied across trials · Varied across trials
Results could not be pooled; small trials were positive, two rigorous trials found no difference, and the largest trial found 0.6 fewer attacks a month. Evidence was rated low quality.