ObjectivesTo assess the effects of botulinum toxin for prevention of migraine in adults.DesignSystematic review and meta-analysis.Data sourcesCENTRAL, MEDLINE, Embase and trial registries.Eligibility criteriaWe included randomised controlled trials (RCTs) of botulinum toxin compared with placebo, active treatment or clinically relevant different dose for adults with chronic or episodic migraine, with or without the additional diagnosis of medication overuse headache.Data extraction and synthesisCochrane methods were used to review double-blind RCTs. Twelve week post-treatment time-point data was analysed.ResultsTwenty-eight trials (n=4190) were included. Trial quality was mixed. Botulinum toxin treatment resulted in reduced frequency of −2.0 migraine days/month (95% CI −2.8 to −1.1, n=1384) in chronic migraineurs compared with placebo. An improvement was seen in migraine severity, measured on a numerical rating scale 0 to 10 with 10 being maximal pain, of −2.70 cm (95% CI −3.31 to −2.09, n=75) and −4.9 cm (95% CI −6.56 to −3.24, n=32) for chronic and episodic migraine respectively. Botulinum toxin had a relative risk of treatment related adverse events twice that of placebo, but a reduced risk compared with active comparators (relative risk 0.76, 95% CI 0.59 to 0.98) and a low withdrawal rate (3%). Although individual trials reported non-inferiority to oral treatments, insufficient data were available for meta-analysis of effectiveness outcomes.ConclusionsIn chronic migraine, botulinum toxin reduces migraine frequency by 2 days/month and has a favourable safety profile. Inclusion of medication overuse headache does not preclude its effectiveness. Evidence to support or refute efficacy in episodic migraine was not identified.