BACKGROUND:Migraine and cluster headaches are severe and disabling. Migraine affects up to 18% of women, while cluster headaches are much less common (0.2% of the population). A number of acute and prophylactic therapies are available. Hyperbaric oxygen therapy (HBOT) is the therapeutic administration of 100% oxygen at environmental pressures greater than one atmosphere, while normobaric oxygen therapy (NBOT) is oxygen administered at one atmosphere. OBJECTIVES:To assess the safety and effectiveness of HBOT and NBOT for treating and preventing migraine and cluster headaches. SEARCH STRATEGY:We searched the following in May 2008: CENTRAL, MEDLINE, EMBASE, CINAHL, DORCTIHM and reference lists from relevant articles. Relevant journals were hand searched and researchers contacted. SELECTION CRITERIA:Randomised trials comparing HBOT or NBOT with one another, other active therapies, placebo (sham) interventions or no treatment in patients with migraine or cluster headache. DATA COLLECTION AND ANALYSIS:Three reviewers independently evaluated study quality and extracted data. MAIN RESULTS:Nine small trials involving 201 participants were included. Five trials compared HBOT versus sham therapy for acute migraine, two compared HBOT to sham therapy for cluster headache and two evaluated NBOT for cluster headache. Pooling of data from three trials suggested that HBOT was effective in relieving migraine headaches compared to sham therapy (relative risk (RR) 5.97, 95% confidence interval (CI) 1.46 to 24.38, P = 0.01). There was no evidence that HBOT could prevent migraine episodes, reduce the incidence of nausea and vomiting or reduce the requirement for rescue medication. There was a trend to better outcome in a single trial evaluating HBOT for the termination of cluster headache (RR 11.38, 95% CI 0.77 to 167.85, P = 0.08), but this trial had low power.NBOT was effective in terminating cluster headache compared to sham in a single small study (RR 7.88, 95% CI 1.13 to 54.66, P = 0.04), but not superior to ergotamine administration in another small trial (RR 1.17, 95% CI 0.94 to 1.46, P = 0.16). Seventy-six per cent of patients responded to NBOT in these two trials. No serious adverse effects of HBOT or NBOT were reported. AUTHORS' CONCLUSIONS:There was some evidence that HBOT was effective for the termination of acute migraine in an unselected population, and weak evidence that NBOT was similarly effective in cluster headache. Given the cost and poor availability of HBOT, more research should be done on patients unresponsive to standard therapy. NBOT is cheap, safe and easy to apply, so will probably continue to be used despite the limited evidence in this review.