BACKGROUND: Most cases of stroke are caused by impairment of blood flow to the brain (ischaemia) which results in a reduction in oxygen available and subsequent cell death. It has been postulated that hyperbaric oxygen therapy (HBOT) may reduce the volume of brain that will die by greatly increasing the oxygen available, and it may further improve outcome by reducing brain swelling. Some centres are using HBOT routinely to treat stroke. OBJECTIVES: To assess the effectiveness and safety of adjunctive HBOT in the treatment of acute ischaemic stroke. SEARCH STRATEGY: We searched the Cochrane Stroke Group Trials Register (last searched 9 January 2004), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, Issue 3, 2004), MEDLINE (1966 to July 2004), EMBASE (1980 to July 2004), CINAHL (1982 to July 2004), and DORCTHIM (Database of Randomised Controlled Trials in Hyperbaric Medicine) (from inception to 2004). We handsearched journals and conference proceedings, searched reference lists of articles, and contacted researchers in an effort to identify additional published and unpublished studies. SELECTION CRITERIA: We included all randomised controlled trials that compared the effect of adjunctive HBOT with no HBOT (no treatment or sham). DATA COLLECTION AND ANALYSIS: Two authors used standardised forms to extract the data independently. Each trial was assessed for internal validity with differences resolved by discussion. Data were extracted and entered into RevMan 4.2. MAIN RESULTS: Three randomised controlled trials (106 participants) satisfied the inclusion criteria. The methodological quality of the trials varied but was generally high. Data could be pooled for a limited number of clinically important outcomes. There were no significant differences in mortality rate at six months in those receiving HBOT compared to the control group (relative risk 0.61, 95% confidence interval (CI) 0.17 to 2.2, P value 0.45). Two of 15 scale measures of disability and functional indicated an improvement following HBOT, both at one year follow up: the mean Trouillas Disability Scale was lower with HBOT (mean difference (MD) 2.2 points reduction with HBOT, 95% CI 0.15 to 4.3, P value 0.04) and the mean Orgogozo Scale was higher (MD 27.9 points, 95% CI 4.0 to 51.8, P value 0.02). These improvements were not reflected in other trials or functional scales. AUTHORS' CONCLUSIONS: This systematic review has not found evidence to show that HBOT improves clinical outcomes when applied during the acute presentation of ischaemic stroke. While evidence from the three randomised controlled trials is insufficient to provide clear guidelines for practice, clinical benefit does not seem likely. Further research is required to better define the role of HBOT in this condition.