BACKGROUND: Mannitol is sometimes dramatically effective in reversing acute brain swelling, but its effectiveness in the on-going management of severe head injury remains open to question. There is evidence that, in prolonged dosage, mannitol may pass from the blood into the brain, where it might cause reverse osmotic shifts that increase intracranial pressure. OBJECTIVES: To assess the effects of different mannitol therapy regimens, of mannitol compared to other intracranial pressure (ICP) lowering agents, and to quantify the effectiveness of mannitol administration given at other stages following acute traumatic brain injury. SEARCH STRATEGY: The review drew on the search strategy for the Injuries Group as a whole. We checked reference lists of trials and review articles, and contacted authors of trials. SELECTION CRITERIA: Randomised trials of mannitol, in patients with acute traumatic brain injury of any severity. The comparison group could be placebo-controlled, no drug, different dose, or different drug. Trials where the intervention was started more than eight weeks after injury, and cross-over trials were excluded. DATA COLLECTION AND ANALYSIS: The reviewers independently rated quality of allocation concealment and extracted the data. Relative risks (RR) and 95% confidence intervals (CI) were calculated for each trial on an intention to treat basis. MAIN RESULTS: Overall there were few eligible trials. In the pre-operative management of patients with acute intracranial haemorrhage the administration of high-dose mannitol resulted in reduced mortality (RR=0.55; 95%CI 0.36, 0.84) and reduced death and severe disability (RR=0.58; 95%CI 0.45, 0.74) when compared with conventional-dose mannitol. One trial compared ICP-directed therapy to 'standard care' (RR for death= 0.83; 95%CI 0.47,1.46). One trial compared mannitol to pentobarbital (RR for death = 0.85; 95% CI 0.52, 1.38). No trials compared mannitol to other ICP-lowering agents. One trial tested the effectiveness of pre-hospital administration of mannitol against placebo (RR for death=1.75; 95% CI 0.48, 6.38). REVIEWER'S CONCLUSIONS: High-dose mannitol appears to be preferable to conventional-dose mannitol in the pre-operative management of patients with acute intracranial haematomas. However, there is little evidence about the use of mannitol as a continuous infusion in patients with raised intracranial pressure in patients who do not have an operable intracranial haematoma. Mannitol therapy for raised ICP may have a beneficial effect on mortality when compared to pentobarbital treatment. ICP-directed treatment shows a small beneficial effect compared to treatment directed by neurological signs and physiological indicators. There are insufficient data on the effectiveness of pre-hospital administration of mannitol to preclude either a harmful or a beneficial effect on mortality.