A Cochrane review [Abstract] 1 included 3 studies with a total of 226 subjects. One trial included patients with presumed ischaemic stroke without computerised tomography (CT) verification, and the other two trials included patients with CT-verified non-traumatic intracerebral haemorrhage (ICH). Data on the primary outcome measure (death and dependency) were not available in any of the trials. Death and disability could be calculated in the larger ICH trial without differences between the mannitol and control groups (Peto OR 1.28, 95% CI 0.64 to 2.56; 1 study, n=128). Case fatality was not reported in the trial of ischaemic stroke. Case fatality did not differ between the mannitol and control groups in the ICH trials (Peto OR 1.03, 95% CI 0.47 to 2.25; 2 studies, n=149). Adverse events were either not found or not reported. The change in clinical condition was reported in two trials, and the proportion of those with worsening (Peto OR 1.20, 95% CI 0.51 to 2.82; 2 studies, n=98) or not improving (Peto OR 1.33, 95% CI 0.58 to 3.03; 2 studies, n=98) condition did not differ significantly between mannitol-treated patients and controls. Based on these three trials neither beneficial nor harmful effects of mannitol could be proved.
Comment: The quality of evidence is downgraded by study quality (unclear allocation concealment and more than 20% loss to follow up in some studies), by indirectness (all three trials used mannitol in a dose below which it does not always reduce intracranial pressure) and by imprecise results (limited study size for each comparison).
Primary/Secondary Keywords