Objective: To evaluate the effect of a single large dose of mannitol on midline tissue shifts after a large cerebral infarction. Background: Theoretically, mannitol use in the largest cerebral infarctions may preferentially shrink noninfarcted cerebral tissue, thereby aggravating midline tissue shifts and worsening neurologic status. To test this theory, we studied patients with hemispheric infarctions using continuous and sequential MRI during administration of a single dose of mannitol Methods: Patients with neurologic deterioration from complete middle cerebral artery (MCA) infarctions and CT evidence of at least 3 mm of midline shift were studied using T1-weighted three-dimensional multiplanar rapid acquisition gradient echo image data sets acquired at' 5- to 10-minute intervals before, during, and after a 1.5 gm/kg bolus infusion of mannitol. Horizontal and vertical displacements were calculated by previously described methods. Glasgow Coma Scale (GCS) and MCA Stroke Scale (MCASS) were measured before and after mannitol administration. Mean changes in tissue shifts were compared using repeated measures analysis of variance. Clinical variables were compared using paired t-tests. Results: Seven patients were enrolled. The final average change in midline shift compared with the initial displacement was 0.0 ± 1 mm for horizontal (F = 0.06, p = 0.99) and 0.25 ± 1.3 mm for vertical displacement (F = 0.06, p = 0.99). Whereas average scores for the group did not change, MCASS improved in two, GCS improved in three, and pupillary light reactivity returned in two patients. No patient worsened. Conclusions: Acute mannitol used in patients with cerebral edema after a large hemispheric infarction does not alter midline tissue shifts or worsen neurologic status.