Cerebral edema after ischemic stroke is frequently treated with mannitol and hypertonic saline (HS); however, their relative cerebrovascular and metabolic effects are incompletely understood, and may operate independent of their ability to lower intracranial pressure. We compared the effects of 20% mannitol and 23.4% saline on cerebral blood flow (CBF), blood volume (CBV), oxygen extraction fraction (OEF), and oxygen metabolism (CMRO(2)), in nine ischemic stroke patients who deteriorated and had >2 mm midline shift on imaging. (15)O-PET was performed before and 1 h after administration of randomly assigned equi-osmolar doses of mannitol (1.0 g/kg) or 23.4% saline (0.686 mL/kg). Baseline CBF values (ml/100g/min) in the infarct core, periinfarct region, remaining ipsilateral hemisphere, and contralateral hemisphere in the mannitol group were 5.0 ± 3.9, 25.6 ± 4.4, 35.6 ± 8.6, and 45.5 ± 2.2, respectively, and in the HS group were 8.3 ± 9.8, 35.3 ± 10.9, 38.2 ± 15.1, and 35.2 ± 12.4, respectively. There was a trend for CBF to rise in the contralateral hemisphere after mannitol from 45.5 ± 12.2 to 57.6 ± 21.7, P = 0.098, but not HS. CBV, OEF, and CMRO(2) did not change after administration of either agent. Change in CBF in the contralateral hemisphere after osmotic therapy was strongly correlated with baseline blood pressure (R (2)= 0.879, P = 0.002). We conclude that at higher perfusion pressures, osmotic agents may raise CBF in non-ischemic tissue. We conclude that at higher perfusion pressures, osmotic agents may raise CBF in non-ischemic tissue.