Investigations in dogs have shown substantial dexmedetomidine (Dex)-induced reductions in cerebral blood flow (CBF) unaccompanied by reductions in cerebral metabolic rate (CMR). If this effect were to occur in humans in areas of injured brain in which CBF is already low, oxygen delivery might be critically impaired. The institutional use of brain PO2 monitoring during neurovascular surgery and the use of Dex as a component of the anesthetic allowed insight into this issue. Data from 5 neurovascular surgery patients, 2 for excision of arteriovenous malformations (AVMs), and 3 for intracranial aneurysm clipping were reviewed retrospectively. All had acute, lesion-related neurologic deficits. During general anesthesia with sufentanil and sevoflurane, with or without N2O, a parenchymal brain tissue PO2 (PbrO2) electrode was placed directly in the territory at risk from the pending neurosurgical intervention. After a stable PbrO2 value was achieved, Dex was administered by bolus (1 μg/kg over 10 min) and infusion (0.5 to 0.7 μg/kg/min). Mean arterial pressure (MAP), heart rate (HR), and PbrO2 were observed continuously for at least 25 minutes. Baseline PbrO2 values were relatively low (≤16 mm Hg) in 4 of the 5 patients, a pattern consistent with antecedent neurologic insult. In the 15 minutes after initiation of Dex administration, the pattern was one of a modest increase in Pbr02 (maximum 11.1%; P=0.0147) occurring roughly in parallel with a modest increase in MAP [maximum 3.5 mm Hg (4.5%); P=0.041]. HR did not change. Clinically significant reduction of PbrO2 did not occur before neurosurgical interventions. These observations provide no support for a direct cerebral vasoconstrictive effect of Dex in humans that is independent of any vasoconstriction that may occur as a consequence of Dex-induced reduction in CMR. At a minimum, any such effect was insufficient to have an adverse effect on oxygen delivery to brain parenchyma.