The effects of propofol on cerebral blood flow, intracranial pressure (ICP) and cerebral oxygen consumption (CMRO 2) were assessed in ten severely head-injured patients undergoing surgery for limb fractures. The patients, aged between 15 and 40 years, were in deep coma, scored 6–7 on the Glasgow coma score. They were mechanically ventilated and sedated with 1 mg · h −1 phenoperidine. Anaesthesia was carried out with a 2 mg · kg −1 intravenous bolus of propofol, immediately followed by a 150 μg · kg −1 · min −1 infusion, which lasted for a mean time of 41.4 ± 7.3 min. Data were collected 5 min before any propofol was given, 15 min after the start of the infusion, and 15 min after its end. A radial artery cannula, a 7.5 Fr thermodilution flow-directed pulmonary arterial catheter, a cerebral intraventricular catheter and a catheter in the jugular venous bulb were used for this purpose. Carotid arterial injection of 133 Xenon was used to determine regional cerebral blood flow (rCBF). Anaesthetic blood concentrations of propofol (3 to 5 μg · ml −1) were associated with a decrease in all the parameters studied : cerebral perfusion pressure, from 82 ± 14 mmHg to 59 ± 7 mmHg (p < 0.001); rCBF, from 35 ± 6 ml · 100 g −1 · min −1 to 26 ± 5 ml · 100 g −1 · min −1 (p < 0.01) ; ICP from 11.3 ± 2.6 mmHg to 9.2 ± 2.5 mmHg (p < 0.001) ; CMRO 2 from 1.63 ± 0.38 mlO 2 ± 100 g −1 · min −1 to 1.18 ± 0.38 mlO 2 · 100 g −1 · min −1 (p < 0.01). Flow-metabolism coupling was maintained (rCBF = 17.09 CMRO 2 + 6.62 ; r = 0.88 ; p < 0.001). Cerebral vascular resistances and the gradient in oxygen content between carotid arterial blood and jugular venous blood remained unchanged, as well as the jugular blood lactate levels. After propofol administration had been discontinued and blood propofol concentrations had theoretically returned to a level corresponding to recovery from anaesthesia, rCBF, ICP, and CMRO 2 returned to pre-induction values without any rebound.