During the last decade several studies of cerebral blood flow (CBF) and metabolism in the acute phase of head injury have been published. It is the aim of this review to describe the dynamic changes in CBF, cerebral metabolic rate of oxygen (CMRO2), cerebral autoregulation (CA), and reactivity to PaCO2 and barbiturate (metabolic reactivity) in the acute phase after severe head injury and to discuss the therapeutical consequences with reference to prolonged artificial hyperventilation, hypothermia, barbiturate sedation, and mannitol therapy. On the basis of present knowledge concerning cerebral circulation and its regulation, the author reviews the literature concerning methodology for experimental and clinical CBF measurements and regulation of CBF and cerebral oxygen uptake. Emphasis is placed on studies of the effect of body temperature (hypothermia) as a therapeutic tool in the control of cerebral metabolism, blood flow, and intracranial pressure. Although hypothermia significantly reduces cerebral metabolism and blood flow, the effect of hypothermia on cerebral blood flow, metabolism, ICP, and outcome after acute head injury has never been investigated in clinically controlled studies. Experimental and clinical studies concerning sensitivity of CBF for changes in PaCO2 are reviewed. The normal CO2 reactivity defined as absolute (delta CBF/delta PaCO2) and relative (% change CBF/delta PaCO2) or delta in CBF/PaCO2 mm Hg are mentioned. In awake normocapnic man the relative CO2 reactivity averages 4%/mm Hg and the absolute CO2 reactivity 2ml/mm Hg. Uncontrolled prospective studies show a therapeutic effect of artificially prolonged hyperventilation on outcome. Only one preliminary controlled study indicates that the outcome is poorer and recovery prolonged. Nevertheless, in the acute phase of HI, artificial hyperventilation is used routinely for control of intracranial hypertension and during the intensive care management of the patients. The steal and inverse steal phenomena are reviewed. Although of considerable theoretical interest these phenomena are without clinical significance in patients with head injury, unless clinical CBF measurements are performed. The frequency of the inverse steal phenomenon in studies of rCBF with a 16-channel Cerebrograph (intraarterial approach) is found to be about 10%. During prolonged hyperventilation experimental studies and clinical studies of apoplexy show an adaptation of CBF and CSF-pH and bicarbonate.(ABSTRACT TRUNCATED AT 400 WORDS)