The effects of the opioids alfentanil (A), fentanyl (F), and sufentanil (S) on cerebral blood flow (CBF) and intracranial pressure (ICP) have been discussed in several recent publications. The purpose of this review is to describe the results of studies in animals, healthy volunteers, and patients with and without intracranial diseases. Clinical relevance and mechanisms of the reported ICP and CBF increases are analysed. METHODS. Approximately 70 original articles and abstracts were retrieved by a systematic literature search using the key word list at the end of this abstract. The cited studies came from computerised database systems like Silver Platter and DIMDI, the SNACC reference list, and the bibliographies of pertinent articles and books. These studies were classified into three groups: significant increase of ICP and/or CBF; no significant or clinically relevant alterations; and significant decreases of ICP and/or CBF. RESULTS. The numerical relationship was 6:7:3 for A, 7:16:9 for F, and 5:11:8 for S. Increases of previously normal or only slightly elevated ICP were registered in some studies in connection with a decrease in mean arterial pressure (MAP). On the other hand, in patients with brain injury and elevated ICP opioids did not further increase ICP despite MAP decreases. In studies monitoring ICP and/or CBF continuously, transient and moderate increases of questionable clinical relevance became apparent a few minutes after bolus injection of opioids. Alterations of systemic and cerebral haemodynamics observed after bolus application were not registered during continuous infusion of A and S. DISCUSSION AND CONCLUSIONS. The cerebral effects of opioids are dependent on several factors, e.g., age, species, ventilation, anaesthesia before and during measurements, systemic haemodynamics, and underlying diseases. The probable mechanism of ICP increase during decreasing MAP is cerebral vasodilatation due to maintained autoregulation. With increasing severity of the cerebral lesion autoregulation is often disturbed. Therefore, ICP often remains unaltered despite MAP decreases. However, the resulting decrease in cerebral perfusion pressure makes such patients more susceptible to develop ischaemic neurological deficits. Induction of somatic rigidity or (with high doses) convulsions, exceeding the upper limit of autoregulation, histamine release, cerebral vasodilatation, increased cerebral oxygen consumption, or carbon dioxide accumulation during spontaneous breathing were discussed as mechanisms for transient ICP/CBF increases. It is concluded that opioids are often beneficial and not generally contraindicated for patients with cerebral diseases and compromised intracranial compliance. However, since negative side effects cannot be excluded, opioid effects and side effects should be monitored (MAP, ICP, cerebrovenous oxygen saturation, transcranial Doppler sonography) in patients at risk. It has to be stressed that opioids should be administered only to patients with stable haemodynamic situations and preferably in well-titrated, continuous infusions.