This study was designed to determine if fentanyl altered MAC-awake (the end-tidal concentration of sevoflurane associated with eye opening to verbal command) in 30 healthy, ASA I patients. During anaesthesia, no other anaesthetics or drugs were given with the exception of sevoflurane. After surgery, end-tidal anaesthetic concentration was maintained constant for at least 15 min. If patients failed to respond to command, the end-tidal concentration was decreased and again maintained constant for 15 min. The anaesthetic concentration midway between the value permitting the response and that just preventing the response was recorded as MAC-awake. Fentanyl was administered at predicted plasma concentrations of 1 and 2 ng mg-1 using a computer-controlled continuous infusion and plasma concentrations of fentanyl were measured at the time of MAC-awake measurements. MAC-awake of the control group in which fentanyl was not administered was mean 0.67 (SD 0.12)% or 0.36 (0.03) MAC, being significantly higher than that of the fentanyl 2-ng ml-1 group (0.57 (0.09)% or 0.30 (0.04) MAC). In the fentanyl 1-ng ml-1 group, MAC-awake (0.65 (0.10)% or 0.34 (0.05) MAC) did not differ from that in the control group. Logistic regression analysis showed that increasing plasma concentration of fentanyl and increasing age significantly reduced the MAC-awake of sevoflurane. Because the reduction was very small relative to the overall scatter of the MAC-awake, a low plasma concentration of fentanyl did not significantly reduce the MAC-awake of sevoflurane.