Endotoxin-induced sepsis in pigs is a recognized experimental model for the study of human septic shock. Generally, pigs are brought into general anaesthesia before sepsis is induced. It is our experience that drug dosages of propofol and fentanyl need to be reduced during endotoxin-induced sepsis, in order to prevent respiratory and cardiovascular depression, but the scientific evidence for this observation is lacking. Therefore, we measured the consumption of propofol and fentanyl at equal level of anaesthesia in pigs with (n = 5) and without (n = 5) endotoxin-induced sepsis, using the cerebral state index (CSI) as measure of anaesthetic depth. Infusion rates of propofol (P < 0.01) and fentanyl (P < 0.05) were significantly lower in septic pigs. Pigs with endotoxin-induced sepsis had an infusion rate of 2.2 mg/kg/hr (S.D. 0.5) for propofol and 12 microg/kg/hr (S.D. 2) for fentanyl, whereas healthy pigs had infusion rates of 3.5 mg/kg/hr (S.D. 0.6) and 17 microg/kg/hr (S.D. 4), respectively. CSI was equal in both groups throughout the experiment, and had a lowest average value of 47 (S.D. 10) at t = 30 in healthy pigs and reached a highest average value of 67 (S.D. 19) at t = 240 in pigs with endotoxin-induced sepsis. Anaesthetic depth was sufficient, assessed clinically, throughout the experiment in both groups. We concluded that the consumption of propofol and fentanyl was significantly reduced in pigs with endotoxin-induced sepsis. In the present study, we adjusted the level of anaesthesia according to clinical signs, and found good agreement with CSI.