BACKGROUND: Skin incision has been used as a standard stimulus in most concentration versus response relationship studies for anesthetics. However, skin incision is not the most intense stimulation and inconvenient method during operation. Mean arterial blood pressure, but not heart rate, is convenient and predicts surgical stress as well as propofol blood concentration. We evaluated the effects of different fentanyl concentration on propofol-fentanyl-N20 anesthesia using mean arterial blood pressure as an indicator of surgical stress during operation. METHODS: Eighty ASA I or II patients (age: 20~55 yrs) scheduled for spine fusion were randomly allocated to four groups according to expected fentanyl blood concentration (Group 1, 2, 3, 4: 0, 1.5, 3.0, 4.5 ng/ml respectively, n=20 for each group). Fentanyl was infused according to isoconcentration nomogram, and propofol infusion rate was titrated by changes of mean arterial blood pressure (0~12 mg/kg/hr). Fifteen minutes before expected end of surgery, propofol and fentanyl infusion were discontinued. Thereafter IV-PCA using fentanyl was applied for postoperative pain relief. Average propofol flow rate, recovery of orientation, verbal rating scale were cheked. RESULTS: Group 2, 3, 4 showed decreased average propofol flow rate, delayed recovery and decreased postoperative 24 hr fentanyl requirement for pain relief gradually compared with group 1. Group 4 showed ceiling effect in terms of average propofol flow rate, recovery of orientation and 24 hr fentanyl requirement for postoprerative pain relief compared with group 1~3. CONCLUSIONS: Keep the fentanyl concentration below 3.0~4.5 ng/ml and titrate propofol flow rate was reasonable method for adequate control of drug infusion during a propofol-fentanyl-N20 anesthesia.