Target-controlled infusion (TCI) incorporates the pharmacokinetic variables of an IV drug to facilitate safe and reliable administration. In this clinical study we investigated the performance of propofol TCI in combination with remifentanil. Fifty-four adult patients scheduled for general surgery lasting longer than 1 h received a combined TCI of propofol (Marsh parameter set; propofol randomly either dissolved with long- or middle-/long-chain triglycerides) and remifentanil. Arterial propofol plasma concentrations and hemodynamic and derived electroencephalogram variables were determined at various stages before, during, and after surgery. Measured propofol plasma concentrations exceeded the predicted values by 59%, and 48% when recalculated with the Schnider parameter set. Pharmacokinetic population analysis showed a small central volume of distribution (3.55 L) and reduced clearance (1.31 L/min) for propofol. ASA status and sex were the only variables that had a significant influence on propofol pharmacokinetics. In a second step, a new pharmacokinetic variable set for propofol was determined in the first 27 patients. Post hoc performance analysis of the remaining 27 patients showed improved accuracy using the new variable set. Our results show that when remifentanil and propofol are combined, the Marsh and Schnider parameter sets systematically underestimate propofol plasma concentrations. Presented, in part, at the Annual Meeting of the European Society of Anesthesiologists, Amsterdam, The Netherlands, June 1, 1999, and the Annual Meeting of the American Society of Anesthesiologists, Dallas, Texas, October 12, 1999.