Predictive performance of computer‐controlled infusion of remifentanil during propofol/remifentanil anaesthesia

Background The predictive performance of the available pharmacokinetic parameter sets for remifentanil, when used for target-controlled infusion (TCI) during total i.v. anaesthesia, has not been determined in a clinical setting. We studied the predictive performance of five parameter sets of remifentanil when used for TCI of remifentanil during propofol anaesthesia in surgical patients. Methods Remifentanil concentration–time data that had been collected during a previous pharmacodynamic interaction study in 30 female patients (ASA physical status I, aged 20–65 yr) who received a TCI of remifentanil and propofol during lower abdominal surgery were used in this evaluation. The remifentanil concentrations predicted by the five parameter sets were calculated on the basis of the TCI device record of the infusion rate–time profile that had actually been administered to each individual. The individual and pooled bias [median performance error (MDPE)], inaccuracy [median absolute performance error (MDAPE)], divergence and wobble of the remifentanil TCI device were determined from the pooled and intrasubject performance errors. Results A total of 444 remifentanil blood samples were analysed. Blood propofol and remifentanil concentrations ranged from 0.5 to 11 μg ml −1 and 0.1 to 19.6 ng ml −1 respectively. Pooled MDPE and MDAPE of the remifentanil TCI device were –15 and 20% for the parameter set of Minto and colleagues ( Anesthesiology 1997; 86 : 10–23), 1 and 21%, –6 and 21%, and –6 and 19% for the three parameter sets described by Egan and colleagues ( Anesthesiology 1996; 84 : 821–33, Anesthesiology 1993; 79 : 881–92, Anesthesiology 1998; 89 : 562–73), and –24 and 30% for the parameter set described by Drover and Lemmens ( Anesthesiology 1998; 89 : 869–77). Conclusions Remifentanil can be administered by TCI with acceptable bias and inaccuracy. The three pharmacokinetic parameter sets described by Egan and colleagues resulted in the least bias and best accuracy.