ABSTRACT
BACKGROUND: In open TCI and anaesthesia display systems, the choice of pharmacokinetic (PK) parameter sets of opioids is clinically relevant. Accuracy and bias of the PK models may be affected by administration mode and the co-administered hypnotic drug. We retrospectively evaluated the performance of eight PK parameter sets for alfentanil in two data sets (infusion and bolus application). METHODS: With the dosing history from two studies in orthopaedic patients anaesthetized with propofol or inhalation anaesthetics the alfentanil plasma concentration over time was calculated with eight PK parameter sets. Median absolute performance error (MDAPE), log accuracy, median performance error (MDPE), log bias, Wobble, and Divergence were computed. Mann-Whitney rank test with Bonferroni correction was used for comparison between bolus and infusion data, repeated measures analysis of variance on ranks was used for comparison among parameter sets. RESULTS: The parameters by Scott (original and weight adjusted) and Fragen had a MDAPE ≤30% and a median log accuracy <0.15 independent of the administration mode, while MDPE was within ±20% and log bias nearly within ±0.1, respectively. The sets by Maitre and Lemmens were within these limits only in the bolus data. All other parameter sets were outside these limits. CONCLUSIONS: In healthy orthopaedic patients, the PK parameters by Scott and by Maitre were equally valid when alfentanil was given as repeated boluses. When given as infusion, the Maitre parameters were less accurate and subject to a significant bias. We cannot exclude that the difference between bolus and infusion is partially because of the different hypnotics used.
