An evaluation of the accuracy of pharmacokinetic data for the computer assisted infusion of alfentanil.

The accuracy of using average alfentanil pharmacokinetic data in a computer assisted infusion pump (TIAC) to predict alfentanil plasma concentrations was tested in 35 patients (divided into three groups) receiving alfentanil and nitrous oxide in oxygen anaesthesia for lower and upper abdominal surgery. By frequently measuring the arterial plasma concentration, it was possible to determine the average prediction error for individual patients and for groups of patients. For the groups, there were no significant systematic over- or underpredictions of the alfentanil plasma concentrations (bias). However, there existed a moderate degree of variability (imprecision) within the groups, caused by deviations of measured and predicted plasma concentrations in the individual patients within each group. As a result, prediction errors of 22.2-32.5% can be expected with the average pharmacokinetic data used in this study to drive TIAC. It was concluded that, as a result of the moderate degree of imprecision, it is unwise to rely totally on the absolute values of alfentanil plasma concentrations predicted by a computer-regulated infusion pump such as TIAC. However, such devices can be used to attain rapidly a relatively stable plasma concentration that can be adjusted (titrated) to the requirements of an individual patient during anaesthesia.