I.Meineke
Dept.Clinical Pharmacology, University of Göttingen Robert-Koch-Str. 40 D-37075 Göttingen
The population pharmacokinetics of morphine were investigated in a group of eight patients. An intravenous infusion of 0.5mg/kg morphine sulfate was administered to each patient over 30min. Plasma and CSF samples were collected during a time interval of 24h. Special emphasis was on the distribution into and the elimination from the CSF. Since the amount of drug present in the CSF is negligible in comparison with total dose, the CSF is unimportant in terms of mass distribution. Therefore a two step procedure was employed to establish a pharmacokinetic model for the concentration-time profile of morphine in the CSF. In a first step a three-compartment body model was fitted to the plasma concentration-time data of all patients with NONMEM (ADVAN11). Then the individual pharmacokinetic parameters for each patient were estimated. Afterwards the CSF concentration-time observations were linked to the plasma pharmacokinetic model. In this second optimization morphine plasma concentrations calculated on an individual basis were used to drive an additional CSF compartment (NONMEM, ADVAN6). The plasma three-compartment model was characterized by a clearance of 1838 ml/min and a central volume of 12.6 L. The intercompartmental clearances were 2085 mlMin and 181 ml/min respectively. The corresponding volume estimates were 115 L and 181 L. The best CSF fit was encountered when the additional compartment was connected with the fast equilibrating peripheral compartment. Moreover, there was evidence for a saturable transfer of morphine into the CSF. The volume of the CSF distributional space was estimated to be 0.094 L and the corresponding elimination clearance yielded 1.8726 ml/min The results presented in this work allow the prediction of morphine CSF levels after an arbitrary iv dose and can further be used to link morphine effects with CSF concentrations. Secondly, the usefulness of the population approach in the analysis of few sets of rich data is appreciated. Modelling uncertainties can effectively be prevented which frequently arise when concentration-time data are analysed on a patient-by-patient basis.
Reference: PAGE 10 (2001) Abstr 176 [www.page-meeting.org/?abstract=176]
Poster: poster