Tomi Hendrayana (1,2), Ralf Axel Hilger (3), Uwe Fuhr (4), Ingo G.H Schmidt-Wolf (5), Yon-Dschun Ko (6), Ulrich Jaehde (1)
(1) Department of Clinical Pharmacy, Institute of Pharmacy, University of Bonn, Germany; (2) School of Pharmacy, Institut Teknologi Bandung, Indonesia; (3) Department of Internal Medicine, University Hospital Essen, West German Cancer Center, Germany; (4) Department of Pharmacology, University Hospital Cologne, Germany; (5) Med. Klinik und Poliklinik III, Center for Integrated Oncology (CIO), University of Bonn, Germany; (6) Department of Oncology, Internal Medicine, Johanniter Hospital in Bonn, Germany.
Objectives: To develop a population PK model of 5-FU accounting for deviating infusion rates at the end of infusion and identify potential covariates explaining PK variability of 5-FU exposure during continuous intravenous (c.i.v.) infusion.
Method: A total of 1023 plasma concentrations from 93 patients after administration of 5-FU c.i.v for more than 18 hours, were analyzed using NONMEM 7.1.2. 5-FU concentrations were determined using HPLC or immunoassay. Population PK parameters were estimated, a flexible infusion rate was included and potential covariates such as sex, age, BSA and dose were tested. Model evaluation was performed using visual predictive checks, bootstrap methods and by Bayesian prediction of 272 plasma concentrations from 25 patients.
Results: A one-compartment model was successfully fitted to the data. Since in clinical routine deviations from the infusion rate frequently occur at the end of infusion, a flexible infusion rate was introduced during the last 10% of the infusion duration leading to a smaller objective function value. Sex was the only covariate found: Clearance (CL) of female patients was estimated to be 77% (95% CI: 67-91%) of male patients' CL. The inter-occasion variability (IOV) of CL was 12.8% (95% CI: 7.8-16.0%) and the inter-individual variability (IIV) of CL was 37.4% (31.5-43.2%). The fact that IOV of CL is lower than IIV suggests that 5-FU target exposure may be achieved by adapting the dose in subsequent cycles based on measured plasma concentrations from previous cycles. Model evaluation and Bayesian estimation using the estimated population PK parameters predicted individual plasma concentrations with satisfactory precision.
Conclusion: Accounting for irregularities of infusion rates is a promising approach to improve estimation of PK parameters of 5-FU during continuous intravenous infusion. Our model should be further evaluated for its potential to individualize 5-FU dosing maximizing efficacy and minimizing toxicity.
First author acknowledges The Directorate General of Higher Education, Department of National Education of The Republic of Indonesia, for financial support.
Reference: PAGE 21 () Abstr 2509 [www.page-meeting.org/?abstract=2509]
Poster: Oncology