8 Population Distribution of Platinum Pharmacokinetic Parameters for Dosage Recommendations of CDDP and Adaptive Control during the Course of Therapy, using different Softwares.

J. Bres, D. Cupissol, J. Nouguier-Soule, Ch. Gestin-Boyer, Ü. Büttner, B. Palomba

1Laboratoire de Pharmacocinetique, UMI, Faculte de Pharmacie, Montpellier, France 2INSERM, Laboratoire de Pharmacologie des Tumeurs, Montpeillier 3URA-CNRS 698, USTL, Montpeillier

This study was undertaken in order to make an evaluation of the performance of several softwares for pharmacokinetic parameters predictions and for dosage recommendations when the same population distribution of Pt pharmacokinetic parameters was introduced in these softwares. In routine clinical practice two softwares are actually used for the adaptive control of cisplatin dosage. A three compartment model (t1/2 = 0.28 day, 1 day and 10 days) (APIS) or a one compartment model (non compartmental approach; CL and V) (MICROPHARM) being fitted to the data. When cisplatin is administered as a continuous infusion during 4 to 5 days the two compartment pharmacokinetic model with only one serum compartment seems to be more appropriate; this was shown by simulation of drug concentration in the tissue compartment.

1. NPEM2 of USC*PACK Programs allows a direct estimation of population- typical (mean) and population-variability (standard deviation) values for Pt pharmacokinetic parameters in a single stage of analysis applied simultaneously to data from many individuals. This population distribution was obtained from 25 patients (48 courses), cicplatin being administered either as a "short", 2 hours, infusion (15 courses with blood samples up to 30 days), or as a continuous infusion over 4 to 5 days (5 courses with blood samples up to 30 days; 28 courses with blood samples up to 4 to 5 days).

2. These population characteristics, for a two compartment model, were introduced in four softwares: USC*PACK, MICROPHARM, ABBOTTBASE PK System, and APIS, after the necessary conversion from microconstants (rate constant of transfer between compartments, rate constant of elimination and initial volume of distribution) to macroconstants (coefficients, normalized to the administered dose, and exponents of the two exponentional terms).

3. A bayesian estimation was performed to compute Pt pharmacokinetic parameters from routine clinical data during the course of the treatment: 12 courses from new patients receiving cisplatin as a 4 days continuous infusion with 4 to 10 samples during 5 days. Clearance, volumes (V1, V2 and Vtotal), rate constants of transfer between compartments k12 and k21, rate constant of elimination, k10, and the terminal half life predictions using the 4 softwares were statistically evaluated.

4. Dosage predictions for the future course, and for a daily adaptive control during the course of therapy, were also compared.