Vladimir Piotrovskij, Eric Snoeck, Achiel Van Peer
International Clinical Research and Development, Department of Clinical Pharmacokinetics, Janssen Research Foundation, B-2340 Beerse, Belgium
Population analysis of liarozole pharmacokinetics was performed using the NONMEM approach. The database consisted in liarozole plasma concentration measurements obtained both from healthy male subjects (totally 46 from 3 studies) and prostate cancer patients (totally 194 from 3 clinical trials) at the steady state. In healthy subjects, trough levels, as well as multiple samples during one dosing interval after 150mg bid were available, while in patients there were only trough levels measured on 75, 150 or 300mg bid therapy. Healthy subject data subset was used to choose a pharmacostatistical model including no covariates. The one-compartment model (parameterized in terms of the apparent clearance, CL, and the volume of distribution, V) with the first-order absorption was selected. Inter-individual variability in all parameters as well as a covariance between clearance and volume of distribution was assumed.
Also using only healthy subject data, a preliminary population model was developed which included effects of the body weight (WT), age and body surface area (BS). CL was shown to be proportional to BS (6.5L/h per m2) and decreased with the age of the subjects (-0.06L/h per year). Before applying the model to patients, fixed and random effect parameters related to V and to the absorption rate constant were fixed at the values estimated in healthy subjects since patient data might not provide any additional information on these parameters. A special technique was developed that allowed to have inter-individual variability in V and correlation between CL and V fixed while the parameter corresponding to inter-individual variability in CL remained open for iterations.
The final population model for CL was built and validated using patient data in combination with healthy subject data. A typical value of CL in 70 yr-old patient with BS of 1.97m2 was about 13L/h. It was shown that CL did not depend significantly on the liarozole maintenance dose. The duration of therapy was also proved not be a significant factor in general. Nevertheless, after 14 weeks of therapy or more CL was significantly lower than at earlier time, however, the difference was minor (CL ratio 0.94). There were significant differences between clinical trials: typical values of CL in LIA-INT-2 and LIA-INT-5 trials was 0.72 and 0.83 of that in others studies and trials, respectively.
Reference: PAGE 5 () Abstr 578 [www.page-meeting.org/?abstract=578]
Poster: poster