C. Ambery(1), M. Beerahee(1)
(1) Clinical Pharmacology Modelling and Simulation, GlaxoSmithKline, United Kingdom
Objectives: Drug discovery is expensive. Early clinical studies need to deliver more by answering key development questions, in particular, the dose response relationship and optimal dosing regimen. Here we present a case study for the determination of a dose response relationship after once daily dosing and subsequent assessment of pharmacodynamic similarity for two dosing regimens for a topically active novel drug undergoing clinical investigation. Data were obtained from a parallel design, placebo-controlled study with 3 different doses given once and twice daily. Trough measurements for clinical response (forced expiratory volume in 1 second – FEV1) were obtained during the course of study. Approximately 50 subjects received each treatment.
Methods: A population Emax model was used to describe the change from baseline dose-response relationship for Drug X following once daily (reference) and twice daily (test) administration.
Responsei = Emax * Dosei * Fx / (ED50 + Dosei * Fx) + η
The parameters are Emax (maximum response), ED50 (dose that achieves 50% of maximum drug effect) and F (relative bioavailability, where X = 0 or 1 indicates once daily or twice daily dosing). Dose is the total daily dose. The model assumed that test and reference had the same Emax and differed only in their ED50 values. Drug X dose levels investigated were QD1, QD2 and QD3 (once daily); and BD1, BD2 and BD3 (twice daily). Diagnostic methods and posterior predictive checks were used to assess the appropriateness of the population model. To evaluate pharmacodynamic similarity for the two dosing regimens a 90% confidence interval for the model parameter F was constructed using bootstrap [3]. A total of 2000 runs were conducted. The software’s NONMEM, PsN and R were used.
Results: Preliminary findings show responses at baseline were similar for each treatment. Diagnostic methods and posterior predictive checks showed the goodness of fit of the dose-response model to the data. The bootstrap median and 90% confidence interval for the model parameter F was 1.13 (0.53, 2.99), this was a reasonable quantification of the variability of the data.
Conclusions: The dose response relationship for Drug X was defined. Although the study was not designed to assess the bioequivalence between once and twice daily dosing regimens the dose scale method nevertheless provided a quantitative assessment of the similarity of the two dosing regimens.
Reference: PAGE 21 (2012) Abstr 2534 [www.page-meeting.org/?abstract=2534]
Poster: Other Modelling Applications