Leveraging biomarker exposure-response in drug development
C. Ambery(1), M. Beerahee(1)
(1) Clinical Pharmacology Modelling and Simulation, GlaxoSmithKline, United Kingdom
Introduction: An anti-inflammatory drug is currently in Phase II development for asthma. Phase I clinical studies and pre-clinical knowledge strongly suggested that this drug could show clinical benefit for patients with asthma. A biomarker-response model was characterised from early clinical studies, however informative linkage of the biomarker to clinical efficacy was not possible due to paucity of clinical data with this anti-inflammatory mechanism. A clear dose rationale for later phase studies to determine the therapeutic dose was not fully established. The purpose of this study was to explore a wide enough plausible range of doses in order to adequately characterise the potential dose response. These data were also put into context with emerging PK data from various formulations being explored.
Objectives: The objective of this work was to support selection of a range of doses of this anti-inflammatory drug in a Phase IIb asthma patients study using an optimal formulation.
Methods: The percentage change from baseline biomarker exposure-response relationships was previously defined by a turnover inhibitory Emax model using data from the earlier Phase I clinical study. Through integration of the potency of the drug with its estimated oral clearance, subsequent simulations of the exposure-response relationship using average steady-state concentration (Cav) were performed using Berkeley Madonna as a simulation engine. The simulation platform was chosen as it provided a more interactive, flexible and dynamic interaction with the clinical team for exploration of various scenarios, population variability and agree on assumptions. Interactive simulations were performed to predict the proportion of subjects exceeding the biomarker IC50, IC80 and IC90. One thousand trial simulations of 100 subjects per dose level were performed for each scenario investigated accounting for various ranges in inter-subject variability in biomarker and Cav. In addition, these simulation scenarios were linked with emerging PK data evaluating optimal clinical formulation(s).
Results: Biomarker exposure-response relationships were simulated over a wide dose range. Simulation results provided a basis for the number and spacing of doses for investigation in a Phase IIb asthma patient study. Four dose levels were selected based upon the proportion of subjects predicted to exceed IC80 at Cav.
Conclusions: Simulations of the biomarker exposure-response relationship facilitated the selection of doses to be investigated in a Phase IIb dose ranging study in asthmatic patients.