Using Modelling & Simulation techniques to optimise the design of a paediatric PK/PD study
Karl Brendel, Clare Gaynor, Cyrielle Dumont, Alexia Blesius and Marylore Chenel
Department of Clinical Pharmacokinetics, Institut de Recherches Internationales Servier,
Objectives: European Paediatric Regulations in force since January 2007 specify that, when applying for a new indication or new route of administration in adults, a corresponding Paediatric Investigation Plan (PIP) must be submitted to the Paediatric Committee. A development plan for use of Servier drug X in the paediatric population is, therefore, underway. The study being planned is a dose-finding PK/PD study which involves a single i.v. bolus administration of the compound to children in the target population. Modelling & Simulation techniques are playing a key role in designing this study to ensure that the ideal dose is identified as early as possible and that blood samples are being taken at optimum times.
Methods: As an adult physiologically-based (PB) PK model has previously been developed for this compound, this existing model was adapted and used to simulate the plasma concentration-time profiles which would result from an intravenous bolus dose of 0.1 mg/kg in children from 2 to18 years. A population PK model fitting both parent and metabolite concentrations, was built on these plasma concentration-time data and linked to a population PD model previously developed in adult, in order to simulate response-time profiles at different doses and in several age classes over the range being considered. According to pharmacological results obtained in juvenile animals and given the mode of action of this compound, it is expected that the PK/PD relationship will not differ greatly from that observed in adults. This PD model is an agonist Emax model with an effect compartment taking into account the activity of both the parent drug and its active metabolite.
According to biomarker baseline values as well as to response and safety targets, 4 age subsets were determined.For each envisaged dose and each subset of age, the proportion of subjects predicted to have a PD response between the efficacy target value and the safety threshold value, were calculated from simulated profiles, at various experimental times after the i.v. bolus.
Results: From PBPK simulations, children are slightly less exposed to the drug than adults. A three compartments-model for the parent and a one compartment-model for the metabolite were used to fit the simulated concentration-time data. Weight effects were applied and estimated on clearances and volumes of distribution. Simulations of the PD response and calculations of the a priori distribution of response rates for each age class end with a range of doses to be tested higher in children than in adults.
Conclusions:Based on this Modelling & Simulation approach, a starting dose and a range of doses to be tested, which comply with requirements both in terms of efficacy and safety can be chosen for each age class.