Emma Boström, Yi-Fang Cheng, Svante Nyberg, Frank Miller, Aleksandra Karin, Aurelija Jucaite, Märta Segerdahl, Eva Nordström
AstraZeneca R&D, Södertälje, Sweden
Objectives: AZD5213 is a histamine 3 (H3) autoreceptor antagonist currently under development for symptomatic treatment of Alzheimer's disease. Sleep disturbances is a well-known class-effect for H3 antagonists (1) and is associated with high H3 receptor occupancy (RO) at night. Therefore the preferred time course of RO during a dosing interval would be high RO during day (for cognitive improvement) and low RO during night. The objective of the modeling and simulation was to investigate if it was possible to obtain large diurnal fluctuations in RO for AZD5213 as well as to suggest doses that provide a wide spread in predicted diurnal RO vs. time-profiles for a Phase IIa study.
Methods: Data from a multiple ascending dose study in healthy volunteers was used to build a population pharmacokinetic model in NONMEM. The final population model was implemented in Berkeley Madonna. The model predicted plasma concentrations (Cp) were used together with the reported Ki-value for H3 RO from a human PET-study in order to calculate the anticipated RO vs. time profile using the formula RO(%) = 100*Cp/(Cp+Ki). For each investigated dose, 1000 stochastic simulations of RO vs. time at steady state were carried out and the results were transferred to R for calculation of medians and 90% prediction intervals and plotting.
Results: The pharmacokinetics was well described using a two-compartment model with 1st order absorption. Diagnostic plots and visual predictive checks proved the model fitted the data well. Using simulations, the predicted plasma concentration and RO vs. time profiles for a number of doses were explored and doses with a wide range of fluctuations in RO over the dosing interval could be identified. Following this analysis, it was decided to proceed to Phase II with three doses; 0.5 mg, 2 mg and 6 mg.
Conclusions: Using population modeling and stochastic simulations of pharmacokinetic data and the Ki-value from a PET study, it was possible to explore the predicted RO vs. time profiles for a range of different AZD5213 doses. This activity was essential for suggesting what doses to use in a Phase IIa study.
References:
[1] R. Parmentier et al. The brain H3-receptor as a novel therapeutic target for vigilance and sleep-wake disorders. Biochemical Pharmacology 73 (2007) 1157- 1171
Reference: PAGE 21 (2012) Abstr 2341 [www.page-meeting.org/?abstract=2341]
Poster: Study Design