Brigitte D. Lacroix (1), Massimiliano Germani (1), David Sciberras (1), Miren Zamacona (2)
(1) UCB Pharma, Braine-l’Alleud, Belgium, (2) UCB Pharma, Slough, UK
Objectives: To develop a quantitative framework integrating pharmacokinetics (PK), receptor occupancy (RO), clinical outcome (PKPD) and tolerability data at the end of the proof of concept (PoC) phase II study, with the intention to propose a rational dose regimen selection for further development of padsevonil (PSL) in drug resistant patients with epilepsy.
Methods: Data from two clinical studies were used to develop the framework: the PoC study and a Positron Emission Tomography (PET) study for SV2A receptor occupancy. The modeling of the PK, clinical outcome PKPD and tolerability PKPD was based on data from the PoC study in 55 drug resistant patients with epilepsy receiving 400 mg bid. A population PK (popPK) model was developed based on rich sampling on 2 dosing occasions and sparse PK samples. The individual predicted daily average PSL plasma concentration obtained from the popPK model (Cavg) was integrated in the PKPD model describing the longitudinal daily seizure count data. The PKPD for tolerability was explored linking the Cavg with the treatment emergent adverse events registered as scores grouped in three categories: psychiatric disorders, nervous system disorders and a combination of the most relevant nervous system disorders and some general disorders such as “fatigue” and “gait disturbance”. The tolerability data were analyzed using a proportional odds model assuming consecutive scores to be highly correlated. The modeling of SV2A RO was based on data from a PET study in 11 healthy subjects who received various single doses of PSL (6.25 to 100 mg). Simulations were performed from the popPK model and the RO model, that were combined with the outcome from the PKPD model (estimated daily average concentration to reach 50% of maximum reduction in seizure count), the outcome of the PKPD tolerability model. Additionally, the expected GABAa receptor occupancy from a previous study was used to inform the modelling outcome in order integrate all information and propose dosing regimens for the further clinical development of Padsevonil.
Results: The PK of PSL was described by a 2-compartment model with first order absorption and first order elimination from the central compartment. An Emax model described the SV2A receptor occupancy as a function of PSL plasma concentration at the time of the PET scan where the EC50 was estimated at 3.1 ng/mL and EC90 at 27.9 ng/mL. The PSL effect on seizure reduction was described using a negative binomial distribution, taking previous-day seizure frequency into account. Baseline seizure rate was estimated at 1.3 day-1, placebo effect at 20% and EC50 at 190.6 ng/mL. A bi-modal distribution was used to separate the population into responders and non-responders, with almost 45% of the patients allocated to the responder group in this drug resistant epilepsy population. The PKPD model for tolerability showed that the probability of moderate to severe adverse events for patients treated with doses lower than 200 mg bid are similar to that of patients treated with placebo. No adverse-event dose dependency was evident when comparing 200 mg to 400 mg bid. Simulations of the PSL PK profiles at doses of 100 mg to 400 mg bid showed that the plasma concentration profiles after intake of 400 or 300 mg overlap largely. At 300 and 400 mg doses, the predicted concentrations remain above the EC90 for SV2A RO over the whole dosing interval at steady-state. At 100 mg bid, about 75% of the population is predicted to remain above EC90 at Ctrough.
Conclusions: An integrated approach has been used to project dose recommendations, using a combination of PET RO-exposure analysis, a population PK model and PK/efficacy effect estimation. Overall the analysis results allowed identification of doses proposed for further clinical development of PSL, including 400 mg bid as a top dose associated with >90% sustained SV2A RO in the entire population and quantifiable GABAa RO, and 100 mg bid as a low dose associated with >90% SV2A RO and lack of quantifiable GABAa RO.
Reference: PAGE 27 (2018) Abstr 8613 [www.page-meeting.org/?abstract=8613]
Poster: Drug/Disease Modelling - CNS