Pharmacokinetic-Pharmacodynamic Modelling & Simulation of Org 52186, a V3 antagonist, in Support of a Challenge Agent Trial Design with dDAVP
Jeroen Elassaiss-Schaap (1), Bill Knebel (2), Ellen Hulskotte (1) and Rik de Greef (1)
(1) PPDM, MSD, Oss, The Netherlands; (2) Metrum Research Group, Tariffville, CT, USA
Objectives: Org 52186, a V3 antagonist, was in development as an agent against depression before being discontinued. The objective of this analysis was to design a Proof-of-Principle trial to show antagonism of the V3 receptor of Org 52186 in vivo in humans, optimizing for trial cost. This trial should follow up on a single rise dose trial of Org 52186.
Methods: The pharmacokinetic-pharmacodynamic (PK-PD) properties of dDAVP were richly sampled in an experimental medicine study in healthy volunteers on the V2- and V3-selective agonist desmopressin (dDAVP). ACTH levels were determined as described by a PK-PD model developed in NONMEM 6. A population PK model of Org 52186 was built on data from healthy volunteers in a first-in-man single ascending dose study. In vitro affinity and plasma protein binding of Org 52186 were used to predict the antagonism of dDAVP-induced ACTH release. Simulations were performed using resampling of residuals of dDAVP PK-PD analysis. The simulated values were used to reconstruct trials of varying designs and analysis approaches in R, to determine an acceptable design of the challenge trial.
Results: The ACTH-releasing activity of dDAVP was quite variable as can be expected for such a pituitary hormone and in light of the dose limiting V2 effects of dDAVP. The PK-PD model that was developed described a subset of healthy volunteers not responding to dDAVP whereas substantial between-subject variability in maximal effects was observed among the other subjects. The resulting model was not yet flexible enough to capture all trends in the longitudinal responses and therefore the residuals were taken into account in subsequent simulations. The exposure to Org 52186 was readily captured by a pop-PK model.
Simulations showed that a classical cross-over trial design was not acceptable due to the variability in ACTH response to dDAVP. The designed screening phase was therefore expanded to include an evaluation of responsiveness to dDAVP, increasing the cost of screening but decreasing over-all cost. An apparent optimum was found when screening for an average ACTH response of 8 ng/ml and including n=12 subjects in the final challenge test. Inclusion of resampled residuals was determined to be essential for conservative estimation of trial sample size.
Conclusions: Innovative resampling approaches and trial design element evaluations were able to select an acceptable design for a dDAVP-challenge trial of Org 52186.