The assessment of convulsion risk: a translational PK/PD modelling approach
Zamuner S. (1), Paronetto M. (1), Moscardo E. (2), Renzulli C. (3), Terron A. (2), Poggesi I. (1)
(1) GlaxoSmithKline, Clinical Pharmacology Modelling and Simulation, Verona, Italy (2) GlaxoSmithKline, Safety Assessment, Verona, Italy (3) GlaxoSmithKline, DMPK, Verona, Italy
Objectives: Among the different central effects monitored during preclinical safety studies, convulsions are the one of the most concerning. After the preclinical characterization, the therapeutic window to be used for convulsions in the subsequent clinical development is typically obtained applying a multiplicative factor (e.g., 10 or 100-fold) to the no adverse event level [1]. This introduces a substantial level of subjectivity in the definition of the convulsion risk. A more thorough interpretation of the convulsion findings in preclinical experiments is recommended to define the safety margin to be applied in humans. In particular, the risk assessment should be more appropriately based on the level of systemic exposure, rather than dose, after having identified the most relevant pharmacokinetic parameters and metrics (e.g., Cmax, AUC) [1].
The objective of this communication is to propose a logistic model to assess the relationship between plasma concentrations and the probability of convulsions observed in preclinical studies and predict the risk for humans.
Methods: The non-clinical convulsion data obtained in different animal species for a compound under development were evaluated using nonlinear mixed effect models with the aid of NONMEM. A logistic model was developed, exploring the potential role of compound plasma concentrations, free fraction in plasma, species (mouse, rat and dog) and gender as independent variables. Model selection was based on statistical tests (Wald Test and Likelihood Ratio Test) and diagnostic plots such as Visual Predictive checks (VPCs) for categorical data [2].
Results: A statistically significant log-linear relationship was observed between total plasma concentrations and the probability of convulsion. Species was identified as statistically significant predictor. According to the model, it was estimated that the risk of human convulsion at the plasma concentrations anticipated to be associated to therapeutic benefit was less than 0.01%.
Conclusions: For the characterization of the convulsion findings, relevant metrics/parameters of exposure should be selected based on a thorough evaluation of the toxicokinetic data. Based on this and on the proposed PK-PD approach, an objective assessment of the hazard can be made, leading to a more robust definition of the safety margin to be applied in the subsequent clinical development.
References:
[1] Cavero I. Exploratory Safety Pharmacology: a new safety paradigm to de-risk drug candidates prior to selection for regulatory science investigations. Expert Opin. Drug Saf. (2009) 8(6):627-647
[2] Bergstrand M, Hooker AC, Karlsson MO (2009), Visual Predictive Checks for Censored and Categorical data (poster). Population Approach Group Europe (PAGE) 18th Meeting
