Amais Ahmad (1), Xavier Pepin (2), Leon Aarons (1), Amin Rostami-Hodjegan (1,3) [on behalf of WP4, OrBiTo, an IMI Project]
(1) University of Manchester, United Kingdom, (2) AstraZeneca, United Kingdom, (3) Simcyp Limited (a Certara Company), Sheffield, United Kingdom
Introduction: Oral drug absorption is a complex process depending on many factors including the physiochemical properties of the drug, characteristics of the formulation and their interplay with dynamic gastrointestinal physiology. The ability to anticipate impact of these processes is of great importance in drug and formulation development in order to accurately predict the systemic exposure. In silico physiological-based pharmacokinetic (PBPK) models provides a systems pharmacology approach to predict plasma concentration-time profiles using in vivo in vitro extrapolation (IVIVE) and other preclinical data. Therefore, the models can help with various aspects of drug development process such as, for example, anticipation of human pharmacokinetics, choice of formulation, impact of physiology on exposure, prediction of sources of variability in exposure, formulation optimisation. Despite recent successes of PBPK in many areas of drug development, an improvement in their utility for evaluating oral absorption is much needed. Three PBPK software packages (GI-Sim, Simcyp®, and GastroPlusTM) were evaluated as part of the Innovative medicine Initiative Oral Biopharmaceutics Tools (OrBiTo) project using “bottom-up” anticipation of human pharmacokinetics.
Objective: Evaluate the overall performance of PBPK software packages (GI-Sim, Simcyp®, and GastroPlusTM)
Methods: Fifty eight active pharmaceutical ingredients (API) were chosen from an OrBiTo database, meeting the defined minimum inclusion criteria [1]. These 58 API represented over 200 human studies, and approximately 800 clinical study arms. Population representative simulations were performed by modellers from project partners. Input parameters were harmonised across different software packages by providing guidance on selection and calculation of input parameters. Pharmacokinetic (PK) parameters (AUC, Cmax, Tmax) of both simulated and clinical data were calculated using automated tools. Overall prediction performance was evaluated based on performance indicators (Fold error-FE, Average fold error-AFE and absolute average fold error-AAFE).
Results: On average, AUC value was over predicted with median FE of 1.57 and Cmax was accurately predicted with median FE of 1.04. The prediction performance was fairly consistent across different software packages with a few exceptions mostly related to different input parameters for the models. Around half of the simulations were within 2-fold error for AUC and around 90% of the simulations were within 10-fold error for AUC. A general trend of overprediction was observed for all formulation with AFE ranging from 1.1 to 2.3 except slightly underprediction with immediate release (IR) suspensions with AFE of 0.9. There were higher percentages of within a certain specified fold errors in case of controlled release (CR) formulations as compared to IR formulations. Moreover, there was less FE variability in case of CR formulations, having AAFE approximately half than of IR formulations.
Conclusion: Average predictive performance did not seem related to software package but there was a very high level of variability in predictions for some APIs. This variability could be related to many factors such as compound specific properties, the quality and availability of information, and errors in scaling from in vitro and preclinical in vivo data to human in vivo behaviour.
References:
[1] A. Margolskee et al., “IMI – oral biopharmaceutics tools project – evaluation of bottom-up PBPK prediction success part 1: Characterisation of the OrBiTo database of compounds,” Eur. J. Pharm. Sci., vol. 96, pp. 598–609, 2017.
Reference: PAGE 27 (2018) Abstr 8620 [www.page-meeting.org/?abstract=8620]
Poster: Drug/Disease Modelling - Absorption & PBPK