I González-GarcÃa (1) ,V Mangas-Sanjuan (2), M Merino-Sanjuán (3), A GarcÃa-Arieta (4), IF Trocóniz (2), M Bermejo (5)
(1) Pharma Mar, Madrid, Spain; (2) Pharmacometrics and Systems Pharmacology. Pharmacy and Pharmaceutical Technology Department. University of Navarra; (3) Pharmacy and Pharmaceutical Technology Department. University of Valencia; (4) Head of Service on Pharmacokinetics and Generic Medicines, Division of Pharmacology and Clinical Evaluation, Department of Human Use Medicines, Spanish Agency for Medicines and Health Care Products; (5) Pharmacy and Pharmaceutical Technology Area. Miguel Hernandez University
Objectives: The main objective of this work is to assess the influence of in vitro inter-batch and in vivo inter-individual variability for the establishment of dissolution specifications based on an in vitro-in vivo correlation (IVIVC) and the in vivo bioequivalence (BE) standards. The final aim is to assess whether the in vitro specifications to ensure BE should be based on average dissolution data and average in vivo profiles or should incorporate in vitro and in vivo variability.
Methods: Dissolution profiles with three different (slow, medium and fast) release rates (first-order kinetics) were simulated and their corresponding in vivo profiles were obtained through a link model implemented in NONMEM assuming a linear level A IVIVC [1]. The medium release rate formulation was taken as reference. Then two test formulations were simulated; Test 1 with a very similar dissolution profile (i.e., BE based on the IVIVC) and Test 2 was a borderline formulation with f2 similarity factor value close to 50. Four scenarios of variability were considered (high (30%) or low (15%) for in vitro and in vivo). Residual variability was set at 5%. All variability terms were described with exponential models. In vivo profiles obtained through the IVIVC were analyzed to obtain Cmax and AUC ratios and their 90% confidence intervals (CI). All simulated scenarios (n=1000) were performed using non-linear mixed-effects modelling implemented in NONMEM 7.3[1]. For graphical and statistical analysis, the R (v.3.2.2) software and RStudio (v.0.99) were used.
Results: For both test formulations BE was concluded based on the average in vivo profiles which were included within BE limits (0.8 to 1.25) in all the variability scenarios. Nevertheless, when variability was accounted for and the BE analysis was based on the 90% CI, the borderline test formulation failed to show BE in the high variability scenarios
Conclusions: These results point out the relevance of taking into account in vitro and in vivo variability for setting dissolution specifications based on a validated IVIVC. Formulations with high inter-batch variability (even within quality control limits) could have average in vivo profiles within BE limits but fail under the 90% CI approach.
References:
[1] Gaynor C, Dunne A, Costello C, Davis J. A population approach to in vitro–in vivo correlation modelling for compounds with nonlinear kinetics. J Pharmacokinet Pharmacodyn 2011;38: 317–32.
Reference: PAGE 25 (2016) Abstr 6033 [www.page-meeting.org/?abstract=6033]
Poster: Drug/Disease modeling - Absorption & PBPK