Examining the Use of a Mechanistic Model to Generate an In Vivo/In Vitro Correlation: Journey Through a Thought Process

A critical review on approaches to generate and validate virtual population for physiologically based pharmacokinetic models: Methodologies, case studies and way forward

PURPOSE

In silico modeling and simulation techniques such as physiologically based pharmacokinetic (PBPK) and physiologically based biopharmaceutics modeling (PBBM) have demonstrated various applications in drug discovery and development. Virtual bioequivalence leverages these computation tools to predict bioequivalence between reference and test formulations thereby demonstrating possibilities to reduce human studies. A pre-requisite for virtual bioequivalence is development of validated virtual population that depicts the same variability as that of observed in clinic. This development, validation and optimization of virtual population is a key attribute of virtual bioequivalence based on which conclusion of bioequivalence is made.

METHODS

Various strategies for optimization of virtual population based on appropriate considerations of physicochemical, physiological and disposition aspects are demonstrated with the help of six diverse case studies of immediate and modified release formulations. Once the virtual population is optimized to match in vivo variability, it can be used for various applications such as biowaivers, dissolution specification justification, f2 mismatch, establishing dissolution safe space, etc. In this review article, we attempted to describe various methodologies and approaches for optimization of virtual population using Gastroplus.

RESULTS

Strategies based on optimization of virtual population with emphasis on specific and sensitive parameters were portrayed. We have further elucidated considerations related to study design, in vivo variability, sample size for optimization of virtual population from Gastroplus perspective.

CONCLUSION

We believe that this review article provides a step-by-step process for virtual population optimization for interest of biopharmaceutics modeling scientists in order to ensure reliable and credible physiological models.

Integrating Drug- and Formulation-Related Properties With Gastrointestinal Tract Variability Using a Product-Specific Particle Size Approach: Case Example Ibuprofen

In the present study, an in vitro-in vivo extrapolation of dissolution integrated to a physiologically based pharmacokinetics modeling approach, considering a product-specific particle size distribution and a self-buffering effect of the drug, is introduced and appears to be a promising translational modeling strategy to support drug product development, manufacturing changes and setting clinically relevant specifications for immediate release formulations containing ibuprofen and other weak acids with similar properties.

In vivo release of peptide-loaded PLGA microspheres assessed through deconvolution coupled with mechanistic approach

In this study, a reevaluation of the in vivo release phases from long-release PLGA-based microspheres is presented, leading to a better characterization of the plasma concentrations/time profile. Microspheres were designed for intramuscular injection releasing a cyclic somatostatin analog over 70 days. Clinical study was performed in 64 healthy subjects receiving a subcutaneous dose of an immediate release solution as reference formulation and an intramuscular injection of microspheres as test formulation. The in vivo input curve was obtained by numerical deconvolution. Results showed that double Weibull function could not fit correctly the tri-phasic (burst, lag, and erosion) in vivo input profile typical for PLGA-based formulations, due to a change in the drug release trend in the terminal phase. Triple Weibull showed a significant improvement in the curve fitting, each term being assigned to one of the following phases: initial (burst/lag), erosion, and terminal phase of drug release. The existence of the additional terminal phase was confirmed by a mechanistic approach as well, which denoted that this phase was, most probably, a consequence of the release mechanism change from erosion to diffusion controlled. The same model demonstrated that the burst release was as well influenced by the polymer swelling, while currently existing theories state that the burst phase is mainly determined by the dissolution of immediately available drug substance and diffusion through surface related pores.