Towards in vitro - In vivo correlation models for in situ forming drug implants.

In vitro-In vivo correlation (IVIVC) is a main focus of the pharmaceutical industry, academia and the regulatory sectors, as this is an effective modelling tool to predict drug product in vivo performance based on in vitro release data and serve as a surrogate for bioequivalence studies, significantly reducing the need for clinical studies. Till now, IVIVCs have not been successfully developed for in situ forming implants due to the significantly different in vitro and in vivo drug release profiles that are typically achieved for these dosage forms. This is not unexpected considering the unique complexity of the drug release mechanisms of these products. Using risperidone in situ forming implants as a model, the current work focuses on: 1) identification of critical attributes of in vitro release testing methods that may contribute to differences in in vitro and in vivo drug release from in situ forming implants; and 2) optimization of the in vitro release method, with the aim of developing Level A IVIVCs for risperidone implants. Dissolution methods based on a novel Teflon shape controlling adapter along with a water non-dissolvable glass fiber membrane (GF/F) instead of a water dissolvable PVA film (named as GF/F-Teflon adapter and PVA-Teflon adapter, respectively), and an in-house fabricated Glass slide adapter were used to investigate the impact of: the surface-to-volume ratio, water uptake ratio, phase separation rate (measured by NMP release in 24 h post injection in vitro or in vivo), and mechanical pressure on the drug release patterns. The surface-to-volume ratio and water uptake were shown to be more critical in vitro release testing method attributes compared to the phase separation rate and mechanical pressure. The Glass slide adapter-based dissolution method, which allowed for the formation of depots with bio-mimicking surface-to-volume ratios and sufficient water uptake, has the ability to generate bio-relevant degradation profiles as well as in vitro release profiles for risperidone implants. For the first time, a Level A IVIVC (rabbit model) has been successfully developed for in situ forming implants. Release data for implant formulations with slightly different PLGA molecular weights (MWs) were used to develop the IVIVC. The predictability of the model passed external validation using the reference listed drug (RLD), Perseris®. IVIVC could not be developed when formulations with different PLGA molar ratios of lactic acid to glycolic acid (L/G) were included. The present work provides a comprehensive understanding of the impact of the testing method attributes on drug release from in situ forming implants, which is a valuable practice for level A IVIVC development.