Abstract:

Setting a shelf-life for medicinal products can be on the critical path for the CMC package required for the approval of new, innovative or improved medicines and thereby can delay the availability of these products to the patients that need them.  Predictive stability models are seen as a solution for some drugs to optimise product development, for which there are many possible applications throughout development.

Many types of stability model for synthetic chemical entities and biologics are being developed for prediction of attribute levels past the point of available stability data at the recommended storage conditions. Models being considered for biological medicinal product application, include those based on product-specific stability data using linear regression analysis, or complex derivatives of the Arrhenius equation, whereas other models are based on prior knowledge that can include use of Bayesian statistics or machine learning algorithms. Meanwhile, the ICH Stability guidelines are under revision, and the publicly available draft version includes principles and best practices for the use of stability models as part of the modernisation of the stability regulatory environment. 

This presentation aims to use case studies that provide an overview for the development of predictive stability models across the biopharmaceutical industry and the current limitations to their use compared to the anticipated future state.  Included will be lessons learned from the COVID-19 pandemic and Regulator experience in managing predictive stability models used to extend product shelf-life.