Predicting drug substances autoxidation

Abstract

Purpose: Chemical degradation and stability in formulation is a recurrent issue in pharmaceutical development of drugs. The objective of the present study was to develop an in silico risk assessment of active pharmaceutical ingredients (APIs) stability with respect to autoxidation. Methods: The chemical degradation by autoxidation of a diverse series of APIs has been investigated with molecular modelling tools. A set of 45 organic compounds was used to test and validate the various computational settings. Aiming to devise a methodology that could reliably perform a risk assessment for potential sensibility to autoxidation, different types of APIs, known for their autoxidation history were inspected. To define the level of approximation needed, various density functional theory (DFT) functionals and settings were employed and their accuracy and speed were compared. Results: The Local Density Approximation (LDA) gave the fastest results but with a substantial deviation (systematic over-estimation) to known experimental values. The Perdew-Burke-Ernzerhof (PBE) settings appeared to be a good compromise between speed and accuracy. Conclusions: The present methodology can now be confidently deployed in pharmaceutical development for systematic risk assessment of drug stability.