Abstract Objective Ultrafiltration and diafiltration (UF/DF) operations have been demonstrated to clear leachables from drug substance, however there is limited data available. Consequently, comprehensive and systematic characterization of leachables clearance during UF/DF is required and essential. Methods To achieve this, the reduction capacity for 28 selected organic compounds spiked into 3 different proteins during UF/DF processes was investigated using liquid chromatography high-resolution mass spectrometry. Selection of compounds was based on their presence in representative biomanufacturing processes. Results Most compounds (24) showed clearance over 98% across the process for the 3 protein materials. The specific protein characteristics and process parameters for each protein had a minimal impact on clearance, with sieving coefficients essentially the same for each one of the 3 protein processes. The sieving coefficient is a parameter that characterizes clearance of compounds during UF/DF. Physicochemical properties of the compounds under study significantly influenced their clearance, with the octanol–water coefficient (Log P) being the most crucial factor. Compounds with Log P 7 showed lower but still significant clearance (> 93%). Other important parameters were established to be molecular weight, polarizability and solvent accessible surface area. Modelling tools based on Orthogonal Partial Least Squares (OPLS) regression were created to predict sieving coefficients. Conclusions The present work has created a strong background to describe the ability of UF/DF to remove potential organic leachables. Application of these modelling approaches becomes critical to support product safety assessments. Demonstration of significant removal along UF/DF operations confirms risk reduction of leachables coming mostly from upstream stages. Graphical Abstract
Dorival-García et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: