Quantitative Aspects of Non-targeted Analysis for Organic Extractables/Leachables; Insights Secured by Considering the FDA CLAP List Compounds

Abstract

Substances released from a drug product’s manufacturing, packaging, and delivery system are termed leachables. Medical device leachables are released during their clinical use. Packaged drug products and medical devices are profiled for leachables (and/or extractables as probable leachables) by screening the drug product or extracts via non-targeted analysis (NTA) employing chromatography coupled with mass spectrometric detection. A leachable’s concentration is important as concentration is establishes a patient’s exposure to leachables during the clinical use of a drug product or medical device. Two issues are relevant when quantifying extractables or leachables (E/L) during NTA: when is quantitation necessary, and how is quantitation accomplished? Quantitation for toxicological safety risk assessment is necessary if the E/L is present at a level that exceeds the Analytical Evaluation Threshold (AET). To address compound-to-compound variation in analytical response, the AET is lowered via an Uncertainty Factor (UF). Estimated concentrations for identified E/Ls are accurate if either a reference standard is available for the E/L or the E/L can be linked to a surrogate reference compound. However, these approaches are not applicable to unidentified E/Ls or when reference compounds are unavailable. The means for calculating the UF and the use of a universal quantitation surrogate, specifically the median relative response factor (RRF) of a data set of compounds, to estimate the concentration of unidentified E/L are considered herein. Using the FDA CLAP list as a reference, this manuscript concludes that the proper calculation of the UF involves the median RRF and the RRF at the 16% percentile. Furthermore, this manuscript establishes that methods coupling liquid chromatography with mass spectrometric detection (electrospray ionization, ESI) are inaccurate when quantifying unidentified E/Ls given the large variation in RRF values. However, as RRF values for GC/MS are less variable, concentrations estimated for unidentified E/Ls using RRFmedian can be accurate and protective.