Development of a non-targeted fingerprint profiling LC-HRMS method combined with chemometric analysis as a new approach for the detection of unexpected impurities.
Schneider Mandy M, Meins Jürgen J, Sandner Viktor V, Schubert-Zsilavecz Manfred M et al.
Valsartan drugs were recalled in 2018 because of the coincidental discovery of unexpected N-nitrosodimethylamine (NDMA). The fact that NDMA was not detected by the established comprehensive quality control methods is mainly due to the targeted approach applied being devoted to the determination of known and expected impurities. This study examines the use of a non-targeted screening approach to monitor the fingerprint profiles of finished products for unexpected impurities. To this end, we developed a profiling method for ramipril tablets using LC-HRMS coupled with SIMCA (soft independent modelling of class analogies) chemometric analysis. Fingerprint profiles were generated using a pentafluorophenyl reversed phase column with a methanol/water gradient coupled to a qTOF mass spectrometer with electrospray ionization in positive and negative mode. By applying chemometric analysis to the untargeted fingerprint profiles, it was possible to distinguish the fingerprint profiles generated from "clean" and spiked ramipril batches at impurity concentrations as low as 10 µg/L demonstrating the methods suitability for trace-level impurity profiling. SIMCA even allowed the identification of the substances responsible for the deviations in the fingerprint profiles. This study shows that untargeted fingerprint profiling coupled with multivariate data analysis is capable of being applied in addition to existing targeted quality control methods as a proactive approach to detect unexpected impurities instead of reactive batch recalls.