Characterization of metformin by-products under photolysis, photocatalysis, ozonation and chlorination by high-performance liquid chromatography coupled to high-resolution mass spectrometry.

RATIONALE: Metformin (MTF) is the most widely prescribed drug for the treatment of patients with type 2 diabetes. Studies involving the removal of MTF from aqueous solutions and detailed information regarding the overall degradation process are scarce. METHODS: The degradation of MTF in aqueous solution induced by direct photolysis, photocatalysis, ozonation and chlorination was evaluated. The process was continuously monitored focusing on the identification and monitoring of the by-products formed by applying high-performance liquid chromatography coupled to high-resolution mass spectrometry in positive ion mode. The cytotoxicity of metformin by-products was evaluated with an MTT assay. RESULTS: The results from the chlorination and ozonation tests indicate metformin removal efficiencies of 60% after 30 min of exposure. On the other hand, direct photolysis (UV-C) and heterogeneous photocatalysis (TiO2 /UV-C) led to a lower degree of metformin degradation, with removal efficiencies of 9.2% and 31%, respectively, after 30 min of exposure. The mineralization rates varied from 20% for ozonation to 0.72% for photolysis, thereby indicating there was accumulation of degradation by-products in all experiments. Mass spectrometric analysis indicated the presence of five metformin by-products. It was not possible to identify any by-product generated in the photolysis, and, in all oxidative assays, the treated samples were nontoxic to HepG2 cells. CONCLUSIONS: It is also observed that all systems exhibited low mineralization rates, with the chlorination process being slightly more efficient in promoting the degradation, whereas the ozonation was more efficient in promoting the mineralization of metformin. Based on these results a route for the chlorination, photodegradation and ozonation of MTF, which comprised of its successive oxidation in the aqueous medium, could be proposed. It could also be concluded that the treated samples were not cytotoxic to HepG2 cells in a MTT assay. Copyright © 2016 John Wiley & Sons, Ltd.