IDF21-0121 NDMA Analytics in Metformin Products: Comparison of Methods and Pitfalls

Abstract

Background: Since late 2019 concerns regarding trace levels of the N-nitrosamine NDMA in metformin-containing pharmaceuticals have been an issue, as they exceeded the maximum allowable intake of 96 ng/day for a medicine with long-term intake. This has resulted in thorough assessment of the safety of metformin products in this regard by manufactures worldwide. From the early focus on trace NDMA levels in drug substance, the scope of concern quickly expanded to drug products, associated excipients, and the potential for nitrosamine formation in packaging processes. As the scope expanded, the trace analysis challenge has grown. Metformin is made from the starting materials 2-cyanoguanidine and DMA, hence low residual amounts of DMA in the API are unavoidable. Inorganic nitrite is present in trace amounts in many excipients and can react with DMA to form the mutagenic N-nitrosamine NDMA, which needs to be measured in complex drug product formulations. The methods developed in the pharmaceutical industry are validated according to the applicable guidelines. Nevertheless, incorrect measurements can never be fully excluded. In food analysis, environmental analysis and in medical diagnostics, proficiency tests are a common tool to identify and eliminate systematic errors by comparing measurement results from individual laboratories. Aim: Here, we describe the outcome of an across-industry method comparison study for determination of NDMA in metformin-containing drug products between four labs, employing orthogonal mass spectroscopic methods, and discuss how it has helped to identify and mitigate artefactual results due to NDMA in-situ formation in the analytical process. Method: Proficiency testing employing orthogonal mass spectrometric methods was performed between four laboratories in three countries through analysis of different branded metformin-containing products. To exclude bias from sample inhomogeneity, fine powders of test tablets were prepared by one of the participating laboratories. To avoid bias from sample aging, experiments were performed contemporarily. Results: A risk of overestimating NDMA levels has been identified. Notably, special attention must be given to the sample preparation to prevent artefactual formation of NDMA in the extracted sample. Metformin medicines contain DMA in the ppm range, and nitrite in the ppb range, which react in the standard organic solvent DCM to form elevated levels of NDMA. This study has demonstrated that both removal of DMA and nitrite via solvent extraction, and “quenching” the reaction by addition of a nitrosation scavenger are effective remedies. Discussion: Without proper mitigation, NDMA in-situ formation causes overestimation of the NDMA content, which may in turn lead to an overestimation of affected batches and, ultimately, the risk to patients.