Date Published: July 1, 2020
Publisher: Springer International Publishing
Author(s): Jingyue Yang, Tim Andres Marzan, Wei Ye, Cynthia D. Sommers, Jason D. Rodriguez, David A. Keire.
A private testing laboratory reported in a Citizen Petition (CP) to FDA that 16 of 38 metformin drug products they tested had N-nitrosodimethyl amine (NDMA) amounts above the allowable intake (AI) of 96 ng/day. Because the FDA had been monitoring drugs for nitrosamines, orthogonal analytical procedures had been developed, validated and applied to detect the following nitrosamines in metformin drug products (if present): (i) NDMA (with a dedicated method) or (ii) NDMA (with a second confirmatory method), N-nitroso-diethylamine (NDEA), N-ethyl-N-nitroso-2-propanamine (NEIPA), N-nitroso-diisopropylamine (NDIPA), N-nitroso-di-n-propylamine (NDPA), N-nitroso-methylphenylamine (NMPA), N-nitroso-di-n-butylamine (NDBA) and N-nitroso-N-methyl-4-aminobutyric acid (NMBA). In contrast to the private laboratory results, FDA testing on the same set of 38 samples with orthogonal procedures observed amounts over the AI in only 8 of the 38 products and generally observed lower values than reported by the private testing laboratory. As described here, the investigation into the cause of the discrepancy revealed that N,N-dimethylformamide (DMF) can interfere with NDMA measurements. The data showed that the use of sufficient mass accuracy in the data acquisition and appropriate mass tolerance setting in the data processing to assure the selectivity of mass spectrometry measurements of NDMA in the presence of co-eluting DMF was necessary to prevent overestimation of the level of NDMA in metformin drug products. Overall, care should be taken to assure the necessary specificity in analytical procedures for adequate assessment of the nitrosamine level in drug products that also contain DMF or other potential interfering substances.
NDMA and other nitrosamines are common contaminants in low amounts in foods, beverages, cosmetics, water, tobacco products, and consumer goods (1–4). In 2018, observations of NDMA and NDEA in angiotensin receptor blocker drugs (ARBs) led to recalls of batches of products which had unacceptable amounts of nitrosamines (5,6). Since the ARB nitrosamine impurities were discovered, there have been additional drugs found to contain nitrosamines in the parts-per-million (ppm or ng/mg) to parts-per-billion (ppb or pg/mg) range (e.g., ranitidine in 2019, metformin in 2020 (7)) each with unique properties in terms of the route and source of their presence.
NDMA standard solution (100 μg/mL in methanol) was purchased from Chem Service (West Chester, PA). Stable isotope labeled NDMA (13C2; D6) (1 mg/mL in CD2Cl2) was ordered from Cambridge Isotope Laboratories (Andover, MA). N,N-Dimethylformamide (DMF) (99.8%) was purchased from Acros Organics, part of Thermo Fisher Scientific (Waltham, MA). Methanol (LC-MS grade) and formic acid (LC-MS grade) were purchased from Fisher Scientific (Hampton, NH). Water (ultrapure, resistivity ≥ 18.2 MΩ.cm) was from an in-house water purification system (ELGA) (Celle, Germany). All the metformin samples tested were commercially available finished drug products (FDP) and are listed in Table 1.
For the analysis of the 38 lots in this study, the FDA compared values obtained from the primary metformin testing method for NDMA (denoted as FDA-1) and the secondary confirmatory method (FDA-2) which was used to confirm the observations from the primary test. The performance characteristic values reported for FDA-1 and FDA-2 analytical procedures were derived from validation experiments and were shown to be fit-for-the-intended purpose of detecting nitrosamines in metformin products with LOQs at or better than 0.03 ppm. By contrast, the procedure and performance characteristics for the private laboratory method were derived from the information they provided in the Citizen Petition and was only partially verified in this work.
The private laboratory reported testing results to the Agency in a Citizen Petition. The FDA responded to the Citizen Petition by testing these products and noted different results (although 8 of the 38 samples from 5 firms did have NDMA amounts above the AI). In this set of 38 samples, only the ER products had detectable NDMA amounts as tested by FDA. The Agency responded by requesting voluntary recalls from the firms manufacturing these products. These firms complied with the request (16). Subsequently, the private laboratory hypothesized that because the FDA had not used a stable isotope-enriched standard for NDMA in the FDA method the agencies results were an underestimation of the NDMA amounts in the US metformin supply. In addition, the private laboratory “crowd sourced” samples from the US market tested them using the same method described above and provided the results in a non-peer reviewed preprint manuscript (15) which reported amounts of NDMA above the AI in 36% of the metformin drugs tested. As described in this study, the use of a stable isotope enriched standard for NDMA as internal control did not prevent the co-elution of DMF from causing overestimation of the NDMA amounts in many products when using the private laboratory approach.
The results of the study described here indicate that care must be taken to ensure analytical procedure specificity in the presence of potential interfering substances in the development and validation of analytical tests for drugs like metformin. Often, extensive knowledge of the drug substance and drug product properties, manufacturing processes, and the potential impurities that could interfere with an analytical measurement are required to assure sufficient analytical specificity. In this case, an isotopic peak associated with an allowed DMF impurity was shown to impact the private laboratory’s method only when insufficient mass tolerance was used. The data show the value of having the necessary measurement resolution to assure the specificity of an analytical procedure for small molecule analytes such as nitrosamine compounds present in ppm (ng/mg) to ppb (pg/mg) amounts. Furthermore, the data show that orthogonal methods provide assurance that matrix effects are not impacting the measurements and serve to confirm NDMA levels and the accuracy of the reported values.