Lessons Learned: How Can the ‘Nitrosamine Crisis’ Inform Future Formulations?

Concerns about nitrosamines—which have been known to be carcinogenic since the 1950s—in pharmaceutical finished products (medicines) have been percolating for decades. Nitrosamines are not an intentional part of a medicine’s chemical composition, but can form from reactions between some active pharmaceutical ingredients (APIs) and components of excipients. An increasing awareness of this phenomenon reached a fever pitch over the past few years. This ‘nitrosamine crisis’, which led to recalls and the complete withdrawal of one ‘over the counter’ active ingredient, should be a wake-up call for our industry. It can provide useful lessons, and perhaps prevent another such crisis in the future. 

Since nitrosamines’ carcinogenic potential has been known since the 1950s. How did the recent crisis develop? In part, the lack of adequate analytical methods was likely a factor. However, other factors also probably had an impact: including a lack of understanding of the composition of excipients and how combinations of certain APIs and excipients could facilitate formation of nitrosamines. In addition, the effects of continued exposure to small amounts of nitrosamines was also not understood. We now know that nitrosamine formation can occur with many drugs. For example, ranitidine was withdrawn because of its propensity to form a nitrosamine. Metformin is also of concern where trace levels of one of the starting materials appears to be the source of nitrosamine formation. Other drugs are also being investigated. 

We know how nitrosamines are formed – it is an oxidative process, typically involving a reaction between nitrous acid and a secondary amine. This has been known since the 1970s, possibly earlier. Nitrite impurities in excipients may be a source of nitrous acid. Many excipients are derived from natural feedstocks such as trees (cellulose products) and cereal crops (starch products), and nitrites and nitrates would be taken up from the soil in which the plants are grown. Nitrate can be converted to nitrite. We also know that nitrosamine formation appears to be more rapid at low pH. 

Given what we do know, what information do we need to predict nitrosamine formation? There are two main points to consider:

• The structure of the API molecule and its functional groups, particularly the presence of secondary amines; and 
• The composition profile of the excipient(s), including: 
  • The presence of nitrite and nitrate residues. 
  • Oxidative impurities such as hydroperoxides and oxidative residues from pulp manufacture. The first point should be straightforward. Any competent medicinal chemist should be able to undertake a functional group assessment of their development compounds to identify any secondary amine groups, etc., or groups that can transform into secondary amines during storage, formulation (manufacture), and handling. 

The second point is probably less straightforward in that we have not traditionally considered excipient composition profiles in such detail. In some cases, we may not have adequate analytical methods. We have methods that will detect nitrite, nitrate, and oxidative impurities, but are they sensitive enough? In addition, the excipient may be insoluble, so how do we release the components of interest from the excipient matrix? Obviously, we need to develop better sample preparation methods and better, more sensitive, detection methods. This could be fertile ground for academic research groups. 

If we think about the next possible crisis involving generation of toxic degradants and/or impurities, are there actions we can take to prevent it, or avoid it? Are there lessons to be learned from the nitrosamine crisis? My answer to both would be yes! 

If we understand what could go wrong, and the reasons it could go wrong, then we can implement measures to negate them, or at least significantly reduce their impact. There are databases that can be searched for compounds of e.g., carcinogenic potential. For example, a quick internet search identified the Carcinogenic Potency Database (CPDB) administered by the National Library of Medicine (NLM) under the auspices of the National Institutes of Health (NIH), and is accessible via TOXNET®. There may well be other databases holding similar information. Once we know the chemical structures of carcinogenic potential, there are other databases that can also inform us of the possible synthetic route(s) for their formation. The databases may not be perfect, but they are a start, and we can build on them.

With this information, and if we know the composition of our excipients in sufficient detail, we will be in position to assess the potential for formation of carcinogenic materials in our drug products. In essence, we could add a further risk assessment to our drug candidate selection process, whereby, in addition to the usual pharmacological assessments, the candidate is assessed for its potential to give rise to carcinogenic impurities (from either the drug itself, residual starting materials, or other impurities). If we have two possible candidates, and one has a much greater potential to form toxic or carcinogenic impurities— unless there are other considerations which outweigh the risk— then it might be prudent to select the candidate with the lowest potential to form such compounds. 

Once we have identified potential risks of carcinogenic impurities, we need to assess how to reduce that risk. There will likely be chemical precautions we can take, including the addition of inhibitors, etc. This will likely mean that the formulations will contain more rather than fewer components. The preceding discussion has focused on new drug candidates. What about the existing drugs on the market? Are there lessons to be learned here, too? I think there are! Using the same databases mentioned above, we should be able to determine which drugs have the greatest potential to give rise to toxic or carcinogenic impurities. We can then assess the potential for such impurities to be formed in the drug products by assessing whether or not the combination of the drug and excipients could give rise to them. If there is no risk identified, then nothing further need be done. If there is an identifiable risk, then further assessment needs to be undertaken and steps taken to reduce the risk below an acceptable level. This may include substituting the API for another similar type of drug, such as was done for Zantac®, which originally contained ranitidine, but now contains famotidine. 

These assessments should not be considered as ‘one-offs’. Knowledge is expanding, new discoveries are continually being made, and search engines are being upgraded. As new carcinogens and toxic compounds are discovered, the previous risk assessments, etc. should be revisited. In addition, in the evaluation of alternate sources of excipients, the risk assessment should be repeated for the new excipient source to take into account different processing conditions and raw material origin that may affect the excipient composition profile. It is better to anticipate problems than to be forced to react! 

The biggest gap is our lack of understanding of the detailed composition of our excipients. As has been discussed above, we need better methods of sample preparation and better, more sensitive methods of analyte detection. Without such developments, excipient users will continue to operate somewhat in the dark.