Recently we received Metformin API from two different supplier and in one supplier NDMA is getting about the 0.034 ppm which is nearing the limit of 0.038 ppm(if we consider the MDD of 2550).
As supplier is not agreeing to us can we increase the limit and what could be the probable justifications.
Dear TAUSEEF SIDDIQI,
just change the supplier 
It is not logical for the API supplier to set the specification limit for the API near the limit.
In similar cases we have demand to set the API specification around 10% of the nominal spec. limit and in most cases this have been accepted, even as an ‘‘internal limit’’.
If your supplier is denying to change the spec. limit, you could demand from him to set an internal limit for the batches which are going to be sent to you e.g. 0.003ppm.
It is not possible to change the limit of NDMA, it is one of the most studied nitrosamines and its limit is well established, at least up to now.
Please, also note that the quantity of DMA (dimethylamine, as starting material) is also critical in the API of Metformin.
best regards
Christos
3 Likes
Hi chris
Many thanks for your reply.
Surely supplier will be changed but to defend himself he shared the analysis with results about 0.003 ppm. When i investigated i found he have only used quantifier whereas there there was no qualifier to confirm the same ndma.
My question is as i have seen different monographs and available methods by authorities they have used both quantifier as well as qualifier. Is it mandatory to apply both or only one is workable??
2 Likes
I believe @chrischar has made a very important point. Remember that the limits of impurities are intended for the finished product. Regulators will evaluate the finished products against those accepted standards.
The risk assessment and robust evaluation of the ‘finished product’ will indicate if the specifications for your suppliers need to be adjusted. You might need to reconsider new suppliers. Please do not focus solely on the nitrosamines in the active ingredient (API), remember the case of DMA in Metformin. One needs to consider the nitrosamines AND the precursors that could form further nitrosamines downstream in the process.
The solution is not to find a justification for a higher limit because the supplier doesn’t want to comply with a tighter internal limit.
2 Likes
dear tsiddiqi
I totally agree with both Christos and Naiffer.
Based on previous experience, I would like to add a note from my side. Please always keep in mind that as Naiffer said, Regulators will evaluate the finished products against those accepted standards. But please note that this is related to your product during stability as well. and perhaps your NDMA go a bit higher during storage. So, such marginal levels in API would definitely be risky for your FP.
What’s more, there are plenty of Metformin suppliers with well controlled NDMA levels in API as to mitigate this risk, but please don’t skip a quick control of your API material from your end too.
kind regards,
Eleni.
2 Likes
Dear tsiddiqi,
as far as i now, qualifier is not mandatory to be presented. But the usage of both quantifier and qualifier secure the quality of the presented method.
Neverthelss, even if the API supplier presents a very good trend of low values of NDMA you cannot be safe if the limit is so high.
For this reason, they should at least agree with you in an internal limit around the 10% of the spec. limit.
best regards
Christos
1 Like
May I invite you to take a closer look at some data?
Take a look at this chromatogram I’m sharing:
It was generated as part of an in-use stability study involving metformin products combined with glibenclamide, glipizide, pioglitazone, alogliptin, and linagliptin. The samples were stored for 3 months at 30 °C / 75% RH and, at the end of this period, were analyzed for the presence of NDMA.
Detection was performed by LC-MS, using ESI in positive mode, with a single quadrupole monitoring the precursor ion at m/z 75.06 (NDMA).
Now I ask you: would you feel confident making a regulatory decision based on this data?
Let me walk you through my reasoning, from someone who’s also routinely evaluating this type of data.
The literature itself — including the article from which this chromatogram originated — shows that DMF can interfere with NDMA detection. Even at much lower concentrations, DMF can produce a signal that suppresses NDMA, masking the real risk. The result may appear to be below the limit, but still be compromised.
And this brings us to the core of your question: is using only a single quantification channel sufficient?
In the article in question, only the precursor ion is monitored. When we’re dealing with potentially carcinogenic impurities and extremely tight limits, relying solely on m/z 75.06 — on a unit-resolution system like a single quadrupole — is risky. Especially since this m/z can be shared with other substances present in the sample.
Another critical point: nitrosamines like NDMA tend to fragment at the ion source, losing NO (30 Da), which compromises sensitivity and increases noise. This makes the scenario even more challenging on instruments with intrinsic limitations in selectivity (single quad).
With that in mind, working with quantification and qualification channels means adding an extra layer of selectivity and confidence. What I bring to this discussion is the understanding that using two channels is not a regulatory requirement or a checklist item — it is a safeguard against false negatives and hidden interferences.
When we’re dealing with health risks, what matters most is trusting the data. And trusting the data doesn’t just mean accepting the final number — it means understanding how that number was produced and whether it holds up under audit or in a critical decision-making scenario.
Lucas Maciel
The article is available here: Patient In-Use Stability Testing of FDA-Approved Metformin Combination Products for N-Nitrosamine Impurity
3 Likes
Dear Lucas
Many thanks for the positive reply and i second you on believing and recommending using as qualifier as well as quantifier both to have precise data.
Can i add one more question to the ongoing query that how come supplier had ion source as EI for GCMS and the precursor ion is 74.
It should be 73 or 75 correct??
Dear Tauseefy,
In this case, the m/z 74 reported by the supplier is correct. The equipment uses an Electron Ionization (EI) source, commonly employed in GC-MS systems.
I won’t go into the ionization mechanism in detail here, but essentially, EI involves ionization of atoms or molecules by electrons, which are typically accelerated in order to remove one or more electrons from the target species.
Because the mass of the electron is negligible compared to the molecular mass, the resulting ion’s m/z closely reflects the actual molecular weight. That’s why we refer to the molecular ion (not protonated molecule) in this context — and in the case of NDMA, this corresponds to m/z 74.
1 Like