EMA Nitrosamines Q&A Rev 12 (10 OCT 2022) - Temporary AI limit

Dear all,

Yesterday EMA published the Rev.12 of the Nitrosamines Q&A guideline. Nitrosamines EMEA-H-A5(3)-1490 - QA Art. 5(3) Implementation_ for October PROM (QA10 and 21) (europa.eu)

In summary, the list of changes includes the following points:

  • Update of Q&A 10 to add 2 new N-nitrosamines to the list and
  • Addition of a new Q&A 21 on approach to control presence of nitrosamine while the AI is being established.

In detail, newly added to Q&As 10 and the list of potential N-nitrosamines is

  • 4-(Methylnitrosoamino)-1-(3-pyridinyl)-1-butanone (NNK) with a limit of 100 ng/day and
  • N-nitrosoduloxetine using a limit of 100 ng/day.

Furthermore, the footnotes for N-nitrosoduloxetine was added clarifying that the limit derived using structure-activity-relationship (SAR) /read-across approach using the TD50 of NNK as point of departure. In accordance with Health Canada evaluation for example.

The new Q&As 21 is clear in the statement that the Non-clinical Working Party (NcWP) and the CHMP are the final instances to decide on control options for the nitrosamine in the finished product.

The application of a temporary AI limit for a max of 12 month would be acceptable for avoiding shortages.

In practice, this means that when competent authorities are notified about a product containing a new N-nitrosamine exceeding the TTC limit of 18 ng/day, no market actions may be required for batches with N-nitrosamine levels ≤178 ng/day at the MDD pending the agreement of the AI. The adoption of the t-AI is not automatic and is evaluated by the relevant authorities at the time of notification. Use of the t-AI beyond 12 months will require additional consultation with competent authorities.

The t-AI should not be used as a target for development of validated analytical methods to quantify new nitrosamines since the long-term limits adopted by CHMP might ultimately be lower than the t-AI.

This is I think a great step into looking for pragmatic solutions for the NA issue.


Hi, @Diego_HM.
Thank you for sharing the information. I perfectly agree with you. This t-AI is helpful.

I want to add the calculation of t-AI. It may be similar to the LTL approach because the duration time is limited.

This t-AI has been derived using TD50 values calculated in the Lhasa carcinogenic potency database and is based on a probabilistic approach that there is a 33% risk that the “true” AI is below the t-AI. It is expected that the t-AI would be used for a period of less than 12 months, as an exposure over this period of time is not expected to increase the theoretical overall lifetime risk above 1:100,000.


I believe the clarifying statement about potential use of t-limits will be welcome by everybody.
Thoughts @conudel @AndyTeasdale @schlinjo1975 @jbercu

1 Like

Una muy esperada actualización por parte de la agencia Europea @LATAM_Community

1 Like

Muchas gracias por la información, es de gran utilidad.

I think it is now hitting them that 30% drugs will need to be withdrawn if they dont change their stance. Again, the industry possibly needs about 3 more years to understand this better and resolve this. Also, hopefully, the agencies start doing their own due diligence on endogenous nitrosation, the TD50s of complex nitrosamines, and think twice before recommending that products be removed from the market.


if any unknown nitrosoamine impurity possible in Drug product then how you calculate limit. as consider AI is 18 or some other way of calculation or if AI consider 18 which guidline u r following