🇪🇺 EMA Q&A Rev. 16 EMA/409815/2020 - MAJOR UPDATE

Modelling software certainly could have an impact - it’s entirely plausible that an electron-poor ring like this is going to be electron-withdrawing, at which point it could be argued that the +1 and -1 would cancel each other out. However, the CPCA is designed to be facile to implement, and conservative, so including this evidence moves out of the CPCA to a weight-of-evidence assessment.

As a result it would need a substantive research project and sufficient evidence to argue this to the point it could be included in a revised CPCA.

Only one of those is a nitrosated aromatic system - the two indolines both have the nitrosamine in non-aromatic rings (though fused to aromatic, yes) and the nitrogen has true single bonds.

image

1-nitrosoindoline would be category 4 by CPCA:
(0,2) for hydrogen count, and in a generic 5-membered ring, for a total of 5 points.

3-methyl-1-nitrosoindoline would be category 4 as well, with the same features but also the beta-methyl for an activating feature and a total of 4 points

I’m unsure why you were asked to test 2-Methyl-1-Nitroso indole (in an aromatic system and completely lacks a-hydrogens!), but would be really interested to see the data if you do - I’d expect a very different Ames strain profile (TA98/100 +/- S9, as described in Ponting and Foster).

Hi david,
you have applied CPCA approach for the “1-nitrosoindoline” (See below the structure) and “3-methyl-1-nitrosoindoline” (See below the structure) as they are considered as non-aromatic ring. and both are falls under category-4 and AI limit would be 1500 ng/day.

But, “2-methyl-1-nitrosoindoline” (See below the structure), which is present as an impurity in the Indapamide API EP monograph with a limit of 5 ppm and corresponding MDD of indapamide is 2.5 mg hence AI limit would be 12.5 ng/day for the “2-methyl-1-nitrosoindoline”.
If we apply CPCA approach for the “2-methyl-1-nitrosoindoline” then it also falls under category-4 and AI limit would be 1500 ng/day.

what i believe is that the toxicity of “2-methyl-1-nitrosoindoline” is not due to the nitroso structure but might be due to some other functionality or re-arrangement as described in Ponting and Foster .

so my question is, all the three indoline structures are more or less similar and “2-methyl-1-nitrosoindoline” is already having AI limit of 12.5 ng/day.
In this case,…
is it the correct way to apply CPCA approach and categorise “1-nitrosoindoline” and “3-methyl-1-nitrosoindoline” as category-4 with 1500 ng/day as AI limit?
or
should we stick to default AI limit of 18 ng/day for the “1-nitrosoindoline” and “3-methyl-1-nitrosoindoline” ?
or
should we consider AI limits 12.5 ng/day for the “1-nitroso indoline” and “3-methyl-1-nitrosoindoline” as per “2-methyl-1-nitrosoindoline”?

1 Like

Unfortunately I don’t have access to the EP monographs to read their rationale, but my understanding is that the limits in them are specification limits taking a number of factors into account, not strictly the regulator-requested AI limits. This compound lacks compound-specific carcinogenicity data to derive a limit from, so it must have been found in some other way, and I would definitely hope that EP take the CPCA into account when they next update the monograph.

My best guess would be that there are multiple nitrosamine impurities, which results in different treatment of the limits (EMA Q&A section 10) in the specification for each, but the AI limits for the compounds in the abstract case are not changed.

As a result, I would say that back-conversion from a DP-specific monograph for general AI limit determination for any compound is not an appropriate, especially as it is in this case implying a limit less than 18 ng/day which was the limit before the CPCA was developed, and argue even more strongly against taking a read-across from this value to other compounds where the default CPCA is 1500 ng/day.

Should a Fischer-Hepp rearrangement occur, as I described in Ponting & Foster, the potency of the resultant C-nitroso compounds is not cohort of concern-worthy - so again the limit if mutagenic would be 1500.

As a result I would argue strongly that the AI limit for all three indoline compounds should be 1500 ng/day in the general case. (as an aside, 2-methyl-1-nitrosindoline is actually category 5, lacking an a-CH2 group)

I’d like to read the justification from the monographs to see where that exceptionally low specification limit came from, if anyone can DM the relevant section to me!

1 Like

Hi David,
it is to clarify that it is not DP specific monograph it is API monograph in EP.
and find the below screenshot of EDQM FAQs on how EP defines the limit of impurities in their monograph.
i hope this helps.

so ultimately now the question is, can we defend the pharmacopoeial limits based on CPCA approach?

1 Like

Thanks, that probably solves my question: “not higher than the limits approved…” and based on batch data, rather than derived directly from the AI limits.

To my eye this suggests that this impurity is well-controlled in currently marketed batches, and the EP/EDQM are of the opinion that a good batch can be reliably controlled to this level, rather than that it is sufficiently potent to require control to this level for patient health?

I’m sufficiently far removed from the routine batch manufacturing process that I can’t comment as to how and whether the current specification could be changed/challenged if required, but this confirms that back-extrapolation is not applicable, because the monograph limits themselves are not derived based on the AI alone.

1 Like

As I see, David has provided a response to your questions. There have been a lot of studies on nitroso indoles which are well known. Many of them are mutagenic. My recommendation will be to look for surrgates with do have nitroso indole group and again, they are not “nitrosamines”. So, dont mix them up and provide separate analysis. Like the nitroso carbamates and nitroso guanidines, they do not need metabolic activation to show their mutagenicity in Ames. Look at this paper, which may give you some idea on nitroso indoles In-vitro testing and the carcinogenic potential of several nitrosated indole compounds - PubMed

2 Likes

Hello ,

Somebody can help me to set Limit for Vancomycin Hydrochloride Nitrosamine impurity .

image

Thanks in advance.

Hi Chirag
In the minutes CHMP PROM May 15, there is information about 2-Methyl-1-nitroso-2,3-dihydro-1H-indole.
image

It can be assumed that EMA does not consider this substance to be a strong genotoxic. Did Health Canada impose any limits on you for these molecules?

1 Like

Dear all,

Now CPCA Approach is available in Japan. I add the translation by DeepL translation.

Q15.
Attachment 2 (2) of the Voluntary Inspection Notice states, "If carcinogenicity test data are not available, limit values should be set using a scientifically valid method, such as setting limits based on structure-activity relationships or genotoxicity tests. However, the EMA guidance has recently been updated and The Carcinogenic Potency Categorization Approach (CPCA) for N-nitrosamines (Annex 2) has been presented. Is it acceptable to use the CPCA for setting limits in Japan?

A15.
For nitrosamines for which there is no sufficient carcinogenicity test data, it is acceptable to set limit values using the CPCA presented by the EMA. In the case of setting the limit values using the CPCA in the case of Attachment 2 (2) (ii) of the Voluntary Inspection Notice, consultation with the MHLW regarding the validity of the limit values is not required. In cases where the limit value can be regarded as scientifically valid, setting the limit value based on the conventional structure-activity relationship is not denied.

2 Likes

Can we extrapolate invivo data of Nitroso Quinapril to other ACE inhibitors like Lisinopril, Ramipril, Enalapril etc. if we have negative results of enhanced Ames test using test conditions described in guidance for their respective nitroso impurity (Nitroso Lisinopril, Nitroso Ramipril etc) to prove that these are not a CoC and can be controlled as per ICHQ3B?, as these all impurity have almost similar chemical environment.

Dear Team,

Reference to EMA Rev No. 16 for Eu and US Federal register on going harmonization,
how do one approach for method validation to comply current limit as per US (e.g. based on 96 ng/day) and Eu (CPCA approach e.g. 400 ng/day) to take care of linearity considering risk assessment with target concentration based on 96 ng and linearity to comply Eu requirement (based on 400 ng)?

Regards, Samir Shah

Dear @David, dear @ASrinivasan,
thank you very much for your clarifications!

@Yosukemino is there any chance to get the recording in youtube in English language? Portuguese i am not aware of…

Hi, @Jaideep,

Now English subtitles are available. Please watch the Youtube channel. Thank you for your help, @lucas10mauriz.

After clicking the subtitles button, you can select English subtitles through the setting button.
image

2 Likes

What α-Hydrogen Score shall be assigned for 0,1 count of Hydrogen Atoms on Each
α-Carbon?. EMA/HC guidance does not defined same. For example, what shall be the CPCA score for below impurity?

image

I suggest following the guide flowchart to assess the structure before jumping into the calculation of the score…

Does NDSRI have more than one alpha-hydrogen on one or both sides of the N-nitroso group?
→ NO >> AI=1500 ng/day

3 Likes

Thanks @Naiffer_Host

Hi Naiffer,
This is regarding “Nitroso Desmethyl Edoxaban” considering CAPA what will be the category?
find the structure below.
image
As per my point of view…

  1. Alpha-carbon is (2,2) which gives the score of 1
  2. Considering 6 membered piperidine ring, can we apply +2? because the ambiguity is related to the double bond in the subsequent 5 membered ring.

So what score shall we consider either 1+2 = 3 or only 1?

Actually the question is when 6 membered piperidine ring is directly fused with thiazole ring, can we consider piperidine ring as non-aromatic or aromatic?

considering the example- 8 of N-Nitroso Locaserin, mentioned in EMA Rev 17 and Health canada guidance. see screenshot below for reference.

Note that in this example a 7 membered ring is considered as non-aromatic and it is directly fused with benzene ring.