doubt on validation

[morse]

Hi everubody.

Can you please give me your opinion on this matter?
I have been asked to write a VP of a chromatographic method which basically detect an impurity in A % (HPLC).
There is no reference std available for this impurity nor for the main product.
My opinion is that we necessarily need a reference std to execute tests like accuracy or linearity.
Anyway, I was asked to see if there's a way to validate the method all the same.
Can you please help?

Thanks a lot,

Morse

[krickos]

Hi

Depending on phase in R&D the level of validation may be different, but when getting close to registration the ICH guidelines provides some information.

From ICH Q2R1:
1.2.2 Impurities are not available
If impurity or degradation product standards are unavailable, specificity may be demonstrated by comparing the test results of samples containing impurities or degradation products to a second well-characterized procedure e.g.: pharmacopoeial method or other validated analytical procedure (independent procedure). As appropriate, this should include samples stored under relevant stress conditions: light, heat, humidity, acid/base hydrolysis and oxidation.
- for the assay, the two results should be compared;
- for the impurity tests, the impurity profiles should be compared.
Peak purity tests may be useful to show that the analyte chromatographic peak is not attributable to more than one component (e.g., diode array, mass spectrometry).

However have no first hand experiance of the above approach.

As for the API would need to make a reference standard yourself i.e characterize structure (NMR, IR etc) and purity and assay. See ICH Q7a §11.18


Cheer Chris

[morse]

Dear Chris,

Thank you very much for answering.

The product at issue is a starting material that is routinely analysed by our QC department. This is why I've shown some concern.

I think I will need to turn a sample of the impurity into a reference std.
I also believe that an absolute assay/purity technique will be required such as DSC or NMR or maybe, though not selective, potentiometric titration.

Of course, the HPLC method is not validated yet and I can't use this method to assess the sample purity/assay.

What is your advice?
thanks a lot agai for your kindness.

M.

[pkenny]

If this is an impurity method that is determining the result in area %, would this fall under the ICH Q2 guidelines for a Limit Test for Impurities? If so then you are not required to perform linearity or accuracy, only specificity and LOD. I would still think you need to have some type of purity reassurance. If you know what the impurity is maybe you could use elemental analysis?

[morse]

Hi Pkenny.

Do you think it might be thought as a limit test?
It actually gives a result in % area. It is not an on/off test.

I'll have a look soon at your reference.
Thanks a lot for answering.

M

[pkenny]

I would say, from what I have seen, most of the time they are limit tests. For example, the spec for that impurity may be < 0.5%. If you don't need to know whether it is really 0.2% or 0.3%, just that it is <0.5%, then you could use a limit test.

[morse]

Thanks a lot.
It actually sound a good idea.
I will try to pur it forward. Hopefully it'll be accepted.

Cheers,
Morse

[HW Mueller]

How does one determine any % if one doesn´t know any of the chemical/physical characteristics of a substance?

[morse]

I totally agree with you...
This is the point I 'm trying to make my boss understand.

M

[Consumer Products Guy]

I'm trying to make my boss understand.

Such an exercise can be challenging, or fruitless....

[krickos]

I'm trying to make my boss understand.

Such an exercise can be challenging, or fruitless....

hehe thats true

OK well still maintaining what I stated previously, but as you now stated it was a starting matrial we may actually slide into an "interpretation" area of GMP (ICHQ7a). As you likely are aware of the guideline "start" with the startmaterial and "GMP" increases towards the final drug substance.

I would not be suprised if there are some different approaches for references used for startmaterials/intermediates unless they also is used for the final drug substance. During commercial manufacturing I would not be surprised that such references are not fully caracterised. More like previous approved batches, commercial attenible materials, sure documented and tested at intervals but perhaps not as closely as a primary reference standard used for the drug substance.

Sorry if confusing this, but as I have seen some slightly different approaches for startingmaterial/intermediates at different API manufactureres I wanted to share it.

[morse]

Kri & all,

You're not confusing at all!
Thanks to you, Guy and HM for replying.

Can you give me any example of the kind of testes you've seen executed to characterize a reference standard for a starting material analysis?
I'm just curious..

Morse

[morse]

Hi All!

Eventually, I managed to make the boss understand the need for the reference standard.
Thnks a lot to all of you.

One more question if you don't mind.
What is your opinion about validating this method giving the result in percentage area?
I am not very favourable on this issue.
I've seen very few validated methods used in QC giving results in %A.
I reckon it's a good chioce during development. It might overestimate the impurity because of compounds not detected. It implies that the response factor is the same as the main compound. LOQ and LOD should be provided in concentration unit.

I don't believe it's a good idea considering that I have to characterize the reference std.

What is your opinion?

Thanks a lot

Morse

[krickos]

One more question if you don't mind.
What is your opinion about validating this method giving the result in percentage area?
I am not very favourable on this issue.
I've seen very few validated methods used in QC giving results in %A.
I reckon it's a good chioce during development. It might overestimate the impurity because of compounds not detected. It implies that the response factor is the same as the main compound. LOQ and LOD should be provided in concentration unit.

I don't believe it's a good idea considering that I have to characterize the reference std.

What is your opinion?

Morse

Hi

Actually when reading ICH guidelines like Q2R1, Q3A in conjuction you will note that using the responsefactor for the API/starting material is OK in a case like this. ie you may assume that responsfactor for mainpeak and impurity is the same and as a consequence you may dilute the start material to investigate LOD/LOQ.

As for area % methods, in opposite I am quite used to have them (ie regulatory approved), on the other hand my company develops "orginal drug substances/products" as such we typically have all relevant impurities availeble at least for drug substance/product sooner or later during development to compare responsfactors.
To be more clear, it it is not unheard of putting a = between area % and w/w % and correct impurities with significant different responsfactors.

I understand that you feel hesitant, I would likely feel the same as your case is sort of unknown to myself, also make sure that you document your approach with references to the appropiate guidelines.


Edit: This issue is also in part discussed under accuracy in a LCGC North America article: Volume 21 Number 12 December 2003. "Validation of Impurity methods, part II".

[morse]

Hi Krickos!

Thank you very much for answering.
I really appreciate your advice.

You mean there's nothing weird in issuing a area % hplc method.
You're right, I feel hesitant as it sounds inaccurate. Actually, it's just a personal feeling. I understand that once you can generate response factors it is practically the same.

Thank you for the reference you texted. I am going to have a look at it soon. Hopefully it's somewhere on the internet.

Thanks,
Morse