Chromatography Forum

Could you please advice in respect to an impurities calculation issue.
We have developed / validated a method where impurities are calculated by the known formula:
%imp= (Atest/Aref)* limit.

Comparison of the % percentage for an unknown imp. with specific rrt with the %area presented in the chromatogram shows really high differences. Is this normal??

eg: impurity calculated with the method formula: 0.015%
impurity as given in the chromatogram %area: 0.22!!!

Forced degradation in acidic conditions also showed possible degradation product with the same rrt.
Is it possible that the formula ''hides'' impurities-degradation products?? --->> problems during stability !!
I mean maybe the detector's response is not correct at the concentration limit. To be noted that linearity was ok during validation..

Thank you for your help!

.

We have developed / validated a method where impurities are calculated by the known formula:
%imp= (Atest/Aref)* limit.

That "known formula" unfortunately is not entirely clear to me . What is "limit"? What is Atest and Aref?
I suppose Atest refers to the peak area in the test solution and Aref to the peak area of a (external) reference solution? Then "limit" should be to the concentration of the external standard? What's the external reference? The API?

If I understood correctly, you see differences between calculations of unknowns when using API as external standard and 100% area method, correct?

There is a wealth of possible reasons:
- with 100% area method you generally need linearity of the method from LOQ to 100% - is it given?
- your unknown impurity might have a significantly different response factor than the API
- some impurities simply might not be visible with your method

Thank you for your reply.

I'm sorry for not being that clear but i needed a quick answer and yes, it finally seems that the ''known formula'' is not unknown for you as for any other scientist that work in industry .
The limit refers to the threshold as defined by ich and maximum daily dosage of the product under which qualification is performed.

Under the term ''%area'' i mean the percentage of impurity resulting if divided with the total of peak areas of the chromatogram. This is stated in the chromatogram.
Results under those two calculations really differ.

In respect to your proposals (thank you):

There is a wealth of possible reasons:

- with 100% area method you generally need linearity of the method from LOQ to 100% - is it given? ---> Yes, from LOQ to 120% if I remember correctly.
- your unknown impurity might have a significantly different response factor than the API ---> Response factors have been defined for known impurities during method validation, how could I define the response factor for an unknown?
- some impurities simply might not be visible with your method ---> Impurity is visible, the way of calculation differs. This is my main question, what might the problem be??

Thanks again,

The limit refers to the threshold as defined by ich and maximum daily dosage of the product under which qualification is performed.

In that case "ARef" should refer to the peak area of a reference solution corresponding to that limit. So again, what is ARef in your case???

Yes!

Sometimes when doing an impurity method, you are overloading the API on the column so that you can get quantifiable results for the impurities. Since that peak is so large/possibly overloaded is can lead to peak shape issues/variable integration/etc. That is why when you use a standard, you use a concentration more comparable to the concentration of the potential impurities or your impurity limit, because this concentration does not overload the column and therefore will lead to better peak shape and more consistent results.

- with 100% area method you generally need linearity of the method from LOQ to 100% - is it given? ---> Yes, from LOQ to 120% if I remember correctly.

Really? Tristane was shooting in the same direction here - deviation from linearity. How did you verify linearity? Just by looking at correlation factor r2 you can be easily tricked. There might be slight deviations at either end of the line when verifying linearity from LOQ to 120% that are hard to see. Did you have a thorough look at the residuals? Did you check the precision of the response factor over the whole range?

- your unknown impurity might have a significantly different response factor than the API ---> Response factors have been defined for known impurities during method validation, how could I define the response factor for an unknown?

You'd need that impurity in pure form, quite hard if it's an unknown .

- some impurities simply might not be visible with your method ---> Impurity is visible, the way of calculation differs. This is my main question, what might the problem be??

Yes, THAT particular unknown impurity is visible, but there might by others that you don't see. So, when using 100% area method, you basically use the wrong value for 100% .

Generally, I'm a strong advocator of using an external reference standard in the same concentration range of the impurities. I'd use %area only for development purposes to get some fast quick-and-dirty values.

ok. thank you so much for your replies.