Impurity reporting by HPLC

[shadow]

I have three samples, each sample contains two APIs (A & B). Each sample is of different strength for A but is the same for B i.e. 1) 5A/B, 2) 10A/B, 3) 15A/B. Samples were prepared to keep B at the same concnetration but A was different. I ran them all in the same sequence and now I need to report the impurities either by quantitation or by area normalization. When I do area normalization the %area of the i mpurities gets affected by increasing are of A and it does not seem right. Wat is the best way to report impurities in this scenario? Thanks for the help.

[danko]

When reporting/calculating purity/impurity of A, B and B related peaks should be excluded/ignored and vice versa when calculating puriry/impurity of B.
So, you should end up with 2 results each related to the specific API.

Best Regards

[shadow]

Thank you for the reply. How about unknowns that were not observed during forced degradation but were observed from accelerated stability of the drug product, I do not know if they are related to A or B, and If I needed to quantitate the impurities present against a reference material of the APIs, how to do it? Any help would be really appreciated.

[danko]

It is a prerequisite to know which peaks are related to either A or B. I would expect some kind of prior investigation as to which main compound is degraded to which degradants/peaks. Meaning before deciding to formulate them together. If not then it’s something that has to be to carried out – just like one investigates the matrix/placebo for potential contribution to appearing peak under stability studies.

Best Regards

[shadow]

Thank you Danko. The problem is that I have some unknowns and I do not know if they belong to A or B. So if everything is perfect I would quantitate the known impurities to the respective API reference standard but what to do with the unknowns? who to calculate them agains? that is my dilema. Thanks for the input.

[HPLCaddict]

Pragmatic approach: Evaluate the unknowns against the API with the lower response, i.e. the API that gives the smaller peak. This is the worst-case evaluation.

[shadow]

Thanks HPLCaddict for your input, would this work in a QC environment? probably not uh?. If any of you guys have an example of an article that describes an impurity method for dual actives using the same wavelength I would really apprciate you sharing it with me, or any suggestion on how to make this method work I would be really thankfull. Regards

[danko]

I'm afraid you don't just accept unknown peaks of certain size and beyond in a regulated invironment.
You need to (read have to) know what these degradants are. There are some differences in requirements but as soon as you reach 0.5 - 1% for a single peak, you'll have to investigate it a bit more than just say: Oh, I see a peak I know nothing of.
I was thinking for instance of accelarated stability for each of the components for them selves and not only for the mixture.
Then if nessessary some characterisation - at least utilizing PDA detection and preferably MS and so on.

Best Regards

[HPLCaddict]

As per Danko's advice: It depends on the amount of those impurities. Refer to the relevant ICH guidelines.
I'd suppose if you take the worst-case approach (using the smallest API peak for quantification) and those unknowns are still under the identification threshold (usually 0.2 % for a finished product, but have a look at the guidelines, this may change according to the daily dosis), then you're fine and no further work involved. If it's above the identification threshold, well, you're in trouble...

[gtma]

I agree with HPLCaddict. You should consider the worst case. It will depends on your wavelength/response. This can be very tricky. Try to obtain the UV spectra of the impurities and see if you can associate the impurities to one of the APIs. You may need mass spec data. I hope your compds are very stable.

[shadow]

Thank you all for the advise. Regards