Alternative approaches for method validation -HPLC for drugs

[Alfred88]

Dear chromatographers,
I would like to bring up these questions to solicit opinion.

For force-degradation studies, is it allowed to use expired drug samples (if available, or samples past accelerated or long-term stability), and use them instead of artificial samples generated by applying stressed conditions? Any restrictions for this approach? Has it been denied by a reviewer/auditor?

For peak purity, can we examine peak shape, and evaluate peak symmetry, plus 1st derivative symmetric to *prove* peak purity? Or this approach cannot be used alone, but in combination with an orthogonal separation, or other approach(es)? In essence, is it possible to demonstrate peak purity/homogeneity without using PDA, MS detectors? PDA is less sensitive than UVD, while MS is not cheap.

Is it possible to combine validation for testing potency/strength, and for testing impurities?

Thank you,
Alfred

[pliva]

Dear chromatographers,
I would like to bring up these questions to solicit opinion.

For force-degradation studies, is it allowed to use expired drug samples (if available, or samples past accelerated or long-term stability), and use them instead of artificial samples generated by applying stressed conditions? Any restrictions for this approach? Has it been denied by a reviewer/auditor?

For peak purity, can we examine peak shape, and evaluate peak symmetry, plus 1st derivative symmetric to *prove* peak purity? Or this approach cannot be used alone, but in combination with an orthogonal separation, or other approach(es)? In essence, is it possible to demonstrate peak purity/homogeneity without using PDA, MS detectors? PDA is less sensitive than UVD, while MS is not cheap.

Is it possible to combine validation for testing potency/strength, and for testing impurities?

Thank you,
Alfred

well , you can use them in my opinion if the samples are really expensive. But with my experience , we usually perform force degradation before the stability starts or in the beginning of the project. So, by the time the long term and accelerated completed, the project will be ready to go for submission and we dont want to wait 6 months for the sample:) After all, commercial research is very much time bound, we have to keep this in mind.

[tom jupille]

For peak purity, can we examine peak shape, and evaluate peak symmetry, plus 1st derivative symmetric to *prove* peak purity? Or this approach cannot be used alone, but in combination with an orthogonal separation, or other approach(es)? In essence, is it possible to demonstrate peak purity/homogeneity without using PDA, MS detectors? PDA is less sensitive than UVD, while MS is not cheap.

I'll give my personal opinion, but what counts is the opinion of whoever audits/reviews the method.

I'm not a great fan of PDA peak purity measurements. I agree with Rob Burgess's post at the end of this thread: viewtopic.php?f=1&t=14138 My recommendation would be LC/MS on two orthogonal separations. MS is not cheap, but it is part of the "price of admission" to the pharma business.

Is it possible to combine validation for testing potency/strength, and for testing impurities?

I don't see any reason why not (assuming you write your validation protocol appropriately), but this would imply that you are using substantially the same conditions for both analyses. My expectation is that that would result in a longer-than-necessary run time for the potency assay (since you don't have to separate impurities from one another).

[Alfred88]

Hi,
I agree with Tom Jupille that LC-MS will be required in pharma, and the acquisition cost should be considered the membership fee to join the country club "pharma biz."
I further propose that as part of confirmation of purity, we should scale up (with preparative column?) the runs, and isolate the peaks to check with LC-NMR for complete assurance.
But we do not operate in an ideal world, and in many cases, the money will dictate what you have and what you can do.

What if you don't have a R/D dept. staffed with scientists to develop new method. And you don't have Validation dept. staffed with scientists to validate the new test.
And you have only analysts who are willing to *run* the tests (and perhaps do *some* validation studies); however, the analysts are required to have a written SOP to follow (per cGMP), and they are required to get training on the methods/SOP (again, GMP).
Then, you are required to be a magician, to deliver the methods (new SOPs) by flipping a hat.

I don't want to cut corners; I would like to comply with all the applicable regulations. But we have to be realistic, because our available resources are limited.
Your feedback about the alternate approach to validation is still needed.

Thank you.

Alfred.

[tom jupille]

I would like to comply with all the applicable regulations. But we have to be realistic, because our available resources are limited.

Alfred, forgive me for saying this, but that sounds very similar to PG&E's position with regard to inspection and recordkeeping on their natural gas pipelines*. A pharmaceutical company operates in a highly regulated environment (appropriately so given the consequences of failure). It's not a case of "we would like to comply . . ."; you *must* comply with the regulations or be shut down. I understand the pressure from management to cut corners, but I also feel that it is part of our role as professionals to resist that pressure when it is inappropriate.

I'm glad I'm not in your shoes!

*for those of you who do not live in Northern California, here is what I'm referring to: http://articles.sfgate.com/2010-12-26/n ... pg-e-blast

[krickos]

Hi

Alfred, it is here where contract laboraties comes to mind. Smaller firms that can not support having an array of high end equipment and the staff to manage it typically employ contracts laboratories for forced degradation studies, analytical procedure development, structure elucidation/qulification of primary reference standards/impurities etcetera.

As for having assay and impurities in one procedure. Yes I have seen that been used. As Tom stated you do not know until you start some initial development if its doeble or practical.

[lmh]

I'd be very, very dubious about using an analysis of peak-shape as proof of purity. Peak shape can prove impurity (a double peak is obviously impure), but it can't prove purity because (a) the impurity might be only weakly detected, in which case it cannot affect the peak; (b) the impurity might coelute exactly.

For the same reason, the PDA purity tests are a bit dubious. Exact coeluters will always be a problem, but even not-quite-exact coeluters are problematic if they have very similar UV spectra to the target compound. Of course the things that are most likely to coelute are those that are structurally very similar, and the things most likely to have very similar UV spectra are those that are structurally very similar...

[HW Mueller]

lmh, a double peak is no proof that the substance is impure.

[Alfred88]

Thanks to Tom J, Mueller, Krickos, lmh!

I just want to share what we got from a recent regulatory inspection.

The auditors noted that for force-degradation tests, we did A, B as stressors, but not C, D, E conditions (per ICH).
They stated that we did not evaluate peak purity; did not specify a target level for degradation in the protocol;
and could not detect and separate impurities when injected.

Lesson learn. More work must be done on our validations. I always consider myself analyst-in-training.

Thanks for reading.
Alfred

[tom jupille]

I always consider myself analyst-in-training.

Welcome to the club!

[lmh]

HW Mueller, apologies, yes, absolutely, I should be more exact in my writing: a double peak can be created by other factors than two things eluting close to one another, but given a system that I know gives single peaks of good shape when using standards standards, and given a chromatogram from that system which contains a peak that looks obviously like the sum of two peaks very close to one another, my first assumption is that there may be two chemicals present. I also feel that in evaluating purity, we should probably err on the pessimistic side.

The point I was trying to make is that it's fundamentally easier to prove impurity than to prove purity.