Validation of GLP analytical methods (m)

[domino1]

I am trying to get my head around the issue of "validating" analytical HPLC methods for potency/purity determination of GLP dosing solutions.

As an analytical chemist, I associate method "validation" with the ICH guidelines. It is something that you do prior to registering your drug (meaning when you are approaching Phase III clinical trials).

A GLP consultant has said that you must "validate" your methods used to support GLP studies. These studies are being conducted before the compound even makes it into Phase I studies.

It does not seem reasonable to validate a method, in the traditional sense, at this state. The method will evolve over the course of the life of a drug and is by no means the final method. Additionally, excipients and dosage forms change dramatically as the GLP studies progress so it would be unreasonable to "validate" for each one.

Could you all give me your perpective on validation of analytical methods for GLP studies? What have you done? What have you heard from other consultants etc?

I fear that the term "validation" is being thrown around when people do not understand the impact to an analytical chemist.

Thank you for your time.

Lisa.

[Consumer Products Guy]

Congratulations upon entering the state of confusion (GLP). I've found that lots of consultants "consult" because they can't really do anything, or got downsized from real jobs. That said, my company has both GLP and cGMP products, and remember that validation is justifying that a certain procedure is "valid" or realistic, for the job at hand. I agree, though stuff like linearity and possibly recovery may not change with your project, stuff like limit of quantitation and detection may vary with matrix changes. Right now I'm battling my pointy-haired boss that ICH, cGMP, and GLP have real, defined criteria that must be met to consider a test procedure "validation", not our company statistician. With GLP you'll need a bunch of government-mandated SOPs, and labels stating such and such instruments are blessed for GLP studies, etc. Our so-called "GLP expert" hired a consultant to "teach" everyone GLP for 1.5 days (when they weren't sleeping), and I got in trouble when I asked why a highly-compensated expert needed to hire in someone, instead of going through the lousy Powerpoint slides himself. Then, there's the QA department: don't get me started.......

[domino1]

Well, it looks like you have been caught up in similar frustrating circumstances!!

I have to think that the consultants are confusing "validation" of bioassays with our ICH defnition of validation.

I feel like we would be doing a big disservice to our clients if we made them pay for a validation of their method before we analyzed GLP samples. A true validation would take a significant amount of time and money. If we claim it is stability indicating, then we would also be doing stress studies etc to determine specificity. That is more time and money. We most certainly will verify that a method is linear and precise, etc, however this will not be done under a protocol.

I would like to hear some real world stories where the FDA came in and expected a full validation of a GLP method. Most of these early stage companies are just not equipped.

The sad part is that most of these drugs will die before they ever get to the NDA stage and I hate to see companies spending a lot of time and money on something that they didn't need to do!

Thanks again,

Lisa.

[josebenjamin]

Dear domino1,

I identify a lot with all your comments and doubts. I feel the key to understand all this is that in the product development process there is room for different levels of validation. As you say it, it makes no sense fully validate methods for products that may not survive Phase I or Phase II stages.

In the early stages I think everybody would agree that a minimum of validation should be done, perhaps some test of specificity and a limited linearity study, plus duplicate samples to get a feeling for precision or accuracy. This approach would also let you develop enough familiarity with the method, and will teach what changes could be adequate once the full validation is done (Phase III, NDA documentation).

I always believe that a method can be valid without being fully validated, but in these circumstances you could not defend the method before a jury, and you do not have to.

Do not let the consultants intimidate you too much, take their words with a grain of salt.

Good Luck,

josebenjamin