HPLC ID tests

[AdrianF]

We routinely use retention times as an ID test for pharmaceuticals. we also carry out another chemical or spectroscopic test. Some of these tests are extremely time consuming and seem to me to be a waste of time.

What chance is there for a substance to have the expected RT , give the expected assay result and be something else? I have never known anything to fail an ID test.

Do others have this belt and braces approach? Is it a regulatory requirement?

[DR]

In pharmaceuticals, you should have a pretty good idea how well a method works before you use it on routine production lots. Validation protocols, if well written and passed, should prevent any nasty surprises from cropping up. If you're using a validated method for its intended purpose and are getting the right sized peaks at the expected retention times, doing orthoganol ID tests is a waste of time.

That does not remove the responsibility for doing ID tests on incoming raw materials before they are used in production, however.

[bartjoosen]

If you use HPLC/PDA, I think it's allowed by ICH guidelines to use this combination to prove identity.

[Consumer Products Guy]

We routinely use retention times as an ID test for pharmaceuticals. What chance is there for a substance to have the expected RT, give the expected assay result and be something else?

Very remote.

Concerning "belt and braces approach": in US, we call them "suspenders", like in "Once Upon a Time in the West" Frank (Henry Fonda) shot Wobbles through the belt buckle and both suspender buckles, stating "you can't trust a man who wears both belt and suspenders - the man can't even trust his own pants".

[Bruce Hamilton]

I disagree somewhat about the waste of time, the requirement for independent ID tests goes back to pharmacopoeia from early last century - long before HPLC tests. If you think modern ones are tedious, try sodium fusion and group separation.

Note that ID tests are usually part of a monograph, and may be used without other tests ( assay, related substances etc. ) , eg for Inwards good testing by downstream formulators etc.,

As a QC analyst, I had several chemicals fail pharmacopial ID tests, and even had raw materials rejected because their labelling/quality was incorrect - as detected by the ID tests, and later confirmed by further work.

Published pharmacopoeia monographs often require at least two discrete ID tests from a selection of several ( eg "one ID test from A or D, and one from B, C, E "), In a few monographs, specific assay tests ( eg GC FAMEs ) may also be the only required acceptable ID.

You haven't said what stage of the pharmaceutical industry you are playing in, but let's investigate a few scenarios. I'm ignoring the requirement that as you developed your production, you should have also validated the quality procedures of manufacturing to ensure they met ICH Q-series requirements.

You're synthesizing an API. You hope you have made 2-OH- gobblydegook, but if you didn't, you may have made 2,4-diOH-gobblydegook, which could have very similar HPLC chromatographic properties ( give acceptable assay ), or made two components - one with double the chromophore, 2-OH-digobblydegook, and the other without the chromophore, 2-OH-degook. You're unlikely to have made 2-0H-jargon, which would have different retention times.

Your producing raw materials, the factory produces antibiotics ( sulphonamides, penicillins, or tetracyclines, ), and holds stocks of product in stores. Your ID tests must discriminate potential mislabelling and cross-contamination. The HPLC conditions may not resolve or detect all the potential contaminants. The ICH Quality guides allow customers to accept your CoA data as one of the ID tests, so quality testing has to be robust.

You're formulating products. Your inwards good have all the raw materials that meet compendial standards. You have to confirm that each container ( statistical sampling ) is true to label. ID tests alone may be used - without the assay. You can even use the supplier's CoA as one ID, provided their quality have been audited by your quality people.

ID tests serve a very specific role in a monograph, and that role is used at different stages in drug manufacture. You could justify not performing ID tests for in-house use, but if you are using compendial monographs, you have to comply with their requirements, and if you are making APIs for others to use, you still have to confirm that the ID tests alone are adequate for each batch of product.

If all systems were perfect, there would be no need for any ID testing, the industry would just use the original CoA, but almost all processes involve humans...

Please keep having fun,

Bruce Hamilton

[Dan]

In the pharma places were I have been, the QA/RA folks have been telling us that the EMEA and FDA want two (orthoganal) ID tests. Although I have not seen any regulatory guidance requiring this. But we analysts take the word of the QA/RA folks and provide the two tests.

We have done HPLC/PDA as mentioned above and it has been accepted. You have the retention time ID and the UV spectra ID. This doesn't add much in the way of taking time for a second test.

At one place, our ID tests were required to be semi-quantitative. The idea was to be able to distinguish between the 20 mg and the 50 mg tablets. The purpose was to be able to check for mislabeling and cross-contamination as Bruce noted.

Regards,
Dan

[Bruce Hamilton]

Some general ID testing requirements for cGMP are defined in Q7 and Q6A, but there may be more in regulatory guides and pharmacopoeia.

My experience was that our Quality groups interpreted Q7A as at least one ID plus a close review of the CoA from an audited supplier, but most of our clients wanted a mininum of two IDs plus the close review of the CoA for audited suppliers, and sometimes full specification for all batches from unaudited suppliers.

Q7

7.3 Sampling and Testing of Incoming Production Materials

7.30 At least one test to verify the identity of each batch of material should be conducted, with the exception of the materials described below in 7.32. A supplier's Certificate of Analysis can be used in place of performing other tests, provided that the manufacturer has a system in place to evaluate suppliers.

Q6A

3.2.1 New Drug Substances

b) Identification: identification testing should optimally be able to discriminate between compounds of closely related structure which are likely to be present. Identification tests should be specific for the new drug substance, e.g., infrared spectroscopy. Identification solely by a single chromatographic retention time, for example, is not regarded as being specific.

However, the use of two chromatographic procedures, where the separation is based on different principles or a combination of tests into a single procedure, such as HPLC/UV diode array, HPLC/MS, or GC/MS is generally acceptable. If the new drug substance is a salt, identification testing should be specific for the individual ions. An identification test that is specific for the salt itself should suffice.

New drug substances which are optically active may also need specific identification testing or performance of a chiral assay. For a drug substance developed as a single enantiomer, the identity test(s) should be capable of distinguishing both enantiomers and the racemic mixture.

For a racemic drug substance, there are generally two situations where a stereospecific identity test is appropriate for release/acceptance testing: 1) where there is a significant possibility that the enantiomer might be substituted for the racemate, or 2) when there is evidence that preferential crystallization may lead to unintentional production of a non-racemic mixture.

[ End Extract ]

Please keep having fun,

Bruce Hamilton

[AdrianF]

Thank you for your replies - I should have made the scenario clearer. I am referring to tablets made under our control in another EU country. We have acccess to all the manufacturing documents. We carry out full testing at our UK facility.

When manufacturing was carried out on this site we didn't carry out any ID tests.

Re -braces/suspenders as someone said 'Two countries separated by a common language'!

[Bruce Hamilton]

I don't understand your problem.

The manufacturing site is required to comply with Q7 ( at least one ID + CoA review, followed by Quality group release ) at the site before the tablets were manufactured. The only exception are raw materails like toxins etc. I'm not sure whether unique tamper-proof containers could give some respite, but I suspect not - especially if materials are shipped off-site.

To detect contamination/mistakes, the tests are required to be performed on samples taken from the manufacturing site before the material is released to manufacture or customer, and efficiency is one reason to establish a QC lab on a manufacturing site.

As a client, you would be expected to audit the maufacturing site, and that would normal include some analytical correlation or check by an independent/separate lab.

The regulators don't require complete duplication of testing, just some sort of systematic and statistical quality assurance. I don't see any belt and braces - unless both sites are duplicating all tests, as the reasons for the tests would appear to be quite different.

You are testing the materials/products to see if they are suitable for your organisation to accept and use ( conform to your specifications and generating data on an approved supplier ).

The manufacturing site is testing to ensure the actual containers they are using and/or manufacturing contain what their labels state. If you pay them enough, they presumably will also perform full specification testing, which means assurance of complete comformance of the products at their site to your specifications

Please keep having fun,

Bruce Hamilton

[AdrianF]

This is another case where a few minutes face to face would clear up any misconceptions.

This is the problem:

1. Product is made in country A with all the normal GMP at a plant which is part of the company I work for.

2. It is then shipped to UK where we carry out full specification testing.

3. We currently carry out two ID tests one of which is usually HPLC retention time.

4. The second test is often very long winded - one for instance requires extraction, decolourisation with charcoal followed by IR and Melting point which can take a whole day.

My question is why should it be necessary to carry out these 2 IDs? When we manufactured on the same site as we tested it we didn't do any ID tests.

[Bruce Hamilton]

OK, I'm expert at misperceiving, so the following may also be off the mark.

I assume that your two ID tests are part of the "full specification" testing that you perform. If so, two ID tests are just part of the consequences of performing the full specification testing.

Your full specification test samples ( presumably from the statistical sampling on arrival at your site ) should be different to the release test samples at the manufacturing plant.

You would be aware that it's usually not acceptable to combine results from one set of samples with results from another set of samples to generate a CoA for a batch. Resampling usually results in retesting, or a lot of trees have to be felled to explain why not.

I'm assuming that when you manufactured the product at your site, the intermediate stages were held in quarantine ( no ID required ), and the sampling for release would also have been used for final product testing.

If you are performing Full Specification testing, and also performing additional ID testing, presumably on different samples, that testing may be something to do with other products that are also held in the quarantiine or release store, or perhaps if you are modifying the packaging in any way, but you should ask why.

Please keep having fun,

Bruce Hamilton