USP 467 Residual Solvents-Procedure A, water soluble artic.

[pcsox]

What do you do for peaks in the test solution that don't correspond/ overlay with any of the class 1 and class 2 mixtures A & B and are also larger than them, in some cases. Do I identify and quantify them too?

Any input is greatly appreciated

[chromatographer1]

Unless you are testing someone else's product you should be able to know the possible solvents in your article.

These should be identified and quantitated. If you are dealing with an unknown matrix then try to find out from the supplier what solvents were used. Look at the article's chemistry and see if any artifacts could be made from heat, water, and time.

If you don't know what the peak(s) are then you must discover what they are. I assume you just added water to your article (test matrix). Have you determined if the peak(s) in question could be artifacts from the heating and dissolution of your matrix?

Good luck.

Rod

[Schmitty]

If you don't know what the peak(s) are then you must discover what they are.

Rod, do you have a page/paragraph of the USP or other reference which states that you must identify all peaks that cannot be associated with the Class 1 or 2 (A, B or C) solvents?

I am under the impression that the 467 reference has identified the most likely contaminants that are used in the processing of test articles and wants those identified/quantified if necessary.

Thanks

[chromatographer1]

USP <467> p183 last paragraph on page:

Unless otherwise stated in the individual monograph, class 3 residual solvents are limited to not more than 50mg per day (corresponding to 5000 ppm or 0.5% under Option 1). If a class 2 solvent limit in greated than 50mg per day, that residual solvent should be identified and quantified.

I submit that if an unknown is present and responds in a peak larger than the known residual solvents in many cases, then the FDA will require that peak to be identified and quantified. Just try to get approval of a NDA with a large unknown volatile present but not measured or identified !

If you dont know what it is, you cannot say it is less than 50mg per day, right?

best wishes,

Rod

[chromatographer1]

Of course, the USP says you must do a loss on drying and a water content analysis. That would determine if the unknown solvent was less than 50 mg per day.

best wishes,

Rod

[Schmitty]

Well, this may bring up an interesting issue for me. I am doing the analysis for Class 1, 2A and 2B simultaneously by dynamic headspace/GC/MS. If I am only monitoring the ions specific to the above mentioned solvent classes, then there is no way I would know that there are extraneous analytes eluting throughout the run. I'll admit that there will indeed be random peaks on specific ions that may be common to many compunds, but you cannot rely on that as an indicator of alternate solvent contamination.

Maybe analyzing all samples for Loss on Drying prior to the chromatographic analysis is the way to go.

[chromatographer1]

ok, one final postscript:

just remember, at your FDA trial, don't appeal your innocence based on "but the guy on the Forum told me it was ok".

You will go to jail, directly to jail, and you will not pass GO. ( for our non-American viewers, this refers to a popular board game called Monopoly )

Seriously, always ask the USP about interpretations of USP directives.

Everyone, please have a good weekend.

best wishes,

Rod

[Schmitty]

But you're not just any guy...you're a chromatographer(1)!

[chromatographer1]

Schmitty

The USP requires that you use a GC-FID, does it not?

best wishes,

Rod

[chromatographer1]

but the USP allows you to write your own procedures and get them approved. I forgot to mention that. And that is really a good way to go about it. Doing the USP 467 for everything would really be exhausting !

Golly, when you get involved with drugs things can get expensive !

best wishes,

Rod

[Schmitty]

The USP method is to use flame ionization, yes. However, any similar method that is validated is also acceptable. I hope it gets expensive, I am the one who makes the money for analyzing the <467> samples.

In reality, I am working on a pricing structure that will allow for many different charges, based on the extent of the analysis required; If the sample fails the LoD test, then I only charge for that, if it passes the LoD I will go on to the Class 1 and 2A+B screen and charge for that. If there is a "hit" that needs to be quantified, further charges are accrued.

My intention is to turn around as many samples as possible, without sacrificing accuracy and quality, therefore the MS was chosen.

[Bruce Hamilton]

Generally, the identity of all anticipated residual solvents should be known, as the process information should have been used to identify the potential for solvents to carry through to final product, and that information confirmed during product development.

Unidentified peaks from the residual solvent analysis should not be from residual solvents, but could be volatile related substances, or degradation products or other impurities. If there were unexpected solvents, that would trigger major incident investigations, probably including product recalls.

However, in some cases, there could be some degradation of residual solvents ( eg mesityl oxide from acetone ) - but once again they should have been identified and sorted during development. That's about the only time I would see MS as having an advantage - during investigations.

MS may throw up a lot of inforamtion that then has to investigated, as any additional peaks should invoke OOS provisions. The use of static headspace, or direct injection, with FID is usually very suitable for routine quality assurance of pharmaceuticals, and avoids some of the issues that use of MS may invoke from clients and auditors.

Bruce Hamilton

[Schmitty]

Bruce,

All the samples we receive would, ideally, be from suppliers that know the source and process by which the sample was prepared. Based on the number of <467> questions here, I am assuming that many raw material suppliers/buyers are aware of the implications of the new requirements and are performing more testing.

I do not see how I would have significantly more peaks on an MS vs an FID. Gases and siloxanes come to mind, but other than that I should see pretty much the same things on both chromats.

Along with simplifying the <467> standard analysis from 3 injections down to 1, I will provide confirmation on the MS, which a simple RT match on an FID cannot do.

I appreciate this banter, and do not take suggestions lightly. Thank you for your input!

[chromatographer1]

The implementation date for this requirement was recently delayed from July 1, 2007 to July 1, 2008, pursuant to a Revision Bulletin. For more information on the delayed implementation date, see Delayed Implementation Date for General Chapter <467> Residual Solvents. Until July 1, 2008, users must meet only the requirements for Organic Volatile Impurities that are specified in individual monographs.


http://www.usp.org/USPNF/notices/generalChapter467.html

Questions and answers are found here:

http://www.usp.org/USPNF/notices/faq.html

This is from the USP and is quite helpful.