Chromatography Forum

I would like some thoughts on the following:

We have quantified two working standards against pharmacopoiea CRS and got purities of 100.3 and 100.6%. All working standard and CRS solutions were prepared in triplicate to reduce analytical error.

Would you assign the determined purity values as quantified (i.e 100.6% relative to the CRS) or say the material can't be more than 100% pure and therefore assign 100.0% as the purity?

Thanks

Hi Chris

Our current practise would be to assign a purity of 100,0%.

Do not think I have ever seen an assigned purity or assay above 100% for references within the pharma industry.

I would like some thoughts on the following:

We have quantified two working standards against pharmacopoiea CRS and got purities of 100.3 and 100.6%. All working standard and CRS solutions were prepared in triplicate to reduce analytical error.

Would you assign the determined purity values as quantified (i.e 100.6% relative to the CRS) or say the material can't be more than 100% pure and therefore assign 100.0% as the purity?

Thanks

Generally, the CRS will not itself be 100% pure. They are produced and tested to a defined set of tolerances and, as long as they fall within these values, they are certified. What you are picking up is minor batch differences in materials.

For example, USP Human insulin is described as 100% on the data sheet (USP dont "do " CoA's or show water content) . In fact, it is ~92% when calculated empirically from the IU/mg data.

Quote the answer you measured, for the following reasons:

(1) we're here to measure the content of something, not deduce it by logic.

(2) if we round all measurements down to 100%, how can we quote an error on our measurement? Our s.d. is automatically smaller than it should be, and our mean will always be less than the correct mean.

(3) and how delighted some less-reputable pharma labs would be with a systematic error of +3%! Nearly all their results would conveniently come out at "100% pure".

Quote the answer you measured, for the following reasons:
...
(3) and how delighted some less-reputable pharma labs would be with a systematic error of +3%! Nearly all their results would conveniently come out at "100% pure".

If you've ever worked in the laboratory of a "less reputable pharma" manufacturer, I can assure you they would be far from delighted. They like high precision analyses so they can formulate to the minimum acceptable consumption of expensive API .

If a monograph states acceptable assay of " 98 - 100.5 ", and stability data permits formulation to 98.5%, they would be far from happy about the waste of 1.5% of expensive API.

I agree that you report what you get, however you may want to ensure that you have treated the CRS as defined in the relevant monograph or CRS documentation, which may specify storage conditions and pre-use drying.

Please keep having fun,

Bruce Hamilton

Typically, in a c-GMP environment:

If a API was quantified to be 100.6% pure, you must assign the deduced value as such - Determination of API potency.

When the same API will be used to quantify samples, its potency can be rounded to a 100.0% to facilitate quantitation of samples.

Please, add comments.

Typically, in a c-GMP environment:

If a API was quantified to be 100.6% pure, you must assign the deduced value as such - Determination of API potency.

When the same API will be used to quantify samples, its potency can be rounded to a 100.0% to facilitate quantitation of samples.

Please, add comments.

Can you quote a compendial reference saying it's OK to ignore an assay result , and arbitrarily assigning 100% to the API after having a valid Assay value?.

That's bizarre, can you round up as well as down?. I'd also wonder how regulators would respond to assignation of purity to nominal values.

If the CoA assay was 100.0%, and you got 100.6%, then surely an investigation is called for, depending on the error bars of the CoA assay. If the CoA says 100.0 +-0.5, you've got a problem, and if it says 100+-1, you're OK, but it would not be a good reference compound..

Please keep having fun,

Bruce Hamilton

I have done these API assays quite a while back, and cannot re-collect how exactly we used to report the numbers.

I will have to investigate and dig on this issue, before I can give my response. But that is a good question.