Chromatography Forum

Hi there,
I have to perform accuracy of drug substance. I know you can do it against a compendial procedure or you can infer accuracy from linearity and precision.

Can someone help me with the third approach. Can accuracy be performed at three separate levels in triplicate with the same reference standard that you use to perform linearity. Or should you use a USP reference standard if your linearity was performed with house standard.

Hope someone can clarify

Thanks

Scio

In that case, how would the accuracy differ from the linearity/range? You would just show that you can recover the same fraction (f. x. 70% - just for the sake of the argument) of the substance at different levels. Is the method accurate or just precise?

Best Regards

Hi Danko, thanks for replying
Ok if we use a secondary standard for the linearity can we use the USP standard, or USP drug substance for the validation to measure accuracy in triplicate.

Hope you can clarify, thanks

Scio

Yes Scio,

If no specific mention is made in your protocol:

You can use the linearity standard (s) as the reference standards for accuracy. But, make sure that all accuracy concentration (s) are bracketed within the standard range.

We use three of the five linearity standards (80120% of nominal) to bracket (determine) the accuracy concentrations. It is acceptable under all valid regulatory guidelines.

Hi Scio,

I’m not sure what mohan_2008 states by:

You can use the linearity standard (s) as the reference standards for accuracy. But, make sure that all accuracy concentration (s) are bracketed within the standard range.

But calibrating with a given standard and subsequently quantifying the very same standard/solution is not going to confirm the accuracy of your method – no matter whether it’s at one or several levels. You might as well inject the standard twice and if you get the same area for both injections conclude that you’ve recovered the whole amount thus proved the accuracy. But how will you know that the value is correct?
The way it should be done is utilizing some kind of certified standard/substance (USP or whatever relevant) and obtaining the expected/certified value.
In situations where there is no such a certified standard (f. x. when you’ve synthesized or isolated a new substance) you should confirm/make the alleged value plausible using at least one orthogonal method.

Best Regards

I agree with mohan

Danko's right that you should do your best to measure your standard purity/potency

mohan, grzesiek, are you suggesting that one calibrates the HPLC with a standard, then run this same standard to check whether results are accurate? What is accuracy then?

The issue is covered in ICH Q2 Validation of Analytical procedures:
Text and Methodology.

If you want to infer accuracy,
" accuracy may be inferred once precision, linearity and specificity have been established. "

Otherwise you have to determine it - more details in the ICH guidance.

The issue may be the standards used to make the test solutions, as you are going use those solutions to confirm "trueness", so the standards used have to have been separately verified ( eg compendial reference substance ), and the solutions also have to represent your proposed samples.

If your linearity solutions are prepared from suitable substances, you could use some of them, but you should always be aware that "accuracy" requirements ( independent confirmation of the trueness ) are different to those for linearity, hence the need for the additional precision and specificity tests if you wish to infer accuracy.

Bruce Hamilton

Okay, Bruce, what *exactly* do we have to do to "infer" accuracy?

accuracy may be inferred once precision, linearity and specificity have been established.

I've always suggested that recovery from a spiked matrix was the most meaningful surrogate for accuracy. And that when no blank matrix is available, that linearity and precision of standard additions would have to "make do". But danko makes a good point: how do we justify the leap from "it's precise" to "it's accurate"?

Okay, Bruce, what *exactly* do we have to do to "infer" accuracy?

accuracy may be inferred once precision, linearity and specificity have been established.

I've always suggested that recovery from a spiked matrix was the most meaningful surrogate for accuracy. And that when no blank matrix is available, that linearity and precision of standard additions would have to "make do". But danko makes a good point: how do we justify the leap from "it's precise" to "it's accurate"?

Refer the ICH guidance...
Section 1 discusses requirements for Specificity.
Section 2 discusses requirements for Linearity
Section 5 discusses requirements for Precision

The extract that I provided was intended to indicate that Linearity and Precision ( as listed by the original poster and others ) were inadequate to infer accuracy, you also have to demonstrate specificity.

I was not suggesting that experimental determinations are inferior, just that people seem to be forgetting about the essential specificity requirement that complements the precision and linear requirements to obtain what used be known as "trueness". If the analysis is precise, linear, and specific, is it not reasonable to infer it's accurate?.

The Guidance also clearly points out the need for independent procedures when available materials are inadequate.

Perhaps I should have posted more of the Accuracy section, so here is...

[ Begin extract ]
4.1. Assay

4.1.1 Drug Substance

Several methods of determining accuracy are available:
a) application of an analytical procedure to an analyte of known purity (e.g. reference material);
b) comparison of the results of the proposed analytical procedure with those of a second well-characterized procedure, the accuracy of which is stated and/or defined (independent procedure, see 1.2.);
c) accuracy may be inferred once precision, linearity and specificity have been established.

4.1.2 Drug Product

Several methods for determining accuracy are available:
a) application of the analytical procedure to synthetic mixtures of the drug product components to which known quantities of the drug substance to be analysed have been added;
b) in cases where it is impossible to obtain samples of all drug product components , it may be acceptable either to add known quantities of the analyte to the drug product or to compare the results obtained from a second, well characterized procedure, the accuracy of which is stated and/or defined (independent procedure, see 1.2.);
c) accuracy may be inferred once precision, linearity and specificity have been established.

[ End extract ]

Please keep having fun,

Bruce Hamilton

Bruce, you did not answer Toms question.

"The extract that I provided was intended to indicate that Linearity and Precision ( as listed by the original poster and others ) were inadequate to infer accuracy, you also have to demonstrate specificity." - I for example assumed specififcity, when inferring accuracy from linearity, donno how others do, but I would expect specificity is first demonstarted (before going to linearity/accuracy studies)

"Bruce, you did not answer Toms question." - I think it has been answered in "so the standards used have to have been separately verified ( eg compendial reference substance )"

But it says: "Several methods of determining accuracy are available:"

So how do you "infer" in method c)?

But it says: "Several methods of determining accuracy are available:"

So how do you "infer" in method c)?

I think I'm missing your point, because the obvious answer is:-

If the individual results for determining Precision, Specificity, and Linearity ( as defined in the appropriate sections ) are acceptable, then it is also acceptable to infer accuracy, rather than experimentally determine it by separate experiments.

However, if your protocols do not require you to perform all of those tests, you obviously can not infer accuracy, but have to determine it, using techniques indicated in a) and/or b).

Bruce Hamilton

"The extract that I provided was intended to indicate that Linearity and Precision ( as listed by the original poster and others ) were inadequate to infer accuracy, you also have to demonstrate specificity." - I for example assumed specififcity, when inferring accuracy from linearity, donno how others do, but I would expect specificity is first demonstarted (before going to linearity/accuracy studies)

.

I agree, however the Q series covers all ICH testing, and whilst Specificity is normally evaluated, there is an alternative option available in Q2 :-
" (2) lack of specificity of one analytical procedure could be compensated by other supporting analytical procedure(s) ".

Remember that there is considerable flexibility in ICH guidances, and my point was that the apparent agony over whether accuracy can be determined solely from Precision and Linearity should disappear when the explicit Specificity requirements are added to the mixture.

Please keep having fun,

Bruce Hamilton