Variations in impurity content of the same batch

[santech]

Hi
I would like to know about determination of related impurities. Does it vary from system to system (HPLC)? Is it required to perform stability analysis, from initial to end of shelf life, (impurities) in the same HPLC system to follow a trend in analysis? If this is true, how does the method conform to validation(Intermediate precision)? A well developed method should behave in any system with little or minimal changes. How much percentage of change is allowed between system to system? Does lamp energy play a role in significant change in related impurities of the same batch of drug product? Please comment.

[krickos]

Hi

I would like to know about determination of related impurities. Does it vary from system to system (HPLC)?

Yes, bluntly put I would say any quantitative analys has a day to day variation. The question is if it is significant/acceptable or not.

Is it required to perform stability analysis, from initial to end of shelf life, (impurities) in the same HPLC system to follow a trend in analysis?

.
No.

If this is true, how does the method conform to validation(Intermediate precision)? A well developed method should behave in any system with little or minimal changes. How much percentage of change is allowed between system to system?

Would not state a general limit as it depends on what level in concentration you do compare at and the individual method. Consider the following example two different persons runs 6 injection at different days on different instrument with different columns. They evaluate largest single unspecified impurity as it turns out it is present at the lower end of reporting limits in for example ICH Q3A (0,03-0,05%).
Analyst 1 gets 6 values of 0,03 and analyst 2 gets 6 values of 0,04, that gives a total RSD (n=12) of what..~15%, and not really suprising. If the RSD were to be calculated on 3 decimal places instead the RSD might drop significantly.
So close to the reporting limit/LOQ I would generally expect a higher RSD for intermediate precision, but it is crucial to consider number of analysis per analyst, how many decimal places and what concentration level to be evaluated to set an acceptance criteria.
For very pure substances/products it might sometimes be better to work with samples solutions spiked with impurities at various specification limits to get a better idea of the intermediate precision.

[Peter Apps]

Surely if a method is to be run on multiple instruments it must be validated on multiple intruments - in other words the inter-instrument reproducibility must be shown to be within the required limits. If a method has been validated on only one instrument, and is then run on another one this would be a major change that requires re-validation of the intrumental parts of the analysis.

And if the resolution / readability (in Krickos' example the number of decimal points) is so low that it contributes significantly to inter-operator or inter-instrument uncertainty then finer resolution needs to be specified in the method.

Peter

[santech]

Thanks for ur reply.
But significant change is reported as +/- 5% in the guidelines. How this will be applicable in case of impurities?
What about lamp energy? will it bring too much change in response factor?

[Bluesman75]

Thanks for ur reply.
What about lamp energy? will it bring too much change in response factor?

If you are worried about lamp energy and/or baseline noise add a sensitivity check to the method.

[Peter Apps]

If samples and standards are run together (i.e. in the same sequence on the same machine) then lamp energy should not have an important impact.

You were not thinking of running standards on one instrument and samples on another were you ?

Peter

[krickos]

Surely if a method is to be run on multiple instruments it must be validated on multiple intruments - in other words the inter-instrument reproducibility must be shown to be within the required limits. If a method has been validated on only one instrument, and is then run on another one this would be a major change that requires re-validation of the intrumental parts of the analysis.
Peter

I agree, I just did not think beyond having 2 identical qualified instruments used in intermediate precision. In the case of having different instruments and if you want to be able to use both/all that have to be accounted for in a validation but also when introducing/verifying suitability of a compendial procedure like a pharmacopiea procedure.

[Peter Apps]

[quote="santech"]Thanks for ur reply.
But significant change is reported as +/- 5% in the guidelines. If you need to relate it to these guidelines I would say that a change from one instrument to another is a 100% change in instrument quote]

[santech]

If samples and standards are run together (i.e. in the same sequence on the same machine) then lamp energy should not have an important impact.

You were not thinking of running standards on one instrument and samples on another were you ?

Peter

Nobody thinks of running standards and samples on different instruments. How about the impurity analysis performed in one system having lamp energy low and another high

[santech]

Thanks for ur reply.
But significant change is reported as +/- 5% in the guidelines. If you need to relate it to these guidelines I would say that a change from one instrument to another is a 100% change in instrument quote]

I am talking about the content of impurities and not the instrument

[HW Mueller]

We used to call this calibrating a system, and everybody seemed to know that it was absolutely necessary to do this. It appears that some people don´t know what validation is. Now back to calibration: If you did it properly you get the right answers from any properly calibrated instrument. That means that an instrument with high energy lamp gives the same (correct) results as one with low energy. Now Peter already told you that.

[Bluesman75]

Thanks for ur reply.
But significant change is reported as +/- 5% in the guidelines. If you need to relate it to these guidelines I would say that a change from one instrument to another is a 100% change in instrument quote]

I am talking about the content of impurities and not the instrument

If you are worried about instrument sensitivity then include a sensitivity check as part of your system suitability, based on the lowest level you need to be able to quantify. You prove that the sensitivity level is valid as part of your validation. Then run this as part of each analysis. This demonstrates that the instrument is giving enough sensitivity to provide reliable results.

[tom jupille]

A properly validated method should include system suitability criteria. These are tests that may include things like retention times, plate counts, resolutions, baseline noise/drift, and repeatability measurements and are run on the instrument you will be using immediately before you run your standards and samples. If your system passes system suitability, it is (by definition) suitable for the intended purpose. so, if lamp energy is too low, then you will likely fail baseline noise or repeatability, and if you pass baseline noise and repeatability, then your lamp energy is adequate.