Robustness- Solution Stability

[NVG]

Hi, I need some help. While testing Stability of Standard/sample solutions as part of robustness, which document needs to followed to set limits. I tried looking in ICH Q2 (R1) and USP 1225 but wasn't of much help. I am confused as in variuos documents I have seen limits to be - %diff between Tn and To is nmt 2%and 1% or absolute value is +/- 10% of To value. So I am confused. Also sometimes people use 2 solutions and sometimes 1 solution only for the shelf life study. Could you please help me as to what is right way to go while determining shelf life of Standard/sample solutions.

Regards
NVG

[Dan]

NVG,

There are a variety of ways to proceed. I don't recall any ICH/USP/FDA reference for how I've done it; I will post a reference later if I find one. Meanwhile here are some ideas/suggestions.

Don't use peak area for comparison from Tn to T0. You need to actually analyze the sample and standard solutions (treat an injection of standard as if it were a sample). I have seen instances where people use peak areas for the standard solution without realizing that peak areas vary for a number of reasons from day to day. They usual use peak area because they forgot to anaylze the standard solution like a sample on T0.

Try to use the same vials for every time point. Refiltering from the original sample prep can add variability, giving you an uncontrolled experiment. However, be aware that there may be evaporation from the sample vial; you can recap the vial with a new septum if possible.

For an active ingredient, you can use a +/- 2% relative or absolute limit. For some actives, the limit may be more or less than 2%.

With an absolute limit, use Tn - T0.

With an relative limit, use (Tn - T0)/T0 X 100%.

For impurities and trace level analytes, a +/- 10% limit is used. However, you can also use a relative limit; example: +/- 0.02% w/w on a relative basis.

Also, there should be an evaluation to see if any extra peaks show up in the stored solutions. The active may degrade by 0.5% and still meet your +/- 2% limit, but it can mean that there is 0.5% of an impurity somewhere and so your solution is stable for the assay analysis but not stable for impurity analysis.

Your choice of limits will depend on your method's ability, your analytes and what your QA group will let you get away with.

Lastly, usual the solutions are stored at ambient conditions, but, for some analytes, you may want to consider storage at 5 degrees C.

Regards,
Dan

[Dan]

Here are a few more comments and clarifications from my previous post:

1) The choice of the value for a protocol acceptance limit will depend on parameters such as the type of analysis (assay, impurity, trace analysis, etc.) and the analysis technique (HPLC, CE, titration, etc.). Therefore, using different limits for different methods that are being validated is an acceptable, and common, practice. For the validation of an HPLC method for the assay analysis of a pharmaceutical drug product, a commonly used limit is +/- 2%. This is generally considered the obtainable accuracy for an HPLC assay method. Yes, HPLC can be better than that, but so far regulatory agencies have accepted the 2% limit and, although there have been pushes to tighten the limit, 2% is still common in this instance.

There should be consistency within and between validation protocols. If 2% is the limit for the solution stability in the protocol, then the protocol would also have 2% limits for the for the accuracy and precision experiments. Any SOP for method validation should contain typical values that are to be used (the typical values can be in and associated document such as a Work Instruction rather than in the SOP itself).

2) It is possible to store solutions at more than one storage condition but only test the solutions that have been stored at one only condition and not the other. The reason for this is that the second condition is a back-up storage condition. So, you keep solutions at ambient and chilled (i.e. refrigerated) but only test the solutions at ambient temp. Why? Well, you may not be sure that you solutions are stable at ambient but you don't want to repeat the experiment with a different storage solution. You would only do this if you suspect the solutions are not stable at ambient, otherwise you don't need the extra work. Whatever is done, the procedure for the experiment would be described in the method validation protocol. There would be a statement to the effect that the refrigerated solutions would only be tested if there were failures at the other storage condition.

3) There are two equations that can be used to calculate a relative difference:

A) (Tn - T0)/T0 X 100%

B) (Tn - T0)/((Tn + T0)/2) X 100%

Equation A would be used for a solution stability experiment as you need to compared the result for the stored solution to that of the initial (time zero) solution. For Equation B, you are making the calculation relative to the mean of two results. Thus, Equation B is NOT appropriate for the solution stability experiment. However, a variant of Equation B is appropriate for the precision experiment:

B) (R1 - R2)/((R1 + R2)/2) X 100% [variant]

In Equation B variant, R1 and R2 are the mean results for analyst 1 and analyst 2. As you are not sure which analyst has the true result, the equation uses the mean of the two results in the denominator.

When you have results that are around 100% (of label claim), then there is little difference in the calculated answer if either Equation A or B is used (i.e. the two equations give nearly the same answer). However, there is still a technical difference between the two equations even in the calculated results are nearly the same.

Note: For a method transfer, Equation A should be used with T0 being the result for the transferring analyst and Tn being the result for the receiving analyst.

4) Documentation, documentation, documentation! Keep your SOPs, Work Instructions and Method Validation Protocols complete. In your method validation protocol, state how you are to perform the experiment, what the acceptance limits are and how the results are to be calculated. Of course, your protocol should follow the procedures outlined in your SOP.


Regards,
Dan

[Sunjay]

Dear All,

There is a referance for the determination of solution stability. In the pharmocoepial forum USP has given a chapter on Dissolution method development and validation in which they have specified the procedure and criterial for solution stability of dissolution saamples.

Regards

[NVG]

Thanks Dan for the reference, I read it and got few questions:
Can these criteria be applied to:
tests other than Dissolution, e.g Assay, since dissolution tests limits are not bracketed.
techniques other than UV/VIS assay i.e. HPLC analysis.
I have seen in some documents that if solutions are found to be stable for, lets say 12days, the method mentions solution stability to be 2/3rd of that period i.e. 8days.
What are your thoughts on that?

[Dan]

Sunjay,

Thanks for the reference. I haven't looked at that USP chapter in a while and almost forgot about it. That chapter should be official in the USP 30. I don't have it available but, if I remember correctly, it is <1092>.

NVG,

You can perform the solution stability validation experiment for different test types: assay, dissolution, impurity, etc. and for different measurement types: HPLC, UV/Vis, etc. The acceptance criteria will most likely change in each case. See my earlier comments for examples.

I don't know why people would sell their methods short. If we find that the solutions are stable for 12 days, then we put that in our method. Perhaps they have a reason for using a shorter time period than they found in the method. It may have to do with the way they performed the experiment. Perhaps they just are being cautious. I don't know for sure.

Regards,
Dan

[NVG]

Thanks Dan. I understand that solution stability studies can be performed for different tests and techniques, what I meant was that can the acceptance criteria mentioned in the USP chapter you mentioned, can be applied/extended to other tests and techniques.