Stability of solution?

[koko]

Hi!
How I confirm the stability of solution?
Which data I need to present? I am not sure if I have to do a t test

[tom jupille]

You prepare the solution and inject samples periodically over whatever time period you want to check. You should get the same response (peak area) every time, and no new peaks should appear.

[mohan_2008]

To further elaborate on Tom's advice:

1. Use a linearity/Medium standard solution that you have prepared for the "intial day zero" validation. I assume that you should have performed a set of five injections to confirm the instrument precision (ICH). Use the "Average" of these five injections as a comparison standard.

2. Store aliquots of the above Medium standard solution at different timepoints as per your protocol.

3. At each time point, take out the solution and inject in triplicate (ICH) and compare area counts to the comparison standard mean.

As your are comparing the area counts/peak response - you need to keep the detector on throughout the length of study.

[HW Mueller]

Is this what the regulators require? Keep the detector on for, lets say a month, if that is the time one chooses? Who guaratees that the lamp doesn´t change, etc., etc.? I understand HPLC as a method which has to be calibrated or checked someway each time one uses it.

[mohan_2008]

Typically,

The time frame for solution stability wraps around 8 days.
This shouldn't be significantly long to affect detector performance due to lamp issues, however,

as per our ICH/OECD guidelines.

[Bruce Hamilton]

as per our ICH/OECD guidelines.

Which ones?. Please be specific about the sections.

That's a bizarre interpretation of most guidelines I've seen.

Besides, the mobile phase and column would appear to be far more likely to change results than the lamp output.

Bruce Hamilton

[JGK]

For tests less than 24 hours I would compare time X against time zero response values.

For tests > 24 hours I would compare responses of the test standard against a freshly prepared STD at each time point, comparing UNIT responses (response per unit concentration) of the test vs "fresh" STDs. testing this way eliminates any differences caused by the system.

If you are using a range of calibration STDs, I would test the highest and lowest STDs at each point.

[mohan_2008]

Hi Bruce,

My answer was a generally enquiry answer to solution stability as performed under - one of our current glp practices under the mentioned ICH/OECD guidelines.

I cannot be specific about the sections since these were interpreted by our sponsor's QA department - and passed around to us. The only thing I know is for sure - they are derived under ICH/OECDs'. I cannot explain the 8-day part since it might be our sponsor's recognition as per their study schedule.

I do not support the scientific validity of this procedure, since you are right and JGK's would be a more plausible solution.

Yes, I forgot to mention but : apart from keeping the lamp on, we need to use the same column (with the same S#) and the same system, with preferably the same mobile phase.

[krickos]

Hi

I have to agree with Bruce Hamilton here. Never ever seen something like that at least in the GMP/pharmaceutical area. and here is the danger ie to separate what actually is stated in a guideline and what someone has interpreteded into their organisation, an easy mistake to make.

From a presonal experiance point of view with the robustness section in the ICH Q2R1 in mind, there is no formal requirement on exactly what data is required. In fact I have gotten away with a ferw sentences in validation reports going to authorities, however internally we typically add tables/diagrams to spot degradation etc.

Like Tom stated early on, evaluate the "time" of intrest ie for sample solutions what do you need? If instrument get a hickup and stops overnight it might be good to know if you just can restart sequence or not so typically just a few days there.

Stock standard/system suitability solutions can be more worthwile to study longer (weeks) in order to save money on standards/solvents and time.

Disclaimer (uhh do not like it but): If Mohan refers to biological samples like serum analysis of pharmaceuticals we are talking about something different (transport of samples from clinical trials to laboratory, stability in biological matrices etc), there might be other guidelines that I am not aware of that cover that particular area of GLP.

So perhaps a misunderstanding