Response factor from degradation experiment

[aceto_81]

Hi,

we have some problems with our mass balance when performing stress tests.
But when we breakdown our API, there is 1 peak which gives always a much higher response compared to my main component.
for example, if I breakdown almost completely my API, I have 2 peaks: my main component and 1 degradation product. Mass balance: 135%.

Is it allowed to setup stress test in such a way that I breakdown for 10%, 20%, 30%, ... take the response for the degradation product and calculate a response factor based on the % degradation of my main component peak?
Do you think this approach is allowed by the regulatories?

Ace

[tom jupille]

If I were reviewing the method, I would not allow it. I would require that the peak be collected, identified, and the relative response factor confirmed (but then, I'm "hard-nosed" about that sort of thing).

[Alex Buske]

You don't have problems with the mass balance, as you don't have the masses. You have a challenge with your assumption that peak area is representative for the masses.
For impurity identification there are limits by the ICH. Just check that the peaks appear in real samples.
For the meantime add the peak ares at different degradation breakdowns and try to apply response factors. Look for the solver function of excel.

Alex

[umoshplc]

I am confusing with your question
Why you need to get the response factor for the degradant? Once your active is degraded at certain test condition, all you need to see is that your main peak of interest is clearly separated from all degradants(specificity) for that you can try peak purity and LC-MS. I hope this is some kind of study you are doing as part of stress study.

[Mattias]

I agree with Tom Jupille. If you want to determine the relative response factor, you will need to identify and synthesise the impurity.

It is another question if you really need the relative response factor. You have shown that the degradation product has a higher response than the main peak. If you use area%, you will always overestimate the impurity (i.e. make it worse for yourself). This kind of argumentation with the authorities is usually very successful.

This only works of course if the impurity is not limiting your stability of real samples.

[aceto_81]

...
This only works of course if the impurity is not limiting your stability of real samples.
...

Mattias, that's exactly the reason for my question.
The peak is in my real samples, and is just on the edge...

@Alex: your solution with the solver function of excel, would this approach accepted by the regulatories?

[Alfred88]

Hi experts.
I am curious about the case when you don't know the response factors, and as a result, you get a big mass-balance deficit (let's say 35%).
How to handle this case, if you cannot resolve with your limited resources. Can two weeks spent on thinking/reasoning be considered sufficient/ adequate effort for an auditor?
Is it common practice to manipulate the "possible" values of response factors of degradants/impurities?

Thanks,
Alfred

[Mattias]

I think the auditor will have major concerns if you have a loss of analyte but you cannot see the corresponding amount of impurities:

1. Are you sure that you can detect all degradation products (are all peaks eluted from the column, do they have UV-absorbance)?
2. Can some impurities be hidden in the front or the main peak?

The only 100% accurate way to prove that your impurity method is OK is to know the degradation products and the response of your degradation products, and by that show that your method has a good mass balance and no peaks are missing.

But I also can see situations where this is not possible or needed.
1. You are analysing large peptides or proteins. The degradation products will by default have very similar absorbance to the main peak (but then you will have OK mass balance already from the start).
2. The degradation pattern is so complex that it is impossible to identify and synthesise all impurities. In that case I think it could be acceptable to run your forced degradation samples with a diode-array detector. Then you can choose the detection wavelength where the mass balance is OK for all stress conditions (typically not the UVmax for the main peak).

The auditors are not impossible, as long as you show that you have done your best and that you have proven that you are not underestimating the amount of impurities. Going back to the original question - I don't think the auditor will be too impressed if you have not identified the degradation product when there is only one single impurity formed...