Forced degradation and wavelength selection

[Mattias]

I have just finalised the forced degradation study of 5-ASA, and the results were that the drug is degraded only at high pH and in contact with oxidants.

When about 15-20% of the substance is degraded, the chromatogram is full of unknown peaks (50-100). Now, the mass balance is very different depending on which wavelength I am looking at (I can extract single wavelengths from DAD-data).

Would it be good practise to chose the detection wavelength of the method where the mass balance is OK? I have no idea what all these peaks really are. Different peaks must be over and underestimated at different wavelengths (using area% of total peak area).

[HW Mueller]

You got your answer already:
"Different peaks must be over and underestimated at different wavelengths (using area% of total peak area)."

[Mattias]

I understand that my original question was of the same type as "how long is rope".

In the ideal world I would have all peaks identified and synthesised. I will just stick to the wavelength maxima of the main peak, and live with bad mass balance. It must be just as correct as anything else.

[HW Mueller]

You mean it should be as equivocal as just about everything else?
Whatever you do, please don´t call it a mass balance if it isn´t.

[grzesiek]

"When about 15-20% of the substance is degraded, the chromatogram is full of unknown peaks (50-100). Now, the mass balance is very different depending on which wavelength I am looking at (I can extract single wavelengths from DAD-data). "

what's going on when about 5-10% is degraded? How do you measure how much is degraded? there are too many peaks, you can forget about doing mass balance there

I'm assuming that you are developing the method, how many known impurities you already have? What wavelenghts do you use for them? What wavelenght for main compound? What is the mass balance at these wavelenghts?

[Mattias]

The recovery of the main peak is determined by external standards. The mass balance is then recovery of main peak (% of nominal) + amount of impurities (area% of total area).

I don't have ID of any of the degradation products at this point. I know that many of them are oligomers of 5-ASA, having very different UV-spectra due to "resonance".

I have an internal guideline that states that one should aim for about 20% degradation in forced degradation studies. At 5-10% degradation, everything looks about the same, the peaks are just smaller.

I don't know what is the right thing to do here? I get very good massbalance with oxidative degradation, but poor with alkaline degradation (20% of the main peak gone - 5% impurities). I have used 232 nm so far, one of the UVmax of 5-ASA. At this (relatively low) wavelength, I can at least see all impurities. At 300 nm (the second UVmax), some of the impurities are not detected at all (but the "mass balance" is still better).

[grzesiek]

"area% of total area" - I don't think it is a good idea, main peak dilution would be more resonable

"At 5-10% degradation, everything looks about the same, the peaks are just smaller" - better mass balance at 232nm?

Did you run MS/ELSD?other to see if something's missing?

"that one should aim for about 20%" - does not mean that one succeds

mass balance can not always be achieved, there are reasons why one does not get mass balance

[tom jupille]

As grzesiek suggested, this is a situation where one of the "evaporative" detectors where response is much less dependent on structural details might be much more useful than UV.