LC-MS/MS drug degradents identification

[meillid]

I used to use LC-MS/MS (API 2000,4000) to do quantification of drug. Currently, we have a project to identify unknown degradents in the drug. After several tries, I found by using only Q1 it is very hard to detect any useful signal beside the huge buffer signals from the mobile phase although the unknown degradents showed clear peaks on the UV chromatography. I just wonder how do you work on this kind of project? What do you do to identify the unknown degradents in the drug? If LC-MS/MS (API 2000,4000) is still useful tool to identify the unknowns, what is the proper to do the experiments to get useful information? Thanks a lot.

[james little]

I use mainly use Micromass instruments, so not so good with Sciex software. On the Micromass, have to make sure UV is lined up with MS signal by setting offset.

Then use the UV signal to determine where to average the signal and subtract background in the mass spec signal.

Many times have to inject much more material to detect small impurities. If impurities elute on the back side of large tail from product, we divert the majority of the product and start right before peak of interest. If you get too much material in the electrospray probe, tends to tail forever.

In injecting more material, best to find a solvent that dissolves the material at high concentrations and inject smaller volumes. We use dimethylformamide a lot, injecting 5-10 uls. Start at low organic to focus, then use gradient to elute components.

Normally run multiexperiment on the cone voltage (DP) with lower postive voltage to get molecular weight, higher for fragments, low negative and high negative plus the diode array. (poor-mans MS/MS)

[meillid]

is there any books that focused on drug degradents identification? Thanks.

[Daren]

I'm pretty sure you should be able to subtract the background noise with the Sciex software. What I would do is focus your efforts of collecting fractions of your unknowns and then trying to gently concentrate them (don't use high heat). THen if you have enough you can do both infusion analysis to see if you see something, try to optimize your MS conditions then confirm it via an LC-MS run and see if RT's match.

[meillid]

Thanks for kind reply. I have detected the mass that fit the mass of the guessed compond. I just wonder does this enough to say that I identified the compound? Do you routinely break done this peak with ms/ms and do liberary search to identify unknown? Since our instrument is dedicated to qantitive analysis, we don't have liberary data base. What else shoud I do? I am really not experienced in this aspect of industry. Thanks for any answer.

[Daren]

Hi Emily,

It really comes down to what you are trying to identify this impurity for. If it is because it is above the ICH guidline for identification, then yes you should be doing more analysis on trying to confirm it's identity, if it's above this level and at or near the qualification threshold than you have to make sure that it is not just identified but has been tested for toxicity ( there are various ways to do this). It sounds like you guys may want to hire a consultant in this area just to outline possible strategies for you.

But to take a first pass I would try to first just a get a decent MS/MS spectrum, see what you get, it may be something simple like an oxidation or reduction reaction of you parent compound, acid catalyzed dehydration is also very common and these are things that you could detect via the MS/MS spectrum without a library. Hope this helps
One other thing make sure you are actually seeing your correct parent ion for this unknown via adjusting your cone/inlet voltage and seeing which ions grow or shrink in intensity
Daren