Chromatography Forum

Dear friends,

We have been working on drug analysis in human biological matrices (plasma, serum, blood, urine) during these last few years using several API 4000 or Waters Premier. Sometimes for unexplained reasons we face interfering peaks in the LCMSMS technique at the retention time and m/z of parent and daughter that are very big (more than 50% of LLOQ). We usually check for the source of this interference, by cleaning the instrument, injecting mobile phase, or injecting the solvents used in extraction (by evaporating the amount usually used and then reconstitution and injecting under the same conditions). Also when this happens we change all the devices and tools, e.g., autosampler by using a manual injector, or even the whole HPLC, micropipettes, etc. Unfortunately these procedures sometimes do not solve the problem.

My question is, if these interferences are not contaminations of the analytes of interest, how come we get a quantifiable signal (more than 50% of LLOQ) at the same retention time, parent and daughter m/z? This is against the concepts of selectivity and specificity of the LCMSMS technique. Is there any explanation of interference in LCMSMS technique that is "opposite" of ion suppression? Would I be expecting extra signals due to plasma components that, although does not have the same m/z, yet still rise up as a quantifiable peak?

What kind of sample preparation/clean-up do you use?

Regards Bert

Dear Bert,

Thanks for the prompt interest. I would say usually we use liq-liq and in very few cases direct precipitation.

My question is really not limited to a single case, we have been facing such strange (unexpected interference) in many cases. The question is more of a kind of fundamentals or principals of the LCMSMS technique. I know how ion suppression phenomena affects the response (thus sensitivity), but, is there an “oppositeâ€

[quote="pepe"] cuz I think that for the unique combination of Rt and m/z (parent and daughter) there should be only one thing that gives a signal and that thing only will be the analyte of interest.
[quote]

Hi Pepe,

I think you can't be 100% sure of that When you deal with biological matrices, the composition of individual samples from different subjects is different. Is your contamination peak for that samples reproducible? Dir you try another sample prep or LC condition? It could be that certain compounds fragment in your source and give that particular response. Another possibility is to monitir two or three daughter ions and check the ratios. This will surely increase your selectivity and reliability of your results.

regards Bert

I can not but appreciate your prompt answers. Yes dear Bert, this is what drives me crazy, that one changes the LC conditions and still get the same shift in Rt as the analyte of interest, which again, indicates that it seems to be the same molecule. If it was not that we control to very last strict step the volunteers who participate in our studies, I would say those volunteers have taken the medication under study without our knowledge, but this issue is not possible under our strict controlled clinical conditions.

I loved very much the idea of "another possibility is to monitor two or three daughter ions and check the ratios". See, this is the kind of help I was searching for. Ideas or references that could help me to experiment how to make sure the mentioned peak (s) are interferences from plasma (or others) or real contamination by the analyte. I really appreciated your recommendations.

Best regards,

Sorry, but I dont have references at hand. What you could do is look for demands for regulatory methods of FDA and EMEA and papers concenrning identification using LC/MS/MS. By the way, did you try to take a scan of the eluting peak and compare it with a scan of your analyte?

regards Bert

Thanks Bert for all your efforts.
I will scan to see what these peaks are.
Best regards,

hi pepe,

We have had exactly the same problem in our lab. We are using API4000 mass specs and have had a similar problem with several different methods. The interfering peak is present at exactly the same RT as the analyte and has the same MRM transition.

We have found that it often originates from the column, especially when running gradients - fast gradients are particularly bad. If we don't inject anything but just run the gradient through the column the peak is there. Are you running a gradient? I suspect that the analyte is building up on the LC column and eluting when the gradient is run. We have found the solution is to run isocratic methods where possible, or a very shallow gradient.

Preparing fresh mobile phase and flushing the aqueous mobile phase line regularly with an organic solvent can also clear the problem.

I'd be really interested to know if anyone else has come across this?

George

Hi Pepe,

From what you are saying, this sound like being the same molecule and probably appears due to carry over problems. As you do not always see it, it means that at some point something (probably in the injector) goes wrong...