Sudden decrease in LC-MS peak area count

[kchua611]

Hi there,

We've recently validated a method, however, we now have an issue with the method whereby there is a drastic decrease in peak area count for the internal standard and analyte. This was not observed during the validation process.

During the validation, the ISTD peak count in matrix remained constant around 250,000 to 350,000. However, it has now decreased from 300,000 to 8,000 over the last few days. We initially thought it might have been a column degradation issue, however, this has been ruled out as it's observed in the new column as well.

We are thinking that it's most likely due to matrix effect as we had just tested 10 injections of matrix extracted sample followed by 20 injections of an aqueous sample (no matrix). For the matrix extracted sample, the ISTD peak count dropped from 55000 to 8000, and upon injection of the aqueous sample, a reverse trend is observed with the peak count increased back from 100000 to 300000.

Any ideas why this might be especially when we did not observe this issue throughout the validation runs (> 10 runs, with >1000 injections in total)? We're really stumped as to why this might be the case.

Thanks!

[Rndirk]

What is the difference between the sample matrix and the matrix you used to validate? Can you be a little more specific? Is it, for instance, urine?

[kchua611]

Hi Rndrik,

Thanks for your reply.

There is no difference between the sample matrix and the matrix that we used to validate. We used the same lot of dog pooled plasma - which adds to our confusion.

[Rndirk]

Official guidelines on LCMS validation concerning matrix effects state that pooled matrices should not be used to validate a method.

This is a possible explanation for your problems. Since your method is already validated, I don't think there's much you can do without (partial) revalidation.

I think it depends on the regulations (if any) that apply to your method if you can continue with routine analysis or not.