decreasing signails

[zxmzxm1]

I am having problem with the reproducibility of LC/MS signals. I use API 5500. I analyze the stability of small drug molecules (MW 200-1000 dalton) in blood or plasma after incubation at 37C for a period of time. The sample was prepared by quenching 20ul blood or plasma using 600ul mixture of Acetonitrile and methanol (50:50). After centrifugation at 4000 rpm for 25 minutes, 70ul supernatant was removed and mixed with 210ul water. Then the sample is ready for injection on LC/MS/MS.

For most of the compounds, I got good results. However, some compounds have bad reproduciabilities. When I repeat 15 injections from the same blood or plasma samples, I see decreasing trend in the MS responses. Can anyone tell me what can be the possible reasons? I know it can be due to matrix effect. Can it be due to dirty MS source? Thanks in advance.

[Alp]

Have you tried something like extracting a blood sample, (using your 50/50 ACN/MeOH ppt) leaving it on the benchtop for about an hour or two, then extracting another sample from the same blood sample, and injecting the 2 hour old ample, then the fresh extract?
In other words, does your analyte/sample have stability over the period you are assessing it?

Alp

[zxmzxm1]

Sorry for the delay in my response. Our area was hit by thunderstorm. We lost power for four days. Stability was not an issue for the compounds having decreasing trend. I know those compounds are stable. Sometimes I also see increasing trend. For example, carbutamide (my internal standard) has good reproducibility in positive mode. But sometimes when I switch to negative mode, it shows increasing trend.

I am wondering what are the possibilities that cause this kind of problem. I know matrix effect can, what else? Thanks.