Bioequivalence and Bioavailability Forum

Dear Hale!

❝ mobile phase: Acetonitrile: 10mM ammonium acetate:formic acid

❝ (50:50:0.530, v:v:v) pH: 4.6

❝ reconstitue solution: water+25mM ammonia+ACN (50:50), pH: 9.9

One thing comes into my mind: A different pH of the analyte's solution and the mobile phase may lead to instable conditions on top of the column. Do you have problems getting the extracts into solution? Maybe ultrasonification helps. IMHO it would be better to dissolve extracts in mobile phase - or even better a solvent with slighlty less organic modifier (results in narrower peaks).

Dear Hale,

❝ I have response variation problem. When injections made from same vial

❝ responses of the analyte and IS changes but not with same proportion.

❝ Also after every six injections in a batch mobile phase is injected four

❝ times and results get better with this procedure.

I can see two different reasons for this problem, in addition to the one mentioned by Helmut, but there may be some others:

- one common problem with LC/MS/MS can be the need to condition the ion source to obtain a stable signal. You may need a number of injections from plasma samples before your signal stabilises. Sometimes it is just half a dozen, but it may happen that you need 20 or 30 injections. Injecting mobile phase regularly doesn't really solve the issue, you rather need to inject some plasma blanks before your samples. If you are in this situation you probably see a trend towards a decrease or increase in the signal of the internal standard over time in the run. It may not be that clear if the IS is not affected much but the analyte is. You may be able to explore this by repeated injections of a QC sample in a run and monitoring the response of the analyte and IS.

- another possibility is the accumulation of phospholipids on the column, which can then "bleed" out of the column in an erratic way, leading to matrix effects issues. Variations in retention times can sometimes also be seen due to this accumulation, which modifies the polarity of the stationary phase. Possible solutions are a better sample clean-up, or use of a mobile phase gradient to clean the column between two injections (will not take longer than your injections of mobile phase).

Regards
Ohlbe

Dear Hale,

In case the problem has not been solved yet, I would suggest the following:

- you would benefit from injecting a pool of extracted samples, say low level QC, rather than blank to "warm-up" the instrument. In this way you can see how the ratio varies and how many samples it takes to obtain a consistent signal.
- If you decide to use a cleaning injection, for consistency that injection(s) should be made after each sample.
- for cleaning phospholipids, I would recommend adding a 1-2 minutes rinsing step with 100% methanol. You would need to determine the time based on the flow rate and all the conditions of your method, by monitoring the phospholipids. They have a common fragment of 184 and a precursor ion scan test will reveal their m/z in Q1. A few m/z that I remember are 496, 524 and 782. They elute fastest with about 80-85% acetonitrile, at higher organic % they start to retain more. Among other reasons, this is why the classic gradient that goes to 100% ACN is not recommended. However, 100% methanol will elute them.
- you have mentioned not having this issue during the validation, but with the samples. There could be a number of reasons:
- a metabolite with the same MRM as the IS (probably not the case, but it could happen)
- a metabolite co-eluting with IS and impacting the ionization
- the formulation components (this can be verified by looking if there is a trend of the IS response over the sampling time, corresponding to the absorption and elimination of the "fillers"). If you are allowed, you can also spike a low conc in a blank sample and a later time point of a placebo - if available.
- inter-subject variability.

All this stands true assuming that the variability of IS response is confirmed to occur in the mass spec, and not during extraction (LLE extractions are notorious for inconsistencies).

Last point: if a validation becomes necessary, and if it is possible, the reconstitution solution should be very close to the composition (including pH) of the mobile phase. Some compounds behave well, but other compounds are impacted by the difference in pH, especially if they have multiple pKa values, and don't have time to equilibrate from one form to another during the elution.
I hope this helps, if not for the current problem, maybe for potential future issues.