bad signal for post-SPE-spiked samples

[Fluterd]

Dear forum-members,

I would like your advice about the following problem.

I am validating my LC-MS/MS method for a basic compound and its metabolite.

I use a C18 column with a gradient of A) water 25 mM ammoniumformate pH 3 B) MeOH containing 0.1% formic acid.
The mass spectrometer is operating in positive ESI, MRM-mode with all parameters nicely optimized.

Injection of a standard solution gives nice, repeatable peaks.

In the future this method will be used for the analysis of urine and whole blood, so sample preparation is obviously needed. I selected an SPE method using mixed mode SPE cartridges (Bond Elut Certify, C18 and a cation exchanger) and I use the protocol that is given by Varian itself. The protocol also ressembles other protocols used in the literature, so I suppose this hould work fine.

However, when validating matrix effects (ME) and recovery (RE), following problem was seen.
I used a standard, a pre-SPE-spiked sample and a post-SPE-spiked sample to calculate RE and ME.
When using urine, RE and ME were bad (about 30-40%). However when using the ratio analyte over internal standard for calculating ME and RE, values were excellent (90-110%). Moreover, when I repeated the same processes with the same matrix, variability for the pre-SPE-spiked samples and a post-SPE-spiked samples was awful... So the internal standard compensates great for ME and RE, but of course can not solve the impact of the variability on sensitivity.

I repeated this experiment with pure water, in orde to exclude any effect of the urine, but the same was seen: good variability for standard, bad variability for different pre-SPE-spiked samples and post-SPE-spiked samples and significant loss of signal in the pre-SPE-spiked samples AND post-SPE-spiked samples!

I already checked some elements.
The glassware and pipets etc used all work fine.
I tested another SPE-protocol with a different elution solvent, this resulted also in bad variability and loss of signal.
When I simply evaporate some elution solvent and then add the standard, variability is again good and there is some signal loss (5-10%), so the elution solvent doesn't seem to be the cause...
When I simply ran some elution solvent over the SPE-column (so no condition, sample application or wash performed), I evaporate the elution solvent and the add the standard, variability is bad and so is signal loss.
Another person of the lab performed the SPE-procedure, the same results were obtained.
Mobile phase was made fresh, another column was used, ... : I think the LC-MS/MS method works fine...

I really would appreciate some advice since the problem is starting to drive me crazy!

Thanks!

[sam.pedraglio]

When I simply ran some elution solvent over the SPE-column (so no condition, sample application or wash performed), I evaporate the elution solvent and the add the standard, variability is bad and so is signal loss.

This test suggests me a very high cartridge phase bleeding or some manufacturing residue eluted from the cartridge.
This is the only reason why if you simply wash your cartridge and the spike the eluted solvent with your analyte you find signal loss.

Try this other test: wash your cartridge with 2 or 3 times the volume of the eluting solvent and then start the extraction procedure from the beginning (conditioning, equilibration and so on) and check RE and ME. If you're lucky you've already washed away the contaminants (or at least the main part). If this is working contact your sales man and ask for a batch exchange.

[Fluterd]

Dear,

Thanks for your suggestion. Since the Bond Elut Certify SPE cartridges are widely used in literature, I didn't suspect them of causing the problems.
These are the results of the experiment, I did everything in triplicate.

(1) 2 ml of elution solvent ran over the SPE-column, this was evaporated and standard added. There was loss of signal and a coefficient of variation higher than 15% (like the previous results thus).

(2) I then ran 2*2ml elution solvent over the SPE-column, towards the waste. The I ran another 2 ml over the column and evaporated this again and added standard. The variability is still bad, but the average signal was higher than that of sample (1).

(3) I then performed the whole SPE protocol (condition, wash etc). There was again high variability and the signal was lower than the standard and it was lower than the signal of (1) and (2).

What could I conclude?

We also have some other mixed mode SPE cartridges of Waters in the lab. Do you think it is a good idea to perform the same experiments on these columns?

Thanks!

[Fluterd]

Dear,

I tried the same SPE protocol on a different column (this time I used mixed mode) Waters Oasis 60 mg cartridges, with the same Varian protocol as I used before.

Variability of pre- and post-SPE spiked samples is better, but for one compound the CV of the different post-SPE spiked samples is around 15%. Is this ok?

Also a difference of 20% is seen between a standard in a vial and a vial that contains the same standard but that was transferred through a silanised tube. Previously, this difference was smaller. We manually silanise the glassware... Could there be a problem here or is it the SPE-cartridge...? What variability is considered as normal when transferring between tubes, vials etc...?

I know, a lot of questions, but I hope someone can help!!! I would be very grateful to the person who helps me solving this mistery...

[sam.pedraglio]

Variability of pre- and post-SPE spiked samples is better, but for one compound the CV of the different post-SPE spiked samples is around 15%. Is this ok?

For our standard procedure, this CV% is at the limit of acceptance, but it's ok. In my mind it could be related to an unefficient extraction procedure.

Also a difference of 20% is seen between a standard in a vial and a vial that contains the same standard but that was transferred through a silanised tube. Previously, this difference was smaller. We manually silanise the glassware... Could there be a problem here or is it the SPE-cartridge...? What variability is considered as normal when transferring between tubes, vials etc...?

A reduction of analyte lost using silanised vial is correct for a basic compound (no ion interaction between the basic center of it and the SiOH of the glass) even if it's very very compound dependent.
If you had better glass pre-treatment in the past check the procedure searching for sources of variability.

Have you the same problem using plastic vial instead of glass?
If you suspect a high unspecific binding to the plastic, you could wash them with water/acn/TFA 90/10/0.1.
We found this mix a good way to remove the slippery agent remained from the plastic production.