Guidelines/Tips for pesticide screening by GC-LC/MS/MS

[Rndirk]

Hello,

Our company wants to offer multi-residue analysis of pesticides in human urine. I have the task to research & develop this method with the following (loose) requirements:

- include as many as possible (hundreds) of relevant analytes
- Preferably quantitative on a relevant level for urine
- 1-2 sample prep / sample, as simple as possible
- Detection with GC & LC-MS/MS (both triple quad)

I'm not asking to do this task for me . I don't have experience with developing chromatography for this amount of target molecules. What I'm looking for here are some general guidelines/tips/experiences with these kind of screenings. Also more specifically:

- How should i deal with metabolization? Find out in literature which pesticides are metabolized by the human body and scrap those?
- Will I need a database?
- How would i deal with quantification? We won't have all the target components in a calibration solution.

Any help is appreciated!!

[richiekichi]

You should contact the manufacturer of your instruments. They will very likely be able to provide you with a solution for a hefty fee.

[mckrause]

1. You need to measure the concentration of the metabolites as well as the parent pesticide. Some pesticides are very rapidly converted to the metabolites; others take longer.

2. There are published databases available. If you have Agilent systems they have a full-blown pesticide analysis program for you (other manufacturers may, as well). Take full advantage of AMDIS if you can.

3. It is quite complex. The LC-MS/MS solutions typically involve time-switched selected ion monitoring; it is not easy. It takes a lot of method validation and maintenance.

4. You can buy commercial standards as sets. I use the Restek sets; others have sets available as well. It is an investment in standards but the only way you can validate your method.

5. I would assume that you could use QuEChERS for sample prep; you'd need somewhere around 5-10 mL of urine. This is the best prep for this type of analysis that I know of. It's fast, cheap and works. You'll need to buffer it, so you'll have to work out a neutralization step in there.

6. Each system is better for certain sets of pesticides. Take full advantage of both systems. For example, using conventional ESI-LC-MS you pretty much can't see any of the chlorinated pesticides but they show up beautifully on the GC-MS. The converse is true for most carbamates, carbendazim, etc. We run about 90 pesticides on the LC-MS and the remainder (>200) on the GC-MS.

[James_Ball]

When I need to add a new pesticide to our list I usually check with Restek first. They have a lot of the pesticides in a spread sheet with a list of the MS/MS transitions and retention times for certain columns, and they have them for both LC and GC MS/MS. They also have mixes that total about 400 pesticides I believe on last count. It is designed for Quechers analysis of foods, but should suit your project as well.

[Rndirk]

Thanks for the helpful replies.

It quite a big task but I'm getting to the point where I can see the forest trough the trees. I made an inventarization of every pesticide we have in the lab (from other analysis procedures) and devided them between LC & GC. I will make an MRM method for each instrument.

I figured Quechers is the only way to combine everything in one sample prep, and I'll probably need to clean up the LC & GC part with a different dSPE.

The Restek mixes are indeed promising to add more pesticides.

5. I would assume that you could use QuEChERS for sample prep; you'd need somewhere around 5-10 mL of urine. This is the best prep for this type of analysis that I know of. It's fast, cheap and works. You'll need to buffer it, so you'll have to work out a neutralization step in there.

Could you elaborate why I need a buffer step? Is this to make sure I'm working with the same pH on different urine samples since it tends to be different?

[mckrause]

The optimum pH for multi residue pesticide analysis is between 4 and 6. You lose some of the pesticides and metabolites to degradation at basic pH levels; others need to be extracted below their pKa. The QuEChERS method does do some buffering (sodium acetate/acetic acid) but it has been my experience that you really need to adjust the pH prior to performing the extraction. We neutralize highly acidic samples (like jalapeños in escabeche, lime juice, etc) and we neutralize any sample with a pH over 8 prior to extraction.

As far as cleanup, yes, you will probably want to use separate LC-MS and GC-MS cleanups. On the GC-MS you are really worried about background, so we generally use a dSPE cleanup that includes GCB (carbon black). Unfortunately you lose several pesticides in this step. Most of them are LC-MS pesticides, so we don't include the GCB in the LC-MS cleanup (plus, it's cheaper). In setting up our methods what I did was run the GC-MS first and figure out which pesticides are not all that great on the GC-MS side, and then I threw them into the LC-MS run. For example, we run acephate, omethoate and diazinon on the LC-MS side because they are so much better than on the GC-MS side.

Depending on your instrument your run type changes. For example, I run an Agilent 6890/5973 for the GC-MS runs. I use AMDIS and extracted ion profiles for quantification, and acquire in full scan mode with large volume injection (CSR) using the Agilent retention time locking method. On the LC-MS I run a MRM experiment with multiple time-switching schedules for transitions. It is quite tricky to balance the LC-MS to get it to all work together, so I try to run the minimum number of compounds possible on the LC-MS. If you are buying a new triple quad look at the scan rates because you want the fastest scan rate you can get.

[Rndirk]

Thanks a lot for sharing this information, this is really helpful. I have a related question about injection:

From my experience with running Quechers extracts with PTV injection in GC, it's such a pain compared to running any other method with PTV injection of hexane extracts.

I've been fine tuning the PTV parameters for acetonitrile. For the calibration standards in acetonitrile it's more or less OK (more tailing for early eluters like dichlorvos, compared to hexane). But for quechers extracts the signals get worse: some compounds added by standards addition are simply lost. I realize this can also be due to the workup. But the chromatography, even with backflushing, tends to degrade after running about 5 samples => higher background, more signals lost.

We suspect the 1% acetic acid in the extraction causes problems with PTV injection and the column. Should i try to reduce the amount of acid, or just go for splitless injection? We are using an MMI inlet. Will splitless be more robust in the long run?