Shifting RT of Pyrazinamide and IS Peaks

[ghie malig]

Hello all:

We are validating a method for the determination of pyrazinamide in human plasma by HPLC UV. We are encountering retention time shift. We already did the following:

- pump check: flow rate is ok
- leak checked: none
- column temperature maintained at 27 degrees
- mobile phase equilibration from 1 to 2 hours at 1 mL / min

M.P.: 20 mM of Phosphate buffer/ Methanol, 96.8:3.2, pH 7.4
Sample Prep: 200 uL of plasma precipitated with 50uL of 20% Perchloric acid, 150 uL of supernatant is neutralized with 1M NaOH. 20 uL injection into the HPLC system.
Column: Phenomenex Luna C8, 4.6 x 250 mm, 5um

Retention times of Pyrazinamide and IS (Sulfanilamide) decrease all throughout the analytical run (e.g. from 9 minutes to 6 minutes). Where could be the problem?
Any feedback is appreciated. Thanks.

-ghie-

[Peter Apps]

Hi Ghie

Your mobile phase has very little methanol in it, and it could be that the stationary phase in the column is "collapsing" (the hydrophobic molecules fold tightly down onto the silica surface).

Could you try a C18 column with more methanol in the mobile phase ?

Peter

[Bryan Evans]

Or try a C8 that can be run in 100% aqueous mobile phase
without phase collapse

[tom jupille]

Or you could have residual protein building up on your column.

Some additional questions:

1. Does the retention time decrease continuously or does it shift erratically?
(erratic shifts would suggest pumping problems; if you are using on-line mixing, you may be on the ragged edge of your system's capabilities).

2. If the decrease is continuous, is it proportional to time or to number of injections? (inject some samples, pump mobile phase for an hour or so, then continue with injections; plot retention time vs. # of injections and then again vs. total time; see which is a better fit). If the shift is proportional to the number of injections, that suggests contamination from the sample. If proportional to total time, it suggests a mobile phase or column issue.

[Mark Tracy]

If you are seeing "phase collapse" (more accurately termed de-wetting) you will also see your void peak shift. De-wetting is reversible: just wash with degassed 100% acetonitrile for 10 column volumes.

[Uwe Neue]

I also suspect slow column contamination. You could improve the sample prep method by adding a SPE step on a mixed-mode cation exchanger. The supernatant after the precipitation is sufficiently acidic to protonate the analyte (and the internal standard) such that it would be retained on an Oasis MCX packing. I would then elute at pH 7 with maybe 20% acetonitrile. Maybe you will not need to go that high, but experiment will tell.

This small step will provide a much cleaner extract, and your column will last longer.