Hello,
I have run into an issue that has sparkled a debate at work: is the LLOQ/ULOQ impacted by injection volume under the following circumstances:
1) Same dilution factor of all samples
2) Volumetric deviations corrected for by internal standard (separate compound from analyte, not labelled)
3) Matrix effects studied and found to be negligible at all injection volumes
4) Good linearity of the regression model; not impacted negatively by inclusion of standards injected at lower volume
The problem arose when I, instead of performing a dilution, chose to rely on the correcting function of the internal standard and injected 1/5:th the volume (0.2 µL rather than 1 µL) of my most highly concentrated samples and standards and evaluated them as part of the same calibration curve as the samples injected at 1 µL. As I mentioned, matrix effects were investigated at all concentration levels as well as injection volumes and they were very low.
The software settings I use employ the "signal response" which is [area analyte/area IS]-ratio and the regression model from the calibration curve to calculate concentrations. Hence, under the circumstances I mentioned above the response is independent from injection volume - providing one does not drop below the background level or saturates the detector.
We ended up in a situation where one of my colleagues claim that the LLOQ for the samples injected at 1/5 should be adjusted and multiplied by the factor difference in injection volume (i.e. x5) and the ULOQ to be divided by the same factor. Another colleague claim it should be done the other way around.
In contrast to both I claim that the volumetric correction by the IS has already compensated for this and that LLOQ/ULOQ should remain at the set concentrations of their respective standards.
Can anyone help me out here?
Best regards and thanks in advance,
// Karl
It's doubtful that your signal response is independent of your injection volume as you stated. I suspect that you are thinking of the response ratio (which should be independent of volume). Ideally, if you double the injection volume, you should see a doubling of the response.
Hi.
Depending on the ionisation mode your detector will give a concentration or amount proportional response. In both cases, you're reducing that response by 5 X due to 1/5 of sample injected. This should be equivalent to injecting 1 uL of sample diluted 5X. Thus, what you can do (assuming no significant matrix effect) is to calculate the concentration of your sample directly from the calibration curve and multiply it by 5 to correct for the reduction in the injected sample.
Cheers!
The answer depends on the calibration that you are using.
You say that the matrix effects and linearity are good and unaffected by injection volume. The chromatographic effects are therefore *almost* negligible.
If you are using a multi-point calibration then your LLoQ is your lowest calibration point and your ULoQ is your highest calibration point. In this case your ULoQ and LLoQ are unchanged.
If you are using a single point model or extrapolating beyond your upper or lower calibrators then I would be very cautious about any claim of a change in calibration range, and I recommend validating this range on your lower injection volume.
The fact that you are doing a dilution suggests that either your calibration range is wrong or that you know the sample in question has a very high signal. A reduced injection volume will not affect the quantitation but will allow you to obtain a better peak shape that does not cause detector saturation or column overload.
I stated above that chromatographic effects are *almost* negligible in the method that you describe. If you have a poor detector response, signal:noise ratio or peak area of your analyte or internal standard then this will lead to increased imprecision which will negatively impact all your results (even within the linear range).
I hope this helps
Richard
- Badges
- Science topic
- Similar topics
- Correction