Chromatography Forum

Do I have to change my LOQ when I dilute a sample, and if so, why?

Seen lab results that when diluted by a factor of 10, it raises the LOQ by a factor of 10 making it an unusable number. Isn't the LOQ a measure of instrument performance and not take into account dilution?

The original is probably not a true LOQ which is why I prefer the statistical calculation of the LOQ based on the sum of the residuals (standard deviation). Factors that influence the LOQ are injection volume, detector, flow, interferences...

The original is probably not a true LOQ which is why I prefer the statistical calculation of the LOQ based on the sum of the residuals (standard deviation). Factors that influence the LOQ are injection volume, detector, flow, interferences...

Right, so if I have my LOQ calculated from LOQ = Blank response + 10*STD Dev

Of a set of blank samples then that value should not change when I dilute a sample, right?

The original is probably not a true LOQ which is why I prefer the statistical calculation of the LOQ based on the sum of the residuals (standard deviation). Factors that influence the LOQ are injection volume, detector, flow, interferences...

Right, so if I have my LOQ calculated from LOQ = Blank response + 10*STD Dev

Of a set of blank samples then that value should not change when I dilute a sample, right?

If you normally see 10ng/ul as your LOQ, then if no prep is involved the LOQ of the sample would be 10ng/ul. If you dilute the sample ten fold, you still see an LOQ of 10ng/ul on the instrument, but that has now become 10ng/0.1ul of sample or 100ng/ul. The LOQ of the instrument doesn't change when you dilute, but you can no longer have the same sensitivity versus the actual sample.

A rule of thumb I have always used, if you inject the sample full strength and have a result that is above the calibration, then dilute to bring the result within the calibration curve, report the original LOQ since you actually have data at that dilution factor of 1X. If you have to dilute because of an interference that prevents you from getting a positive confirmation of detection in the undiluted sample, then you have to report the higher LOQ taking the dilution factor into account.

LOQ for a sample has to take into account all of the preparation steps, so if you change anything it must be accounted for in the final result.

The LOQ belongs to the method and its validation using a particular sample-type. It is stating: "We can measure down to this level IF we use the following type of sample AND we prepare it in this way AND we run it in this way ON that instrument AND analyse the data in the following way".

If you change the dilution, it might or might not change the overall LOQ. If the sensitivity depends on the instrument's ability to discern the target compared to electronic noise, then dilution will reduce the LOQ in simple proportion. If the LOQ is limited not by the instrument, but by coeluting stuff in the sample (which might compete for detection, or which might make it hard to integrate small peaks in the presence of nearby big peaks) then dilution probably won't affect the overall method LOQ at all.