We are determining recovery for a drug in plasma. We have spiked plasma samples which are liquid/liquid extracted, evaporated and redissolved in mobile phase. We compare these to spiked mobile phase samples. The question is: if, say, 4 ml extractant is used and 3.4 of the organic layer is dried down, does that 15% loss calculate into recovery loss or is it a known factor which does not factor into the loss.

You should take into account the lower volume you recover, i.e. you should multiply the response of your extracted samples by 4/3.4.
In addition, the correct way of measuring recovery from complex matrices such as plasma is somewhat different from the procedure you adopt: you should compare the response of the extracted samples with the response of blank matrix fortified with the analyte after extraction, not with "spiked mobile phase samples". In this way you take into account the possibility of coeluting, interfering peaks

Thanks

I disagree with Anon 7:46 am. Anon, I know you know what you are mean, but it didn't come out quite right on the screen.

Anon 7:12, you may need to do both. Your "spiked mobile phase samples" are essentially external standards, assuming you mean mobile phase with added analyte. Using chromatography presumes the analyte is being separated from all others, and comparing to external standards will help tell you this. If there are "coeluting, interfering peaks", they will appear to be the analyte. If you think the suggested technique can compensate for interference peaks, try spiking 0.01 mg into a 'blank' sample that has 10 mg of interference. The 'blank' and 'spike' will be indistinguishable. You may not be able to tell the difference until, say, 1 mg (10%) is added. And what if the interference is not constant?

Typically, this 'background', if it can't be separated, is quantitatively characterized (low-high range for the project) and subtracted or otherwise compensated for in the project report.

And that is why more and more sophisticated detectors were built, culminating with mass spectrometers. They help, but do not guarantee that we can find that 0.01 mg in the 10 mg of interference.

Mostly used in GC analysis, spiking into a blank matrix after extraction compensates for response enhancement or quenching effects, presumably. In a perfect world, these "spiked matrix" samples would look identical to "spiked mobile phase samples", and it is important to know if the matrix is causing these effects, and to what extent.

This is an imperfect solution, but sometimes the only cost-effective one available. If this approach must be used, 'blanks' can be very difficult to come by, and very valuable when found. And to find those 'blanks', samples must be compared to "spiked mobile phase samples".