Chromatography Forum

I would like to know how to calculate the percent recovery of spiked samples in accuracy.
Thanks for any suggestions,

% recovery = (measured mass)/(added mass) X 100

where "measured mass" is the result you obtained on the spiked sample, and
"added mass" is the amount you spiked. This assumes that the matrix sample you used had no analyte in it to begin with.

Is “measured massâ€

Yes. To expand a bit:

1. Run your standards and establish your calibration plot
2. Run your spiked sample and determine the amount of analyte using that calibration plot
3. The ratio of the amount of analyte determined in step 2 to the amount you actually added (times 100) is the percent recovery.

Tom, thank you for your help.

You're welcome!

Hi ,
How can I get the recovery if the matrix sample I use has an amount of the analyte in it ?

It's called "standard additions". You spike at multiple levels and then extrapolate back to zero added to get the amount in the matrix. The whole concept of "recovery" is probably hazy in this situation.

Thanks a lot tom but why do you think it is hazy?

Like Tom infers, I would only use standard additions as a last resort. If you're in the business, why can't a placebo matrix be prepared for you, which you could then spike? That's what we do.

The idea behind "recovery" is to be able to account for any bias in the results (for example, some of the analyte sticking irreversibly to the matrix). Standard addition compensates (hopefully!)for the bias (in effect, your actual sample is included in all the calibration runs), but it doesn't account for it in the sense that you don't know how much occurred. The big drawback is that, in effect, you have to run several calibrators for each sample. As Consumer Products Guy (can we just call you "CPG" for short?) points out, it's really a last resort; used most often when analyzing endogenous compounds for which a "blank", per se, doesn't exist.

Thank you all.
It seems to me that I have to use the standard addition method as my sample is a growth medium for a bacterial species which is analysed for the analyte at different periods of time.
Don't you think so?

Can you make or obtain the same medium without the analyte?

Yes. But the real sample after a period of time will have other compounds produced by the bacteria and some others may be consumed.
Don't you think that this will cause matrix differences?

Yes, but those differences are what you're trying to measure.

Please don't misinterpret what I was trying to say; there is nothing inherently wrong with using standard additions, but it does represent much more work than the usual "external standardization". In "standard addition", you essentially have to generate a separate calibration plot for each sample. In "external standardization", you generate a single calibration plot which you use for many samples.