recovery problem

[chemist-pharma]

I have developed an HPLC method. It works fine for all the standard. When I spike the sample, it gives me low recovery. If I dilute the sample 10 times, then spike, the recovery is over 90%. Can anybody explain? The samples are citrate buffer with pH~5-6 with protein concentration about 50-100mg/mL. Analyte is lipoic acid (thiotic acid). Mobile phase is phospate buffer with 40%ACN.
Thank you.

[bert]

Do you have any sample preparation or do you inject the sample containing the protein directly. What kind of detection do you use??

regards Bert

[chemist-pharma]

I use molecular weight cutoff filter(10,000 cutoff) to filter out protein and then inject.

[bert]

It seems to me that there is a interaction between your analyte and matrix or matrix and detector.

Did you spike the diluted matrix sample with the same conc. analyte?

Do you use MS?

To evaluate sample prep problems, try to remove the protein by acetonitrile or acid precipitation and look at the effect on recovery.

[Mark Tracy]

I agree with Bert about the interaction with the protein matrix.

Your sample prep is a good example of an assay for "free" thioctic acid. Since you get better recovery at high dilution, it looks like an equilibrium process. It is probably some kind of ionic interaction that you can manipulate by pH.