Page 1 of 1

Sample intake for premixtures

Posted: Wed May 08, 2019 7:23 am
by Rndirk
I'm working on developing an LCMS analysis of fat-soluble vitamins A, D and E to replace our routine HPLC analysis. Most of our samples for this analysis are feed and premixtures. Premixtures contain vitamins in about 1000-2000x higher amounts than the feed they're being mixed into. In our HPLC procedure, we follow the same procedure for feed & premixtures and just dilute the premixture until it falls in the calibration curve.

The LCMS analysis in development, however, will make use of isotope labeled internal standards that should be added before the (saponification) sample prep. It is unrealistic, and like a waste really, to add like 1000x more internal standard to the premixture just to be able to dilute it to the calibration curve level in the end.

I solved a similar issue for water-soluble vitamins in supplements, by dissolving 100mg in 100mL buffer, and taking 1mL of this homogeneous mixture through the sample prep (adding the internal standard at this point).

I have less experience with premixtures. Is there any solvent (combination) that I can use that should make a homogeneous "solution" of a premixture that allows me to pre-dilute these kind of samples?

Re: Sample intake for premixtures

Posted: Wed May 08, 2019 11:38 pm
by Consumer Products Guy
I'm thinking that the premix assay is an internal test. So I don't see the need to use the expensive internal isotope standard test and LCMS on the premix, why not just keep that assay on regular HPLC using external standard quantitation?

And just use the LCMS test on the finished product.

Re: Sample intake for premixtures

Posted: Thu May 09, 2019 1:31 pm
by James_Ball
Because of the expense of the internal standards I would do a study using known spikes of the target analytes into the premix matrix then process and add internal standard just before injection. The study is to determine the efficiency of the extraction by calculating %recovery, but the internal added just before analysis will still account for any matrix suppression/enhancement, which is the main reason for using a labeled internal standard.

Re: Sample intake for premixtures

Posted: Thu May 09, 2019 2:53 pm
by Rndirk
Consumer Products Guy wrote:
I'm thinking that the premix assay is an internal test. So I don't see the need to use the expensive internal isotope standard test and LCMS on the premix, why not just keep that assay on regular HPLC using external standard quantitation?

And just use the LCMS test on the finished product.


You're right that isotope labeled internal standards are probably an overkill since (correct me if i'm wrong) premixtures are not a very complex kind of matrix + the fact that it will be diluted a lot. However there are some issues in our HPLC procedure and I need to make 1 single procedure on 1 single instrument and do a proper validation (this time).

James_Ball wrote:
Because of the expense of the internal standards I would do a study using known spikes of the target analytes into the premix matrix then process and add internal standard just before injection. The study is to determine the efficiency of the extraction by calculating %recovery, but the internal added just before analysis will still account for any matrix suppression/enhancement, which is the main reason for using a labeled internal standard.


This is not a bad idea. I'm going to do some tests like this.

Another idea a colleague gave me was, instead of making a liquid "pre-dilution", to 'dilute' the premix powder with another, inert, powder. For example, I could take 100mg premix + 100g of inert powder, mix it, and proceed to take 1g of this mixture in analysis (which gives me the 1000x dilution i'm looking for compared to taking 1g of feed in analysis).

I'm not sure if this is practical, and this would mean I take 1 milligram of premix in analysis. This sounds like a scary low amount in terms of being representative for the whole sample, or am I missing something?

Re: Sample intake for premixtures

Posted: Thu May 09, 2019 7:34 pm
by James_Ball
Rndirk wrote:
Consumer Products Guy wrote:
I'm thinking that the premix assay is an internal test. So I don't see the need to use the expensive internal isotope standard test and LCMS on the premix, why not just keep that assay on regular HPLC using external standard quantitation?

And just use the LCMS test on the finished product.


You're right that isotope labeled internal standards are probably an overkill since (correct me if i'm wrong) premixtures are not a very complex kind of matrix + the fact that it will be diluted a lot. However there are some issues in our HPLC procedure and I need to make 1 single procedure on 1 single instrument and do a proper validation (this time).

James_Ball wrote:
Because of the expense of the internal standards I would do a study using known spikes of the target analytes into the premix matrix then process and add internal standard just before injection. The study is to determine the efficiency of the extraction by calculating %recovery, but the internal added just before analysis will still account for any matrix suppression/enhancement, which is the main reason for using a labeled internal standard.


This is not a bad idea. I'm going to do some tests like this.

Another idea a colleague gave me was, instead of making a liquid "pre-dilution", to 'dilute' the premix powder with another, inert, powder. For example, I could take 100mg premix + 100g of inert powder, mix it, and proceed to take 1g of this mixture in analysis (which gives me the 1000x dilution i'm looking for compared to taking 1g of feed in analysis).

I'm not sure if this is practical, and this would mean I take 1 milligram of premix in analysis. This sounds like a scary low amount in terms of being representative for the whole sample, or am I missing something?


I would expect a lot of variance depending on how well the two powders mix together. If you could run 10 aliquots and have low RSD then possibly it would work, but it still puts the performance on how well it is mixed.