B Vitamins HPLC Analysis

[Sallybeetle]

Is there a good, reliable method for the HPLC analysis of the water-soluble B Vitamins?

I used a couple of Waters Corporation application notes to develop and validate a method for niacinamide, thiamine, pyridoxine, and riboflavin combined in a product. The method uses the Atlantis dC18 column and a gradient with mobile phase A (0.2% v/v TFA) and mobile phase B (10:90 0.2% v/v TFA to acetonitrile). The samples and external standards are diluted in purified water and analyzed at 275 nm.

The gradient is:

Time MP-A MP-B
0 100% 0%
5 100% 0% isocratic hold, separates niacinamide and thiamine
15 89% 11% ramp to elute pyridoxine
28 89% 11% isocratic hold to elute riboflavin
29 0% 100%
33 0% 100% isocratic hold, flush column
34 100% 0%
44 100% 0% isocratic hold, re-equilabrate column

I get great analyte peak and degradant peak separations, peak symmetry, and generally good %RSD. But, I sometimes get very variable pyridoxine results (82 - 108% label claim).

When the pyridoxine results are low, all the separately prepared samples in the run are low. The external standard peak areas and the other analytes in the samples are generally very comparible between the runs that exhibit lower and higher results.

Does the pyridoxine stick to glassware or HPLC components in a variable fashion? Should I be using the expensive HPLC vials? Does it react differently to the ion pairing reagent each time of sample preparation?

Some here have heard that the pyridoxine does not like ion pair reagents. Also the same with these vitamins in general.

Any advice, wisdom, and / or prayers would be greatly appreciated.

[Bruce Hamilton]

For anybody analysing vitamins, I'd highly recommend starting with :-

" Modern Chromatographic Analysis of Vitamins " by Andre P. De Leenheer, Willy E. Lambert, and Jan F. Van Bocxlaer.
The third 3rd Edition, ISBN 082470316-2 ( 2000 ), has a chapter on the analysis of pyridoxine.

If you are analysing formulated multi-vitamin products, then there proably will be details of the analytical methods in Pharmacopeia or Food Codex, which should also desciribe any precautions needed for quantitative results.

Bruce Hamilton

[Bryan Evans]

Below are chromatograms of water soluble vitamins
separated on Unison UK-C18:

http://www.silvertonesciences.com/files/TI269E.pdf
http://www.silvertonesciences.com/files/TI268E.pdf

[Sallybeetle]

Bruce:
Thanks for the book reference. I am going to the library tonight to check it out.

Bryan:
The chromatograms show good Tf and Rf just like mine; and, the method is very similar to mine. But, has this methodology / column been used repeatably and reliably for the vitamins? No insult intended towards your column; but, my method / column worked for the first few months, then started giving these crazy results.

I validated using two different gel lot numbers. We have tried at least 4 different columns using these gel lot numbers.

[Uwe Neue]

My most important question is, what does the column have to do with the results reported!!!

I was very annoyed about the rather blatant advertisement by Brian Evans with the underlying notion blaming the column for the event reported. This was below the belt! However, I initially did not feel that I should make a big deal out of it.

Let us try to solve Sallybeetle's problem instead!

I am puzzled by the problem as well. I do not have the answer, but I would look for adsorption to septa, vials, even parts of the injector. There is nothing extraordinary about pyridoxine that I can see, but maybe the literature recommended by Bruce can shed some light on possible stability issues. I also do not think that TFA is to blame.

[goxy43]

Hello,
I did some vitamin B separation on PepMap columns (Dionex) and did experience the same problem. I tried different batches of PepMap columns and different wash procedures for the autosampler, the trap column ( we did some trapping before separation) and of the separation column. Nothing did really help. What I have found was that sample vials (purchased under identical P/N and from the same supplier) were produced using a different recipe and the manufacturer sold them without notifying the change. I have used polyprop vials at that time. For the glas vials: I know that some glas vials really do bind a lot of material and you need either to use sylanize them or try the polypropvials from eppendorf as they proved very reliable.
Success
G.

[Bryan Evans]

Hi Sally -

Thank you for your question. Yes, we have customers using Unison UK-C18 for this
application and they are very happy with it.

I agree that the sample vials, autosampler, ect. should be investigated.

Bruce - thank you for the book reference. It looks interesting, I might
have to purchase it.

[juddc]

If I understand you correctly, your standards are consistent in all cases (whether or not your samples have suitable recovery) and other analyte peaks in the samples are OK as well, meaning that your weighing and pipetting are likely spot-on.

I would think that if your standards are OK, your LC system should be OK and that if adsorption was occurring with glass or your LC components it should be detectable there as well.

I suspect matrix / extraction issues.

Is there anything in your sample that does not dissolve in your extraction solvent?

Is the pH of your standards consistent with the pH of your prepared samples?

Have you assayed your sample extracts over time?

Have you tried doing a recovery on a spiked solution?

I would investigate whether or not the pyridoxine (HCl or free base?) might be sticking to some undissolved particulate (if present), making a mess of your recovery. Instead of diluting samples in DI water, you might dilute them in your mobile phase A and see where that gets you. Alternately, if you have particulates in your sample preparations, try to dissolve or at least reduce them too.

Good luck!

[Bruce Hamilton]

Not sure if you managed to access the recommended text, but if you did, you'll understand why there's a whole chapter devoted to Vitamin B6, as naturally, it's often a whole family of compounds.

I didn't read the chapter, but a couple of comments stood out.
B6 can occur in 6 different forms in nature, but pyridoxine is used for supplements because of it's relative stability compared to other forms, and I assume that you are measuring a formulated/supplemented product.

It's photosensitive and heat sensitive - look closely at your sample preparation - it may be slightly inconsistent.

Several of the methods of multi-vitamins used sodium hexane sulphonate as the ion-pairing agent.

Please keep having fun,

Bruce Hamilton

[Sallybeetle]

Thank you to all who provided much needed and greatly appreciated advice. I am implementing a number of your advice steps into the procedure. I hope these cure the problem.

And. . . yes, Bruce. I read the 'bible on vitamins' you recommended. So, I'm sure that I will be having much fun in the future with vitamin analysis.