How do you deal with flavour peaks?

[Mattias]

I have been asked to improve an HPLC method for sodium picosulfate. The drug is present in a mix with a natural flavour, and as expected the chromatogram is filled with flavour peaks.

The method must detect known (two peaks) and unknown degradation products of picosulfate. I have been struggling to move around the flavour peaks to areas of the chromatogram where I do not expect any picosulfate peaks. This works fine for a while - but when I stress the sample or get a new batch of flavour I am back to scratch again.

What I really would need is a detection technique that is more specific, but I cannot measure any fluorescence at all from this compound. LC/MS is not an option since this will be used at QC. The main compound and the impurities all have the intact pyridine group - could that be used for some specific derivatisation?

The obvious solution is to run a flavour sample for each batch of product (using the same batch of flavour as in the product). But not even that would be foolproof since the flavour ages differently in the product and as a substance.

I guess this is not an unusual situation for many of you, but it is the first time for me. All ideas are welcome!

[prepcolumn]

Don't you have any possibility to extract picosulfate from flavors with extraction?

[Mattias]

It seems hard, since picosulfate has two sulfate groups (very anionic) and one of the impurities has none (not anionic at all). I cannot think of any ion-exchange SPE that would work here? I assume that reversed-phase SPE would capture everything, including the flavor peaks.

I would love if someone had a brightly flourescent agent that would react with the nitrogen in pyridine (and pyridine only) in water solution. I guess that does not exist.

[Consumer Products Guy]

Welcome to my world of fragranced OTC pharmaceuticals.

Not exactly the same as your issue, but nonetheless an area that the FDA/USP don't address very well, as the vast majority of pharmaceuticals are not flavored or fragranced. Your active is very polar, sounds like its degradation product is not. Most fragrances are not nearly that polar. You could drive off most flavor components in the oven, if all your other stuff is stable (likely not).

What interference area compared to the area of the sought-for components are you finding?

[Mattias]

Welcome to my world of fragranced OTC pharmaceuticals.

Not exactly the same as your issue, but nonetheless an area that the FDA/USP don't address very well, as the vast majority of pharmaceuticals are not flavored or fragranced. Your active is very polar, sounds like its degradation product is not. Most fragrances are not nearly that polar. You could drive off most flavor components in the oven, if all your other stuff is stable (likely not).

What interference area compared to the area of the sought-for components are you finding?

I get flavor peaks of all sizes, from 0.01 - 50 area% of the active compound. We have quite tight limits for unknown peaks (0.2%) and it is a hazzle to figure out if what peaks that can be ignored or not. Good idea with the oven, never thought of that. It should stay stable for an hour or so.

The main peak and the two impurities have very different polarity, and so far I have used very steep gradients to get them out. But a steep gradient also reveals everything else in the sample...

Today I managed to get all three "real" peaks very close to each other (by using tetrabutylammonium sulfate). Close enough to be able to separate them with a 5 minute isocratic run. It appeared as the flavor peaks did not disturb in this small retention window. They were eluting in the front or stuck on the column. But I will check with more flavor batches tomorrow, I keep my fingers crossed.

[Vlad Orlovsky]

Mattias,

You can design procedure where you disulphate peak is trapped on the anion-exchange guard and washed off during the run (need a switching valve). We did couple of methods where we analyzed amino acid in tooth paste which contained dozen of excipients. by carefully designing guard and column we managed to isolate only one peak (target). You can do the same, by exploring difference between 2 sulfates, one sulfate s and non. I will send you "Toothpaste report" by email.
It uses this approach:

http://www.sielc.com/upload/file/pdf/SI ... r_2004.pdf

P.S. Are you still at Ferring?