Specificity of a method: late eluters

[aceto_81]

Hi,

when performing a method validation for an API, we have to perform specificity against related compounds.
Now we have a API with a few related components which are strongly retained by the column. So strong that they require an additional gradient to retain them from the column.

How do we justify this to the regulatories?
1. Just inject them in the isocratic method, run for 3,4,5.. times the retention time of the main component and write:"undetected, so method is specific"
2. Add a gradient to the method so that they are eluted. But do we have to use the gradient method for our samples, as the degradation products of our API don't require the gradient (and the related substance don't show up on the API), so the gradient is only a waste of time.
3. maybe other ideas?

How do you guys and girls deal with this sort of problem?

Kind regards

Ace

[placetoo]

Greetings Ace,

I think it is important to remember that we are not only trying to justify our craft to the regulatory bodies, but also to ourselves. We always want to demonstrate that we are, confidently, identifying the target compounds to the exclusion of all others.

I believe that you have, really, answered your own question. The price you are going to pay is time in any case. I wouldn't recommend the isocratic approach because if the late eluters are ever, truly, present, they will show up a few injections later unless you always run 3 to 5 times longer than the k' of your target. Running a gradient will add time, but remove any ambiguity. If you do not feel compelled to, individually, identify the late eluters, you could take a hybrid approach and program a "step" gradient at a k', or so, after your target compound(s). That way, you would be certain to remove the late eluters, but take less time than a linear gradient.

Hope this helps,

pcb

[Bruce Hamilton]

If the late eluters are related compounds, then your method has to detect them at concentrations down to the specified limits.

Currently, your method seems to be inadequate, so more development work ( eg gradient ) is required.

I suspect you knew this, but were hoping for another outcome.

Some compounds still have TLC as the related compounds method just because some relations didn't elute off a suitable HPLC column during development.

Bruce Hamilton

[tom jupille]

aceto_81, could you clarify what you are trying to do?

I read "specificity against related compounds" as meaning that you are analyzing for your compound of interest and have to demonstrate that you are not mistakenly seeing something else. In that case, if you run your isocratic+ gradient segment as part of the validation and show that those related compounds come out much later, there is no further need to worry about them.

If you meant that you have to analyze for the related compounds, then I would concur with placetoo and Bruce Hamilton's comments.

[reyoungs]

Aceto,

I also think we need to determine if you are at a stage where you can redevelop your method or if you feel locked in. If you are locked in, the your choice is taking time or leaving peaks out and hoping the agency won't feel your impurities are toxic enough to need to be tracked.

My first thought is that the method was not designed for the target analyte and impurities properly. By this, I am thinking that we may be using a method that maximizes retention differences based on the primary retention mechanism rather than using a more selective phase that allows the analyte and impruities to elute closer to each other.

You can sometimes convert a gradient method on C18 to an isocratic method on a shorter-chain phase. This has been well described in literature showing the effect of changing from C18 to C8 to phenyl to C4. I call this "chromatograpic compression," but many folks do not like my terminology.

Have you explored the use of solvent class (ACN, THF and DME or THF) to see if the change in conditions helps change the difficult gradient in to either a sjimpler gradient or even an isocratic method? Extremely powerful if you haven't explored it yet and have the time. (This is a good time to put the method on a UPLC system to learn as much as possible in as short a time as possible.)

If you are not locked in, you can also consider changing to a mixed mode system in which you may have two parameters (lipophilicity and ionic strength) to help control the retention, such than the primary factor controls retention for the entire family of compounds and the second parameter provides the selectivity.

If you do a mixed mode column, I strongly encourage some form of DOE to see what variables control each component of the mixture.

Good luck!

[Consumer Products Guy]

Hi,

when performing a method validation for an API, we have to perform specificity against related compounds.
Now we have a API with a few related components which are strongly retained by the column. So strong that they require an additional gradient to retain them from the column.

How do we justify this to the regulatories?
1. Just inject them in the isocratic method, run for 3,4,5.. times the retention time of the main component and write:"undetected, so method is specific"
2. Add a gradient to the method so that they are eluted. But do we have to use the gradient method for our samples, as the degradation products of our API don't require the gradient (and the related substance don't show up on the API), so the gradient is only a waste of time.
3. maybe other ideas?

How do you guys and girls deal with this sort of problem?

Kind regards

Ace

If you're just inquiring about specificity here, I'd lean towards choice 1.

[Vlad Orlovsky]

if you are not looking to quantify late eluters you can run isocratic method for your API and trap "unwanted" impurities on a guard column. Guard column is switched as soon as your API elutes from it and then you wash guard in opposite direction during your run:
http://www.sielc.com/pdf/SIELC_September_2004.pdf

Regards,

[aceto_81]

aceto_81, could you clarify what you are trying to do?

Oké, here we go:
the method is designed to detect the impurities and degradation products, but for specificity we have to inject also some related compounds. (eg for cephalosporins, there is cefuroxime, cefotaxime, cefazolidin, cefadroxil, ...)

Suppose I made a method to analyse cefuroxime, I can clearly detect the impurities and degradation products in my sample of cefuroxime with an isocratic method. For my specificity, I have to inject the other cephalosporins. Now I found that cefazolidin doesn't elute and is strongly retained by the column, only showing up after a strong gradient wash.
Returning to my question: cefazolidin should be never in my samples, so the gradient is a waste of time, as all the relevant peaks are in the isocratic part. Can this be justified to the regulatories, or should I always run the gradient for my samples.

Thanks for your time

Kind regards

Ace
btw: the cephalosporins is just an example, its possible that cefazolidin is in the injection peak, I don't know, but that's not the quesion

[krickos]

"removed" due to late posting

[tom jupille]

the method is designed to detect the impurities and degradation products, but for specificity we have to inject also some related compounds. (eg for cephalosporins, there is cefuroxime, cefotaxime, cefazolidin, cefadroxil, ...)

That's the way I interpreted it. Your demonstration that these other compounds do not elute in the same region of the chromatogram (or at all!) should be sufficient for specificity. That said, you are still open to questions about how certain you are that these compounds would not appear (as contaminants, for example) in your samples, so be prepared to document your position !