Broad peaks in middle of gradient chromatography run?

[Rob Burgess]

Hi All,

I have a bit of a chromatography problem and am after some comments / thoughts from the wise users of this board. The chromatogram above shows the problem I am encountering. With this particular chromatographic impurity method we see an unusual large "broad lump" midway (tR ca. 9 mins) through the run for our degraded drug product (DP) samples (black chromatogram). Our degraded drug substance sample (DS) (green chromatogram) does not show this "broad lump" and nor does the blank sample solvent (maroon chromatogram).

The drug product is a lactose-drug blend that is filtered. A filtered blank does not show this "broad-lump peak either".

Could this broad lump of peaks be a whole bunch of unresolved peaks. It just looks very strange. Any ideas what could be causing this anyone?

PS> if you require further info fire away and I'll get back to you. Cheers.

[Bryan Evans]

Could be a number of things - air bubble in detector, problem with
autosampler needle, local glassware contamination (residual detergant from glassware washer), HPLC septum being pushed in, ect.

Is this for the pharmaceutical industry? Do you have stability data on the product, and are you within known sample stability? Did your bracketing standards pass system suitability? When we had extra unknown
peaks in a run that passed system suitability - we did the following:

1. Initiate an investigation (OOS) and document everything
2. Re-inject a sample that gave an "in spec" result (i.e. your first sample
prep) - this serves as a control - as long as your initial and re-injected
"in spec" samples are comparable, it is justification to accept original
results from your 1st run.
3. Re-inject the sample in question (from the original HPLC vial), along
with the original working solution, as well as re-diluting all stock
solutions. It's a way of backtracking and trying to figure out
where the extra peak/local contamination came from.
4. At some point - you should get a chromatogram similar to your 1st
sample prep.
5. Having 2 sample preps is great also because if the extra peak were
sample related, it would occur in both peaks, not just in one replicate.

Hope that helps!

[tom jupille]

Rob, you've probably already eliminated this, but just to be sure: one possibility is late-eluting garbage from a previous injection (that would explain the breadth of the peak). If you always see the peak in a sample injection that was preceded by a blank, and never see the peak in a blank that was preceded by a sample, then you can eliminate it as a possibility.

Your speculation that it is an "envelope" containing a distribution of related compounds is defintely possible. Easiest way to check would be to make a big (concentrated) injection of sample, collect the broad peak and reinject.

I don't know what to make of it, but I'm intrigued by another anomaly in the sample chromatogram: the negative dip around 4.5 minutes.

[Bryan Evans]

I've seen "lactose adduct formations" which were large humps
like that created when the product was heated in an acidic (or basic
environment, can't remember). Do both chromatograms
contain lactose - or is it just the one with the large hump?

[Uwe Neue]

Rob, It looks as if you are running isocratic. If that is the case, I agree with Tom: your peak is probably a late eluter from an old injection. A brief wash in between injections should help.

[Rob Burgess]

Dear All, Thanks for your comments.
To clarify, this is a gradient run with an isocratic hold at the end of the run. The gradient profile is:

T/ min %B
0.0 30
8.0 70
17.0 70
17.1 30
20.0 30

To make clear we only see this peak in the drug product runs. So far I have seen it in all our dp runs, never in our drug substance samples, and never in the blanks. Thus it appears like it could be a real sample related peak. The green chromatogram (ds) does not contain lactose. Re-injection confirms original result for all drug product injections. Sample preps are prepared in duplicate (8 samples altogether) and this strange peak phenomenem occurs in all of them. Obviously this trongly indicates that it is a true peak coming from the drug product samples.

An orthogonal HPLC method seems to suggest that any lactose adducts elute before the main peak, not after as in this case.

[Rob Burgess]

Dear All, Thanks for your comments.
To clarify, this is a gradient run with an isocratic hold at the end of the run. The gradient profile is:

T/ min %B
0.0 30
8.0 70
17.0 70
17.1 30
20.0 30

To make clear we only see this peak in the drug product runs. So far I have seen it in all our dp runs, never in our drug substance samples, and never in the blanks. Thus it appears like it could be a real sample related peak. The green chromatogram (ds) does not contain lactose. Re-injection confirms original result for all drug product injections. Sample preps are prepared in duplicate (8 samples altogether) and this strange peak phenomenem occurs in all of them. Obviously this trongly indicates that it is a true peak coming from the drug product samples.

An orthogonal HPLC method seems to suggest that any lactose adducts elute before the main peak, not after as in this case.

[Bryan Evans]

Rob,

Sounds like you're going to have some fun with this one!
You can see some degradates in your drug product samples,
but it doesn't look to be too bad.

Under what conditions were these samples stored at? Are these samples
capsule, tablet, liquid, ect. What packages are they in? Did you see this
same peak in all packages?

I've seen some very large unknown peaks grow from
IV samples under stressfull conditions (i.e. 40c/75%). Usually
this occurred from certain rubber plungers/caps dissolving in solution under stressfull conditions.

Usually have not seen these large extra peaks from solid dosage
forms - but that's not to say it hasn't happened.

[Ary]

Rob

For what its worth....

Peak isnt in the blank or DS but it is in DP. This suggests some sort of interaction between the matrix and the DS. Logical next step would be to dose the DS with the major components of the matrix to see if your peak reappears.

I have seen adducts form these kind of broad peaks before so just guessing this would be what I'd expect you to find at the end of the trail as you seem to have eliminated all the other obvious stuff.

One last thought is that you could be getting some weird pH type effects from the degradation product(s). You dont say what your eluent buffer conc is so that may be something to play with???

[Yorkie31]

Hi,

It definitely looks like an excipient related problem in your DP.

I presume you've performed some form of specificity experiments in the validation of this method?

I think as someone previously suggested, it would definitely be easiest to prepare solutions of the the excipents at the working concentrations to see where they elute.

You could also then spike these into a DS solution to see if there is some form of interaction between the DS and the excipients.

Anthony

[DR]

How do the placebo formulation samples look? (You did put matrix-active on stability, right?)

If there's nothing similar in the placebo, then you've probably got an interaction. Time to do a DOE w/ constant API, hi & low levels of ea. excipient (relative to formulation) - store under the conditions that bring about the broad peak (or a little worse for less time) and assay everything.

My first guess is that there's an exposed amine or other somewhat reactive functional group engaging in some lactose chemistry (Mallard rxn). If do, time to either reformulate or figure out a process that keeps lactose away from API.