Reduced recovery when analyzying Chlorhexidine Gluconate

[Labmonkey]

Just wondering if anyone out there has any experience in analyzing chlorhexidine gluconate?? During the analysis of the method we are experiencing an under recovery using the waters HPLC alliance and H-class systems. The issue was identified during an inter-lab transfer of a method, where we cannot reproduce the results of the orginating laboratory. There is almost a 5% difference in results All materials right down to the lot numbers of column and reagents are the same, the only difference is that the orginating laboratory used a shimadzu prominance system. We have prepared reference and sample solutions at our lab and had them analyzed at the orginating lab and the results generated confirmed the results they obtain. It is only when we run them at our lab we see the difference.

Mobile Phase consists of
73% Methanol
27% Water
2.2g Octane 1-sulphonic acid sodium salt
pH adjusted to 3.0 with glacial acetic acid
C18 µbondapak column
Flow rate: 2ml/min
Wavelength: 254nm
Column Temp: 40C
Sample Cooler: 10C
Run Time: 8 mins
Injection volume: 20µL
Reference Material: chlorhexidine diacetate (recommended as per monograph)
Reference and Sample solution made up in 73% methanol:27% water

Any help would be greatly appreciated, as we cannot seem to identify where we are going wrong, as it is a consistent under recovery, no noticable carryover, peak shape is consistent with little tailing or fronting.

[unmgvar]

where is the difference?
in the peak height, area, concentration?

[Consumer Products Guy]

We formerly used our in-house test procedure for CHG. But we changed a few years ago to the USP procedure, and now use that as written. OK, it's a little cumbersome, but we've been fine with the chromatography and results. And that's one of the few USP assay procedures I feel that at least someone used good science with.

[Labmonkey]

The difference is in peak area, which results in a difference in the concentration when it is used in the calculation. We have a limit of 3.8 to 4.2% w/w for the product, the orginating lab achieved 4.14% w/w and we can only achieve 3.84% w/w at best, for the same batch.

I have run analysis of the solution under the USP monograph specification and am still achieving the same result. I also performed a recovery study and could only achieve 97%, so i believe i am losing it somewhere, but i am unable to determine whether as it appears so consistent, and i cannot see any carryover peaks, had it not been a inter-lab transfer we possibly would not even have picked it up. I have asked suppliers of the HPLCs as to what possible difference there could be, and no one can pin point what would be causing this.

[unmgvar]

so the standard concentration stays the same?
it is the samples that you see that are less in concentration?

2 things comes 2 mind at first
check the filters
maybe something in the sample preparation is not done exactly the same

i would blame a system here only if the chromatogram i would get had a very low or very high signal, then i would look into the linearity precision of the UV.
another thing maybe, would be that the injection settings are not suite for 20ul, it is after all a little more than usual for an H-class system
hope this helps

[cody84]

I'm working on the related compounds for this method (USP) again right now.

I would first check your detector setting between both. Check the data points per peak and see if they are close to get an initial idea if your detectors are doing the same thing. Then I would just try and get them to match and run again. Are you calculating results using external standard and by software or manually?

We use 239 nm for this assay, maybe that will help?

[Labmonkey]

Yes, unmgvar, the standards stay the same concentration, but the samples are lower. I have played with different injection volumes on the H-class and it made no difference to the result obtained. Will have another check of the filters on the system and see if there is something i missed.

Hi cody84, we are calculating both manually and via software (we used Empower 3) but we don't use an external standard. Will looking into the related compound method and see what similarities there are to the assay method, maybe 239nm may help.

Thanks for all the ideas.

[LCFan]

Dear Labmonkey,

I just read in J.Pharm.Biomed. Anal. 14 (1996), 1327-1334:
Published HPLC assays suffer from quantitation problems caused by irreversible adsorption of chlorhexidine onto the silica-based reversed-phase HPLC column. The problem can be overcome by adding NaCl to the mobile phase as described in [8].

And [8] is: Journal of Pharmaceutical sciences, Vol. 75, No. 1, January 1986, 83 - 86

The citations might help.

Regards

Florian

[HW Mueller]

By what mechanism would only the unknown (sample) analyte get stuck on the column, but not the standard?

[Labmonkey]

That is what has me and everyone else i have spoken to stumped, the only thing i can put it down to is the sample is chlorhexidine gluconate and the reference is chlorhexidine acetate, however i thought that the regardless of whether it was in the presence of acetic or gluconic acid ions that the chlorhexidine ion dissassociates to form the free ion and that is what measuring. I cannot explain why i am seeing it, i cannot see where it would be going, and i cannot explain why the originating laboratory doesn't have a problem with it. And i have tried it on different HPLCs ie. shimadzu, waters alliance and H-class and am still seeing it. Had we not had the data from the orginating laboratory i would just have assumed that the assay was low as the chromatograms look fine.

[cody84]

The only time I've had a problem like this, where the standard and samples were different, it turned out it was the syring draw speed (somehow). The only different between the sample and standard was the viscosity I guess, when you consider all the way back to the product itself before dilution.
That was on a Waters Alliance. So it might be worth looking into. My syringe was set to slow draw, but when put back to normal there was no problems. I think the motor itself had something wrong with it, because if anything it should have worked better drawing sample slower. We did EVERYTHING for this...new injector assembly, new sample and standard preps, new check valves, new needle and seal washes, new mobile phases....and when I say new for a lot of this, I mean method development, not new preparations...it was not fun. Anyways, just a heads up! Good luck!

[Labmonkey]

Thanks for the idea, i will look into that, and check out what it is actually set at, i had completely over-looked it. It hasn't been fun, everytime i think i am making head way, i am back to square one. And it is such a simple method i can't understand why it is going wrong.

[magda roth]

We are analyzing CHA routinely with an in-house method: a C18 column, MeOH/Water (acidified with TFAA 0.1%) solvent gradient and internal standard. Got very good linearity, spike recovery, reproducibility. Granted, di not have any method transfer issue: the "advantage" of having only one instrument!
I would make a sample in the lab and check if any of the ingredients may have any small absorption at the RT/wavelenght of interest. Are you using the same column? Could one column be older and have some "junk" absorbed on it? It would react differently with the sample which contains (I assume) lots of other stuff besides CHG.

Good luck

[cody84]

I was recently asked to validate the USP related compounds method on a lot of products...well anyways I'm finding a lot of carryover on a brand new column. Have to run blanks between every injection and do an injection delay..this method is ridiculous, I hate it.

To your problem (if you haven't figured it out) are the columns the same? Guard columns? Inlet filters maybe have something grabbing the CHG in one system not the others? Syringe filters the same for everyone (supplier, not type like nylon)?

[Labmonkey]

Hi there, i haven't figured out the problem yet. I am going down the way of thinking that it possibly has to do with the needle and sample loop components. And that the chlorhexidine is sticking to the needle and then getting washed was during the rinse process and that is why we appear to be loosing concentration and no apparent carryover can be noted.

The only difference between the originating laboratory and ours was the HPLC that was used, i purchase everything else (salts, columns, methanol etc) exactly the same, right down to the same lot numbers.

My current theory is that chlorhexidine is quite a sticky substance and will stick to stainless steel which is what our systems needles are, however in the shimadzu prominence the standard needle is platinum. I am in the process of trying to find out if shimadzu have any platinum needles that will fit into our old Shimadzu HPLCs.