Check Standard duplicate injections are very different

[JasonWellsSNN]

Hello All,

Thanks for taking a look at this and seeing if you can help, I really appreciate it.

I have been running a fairly simple HPLC analysis determining Preservative Content in our product for the last 2 years or so. We are quantifying M-Cresol, Methylparaben, and Propylparaben using a water:ACN gradient and just water to dissolve the samples, and methanol and water to dissolve the STD and STD Check preparations.

I run 2 blanks, 6 STDs, and then 2 STD Check injections, and another STD. The specification for all 3 compounds is 98-102% and the M-Cresol and Methylparaben have been doing fine, but the Propylparaben keeps falling out of that range, usually 103-104%. The issue I keep seeing is that the duplicate injections of the STD Check give very different Peak Areas for Propylparaben, but not for M-Cresol or Methylparaben, and then the average makes the calculation fall out of spec. I get about 169 for all the STD injections, with RSD passing easily even for Propylparaben in the STD injections, but then with 2 duplicate injections of of the STD Check, I will get like a 182 and 169. If both injections were 169, STD Check would pass and everything would be fine. The STD Check is prepared the same way as the STD and then loaded into 2 different vials from the same syringe and filter for duplicate injections.

I have had this problem the last 4 times of trying this assay and I have no idea why this keeps happening. I am making sure there is no vacuum so it won't be an injection issue, all mobile phases and reagents are fresh and new, and the issue is basically just on 1 peak of 1 injection in the entire 18 injection run.

Any help would be great since I am lost as to why this keeps happening. Thank you!

[Consumer Products Guy]

I have been running a fairly simple HPLC analysis determining Preservative Content in our product for the last 2 years or so. We are quantifying M-Cresol, Methylparaben, and Propylparaben using a water:ACN gradient and just water to dissolve the samples, and methanol and water to dissolve the STD and STD Check preparations.

I've had tons of experience with preservatives in finished consumer products by HPLC, GC, and GCMS. We'll stick to HPLC here, and my experience with chloroxylenol, methyl paraben, and propyl paraben.

I would use mobile phase with 0.5% acetic acid or 0.5% phosphoric acid in place of the water.

I also suggest using methanol as the solvent for all; propyl paraben especially is not that water soluble (some of our formulations had issues with propyl paraben precipitating, but that's a different issue).

The specification for all 3 compounds is 98-102%

And if you're in USA, your specification range for all 3 compounds is 98-102% is tighter than it needs to be. ORA 4.5 regulations permit recoveries of active in finished products to be 95 - 105%, so preservative specifications don't need to be tighter than the active.

[HPLC chemist]

Your method is comparing 'apples and oranges'! You have to treat both the samples and standards the same so they are the have the same chromatographic profile (this may even mean spiking samples). You may also have to use IPA instead of ACN or MeOH in order to sharpen the late eluting peaks.

[JasonWellsSNN]

To be clear, the STDs are dissolved in MeOH first and then diluted to volume with water, and the samples are digested with cellulase first and then diluted to volume with water only, no MeOH. I sonicate the STDs as well before adding any water.

Unfortunately, this is a validated method and I cannot change Mobile Phase, preps, or the specification. I wish they would have left the spec at 95-105% but that's not what is in the method and filling.

I guess my big issue is that it seems that what is coming out of the syringe and filter ends up being different between the 2 vials, and I do not know how to fix that, even though I have run this method many times before these issues and have had minor STD Check issues before with Propylparaben, but not ongoing issues like this.

[Consumer Products Guy]

I guess my big issue is that it seems that what is coming out of the syringe and filter ends up being different between the 2 vials, and I do not know how to fix that, even though I have run this method many times before these issues and have had minor STD Check issues before with Propylparaben, but not ongoing issues like this.

I'm guessing some propyl paraben is precipitating out, and then being caught by the filter.

Just my knowledge and experience talking, though.

I suggest to modify and revalidate the procedure. Being tighter (98-102%) than required by regulations is NOT a good thing. Yes, I had to convince our own QA Director he was wrong when he wanted to stick with 98-102% recovery.

Of course, I later had to fight with my own supervisor who wanted me to have 95-105% recovery for levels of analyte under 10ppm unless I could find written documentation allowing for more leeway. I did find guidance, but that was in veterinary pharmacopeia so he dismissed that. That procedure ended up never being validated because of that/him.

[BostonFSE]

Rsd issues are typically caused by the sampler. Examine the rotor seal, the groove that connects the sample loop can be partially clogged or deformed. This causes inconsistent peak area counts and poor RSD
To test if hardware related. Inject caffeine at 273nm isocratic with water. Overlay the results, if peak areas are inconsistent change rotor seal and possibly stator face (or sonicate both in water:meoh for 10 min)
If peak areas are good with simple caff test, its a chemistry/method issue
The pump seals and rotor seal are the most important part of maintenance and should be replaced annually for best results

[Consumer Products Guy]

Rsd issues are typically caused by the sampler.

True. But see below.

I guess my big issue is that it seems that what is coming out of the syringe and filter ends up being different between the 2 vials, and I do not know how to fix that.

Check out the autosampler itself as Boston suggested. But in this instance, your preparation might be causing more propyl paraben to be filtered out sometimes; that's why you may be better off validating a change.

Have you tried your end injections from the original vial?

[Ice9]

I had a similar issue which was much large. On a sample with 5 components, one was coming out at a greatly reduced amount.(@75%percent) it was system contamination and cleaning/passivation resolved the issue.