HPLC system suitability

[wrighty]

Afternoon all,

i am in the middle of war of words at the moment regarding one of our HPLC instruments system suitabilty method.

The method in question involves the analysis of marine oil via GPC. Their is a system suitabilty standard which caontains a mixture of TG-DHA, DG-DHA and MG-DHA (which elute in that order). It has been noticed that a small peak is eluting between DG and MG and this is what is starting all of my troubles.

I need to quote the resolution between DG and MG but this little peak is not letting me do this so my options are:

1: Quote the resolution between DG and the mystery peak
2: Quote the resolution between the mystery peak and MG
3: Re-process the trace so that the mystery peak is "ignored"
4: Add the resolutions obtained in choices 1 and 2.

In my opinion, if a peak is present between DG and MG then that should form part of the resolution and certainly not be ignored but it is trying to get my point across. Does anbody know where i can find clear guidelines into what to do with this situation?

Additionally, somebody else wants to change the system suitability method so that the integration parameters ignore the peak. But from my understanding, the system suit method is used to show that the system is performing ok and that the method the sample is being run on is performing well.

Sorry about the rant, i hope this is in the right section. if anybody can provide some guidance then that will be truly appreciated.

Thanks

Steve

[tom jupille]

First of all, I think your post isn't a "rant", but rather a clear statement of the problem.

Second, you're in the right place, even though the problem is more a regulatory issue than a chromatography issue.

I think the answer hinges on how tightly regulated is your working environment/industry, and on the significance of that "extra" peak.

From a straight bureaucratic perspective, if system suitability is defined in a certain way, and you meet system suitability, your system is suitable (don't you just love the tautology in that!). That suggests that your option 3 would be the way to go (reprocess, ignoring the extra peak). From a scientific perspective, this would be justified only if you "know" that the presence of an extra peak will not affect your results. What you have to do to get that knowledge will depend on the regulatory tightness (you would have to document it a lot more thoroughly in pharmaceuticals than you would in petrochemicals, for example).

If that extra peak can affect the results, then you have a more serious problem: the existing system suitability test is inappropriate. That means modifying the system suitability test to incorporate the additional peak (i.e., either your option 1 or 2 -- or both!); depending on your SOP's and (again) on the level of regulatory tightness, this may involve and extensive revalidation. In the meantime, carry out your options 1, 2, and 3 so that you're not hiding anything when the fæco-ventillary collision occurs (1 and 2 give you information about the actual "worst-case" resolution, while 3 gives you the information relevant to the current system suitability test).

Option 4 is a no-go in any case; unless the peak widths are exactly equal, the resolutions are not necessarily additive.

[DR]

I'm pretty much in agreement -

Provided the existence of the little peak is not going to cause the standard material itself to be called into question (I'm assuming that the new peak is either an impurity or degradant in/of a standard component), option 3 is the way I'd go.

If that doesn't silence the critics, procure new standard material or, if the new peak isn't coming from the standard, check mobile phase cleanliness, clean and passivate the system, replace the column (probably not in that order) etc. etc. until the source of the peak is found and eliminated.

[Bruce Hamilton]

Has the small peak always been there?.

If so, Tom has pointed out your options - the documented basis of acceptance could be as an improvement to the protocol - that shouldn't trigger any investigations...

If it has just or gradually appeared, your standard is stuffed, and you need a new standard. This would be a larger can of worms, as you would have to redefine acceptance criteria for your standard to prevent such issues in the future..

Personally, I would be quite concerned about an unallocated/unidentified peak in a DHA lipid standard, as the possibility of analyte degradation is quite high for such a polyunsaturated lipid.

Please keep having fun,

Bruce Hamilton

[wrighty]

thanks guys.

i've identified the source as being an impuity of the MG-DHA. well, when i say impurity it is 1% and the relevant cof a quotes the purity of >99% so i cant question the appropriateness of using this standard.

this standard type has been used in this method since year dot. to the best of my knowledge this peak has always been present and we (the instrument technicians) have only found out that there is a problem now that the odd system suitability is failing.

i am currently trying to separate the impurity from the MG-DHA and it is running overnight (UK time) so the plan is to remove this extra peak from the trace. i'll post an update when i'vegot some data.

thanks

steve

[krickos]

Hi

Would agree with Tom and the other speakers.

The key issue is first to find out if this mystery peak is a by product only present in standards or actually can be present in your sample and then act accordingly.

[wrighty]

Finally, i can put an end to this one. The bump is exclusive to the standard but as this is does not impact on the certified purity of the standard then there is no contamination qualms. After extensive testing, this bump does not appear to any samples tested.

The conclusion of this work is that the chemstation integration parameters have been altered to "ignore" this peak.

Everybody is now happy, i'm just waiting for the next crisis

Thanks for all you r contributions!