Qualifying a Method on HPLC

[putnam]

I am reviewing validation and qualifications at my new company. I have seen something that doesn't quite make sense to me.

All of the HPLC equipment is the same ... except some instruments have third party column heaters and others have integrated column heaters, and the instruments are disbursed between two laboratories in the same building.

The HPLC methods have been validated and have, for the most part, complete validation packages. However, for some reason, some of the methods have been qualified on each instrument that they are run on. This qualification has consisted of linearity and precision studies - essentially performing a mini-validation, sometimes up to seven times!!

I could perhaps understand this if done on different instrumentation, but why would this have been done if all the instruments are of the same make and configuration? I was told that they had problems with achieving satisfactory results when using a different instrument. Looking back at the method history, I think this was primarily due to a poorly developed method which has since been optimized.

Equipment qualification of this nature has persisted and I just think that it is a waste of time and resources. If the method is properly developed and validated then qualifying it under the circumstances we have in the lab here seems counterproductive. Am I correct?

[tom jupille]

If the method was properly developed/validated/documented, it should have system suitability criteria. These may include retention time or k' targets, resolution limits, tailing limits, and/or plate counts, and will almost always include repeatability for some number of replicates. In principle, system suitability is checked at the beginning of every run.

If you fail system suitability, then "qualification" of the instrument is irrelevant; the results are invalid. If you meet system suitability, then the qualification of the instrument is unnecessary, since you have demonstrated suitability.

The catch is that if you're going to fail system suitability, you do not want to become aware of that fact when you have a hundred samples (or whatever) to be run. That's a good reason for "qualifying" an instrument ahead of time: you minimize nasty surprises.

[unmgvar]

Hy Putnam,
like said before you should have for a properly validated method a section for system suitability check.
the purpose of the check is not for the smaple preparation and for the quality of the separation per say. this was checked in the validation (robustness and so forth).
yet there are enough parameters that change every time you work with that method.
1. the analyst
2. the instrument.
3. the column.
4. the reagents.

all of them can influence your work.
therefore you commit a set of tests that will check all of these parameters.

for the analyst there is generally a check between 2 standards preparations. not more then 2% difference is generally acceptable

for the instrument you generally have an RSD check for the a repeated injcetion 2% for assay tests generally. you also have an RSD check for all the std injected throu out the work or system check between last std in sequence to the first std injections

for the column we generally go for NTP counts, rrt or resolution and assymetry factors, k factor swpecific retention times.

the major problem with most reagent is the grade quality. Not all companies provide clean enough reagents. therefor a clean blank or at least a blank injection with reported system peaks is very important.
the blank injection is also an important matter to prove that your LC does not show carry over on your std peak of interest. generally only a <0.5% carry over in a blank injection is acceptable.

the reason of the SST test is to prove that your equipment is up to the task at end. it is a major suplement to your regular OQ testing. Consider that next time you check your instrument, your OQ fails. does it mean that all your work until then is lost. the answer is no since you have an SST test to support that at the given time your instrument was performing within acceptable parameters to perform the work given to it.
All in all SST testing is a vital part of your daylly GLP environment. you will aslo find that the EP pharmacopeia seriously adresses the subject as well. USP and fda at least to my knowledge also regard it as a good guide lines allthou to the best of my knowledge they have not yet taken a formal step.

[putnam]

Yes, I agree with everything Tom has to say.

We develop system suitability requirements during validation and incorporate them into the method. They must be met before samples are analyzed for the run to be valid.

In addition, the method validation includes intermediate precision in which, among other things, system suitability is demonstrated by multiple analysts on multiple machines (usually two of each, though sometimes three of each if time and resources permit). We are pretty confident of the system suitability parameters by the time the validation package is complete, so I still don't see the utility in qualifying other systems before the method is "allowed" to be run on them.

It is unfortunate that we don't have HPLCs from multiple vendors. All of the gradient methods I have inherited, while properly validated, were not developed with transfer to other equipment in mind. This has become a problem with some recent technology transfers were have been engaged in. Each time the problem has been related to differences in Void Volume between our HPLCs and that of the receiving laboratory (who use HPLCs from different manufacturers). The issues are quite easy to resolve ... but still a pain in the xxx!