Chromatography Forum

Hm, they used to regulate emitted light intensity with slits and double beams, now I thought they did it via double beams and/or curve correction. Varying power to the lamp to level its output .... never heard of it... have to check my manuals more carefully. Or is this completely newfangled?

I can't believe Tom et al aren't in here talking training w/ Wanda (yet).

I agree, Wanda. Your people need more training. Someone needs to pop for a troubleshooting course. Also - Back in the day, chemists were taught how to prepare solutions, use balances, read vernier scales and interpret their results w/ wet chem. tests. Lucky ones got to touch more advanced instruments (once in a while). Now it seems that all they do is watch others do the work and they're asked to do the math.

I've seen recent grads try to prepare mobile phases by stacking aqueous & organic components in the same grad. cyl. I've seen them spend hours worrying about optimizing integration for <10 pixels worth of peak area on a peak that's about 40 stories tall. I've seen horrendous volumetric techniques employed. And I've seen all of the above executed while a cell phone was in use. No wonder I'm getting to be a crabby-old-fart.

I can't believe Tom et al aren't in here talking training w/ Wanda (yet).

That's because we try to be subtle (but thanks for the plug!).

Well, here is another approach, at least for routine measurements...

Have the instrument measure the flow rate!!!

Here is how: Inject a (well defined) unretained peak. Have the instrument measure the retention time. If it is within xxx seconds of the specified time for this column, the flow rate is OK.

PS... I know I am lazy...

I agree with you all, that taking the equipment seriously is not one of the things one experiences nowadays.
But is checking the flow rate and lamp energy really a part of the validation of a given analytical procedure?

In my humble opinion, this should be clarified in the equipment qualification procedure. Provided it is done of course. If not, then one should include many more tests in the validation of the analytical method: E.g. injection volume precision, wave length accuracy, column temperature accuracy and so on – just to mention some randomly chosen tests.

Analytical procedure’s validation should be conducted under the same conditions as the daily routines! Who determines the flow rate accuracy before setting an analytical series daily?

I’m not saying, these things shouldn’t be verified, but it belongs to the periodical control/calibration of the hardware and the frequency of this should be well defined according to the needs of the lab/company.

Best Regards

this should be clarified in the equipment qualification procedure.

Actually, that's a really good point!

Analytical procedure’s validation should be conducted under the same conditions as the daily routines!

I'm not sure I agree with that. Validation should be more rigorous than the routine analysis, if for no other reason that a big part of validation is establishing system suitability criteria. System suitability, in turn, is used to confirm suitability (a bit of a tautology there! ) on a routine basis.

I'll repeat my earlier statement that if retention time is "correct" (or, following Uwe's comment, if t0 is "correct") then the flow rate is, by definition, correct. Having a separate flow rate measurement is useful only if retention times are incorrect.

Thanks everyone for a great discussion.

My main concern was that the chemists performing the validations assume that since their instrument met calibration specs 9 months ago, that everything is still fine. (we calibrate once a year, drives me nuts, but...)

And, everyone addressed my concerns.

Instrument quailification and analytical method validation are separate things. But, they can have overlap because of the system suitably test requirements for the method.

In a discussion I had with an FDA representative, it was made clear that the FDA considers the instrument PQ to be an ongoing process. You have the scheduled (one or twice a year) system performance checks (such as flow rate accuracy) and then you have the routine and non-routine repairs and system suitability results for the analysis. All of these comprise the PQ for the instrument (and, of course, all must be documented.)

The flow rate can be checked as part of the scheduled PQ as an actual mL/min measurement but it can also be checked indirectly as a system suitability requirement by looking for the correct retention time for a peak. Lamp energy is indirectly checked by assuring proper quantitation results (the USP is adding a detector sensitivity requirement for the system suitability test). (Tom made these same devil's advocate comments and they are valid comments/considerations.)

Unless the performance check is part of the system suitability of the method, I would not include it in the method validation protocol simply because these are separate issues.

I think that what needs to be considered is the frequency of the scheduled PQ. It is all a bit of a balancing act. How long can you go while still being assured that the instrument and all its components are operating as it was at its last PQ? You want to be assured that all data collected between the scheduled PQs is valid data. One thing that helps provides that assurance is the system suitability test.

So, in summary, yes performance checks are needed, but unless they are part of the system suitability requirements, the performance checks should be separated from the method validation protocol; keep the performance checks separate and perform them in accordance with the instrument qualification procedure (i.e. in an instrument SOP).

Regards,
Dan

Unless the performance check is part of the system suitability of the method, I would not include it in the method validation protocol simply because these are separate issues.

I agree with Dan. The previous discussions seemed to focus on the importance of instrument performance check, not considering that this requirement is now being put into a method validation protocol.

If it is part of the validation protocol, then the method used to check the flow rate should have been demonstrated to be accurate first. I don't think a stop watch and a graduate cylinder will be acceptable here. I am not sure if Wanda50 will accept 5% variation in this case. at a minimun, used the approach that Alfred88 mentioned (see below). Or follow the specific instrument vendor's procedure. Injecting an unretained peak is not acceptable here (unless you also validate that first).

As a practical note, you need to use traceable timer (get one from VWR, or calibrate your timer with tone from NIST), and calibrated vol. flask for the measurement.

It won't be a simple 15 minute task. it may only take 15 min to get the data. Since it is part of a validation, you need to document it properly, have someone counter sign it. then write it in the validation report, have someone to verify the correct transcription of data. then you need to send it to QA for review.

Wanda50,

does your protocol specified how the flow rate and lamp energy should be measured? if not, you may not get the answer you wanted. is there another SOP in your company that specifies how such measurements should be made?

A method validation protocol should focus on the method. I have not seen any protocol that requires instrument verification first. Those are important considerations, but they don't belong in a protocol.

Now, whilst I agree that Instrument Qualifications and Method Validations should be separate, I'm less certain that simple checks should be abandoned solely because they could trigger some Quality overview.

I'm concerned that too much precision is being applied to what should be simple performance checks that can be performed during routine operation of HPLCs. I'm not sure why anybody should worry about a simple operational check, provided the data was not used to modify the method.

It's not even a system suitability issue, which is more about the chromatography, because flow issues can be caused by many simple events. For example, gas bubbles in pumps tend to be a problem after changing columns and/or mobile phases, especially those with high water and methanol. Having simple checks before starting a run is often easier than having to explain different peak retention times later.

If additional performance checks are specified because the instrument is being used for method validation, then I have to ask why?. Does the nausea of validating methods that many chromatographer feel also get transferred to the instrument, which then spits the dummy?.

According to the IQ on my Agilent 1100, the Flow Accuracy Acceptance Limits for individual pump A at 1 and 2 ml/min are less than, or equal to, 5%.
I've no idea of the HPLCs used by Wanda's people, but a graduated cylinder and stopwatch can achieve 5% fairly easily.

IQ also requires ten flow rate readings with a precision of 0.5%, but that can surely be relaxed for routine checks. If "Quality" insists on making mountains out of molehills, the likelyhood of reaching your destination on schedule will be greatly diminished.

Bruce Hamilton

Taking it out of the protocol does not prevent you from checking the instrument. You just don't need to deal with the paper work, and can do it any way you feel approriate. on the other hand, if you put it in the protocol, you have to document it properly to demonstrate that you have complied with such requirements.

we actually purposely vary the flow rate by +/- 10% during method development to check method robustness. the method should still pass system suitability under those conditions. there's very little reason to worry about small flow rate changes during validation.