Flow Rate Check for Validations

[wanda50]

Part of my validation protocols is a flow rate and lamp energy check prior to validation experiments. I am meeting with extreme resistence with my chromatographers saying that the yearly calibration should be enough.

should protocols include these checks? I cannot in good faith eliminate that criteria due to years of experience, but I need some hard facts to present to these misguided folks.

thanks!

[DR]

Given the trivial amount of time it takes to execute and document flow rate & lamp energy readings, I would think a simple
"Because I said so ."
or
"Because that's how we've always done it." should suffice.

Seriously, in the likely event of method transfer issues, flow rate and lamp energy allow you to assert, with confidence, that your system(s) were in spec. when the test was run in your lab (so go figure out what's broken in your lab, then rerun your samples).

For gradient methods, system dwell volume should also be known...

[Mark Tracy]

Because a check valve can become leaky at any time, a flow rate check prior to a high-stakes validation study is cheap insurance. Any decent chromatography data system and detector can report lamp energy for every run; put that to use.

[wanda50]

Guys, we're all on the same page. I feel like I'm Robert Pirsig trying to explain motorcycle maintenance to John and Sylvia.

thanks for your support!

[tom jupille]

Let me play "devils advocate" here.

If your retention times are "on target", then your flow rate must be correct.

And if your quantitation (repeatability, linearity, LLOQ, LOD [as appropriate]) is "on target", then your lamp energy must be adequate.

Under those conditions, running separate tests for flow rate accuracy and lamp energy is redundant. Those tests are most useful if you fail system suitability, for finding and fixing the cause of the problem.

The real question is "do we run those tests 'preemptively' before validation or wait until problems occur and then run the tests".

That said, my vote would be that it's more cost-effective to run an hour's worth of tests "preemptively" rather than have to re-run all the calibrators and replicate samples when there's a problem. Tell them it's like buying insurance: they're trading off a consistent minor annoyance in return for avoiding an occasional catastrophic trauma.

[Bruce Hamilton]

Maybe you could approach the issue a different way.

Discuss the issues. Lamps have nominal lifetime 1000-2000 (long life) hours - can be longer if normally used at wavelengths of maxium energy, and less if used at wavelengths of lower energy. Output energy degrades gradually over that time.

A typical employee's working year is about 1600 hours - not even allowing for overtime or overnight instrument runs, so ask them how they are going to assure that the lamp is performing to specification with annual checks?.

The flow rate issue is a little more complicated. If the peaks appear at the expected retention time, that is a reasonable indirect indication that the flow is OK, assuming they put the correct column in and the mobile phase composition is correct - something that pressure monitoring may also be showing.

But those are assumptions, and isn't it easier to actually measure the flow?. Gas bubbles in pumps are pervasive and wilful, usually appearing because people are in a hurry. It's all about risk management, and ensuring that the minimum energy path is followed.

Consider reviewing why flow measurement is causing them grief. Routine flow rate measurement should be quick and easy, is yours?.

Surely a stopwatch, measuring cylinder, and a few minutes would be adequate. But if somebody has specified a painful procedure of individually calibrating Pump A, B, C etc, then they may well become more revolting. Consider the information required, and only use simple procedures to confirm the performance.

Annual tests are intended to systematically requalify the instrument, and will inevitably be more painful than simple confirmatory tests.

Bruce Hamilton

[wanda50]

I really appreciate all the views on this question.

My honest opinion is that the chromatographers believe that they have the best equipment available that will never get air bubbles or faulty check valves. They do not know how to troubleshoot or repair simple things. One person was amazed to find out the average lamp life - didn't even have a clue they had a "life". I feel that it is more of a "head in the sand" situation, assuming that everything is fine until a problem occurrs, and then it is panic time.
In one case, a chromatographer made up about 6L ACN containing MP, and didn't degas it since there is a degasser on the instrument. Let me tell you what a mess that was!

So, everyone is confirming my gut feelings, and is showing me that there needs to be some more training.

Thanks

[Bruce Hamilton]

In one case, a chromatographer made up about 6L ACN containing MP, and didn't degas it since there is a degasser on the instrument. Let me tell you what a mess that was!

So, everyone is confirming my gut feelings, and is showing me that there needs to be some more training.

I obviously can't comment on who needs training, but are you listening and responding to their reasons, or just seeking confirmation of your position?.

With regard to degassing mobile phases, I haven't for the last 4-5 years, as the whole point of having degassers on the instrument is to perform that role. If the degasser isn't working, maybe it should be fixed.

Just in case somebody wonders if I achieve the same result by vacuum filteration, I don't filter mobile phases that consist only of miscible HPLC grade solvents and/or Milli-Q water. Only mobile phases containing buffers or IP reagents are filtered.

Bruce Hamilton

[wanda50]

I was seeking guidance for why flow checks are done, or even if others have the requirement in protocols. If the requirement is there, how do you answer a question of "why".

No, it was not confirmation of my views that I was seeking. I was just so astounded that anyone would question such a simple 15 min action that I couldn't come up with a reason why. As someone said, it is so little time to assure that a validation will go forward.

[Consumer Products Guy]

So I'm not the only one who has to deal with protocols from a QA department that likes to detail all possible issues....tell them you can use a graduated cylinder or a balance to calibrate...if gravity is working that day so the eluent doesn't float away...

[Alfred88]

Dear wanda50:
If you need guidance document, I would like to quote from an FDA document, "Validation of Computerized Liquid Chromatographic Systems":

Pump
Are flow rates accurate and reproducible?

(source: http://www.fda.gov/ora/science_ref/priv ... /jaoac.htm)

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.

For CPG: Our QA/QC Manager is one person (non-chemist). Our former QA/QC Manager once observed that I was filtering mobile phase and asked, "what are you doing?" I told her I prepared MP for an assay on HPLC, and she asked, "What is MP? What for?" (She believed that HPLC could run by turning a switch like a light bulb!). She had left to become a senior auditor for a large co. And last year, we hired a QA specialist. The young woman asked me to show her our HPLCs, because she never saw one. If your QA people are able to write protocols for your dept., you should be proud of their expertise!

Alfred.

[Bruce Hamilton]

Dear wanda50:
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's too easy to get distracted with volumetric flasks, you only get one shot at the measurement. I prefer a 5 or 10 ml with 0.1 ml/divisions graduated cylinder. It's easier to do other things, and come back after 4 - 8 minutes to record the volume at the next full minute - assuming flow is 1 ml/min.

I expect a cylinder with 0.2 ml/division would be fine as well, provided the procedure also says to sample more that 4 mls, and that acceptable measured flowrate variation is 5%.

Weighing just allows quality people to raise other issues, so measuring volume directly is the minimum energy path. It's important that procedures are easy and flexible, as well as accurate, so that expensive instrument and staff resources are not wasted.

I'd say "please keep having fun", but....

Bruce Hamilton

[Consumer Product Guy]

Alfred88 - you've misunderstood: my QA department can't even spell "HPLC" - but that doesn't stop them from writing protocols.

[HW Mueller]

This has been mentioned before: I need to check the flowrate quite often so I just cut the tip off a graduated glass pipet, attached a T-connection at the bottom so that it can be hooked up behind the detector and a air bubble be pushed in....takes about 20 sec. for a flow rate determination.

Bruce, I wonder about this lamp life depending on the wavelength used?

[Bruce Hamilton]

Bruce, I wonder about this lamp life depending on the wavelength used?

Oops, I probably shouldn't generalise so much, as different UV detectors will behave differently, and later lamps are supposed to be superior to older models.

My limited experience has been that if you operate a UV detector with transparent solvents in the region of maximum lamp energy ( 210 - 250 for lamp, combined with 230+nm for typical solvents, yielding ~ 230 - 250nm ), the lamp life consistently goes well past the nominal, eg to 3000 hours for longlife, 1700 for standard.

If I routinely operate in a region where mobile phase reduces the energy to the detector, and the lamp has lower energy ( eg 205-210nm ), lamp life can be less than nominal, down to 1900. This year, the longlife DAD lamp failed at 1998 hours!.

The local rep also commented that lamp life appears to correlate the wavelength used, and I've always assumed it was related to detector energy feedback affecting the temperature ( which I believe is the critical parameter for lamp life ). I haven't checked the literature to confirm my assumption, and I may have inappropriately extrapolated my experience.

Sorry for any confusion.

Bruce Hamilton