FDA auditor he%$

[featherland]

I'm the head of QA for a contract lab. Please put down that shoe.

An FDA'er is going over our data and insisting that once you run a "STOP" sequence (eluent flow stops) during an HPLC run, you have to repeat full system suitability. Instead, we reinject the system suitability standard and calculate RSD based on the last five injections (including the one after restarting the instrument the next morning). If the RSD is acceptable, the samples are analyzed.

She insists that all five injections must be acquired after the eluent flow is restarted.

Does anyone have a reference to some kind of official guidance on this matter? USP <631> gives very little guidance on this, of course.

[mohan_2008]

If you are a pharmaceutical CRO - Yes, the FDA person is right.

And things are worse - if you are a c-GLP/c-GMP oriented facility.

All guidelines ICH, FDA and OECD suggest that a system performance criteria need to be met, prior to acquring any data.

A system suitability confirms an appropriate system performance at the time of USE, and has to be performed everytime when you initiate a sample acquisition.

Which means, you need to inject all standard injections used to calibrate your samples/ precision injections when you stop/or start an acquisition.

[Dan]

I think that you will find that many will be on the side of the FDA representative.

I am not aware of any guidance that will provide direction on this other than internal SOPs. The FDA may even have their own internal document for this issue.

I have been in the GMP lab for many years and I am now in QA. Internal SOPs address this. Those documents describe situations for which a change can occur and the system suitability does or does not need to be repeated. It all depends on the change. Stopping the flow is a major change.

Basically, it is as Mohan stated:

A system suitability confirms an appropriate system performance at the time of USE

If you make a major change to something in that system, such as stopping the flow (or even slowing the flow), then you have changed the system and the previous system suitability is no longer valid after the change. Addition of more of the same batch of mobile phase is not a major change, so no need to repeat the system suitability for that.

It can all be in how you design/write your SOP. Of course, the FDA can still not agree with your SOP.

Regards,
Dan

[mohan_2008]

Internal SOPs should guide through the System suitability requirements as appropriate.

But still, a Pharma company should incorporate the FDA guidelines or appropriate regulatory concepts while designing their SOPS - failure to which can be a potential problem for the organization.

[krickos]

Hi

Would agree with the other speakers. As touched on by another, the only way I can see that you can limit the number of SST injection is if you refill the mobile phase from a stock solution and keep the system running and "bracket" SST solutions to monitor that everything is OK. FDA inspectors tends to like bracketing

As less funny comment on the related issue: My current group took over a bunch of responsibilites from the R&D department in around 2000, knowing that SSTs was not optimal. We had started sorted a number of them but still we got hit by a FDA inspector just a year later (a lab manager on top of that), so in the end we did adjust/try/document/implement SSTs for 110 different analytical methods.
Lets just say we could have spent the time doing funnier things.

[featherland]

I'm just stunned that you folks appear to support the point of view that pausing eluent flow and turning off the lamp for the night is a "major system change".

So your companies spend up to six hours running system suitability *every day* before you can run two samples and call it a day?? This is just not making sense to me here.

Still interested in any official FDA guidance here. Found a reference to "splitting" system suitability (an acceptable practice), running half before and half after the samples. The premise was that because split system suitability is actually harder to meet this way, that this supports demonstrating system suitability throughout the run. That same argument would be true of multiple day runs, wouldn't it? If the last five consecutive standard injections are meeting system suitability requirements, don't I have de facto proof that the system is still in control, despite the fact that we only work a single shift here?

We are not a Pharma company or CRO, we are a third party contract lab (about 50% Pharma work, 50% environmental and consumer products), so I am writing these SOPs from scratch.

[krickos]

Hi a few FDA links/references:

http://www.fda.gov/cder/guidance/2396dft.htm

A draft from 2000 of Analytical Procedures and Methods Validation. Yes it is a draft but there are evidence (warning letters, 483s) that FDA enforces certain parts of it (blank titrations, SSTs for chromatographic methods etc).

One quote:

The RSD is normally performed at the beginning of the run. However, for assays with
lengthy run times or as otherwise justified by the applicant, the reported average may
be taken from injections at the beginning and end of the run, or beginning, middle, and
end of the run.

Also if recalling right the draft also refers to the old Reviewer guidance "Validation of chromatographic methiods" from Nov 1994.
(http://www.fda.gov/cder/guidance/cmc3.pdf)

[Peter Apps]

I'm no expert on regulation (thank goodness) but I would have thought that this could be covered by ruggedness testing during validation - prove that shutting down and starting up again is seamless from the outset, and then overnight shutdowns will not count a system changes.

Peter

[unmgvar]

from my experience,
the main problem that you need to face first is semantics in your SOPs,
if you have written that in such cases SST needs to be repeated then yes you will need to reinject all the SST (including the 5-6 injections)

an interesting question to you is based on what you pointed out regarding long batch works of several days, what do you implement in order to "monitor" the system after the SST has been injected several days back:
do you have a general SOP regarding:
retention time deviation?
do you inject standards in between the work, at fixed intervals; have you set a test for the deviation and what to do in case of deviation problems?
have you check the stability of the standards preparation during method validation so has to show that indeed you can use them over a period of time safely?

the 5-6 repeated injections come to see system repeatability which will mainly show the sampler behavior but not only. some methods require long equilibration before column behaviours gives repeatable chromatograqphy which can otherwise influence results.

remember that after you stop the flow and turn off the lamp you do re-equilibrate your system in order to work again, SST parameters are there to show that you have appropriate system performance so straight forward you will need to re-check the SST.

still if you create an adequate SOP that covers a range of possible problem and solution then you can do with only a 'partial SST" in very specific cases.

we never stop the flow entirely but go to a lower flow. we never stop the lamp or any other module. we never wash the systrem and column for storage.
especially stopping the flow with buffers means that you might clogg the system and or column which can cause problem that will requires troubleshooting and do effect sometimes the column stability.
when continuing work we reinject all SST solutions once again and the work standard at least once more, and after a certain interval we reinject the standard always and we close with a standard.
RSD needs to never go above 2.5%(for assay UOC, DIS) else we stop the work and recheck completely the SST including 5-6 injections
retention must not deviate by 5% also over the range of the work. otherwise we recheck the entire SST.

[Bruce Hamilton]

I'm just stunned that you folks appear to support the point of view that pausing eluent flow and turning off the lamp for the night is a "major system change".

So your companies spend up to six hours running system suitability *every day* before you can run two samples and call it a day?? This is just not making sense to me here.

Unfortunately, the "4Q" model has become so pervasive that obvious exceptions instantly trigger angst from internal QA and regulators.

The "new" FDA is supposed to interduce risk assessment into processes, but the entrenched quality systems are not going to yield easily.

For example, the FDA has finally recognised ISO 17025 laboratory accreditation for some analytical roles, but many of the trainers now leaping on the bandwagon are pushing versions of the "4Q" model, rather than the internal assessment and result uncertainty determinations that 17025 permits. They are
overspecifying ISO 17025 requirements.

If your laboratory is performing both cGxP and research work, your systems will be determined by the contracts with clients. You have to meet their needs, but it's possible to run variable intensity Quality systems cost-effectively.

The obvious answer for system suitability is to select tests and stds that can be quickly analysed using the appropriate number of replicates.

In large companies, the solution is to purchase more instruments; but for small companies, the solution can be to simplify the protocols, and match them to client requirements.

You have to document the processes, but auditors are more accepting of simpler tests than no/inappropriate tests.

Turning a pump/detector off and on is a major event, but I would go for faster SSTs at the start of each run, rather than continuing with low flow settings etc.. D2 lamps only have 1000-2000 hours life.

If you want more details about HPLC quality systems and regulators, I highly recommend www.labcompliance.com

Please keep having fun,

Bruce Hamilton

[mohan_2008]

Lorraine,

Make sure you keep the lamp on, even if you stop the flow if you insist on doing an intermittent system suitability.

Area response/counts are a major factor that may differ, when the lamp is turned off, and back on.

If the flow is stopped momentarily it should be O.K. because it shouldn't technically affect any output, since the column would be already equilibrated.

This is stringent with PDA detectors, where warm-up time is significant.

If you need to convince an FDA guy, your argument should be scientifically sound and should clear any queries.

[Bruce Hamilton]

Make sure you keep the lamp on, even if you stop the flow if you insist on doing an intermittent system suitability. ....

If you need to convince an FDA guy, your argument should be scientifically sound and should clear any queries.

I disagree. She is apparently turning the instrument off overnight. In which case, leaving the lamp on is probably a waste of lamp lifetime, and will have minimal effect on whether any regulator/client auditor would accept an unconventional SS.

I would be loathe to accept a SS performed on the previous day when the instrument has been idling for several hours, unless that behaviour had been part of the initial qualification protocols - such as for a long IPC method.

She's in a contract lab, the quality contracts with clients should specify what is acceptable, if not - she should be asking the client's Qulaity team. If they say "conformance with FDA requirements", then the full SS should be performed as specified, and the cost passed to the client.

If they say "research only", then she can follow in house protocols that the client's Quality people have approved. Such procedures can have much more flexibility with regard to qualification, method modification etc. etc.

Bruce Hamilton

[bakej0]

I'm with you Lorraine. I work in QC support for a contract manufacturing pharmaceutical company and we do not repeat full system suitability after stopping the flow. Having said that, we tend not to stop the flow between analyses although we sometimes slow the flow rate to conserve mobile phase.

We reinject the standards from the initial SST before and after our test samples and check the RSD of all standards back to the initial SST. In my opinion, this is sufficient to show that the system has not changed since the initial SST.

[mohan_2008]

That is right.

The problem here is, we need to find a way to convince the FDA.
If you stop the flow in between analyses - there is no question of system suitability.

System suitability implies the system performance monitoring the entire run - including Bracketing standards.

If you stop it - you need to do it all over again!