Weekly performance verification

[qfiviera]

Hi again guys.

In my lab (well, it`s not mine, I work in it) I have to check every Monday 14 HPLC as a weekly performance verification. I wash them to remove solvents or salts first from all channels. So then I run 5 blank runs to determine the noise, and 5 runs of an 5 mg/mL in methanol of a caffeine solution, column 125 x 4.6 RP 18, MP Water:Methanol (53:47) taken from manuals. This way I`m checking all modules, and put all results in a control chart. I have an specific column and an specific channel in each HPLC. Noise, retention time and Area %RSD are checked. We started this year. To fix limits in control chart, we took results from about 20 Mondays and use statistics (+/-2 sigma and +/-3 sigma from mean value). But during the year, the equipment need maintenance so you change old lamps or perhaps some PEEK tubing broken. And the values changes. For example Noise reduces, or retention times changes a bit, but for Caffeine (about 1.7-1.9 min rt) it is significant.
So, what do you think? Is there a better method for a weekly PV or shorter because it takes me all Monday? Or may I have to recalculate my charts with a values from all year so I can introduce all the variables (consumables changed)? What about the columns? Because in one point I will have to change them, so will I have to build new charts?

I want to hear some ideas from anyone. Thanks for your time.

[unmgvar]

how do you know that the column is not faulty?
these days companies use a long tubing instead of the column, that creates back pressure as well
you do not say if you are checking for the gradient of the pump, you are assuming of course that the flow rate is correct since you are using a mobile phase mixture
these days the runs are made with a 100% of a pure solvent like water or MeOH, in that case with the combination of the tubing instead of the column, the RT for a given instrument should be closely the same and will allow you to check flow accuracy
it looks like there are many more tests that you do not do in that PV. and why do you do this procedure every week? does it help you to reduce trouble shooting of the instruments?
overall every HPLC method has a set of System suitability procedures that if within range show that the system and column are working within the needed range to perform the test
we check a full PV every year only, after changing all the consumable parts like pump seals. the valves and lamp are changed when needed only.

one thing that i think is great that you do is that you do a pre-defined washing of the system. this always helps as a preventing action, i would keep it.
the RSD of area for the sampler should be checked each time so i would not do it every week as well further.
the noise is more important for the impurity applications, this is also column dependent, so there should be a test within the application with a LOD or/and LOQ solution for each application
the lamp itself should be capable of running at least 2000 hours, and generally they go 4000 and a little above. all instruments can also give you the energy of the lamp under a given condition. monitor that based on the age of the lamp and change if necessary based on the supplier technical notes.

washing the instrument from salts, especially the pump, the sampler from left over sample residues and the flow cell reduces downtimes and troubles, this way you also force the removal of old left over mobile phase bottles that is generally left open and that accumulates dust and other things that go in the frits, filters and column after that.
again we rely more on the SST requirements of the applications and do a full PV once a year.
keep washing the HPLCs

[qfiviera]

First of all thanks for your answer umgvar.
I work in a Pharma lab. Well, all of this came out because auditors asked us how were we sure the instrument is working properly at the start of the week. Of course we do like everyone anual maintenance and OQ/PV which are more extensive than this and change consumables when need to be changed. So this is only a way to check the entire instrument, to check it does not appart from the normal behavior, and to check the modules are working well together. An year is too much time to only do a PV, what happens if your HPLC apparted from normal behavior and you didn`t know when it was? Well, this is one way to do it, as part of a cGMP program. We also check the flow and DAD every 6 months.
But you gave me some ideas like the tubing instead of using columns, that was a good one.

[unmgvar]

your auditor asked the completely wrong question.
shows he does not know much i am sorry to say.
why should the instrument be verified every week? and why not every day? from knowledge we know after each work something can go wrong in an HPLC
the question is what and at what frequency and how does it influence your work
for example
for assay, UOC or dissolution tests by HPLC you can use the lamp until it goes and there is no effect of the noise in the application. noise becomes an issue for some impurities applications. now why would the lamp and column behave correctly from week to week and be faulty only on a monday?
what do you do if the noise test fails for you? do you redo all the tests for the last week?
do you do a carry over test as well for example. again for impurities it is important, but for assay you can use a broader ranger, but it is still important to look at that for every work and not only once a week or every year.
or maybe you are smart to include for each application the system suitability characteristics that will allow you to check that the HPLC and column are working within the necessary ranges in order to perform the current application. and only if the SST cannot be meet to see if the instrument is faulty.
USP and EP require that you check RSD of area for every work, not only every week.
but flow accuracy, gradient accuracy temp of the oven, do not need
remember that Performance verification is meant to show that your instruments are working within the guidelines of the required applications, and in HPLC each application is different and requires a different set of tests, so you cannot go generic. hence the system suitability test in the application is the PV that is done each and evry single time to show that the instrument is working within the necessary ranges of the applications
my advise would be to dump the weekly testing and have a smart PV/SST test within each application.
but keep the weekly maintenance of washing the instrument, it yes saves you money

[tom jupille]

Many years ago, a director at the USFDA said something to the effect that a validated method should have a system suitability test. If your system passes, it is suitable for the analysis, and what part of 'the system is suitable' is unclear?

[Consumer Products Guy]

Your plan sounds like way overkill to me. Our lab gets yearly preventive maintainence from a service company (seals, autosampler rotor, lamps replaced), and then a qualification. We run system suitability with each cGMP or GLP sequence; if SS doesn't pass, then there is NO cGMP or GLP data, no OOS investigation because there is no data.

Our test procedures are 98% in-house developed and validated, and our assays RARELY do not pass SS. Typically measured are retention time RSD, peak area RSD, tailing factor, and resolution.

[DR]

I agree w/ TJ & CPG. I don't care what they do w/ caffeine on Monday. I care what my LC does with system suitability standards and resolution solutions immediately before samples are run. Anything else on a weekly basis = wasted time.

[sharma]

I think you are overdoing the performance verification. It is really not required. US FDA even put more emphasis on daily system suitability done rather than half yearly calibration.

So if your system suitability is passing that is a performance verification of the system,,,,,,,,,,,,,

[qfiviera]

But not all monographies from USP or BP or EP has SS. Some doesn´t. And perhaps your HPLC SS complies before a sequence run, but it does not mean that all your system works OK. For example, may be the flow is in 0.9 instead 1.0, and your SS (depending on the SS) complies, but your pump is not OK. And an yearly PM does not guarantee that your HPLC will work well all year and what happens with all the data generated? When did it broke? Understand what I mean?

[rwjohnson]

Do you have control samples? You can run a control sample with an analysis of the product. This would save you some time since you will be setting up the instrument to run the analysis anyways.

Regarding use of the control chart, you don't want for it to just look pretty even thought that is what often happens. The control chart can tell you some important information about your system and when you need to start looking for issues. Ideally, you should only change the limits when you intentionally make a change.

Ron J

[unmgvar]

for all regulatory bodies there are:
SST of the method
SST of the general guide lines

so if not specified in the method for example the RSD for 5 injections should be less than 2%
tailing factor should fall between 0.8-1.5

the example of the pump is a nice one,
in many cases if your pump is 0.9 or 1.0 will not matter, as long as it is constant. since you are running the work against a standard. do not limit yourself to retention time for your chromatography, you have K' and Relative RT as well
it is by far more important for you to spot flow instability during the run, so RT shifts and RSD of area can spot that, and if you monitor pressure (a never requested SST by the way), then you will know easily if the flow as been moving from 0.9to 1.0
or injection volume, 10 or 11 or even 13 does not matter if it is the same all the time, more important is if it shifts from 9 to 11 in the work

this is why you need SST to spot shifting inaccuracies,
once a year look also that the instrument is precise on the hardware side, from history we see that it is enough.

FDA, EP, JP, BP do not want to know that the instrument is working well the beginning of the week. they want you to show it at the start of every work with the SST tests. they tell you that the instrument is working well enough to perform what you intend it to do.

again you auditor ask a wrong question

[LC_labrat]

I am surprised some of the USP/BP/EP mongraphs do not have System Suitability requirements. As I read a couple of USP monongraphs they were fairly simple, tailing, RSD, maybe resolution if there is a potential peak interference. Now these requirements do not include everything that could go wrong in an HPLC, just the critical parameters for each method.

I am assuming you use a standard to quantitate your results? I would write an in-house SOP stating some generic suitabilty requirements. As this would be a better example of verifying the instrument is working properly at the time of analysis. For example sample retention time is +/- 5% of the standard, %RSD is not more than 2.0%, tailing not more than 2.0, resolution not less than 1.5. Unless otherwise stated in the monograph.

Not all auditors understand chemistry.

[unmgvar]

qfiviera re-reading i saw that we did not comment on one very important point
you must run standards during the work as well.
this way you can spot trends of bad behavior. in some cases we run close to as many standards as we run samples.
monitoring the behavior of those standards during the work is part of the SST you can set for yourself
for example, USP says 5 injections-rsd not more then 2.0%, and you need for the entire work with 20 standards, NMT 2.5% or 3.0% for the area, for the RT, check tailing again in the last standard maybe, resolution,
then you at least can more easily spot where things seem to have started to go wrong.
we always when in doubt question the results from the last good standard.
everything before it is good and we see what goes with all the rest and we start to investigate the rest.