System Suitability Invalidation

[gyntar]

Hi,

I don't understand why HPLC system suitability (in our lab) is invalidated in 2 cases:

1. the flow of the mobile phase is stopped through the column by mistake for few seconds, however pumps, detector and other critical part of HPLC system are still working.

2. the composition of mobile phase pumping through the column is different from those in the compendial method (for exaple: H2O and ACN applied as a wash in the middle of the sequence to wash out buffer and prevent precipitation of salts, so that more samples can be run).

I couldnt find anywhere the reasonable answer or guidelines for it.
Did anyone come accross similiar issue?

Piotr

[AdrianF]

Could you elaborate so we can understand the problem more clearly.

[wanda50]

In the US, it used to be acceptable to put a wash after the stated compendial gradient has completed. Of course, you have to change the requilibration time to accomodate the wash out. I haven't had the question come up recently, so I cannot stated for certain that my statement is still true. Secondly, I have not had the quesiton come up for an EP method.

Was the system suitability replicate injections? If so, there needs to be a minimum of 5 contiguous assay standards or 6 related substance standards. If the system stopped in the middle of the replicates, it must be started over.

if the system suitability was one injection for RT, RRT, tailing, resolution, etc., of course it needs to be repeated. the clock ticking on the data collection will give false conclusions for RT, etc.

Try to think like a reviewer - what is this company going to try to put over on me?

So, always make sure your analyses are the best they can be so there is no room for doubt. Do it right until it hurts.

[Rande]

To give a quick answer, because your SOP says so!

System Suitability criteria are created to show that you have control of your analytical method.

Try to imagine explaining to a reviewer that:

" …I know that the mobile phase stopped flowing, but believe me – that did not affect my results! "

If you have good scientific reason to justify a deviation, document everything and write a deviation.

Remember, most FDA 483’s are for failure to follow your own procedures.

[gyntar]

Could you elaborate so we can understand the problem more clearly.

The SOP we have doesn't mention about invalidation system suitability in case of stopping flow or using wash between injections with the composition that is different from mobile phase, so the explanation that system is not suitable for use after stopping the flow sounds more like superstition.

The main reason I wanted to stop the flow is to get rid of phosphoric buffer cristallisation that clogs up the pipes by using wet prime option on Waters HPLC but it means to stop the flow through the column.
Thanks for all replies!

Piotr

[Mark Tracy]

If you are having a problem with buffer precipitation, your method is already faulty. Investigate permissible method changes that might reduce the problem of buffer precipitation. One thing that sometimes causes precipitation is a step change from high % organic to high % buffer at the end of the run; use a short gradient instead of a step change.

[Bruce Hamilton]

The issue of buffer precipitation should have been sorted at the time fo the method development.

If you are following a compendial method( EP, USP, BP etc ) then the problem is likely to be either environmental ( mobile phase bottle temperature is varying ), quality of reagents - especially the salts used to make up the buffer ( use HPLC grades that state they produce clear solutions in gradients ), or mobile phase preparation ( is it filtered after allowing to stand for an hour or so after preparation ). If the method is compendial, it's very likely the method worksin sequences without washes.

If the method is in-house, I would address the issue by fixing the problem ( eg potassium salt buffers are often more soluble than sodium, changing the gradient, etc. ). There is also no reason why a sequence can't incorporate multiple rinsing flush methods, provided that you have included a confirmation of instrument performance ( system suitability ) throughout the run after each flushing, or duringa full method validation.

You can't just change a sequence when you feel like it, you must put the washes and system suitability tests into the sequence before starting it - hence it's usually better to fix the method.

Please keep having fun,

Bruce Hamilton

[tom jupille]

the explanation that system is not suitable for use after stopping the flow sounds more like superstition.

That's not the explanation. The explanation is that they are not sure the system is suitable for use after stopping the flow or washing out the crud.

I don't think it's superstition at all (paranoia, yes, but not superstition). Look at it this way:

• Never walking under a ladder is superstition.
  
  Not walking under a ladder when you can see a paint can perched on top is common sense.
  
  Not walking under a ladder when you can't see what's perched on top is merely paranoia.

[LC_labrat]

Stopping the flow of mobile phase is a "change to your system". Therefore you need to re-establish that the system is still suitable. You need to prove to your audience (reviewer/client/FDA) that it was working.

Adding a wash to a validated method is a "change to the method" and brings in to question whether or not this change effects the validation of it. Was the change within the acceptable limits for the method? Was the re-equilibration long enough? Will it produce the same results? Most likely yes, but you have to prove it.

[gyntar]

I will try to do some small changes that the EP method allows. I didn't know that the buffer preparation may prevent salt precipitation.
Thank you for all suggestions and advices.

Piotr

[Rande]

Just to add my 2 cents to this discussion again:

Stopping and restarting the flow of mobile phase should not be a system suitability criteria (if your SOP specifies it - follow your SOP until you can revise it). System suit should be based on measurable values with defined criteria (such as the CV of replicate injections, as well as some expected values for standards).

There is a good chance that stopping the flow will cause a strange baseline about one column volumne after restarting. If you add a buffer blank injection - after a pause, then there is no problem. If this "strange baseline" comes in the middle of a sample peak - that could be a problem.

For our HPLC assay, one of the "Assay system suit" tests is based on product reference standards that are run before and after the samples. Based on USP guidelines, we use the average of 6: 3 before and 3 after. If these fail - the entire assay is not valid and you do not even look at the results for any of the samples.

There are also sample "acceptance criteria" for the replicate injections of a sample. If these fail - the results for that sample are not valid, but other samples in the same assay run, might have reportable results.

As my boss always reminds me, keep the system suit acceptance criteria fairly tight. If anything is a little funny with the assay, you want to be able reject the assay as a system suit failure (SSF), and be able to re-test the samples. If your acceptance criteria is always easy to pass, even if things are a little off - you run the risk of an Out Of Specification (OOS) result for your samples. It is much better to repeat a few assays for a SSF - than to require an OOS investigation!