Failed System Suitability Test

[assyya]

Dear Forum,

We are trying to define in an internal procedure what to do when an SST fails. We usually repeat the SST until we get a pass & keep the data from all failed SSTs. Do you guys raise a deviation (& root cause investigation) each time an SST fails?

Many Thx for your help

[Rndirk]

This is a very general question. In what type of environment do you work (pharmaceutical, food,...?).

It depends on the instrument and on the analysis. Just repeating the SST until it passes is, in my opinion, not a good practice. You have to define the criteria of the SST so that, when it doesn't pass, something is wrong and needs to be fixed/ajusted.

If just re-injecting makes it pass, then your criteria are too narrow or the instrument is too unstable. I would advice to not take your criteria too narrow, but of course stay within the limits of legislation (if any). Take a good look at your validation data to define these. Which analyte, for instance, is most sensitive to the state of the instrument?

Most of the time people here recognize the issue or contact someone who is more experienced with the instrument and/or the specific application. Actions depend on the symptoms of course. An example in the case of an unstable HPLC pressure: changing pre-column, longer conditioning, rinsing column with high organic,....

[dap]

Hello.

We don't have a SOP for such cases. If resolution or plates or symmetry fails usually column guard is replaced by new one than column can be replaced or mobile phase flow can be reversed. If RSD fails it is usually an injector or detector (or method also) failure that has to be resolved somehow. If signal/noise ratio fails mobile phase is remade at first than see p1. And so on…

[HPLC chemist]

You should have an SOP on SST and chromatography that is compliant with USP section <621>. In short a failure of SST means that you have demonstrated that the instrument (HPLC, GC, CE...) is not capable of performing the required analysis and that ANY results are not valid and should not be reported. In addition, a CAPA should be issued and the method should be examined for validation per ICH Q2(r1).

[dap]

and the method should be examined for validation per ICH Q2(r1).

Of course if there is a SOP for columns testing and if you are 100 % sure the column is OK.

[tom jupille]

As already stated, that is a *very* general question that is not easily answerable in a post. What you do depends on why/how you failed SST. We will actually teach a one-day (actually, three 2.5-hour sessions) course to cover it in October: Making LC Methods Work