re-validation

[hyenake]

Thanks in advance.

Is it necessary to re-validate a HPLC method? I will tell what happened.

Our hplc instrument wit an UV detector is broken (pomp) , and will not be repaired anymore.

I take initiative to take the samples to another HLPC, with an DAD detector so that the death-line is still ok.

I take the same method parameters from the broken instrument and put them in the other HLPC. Use the same solvents, isocratic flow, same wavelenghts, the only thing that is different is it's another instrument. The cleanup stays the same.

Today they told me that I had to do a kind of validation that the two instrument give the same results, but thats impossible. Now a new validation would be in order.

Is this necessary?

[dblux_]

Yes - IMHO.

[Perreman]

If your new instrument set up is not equivalent to the prior one, then I would do a re-validation.
You should look upon this as a kind of method transfer, therefore you will need to perform some kind of testing to evaluate if the new instrumentation is suitable for your analysis.

[tom jupille]

I'll give a slightly different response: If the method sspecifies a *specific* instrument configuration (e.g. it specifies a variable wavelength detector), then a different instrument configuration requires revalidation. However, if the method is not that specific, my opinion would be that you should demonstrate "equivalent" results, which is less than a full validation. Obviously you must meet system suitability, and I would also verify linearity, recovery, and LOQ/LOD (if those are important).

That said, remember that *my* opinion does not matter; it's the opinion of an auditor/reviewer that is important!

[Consumer Products Guy]

I had 4 decades experience with OTC products and cGMP.

I'm to agree with Tom Jupille's remarks above, no need to do a re-validation. I'd just document everything in lab notebooks or similar, and explain why stuff was done.

Of course, my pointy-haired boss would go opposite to that, as he always chose opposite to my opinions.

cGMP, USP, ICH - none specify a brand or type of instrumentation. Everything is up for interpretation by an FDA inspector, whether he/she got lucky, had a bad day, etc.

For example: cGMP guidelines say an instrument must be requalified if moved, but what constitutes a move? Is 6 inches down the bench, to an opposite lab bench, etc., a move???

Does 280nm using a UV detector mean 280nm on a DAD detector is not equivalent? Is 0.5ml/min from a quaternary pump different than 0.5ml/min from a binary pump? What if an isocratic method used pre-mixed solvent - is that different than having the pump mix them? What if the mixing valve on a quat pump was set to 100% as opposed to bypassing the mixing valve? It's pretty much impossible to specify EVERYTHING, one could never get any work done.

USP <621> even details what changes can be made in USP procedures without re-validation, and FDA ORA 4.5 even broadens that to in-house validated methods. I read those documents from the printed word, but pointy-haired boss liked to be devil's advocate 100% of the time (that baseline isn't EXACTLY flat, that calibration curve looks a little quadratic, etc.).

Enough said.

[DR]

Have a look at the original validation protocol and report. With any luck, some of it was done on a DAD system (development and/or intermediate precision). If so, you're done. If not, you need to fins another system similar to the broken one and run a sample set on both systems to demonstrate suitability and (maybe) the other stuff Tom mentioned. I would also look into validation practices to see that you're never again "stuck" using a particular setup.

[dblux_]

IMO it is important whether it is pharma lab or ISO 17025 accredited lab.
There was a change of the type of detector from UV to DAD.
They certainly may have different sensitivities and different spectral properties.
These may have impact on LOD, LOQ, selectivity, accuracy, linearity and so on.
These validation parameters have to be determined for new detector.
Can we call it partial revalidation ?
Worth to mention is assessment of uncertainty of the method (in 17025 environment) which has to be recalculated for new detector.

[HPLC chemist]

I agree with Tom. I would do a partial revalidation, particularly LOD and LOQ since they will be different than the original HPLC.