How much validation is required?

[Consumer Products Guy]

Opinions requested: How much validation is required?

I've been using a validated isocratic HPLC-UV method for over 20 years (on Type A RP-18, that's what was available then) with an OTC pharmaceutical product . It may turn out that we may need to assay for a new material as well, which also would use HPLC-UV. Because the polarities of the two analytes are so different, we would need to use a gradient (with the same column, same mobile phases, same flow rate), just starting with lower percentage of organic, to assay for both components in a single run. I'm thinking that this is not a major change, sample preparation would be the same, the only change would be to use a gradient instead of isocratic. The new component would not be an active, just for internal QC use. I'm hoping all I'd need to do would be to assay a few samples both ways, document equivalent results, and write a letter to the file, and term this tweaking. Any thoughts what the FDA might think? Surely they understand that much about chromatography. Don't say to ask my QA department because they feel that if one sneezes one time, that one must sneeze every time or that's a change.

[Alex Buske]

I would regard a change from an isocratic to a gradient method as a major change.

Alex

[Hollow]

also think its a major change.

At least I would check for:
- specifity (peak purity)
- robustness (change in starting conditions and/or gradient slope, maybe small change in the flowrate too)
- verification of equivalency (different conc levels, method precision at 100% level with both methods, mabe done by two persons each)

With the rest (sample prep, linearity) I would reffer to the old method, as there is no change

[Mattias]

If you could keep the original isocratic method and then just add a gradient step in the end - you should be able to get away with a small validation concerning the peaks in the gradient.

Otherwise I agree with the other replies, you will be doing a major change to the separation.

[Dan]

I also agree that it would be a major change and a new validation is needed. (No, I am not a QA person, I am an analyst.)

The change to a gradient may also mean a change in the values for the system suitability parameters and that alone would be justification for a new validation.

I would also add that this may be a good time to update the method to a newer column technology. I would say that 20 years is a long enough time that a method remediation would be a good thing to consider. Just realize that updating the method can improve the analysis (time and efficiency, etc.) but may have implications for the product history (specifications, purity profile, etc.) and QA and RA may not like you to mess with the product history.

Regards,
Dan

[Consumer Products Guy]

I would also add that this may be a good time to update the method to a newer column technology.

Actually, I had already discussed that same topic with a co-worker this morning: that if re-validation would be required, it would make sense to update to a modern Type B column. Thanks, running some scouting assays on modern column as I type.