When I’m going to change a method (e.g. changing the length of column, change from C18 to C18 endcupped, adding to MP 1% THF etc.) have I to validate my method as a new method or it is possible to work without validation?
I’m looking for parameters those are major and it is necessary to validate the method, and those are minor.
May somebody help me?

In general, you are allowed to "adjust" a method in order to meet system suitability specs, but if you "modify" the method, it must be revalidated. The catch is that there is no clear definition of where adjustment ends and modification begins. To be safe, if you are developing a method, you should specify an allowable range of adjustment. If you are using methods developed by others, you should have SOP's in place that define what you are and are not allowed to adjust.

Absent all these things, changes in column length, particle size, and flowrate are often considered to be adjustments (remember that you must still meet system suitability) -- but an overzealous inspector may still call you on them!

Changes in column packing typically require re-validation (unless "or equivalent" was specified, in which case you can *probably* get away with demonstrating that you meet system suitability). Adding an organic modifier typically requires re-validation.

Hope this helps!
-- Tom Jupille / LC Resources

By the way, has anyone here had *direct* experience in this area with regulators that they could comment on?

Splutter!!! Where's my soapbox, let's see if I can get myself fired for this one....

Last year the FDA floated a draft Guidance for Industry document "Bioanalytical Methods Validation for Human Studies". Our regulatory group wrote a polite letter to the FDA asking them to reconsider the document based on a group of method developers, including me, who impolitely shredded it.

On page 2 of the draft, it said, "When changes are made to a previously validated method, the analyst should exercise judgment as to how much additional validation is needed." It should have stopped there, but it didn't. It goes on, "For minor modifications, such as a change in the ratio of solvents for elution, a change in buffer system, the number of extractions of the biological matrix, or a small change in the column temperature to obtain better separation, only limited validation may be recommended. For major modifications, such as change of an instrument, solvent system, detector, or temperature, full validation of the modified method should be performed."

Okay, quick show of hands, for all us HPLC geeks who've taken a class from LCRecources or the equivalent, how many of you would categorically consider a change of buffer to be a "minor" change? Hmmm, I see no hands. Good, nobody slept through class. And as to that list of "major" modifications, what do they mean? Change of "instrument" or "detector" means what? Let's say I've got two HP 1090s (I don't, but hypothetically speaking), and one breaks down. I move the method from one 1090 to the other. I need to do a full revalidation??? Oh, maybe that isn't what they mean by change of instrument. Maybe we could get Bill Clinton to define the words for us....

Oops, here comes our HR group.... gotta run.

End of rant.

better try to go "anonymous" on this one

That's the advantage I have working for a small (and tolerant) company -- I don't have to stay anonymous! Seriously, your points are very good.

A number of the issues involved (and some draft suggestions) were discussed in Pharmaceutical Technology (June '98, starting on page 58, authors were Fuhrman, Dorsey, and Snyder -- sorry I don't have the exact reference at hand). One aspect of this whole issue is the observation that it represents a major change in the way a "method" is defined, a change which puts *far* more weight on system suitability criteria.

As it stands now, a "method" is defined by the process used to carry it out, and the system suitability is, in effect, a reality check to make sure the results are as expected. Hence, any but the most trivial changes in the process represent a change in the method, which must be revalidated.

If a wider range of adjustments is permitted, then the method is, in effect, defined by the system suitability criteria. The "process" (column, exact mobile phase composition, etc.) becomes much more advisory in nature (obviously, within limits!). As I said above, this puts more weight on good system suitability criteria and on providing robustness data with the method as a guideline to the analyst concerning how much "adjustment" is permitted.

Having said all that, I would *not* want to be in industry in the forefront of the movement towards "system-suitability based" methods (what's the old joke about pioneers are often found face-down with arrows sticking out of their backs?).

-- Tom Jupille / LC Resources Inc/

As for the pharmaceutical industry, there is a guideline for method revalidation requirements viz., Drug Directorate Guideline - Acceptable Methods by Health Protection Branch, Canada, 1994, though it is not entirely satisfactory. With regard to the USP methods, I have come across several seminars and papers (I do not have them in hand) discussing the limits of changes allowed for USP methods. Theoretically it is not required to revalidate the USP method if it is not considered as changed. However in most cases, to meet the regulatory requirements, a partial or full method revalidation on the 'modified' methods should be performed. The extent of method revalidation should be scientifically justified and the Acceptable Methods is a good reference.

Hi, Wilhelm! Do you know if that Guideline is accepted down here south of the border? And does HPB have a web site from which it could be downloaded?

Thanks!

-- Tom Jupille / LC Resources

Hi, Tom. It's nice to hear from you. Health Canada has a web site http://www.hc-sc.gc.ca. I don't know the document is in there or not. Since FDA does not has a guideline, I think personally this is a good regulatory reference point in addition to scientific justifications, until FDA comes up with one.
Wilhelm

Many thanks!

-- Tom

As long you have all of your peaks clearly identified and quantified, and those results are the the same, there is no reason, why not to use an adjusted method, even without a complete new validation. It is much more important to learn, which changes are critical and how they can be corrected, than to stick to a nonrobust method, giving you headache every other day.
Never except a method from somebody, without detailed explanation for the reasons to set the working conditions as they are, f.e. the pH to 3.5. There should be an explanation.
DryLab can deliver such explanations in the form of the relative resolution maps (RRMs).

You should give yourself reasonably set windows, in which you can allow the change of a single parameter, if necessary.
You can't stop the production process, because you have the critical resolution below 1.5, if you can correct this by changing the temperature of let say 5°C.