Method changes

[AdrianF]

I work in the pharmaceutical industry - many of the methods are old and could be greatly improved.

However I am told that even a minor change will require a licence variation, therefore it is generally not worth bothering - what are other peoples experience?

Will some products still be analysed with 30cm uBondapak in 20 yrs time!!

[Bruce Hamilton]

If you want to use an alternative method, you can. You just have to demonstrate that the method is equivalent to the approved method. In other words, pretty much a full method development and validation and supporting data pack that will be submitted for approval, if requested.

That's not a cheap exercise for any employer, but if they will make significant savings overall, the cost would be justified. Casual method users are less likely to justify the investment.

The various pharmacopoeia/drug regulation authorities have programmes for updating methods, but the programmes are prioritised by the industry, and also driven by the global harmonisation process.
They have the advantage that proposed changes can be evaluated by large numbers of producers concurrently, ensuring the new method is appropriate for the whole industry.

If the most cost effective solution for your employer is to use an existing method that matches all regulatory requirements, why invest significant time and money in changing the method just because it's been around a while?.

Bruce Hamilton

[AdrianF]

I realise that a completely different method will require revalidation and a licence variation.

The question I am asking is how much change can be made without this laborious procedure?

[Jumpshooter]

"Validate Reference Methods" are viewed by both Compliance Dept' and Vice Presidents of R & D as pretty much "carved-in-stone" documents. It is laborious, litigenous, and enervating to attempt to change these SOPs--even IF your suggested revision ("Candidate Method") has been shown to provide equivalent or better results and saves time & money. If you find yourself burdened by this "carved-in-the-stone" perspective, then you had better find a job in academia instead of government or private industry. I realize that this may seem like an anti-intellecutal view b/c in grad school we were trained to devise, improve, optimize, and critique current methodology--but this is the "real world" (i.e., the "looking glass world" to quote Alice in Wonderland--heheh).

[Bruce Hamilton]

The question I am asking is how much change can be made without this laborious procedure?

Any parameter can be adjusted to achieve the desired separation whilst still within the original validation range of that parameter.

Each regulator/control authority has their own guidelines on how much change is acceptable, and they are usually defined in the general information, or occasionally, in the specific monograph.

I think the EP ( which I don't have in front of me ), permits something like:-
Permitted Column Length Variation: -50% to +100%
Column Particle Size: 50% of nominal
Column ID: ±25%
Mobile Phase Flow Rate: ±50%

It would pay to check the particular regulator for the actuial values.

Bruce Hamilton

[JA]

I think the EP ( which I don't have in front of me ), permits something like:-
Permitted Column Length Variation: -50% to +100%
Column Particle Size: 50% of nominal
Column ID: ±25%
Mobile Phase Flow Rate: ±50%

With the appreciation that the pharmacopoeia may state different numbers than actually quoted, I would like to ask for clarification if the regulatory bodies would consider it necessary to have already investigated these variations during a validation?

To ask more specifically: if a method has been satisfactorily validated using 150 mm columns only and at a later date a system suitability resolution criterion is not quite met in a batch analysis, can I simply swap in a 250 mm column of identical stationary phase?

[shaun78]

To ask more specifically: if a method has been satisfactorily validated using 150 mm columns only and at a later date a system suitability resolution criterion is not quite met in a batch analysis, can I simply swap in a 250 mm column of identical stationary phase?

According to the USP (which Bruce's post is definately correct for): yes. What was left out is that the column length et al is only allowed to be modified within the allowable parameters if a clear improvment in chromatograpy/system suitability is made.

In your case you went from not meeting resolution requirments to meeting resolution requirements. That clearly improves chromatography and you went from failing system suit to passing system suit...

[JJG]

To ask more specifically: if a method has been satisfactorily validated using 150 mm columns only and at a later date a system suitability resolution criterion is not quite met in a batch analysis, can I simply swap in a 250 mm column of identical stationary phase?

In an FDA regulated environment, the answer is no. If the method was satisfactorily validated using a 150mm column, and system suitability is suddenly not met, then you must investigate why system suitability was not met. A validated method means the method worked with a 150mm column, so the column length should not be why system suitability is not met. Now, it could be the column went bad, so you could replace it with another 150mm column, but the FDA will not like it if your SOP says 150mm and you're using 250mm.

Unfortunately, this is why some methods go relatively unchanged for many many years in a QC department, because it is easier and you don't have to justify keeping the same 20 year old method, whereas if you do change the method, you do have to justify it. From my experience, old methods don't change unless issues start arising. Then you have the opportunity to address the ancient method.

[tom jupille]

In an FDA regulated environment, the answer is no.

Unless you work for the FDA, in which case the answer is "yes". Read page 10 of this document:
http://www.fda.gov/ora/science_ref/lm/v ... _04_05.pdf

It allows an FDA analyst to adjust column length by as much as +/- 70% so long as they meet system suitability.

[JA]

So for clarity, it is correct that:

• Method changes, such as those suggested in Bruce's second reply, are applicable to compendial methods straight out of the Pharmacopoeia (which do not require validation at each site before use).
• The ORA are permitted to make other changes, per the document referenced by Tom, when they perform some sort of verification of a method submitted by a pharmaceutical company in an NDA?
• A lab performing analyses can only make tweaks to their method within the scope of the robustness parameters of the validation, of which column length, diameter and particle size would not be included?

Sorry for dragging this on

[Bruce Hamilton]

So for clarity, it is correct that:

• Method changes, such as those suggested in Bruce's second reply, are applicable to compendial methods straight out of the Pharmacopoeia (which do not require validation at each site before use).

Yes, noting each regulator/authority will detail the pemitted variations along with qualifiers- such as for column length mentioned by Shaun78.

[*]The ORA are permitted to make other changes, per the document referenced by Tom, when they perform some sort of verification of a method submitted by a pharmaceutical company in an NDA?.

Yes. There can be organisation-specific limits, and they can even exceed those limits provided they demonstrate that precission and accuracy of the method are not affected, but have to experimentally confirm that claim.

[*]A lab performing analyses can only make tweaks to their method within the scope of the robustness parameters of the validation, of which column length, diameter and particle size would not be included?[/list]

Depends on the organisation, but why would any company routinely waste resources validatiing their method for multiple sizes of columns?.
If the method is fit-for-purpose, stop spending, and give staff salary increases .

The usual solution is to choose columns that are likely to be readily available for method transfer for the projected life of the product - say 10-15 years before become generic.

Pharmaceutical compendial methods have existed for over 100 years because they provide a cost-effective solution to the need for consistent quality criteria for acceptance or rejection of batches. They usually travel through a more comprehensive evaluation and continual improvement programme.

Bruce Hamilton

[Mary Carson]

The ORA are permitted to make other changes, per the document referenced by Tom, when they perform some sort of verification of a method submitted by a pharmaceutical company in an NDA?

The FDA does do analyses other than those submitted as part of an NDA. I would not assume the ORA laboratory manual is referring to solely dosage form drug analyses in this section.

[tom jupille]

I think that the intent of the FDA ORA Lab 5.4.5 de facto limits as well as the proposed (but, as far as I know, not officially implemented) USP de facto limits, was to allow "adjustment" of separation conditions in order to meet system suitability requirements. This is not the same thing as speeding up a method or moving to newer technology. In fact, if system suitability specifies a retention time window, changing to a significantly shorter column might well make the method fail on that basis alone.

Making a large change in the method in order to improve resolution or speed up the method beyond the system suitability specification for that method would, in my mind, clearly be a modification to the method and thus require revalidation.

[AdrianF]

So if you can get from A to B using a cart track you are best sticking too it rather than use the Motorway!

Tom when you say you are required to revalidate - how much would you have to do? ie recovery linearity reproducibility etc.

eg say you have a method on Microbondapak and you find a 5u column with the same separation characteristics using a 10cm column

[shaun78]

eg say you have a method on Microbondapak and you find a 5u column with the same separation characteristics using a 10cm column

Most of the time I try to demonstrate equivalency between the two columns and methods. Depending upon where I have worked, that has ranged from anything from:

1. Demonstrating method specificity. Six sample preps tested on the same day using the two different methods. Agreement between results had better match/conform to what was stated for method ruggedness section of protocol. Perform stress degradation using condition that degraded the product the most.

2. Re-execute the entire validation protocol again for the new method.

3. Re-execute entire protocol, but only perform stress degradation under the condition that degraded the product the most. Also, as long as the matrix was not changed, no need to re-evaluate sample/standard stability.

I hope this helps!