Method validation - a silly question!

[sksrinivas]

Hello All:

Been a long time since I last posted on the forum.

I have an apparently silly question, that I haven't been able to answer myself or get a suitable answer for:

Pharmacopoeial methods permit small changes or adjustments. Typical permissible limits include +/- 10% change in pH, 50% change in flow rate, 100% change in injection volume, and so on.

What baffles me is the permissible changes in mobile phase composition i.e. "+/- 30% relative or +/- 2% absolute, whichever is less."

What exactly does this statement mean? I get a different answer each time I ask this question.

Could anyone show me the light?

Warm regards to all,

S.K. Srinivas

[Alex Buske]

hi,

Thats rather easy:
imagine a method using 40% MeOH in Water +-30%relative would be: 28-52%((0.7*40) - (1.3*40)), 2% absolute would be 38%-42% ((40-2) - (40+2)) so in this case the 2% absolute applies.
Imagine on the other hand a method with 1.5% IPA in hexane: -+ 30%relative would be 1.05-1.95% IPA, 2% absolute would allow for -0.5% -3.5% IPA. In that case the 30% relative apply.

Alex

[vicentexxx]

hi alex,

i agree with your tip but can you tell me what is the limit to consider 2% or 30% (3% 5% 10%)? and then i can't grasp the meaning of relative and absolute...
can you or someone light me?

thank you
Vincenzo

[sksrinivas]

HI Alex:

Thanks for your quick reply.

I'm still a bit confused though. I'll try and work it out, using your reply as a starting point.

Thanks!

[PJ8]

It is a good idea to run robustness studies on finalised HPLC methods. This provides you information to work on and evidence your adjustments wont affect the method adversely rather than relying on benchmark figures that may or may not be suitable for your methods and re-validation!

[tayzyboy]

It seems you are conused by the terms "absolute" and "relative" ? . (sorry if you're ok with this)

2% absolute means that you alter the mobile phase composition by 2% . Eg. changing from a) 5% to 7% or b) 50% to 52%.

30% relative means that you alter the mobile phase by 30% of the original value.
Eg. changing from a) 5% to %6.5 or b) 50% to 65%

<2% absolute or <30% relative whichever is less.... in example a) you would use "30% relative" and in example b) you would use "2% absolute"

hope thats of some help.

James

[sksrinivas]

Thanks James.

Let me see if I've got this right:

Assuming my mobile phase is ACN/water, 70/30;

- +/- 30% relative with respect to ACN means I can change ACN composition from 49% (0.7*70) to 91 (1.3*70).

- +/- 2% absolute with respect to ACN means I can change ACN composition from 68% to 72%.

Right?

If so, then I have finally seen the light!

Warm regards to all,



Srinivas

[Peter Apps]

Cat among the pigeons

starting with ACN/Water 70/30

Change the water by 30% relative and you get ACN/water 89/21 or 61/39, which is a much tighter range than allowed by changing the ACN by 30% relative.

Starting with ACN/Water 30/70

Change the water by 30 % relative and the ACN/water can change to anything from 9/81 to 51/49 !

Peter

[tayzyboy]

Intersting take on the problem Peter.

This would be further complicated if you had 3 components to the mobile phase. Its far too early in the morning to think about the calculations.

Its just a case of making sure you alter as many different components as possible to check them all out.

Good luck!

[vicentexxx]

as i understand,

you have to apply the less variation possible so you can apply the rule as you want (water or CH3CN) becouse at the end your choose is that which changes less the composition.

[Alex Buske]

Peter,

In your models the 2% absolute would apply.
However, thinking about the topic I would say you have to consider the minor component. Only this makes sense (as much sense general rules can make).
+-30% of a 90%mixture component wouldn't make any sense.
PJ8,
The whole thing is a rule how far can you go without new validation. Hopefully there is a system suitability check.
On changing the parameters the _responsible_ analyst would assure that everthing is working as intended.

Alex

[PJ8]

Thanks Alex, that's what I was trying to say. The percentage figures are there as guidelines. It's possible that your method may be more (or less) sensitive than the changes permitted by the figures so it is impotant to test the method once changes have been made.

If you perform robustness studies on your LC methods once they development is complete, if the need for adjustment later arises you have pre-tested what changes in parameters your method can tolerate.

[tom jupille]

pj8, you make a good point that needs amplification. Those guidelines are and apply in the absence of specific guidelines in the method. That means that it is important to specify in your method how much adjustment is allowable (obviously, supported by validation data). That may be either tighter or looser than the default guidelines. If you do not specify the allowed adjustment, then the default guidelines apply.

As a further note, as far as I know, the USP has not officially adopted the 2004 proposed adjustment guidelines. The US FDA has a set of guidelines for their internal use (http://www.fda.gov/ora/science_ref/lm/v ... _04_05.pdf), but you should probably have an SOP in place adopting those guidelines for your own use to be in the clear if questioned by an overzealous auditor.