HPLC method validation

[ananda]

Recently someone asked me what is the most important HPLC validation parameter. I think even ICH or CBER has not pick one parameter over the others to say that is the most important. To me, all the parameters are important since any of the paramaters fail, the method fails. However, can it be the Linearity? Specificity?Precision?Accuracy? Any comment would be useful.
Thanks
Ananda

[Dan]

Robustness

Methods for use in the pharmaceutical industry must be used by many different analysts in a number of different labs. You need a robust method for it to work well in all those situations. The less you see an OOS, the better.

Regards,
Dan

P.S. All validation parameters are important, but If I had to choose just one; I'ld go with the robustness.

[ananda]

Thanks Dan. Robustness makes sense but how about specificity or linearity?

Ananda

[bartjoosen]

specificity could be a part of the robustness....

Linearity is important, LOD & LOQ is important, accuracy is important, ...

As we have only one lab, I would go for accuracy.


Bart

[Tim_van_Dyke]

Precision is the most important parameter.

[ym3142]

of course, specificity.
there are not many reasons making a well separated peak not accurate, precise or robust or linear or any others unless there are sample prep issues

[Alex Buske]

The method wil not work when specificity is not given. Results will suffer when linearity and accuracy don't work out. As a lab manger you really woldn't like a method that isn't robust. And your customers would hat scattering results due to poor precision.
There are for special cases some exceptions.

Alex

[ym3142]

My question:
if I have a method to give a peak at 5min , which is not interferred by anything else at all, could you please tell me the reasons which cause a bad accuracy, precision, robustness, linearity or anything else, if the LC is still meeting its qualification and the sample preparation does not cause issues?

[tom jupille]

could you please tell me the reasons which cause a bad accuracy, precision, robustness, linearity or anything else, if the LC is still meeting its qualification and the sample preparation does not cause issues?

Here are a dozen for starters.
- your detection wavelength is on a slope of the absorbance spectrum
- you are operating near the pKa of your analyte, and the chromophore is coupled to an ionizable group (UV spectrum shifts with pH)
- your analyte is only marginally stable in solution
- your analyte can equilibrate between multiple forms in solution
- your injection volume is small relative to the volume of the autosampler syringe
- your injection volume is large relative to the volume of your injection loop
- your analyte sticks to tubing (stainless steel or PEEK, as the case may be)
- your retention depends strongly on secondary interactions
- you are at or near LOQ
- you are at or above the upper end of the linear range
- your sample diluent is stronger than your initial mobile phase

I'm sure people can come up with lots more.

[ym3142]

Tom, Thank you for taking time to get this long list. I really hope other people will extend this list. I am very interested in this topic, which should help us to prevent problems or do troubleshooting.

I have some comments about these senerio:
1-your detection wavelength is on a slope of the absorbance spectrum
Re, when I decide a wavelength I look at the UV spectrum. I will avoid a steep slope. A general slope is OK I believe.
2- you are operating near the pKa of your analyte, and the chromophore is coupled to an ionizable group (UV spectrum shifts with pH)
Re, preparation issue. so avoiding pka value pH is not only for MP but also for prep.
3- your analyte is only marginally stable in solution
Re, solution stability study takes care of this;
4- your analyte can equilibrate between multiple forms in solution;
RE, Prep or stability study
5- your injection volume is small relative to the volume of the autosampler syringe
6- your injection volume is large relative to the volume of your injection loop ;
Re hope a common sense will take care of this 5 and 6;
7- your analyte sticks to tubing (stainless steel or PEEK, as the case may be)
Re, this can be real trouble
8- your retention depends strongly on secondary interactions ;
RE this will be robustness issue;
9- you are at or near LOQ ;
RE, or even below LOD,
10- you are at or above the upper end of the linear range
RE, So UV is not linear if >2AU;
11- your sample diluent is stronger than your initial mobile phase;
RE, this is OK unless peak splits.(right?)

Tom, I value all of the 11 possibility. Thinking chemists are fighting for resolution/separation/specifity/selectivity every day and facing these above issues only once in a year at most(or even none in ten years) I'd like insist specificity is the most important in general for all method in this world though specific method usually has its own specific shortcomings.

[tom jupille]

I'd like insist specificity is the most important in general for all method in this world though specific method usually has its own specific shortcomings.

While I understand your point, I disagree with it.

Asking which of the parameters is most important is like asking a fire fighter "which is more important: thermal protection or an air supply?". The fact is that lack of either one can kill you, and both need to be assured for survival.

Likewise, validating a pharmaceutical regulatory method (e.g., one conforming to ICH guidelines) requires that certain parameters be characterized. As far as I know, there are no "degrees" of validity. From a regulatory perspective results either are or are not valid. Failure on any point renders the results not valid.