Method validation for injection volume and concentration

[sisay]

Hi all

It is my first time to be here, hope you can provide any information or reference for my question, thanks!

Recently i faced a problem about method validation for injection volume and sample concentration.

For example, i have validated a HPLC test method. The injection volume is 20 uL and the sample concentration is 1mg/mL.

If I change the injection volume to 10 uL and smaple concentration to 2mg/mL, what validation elemnets (like linearity, accuracy, precision...etc.) should be evaluated again and why? Notice that the amount of sample is the same.

Thanks!

[Consumer Products Guy]

I'm not a validation, but have lots of cGMP experience. FDA ORA and USP 621 allow for smaller injection volumes without need for revalidation, but don't address whether concentration can be increased accordingly.

To be safe, I'd re-do linearity, and maybe assay a certain number of samples both ways (from same vials) and see if results are statistically indistinguishable (whatever that really means). Of course you'd be detailing system suitability under each set of conditions, and documenting everything.

Maybe someone more expert can comment.

Scientifically, your idea is sound, injecting same number of molecules onto the column.

[danko]

Hi Sisay,

The answer to your question could become more obvious if you tel us (and actually tel yourself) what made you alter the injection volume.
It could be to inject less liquid and thus improve the peak shape. It could be to use less solvent under sample solution preparation etc. Then it'll be much easier to identify what's actally relevant to investigate and validate.

Best Regards

[sisay]

Hi Sisay,

The answer to your question could become more obvious if you tel us (and actually tel yourself) what made you alter the injection volume.
It could be to inject less liquid and thus improve the peak shape. It could be to use less solvent under sample solution preparation etc. Then it'll be much easier to identify what's actally relevant to investigate and validate.

Best Regards

Thanks for all the response,

To danko, the reason to change both of the parameters is that we found the system suitability is more stable (less RSD) than previous parameters did. We thought the amount of analyte injected is the same, therefore, maybe not all validation elements will be evaluated again. hmm.., to save time and make the method still reliable.

[danko]

Something’s not as the theory predicts.
If you inject the half of the volume as you used to i.e. 10 instead of 20 µL, the RSD should increase instead of decrease – according to theory. Unless something mechanical (a malfunctioning injector) causes the phenomenon you experience.
It might be a good idea to investigate the reason for that before going further.
Otherwise – with regard to the validation needs – you’ll have to think sample preparation thoughts:
Solubility (you’re doubling the concentration)
Sample solution stability (derived from the above with precipitation in mind)
Accuracy
Precision
And
Recovery

Best Regards

[Peter Apps]

Something’s not as the theory predicts.
If you inject the half of the volume as you used to i.e. 10 instead of 20 µL, the RSD should increase instead of decrease – according to theory.

Strictly this is true only if the repeatability of the injection volume is the main contributor to variation in peak area.

If, for example, the sample solvent was too strong for the mobile phase a smaller injection might give sharper peaks that are easier to integrate, and possibly better resolution which makes peak onset and end easier to detect - either of which could reduce rsd of peak area even though the rsd of the injection volume is increased.

Peter

[danko]

He, he………………… Peter – it was far fetched
But OK let’s take it in consideration.
Best Regards

[Peter Apps]

Hi Danko

Not so far fetched - we get plenty of posts complaining about distorted peaks when it turns out that the sample is dissolved in pure methanol. or has an extreme pH.

Peter

[danko]

I know that Peter - believe me.

But distorted peaks are reproducible too. Anyway, when we’re talking main peak.
Besides, there was nothing mentioned about peak distortion. Just RSD and that is weird.
I would qualify/validate the injection system before taking any steps towards altering method parameters if I were to solve the present issue.

Best Regards

[bunnahabhain]

OT: danko, the link in your signature does not work.

==> http://www.stella-chroma.com/

[danko]

@ bunnahabhain

Oh…. I’m sorry, but I had forgotten to remove it.
Thanks for drawing my attention to it.
I closed that business for some time ago.
Now working for another consulting firm.
If you need any contact, you can potentially get my e-mail address.

Greetings