We are currently working on a gradient method with Ammonium Acetate Buffer and Acetonitrile. Detection is by UV at 230 nm. As would be expected, there is a substantial drift of the baseline due to the gradient.

We are using Turbochrom software which has the ability to specify a "baseline" file which is subtracted from the "raw data" file to give a "modified data" file. Obviously, this straightens out the baseline and makes peak integration easier, not to mention nicer to look at.

The question is, does anyone have any experience with this type of data correction in a regulated (GMP) environment. How is this looked at? Is it more regulatory trouble than it is worth?

Thanks in advance.

Kevin Kolodsick
Analytical Method Development
Ferndale Laboratories

i do not have any experience with the GMP-acpects, but i have used the 3d-blank-substration with our PDA's. its nice but be carefull. especialy with gohtspeaks. the retentiontimes chenge a litlebit from injection to injection an this can give strange baselines. with UV-spectra it only works until a certain peakhight. under that it makes thing only worse.

Agreed. I have played around with the feature a little bit and have found exactly the same problem. If one is trying to clean up a baseline with some small ghost or impurity peaks, this will probably cause more problems than anything else, especially in the case of smaller peaks, and certainly if there is any retention time variations.

The situation we have now does not have any extraneous peaks, just a steep baseline rise that causes some integration problems and makes looking at low level impurities difficult.

Thanks for your input.

Kevin Kolodsick
Analytical Method Development
Ferndale Laboratories

I have been using blank substraction in a FDA regulated environment for many years. I must say that I never got very specific questions about this practice during audits, but I am confident it is acceptable provided all the requirements about traceability, control of time/nb of injections between blank and sample, and method validation including this technique are satisfied.

Is there a standard acceptable number of samples between blank injections, or do you just run a new blank with each set of recalibration standards?

Thank you for your advice.

Kevin Kolodsick
Analytical Method Development
Ferndale Laboratories

We use this type of methods only for area % impurity profiles, so we have no standards and blanks cannot be linked to them.
Our methods usually specify running 2 blanks (discarding the fist one) and n samples (+ a blank and n samples again if necessary).
n must be determined and demonstrated to be suitable for every method.

Beppe,

Thank you, this will give us something to work with. I appreciate your discussion and assistance with this subject.

Kevin Kolodsick
Analytical Method Development
Ferndale Laboratories