validation of biological samples, resolution?

[hahe0104]

Hi,
I was wondering if anyone had any information about method validation for biological samples. Are the resolution requirements the same as they are for non-biological samples? I am doing liquid/liquid extraction of plasma and I have two peaks that come up very close to my peak of interest, RT 3.5., the best resolution I can achieve is about 1.4. I have changed the pH, changed column temp (If I decrease my peak runs in to a peak to the left and if I increase it runs in to the peak to the right of it) and attempted to use a gradient (If I use any organic my compound does not stay retained).
Any help would be great.
Thanks

[Gerhard Kratz]

More details please, what column, what mobile phase composition etc.?

[tom jupille]

I don't think there is a specific *required* resolution (while the USFDA *suggests* that Rs should be > 2, it is in fact a suggestion, not a requirement). The real issue is whether resolution is good enough to allow quantitation with the required accuracy and precision, and that is what you have to demonstrate during validation.

If all three peaks are , comparable in size and well-formed, you may be able to get away with 1.4 (by way of reference, Rs of 1.5 provides a 1% overlap between two equal-sized Gausian peaks). On the other hand, if you have a small analyte sandwiched between two large interferences, you may need a resolution of 3 or more.

As Gerhardt indicated, it's hard to make specific comments without knowing more about the separation, but as a general approach, you might try simply coupling two columns in series. That should double the plate count and increase resolution by 40% (to around 1.9) at the cost of doubling your run time and pressure. It's easy to do isocratically; if you run a gradient you have to remember to also double the gradient time.