Method Validation

[knicoleb]

I have never validated a method, but I am attempting to validate a method that is currently being used in an R&D project. I calibrate a GC (FID) for about 20 analytes. Should I run each of these analytes at the lowest level for this method? Or will it be suitable to just to have a few analytes run multiple times to acquire the necessary information for calculations?

[chromatographer1]

LOD and LOQ, limit of detection and limit of quantification are factors which should be documented in a validated method.

Either you know the values or you will be guessing.

Which do you think the reviewers/regulators would prefer?

What happens when you present a 'validated' method that does not show what the actual values are?

You should read the literature (especially, the USP in the USA) about method validation requirements.

The legal issues about inadequate validation work you PERSONALLY will not find agreeable.

best wishes,

Rod

[Johnny Rod]

Look up the ICH website and there is information there on a common approach to method validation. You can most likely do most of the validation (such as precision, linearity, accuracy etc.) for just one or two analytes, and these results will then indicate that the others will follow. For LOD and LOQ you really need to do all analytes as they will have different detector responses.