Chromatography Forum

Hi,

I have a question on method validation guidelines, specifically EDMA guidelines, regarding Between Accuracy and Precision.

It says that:
".... LOQ, low, medium and high QC samples from at least 3 runs analysed on at least 2 different days should be evaluated...."

But, my question is: How many replicates for each level?

I usually analysed 5 replicates by level, I calculated the mean, and either Between-Accuracy or Precision.
But, could I analysed only 1 replicate by level from 3 or more runs, and calculate Between-Accuracy and Precision from these values?

Thanks for your help!!!

Hi again,

I mean, could I calculate inter-day accuracy by comparing several measurements, for example 6, from several days, 6 days, but only from 1 replicate each day?

Or otherwise I should calculate inter-day accuracy by comparing several measurements (n=6) from several days (6 days), from severals replicats each day (for example, 5 replicates)? So that, comparing the mean of each day?

Maybe somebody knows about some excel template to do a Method Validation?

Thanks again!!

First, let's clarify that you're working on a bioanalytical method. The scope of this guideline reads, "This guideline provides recommendations for the validation of bioanalytical methods applied to measure drug concentrations in biological matrices obtained in animal toxicokinetic studies and all phases of clinical trials. As ligand binding assays differ substantially from chromatographic analytical methods, separate validation recommendations for ligand binding assays are provided." If you're working on something that isn't in a biological matrix, this guideline doesn't strictly apply to you.

Also noteworthy is that this guideline is recommendation. You can do what you want, you just have to be able to defend it.

Hi.
You can perform the assay repeatedly, over 2-3 days, on at least 3 concentrations each day.
Same data set can be used to analyzed for accuracy, and precision, for each concentration level.

If you have bio-matrix, set the limits realistically, perhaps +/- 25% error (or could be higher).
Best
Alfred