James_Ball wrote:
Ok, so I am now having a problem with our new QA/QC officer on methods EPA300.0 and EPA300.1. He is saying that since the method says we must inject a well mixed sample, that we MUST shake the sample prior to injection and not filter it before it goes on the instrument. When we argue the point that it needs to be filtered he says that the EPA chemist are more experienced than we are and if it needed to be filtered they would have put that into the method.
Well, before I was retired, I had to deal with similar, from both QA and from my pointy-haired boss (ala Dilbert comic strip). Of course the sample should be mixed. And EPA, etc. don't generally specify removing particles and undissolved stuff by a certain procedure anyway, that is just good general practice. Sure, one should investigate (like in the quote below) as to whether filtration alters the analyte level, or introduces extraneous interferences etc., but that's all understood as part of good practices. Just like EPA doesn't specify "JT Baker" or other brand of solvent, or filter or autosampler vial, doesn't mean that such should be checked out.
And QA should NOT assume that EPA guy/writer knows more about that procedure than you do, could be right, could be wrong. When I worked with USP and house-developed and house-validated methods, had to fight with QA and supervisor all the time. Stuff like how "flat" the baseline of a placebo needed to be in the elution time of the analyte; when magnified enough, anything looked like little bumps (remember: we had fragranced consumer products, and most pharmaceuticals were not fragranced).
We had hand sanitizer products with 60% or more ethanol, we still had to validate the procedure for our product because USP only had for water-ethanol solutions, and we also had to determine ethanol detection limits even though that was a zillion orders of magnitude less than our working concentrations.
Hornet wrote:
I would analyze the same sample twice. The first is the filtered sample and the second one after centrifuge and not mixed.
If the difference between the two results is less than your repeatability limit then you know your procedure is good.
I still feel that filtration (after checking for analyte loss or interferences) is a common practice and does NOT need to be validated every for every procedure. But hopefully what Hornet suggested should convince QA.
One time I had to document in a report that QA Director was present at such meeting on certain date and decided for the company that he wanted the active reported as total active, even though some of the active (an acid) esterified with the substrate (alcohol) in situ to form esters.
I was not comfortable with that, not the same as (for example) reporting salicylic acid content as the amount of sodium salicylate acidified to be salicylic acid so all could be quantified....