Comments on USP 467 From a USP Employee

3 posts / 0 new
Last post
Dr. Analytical
Dr. Analytical's picture
Comments on USP 467 From a USP Employee

While visiting Pittcon last week, I talked with a colleague who now works for USP.  When I complained about the reliability of the written USP method, his (paraphrased) comment was,
"the USP method was never meant to be used.  Labs are expected to develop and validate their own procedures. ... It was just published as a guide."
I was wondering if anyone else has come across this attitude, either directly, or as part of a training seminar?  Certainly the method as written in USP is difficult, if not impossible, to valildate, but this is an unfortunate situation when a standard method is published but isn't expected to be actually used?
How is everyone else doing with their validation experiments?  Any comments, questions, problems?
Post them here.

Ivan Delgado
Ivan Delgado's picture
My experience with USP

My experience with USP methods is actually quite negative (unfortunately). Nevertheless I am surprised by the statement you paraphrased. I assumed that USPs were written the way they were mainly because of legacy problems: by the time a method becomes a USP method there is a good chance that there is better technology to perform a similar kind of analysis.
I can't remember which USP we were dealing with last, but it was an ELISA-like method that asked for the use of huge beads (in the order of a few millimeters in diameter). You had to bind an antibody to the beads (four or five of them in a tube) and then essentially run the ELISA-like experiment. The analysis was performed visually, with a yellowing-brown color denoting a positive signal and no color being negative (of course what you really got was shades of yellow-brown in all vials). The trick was to run controls and then compare the negatives and positives to your sample. After a lot of work and not being able to get it to work, we decided to transfer the method from these beads (which by the way could only be bought from a single source) to an ELISA plate. The results we got when the antibody was immobilized on an ELISA plate were like night and day. The experiments were a lot easier to run and the results were as clear as day. Since the USP method was vague enough we were able to use the new method and drop the old one. In addition since this work was defined by the customer, there was little resistance to perform the experiment using the new method (that worked like a charm) instead of the old method (that never worked). 
I do not know how relevant this is to your applications, but at least in my experience there is room to "work around the USP" to get it validated. The USP we dealt with was literally impossible to validate. Luckily the technology involved was very simple so it was just a matter of us "updating" the protocol to what we knew works today. Having said this I can imagine that the number of USP methods that could be validated in this way is limited at best.

dogcatlike's picture
i‘m using 467 method in

i‘m using 467 method in residual solvents assay almost 2 year,not smooth going,very gloomy?peak area repeatability often bad?

i wonder, what's your problem in usp467 method?it may have some shortcomings,but most of it maybe reasonable,how to make some modification to be better?