Chromatography Forum

Can anyone point me to industry reg's which would tell me what acceptable accuracy levels are for LOQ? The method I am currently working on provides 85 - 90% recoveries at the LOQ. I had hoped for greater than 90% recovery, and will work towards that end if need be. However, in the reg's that I have read, they don't specify a specific accuracy level for LOQ.

Thanks in advance for comments/thoughts.
Shawn

Why are you doing recoveries on your LoQ? Are these impurites or something? I'm not sure if there is even any guidance on this topic.

Rob-
This is an impurity analysis. According to ICH guidelines I need to have a method capable of quantitatively evaluating any impurity >/= 0.1% of label claim. The current method provides ~90 - 95% recovery at 0.01 mg/g (0.1% of label claim). However, at 0.008 mg/g my recoveries are approximately 10% less.

Because I have to define an LOQ for this method I have chosen 0.008 mg/g because my standard curve gives good linearity through this range and I want the bottom of the curve to cover the 0.01 mg/g concentration. Theoretically I could move the bottom of the curve up to 0.009 and this may improve my recoveries.

According to the USP, "LOQ is the lowest concentration of analyte in a sample which can be quantitatively determined with suitable accuracy and precision". ICH guidelines are almost verbatim.

The question becomes what is "suitable" and where is it defined for validation?

tnx again-
shawn

As far as I have understood the guidelines (ICH Q2) there is no specific point for "accuracy at LoD". For the impurity you have to detemine the Accurcy over a certain range (i.e. from LoQ to 150% of spec.) and in a second experiment (or at least in a different section of your report) you deal with the LoD/LoQ. For LoD/LoQ some tests and acceptance criteria are given in the guideline and passing one this tests and establishing accuracy not to far from LoQshould be enough.
Thats a personal opinion, however.

Alex

I agree with Alex here, on this point. What is the value of accuracy of an LoQ component for a designated impurity? I would expect to evaluate accuracy (re: recovery) of your main component peak only at levels of say 50 - 150% of a nominal concentration but not on a low level impurity... how would you spike accurately such a low amount anyway without introducing appreciable errors? The ICH guidelines may state that you have have to quantitate down to the 0.1% level of the main peak, but does it say also that you have to specifically measure recoveries at this level? What is suitable recovery at this level even. AS per usual the wording is rather vague. For recovery precision, our own internal guidelines for pharmaceutical products specifies a precision of 20% rsd for impurity peaks. I would say your recovery values are rather reasonable at the level you are trying to analyse!

Shawn,

It sounds like your method is fine, but I would set your LOQ to 0.05% of label claim if possible. 0.08% does not give you enough room for accurately tracking impurity levels, if this something of importance to you (stability studies, etc.). If you only validate to 0.08% you will not be able to report values on an impurity below that level. You can widen your recovery window for validation to 80-120% for the the entire imp. conc. range you are quantifying (or even wider for anything below 0.1%, It's really up to what you want out of the method). I typically check my LOQ by making sure I get a peak area %RSD of less than 10% for six replicate injections of my LOQ sample, and I like to have 80-120% acc. at my LOQ, but I have accepted methods that did not fit that criteria. Are you validating/qualifying this method for known as well as unknown imps? If they are known and you have reference standard make sure you check their response factor and adjust accordingly.

Daren

Thanks to you all for taking the time to throw in 2 cents.

I think my plan is to:
Establish the 0.1 mg/g as my LOQ (>90% recoveries) and report all values which have a lower concentration as less than 0.1 mg/g.

Daren - liked the % RSD thoughts for the LOQ.

Again, thanks for all of the input, and apparently the answer to my original question is most likely "Suitable accuracy and precision are going to be defined by the chemist who developed the method and if the FDA has an issue with that, the FDA will let you know".

Again many thanks for weighing in. . . .

s