Establishing reference ranges for Factor VIII testing

[kholshoe]

Hi everyone! 

 

We currently perform Factor VIII testing, but we do a very small volume.  The testing isn't batched, but we just do them as they are ordered.  Looking back over the last few months - we normally perform <5/month.  (!!)

 

With new lots of Factor VIII deficient plasma - what's the best way to establish new reference ranges for this test?  Any thoughts?

[Auntie-D]

The standard way is to calibrate you machine based on already established control material, run your new batch of control at least 10 times establishing the values are in the acceptable range provided and then set your mean and standard deviations from this.

Be very careful about using the standard deviations provided by the manufacturers as some of these can actually be 3sd not 2! I was horrified when I found out...

[nziegler]

your original question mentioned "reference ranges" - just to clarify, you don't need to validate reference ranges for patient results when a lot number switches.  regulatory agencies prefer you create your own control ranges, but there is also an "out" for tests that are not performed frequently enough (or for tests in which it would be very expensive) to establish these ranges.

for any of my special coag testing, we use assayed controls, we use the manufacturer mean, but we do tighten the acceptable range to 2sd.  as auntie-d mentioned, many published ranges are, indeed, 3sd - and (are you sitting down?) IL published ranges are a whopping FOUR sd!

[kholshoe]

Thanks everyone!  We have a procedure in place to establish control means and ranges, but my question was actually referring to the test reference range itself.  Ie...how do you establish a new "normal" reportable range with a new lot of reagent?  I'm not sure I'm being clear =P

 

Nziegler - when you stated: " you don't need to validate reference ranges for patient results when a lot number switches" - are you referring to control ranges or assay ranges?  Can you point me to some documenation supporting this?  That would be great!

 

Thanks again =)

[Auntie-D]

You wouldn't change your reportable reference range with each lot, this is only done if you change technique. Your 2sd more than covers batch variability.

[sosier]

You can use a document like CLSI H48A to determine reference ranges or you can use any that are published in the literature with the caveat about S.D.

[SMILLER]

We re-validate reference ranges when we switch lot numbers. (But we do not run factors here.) We run at least 25 presumably normal (hospital associate volunteers not on birth control, coumadin, etc.) to establish a reference range. In general, if the range is very close to what we have been using with the old lot number, our pathologist will not chamge it.

Scott

[kholshoe]

Thanks, SMILLER!  That's exactly what I was looking for.