Alternate, in this context, means another way than by buying commercial available AMR samples.
Likely, most of your analytes are covered by linearity surveys from CAP...but not all!
LDL, Urine Protein, CSF Protein, Lactic Acid and the list goes on....
So what are some "Alternate Methods". My favorite is using a calibrator of a different lot used to calibrate the assay in question. But sometimes a calibrator does not come close enough to the high end of AMR. I like to use the "rule" of 10% as being "close enough".
Secondly I like to use a high patient sample diluted to cover the AMR. Using calculated target values the points covering the low, mid, and high can plotted to validate the analyte over the AMR of the analyzer.
How? Read on and feel free to copy and modify the following your your procedures :)
Note: I got the jist of this from the folks at Binding Site who make our Kappa and Lambda Free Light assay reagents (used on the Roche 8000).
Making a Serial Dilution or Preparing
Samples for AMR Validation
When you don’t have linearity samples to validate your
Analytical Measuring Range then you can prepare dilutions of a high sample that
covers the range, all the way down to 2% of the upper value. Using this method avoids making repeatedly
higher dilutions from the high sample and the use of too many different
pipettes…both likely sources of error.
Needed:
·
Class A pipettes or NIST certified pipettes (an
adjustable pipette delivering 250 and 750uL works best.
·
Sample of at least 2.25mL containing the analyte
being tested.
·
Appropriate diluent.
If the sample you have is (ideally) above the upper
measuring range of your analyzer you will need to spike it to be under the
upper limit AND with-in 10%. Spike and
retest until this is achieved. Remember
you will need at least 2.25 mLs of this sample.
Once this “100%” solution is achieved, measure 3x and
take the average. From this sample and
value you will make dilutions and calculate the expected recovery. Note that not all levels are necessary for the
AMR study but the lowest, mid -range, and highest at a minimum.
Arrange 7 tubes in a rack and label with the final
dilution as stated in the chart below, and prepare the indicated
dilutions. Use a fresh tip between
samples and good pipetting practices.
Fill in the expected recovery.
|
Final Dilution
|
Expected
Recovery
|
Volume of 100%
solution
|
Volume DIluent
|
|
75%
|
|
750uL of 100%
dilution
|
250uL
|
|
50%
|
|
750uL of 100%
dilution
|
750uL
|
|
25%
|
|
750uL of 50%
dilution
|
750uL
|
|
12.5%
|
|
750uL of 50%
dilution
|
750uL
|
|
6.25%
|
|
750uL of 12.5%
dilution
|
750uL
|
|
3.13%
|
|
750uL of 6.25%
dilution
|
750uL
|
|
1.56%
|
|
750uL of 3.13%
dilution
|
750uL
|
Run samples in duplicate, if may be beneficial to extend
the analyzer’s reportable range in the event the lower samples exceed the AMR.
Crunch the data in a statistical program such as EP
Evaluator from Data Innovations. In this
case use the statistical module Accuracy & Linearity- Linearity and
Calibration Verification, choose Confirm Linearity and Confirm Accuracy. It is ideal to set limits for error based on
CLIA, CAP, API, or other sources and this should be approved by the laboratory
director.
For example: The total allowable error (TAE) for ALT is
20% (CLIA). IF using EP Evaluator enter
also a budget for systematic error.
Hopefully, no more than 50% for systematic error budget is used.
The statistical software will indicate if the experiment
passes or fails. Alternatively, a graph
can be plotted and each point can be assessed if it is with-in the TAE. Just add a column to the above chart next to
expected recovery for the result and another column for the acceptable range
(ALT: up to +/-20% from the Expected Recovery).
Comments? Suggestions?
email: joe@abetterlabnow.com
No comments:
Post a Comment