Agriculture Reference
In-Depth Information
can be chosen. Ideally, 60 blank samples and 60 spiked samples of one particular
matrix should be taken in order to determine CC
. The sample numbers may be
reduced in accordance with the STC:AL ratio as stipulated in the bullet points above.
Each target analyte should be tested separately if the method does not produce the
necessary speci
β
city to distinguish between analytes.
The blank and spiked samples should be analyzed on different days and preferably
using different analysts using the prede
of
the method can be judged using the numbers of false-compliant results that are within
the criteria speci
ned method. The detection capability CC
β
ed above.
3.6.4 Determination of the Applicability
3.6.4.1 Same Matrix and Different Species
It cannot be assumed that CC
will be the same, even for the same matrix between
species, for example, bovine muscle and porcine muscle. Therefore, CC
β
must also be
established for all the analyte(s) for the additional species. However, provided that the
AL is the same for all species and the matrix stays the same, then the numbers of
samples tested can be reduced from 20 blank and 20 spiked to 5 blank and 5 spiked for
each additional species.
β
3.6.4.2 Different Matrix and/or Different Species
Again for the same species it cannot be assumed that CC
will be the same for
different matrices, for example, bovine muscle and bovine liver. Therefore, CC
β
must
β
also be established for all the analyte(s) in each new matrix. CC
can be determined
for each new species/matrix combination by analyzing 20 blank samples, for example,
porcine livers, and the same 20 blank porcine livers overspiked at the STC.
Ruggedness studies (effects of variations in methodology that might affect the
results) are described in Commission Decision 2002/657/EC [18].
β
3.6.4.3 Continuous Veri
cation Using Quality Control Samples
As with all methods, initial method validation must be supplemented with ongoing
quality control to ensure that method performance remains acceptable. Each batch of
analyses should include a
“
blank matrix
”
(screen-negative control sample) and a
“
(screen-positive control sample). The spiking concentration
should be at the STC. If the
spiked blank matrix
”
“
spiked blank matrix
”
produces a negative result (i.e., it is
below the cutoff level) or the
gives a positive result (i.e., it is greater
than the cutoff level), then the batch of analyses should be discarded. Results from
these QC samples should be monitored to verify that the screening method is working
reliably and has a false-compliant rate of no more than 5% for all target analytes. A
minimum of 20 QC results should be produced annually and reviewed to check that
the method is continuously working reliably and that no more than 5% of the
“
blank matrix
”
“
spiked
blank matrix
”
samples have fallen below the cutoff level.
