66 / 71
S
chneider
&
A
ndersen
:
J
ournal of
AOAC I
nternational
V
ol
.
98, N
o
. 3, 2015
665
BG treatment bath for 1 h (12). By comparison, the salmon and
catfish incurred for this collaborative study were exposed to
only 2 µg/L of BG for 1 h.
Preliminary Data Analysis
All reported data were compiled and examined for validity.
Data were omitted for cause when participants reported specific
difficulties, such as a sample being lost due to spillage, or
inadvertent combination with another sample. While Method
2012.25
allows the measure of linear correlation to be as low as
R
2
= 0.95, in this collaborative study, single outlier calibration
points were excluded from the calibration when R
2
was less
than 0.99 and the omission of one point would result in R
2
≥0.99
correlation for the remaining five calibrants. In three instances
(CV in catfish for one laboratory, MG and BG in shrimp for
another), linear correlation could not be achieved by deletion of
a single calibration point. For these three, all data reported by
the laboratory for the particular analyte in the particular matrix
were omitted from the statistical analysis.
Statistical Analysis
A determination of repeatability and reproducibility was
performed using the AOAC International Interlaboratory Study
Workbook for Blind (Unpaired) Replicates, v. 2.1 (14), which
was developed to implement the AOAC INTERNATIONAL
guidelines for the AOAC Official Method Program (15).
Table 4(a). CC
α
and CC
β
for triphenylmethane dyes and metabolites in seafood matrix
CCα, μg/kg
CC
β
, μg/kg
MG
LMG
CV
LCV
BG
MG
LMG
CV
LCV
BG
Salmon
0.24
0.17
0.18
0.29
0.35
0.27
0.19
0.20
0.33
0.38
Catfish
0.27
0.17
0.21
0.15
0.41
0.31
0.19
0.24
0.17
0.45
Shrimp
0.29
0.14
0.32
0.28
0.43
0.33
0.16
0.36
0.32
0.49
Table 4(b). MDL and LOQ for triphenylmethane dyes and metabolites in seafood matrix
MDL, μg/kg
LOQ, μg/kg
Salmon
0.12
0.07
0.07
0.11
0.21
0.47
0.27
0.26
0.45
0.85
Catfish
0.13
0.06
0.12
0.12
0.30
0.53
0.24
0.47
0.47
1.22
Shrimp
0.33
0.09
0.23
0.09
0.34
1.35
0.38
0.91
0.35
1.35
Table 5. Comparison of the accuracy (trueness and precision) of 0.9 µg/kg fortified samples using three different
calibration methods for residue quantification; data represents duplicate matrix spikes from 10 laboratories (
n
= 20)
MG
LMG
CV
LCV
BG
Trueness,
(avg.
recovery, %) RSD, %
Trueness,
(avg.
recovery, %) RSD, %
Trueness,
(avg.
recovery, %) RSD, %
Trueness,
(avg.
recovery, %) RSD, %
Trueness,
(avg.
recovery, %) RSD, %
Salmon
Extracted matrix
a
95.3
17.7
101.1
11.9
99.5
8.7
102.4
8.8
99.1
14.8
Post-extraction
fortified matrix
b
96.8
19.2
112.4
11.3
113.4
8.4
116.5
11.0
162.8
28.0
Solvent
c
84.6
14.6
101.2
18.5
104.1
10.5
103.2
10.8
128.4
40.3
Catfish
Extracted matrix
77.5
24.2
107.7
6.3
101.8
8.6
105.2
6.7
90.4
26.2
Post-extraction
fortified matrix
88.8
22.9
116.2
5.7
115.1
6.0
115.2
4.7
163.2
30.5
Solvent
79.3
25.9
104.2
7.4
113.1
10.8
95.5
18.8
140.1
42.6
Shrimp
Extracted matrix
95.3
17.7
105.4
7.8
105.6
15.9
105.9
5.6
101.0
24.7
Post-extraction
fortified matrix
105.5
15.3
118.2
7.2
125.7
21.0
113.3
7.0
145.2
19.5
Solvent
92.9
14.5
102.0
12.0
117.8
27.8
102.9
19.7
120.3
27.6
a
Matrix fortified at calibration concentrations and then extracted to produce a set of six standards: 0, 0.25, 0.5, 1.0, 2.5, and 5.0 µg/kg (tissue
equivalent
d
).
b
Calibration standard based on one portion of tissue extracted and the matrix extract fortified at a concentrations of 1.0 µg/kg (tissue equivalent).
c
Six standards prepared in solution at tissue equivalent concentrations.
d
Concentrations are equivalent to the amount present in the 2 g sample portion. Method results in a 2.5-fold concentration of residues in the extracts.