2. AOACRIChemContMethods-2018Awards

664  S chneider & A ndersen : J ournal of AOAC I nternational V ol . 98, N o . 3, 2015

Table 3. ( continued )

Analyte added, µg/kg

Statistical outlier lab No. a

Mean concn, µg/kg

No of labs

No. of replicates b

Trueness (recovery, %)

RSD r

,

RSD R

, % HorRat

Analyte Matrix

s r

s R

%

Shrimp

0.42 0.90

None None

14 14

28 28

104.8 106.7

0.44 0.96

0.03 0.03 6.48 7.81 0.04 0.06 3.83 6.11

0.15 0.13

1.75

None

14

28

102.3

1.79

0.08 0.12 4.36 6.52

0.16

Inc:1.18

None

14

28

98.3

1.16

0.04 0.06 3.39 5.40

0.12

10 c , 2 d

BG Salmon

0.42

12

24

104.8

0.44

0.04 0.08 8.02 19.07 0.37

0.90

None

14

28

101.1

0.91

0.07 0.15 7.55 16.96 0.37

1.75

None

14

28

96.0

1.68

0.09 0.22 5.31 12.98 0.31

Incurred

None

14

28

1.48

0.12 0.31 8.21 20.75 0.49

12 c , 11 e

Catfish

0.42

12

24

104.8

0.44

0.02 0.09 5.17 21.51 0.42

0.90

None

14

28

101.1

0.91

0.05 0.36 5.90 39.36 0.86

1.75

None

14

27

97.7

1.71

0.14 0.59 8.23 34.46 0.83

2 e

Incurred

13

26

1.07

0.05 0.25 4.98 23.66 0.53

Shrimp

0.42

None

13

26

107.1

0.45

0.04 0.14 9.57 30.92 0.61

0.90

None

13

26

107.8

0.97

0.10 0.28 10.13 28.90 0.64

1.75

None

13

26

99.4

1.74

0.18 0.70 10.57 40.48 0.97

Inc:1.5 0.07 0.42 4.31 27.57 0.65 a Laboratory numbers were randomly assigned and do not correspond to the sequence listed on either the title page or in the Acknowledgments . b No. of replicates after removal of invalid data and outliers. c Outlier by the Cochran test. None 13 26 102.7 1.54

d Outlier by Grubbs 1 test. e Outlier by Grubbs 2 test.

the composition of acetonitrile from 90 to 40% from 6–7 min and held at 40% for a total run time of 12 min. Compared to Method 2012.25 (20 min run time), overall chromatographic run times for the six laboratories that adjusted the mobile phase ranged from 7 to 25 min. Many participants made adjustments to the injection volume, with eight laboratories reducing the injection volume from 20 to 2, 3, 5, or 10 μL. Two laboratories reduced the column flow rate from 250 to 200 μL/min, one increased flow rate to 300 μL/min, and three laboratories used column temperatures of 25 or 35°C instead of the method stated 30°C. Twelve participants used the exact column specified in the protocol, and the remaining laboratories used a Phenomenex (Torrance, CA) Prodigy ODS-3 100 × 2 mm, 3 μm or a Waters Corp. Atlantis dC 18 100×2.1 mm, 3.5 μm column. Most participants used a guard column, although four did not. Avarietyof triple quadrupolemass spectrometerswere usedby the participating laboratories, includingAB Sciex (Framingham, MA) 5500 QTrap (5), AB Sciex 4000 QTrap (2), AB Sciex 3000, Agilent 6490 (2), Agilent 6460, Thermo Scientific (San Jose, CA) TSQ Vantage, Thermo Scientific Quantum Discovery Max, Waters Corp. Micromass Quattro Micro API, Waters Corp. Acuity TQD, and Waters Corp. Quattro Premiere XE. All LC-MS/MS systems used in this study provided suitable data. Participants were requested to optimize mass spectrometer source and ionization parameters to yield acceptable response for the desired analyte range. Two participants analyzed extracts from each test sample on both Agilent and AB Sciex systems. These laboratories obtained comparable study results

from their Agilent 6460/AB Sciex 4000 or Agilent 6490/AB Sciex 5500 QTrap analyses, further illustrating the suitability of this method for varied instrumentation. Mass spectrometric transitions provided in the method worked well for most analytes (Table 2). In the case of CV, some participants found that the qualitative transition ( m/z  372→251) provided a peak with low signal. As a result, four participants used a substitute transition for qualitative purposes ( m/z 372→340). One laboratory inadvertently used this alternate transition for quantification of the catfish and shrimp matrixes. Transition m/z 372→340, when used, provided acceptable results. Although MS/MS parameters and retention times were listed in the First Action method and in Table 2 for the metabolite leucobrilliant green (LBG), this analyte was not specifically included in Method 2012.25 LC-MS/MS validation nor required for inclusion in the collaborative study. Analytical standards of LBG are not commercially available and have limited stability (12, 13). Nevertheless, three of the participating laboratories submitted data for the LBG transitions. None of the study samples were fortified with LBG, but as the expected metabolite from BG exposure, LBG may be present in incurred samples. On review of the data from three laboratories, potential responses for the LBG transitions for incurred salmon and catfish were not large enough to be distinguished from nonincurred samples. Without an LBG standard, MS/MS optimization and retention time comparison could not be performed by these laboratories. The First Action method authors detected the LBG metabolite in the concentration range 8 to 18 μg/kg for trout placed in a 100 µg/L