S chneider & A ndersen : J ournal of AOAC I nternational V ol . 98, N o . 3, 2015  659

well as the published AOAC First Action 2012.25 method (8). A detailed study protocol based on 2012.25 was also provided to the study participants for additional guidance. The collaborative study was designed to meet the requirements of a U.S. Food and Drug Administration (FDA) Foods Program level four full collaborative study chemical method validation (9). Control filets of Channel catfish ( Ictalurus punctatus ) and Atlantic salmon ( Salmo salar ) were obtained from the FDA Center for Veterinary Medicine (CVM) Office of Research Aquaculture Program. Residue incurred catfish and salmon were produced at the CVM by exposing one catfish (2.2 kg) and one salmon (2.8 kg) to individual treatment baths at 25 and 12°C, respectively, containing a mixture of MG, CV, and BG each with a concentration of 2 μg/L. Each treated fish was placed in an exposure tank for 1 h, removed to a clean water tank, and then sacrificed 1 h after the exposure period. One control catfish and one control salmon were also collected from clean water tanks. To prepare homogenized portions of control and incurred muscle, skinless catfish filets (approximately 700 g) and salmon filets with attached skin (approximately 950 g) were separately ground with dry ice to produce fine powders. After overnight dry ice sublimation at –20°C, homogenized samples were sealed and stored at –80°C. Frozen, peeled, and deveined white shrimp ( Litopenaeus vannemei ) were purchased commercially. This product was labeled as a product of Ecuador, farm-raised without the use of antibiotics or growth hormones. Approximately 900 g of these shrimp were ground with dry ice to produce a homogeneous control matrix as described above. Residue incurred shrimp was not available for this study. Surrogate incurred shrimp samples were produced by fortifying 2.00 g (±0.02 g) weighed portions of homogenized control shrimp with analytes to yield the following concentrations: 0.85 µg/kg MG, 0.75 µg/kg LMG, 1.18 µg/kg CV, 0.76 µg/kg LCV, and 1.50 µg/kg BG. Study samples and standard solutions were shipped to each laboratory on dry ice, with instructions to store samples at –80°C and standards at –20°C until analysis. All analytical standards were obtained from Sigma-Aldrich (St. Louis, MO) including MG oxalate, LMG, CV chloride, LCV, BG (bisulfate salt), D5-MG picrate, D5-LMG, D6-CV trihydrate, and D6-LCV. All were indicated as Fluka analytical grade standards, except for LCV and BG that were available at the time only as reagent grade. Individual stock solutions of the dyes (MG, CV, and BG), leuco metabolites (LMG and LCV), and internal standards (MG-D5, LMG-D5, CV-D6, and LCV-D6) were prepared in acetonitrile with nominal concentration 100 μg/mL. Mixed standard solutions (1.000 µg/mL) were prepared in acetonitrile by combining the necessary volume of each stock solution to yield the exact final concentration for all the compounds. One mixed standard solution was prepared for the analytes (MG, LMG, CV, LCV, and BG) and one mixed standard solution for the internal standards (MG-D5, LMG-D5, CV-D6, and LCV-D6). Participants received vials of both mixed standard solutions along with instructions to prepare fresh working solutions from these standards on each day of analysis. Three sets of fish samples consisting of salmon, catfish, and shrimp matrix were also provided to each participating laboratory. Each set consisted of 17 tubes containing weighed portions (2.00 g ± 0.02 g) of homogenized matrix including: six tubes of negative control matrix labeled as calibrants, one tube of negative control matrix labeled as a QC, and 10 tubes of randomly numbered blinded

test samples. Participants were instructed to fortify the six matrix calibrant samples prior to extractionwith the appropriate amounts of working standard solutions to produce matrix fortified with 0, 0.25, 0.5, 1.0, 2.5, and 5.0 µg/kg of the analytes and 2.0 µg/kg of the internal standards. Participants were instructed to fortify the QC sample with analytes (1.0 µg/kg) and internal standards (2.0 µg/kg) after completion of the extraction procedure at the final extract reconstitution step. The QC sample was a post-extraction fortified calibrant that could be used to measure extraction losses when compared to the set of pre-extraction fortified calibrants. For further calibration comparisons in this collaborative study, participants were also requested to prepare and analyze a set of six solvent calibrants with concentrations to match the extracted matrix calibrants (0, 0.25, 0.5, 1.0, 2.5, and 5.0 μg/kg as tissue equivalents). Both the QC sample and solvent calibrant solutions were prepared in acetonitrile with an ascorbic acid concentration of 0.01% to match the composition of the extracted matrix calibrants. The 10 blinded test samples included two negative control matrix samples; six fortified samples with concentrations 0.42, 0.90, and 1.75 µg/kg (in duplicate); and two residue incurred samples (for salmon and catfish). For the shrimp matrix, for which residue incurred tissue was not available, two surrogate “incurred” samples were included in the set of blinded samples, each fortified with 0.75 µg/kg MG, 0.76 µg/kg LMG, 0.85 µg/kg CV, 1.18 µg/kg LCV, and 1.50 µg/kg BG. Study participants were instructed to fortify all 10 blinded samples with internal standard working solution (2.0 µg/kg) prior to extraction . Test samples for each of the three matrixes were labeled with unique letter and color coding (randomized between laboratories), which allowed these to be matched with the corresponding matrix calibrant tubes. The identity of the matrixes was not made known to the participants. Both the 2012.25 method and the collaborative study protocol provided sufficient details to perform the LC-MS/MS analysis; however, participants were given flexibility to choose their analytical instrumentation and optimize the performance of their chosen system. Requirements given to the participants in order to select an appropriate triple quadrupole mass spectrometer were that the instrument would provide sensitivity to detect solvent solutions of the analytes with a concentration of 0.5 µg/L, and that two product ions would be collected for each analyte and one for each internal standard. In addition to the Waters Corp. (Milford, MA) LC-MS/MS system described in the 2012.25 method, participants were provided with optimized source parameters for an Agilent (Santa Clara, CA) 6490 LC-MS/MS system in the study protocol. Finally, study participants were cautioned that triphenylmethane dyes and metabolites are light sensitive and require efforts to reduce background contamination. Participants were cautioned to reasonably protect samples and solutions from excessive light exposure (e.g., place in the dark or cover with foil when not in use), avoid black markers that are a known source of CV, and minimize instrument carryover by using an injection needle wash and/or injecting a blank water sample between blinded samples. Each participating laboratory received a customized report spreadsheet matched to sample matrix and blinded sample coding that automatically displayed calibration curves and provided calculated concentrations as data was entered. On completion of the analyses, the participants returned this completed report sheet, as well as a summary of specific

Made with FlippingBook flipbook maker