ERP Micro December 2019

1608  B astin et al . : J ournal of AOAC I nternational V ol . 101, N o . 5, 2018

throughout Europe participated in this study. The study was organized and managed by ADRIA Développement (France). All 15 laboratories had one participating analyst. Each collaborator analyzed 24 blind-coded organisms: 16 target organisms and 8 nontarget organisms. For the Cronobacter portion, a total of seven laboratories throughout the continental United States participated in this study. The study was organized and managed by Q Laboratories, Inc. (USA). Five of the laboratories had two separate analysts participate, one laboratory had three analysts participate, and one laboratory had one participant. Each participant analyzed 24 blind-coded organisms: 16 target organisms ( Cronobacter or Salmonella ) and 8 non- Cronobacter or non- Salmonella organisms. Tables 2017.09A – D and Tables 2017.09E – H summarize the interlaboratory results per isolate for Salmonella and Cronobacter . Detailed results for each collaborator are presented in Tables 2017.09I – L and Tables 2017.09M – P . Individual collaborator, sample results, and type of target, i.e., reusable stainless steel target or disposable Biotarget 96, used for the analyses are presented in Tables 1–4 of the Supplemental Information. Inclusivity results .— MALDI Biotyper method .—For the target organisms (inclusivity) analyzed, 719 out of 719 isolates were correctly confirmed and/or identified following the Bruker MALDI Biotyper method, to produce a correct identification rate of 100.0%. For one of the participants, S. Newport Ad540 was not tested on RSA. Inclusivity results .— Reference method .—When analyzed following the reference method, 238 out of 239 isolates were correctly confirmed and/or identified, to produce a correct identification rate of 99.6%. S. Saintpaul A00C002 was misidentified as Aeromonas hydrophila by one participant following the traditional reference methods. Exclusivity results .— MALDI Biotyper method .—For the nontarget organisms (exclusivity), 355 out of 355 isolates were correctly confirmed as non- Salmonella organisms and correctly identified to the organism’s true identification following the Bruker MALDI Biotyper methodology, to produce a correct confirmation rate and identification rate of 100.0%. For multiple participants, four of the organisms did not grow on RSA and one organism did not grow on XLD and were therefore not analyzed from those media. Exclusivity results .— Reference method .—When confirmed following the reference method, 117 out of 120 isolates were correctly excluded as Salmonella species, to produce a correct confirmation rate of 97.5%. For all isolates that grew, the reference method correctly identified 105 out of 120 isolates, to produce a correct identification rate of 87.5%. The traditional confirmatory reference methods had significant difficulty identifying Escherichia hermanii Ad457, with 10 incorrect identifications. Yersinia enterocolitica A00C066 was misidentified once as Klebsiella oxytoca , C. sakazakii Ad1418 was misidentified as Enterobacter cloacae in one case, and A. hydrophila Ad1570 was also misidentified by one participant as Vibrio fulvialis following the traditional confirmatory reference methods. There were three instances when Citrobacter braakii Ad2701 was incorrectly identified as S. enterica arizonae following the traditional confirmatory Salmonella

reference methods. Some discrepant results were also observed for the oxidase test or the Gram staining for, respectively, three and one participants. Detailed results are presented in Tables 2017.09A – D , Tables 2017.09I – L , and Table 5 of the Supplemental Information.

Cronobacter

Inclusivity results .— MALDI Biotyper method .—For the target organisms (inclusivity), 671 out of 671 isolates were correctly confirmed and/or identified following the Bruker MALDI Biotyper method, to produce a correct identification rate of 100.0%. For one of the participants, C. condimenti QL17031.1 did not contain growth on CCI and was not analyzed by the alternative method. Inclusivity results .— Reference method .—When analyzed following the reference method, 214 out of 224 isolates were correctly confirmed and/or identified, to produce a correct identification rate of 95.5%. The misidentifications following the reference method resulted for the following species: E. amigeneus, E. cloacae, and E. gergoviae . Exclusivity results .— MALDI Biotyper method .—For all isolates that grew, 322 out of 323 isolates were correctly identified following the Bruker MALDI Biotyper method, to produce a correct identification rate of 99.7%. One participant did not provide an identification result for the strain Serratia marcescens QL110071 isolated on CCI agar. The DT procedure was repeated, and the eDT was run as described in the MALDI Biotyper workflow, but despite no identification result being obtained, no extraction protocol was run to complete the described MALDI Biotyper workflow. Across several participants, there were a total of 13 organisms for which no growth was observed on selective culture media. These organisms could not be analyzed by the alternative method. Exclusivity results .— Reference method .—When confirmed following the reference method, 112 out of 112 isolates were correctly excluded as Cronobacter species, to produce a correct confirmation rate of 100.0%. For all isolates that contained growth, the reference method correctly identified 105 out of 112 isolates, to produce a correct identification rate of 93.8%. Four of the participants obtained a Chromobacterium violaceum result for Pseudomonas aeruginosa ATCC 35032, one participant obtained a result of C. farmeri for E. coli ATCC 10536, and seven participants obtained a result of E. cloacae for C. farmeri ATCC 51633 following the traditional confirmatory reference methods. Detailed results are presented in Tables 2017.09I – P and Table 6 of the Supplemental Information. No negative feedback was provided regarding the performance of the Bruker MALDI Biotyper method by the participants, which included academia, industry, and regulatory agencies located within the European Union and the United States. As described by Elbehiry et al. (6), approximatively 30 min per isolate are required from target plate to the final results, and nearly 2 h to investigate a full 96-spot target plate. Only one isolate of 2068 identification results (0.05%) was not identified by the Bruker MALDI Biotyper method in the Discussion