ERP Micro December 2019

B astin et al . : J ournal of AOAC I nternational V ol . 102, N o . 5, 2019  1615

the organisms’ true identities were correctly identified following the Bruker MALDI Biotyper methodology, producing correct confirmation and identification rates of 100.0%. Detailed confirmation and identification results are presented in Tables 2017.09I and J and Table 3 of the Supplemental Information.

comparison to an appropriate 16S database (Accugenix or GenBank) by analyzing up to 500 base pairs. The data obtained during the precollaborative and OMA evaluations indicated that all Cronobacter, Salmonella , and Campylobacter species can be identified to the genus level. The precollaborative and OMA data generated also demonstrated that the non- Cronobacter, non- Salmonella , and non- Campylobacter organisms can be identified to the species level. Regardless of the platform (microflex LT/SH or microflex LT/SH smart), the support device for the sample preparation (Galaxy) software, or target plate being used, the method produced robust and reliable results.

Discussion

No negative feedbackwas 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. Only one isolate of 2068 identification results (0.05%) was not identified by the Bruker MALDI Biotyper method in the exclusivity panel of the Cronobacter evaluation, but the participant did not follow the MALDI Biotyper complete workflow required in that specific case. If the workflow is not completely followed, the results can be less reliable. The smear of the isolate on the target plate is technique driven. The proteins from the agar may cause a misleading result; therefore, it is critical to select a pure colony not containing any of the agar. All of the other 2067 isolates were correctly identified, leading to a 100.0% correct identification rate. In comparison, 24 of 695 isolates (3.3%) were misidentified using traditional confirmatory procedures. The eDT was only run for 29 of 2068 isolates. The DT procedure provided a reliable identification result at the species level for most of the Gram-negative isolates (98.6%), and the MALDI Biotyper method proved to be robust, as various growth temperatures, media, culture ages, and different operators had no notable impact on the bacterial identification rate (14). During the testing of the method extension ( Campylobacter collaborative), it was determined that two of the participants (participants 16 and 17) required a software patch update to allow for use of disposable targets. This delayed testing slightly for these two participants, which may have resulted in the issues they had with recovery on the mCCDA. Therefore, their data for mCCDA have been removed; however, the data generated for the additional agars, which produced viable growth for each isolate, were included in the overall results. Another participant (participant 3) was not able to complete the full workflow because they had an unexpected emergency schedule change. For the isolates that they had completed, the data were included in the statistical analysis. The isolates that were not able to be analyzed by the entire workflow were removed from the statistical analysis. All 1340 isolates were correctly identified, leading to a 100.0% correct identification rate. During testing, all results were obtained using the DT method. The eDT and EXT procedures were not utilized by the collaborators. The DT procedure provided a reliable identification result at the species level for all of the Gram-negative isolates (100.0%). Overall, the data generated during this evaluation demonstrate the reproducibility of this method using both disposable and reusable target plates. The candidate method produced a higher percentage of positive agreement when compared to the confirmation procedures of the reference methods (ISO, BAM, and MLG) for the confirmation and identification of Cronobacter and Salmonella. All non- Cronobacter, non- Salmonella , and non- Campylobacter Gram-negative organisms were compared by following the confirmation procedures and

Recommendations

It is recommended that the scope of the Bruker MALDI Biotyper AOAC OMA be extended to claim the confirmation and identification of Salmonella species, Cronobacter species, Campylobacter species, and Gram-negative organisms from select agars.

Acknowledgments

We would like to extend a sincere thank you to the following collaborators for their dedicated participation in this study: Salmonella species: Maryse Rannou, Claudie Le Doeuff, and Sarah Peron, Alexander Leclerc, Institut Pasteur, Centre National de Référence & Centre Collaborateur de l’Organisation Mondiale de la Santé pour Listeria (Paris, France) Benoit Thuillier, LABOCEA (Quimper, France) Michael Treilles, Laboratoire d’Analyses Sèvres Atlantique (Champdeniers, France) Elodie Cauvin, LABEO (Saint Lo, France) Alice Marthino, Groupe CARSO (Venissieux Cedex, France) Florence Crombe, Institut Scientifique de Sante Publique (Brussels, Belgium) Annet Heuvelink, GD Animal Health (Deventer, The Netherlands) Greetje Castelijn, Nederlandse Voedsel en Waren Autoriteit (Wageningen, The Netherlands) Roy Betts, Campden & Chorleywood Food Research Association Group (Chipping Campden, United Kingdom) Matteo Capocefalo, ALS (Rotterdam, United Kingdom) Catherine Cockcroft, Eurofins (Wolverhampton, United Kingdom) Thomas Koch, Eurofins Central Laboratories Friedrichsdorf Specialized Nutrition Testing Services (Friedrichsdorf, Germany) LotharKrockel,MaxRubner-Institut–Bundesforschungistitut fur Ernahrung und Lebensmittel – Standort Kulmbach (Kulmbach, Germany) Uwe Schroder, Intertek Food Services GmbH – Standort Bremen (Bremen, Germany) Cronobacter species: William Glover, Raymond Gee, and Michael Tran, Washington State Department of Health (Tumwater, WA) Paul Park, California Department of Public Health (Richmond, CA) ADRIA Développement (Quimper, France) Thomas Charrier, Eurofins (Nantes, France)