Résumé
We compared the accuracy of 6 commercial systems for Aeromonas identification by testing 87 clinical isolates in routine conditions, using partial rpoB gene sequencing as the reference standard. The systems were API-20E, API-32GN, the ID-ON card with the Vitek2 system (bioMerieux, Marcy l'Etoile, France), the identification portion of the NFC47 panel (Micro Scan Walk/Away system; Siemens Healthcare, Sacramento, CA), ID69 (Phoenix system; BD Diagnostic Systems, Sparks, MD), and GN2 microplates (Omnilog system; Biolog, Hayward, CA), for which 67 (77.1%), 80 (91.9%), 72 (82.7%), 70 (80.5%), 64 (73.5%), and 59 (67.8%) isolates, respectively, were correctly identified at the genus and species level. Confusion with Vibrio affected 6.9% and 16.1% of results obtained with NFC47 and API-20E, respectively. Overall, the accuracy of identification for aeromonads was hampered by outdated databases and taxonomy, weak algorithms, and impractical additional tests. Commercial identification systems should be redesigned to make Aeromonas identification algorithms more robust and to cover infrequent clinical species of this genus. (C) 2010 Elsevier Inc. All rights reserved.