Evaluation of an automated system for identification of anaerobic bacteria

Publisher Website
Google Scholar
Abstract
A fully automated computer-assisted system (ATB system, bioMérieux, France) which uses disposable microenzymatic panels was evaluated for its ability to identify 215 strains of anaerobic bacteria (clinical isolates and reference strains). All strains were examined using conventional identification protocols and by gas chromatographic analysis of short-chain fatty acids. Automated reading of Rapid ID32A test kits (bio-Mérieux, France) by the ATB system gave correct identification for 195 strains (90.7 %): 92.25 % of gram-negative anaerobes (116 strains) and 89 % of gram-positive anaerobes (99 strains) were correctly identified. Twelve strains (5.6 %) were incorrectly identified and 8 strains (3.7 %) were not identified by the system. For some strains in theBacteroides fragilis group, forClostridium difficile and for theFusobacterium genus, additional tests suggested by the ATB software were necessary to reach a final identification at the species or genus level. On the basis of the high incidence of correct identifications and the comparison of these results with those obtained previously using other commercially available kits, the ATB system was found to be a reliable method for identification of anaerobic bacteria in clinical laboratories.