Accuracy and reproducibility of the 4-hour ATB 32A method for anaerobe identification.

The ATB 32A system (API System SA, La Balme les Grottes, Montalieu-Vercieu, France) was evaluated for use in the identification of 214 anaerobes. Organisms included 73 isolates of the Bacteroides fragilis group, 24 Bacteroides spp., 10 fusobacteria, 43 clostridia, 28 cocci, and 36 gram-positive, nonsporeforming rods. With the concomitant use of Gram stain, pigmentation, catalase testing, and aerobic growth, the ATB 32A system correctly identified 97% of the B. fragilis group isolates, 88% of Bacteroides spp., 50% of fusobacteria, 74% of clostridia, 100% of cocci, and 86% of the gram-positive, nonsporeforming rods. Overall, 188 strains (88%) were correctly identified, with 18 (8%) requiring extra tests, other than the four mentioned above, for correct identification. Eight strains were misidentified, including one Bacteroides sp., three fusobacteria, one Clostridium sp., and three gram-positive, nonsporeforming rods. Reproducibility was very good, with 12 of 14 strains (86%) tested in triplicate yielding identical correct results on each of three occasions and 2 strains (14%) yielding identical correct results on two occasions. There was a low-probability identification (including the correct species) on the third testing. The ATB 32A system represents a worthwhile advance in systems used for the identification of clinically significant anaerobic bacteria.