Direct comparison of two mechanized systems for identification of gram-negative bacilli. Autobac ID system versus the auto microbic system (with EBC plus).

The Auto Microbic System (AMS) is an almost completely automated system, capable of identifying Enterobacteriaceae after 8 hours, and some glucose nonfermenters after 13 hours of incubation. The Autobac ID system is a mechanized, computer-assisted system capable of identifying Enterobacteriaceae and many nonfermentative gram-negative bacilli within 3-6 hours. The present report combines the results of two independent studies, both of which evaluated the AMS and the Autobac ID system compared with standard reference tests. Among the 1,510 isolates that were tested, both systems reported equivocal identifications (low confidence values) with 5-6% of the strains. AMS produced fewer erroneous identifications (3.8% vs. 4.9%) but more equivocal test results (5.7% vs. 4.9%). Reproducibility of the two systems was compared by triplicate testing of 88 selected strains in both laboratories. AMS was somewhat more reproducible than the Autobac ID system. The AMS was capable of identifying more species with greater accuracy and reproducibility, but the Autobac ID system was more rapid. Both systems demonstrated excellent accuracy and reproducibility and both could be used efficiently in the clinical laboratory.

Sensitivity, specificity, and reproducibility of the automicrobic system (with the Enterobacteriaceae-plus biochemical card) for identifying clinical isolates of Gram- negative bacilli.

Two independent laboratories tested 1,743 clinical isolates by using the Enterobacteriaceae-plus Biochemical Card in the AutoMicrobic system (AMS) and identical standard reference methods. Included were 55 isolates representing 11 species that cannot be identified by the enterobacteriaceae-plus Biochemical Card computer program; 3 or these isolates were incorrectly identified as Pseudomonas cepacia. With the other 1,688 isolates, 5% of the AMS identifications were considered to be equivocal (probability value, less than 0.80), and the remaining test were 97% accurate (sensitive), Difficulty was observed in the ability of the AMS to identify some H2S-negative Citrobacter freundii species. An AMS response of P. cepacia was also considered nonspecific, because several other organisms were misidentified at P. cepacia. Reproducibility of the system was documented by testing 125 strains on 3-separate days; only 6 strains produced significantly variable results. The AMS (with the Enterobacteriaceae-plus Biochemical Card) was a very satisfactory, automated system for accurately identifying most gram-negative bacilli within 8 to 13 h.