OBJECTIVE In order to reduce the turnaround time for laboratory diagnosis of bacteremia, the efficacy of identification and antimicrobial susceptibility testing using samples taken directly from positive BacT/ALERT standard aerobic and standard anaerobic blood culture bottles was evaluated. METHODS 160 positive blood culture bottles were examined and incubated at 35 degrees C in 5% carbon dioxide for 4-24 h, and an aliquot of the culture fluid was Gram stained. Samples containing Gram-negative bacilli were inoculated on VITEK 2 ID-GNB (identification-Gram-negative bacilli) and AST (antimicrobial susceptibility testing)-GN04 cards, and those containing Gram-positive cocci were inoculated on ID-GPC (identification-Gram-positive cocci) and AST-P526 cards. The same samples were also examined by the standard method, involving subculture from positive BacT/ALERT standard blood culture bottles. RESULTS Eighty seven of 97 Gram-negative bacilli (89.7%) and 21 of 63 Gram-positive cocci (33.3%) were correctly identified to the species level. For antimicrobial susceptibility testing, the direct method had an overall error rate of 5.4% for Gram-negative bacilli, with 0.9% very major, 0.9% major, and 3.6% minor discrepancies compared to the standard method. The overall error rate in antimicrobial susceptibility testing for the 13 Staphylococcus spp. was 10.3%, with 6.0% very major, 2.6% major, and 1.7% minor discrepancies. CONCLUSIONS These data suggest that VITEK 2 cards inoculated with samples taken directly from positive Bact/ALERT blood culture bottles would provide acceptable identification and antimicrobial susceptibility testing results for Gram-negative bacilli, but not for Gram-positive cocci. Compared to the standard method, the direct method would reduce turnaround time by at least 24 h.