Background: Procedures for rapid identification and susceptibility testing by direct inoculation (DI) from positive blood culture bottles into an automated system have not been standardized. This study was purposed to evaluate DI from BACTEC 9240 blood culture system (BD, USA) into MicroScan (Dade Behring, USA) or Phoenix (BD, USA).
Methods: From May to June 2006, bacterial pellets from positive aerobic bottles showing gram-positive cocci (GPC) or gram-negative rods (GNR) of single morphology were directly inoculated to MicroScan PosCombo1A and NegCombo32 and to Phoenix PMIC/ID-107 and NMIC/ID-53. In addition, the automated instruments were also inoculated from subcultures (standard inoculations, SI). Species identification and susceptibilities were compared between DI and SI and between MicroScan and Phoenix.
Results: A total of 108, 104, and 78 specimens were tested with MicroScan, Phoenix, and both, respectively. When DI and SI were matched, 94.8% of GPC were correctly identified with MicroScan, compared to 80.7% with Phoenix, and 93.9% of GNR were correctly identified with MicroScan, compared to 95.7% with Phoenix. DI with MicroScan and Phoenix showed correct susceptibilities in 94.6% of 1,150 and 96.5% of 660 tests (with very major error [VME] of 1.1% and 1.1%), respectively, among GPC and in 94.4% of 942 and 96.3% of 781 tests (with VME of 0.6% and 0%), respectively, of GNR. Correlation of identification/susceptibilities between MicroScan and Phoenix using DI were 81.8%/98.0% for Staphylococcus aureus and 100.0%/95.6% for Escherichia coli.
Conclusions: DI warrants a reliable method for identification and susceptibility testing of both GPC and GNR in MicroScan, and those of only GNR in Phoenix.