Comparison of the automated Bactec NR-660 with a conventional blood culture system

Abstract

The Bactec NR-660 blood culturing system was compared with a conventional system using 347 consecutively obtained clinical blood samples tested simultaneously. Of 46 clinically relevant isolates, 12 were detected by the conventional system only, 5 by Bactec only and 29 by both methods (0.05 less than p less than 0.10). The difference could not be explained by the results of additional in vitro tests. Of 12 isolates considered contaminants, 7 were isolated in the conventional system only, 2 in the Bactec system only and 3 in both systems (0.05 less than p less than 0.10). The two systems were approximately equal in speed of detection. A high rate of false indications of growth in the Bactec system could be reduced by applying different CO2 cut-off values.