Abstract
The 10-ml Isolator system (E.I. du Pont de Nemours & Co., Inc., Wilmington, Del.) was compared with the BACTEC 16A-17A nonradiometric resin system (Johnston Laboratories, Inc., Towson, Md.) for isolation of organisms from 6,839 paired blood cultures. Equal volumes of blood (6 to 10 ml for each Isolator and 3 to 5 ml for each BACTEC bottle) were cultured in parallel in the two systems, and 600 isolates that were judged to be clinically significant by chart review were recovered during the study. The BACTEC resin system detected 510 (85%) and the Isolator system detected 435 (72%) of the clinically significant isolates (P less than 0.001). Of 45 polymicrobial blood cultures, the BACTEC system detected 32 (71%) and the Isolator system detected 21 (47%) (P less than 0.05). Of 253 gram-negative bacilli isolated during the study, 30% were detected only in the BACTEC system and 16% were detected only in the Isolator system (P less than 0.001), and of 56 nonfermentative or fastidious gram-negative bacilli detected, 46% were recovered only in the BACTEC system, while 14% were detected only in the Isolator system (P less than 0.001). Of 86 streptococci isolated during the study, 30% were detected only in the BACTEC system, and 4% were detected only in the Isolator system (P less than 0.001). Recoveries of anaerobic bacteria, staphylococci, and yeasts were equivalent in the two systems. Organisms judged to be contaminants were detected in approximately 1% of the cultures in each system. The results suggest that use of resin media renders the BACTEC nonradiometric system equivalent or superior to the Isolator system for detection of clinically significant organisms in blood cultures.