The recovery of pathogens and the speed of their detection were determined for our conventional blood culture system (an Isolator [Wampole] and a 100-ml Thiol bottle [Difco]) compared with automated ESP aerobic and anaerobic bottles (80 ml each; Difco). Each of the four culture devices was inoculated with approximately 10 ml of blood from symptomatic patients weighing more than 80 lb (ca. 36 kg). From 7,070 sets of cultures for 2,841 patients, 607 clinically significant isolates were recovered: 456 (75.1%) from the Isolator, 353 (58.2%) from Thiol, 377 (62.1%) from ESP aerobic bottles, and 346 (57.0%) from ESP anaerobic bottles. Of the 607 isolates, 149 (24.5%) were detected only with the conventional system (Isolator and/or Thiol), and 65 (10.7%) were detected only with the ESP two-bottle system (P < 0.001). Our conventional system allowed for detection of significantly more isolates of members of the family Enterobacteriaceae (P < 0.001), Staphylococcus aureus (P < 0.01), Staphylococcus spp. (coagulase-negative) (P < 0.01), and Enterococcus spp. (P < 0.05), and ESP facilitated detection of significantly more isolates of S. pneumoniae (P < 0.01). When all four devices in a culture set were positive for the same isolate, no microbial species or group was detected significantly earlier (.24 h) by either blood culture system. The Isolator contamination rate (4.8%) was .6 times the rate for any of the bottles. Of pathogens detected by the Isolator, 50% were recovered in counts of '1.0 CFU/ml and 18% were recovered only as a single colony. The ESP system offered an automated, less labor-intensive blood culture system for which routine subcultures were not required, but the important considerations of culturing large volumes of blood and of obtaining at least two sets from each patient in our population were reemphasized.

To document the incidence of low-level bacteremia in the patient population of this study, two blood culture sets were collected from symptomatic patients weighing more than 80 pounds. Each blood culture set consisted of a lysis-centrifugation tube and three bottles containing different culture broths, each inoculated with 10 mL blood. Pathogens from 63 (26.4%) and 48 (20.1%) of the 239 culture-positive patients were recovered from only one and two of the eight culture devices, respectively, representing low-level bacteremia. Isolates from another 60 (25.1%) of the 239 patients were recovered from all eight of the culture devices, representing high-level bacteremia. Whether patients had low-level or high-level bacteremia, there were mostly insignificant differences in the types of species recovered, in the percentages of patients for whom therapy was initiated or changed following the laboratory's reports, and in the clinical signs, symptoms, and characteristics of the patients. Clinically documented, low-level bacteremia is relatively common in this community hospital's patient population. Culturing of up to 80 mL of blood was required for detection of all pathogens from patients weighing more than 80 pounds.

No abstract