INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Three instrument-based continuous monitoring blood culture systems have been introduced for use in the United States: the BacT/Alert system (Organon Teknika, Durham, N.C.), the BacTec 9240 System (Becton Dickinson Microbiology Systems, Cockeysville, Md.), and the ESP System (Accumed Diagnostics, Cleveland, Ohio). Detection of bacteremia and fungemia by use of a continuous-monitoring blood culture device is commonly exploited in clinical microbiology laboratories. Each of the three systems noted above has been evaluated extensively, often in controlled clinical trials, and the results have been published in the literature (2, 5, 7-9, 11-17, 19, 20-25). In general, it can be said that continuous-monitoring blood culture systems afford more rapid detection of bacteremia and possibly fungemia than is possible with non-instrument-based manual methods (2, 8, 9, 11, 12, 15, 17, 19). In addition, differences in detection sensitivity have been noted when these systems have been compared with each other (5, 7, 13, 16, 21, 22, 25).

Recently, a new medium, FAN, has been introduced in both aerobic and anaerobic formulations for use with the BacT/Alert system (4, 20, 24). FAN medium contains adsorbent material referred to as Ecosorb (i.e., Fuller's earth plus charcoal), which binds antimicrobial agents, thus facilitating detection of bacteremia and fungemia in patients receiving antimicrobial therapy at the time blood cultures are performed. FAN media and other resin-based antimicrobial-binding blood culture systems have previously been shown in several investigations to accomplish such enhanced detection (1, 3, 12, 15, 20). A second notable feature of FAN bottles is their ability to accommodate 10-ml volumes of blood despite containing only 40 ml of broth medium.

The intent of the current investigation was to compare rates of recovery and detection times in FAN aerobic medium processed with the BacT/Alert system with those obtained in a second, high-volume aerobic medium, 80A, processed with the ESP continuous-monitoring blood culture system. The 80A bottles contain 80 ml of broth medium and rely on dilution as a means of minimizing antibiotic suppression of blood culture growth.

(Preliminary results of this investigation were presented at the 97th Annual Meeting of the American Society for Microbiology held in New Orleans [4].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Specimen collection. This study was conducted between November 1995 and August 1996. An aliquot of ca. 20 ml of blood per culture was routinely collected by house officers, nurses, or phlebotomists following preparation of the venipuncture site with 70% isopropyl alcohol and 10% povidone iodine. By aseptic technique, equal aliquots of the blood specimen were inoculated immediately into both a BacT/Alert aerobic bottle containing 30 ml of FAN medium and an ESP 80A aerobic bottle containing 80 ml of tryptic soy-based medium. Both bottles are designed to accommodate at least 10 ml of blood. In selected cases, an additional 10-ml aliquot of blood was obtained and a third blood culture bottle, containing ESP 80N anaerobic medium, was inoculated. This investigation, however, was restricted to a comparison of FAN versus 80A. Blood culture bottles were transported to the laboratory within 1 h of specimen collection. Review of computer records of patient medication(s) was used to determine if patients were receiving antimicrobial therapy at the time blood cultures were drawn.

Processing in the laboratory. Upon receipt in the laboratory, between the hours of 7:00 a.m. and 12:30 a.m., blood culture bottles were processed immediately; between the hours of 12:30 a.m. and 7:00 a.m., bottles received in the laboratory were incubated at 35°C in ambient air, and these were batch processed between 7:00 a.m. and 7:30 a.m. To ensure that adequate volumes of blood had been cultured, the weights of all bottles were determined and compared with those of uninoculated bottles. Only culture sets with evidence of an inoculum of 6 to 12 ml of blood in both FAN and 80A were considered evaluable and were included in the analysis presented here.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

A total of 13,640 blood cultures were performed during the one-year period of this study. In 9,131 cases, both a FAN bottle and an 80A bottle were inoculated. In 6,305 instances, both FAN and 80A bottles were judged to have been inoculated with adequate volumes of blood (i.e., 6 to 12 ml) and thus were included in the analysis that follows. These 6,305 blood cultures were used to evaluate 3,002 presumed septic episodes in 2,612 patients.

A total of 433 of these 6,305 cultures (6.9%) were positive with an organism(s) judged to be of probable clinical significance (significant positive cultures). These 433 presumed significant positive blood cultures yielded a total of 490 bacteria or yeasts and had been obtained during the evaluation of 301 septic episodes in 235 patients. Two hundred seventy-five of the 433 positive blood cultures (63.5%), representing 195 septic episodes, yielding a total of 301 isolates, were positive in both FAN and 80A. One hundred nine of the significant positive blood cultures (25.2%) representing 70 septic episodes were positive only in the FAN bottle. One hundred twenty-six organisms were recovered from these cultures. In contrast, 49 significant positive blood cultures (11.3%), representing 36 septic episodes, yielded 63 isolates and were positive exclusively in the 80A bottle of a pair. The higher rates of significant positive cultures and numbers of septic episodes documented in FAN bottles versus 80A bottles were statistically significant (P < 0.05).

A breakdown of isolation rates of individual organisms judged to be clinically significant is provided in Table 1. Significantly enhanced recovery was noted with Staphylococcus aureus and Klebsiella spp. in FAN bottles. In all other cases, the differences between recovery rates in FAN bottles versus 80A bottles were not statistically significant. With certain organism groups, however, despite the fact that differences were not statistically significant, there was a clear trend towards enhanced rates of recovery in FAN bottles (e.g., Enterobacter spp. and miscellaneous gram-negative bacilli) and in 80A bottles (e.g., -hemolytic streptococci). When the total numbers of organisms recovered in these two bottles were compared, i.e., 427 in FAN versus 364 in 80A, the difference was highly statistically significant (P < 0.05).

A conspicuous difference between FAN bottles and 80A bottles is the presence in FAN of an antimicrobial-adsorbing material. In view of this difference, it was of interest to know if the enhanced rate of recovery noted in FAN was accounted for by patients receiving antimicrobial therapy at the time blood cultures were obtained. This appears to have been the case. The numbers of presumed clinically significant isolates in FAN and 80A, FAN only, and 80A only in patients not receiving antibiotic therapy were 158, 35, and 23, respectively (P = 0.15); in patients receiving therapy, these values were 143, 91, and 40, respectively (P < 0.05).

Another measure of detection sensitivity with blood culture systems is length of time to positivity. In Table 2, the average length of time to an instrument signal of positivity is listed for clinically significant isolates recovered in both FAN and 80A. The only significant differences between FAN and 80A with respect to detection times occurred with streptococci and enterococci, in which case 80A bottles became positive significantly faster than FAN bottles.

Three hundred thirty-eight of the 6,305 blood cultures included in this survey (i.e., 5.4%) were judged to be probably contaminated. Of these 338 blood cultures, in 90 cases (26.6%) both FAN and 80A bottles were contaminated, in 143 cases (42.3%) only the FAN bottle was contaminated, and in the remaining 105 cases (31.1%) only the 80A bottle was contaminated (P < 0.05). A listing of probable contaminants is presented in Table 3. The difference between overall recovery rates of probable contaminants in FAN bottles versus 80A bottles was largely accounted for by the large numbers of non-S. aureus staphylococci that were recovered only in FAN bottles. For the 90 probable contaminants recovered in both FAN and 80A bottles, the mean length of time to detection in FAN bottles was 28.7 h and in 80A bottles was 29.6 h. This difference was not statistically significant.

During the course of this study, 46 FAN bottles signaled positive with the BacT/Alert instrument yet failed to yield an organism on subculture. The Gram stain results for these 46 bottles were also negative. These FAN cultures were considered false-positives (i.e., false-positive rate with FAN bottles, 0.7%). Conversely, 191 false-positive 80A cultures (3.0%) were identified. The difference between the false-positive rates with the two blood culture systems was highly significant (P < 0.05).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

It is clear from this and other studies that no single blood culture system optimizes recovery of all organism groups. In the current study, the BacT/Alert continuous-monitoring blood culture system with FAN aerobic medium was clearly superior to the ESP system with standard 80A aerobic medium as a means for detecting bacteremia in a tertiary-care referral hospital. Significantly greater numbers of positive blood cultures with S. aureus and Klebsiella sp. bacilli were obtained with this system. Conspicuous trends towards higher rates of recovery in FAN (albeit lacking statistical significance) were noted with Enterobacter spp. and miscellaneous gram-negative bacilli. In contrast, enhanced detection of -hemolytic streptococci was noted with the ESP system using 80A bottles.

The incremental increases in blood culture recovery rates noted with BacT/Alert were largely accounted for by patients receiving antimicrobial therapy at the time blood cultures were obtained. This is not surprising insofar as the FAN aerobic medium employed with the BacT/Alert system contains an antibiotic-neutralizing material that is not present in the ESP 80A blood culture bottles. These observations are similar to those of Welby-Sellenriek et al., who demonstrated similarly enhanced rates of detection of bacteremia in FAN aerobic bottles in comparison to ESP 80A bottles in a pediatric patient population (21).

Some controversy has existed as to the overall clinical value of using a blood culture system that attempts to diminish antibiotic effect as a means of facilitating detection of bacteremia (6). There is no doubt, however, that higher rates of bacteremia detection can be achieved by either pretreatment of blood specimens with a resin device or use of blood culture systems which employ bottles that contain either antimicrobial-binding resins or an absorbent material such as the Ecosorb that is present in aerobic FAN bottles (1, 3, 12, 15, 20). Two previous investigations that are particularly relevant to the current study compared the rates of detection of bacteremia and fungemia in FAN aerobic and anaerobic medium to those in standard O/T aerobic and anaerobic medium (20, 24). In both studies significantly enhanced detection rates were obtained by use of FAN media.

Length of time to detection is another important measure of the relative sensitivity of continuous-monitoring blood culture systems. Interestingly, in the current study, despite generally higher detection rates in FAN versus 80A, roughly comparable detection times were noted with the two systems. The only exception was significantly shorter times to detection of streptococci and enterococci observed in 80A bottles.

Overall rates of contamination in the current study were noted to be higher in FAN than in 80A. Non-S. aureus staphylococci, recovered exclusively in FAN, accounted for most of the difference in contamination rates seen between the two systems. It is possible that the use of what appears to be a more sensitive system for detecting clinically significant bacteremia, such as FAN, may also result in recovery of more contaminants.

Finally, one important consideration for laboratories which utilize continuous-monitoring blood culture systems is the frequency with which instruments generate false-positive signals. False-positive signals lead to needless expenditures of time, effort, and money in working up negative blood culture bottles. In the current study, the false-positive rates with the two systems that we evaluated were 0.7% (BacT/Alert-FAN) and 3.0% (ESP-80A). This difference in rates was statistically significant.

We conclude from the results of this study that the use of FAN aerobic bottles in conjunction with the BacT/Alert blood culture system generally provides higher recovery rates in a tertiary-care referral hospital laboratory than does the use of 80A bottles processed with the ESP system. In addition, fewer false-positive results are obtained. Use of the BacT/Alert system with FAN aerobic bottles is, however, associated with higher rates of contamination.