Search

New blood culture media containing antibiotic-binding polymeric beads have been developed for the BacT/Alert (bioMérieux, Inc., Durham, NC) blood culture system. To assess the performance of these new media, we compared the new BacT/Alert aerobic medium (FA Plus) with resins to BacT/Alert FA medium with activated charcoal and the new BacT/Alert anaerobic medium (FN Plus) to BacT/Alert FN medium at 3 tertiary care medical centers. Study bottle pairs were inoculated with a target volume of 6 to 10 ml of blood from adults and incubated in the BacT/Alert 3D blood culture instrument. In the FA Plus versus FA comparison, there were 1,507 study pairs. Among 170 isolates causing true bloodstream infections (BSIs), significantly more Staphylococcus aureus (P<0.001) and total microorganisms (P<0.01) grew in the FA Plus bottle than in the FA bottle. Fewer coagulase-negative staphylococcal (CoNS) contaminants grew in the FA Plus bottle than in the FA bottle (10 versus 22; P=0.05). In addition, growth was detected earlier in the FA Plus bottle than in the FA bottle (P<0.001). In the FN Plus versus FN comparison, there were 2,386 study pairs. Among 201 isolates causing true BSIs, significantly more S. aureus (P<0.001), CoNS (P<0.005), and total microorganisms (P<0.001) grew in the FN Plus bottle than in the FN bottle. Also, significantly more CoNS contaminants grew in the FN Plus bottle than in the FN bottle (P<0.001). Overall, microorganisms were detected earlier in the FN Plus than in the FN bottle (P<0.001). Medical technologists at all 3 study sites preferred the new media for Gram stain interpretation. We conclude that the FA Plus and FN Plus media provide improved and earlier detection of microorganisms compared with the FA and FN media and are preferable for Gram stain interpretation as well.

To determine the optimal anaerobic companion bottle to pair with the BacT/ALERT (bioMerieux, Durham, N.C.) nonvented aerobic FA (FA) medium for recovery of pathogenic microorganisms from adult patients with bacteremia and fungemia, we compared the BacT/ALERT FN (FN) anaerobic bottle with the standard BacT/ALERT SN (SN) anaerobic bottle. Each bottle, FA, FN, and SN, was filled with 8 to 12 ml of blood. Of 11,498 blood culture sets received in the clinical microbiology laboratories at two university medical centers, 7,945 sets had all three bottles filled adequately and 8,569 had both anaerobic bottles filled adequately. Of 686 clinically important (based on previously published criteria) isolates detected in one or both adequately filled anaerobic bottles, more staphylococci (P < 0.001), including Staphylococcus aureus (P < 0.001); members of the family Enterobacteriaceae (P < 0.001); and all microorganisms combined (P < 0.001) were detected in FN bottles. In contrast, more Pseudomonas aeruginosa isolates (P < 0.01) and yeasts (P < 0.001) were detected in SN bottles. More Bacteroides fragilis group bacteremias were detected only in the FN (six) than in the SN (one) anaerobic bottle (P = not significant). Overall, the mean time to detection was shorter with FN (16.8 h) than with SN (18.2 h). This difference in time to detection was greatest for the B. fragilis group: FN, 28 h, versus SN, 60.0 h. Many of the facultative microorganisms recovered in either FN or SN were also found in the companion FA. When microorganisms found in the companion FA bottle were omitted from the analysis, significantly more staphylococci (P < 0.001), including S. aureus (P < 0.001), and Enterobacteriaceae (P < 0.005) still were detected in FN bottles, whereas there were no significant differences for P. aeruginosa and yeasts, which were found as expected in FA bottles. We conclude that the companion anaerobic FN bottle detects more microorganisms than does the anaerobic SN bottle when used in conjunction with the nonvented aerobic FA bottle in the BacT/ALERT blood culture system.

To determine the optimal anaerobic companion bottle to pair with BACTEC Plus Aerobic/F medium for recovery of pathogenic microorganisms from adult patients with bacteremia and fungemia, we compared Plus Anaerobic/F bottles with Standard Anaerobic/F bottles, each of which was filled with 4 to 6 ml of blood. The two bottles were paired with a Plus Aerobic/F bottle filled with 8 to 12 ml of blood. A total of 14,011 blood culture sets were obtained. Of these, 11,583 sets were received with all three bottles filled adequately and 12,257 were received with both anaerobic bottles filled adequately. Of 818 clinically important isolates detected in one or both adequately filled anaerobic bottles, significantly more staphylococci (P < 0.001), streptococci (P < 0.005), Escherichia coli isolates (P < 0.02), Klebsiella pneumoniae isolates (P < 0.005), and all microorganisms combined (P < 0.001) were detected in Plus Anaerobic/F bottles. In contrast, significantly more anaerobic gram-negative bacilli were detected in Standard Anaerobic/F bottles (P < 0.05). Of 397 unimicrobial episodes of septicemia, 354 were detected with both pairs, 30 were detected with Plus Aerobic/F-Plus Anaerobic/F pairs only, and 13 were detected with Plus Aerobic/F-Standard Anaerobic/F pairs only (P < 0.05). Significantly more episodes of bacteremia caused by members of the family Enterobacteriaceae (P < 0.05) and aerobic and facultative gram-positive bacteria (P < 0.025) were detected with Plus Anaerobic/F bottles only. In a paired-bottle analysis, 810 of 950 isolates were recovered from both pairs, 90 were recovered from Plus Aerobic/F-Plus Anaerobic/F pairs only, and 50 were recovered from Plus Aerobic/F-Standard Anaerobic/F pairs only (P < 0.001). Paired Plus Aerobic/F-Plus Anaerobic/F bottles yielded significantly more staphylococci (P < 0.001), streptococci (P < 0.05), and members of the family Enterobacteriaceae (P <0.001). We conclude that Plus Anaerobic/F bottles detect more microorganisms and episodes of bacteremia and fungemia than Standard Anaerobic/F bottles as companion bottles to Plus Aerobic/F bottles in the BACTEC 9240 blood culture system.

Iodophor and alcohol pledgets were compared with the Medi-Flex Prep Kit II for skin disinfection before venipuncture. Of 12,367 blood cultures collected, 6,362 were done with conventional pledgets and 6, 005 were done with Medi-Flex kits. Contamination occurred in 351 of 6,362 blood cultures (5.5%; range, 3.7 to 8.1%) with conventional pledgets versus 328 of 6,005 (5.5%; range, 3.5 to 7.5%) with Medi-Flex kits.

Two recent multicenter blood culture studies found that BacT/Alert FAN (FAN) bottles (Organon Teknika, Durham, N.C.) had increased yields in detecting bacteremia and fungemia compared with standard BacT/Alert (STD) bottles. Because the clinical importance of this increase in microbial recovery is unknown, we performed a retrospective analysis to determine the frequency with which FAN bottles were the sole means of detecting an episode of bacteremia. There were 1,047 positive blood cultures in which both study bottles were adequately filled and the organism isolated was judged to be the cause of sepsis: 240 (23%) were positive only in FAN bottles and 73 (7%) were positive only in STD bottles. Of a total of 664 episodes of bacteremia, 126 (19%) were identified only by FAN bottles and 43 (7%) were identified only by STD bottles (P < 0.0001). Episodes detected only by FAN bottles more often were recurrent events (23 of 126, or 18%) than episodes detected only by STD bottles (2 of 43, or 5%) (P < 0.05) and more commonly occurred in patients receiving theoretically effective antibiotic therapy (33 of 126 [26%] versus 4 of 43 [9%]) (P < 0.05). The medical records for patients with 127 of these episodes (92 FAN bottles only; 35 STD bottles only) were available for review. More than half of both FAN bottle-only (60 of 92, or 65%) and STD bottle-only (20 of 35, or 57%) episodes were judged to be clinically important. We conclude that FAN bottles improve the detection of bacteremia and that the majority of the additional episodes detected are clinically important. The benefits of the greater yield in specific patient populations must be balanced against the higher costs of FAN bottles.

FAN medium was formulated to improve microbial recovery, particularly for fastidious microorganisms and for microorganisms causing sepsis in patients receiving antimicrobial therapy. In a controlled clinical evaluation performed at four university-affiliated hospitals, FAN anaerobic bottles were compared with standard anaerobic bottles for yield, speed of detection of microbial growth, and detection of septic episodes. A total of 10,431 blood culture sets were received; both anaerobic bottles of 7,694 blood culture sets were adequately filled with blood. Altogether, 925 isolates were recovered: 557 that were the cause of sepsis, 99 that were indeterminate as the cause of sepsis, and 269 contaminants. More Staphylococcus aureus (P < 0.001), coagulase-negative staphylococci (P < 0.001), Escherichia coli (P < 0.02), and all microorganisms combined (P < 0.005) were recovered from FAN bottles; more nonfermentative gram-negative bacilli (P < 0.05), Torulopsis glabrata (P < 0.001), and other yeasts (P < 0.01) were recovered from standard bottles. Growth of S. aureus (P < 0.001), coagulase-negative staphylococci (P < 0.001), Enterococcus faecalis (P < 0.025), streptococci other than Streptococcus pneumoniae (P < 0.01), and all microorganisms combined (P < 0.001) was detected earlier in standard bottles; growth of more isolates of E. coli (P < 0.05) and anaerobic bacteria (P < 0.01) was detected earlier in FAN bottles. The mean times to detection were 14.2 and 16.1 h for standard and FAN bottles, respectively. More septic episodes caused by S. aureus (P < 0.001), coagulase-negative staphylococci (P < 0.005), members of the family Enterobacteriaceae (P < 0.02), and all microorganisms combined (P < 0.02) were detected in FAN bottles; more septic episodes caused by nonfermentative gram-negative bacilli (P < 0.025) and yeasts (P < 0.005) were detected in standard bottles. In summary, more isolates (except for strict aerobes) were recovered from FAN bottles than from standard anaerobic bottles. Similarly, significant more septic episodes (except for those caused by strict aerobes) were detected with FAN bottles than with standard anaerobic bottles. With the exception of E. coli and anaerobic bacteria, growth of more isolates was detected earlier in standard anaerobic bottles.

Bottles developed for use in the BacT/Alert automated blood culture system (Organon Teknika Corp., Durham, N.C.) can accept up to 10 ml of blood without falling below a 1:5 ratio of blood to broth. We compared the yield and speed of detection of microorganisms in 13,128 adequately filled, paired, aerobic bottles inoculated with 5 versus 10 ml of blood at three university hospitals. A total of 798 microorganisms causing sepsis grew in one or both bottles. The overall recovery of microorganisms from 10-ml samples exceeded that from 5-ml samples (P < 0.001); the increased yield attributed to the additional 5 ml in the samples was 7.2%. The increased yield from 10-ml inocula was most marked for Escherichia coli (P < 0.01) and other members of the family Enterobacteriaceae (P < 0.001). Ten-milliliter samples did not yield more gram-positive bacteria, nonfermentative gram-negative rods, or yeasts. When both bottles were positive, the bottles inoculated with 10 ml of blood showed growth sooner (P < 0.001). Earlier detection with 10-ml inocula was especially notable for coagulase-negative staphylococci (P < 0.001), streptococci (P < 0.001), E. coli (P < 0.025), and other members of the family Enterobacteriaceae (P < 0.025). We conclude that an increase in the volume of blood inoculated into BacT/Alert aerobic blood culture bottles from 5 to 10 ml will increase the overall yield and the speed of detection of clinically important blood pathogens.

Recently, we published a comparison of the BacT/Alert blood culture system with the BACTEC 660/730 nonradiometric blood culture system using blood inocula of 5 ml per bottle. By reanalyzing data collected during that study, we found that, for true-positive isolates causing bacteremia or fungemia, 363 (97.6%) of 376 and 341 (97.7%) of 349 isolates were recovered by the end of day 5 of testing, and 364 (97.9%) of 376 and 343 (98.3%) of 349 isolates were recovered by the end of day 6 of testing for aerobic and anaerobic bottles, respectively. Most isolates recovered on days 6 (24 of 27) and 7 (20 of 25) of testing were either contaminants or indeterminate as a cause of sepsis. When used as recommended by the manufacturer, only six (1.3%) of 464 clinically important isolates recovered on test days 6-7 would have gone undetected had testing been limited to 5 days and four (0.9%) of 464 had testing been limited to 6 days. We conclude that BacT/Alert bottles can be tested for as few as 5 days and then discarded with minimal loss of true-positive isolates and maximal reduction of contaminants.

During a 5-month period, we evaluated the sensitivity, specificity, and predictive value of qualitative cultures of intravenous catheters submitted to a university hospital microbiology laboratory. Of 36 catheters submitted for culture from nonseptic patients, 10 (28%) grew one or more microorganisms on qualitative culture. Of 44 catheters cultured from septic patients, 20 (45%) grew one or more microorganisms, but only 5 grew microorganisms that also were isolated from blood cultures. The sensitivity, specificity, and positive predictive values for catheter-related bacteremia were 71%, 96%, and 17%, values similar to those reported for quantitative methods. Although acceptable in terms of sensitivity and specificity, the very low positive predictive value suggests that catheter-related bacteremia cannot be predicted reliably by this test. We conclude, therefore, that nonstandardized qualitative IV catheter cultures have minimal value as predictors of bacteremia and recommend that they not be performed in clinical microbiology laboratories.

The one-bottle Oxoid Signal blood culture system altered to provide a more aerobic bottle headspace was reassessed in a comparative study versus the two-bottle BACTEC radiometric system in 5,426 blood cultures. The BACTEC system detected more microorganisms (P less than 0.02), particularly anaerobes (P less than 0.05) and fungi (P less than 0.05).

We compared BACTEC radiometric blood culture media with (8B) and without (6B) 10% sucrose for the detection of bacteremia and fungemia in pediatric patients at four university teaching hospitals that used identical methods for obtaining and processing specimens. Overall, the yields of microorganisms from 5,714 blood culture sets were no different in the two media, although a trend was noted favoring 6B for the detection of pneumococci. Speed of detection of positive results was faster in the 6B than in the 8B medium (P less than 0.05), largely due to the faster detection of Staphylococcus aureus in the 6B medium. We conclude that, overall, with pediatric patients the hypertonic 8B radiometric medium has no advantage and that it possibly has a modest disadvantage, compared with isotonic 6B radiometric medium.

To evaluate the role of atmosphere of incubation in the detection of clinically important bacteremia and fungemia in adults, we compared the yield of microorganisms from 10,541 paired 5-ml samples of blood incubated aerobically and anaerobically. The medium, supplemented peptone broth (SPB) with 0.03% sodium polyanetholesulfonate, and the ratio of blood to broth (1:10) were the same for all cultures. Only cultures with adequate blood samples (greater than or equal to 80% of stated volume) were compared statistically. More fungi (P less than 10(-7) ) grew in continuously vented bottles of SPB. Aerobic incubation also favored (P less than 0.01) isolation of Neisseria gonorrhoeae and Eubacterium; more than 80% of these bacterial organisms were detected only in vented bottles. Anaerobic incubation (plugged venting units) did not significantly favor the isolation of any genus of microorganisms, although an estimated 11% more Bacteroidaceae grew in the unvented bottle of SPB. By comparison of our data with published results for other media, we conclude that the need for both aerobic and anaerobic incubation of blood cultures is dependent upon the medium used and the microorganisms likely to be encountered. Vented incubation of blood cultured in SPB is crucial for detection of fungi and some bacteria. Routine use of an unvented bottle of SPB may not be worthwhile for patients in whom Bacteroidaceae cause bacteremia infrequently. However, when Bacteroidaceae are suspected as the cause of sepsis, use of an unvented bottle of SPB is prudent.

The Oxoid Signal (Oxoid U.S.A., Inc., Columbia, Md.) blood culture system is a newly described, innovative method for visually detecting growth of microorganisms (D. Sawney, S. Hinder, D. Swaine, and E.Y. Bridson, J. Clin. Pathol. 39:1259-1263, 1986). We did 5,999 paired comparisons of equal volumes (10 ml) of blood in the Oxoid Signal and BACTEC (Johnson Laboratories, Towson, Md.) radiometric blood culture systems at two university hospitals that use identical methods of obtaining and processing specimens. Overall, more microorganisms were detected in the BACTEC system (P less than 0.001), in particular, streptococci (P less than 0.01), fungi (P less than 0.001), and nonfermentative gram-negative rods, especially Acinetobacter species (P less than 0.001). Trends favoring the BACTEC system for detection of Pseudomonas aeruginosa, Haemophilus species, and Neisseria species were noted. There were no differences in the yield of staphylococci, members of the family Enterobacteriaceae, and anaerobic bacteria. When both systems detected sepsis, the BACTEC did so earlier (P less than 0.001). This advantage was most notable at 24 h (70% of BACTEC positives detected versus 48% of Oxoid positives). The proportion of positives detected after 48 h, however, was similar (BACTEC, 84%; Oxoid, 78%). Revisions in the Oxoid Signal system itself or in the processing of Oxoid bottles appear to be necessary to improve its performance in detecting certain microorganism groups, especially fungi.

The in vitro activity of lomefloxacin, a new fluorinated quinolone antimicrobial, was compared to that of enoxacin, norfloxacin, ofloxacin, imipenem, cefotaxime, ceftazidime, and tobramycin against 597 microorganisms isolated from bacteremic patients at a university hospital. Overall, lomefloxacin had activity similar to that of enoxacin and norfloxacin but less than that of ofloxacin. Lomefloxacin had excellent activity against members of the family Enterobacteriaceae and Haemophilus influenzae, and good activity against staphylococci, including oxacillin-resistant strains, as well as many strains of Pseudomonas aeruginosa, P. maltophilia, and Acinetobacter calcoaceticus anitratus ssp anitratus.

To evaluate the role of the volume of blood cultured in the detection of clinically important bacteremia and fungemia in adults, we evaluated the yield and speed of detection of microorganisms from 5,317 paired 2- and 5-ml samples of blood. The same kind of medium (supplemented peptone broth with 0.03% sodium polyanetholsulfonate) and atmosphere of incubation (open venting units) were used for all blood cultures. Only adequately filled (less than or equal to 80% of stated volume) sets (20-ml tube and 50-ml bottle) were compared statistically. Significantly more bacteria (p less than 0.01), Pseudomonas spp. In particular (P less than 0.05), were isolated from the 5-ml samples of blood. We conclude that the volume of blood cultured is a critical factor in the detection of septicemia. Consequently, valid evaluation of other factors influencing the detection of septicemia must be based on comparisons in which equal volumes of blood are cultured.

Both Neisseria meningitidis and Haemophilus influenzae are important isolates recovered in blood cultures from septicemic children. Sodium polyanetholsulfonate is present in most blood culture media and can inhibit the growth of certain bacteria, including N. meningitidis. The addition of gelatin to blood culture media neutralizes this inhibition. The growth of H. influenzae is enhanced by specific growth factors such as hemin and NAD. The addition of gelatin and V-factor-analog (a proprietary supplement for enhancing the growth of H. influenzae) might have a positive effect on the yield and on the speed of detection of septicemia in children. To evaluate this possibility, we did 4,565 paired comparisons of blood cultured in BACTEC 6B (aerobic) medium with and without the addition of both 1.2% gelatin and V-factor-analog. More aerobic and facultative bacteria grew in the 6B than in the 6B-gelatin-V-factor-analog medium (P less than 0.01). Only seven isolates of Neisseria spp. were recovered during this study period, with the 6B medium performing as well as the supplemented medium. When microorganisms grew in both bottles, they did so at the same time except for H. influenzae and Candida albicans. H. influenzae was recovered earlier from the 6B-gelatin-V-factor-analog bottle (P less than 0.01), with a mean time to detection of 8.5 h compared with 15.9 h for the 6B bottle. C. albicans was recovered earlier from the 6B bottle (P less than 0.02), with a mean time to detection of 34.9 h compared with 71.6 h for the 6B-gelatin-V-factor-analog bottle. We conclude that the 6B medium in its present formulation is superior to bB supplemented with gelatin and V-factor-analog.

A commercially available biphasic blood culture system that utilizes an attachable agar slide paddle and Trypticase soy broth (BBL Microbiology Systems, Cockeysville, Md.) was compared with a conventional Trypticase soy broth blood culture bottle in 6,867 paired blood cultures from adult patients. Both systems were inoculated with equal volumes of blood (5 ml) and incubated aerobically (vented) for 2 weeks. More clinically important bacteria and fungi, including Staphylococcus epidermidis, streptococci, Escherichia coli, Klebsiella spp., other Enterobacteriaceae, and Pseudomonas aeruginosa, were recovered from the biphasic system (P less than 0.001). In contrast, more anaerobic bacteria of importance were recovered in the conventional bottle (P less than 0.01). Staphylococci (P less than 0.001), gram-negative facultative and aerobic bacteria (P less than 0.001), and fungi (P less than 0.001) were detected 1 or more days earlier in the biphasic system, whereas pneumococci (P less than 0.05) were detected earlier in the conventional bottle. Of 603 clinically important microorganisms that grew in the biphasic system, 601 (99.7%) were detected by day 7 of incubation, but only 403 of 490 microorganisms (82.2%) were detected by day 7 in the conventional bottle. Overall, the biphasic system was superior to the conventional bottle. For optimal detection of anaerobic bacteremia, however, the biphasic system should be used in conjunction with a complementary anaerobic conventional bottle.

Five hundred episodes of septicemia were reviewed, with emphasis on laboratory and epidemiologic findings. The isolation of facultative and anaerobic gram-negative bacilli, fungi, and gram-positive cocci (except viridans streptococci and Staphylococcus epidermidis) almost always indicated true bacteremia, whereas the isolation of aerobic and anaerobic gram-positive bacilli, including Clostridium species, often represented contamination. More than 99% of all episodes were detected when two samples of blood (a total of 30 ml) were cultured. The five most common isolates were Escherichia coli, Staphylococcus aureus, Streptococcus pneumoniae, Klebsiella pneumoniae, and Pseudomonas aeruginosa. The incidence of septicemia was highest among medical patients and lowest among obstetric-gynecologic patients. Two-thirds of all episodes were nosocomial; S. aureus, enterococci, facultative gram-negative bacilli, and fungi were especially common nosocomial pathogens. The microorganisms isolated varied with the hospital service; polymicrobial episodes were especially common among surgical patients and transplant recipients. The most common sources of bacteremia were the respiratory, genitourinary, and gastrointestinal tracts; however, the source was unknown in nearly one-third of episodes. Microorganisms causing septicemia in neutropenic and nonneutropenic patients were not different; however, polymicrobial infections were more frequent in the presence of neutropenia. After antimicrobial susceptibility data became available, therapy was appropriate greater than 90% of the time.

In an initial evaluation, we found the Oxoid Signal blood culture system inferior to the BACTEC radiometric system for detection of some microorganisms causing septicemia (M. P. Weinstein, S. Mirrett, and L. B. Reller, J. Clin. Microbiol. 26:962-964, 1988). To determine whether modified processing of the Oxoid Signal blood culture system could improve its yield and speed of detecting positive cultures relative to the BACTEC radiometric system, we agitated all Oxoid bottles during the first 24 to 48 h of incubation and performed aerobic and anaerobic subcultures of all Oxoid bottles negative after 7 days of incubation. These modifications improved the overall performance of the Oxoid system, particularly with regard to the yield of streptococci, members of the family Enterobacteriaceae, and Haemophilus, Neisseria, and Acinetobacter spp. The speed of detecting positive cultures also was improved, especially within the first 24 h of incubation. However, the BACTEC system still detected more positive cultures (P less than 0.005), especially of obligate aerobes such as Pseudomonas aeruginosa (P less than 0.05) and yeasts (P less than 0.005). The BACTEC system also detected positive cultures earlier than the Oxoid system (e.g., at 24 h of incubation, 70.5% of BACTEC positive cultures detected versus 62.1% of Oxoid positive cultures detected). Further modifications of the Oxoid system which might include a revised medium, additional processing modifications, altered headspace atmosphere, or a complementary second broth medium should be considered, since the system is attractive in concept and is easy to use in the clinical laboratory.

A commercially available agar-slide blood culture bottle (Septi-Chek; Roche Diagnostics, Div. Hoffman-La Roche, Inc., Nutley, N.J.) was compared with the radiometric blood culture system (BACTEC; Johnston Laboratories, Inc., Towson, Md.) in 8,544 paired blood cultures from adult patients. The systems were inoculated with equal volumes (10 ml) of blood. Overall, there was no statistically significant difference between the two systems in the recovery of clinically important microorganisms, but significantly more members of the family Enterobacteriaceae other than Escherichia coli were detected by the agar-slide system (P less than 0.005). The agar-slide system detected more fungi, and the BACTEC detected more anaerobic bacteria; however, small numbers of recovered organisms precluded statistical significance. When microorganisms grew in both systems, their presence was detected one or more days earlier in the BACTEC (P less than 0.001). More contaminants grew in the agar-slide system (P less than 0.001). Both systems performed well, and either system should provide high yield and prompt detection of positive blood cultures in patients with bacteremia and fungemia if used in an optimal way as recommended by the respective manufacturers.

Although the addition of 1.2% gelatin to broth blood culture media containing sodium polyanetholesulfonate has been shown to enhance detection of certain bacteria, including Neisseria meningitidis, N. gonorrhoeae, Peptostreptococcus anaerobius, and Gardnerella vaginalis, the effect of such supplementation on the detection of other microorganisms causing bacteremia and fungemia is not known. Therefore, we studied BACTEC 6B medium with and without gelatin in 6,833 paired comparisons to examine the effects of supplementation on both the yield and the speed of detection of sepsis. More aerobic and facultative bacteria grew in the 6B than in the 6B-gelatin medium (P less than 0.001), especially staphylococci (P less than 0.01), Escherichia coli (P less than 0.01), other members of the family Enterobacteriaceae (P less than 0.05), and Acinetobacter spp. (P less than 0.05). When microorganisms grew in both bottles, they did so earlier in 6B than in 6B-gelatin (P less than 0.001). We conclude that the 6B medium in its present formulation is superior to 6B medium supplemented with 1.2% gelatin.