INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

VITAL (bioMérieux Vitek, Hazelwood, Mo.) is a continuous-monitoring blood culture system that is similar in design and function to the BacT/Alert (Organon Teknika Corporation, Durham, N.C.), BACTEC 9000 (Becton Dickinson Microbiology Systems, Sparks, Md.), and ESP (Accumed International, Inc., Detroit, Mich.) systems (9, 17). The primary difference between VITAL and other systems is the mechanism of detection. VITAL blood culture bottles contain a broth medium containing a soluble fluorescent detector molecule that decreases in fluorescence as the pH and redox potential change and as electrons are produced.

We report here the results of a controlled clinical comparison of the VITAL and BACTEC NR-660 blood culture systems for the diagnosis of bacteremia and fungemia in adult patients (13). Results of a comparison of the two systems for the diagnosis of bacteremia and fungemia in pediatric patients have been published elsewhere (20). VITAL was compared with the BACTEC NR-660 rather than the current BACTEC 9000 series to provide the same type of comparison that has been published for the other continuous-monitoring blood culture systems (2, 12, 17).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Background information. The study sites were Duke University Medical Center (DUMC), Denver Health Medical Center (DHMC), and Robert Wood Johnson University Hospital (RWJUH). Blood cultures were drawn from adult patients per physician order as part of routine patient care.

Blood culture collection. Venipuncture sites were disinfected with 10% povidone-iodine followed by 70% isopropyl alcohol. Twenty milliliters of blood was withdrawn with sterile needles and syringes or, if necessary, with a "butterfly" apparatus. Once drawn, 5 ml of blood was inoculated immediately into each of the four blood culture bottles; when insufficient blood was drawn to adequately fill all four bottles, BACTEC bottles were filled first to ensure that patients received a blood culture by a standard method. Inoculated bottles were then labeled and transported to the microbiology laboratories.

Laboratory processing. Cultures were accessioned according to each laboratory's routine. Each bottle was compared with prefilled volume standards, and volumes were recorded. Bottles were categorized as underfilled (<4 ml), adequately filled (4 to 6 ml), or overfilled (>6 ml). All bottles containing any volume of blood were processed for purposes of patient care, but only adequately filled bottles (i.e., bottles containing 4 to 6 ml of blood) were included in the study. Upon receipt, bottles were examined visually for macroscopic evidence of growth; those with visible evidence of growth were processed immediately as suspected positive bottles without being placed in the instruments. Bottles without visible evidence of growth were placed in the instruments. Per each manufacturer's instructions, bottles were not vented to room air.

Assessment of clinical importance of isolates. Positive blood cultures were categorized as true positives, contaminants, or of unknown clinical significance based on the published criteria of Weinstein et al. (16). In brief, judgments were made by infectious disease clinicians, taking into account the patient's clinical history, physical findings, laboratory test results (including the results of cultures of specimens from other sites), imaging studies, and clinical course and response to therapy.

Data analysis. Information recorded and collated included (i) which bottle types were received with each blood culture set, (ii) adequacy of bottle fill, (iii) which bottle(s) in each set grew microorganisms, (iv) the means by which bottles were found to be positive (i.e., by visual examination, instrument flagging, or terminal subculture, (v) the time required for microbial growth to be detected, (vi) the identity and clinical importance of microbial isolates, (vii) what antimicrobial therapy (if any) patients were receiving when blood cultures were drawn, (viii) a record of false-positive bottles, (ix) a record of contamination rates for the two systems, (x) the number of septic episodes detected by each bottle, and (xi) whether each septic episode was unimicrobial or polymicrobial. Septic episodes were defined as detection of a clinically important blood culture isolate (i.e., the cause of sepsis) without recovery of a different microorganism during the succeeding 7-day period. If a different microorganism was recovered within 72 h, the two isolates were considered to be part of a polymicrobic septic episode. If a different microorganism was recovered after 72 h, the two isolates were considered to be the causes of two septic episodes. Data were entered into a database program (Paradox; Corel, Farmingdale, N.Y.); statistical analysis was by the chi-square test modified by using Yates' correction when n was 20 (10).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Overall, 9,446 blood specimens were submitted during the study. A total of 8,943 specimens with both aerobic bottles filled adequately were collected, including 5,669 at DUMC, 1,793 at DHMC, and 1,481 at RWJUH. Of these, 885 yielded 1,016 microorganisms, including 622 isolates (61%) considered to be the cause of sepsis, 337 isolates (33%) that were contaminants, and 57 isolates (6%) that were indeterminate as the cause of sepsis. The comparative yields for the two aerobic bottles are shown in Table 1. Significantly more coagulase-negative staphylococci (P < 0.005), viridans streptococci (P < 0.01), members of the family Enterobacteriaceae (P < 0.02), Candida albicans isolates (P < 0.025), Candida tropicalis isolates (P < 0.02), Torulopsis glabrata isolates (P < 0.005), and all microorganisms combined (P < 0.001) were recovered from NR6 bottles. Significantly more Staphylococcus aureus isolates (P < 0.001) were recovered from VITAL aerobic bottles.

A total of 8,647 blood cultures with both anaerobic bottles filled adequately were collected, including 5,386 at DUMC, 1,794 at DHMC, and 1,467 at RWJUH. Of these, 655 yielded 740 microorganisms, including 486 isolates (66%) considered to be the cause of sepsis, 215 isolates (29%) that were contaminants, and 39 isolates (5%) that were indeterminate as the cause of sepsis. The comparative yields for the two anaerobic bottles are shown in Table 2. Significantly more S. aureus isolates (P < 0.02), coagulase-negative staphylococci (P < 0.05), Acinetobacter baumanii isolates (P < 0.025), Pseudomonas aeruginosa isolates (P < 0.001), C. albicans isolates (P < 0.02), and all microorganisms combined (P < 0.001) were recovered from VITAL anaerobic bottles.

Average times to detection of microbial growth were calculated when microorganisms were detected during the first 72 h of incubation and testing (Table 3). For aerobic bottles, VITAL detected microbial growth in a mean time of 19.8 h, versus 21.6 h for BACTEC; for anaerobic bottles, VITAL detected microbial growth in a mean time of 17.6 h, versus 30.9 h for BACTEC.

A system-versus-system comparison was made when all four bottles were received adequately filled (Table 4). Significantly more coagulase-negative staphylococci (P < 0.05), viridans streptococci (P < 0.025), C. albicans isolates (P < 0.05), C. tropicalis isolates (P < 0.02), T. glabrata isolates (P < 0.02), and all microorganisms combined (P < 0.001) were recovered from the BACTEC system, whereas significantly more S. aureus isolates (P < 0.005) were recovered from the VITAL system.

Of 304 septic episodes detected, 253 were unimicrobial and 51 were polymicrobial. Significantly more septic episodes caused by yeasts (P < 0.02) and all microorganisms combined (P < 0.005) were detected by the BACTEC system.

For the 337 contaminants recovered from aerobic bottles, significantly more coagulase-negative staphylococci (P < 0.005), all staphylococci combined (P < 0.001), other gram-positive cocci (P < 0.05), and all microorganisms combined (P < 0.001) were recovered from NR6 bottles. There were no significant differences in recovery for the 215 contaminants recovered from the two anaerobic bottles.

For NR6 bottles, 398 false-positive instrument readings occurred, as opposed to 216 false-positive readings for VITAL aerobic bottles (P < 0.001). A total of 231 false-positive instrument readings were generated for NR7A bottles, versus 240 for VITAL anaerobic bottles (not significant).

Terminal subcultures of negative companion bottles (blood culture sets where at least one bottle was positive, regardless of adequacy of fill) yielded 17 isolates from NR6 bottles versus 41 from VITAL aerobic bottles (P < 0.005) and 57 from NR7A bottles versus 56 from VITAL anaerobic bottles (not significant). Microorganisms recovered by terminal subculture included 28 coagulase-negative staphylococci, 27 S. aureus isolates, 15 P. aeruginosa isolates, 12 C. albicans isolates, 10 T. glabrata isolates, 9 Propionibacterium acnes isolates, 8 C. tropicalis isolates, 6 A. baumanii isolates, 4 viridans streptococci, 3 diphtheroids, 2 corynebacteria, 2 Escherichia coli isolates, 1 Klebsiella ozaenae isolate, 1 Pantoea agglomerans isolate, 1 Serratia marcescens isolate, 1 Acinetobacter sp. isolate, 1 Candida parapsilosis isolate, 1 Enterococcus faecalis isolate, and 1 group D streptococcus. All 15 P. aeruginosa isolates were recovered only from NR7A bottles; none were recovered from VITAL anaerobic bottles by terminal subculture. For yeasts, 5 of 31 (16%) were recovered from NR7A bottles only by terminal subculture, whereas 22 of 31 (71%) were recovered from VITAL anaerobic bottles only by terminal subculture. Only five times did isolates recovered from the companion bottles by terminal subculture differ from those detected by the instruments. In three of these cases different contaminants were recovered from the two bottles, in one case P. aeruginosa was detected by the instrument and a contaminant was recovered by terminal subculture, and in one case a contaminant was detected by the instrument but T. glabrata was recovered by terminal subculture. In the last case, however, the patient had numerous other cultures in which T. glabrata grew and was detected by the instrument. Thus, in no case was terminal subculture the only means by which a pathogen was detected.

For patients receiving antimicrobial therapy at the time blood was collected for culture, significantly more S. aureus isolates (P < 0.01) were recovered from VITAL bottles, whereas more yeasts (P < 0.005) were recovered from BACTEC bottles. For patients not receiving antimicrobial therapy at the time blood was collected for culture, significantly more viridans streptococci (P < 0.025), E. coli isolates (P < 0.05), C. tropicalis isolates (P < 0.05), other fungi (P < 0.01), and all microorganisms combined (P < 0.001) were recovered from BACTEC bottles.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this controlled clinical comparison of the VITAL and BACTEC NR-660 blood culture systems, the BACTEC system was found to recover significantly more microorganisms and to detect significantly more septic episodes than did the VITAL system. More microorganisms were recovered from BACTEC aerobic bottles than from VITAL aerobic bottles. In contrast, more microorganisms were recovered from VITAL anaerobic bottles than from BACTEC anaerobic bottles. When isolates were detected by both systems, VITAL detected microbial growth earlier than did BACTEC and recovered fewer contaminants.

The reason(s) for superior microbial recovery of all microorganisms combined from BACTEC NR6 bottles is unknown. Both BACTEC NR6 and VITAL aerobic bottles contain supplemented soybean-casein digest broth medium, with similar concentrations of sodium polyanetholsulfonate (SPS) (0.03% for BACTEC and 0.025% for VITAL) as an anticoagulant. The agitation systems differ slightly, but this difference is unlikely to account for the observed difference in microbial recovery rates. Similarly, there was no pattern of microbial recovery (e.g., strict aerobes) that would account for differing abilities to recover specific microorganisms. S. aureus isolates were recovered more frequently from VITAL aerobic and anaerobic bottles, but clinically important isolates of coagulase-negative staphylococci had a greater yield in the BACTEC aerobic bottle than in the VITAL aerobic bottle. This pattern of microbial recovery is unusual, since previous clinical studies of BACTEC NR bottles have shown parallel recovery of S. aureus and coagulase-negative staphylococci within a given bottle (4, 5, 7, 14, 18). Data from this study do not provide an explanation for this pattern of microbial recovery.

The most likely explanation for superior recovery of yeasts (and possibly all microorganisms combined) from BACTEC NR6 bottles is that NR6 bottles contain higher oxygen concentrations than do VITAL aerobic bottles. This hypothesis is supported by the observation that significantly more P. aeruginosa isolates were recovered from NR6 bottles, and recovery of strict aerobes such as yeasts and pseudomonads is highly dependent on the concentration of oxygen present in the bottle (1, 3, 4, 6). Higher oxygen concentrations may also explain the recovery of significantly more yeasts and pseudomonads from VITAL anaerobic bottles than from NR7A bottles, an observation underscored by the recovery of 15 P. aeruginosa isolates from NR7A bottles by terminal subculture only, while none of the same 15 isolates were recovered from VITAL anaerobic bottles by terminal subculture. It is notable that of 31 yeasts recovered only by terminal subculture, 16 were recovered from VITAL aerobic bottles and 22 were recovered from anaerobic bottles, compared to 1 isolate recovered from NR6 bottles and 4 isolates recovered from NR7A bottles. This observation may be explained by the presence of an intermediate redox potential in VITAL bottles, which makes for a slightly more anaerobic state in aerobic bottles and a slightly more aerobic state in anaerobic bottles (20). A similar phenomenon has been observed with other blood culture systems (15). When bottles processed on BACTEC NR instruments are tested, an aliquot of headspace atmosphere is removed for testing and an equal volume of gas is backflushed into the bottle headspace. By contrast, VITAL bottles are not sampled. Thus, headspace atmospheres in NR6 bottles may be relatively more aerobic and headspace atmospheres in NR7A bottles may be more anaerobic than VITAL aerobic and anaerobic bottles, respectively. A less anaerobic environment in VITAL anaerobic bottles would also explain why significantly more microorganisms were recovered than with the NR7, since a relatively more aerobic environment would favor growth of common microbial pathogens. The improved yield of VITAL anaerobic medium was due to better recovery of aerobic and facultatively anaerobic microorganisms. Insufficient numbers of anaerobic bacteria were recovered from either anaerobic bottle to permit statistical comparison of the two bottles for relative rates of recovery of these bacteria.

In previous comparisons, continuous-monitoring blood culture systems, with their superior detection technology, have been shown to detect microbial growth earlier than the BACTEC NR-660 and NR-730 systems (12, 17, 19, 20). Not surprisingly, this difference occurred again during this study. Even with the advantage of continuous-monitoring technology, however, VITAL did not detect growth of C. albicans isolates earlier in aerobic bottles, and the magnitude of the difference in speed of detection overall was not as great as that seen with other continuous-monitoring blood culture systems (12, 17, 19, 20). Lelievre et al. (8), in a comparison of VITAL with the BACTEC 9240 blood culture system, demonstrated superior speed of detection with BACTEC 9240 bottles. Although the observed difference in that study may be explained by the use of resin-containing media in BACTEC bottles, it raises the possibility that VITAL does not detect microbial growth as fast as other continuous-monitoring blood culture systems. The difference in time to detection for all microorganisms combined was less than 3 h, a difference that is unlikely to be clinically important.

Significantly more contaminants, particularly coagulase-negative staphylococci, were recovered from BACTEC aerobic bottles. This is most likely due to the BACTEC invasive detection mechanism and frequent handling of BACTEC nonradiometric bottles throughout incubation and testing. Zaidi et al. (20), in their comparison of VITAL aerobic bottles with BACTEC PEDS PLUS bottles, also found a lower contamination rate when VITAL aerobic bottles were compared with BACTEC NR PEDS PLUS bottles. Lelievre et al. (8) reported low contamination rates in their comparison of VITAL with BACTEC 9240.

A disproportionate number of isolates recovered from VITAL aerobic bottles (compared with BACTEC NR6 bottles) were recovered only by terminal subculture. Zaidi et al. observed a similar phenomenon in their comparison of VITAL aerobic bottles with BACTEC PEDS PLUS bottles (20). Such a discrepancy was not observed for the anaerobic bottles. The reason(s) for this phenomenon is unknown. Microorganisms were present in the bottles, as evidenced by their recovery, but either they failed to grow to detectable numbers or their growth went undetected. Neither explanation is satisfactory, however, since the majority of microorganisms, both clinically important as well as contaminant, grew and their growth was detected. Because some microorganisms recovered only by terminal subculture were clinically important, this phenomenon merits further investigation.

When patients were receiving antimicrobial therapy, the detection advantage of BACTEC was diminished. This observation is to be expected, as VITAL bottles contain more broth medium than do BACTEC NR bottles (40 versus 30 ml), thereby diluting to a greater degree any antimicrobial agents present in blood. Better recovery of fungi from BACTEC NR6 bottles for patients both receiving and not receiving antimicrobial therapy indicates that the superiority of BACTEC medium for fungi outweighed the effect of antimicrobial agents. The numbers of microorganisms recovered from patients on therapy were small, however, precluding direct comparisons for some microorganism groups.

The results of this evaluation and of those of Zaidi et al. (20) and Lelievre et al. (8) indicate that modifications to the VITAL system are necessary for it to be competitive with other commercially available continuous-monitoring blood culture systems. Suggested areas of research might include improving the recovery of yeasts via modification of the aerobic medium, determining the utility of lysing agents, determining the practicability of agents that bind antimicrobial agents, adjusting the redox potential in aerobic bottles, and modifying the computer algorithm to decrease the number of isolates recovered only by terminal subculture. The effect of such changes on overall recovery and speed of detection of microbial growth would require evaluation in a controlled clinical study.