Effects of Delayed-Entry Conditions on the Recovery and Detection of Microorganisms from BacT/ALERT and BACTEC Blood Culture Bottles
http://www.100md.com
微生物临床杂志 2006年第4期
Pinnacle Health Laboratories, Harrisburg
Geisinger Medical Center, Danville, Pennsylvania
ABSTRACT
Manufacturers generally recommend that blood culture bottles be loaded into instruments within a short time of collection. However, in our experience, delays often occur prior to loading the bottles. We examined the effect of holding bottles under various temperatures (T)—room temperature (RT), 4°C, 37°C, and RT for 2 h following incubation at 37°C (to simulate transit [TR])—and for various holding times of 4, 12, and 24 h. We utilized the BacT/ALERT system with FA and FN bottles and the BACTEC system with Plus (PL) and Lytic 10 (LY) bottles. Standardized inocula and 5 ml of blood were added to each bottle. Fifteen organisms were evaluated based upon expected performance: aerobic (FA and PL), anaerobic (FN and LY 10), and facultative (all bottles). Based upon expected performance, the FA and FN bottles recovered 458 of 468 organisms and 282 of 288 organisms, respectively, whereas the PL and LY bottles recovered 453 of 468 organisms and 257 of 288 organisms, respectively (P = <0.001, FN versus LY). There were 3, 11, 21, and 27 false-negative results for bottles held at 4°C, RT, 37°C, and TR, respectively. There were 4, 8, and 50 false-negative results for bottles held for 4, 12, and 24 h, respectively. Our results support holding these four bottle types at 4°C or at RT for up to 24 h and at 37°C for up to 12 h. We propose that manufacturers only need to make claims for "delayed entry" when these bottles are held for more than 24 h at 4°C or at RT or for more than 12 h at 37°C.
INTRODUCTION
For patients with suspected bacteremia or fungemia, the inoculation of blood culture bottles for the detection of bacteria and yeast often occurs at the patient bedside. Bottles are then transported to the laboratory and loaded into blood culture instruments or incubators. Continuous-monitoring blood culture instruments are widely used for this purpose in the United States. Ideally, the time in transit from the patient to the instrument is kept to a minimum. Although this may have been a routine occurrence at one time, the proliferation of satellite laboratories, core laboratories, and increased dependence on reference laboratories for the performance of routine microbiology testing often results in prolonged delays between the time of specimen collection and the loading of the blood culture bottles into the instruments.
There have been very few studies that have examined the effects of transport time on organism recovery from blood culture bottles. The purpose of the present study was to evaluate the ability of two different continuously monitoring blood culture instruments to detect organisms from seeded blood cultures that were stored under a variety of temperatures for various lengths of time prior to loading into the instruments.
MATERIALS AND METHODS
The present study was conducted in two microbiology laboratories, Geisinger Medical Center (GMC) and Pinnacle Health Laboratories (PHL), with a standard protocol in use at both institutions. The GMC laboratory utilized the BacT/ALERT blood culture system (FA and FN bottles; bioMerieux, Durham, NC), whereas PHL utilized the BACTEC blood culture system (Plus [PL] and Lytic [LY] bottles) (Becton Dickinson Microbiology Systems, Cockeysville, MD).
Blood used for seeding the bottles was obtained from patients donating blood for therapeutic phlebotomy. To ensure that there was no clotting of the blood from these patients, who are more prone to clot than normal patients (hemachromatosis), sodium polyanetholesulfonate was added to attain a final concentration of ca. 0.060%. A solution of SPS was prepared by adding 7,920 mg (7.92 g) of SPS (provided by John Walsh, bioMerieux, Inc.) to 1,200 ml of 0.85% sterile saline. The SPS solution was filter sterilized, and 40 ml was added to each sterile blood collection bag used for blood collection. Anticipating a minimum of 400 ml of blood per patient, 264 mg of SPS was needed per blood collection bag (400 ml of blood plus 40 ml of SPS solution) to attain an SPS concentration of ca. 0.060%. All blood collection bags with SPS were prepared in one batch at GMC and were used by both the GMC and the PHL microbiology laboratories. Blood was collected by using the standard protocols used by both institutions and stored at 4°C prior to use in testing. Blood was used 5 to 8 days after collection at GMC and 1 to 6 days after collection at PHL.
Fifteen recent clinical isolates of ten different microorganisms were selected for use to include some common clinical organisms, as well as some organisms known to be more difficult to recover ("challenge" organisms). Suspensions of microorganisms were prepared in sterile saline from 24- to 48-h cultures, diluted in saline, and then added to a sufficient volume of blood to inoculate all of the bottles. The intended concentration of organisms was 50 CFU/ml in the inoculated blood. A total of 5 ml of the inoculated blood was added to each blood culture bottle, with a final intended inoculum of 250 CFU/bottle. The choice of which types of bottles to inoculate with the 15 organisms was based upon expected performance. For the purpose of the present study, expected performance was defined as follows. Aerobic and facultative anaerobic organisms were expected to grow in the aerobic bottles, while anaerobic and facultative anaerobic organisms were expected to grow in the anaerobic bottles. The results were assessed based upon this definition of expected performance. Aerobic bottles only (FA and PL) were inoculated with seven organisms: Pseudomonas aeruginosa (two isolates), Neisseria meningitidis (two isolates), Acinetobacter baumanii (one isolate), Candida albicans (one isolate), and Cryptococcus neoformans (one isolate). Anaerobic bottles only (FN and LY) were inoculated with two organisms, Bacteroides fragilis and Fusobacterium nucleatum, and all bottles (FA, FN, PL, and LY) were inoculated with six organisms: Staphylococcus aureus (two isolates), Streptococcus pneumoniae (two isolates), and Escherichia coli (two isolates). Both GMC and PHL utilized the same organisms, but the inoculum preparation was done independently in each laboratory.
After inoculation, bottles were held at four different temperatures (refrigerated at 4°C, incubated at 37°C, ambient room temperature [RT] and, to simulate transport to the laboratory, incubated at 37°C for "X" hours and then held at ambient temperature for 2 h [TR]). For example, with the TR bottles, if a bottle was to be held for a total of 24 h, it was held for 22 h at 37°C and the last 2 h at ambient temperature. Ambient temperature in the two laboratories was between 23 and 26°C. Bottles were held for 4, 12, and 24 h for each of the storage temperatures. Bottles were inoculated in triplicate for each bottle-temperature combination. Thus, for an individual organism, there were four temperatures and three storage periods in triplicate for a total of 36 bottles. Three bottles were also inoculated as zero time controls for a total of thirty-nine bottles for each organism-bottle type combination used in the study.
All bottles that did not signal positive within the routine 5-day incubation period were subcultured with 1 or 2 drops from each bottle onto standard bacteriology media.
RESULTS
All test results were evaluated based upon expected performance for aerobic and anaerobic bottles as described in Materials and Methods. Aerobic bottles were inoculated with a total of 13 organisms, whereas anaerobic bottles were inoculated with a total of eight organisms. Thus, the maximum number of positive bottles was 468 for the aerobic FA and PL bottles and 288 for the anaerobic FN and LY bottles. Using these criteria, the FA and PL bottles recovered 458 of 468 organisms and 453 of 468 organisms, respectively, while the FN and LY bottles recovered 282 of 288 organisms and 257 of 288 organisms, respectively (P = <0.001, FN versus LY) (Table 1).
When results were analyzed based upon the temperature of incubation prior to loading into the blood culture instruments, the best recovery was obtained for bottles held at 4°C (Table 2), and the number of false-negative results increased as the temperature of preincubation increased. There were 3, 11, 21, and 27 false-negative results for bottles held at 4°C, RT, 37°C, and TR. In contrast, the time to detection (TTD) was inversely influenced by the holding temperature. The mean TTDs were 37.3, 33.3, 28.5, and 28.5 h at for bottles held at 4°C, RT, 37°C, and TR, respectively.
When the results were analyzed as a function of the holding time, false-negative results increased with the holding time. There were 4, 8, and 50 false-negative results for bottles held for 4, 12, and 24 h, respectively (Table 3). The TTD was inversely influenced by the holding time. The mean TTDs were 26.1, 30.2, and 39.7 h for bottles held for 4, 12, and 24 h prior to loading, respectively.
The interrelated effect of the holding time and holding temperature on the overall recovery is shown in Table 4. False-negative results do not significantly change for bottles held at 4°C for up to 24 h. At RT, the number of false-negative results increases from one bottle at 4 h to four and six bottles at 12 and 24 h, respectively. At 37°C, the number of false-negative results increased dramatically from 0 and 2 at 4 and 12 h, respectively, to 19 at 24 h.
The intended concentration of microorganisms was 50 CFU in the blood that was added to each of the blood culture bottles. Since 5 ml of blood was added to each bottle, the final intended concentration of microorganisms was 250 CFU/bottle. The actual mean concentrations of microorganisms in the blood were 64 CFU/ml (range, 30 to 166) at the PHL and 42 CFU/ml (range, 17 to 38) at the GMC.
The two isolates of N. meningitidis, the single isolate of B. fragilis, Pseudomonas aeruginosa isolate 1, and the single isolate of Candida albicans were detected in all of the bottles tested. There were false-negative results for all of the other isolates tested with differences between the four bottle types for which organisms produced the most false-negative results. Of the 10 false-negative FA bottles, 3 were caused by Streptococcus pneumoniae isolate 1 and 3 were caused by P. aeruginosa isolate 2. Of the 15 false-negative PL bottles, 6 were caused by E. coli isolate 1 and 6 were caused by E. coli isolate 2. Of the six total false-negative FN results, four were caused by the F. nucleatum isolate. Of the 31 total false-negative LY bottles, 13 were caused by S. pneumoniae isolate 1.
Overall, mean TTD was 31.1 h for BACTEC bottles and 32.8 h for BacT/ALERT bottles. However, the data suggest that the organism mix influences the relative performance of the BACTEC and BacT/ALERT bottles. BACTEC bottles were faster for Acinetobacter anitratus, B. fragilis, C. neoformans, and F. nucleatum, while BacT/ALERT bottles were faster for C. albicans, and trending toward significance for S. pneumoniae and N. meningitidis. Little difference was noted for S. aureus, E. coli, or P. aeruginosa.
Subcultures were performed after 5 days for all bottles that were not signaled positive by the instruments. The results were as follows: FA, 6 positive subcultures from 10 bottles; Plus, 15 positive subcultures from 15 bottles; FN, 0 positive subcultures from 6 bottles; and LY, 19 positive subcultures from 30 bottles. The organisms yielding negative subculture results were as follows: 3 S. pneumoniae and 1 A. anitratus from FA bottles; 4 F. nucleatum and 2 S. pneumoniae from FN bottles; and 11 S. pneumoniae from LY bottles.
DISCUSSION
Ideally blood culture bottles should be loaded into continuous-monitoring blood culture instruments as soon as possible after collection in order to minimize the time to detection of microorganisms. Intuitively, one might suspect that false-negative results could result if there was a delay in the loading process. At one time, this may have been more of a theoretical rather than a practical problem since blood culture bottles were usually delivered to the laboratory within a couple of hours after collection, but the proliferation of satellite laboratories, core laboratories, and increased dependence on reference laboratories for the performance of routine microbiology testing now often results in prolonged delays between the time of specimen collection and loading of the blood culture bottles into the instruments.
The actual time at which a reduction in sensitivity occurs because of delays in loading bottles is not clearly established. Although the term "delayed entry" has been frequently utilized when blood culture bottles are not loaded into the instruments in a timely, but not actually defined time period, we are aware of no specific agreed-upon definition of the term "delayed entry." Is delayed entry 30 min, 1 h, and 8 h or some greater time period Should the working definition of delayed entry incorporate a holding temperature We propose that the term delayed entry be specifically reserved for a time period greater than the time period for which there is an acceptable rate of recovery for a particular bottle type and blood culture instrument system. For example, based upon our results, a bottle that was in transit for 2 h prior to loading into an instrument would likely not require a delayed-entry claim.
The acceptable time period prior to which a delayed-entry claim would not be required is open to discussion. In the present study, there was a noticeable decrease in recovery between the bottles held for 4 and 12 h compared to the bottles held for 24 h. However, when the data are examined more carefully, it is apparent that the holding temperature is at least equal to if not more important than the length of time prior to loading of the bottles. Our results support holding BacT/ALERT FA and FN bottles and BACTEC Plus and Lytic bottles at 4°C or at RT for up to 24 h and at 37°C for up to 12 h. We propose that manufacturers make claims for "delayed entry" for these bottles when they are stored for greater than 24 h at 4°C or at RT or greater than 12 h at 37°C. Shorter periods of time at the respective temperatures would be deemed to be normal product use and not require a delayed-entry claim.
While our study was limited to holding bottles for 24 h, it is possible that bottles could be held longer, particularly at 4°C, prior to loading, with no significant decline in recovery. If additional studies support this use, this could be made as a delayed-entry claim. We acknowledge that recovery may be acceptable for some bottles under certain conditions for a time period greater than 24 h. Nonetheless, we would favor the need for a delayed-entry claim for periods over 24 h, if only to emphasize the importance of getting the bottles loaded into an instrument within 24 h. For example, even if recovery is acceptable at 36 h, this still results in a significant delay to detection and reporting of positive blood cultures, a practice which should be clinically discouraged.
We are aware of no other studies that have examined in detail the effect of storage at 4°C prior to loading in the blood culture instruments. Indeed, the overall greatest sensitivity for the temperatures evaluated in the present study was for the bottles stored at 4°C. This could be very useful information, even for laboratories who do not want to routinely transport blood cultures at this temperature (4°C). Many laboratories, including the Geisinger Medical Center, transport specimens from clinics and other hospitals using three coolers with different temperatures (ambient temperature, refrigerated temperature, and frozen). While we may opt to routinely transport blood culture bottles at ambient temperature, these results indicate that bottles inadvertently transported at 4°C can be accepted with no significant loss of sensitivity.
There were significant differences between the performance of the anaerobic bottles tested in the present study. This was most noticeable with facultative organisms where the growth of both aerobic and anaerobic organisms should be anticipated. We can offer no explanation for the superior performance of the FN bottles compared to the LY bottles.
There are few studies that have examined the effects of holding time and temperature prior to loading of blood culture bottles and overall recovery. The data generated from older instrumentation, older software versions, and different medium formulations or bottle types may or may not be applicable to what is currently in use.
Klaerner et al. reported that BacT/ALERT FAN aerobic bottles failed to detect 46.9% of P. aeruginosa isolates, as determined by terminal subcultures (3). Moreover, seeded FAN bottles held at 36°C for as few as 4 h were not able to detect P. aeruginosa strains in the BacT/ALERT instruments (3). However, a more recent study from the same laboratory reported that P. aeruginosa isolates could be recovered from seeded BacT/ALERT FA bottles when held at 36°C for up to 8 h and for 24 h at RT (5).
Two studies have examined delayed entry with the BACTEC 9240 system (1, 4). Chapin and Lauderdale reported that 9240 vials can be preincubated at 35°C for up to 24 h with no significant loss of detection, whereas bottles could be maintained at RT for at least 48 h with no significant loss of recovery (1). Lemming et al. reported that there was a significantly higher false-negative rate for BACTEC bottles preincubated at 35°C compared to bottles preincubated at RT (4). The time of preincubation varied in that latter study.
There are no consistent recommendations from experts concerning delayed entry of blood culture bottles. Cumitech 1B recommends that blood culture bottles be delivered to the laboratory as soon as possible to prevent any delays in detection (2). The Manual of Clinical Microbiology (8th ed.) recommends a storage time of <2 h unless manufacturers specify otherwise (6). The package insert for the BacT/ALERT blood culture bottles recommends that bottles be loaded immediately after inoculation; however, a recent memo to customers states that inoculated bottles may be held at 25°C for up to 24 h prior to loading (Randy Turner, bioMerieux, Inc., personal communication). BD Diagnostics states that BACTEC Aerobic/F Plus and LYTIC/10 Anaerobic bottles can be held for up to 20 h at incubator temperature (temperature not specified) or up to 48 h at RT (Mike Borlet, BD Diagnostics, personal communication).
In conclusion, we have demonstrated differences in organism recovery depending upon the temperature and time of storage prior to loading of blood culture bottles into continuous-monitoring blood culture instruments. Moreover, there is an important relationship between the storage time and storage temperature, such that they should be evaluated in concert. There were differences in bottle performance as well. Although the FA and PL bottles yielded comparable results, the yield with FN bottles was significantly better than the yield with LY bottles.
We acknowledge that our conclusions are based upon the data generated from 10 microorganisms with "seeded" blood cultures and controlled inocula. However, this may represent the limits of what can be practically achieved as evaluations utilizing patient specimens would be very difficult or impossible to perform.
ACKNOWLEDGMENTS
This study was supported in part by a grant from bioMerieux, Inc.
Present address: Carolinas Pathology Group, Carolina Medical Center, P.O. Box 34455, Charlotte, NC 28234-4455.
REFERENCES
Chapin, K., and T.-L. Lauderdale. 1996. Comparison of Bactec 9240 and Difco ESP blood culture systems for the detection of organisms from vials whose entry was delayed. J. Clin. Microbiol. 34:543-549.
Dunne, W. M., F. S. Nolte, and M. L. Wilson. 1997. Cumitech 1B-blood cultures III. American Society for Microbiology, Washington, D.C.
Klaerner, H.-G., U. Eschenbach, K. Kamereck, N. Lehn, H. Wagner, and T. Miethke. 2000. Failure of an automated blood culture system to detect nonfermentative gram-negative bacteria. J. Clin. Microbiol. 38:1036-1041.
Lemming, L., H. M. Holt, I. S. Petersen, C. Ostergaard, and B. Bruun. 2004. Bactec 9240 blood culture system: to preincubate at 35°C or not Clin. Microbiol. Infect. 10:1089-1091.
Seegmuller, U., U. Eschenbach, K. Kamereck, and T. Miethke. 2004. Sensitivity of the BacT/ALERT FA-medium for the detection of Pseudomonas aeruginosa in preincubated blood cultures and its temperature-dependence. J. Med. Microbiol. 53:869-874.
Thomson, R. B., and J. M. Miller. 2003. Specimen collection, transport, and processing: bacteriology, p. 286-330. In P. R. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology, 8th ed. American Society for Microbiology, Washington, D.C.(R. L. Sautter, A. R. Bill)
Geisinger Medical Center, Danville, Pennsylvania
ABSTRACT
Manufacturers generally recommend that blood culture bottles be loaded into instruments within a short time of collection. However, in our experience, delays often occur prior to loading the bottles. We examined the effect of holding bottles under various temperatures (T)—room temperature (RT), 4°C, 37°C, and RT for 2 h following incubation at 37°C (to simulate transit [TR])—and for various holding times of 4, 12, and 24 h. We utilized the BacT/ALERT system with FA and FN bottles and the BACTEC system with Plus (PL) and Lytic 10 (LY) bottles. Standardized inocula and 5 ml of blood were added to each bottle. Fifteen organisms were evaluated based upon expected performance: aerobic (FA and PL), anaerobic (FN and LY 10), and facultative (all bottles). Based upon expected performance, the FA and FN bottles recovered 458 of 468 organisms and 282 of 288 organisms, respectively, whereas the PL and LY bottles recovered 453 of 468 organisms and 257 of 288 organisms, respectively (P = <0.001, FN versus LY). There were 3, 11, 21, and 27 false-negative results for bottles held at 4°C, RT, 37°C, and TR, respectively. There were 4, 8, and 50 false-negative results for bottles held for 4, 12, and 24 h, respectively. Our results support holding these four bottle types at 4°C or at RT for up to 24 h and at 37°C for up to 12 h. We propose that manufacturers only need to make claims for "delayed entry" when these bottles are held for more than 24 h at 4°C or at RT or for more than 12 h at 37°C.
INTRODUCTION
For patients with suspected bacteremia or fungemia, the inoculation of blood culture bottles for the detection of bacteria and yeast often occurs at the patient bedside. Bottles are then transported to the laboratory and loaded into blood culture instruments or incubators. Continuous-monitoring blood culture instruments are widely used for this purpose in the United States. Ideally, the time in transit from the patient to the instrument is kept to a minimum. Although this may have been a routine occurrence at one time, the proliferation of satellite laboratories, core laboratories, and increased dependence on reference laboratories for the performance of routine microbiology testing often results in prolonged delays between the time of specimen collection and the loading of the blood culture bottles into the instruments.
There have been very few studies that have examined the effects of transport time on organism recovery from blood culture bottles. The purpose of the present study was to evaluate the ability of two different continuously monitoring blood culture instruments to detect organisms from seeded blood cultures that were stored under a variety of temperatures for various lengths of time prior to loading into the instruments.
MATERIALS AND METHODS
The present study was conducted in two microbiology laboratories, Geisinger Medical Center (GMC) and Pinnacle Health Laboratories (PHL), with a standard protocol in use at both institutions. The GMC laboratory utilized the BacT/ALERT blood culture system (FA and FN bottles; bioMerieux, Durham, NC), whereas PHL utilized the BACTEC blood culture system (Plus [PL] and Lytic [LY] bottles) (Becton Dickinson Microbiology Systems, Cockeysville, MD).
Blood used for seeding the bottles was obtained from patients donating blood for therapeutic phlebotomy. To ensure that there was no clotting of the blood from these patients, who are more prone to clot than normal patients (hemachromatosis), sodium polyanetholesulfonate was added to attain a final concentration of ca. 0.060%. A solution of SPS was prepared by adding 7,920 mg (7.92 g) of SPS (provided by John Walsh, bioMerieux, Inc.) to 1,200 ml of 0.85% sterile saline. The SPS solution was filter sterilized, and 40 ml was added to each sterile blood collection bag used for blood collection. Anticipating a minimum of 400 ml of blood per patient, 264 mg of SPS was needed per blood collection bag (400 ml of blood plus 40 ml of SPS solution) to attain an SPS concentration of ca. 0.060%. All blood collection bags with SPS were prepared in one batch at GMC and were used by both the GMC and the PHL microbiology laboratories. Blood was collected by using the standard protocols used by both institutions and stored at 4°C prior to use in testing. Blood was used 5 to 8 days after collection at GMC and 1 to 6 days after collection at PHL.
Fifteen recent clinical isolates of ten different microorganisms were selected for use to include some common clinical organisms, as well as some organisms known to be more difficult to recover ("challenge" organisms). Suspensions of microorganisms were prepared in sterile saline from 24- to 48-h cultures, diluted in saline, and then added to a sufficient volume of blood to inoculate all of the bottles. The intended concentration of organisms was 50 CFU/ml in the inoculated blood. A total of 5 ml of the inoculated blood was added to each blood culture bottle, with a final intended inoculum of 250 CFU/bottle. The choice of which types of bottles to inoculate with the 15 organisms was based upon expected performance. For the purpose of the present study, expected performance was defined as follows. Aerobic and facultative anaerobic organisms were expected to grow in the aerobic bottles, while anaerobic and facultative anaerobic organisms were expected to grow in the anaerobic bottles. The results were assessed based upon this definition of expected performance. Aerobic bottles only (FA and PL) were inoculated with seven organisms: Pseudomonas aeruginosa (two isolates), Neisseria meningitidis (two isolates), Acinetobacter baumanii (one isolate), Candida albicans (one isolate), and Cryptococcus neoformans (one isolate). Anaerobic bottles only (FN and LY) were inoculated with two organisms, Bacteroides fragilis and Fusobacterium nucleatum, and all bottles (FA, FN, PL, and LY) were inoculated with six organisms: Staphylococcus aureus (two isolates), Streptococcus pneumoniae (two isolates), and Escherichia coli (two isolates). Both GMC and PHL utilized the same organisms, but the inoculum preparation was done independently in each laboratory.
After inoculation, bottles were held at four different temperatures (refrigerated at 4°C, incubated at 37°C, ambient room temperature [RT] and, to simulate transport to the laboratory, incubated at 37°C for "X" hours and then held at ambient temperature for 2 h [TR]). For example, with the TR bottles, if a bottle was to be held for a total of 24 h, it was held for 22 h at 37°C and the last 2 h at ambient temperature. Ambient temperature in the two laboratories was between 23 and 26°C. Bottles were held for 4, 12, and 24 h for each of the storage temperatures. Bottles were inoculated in triplicate for each bottle-temperature combination. Thus, for an individual organism, there were four temperatures and three storage periods in triplicate for a total of 36 bottles. Three bottles were also inoculated as zero time controls for a total of thirty-nine bottles for each organism-bottle type combination used in the study.
All bottles that did not signal positive within the routine 5-day incubation period were subcultured with 1 or 2 drops from each bottle onto standard bacteriology media.
RESULTS
All test results were evaluated based upon expected performance for aerobic and anaerobic bottles as described in Materials and Methods. Aerobic bottles were inoculated with a total of 13 organisms, whereas anaerobic bottles were inoculated with a total of eight organisms. Thus, the maximum number of positive bottles was 468 for the aerobic FA and PL bottles and 288 for the anaerobic FN and LY bottles. Using these criteria, the FA and PL bottles recovered 458 of 468 organisms and 453 of 468 organisms, respectively, while the FN and LY bottles recovered 282 of 288 organisms and 257 of 288 organisms, respectively (P = <0.001, FN versus LY) (Table 1).
When results were analyzed based upon the temperature of incubation prior to loading into the blood culture instruments, the best recovery was obtained for bottles held at 4°C (Table 2), and the number of false-negative results increased as the temperature of preincubation increased. There were 3, 11, 21, and 27 false-negative results for bottles held at 4°C, RT, 37°C, and TR. In contrast, the time to detection (TTD) was inversely influenced by the holding temperature. The mean TTDs were 37.3, 33.3, 28.5, and 28.5 h at for bottles held at 4°C, RT, 37°C, and TR, respectively.
When the results were analyzed as a function of the holding time, false-negative results increased with the holding time. There were 4, 8, and 50 false-negative results for bottles held for 4, 12, and 24 h, respectively (Table 3). The TTD was inversely influenced by the holding time. The mean TTDs were 26.1, 30.2, and 39.7 h for bottles held for 4, 12, and 24 h prior to loading, respectively.
The interrelated effect of the holding time and holding temperature on the overall recovery is shown in Table 4. False-negative results do not significantly change for bottles held at 4°C for up to 24 h. At RT, the number of false-negative results increases from one bottle at 4 h to four and six bottles at 12 and 24 h, respectively. At 37°C, the number of false-negative results increased dramatically from 0 and 2 at 4 and 12 h, respectively, to 19 at 24 h.
The intended concentration of microorganisms was 50 CFU in the blood that was added to each of the blood culture bottles. Since 5 ml of blood was added to each bottle, the final intended concentration of microorganisms was 250 CFU/bottle. The actual mean concentrations of microorganisms in the blood were 64 CFU/ml (range, 30 to 166) at the PHL and 42 CFU/ml (range, 17 to 38) at the GMC.
The two isolates of N. meningitidis, the single isolate of B. fragilis, Pseudomonas aeruginosa isolate 1, and the single isolate of Candida albicans were detected in all of the bottles tested. There were false-negative results for all of the other isolates tested with differences between the four bottle types for which organisms produced the most false-negative results. Of the 10 false-negative FA bottles, 3 were caused by Streptococcus pneumoniae isolate 1 and 3 were caused by P. aeruginosa isolate 2. Of the 15 false-negative PL bottles, 6 were caused by E. coli isolate 1 and 6 were caused by E. coli isolate 2. Of the six total false-negative FN results, four were caused by the F. nucleatum isolate. Of the 31 total false-negative LY bottles, 13 were caused by S. pneumoniae isolate 1.
Overall, mean TTD was 31.1 h for BACTEC bottles and 32.8 h for BacT/ALERT bottles. However, the data suggest that the organism mix influences the relative performance of the BACTEC and BacT/ALERT bottles. BACTEC bottles were faster for Acinetobacter anitratus, B. fragilis, C. neoformans, and F. nucleatum, while BacT/ALERT bottles were faster for C. albicans, and trending toward significance for S. pneumoniae and N. meningitidis. Little difference was noted for S. aureus, E. coli, or P. aeruginosa.
Subcultures were performed after 5 days for all bottles that were not signaled positive by the instruments. The results were as follows: FA, 6 positive subcultures from 10 bottles; Plus, 15 positive subcultures from 15 bottles; FN, 0 positive subcultures from 6 bottles; and LY, 19 positive subcultures from 30 bottles. The organisms yielding negative subculture results were as follows: 3 S. pneumoniae and 1 A. anitratus from FA bottles; 4 F. nucleatum and 2 S. pneumoniae from FN bottles; and 11 S. pneumoniae from LY bottles.
DISCUSSION
Ideally blood culture bottles should be loaded into continuous-monitoring blood culture instruments as soon as possible after collection in order to minimize the time to detection of microorganisms. Intuitively, one might suspect that false-negative results could result if there was a delay in the loading process. At one time, this may have been more of a theoretical rather than a practical problem since blood culture bottles were usually delivered to the laboratory within a couple of hours after collection, but the proliferation of satellite laboratories, core laboratories, and increased dependence on reference laboratories for the performance of routine microbiology testing now often results in prolonged delays between the time of specimen collection and loading of the blood culture bottles into the instruments.
The actual time at which a reduction in sensitivity occurs because of delays in loading bottles is not clearly established. Although the term "delayed entry" has been frequently utilized when blood culture bottles are not loaded into the instruments in a timely, but not actually defined time period, we are aware of no specific agreed-upon definition of the term "delayed entry." Is delayed entry 30 min, 1 h, and 8 h or some greater time period Should the working definition of delayed entry incorporate a holding temperature We propose that the term delayed entry be specifically reserved for a time period greater than the time period for which there is an acceptable rate of recovery for a particular bottle type and blood culture instrument system. For example, based upon our results, a bottle that was in transit for 2 h prior to loading into an instrument would likely not require a delayed-entry claim.
The acceptable time period prior to which a delayed-entry claim would not be required is open to discussion. In the present study, there was a noticeable decrease in recovery between the bottles held for 4 and 12 h compared to the bottles held for 24 h. However, when the data are examined more carefully, it is apparent that the holding temperature is at least equal to if not more important than the length of time prior to loading of the bottles. Our results support holding BacT/ALERT FA and FN bottles and BACTEC Plus and Lytic bottles at 4°C or at RT for up to 24 h and at 37°C for up to 12 h. We propose that manufacturers make claims for "delayed entry" for these bottles when they are stored for greater than 24 h at 4°C or at RT or greater than 12 h at 37°C. Shorter periods of time at the respective temperatures would be deemed to be normal product use and not require a delayed-entry claim.
While our study was limited to holding bottles for 24 h, it is possible that bottles could be held longer, particularly at 4°C, prior to loading, with no significant decline in recovery. If additional studies support this use, this could be made as a delayed-entry claim. We acknowledge that recovery may be acceptable for some bottles under certain conditions for a time period greater than 24 h. Nonetheless, we would favor the need for a delayed-entry claim for periods over 24 h, if only to emphasize the importance of getting the bottles loaded into an instrument within 24 h. For example, even if recovery is acceptable at 36 h, this still results in a significant delay to detection and reporting of positive blood cultures, a practice which should be clinically discouraged.
We are aware of no other studies that have examined in detail the effect of storage at 4°C prior to loading in the blood culture instruments. Indeed, the overall greatest sensitivity for the temperatures evaluated in the present study was for the bottles stored at 4°C. This could be very useful information, even for laboratories who do not want to routinely transport blood cultures at this temperature (4°C). Many laboratories, including the Geisinger Medical Center, transport specimens from clinics and other hospitals using three coolers with different temperatures (ambient temperature, refrigerated temperature, and frozen). While we may opt to routinely transport blood culture bottles at ambient temperature, these results indicate that bottles inadvertently transported at 4°C can be accepted with no significant loss of sensitivity.
There were significant differences between the performance of the anaerobic bottles tested in the present study. This was most noticeable with facultative organisms where the growth of both aerobic and anaerobic organisms should be anticipated. We can offer no explanation for the superior performance of the FN bottles compared to the LY bottles.
There are few studies that have examined the effects of holding time and temperature prior to loading of blood culture bottles and overall recovery. The data generated from older instrumentation, older software versions, and different medium formulations or bottle types may or may not be applicable to what is currently in use.
Klaerner et al. reported that BacT/ALERT FAN aerobic bottles failed to detect 46.9% of P. aeruginosa isolates, as determined by terminal subcultures (3). Moreover, seeded FAN bottles held at 36°C for as few as 4 h were not able to detect P. aeruginosa strains in the BacT/ALERT instruments (3). However, a more recent study from the same laboratory reported that P. aeruginosa isolates could be recovered from seeded BacT/ALERT FA bottles when held at 36°C for up to 8 h and for 24 h at RT (5).
Two studies have examined delayed entry with the BACTEC 9240 system (1, 4). Chapin and Lauderdale reported that 9240 vials can be preincubated at 35°C for up to 24 h with no significant loss of detection, whereas bottles could be maintained at RT for at least 48 h with no significant loss of recovery (1). Lemming et al. reported that there was a significantly higher false-negative rate for BACTEC bottles preincubated at 35°C compared to bottles preincubated at RT (4). The time of preincubation varied in that latter study.
There are no consistent recommendations from experts concerning delayed entry of blood culture bottles. Cumitech 1B recommends that blood culture bottles be delivered to the laboratory as soon as possible to prevent any delays in detection (2). The Manual of Clinical Microbiology (8th ed.) recommends a storage time of <2 h unless manufacturers specify otherwise (6). The package insert for the BacT/ALERT blood culture bottles recommends that bottles be loaded immediately after inoculation; however, a recent memo to customers states that inoculated bottles may be held at 25°C for up to 24 h prior to loading (Randy Turner, bioMerieux, Inc., personal communication). BD Diagnostics states that BACTEC Aerobic/F Plus and LYTIC/10 Anaerobic bottles can be held for up to 20 h at incubator temperature (temperature not specified) or up to 48 h at RT (Mike Borlet, BD Diagnostics, personal communication).
In conclusion, we have demonstrated differences in organism recovery depending upon the temperature and time of storage prior to loading of blood culture bottles into continuous-monitoring blood culture instruments. Moreover, there is an important relationship between the storage time and storage temperature, such that they should be evaluated in concert. There were differences in bottle performance as well. Although the FA and PL bottles yielded comparable results, the yield with FN bottles was significantly better than the yield with LY bottles.
We acknowledge that our conclusions are based upon the data generated from 10 microorganisms with "seeded" blood cultures and controlled inocula. However, this may represent the limits of what can be practically achieved as evaluations utilizing patient specimens would be very difficult or impossible to perform.
ACKNOWLEDGMENTS
This study was supported in part by a grant from bioMerieux, Inc.
Present address: Carolinas Pathology Group, Carolina Medical Center, P.O. Box 34455, Charlotte, NC 28234-4455.
REFERENCES
Chapin, K., and T.-L. Lauderdale. 1996. Comparison of Bactec 9240 and Difco ESP blood culture systems for the detection of organisms from vials whose entry was delayed. J. Clin. Microbiol. 34:543-549.
Dunne, W. M., F. S. Nolte, and M. L. Wilson. 1997. Cumitech 1B-blood cultures III. American Society for Microbiology, Washington, D.C.
Klaerner, H.-G., U. Eschenbach, K. Kamereck, N. Lehn, H. Wagner, and T. Miethke. 2000. Failure of an automated blood culture system to detect nonfermentative gram-negative bacteria. J. Clin. Microbiol. 38:1036-1041.
Lemming, L., H. M. Holt, I. S. Petersen, C. Ostergaard, and B. Bruun. 2004. Bactec 9240 blood culture system: to preincubate at 35°C or not Clin. Microbiol. Infect. 10:1089-1091.
Seegmuller, U., U. Eschenbach, K. Kamereck, and T. Miethke. 2004. Sensitivity of the BacT/ALERT FA-medium for the detection of Pseudomonas aeruginosa in preincubated blood cultures and its temperature-dependence. J. Med. Microbiol. 53:869-874.
Thomson, R. B., and J. M. Miller. 2003. Specimen collection, transport, and processing: bacteriology, p. 286-330. In P. R. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology, 8th ed. American Society for Microbiology, Washington, D.C.(R. L. Sautter, A. R. Bill)