Three Days of Incubation May Be Sufficient for Routine Blood Cultures with BacT/Alert FAN Blood Culture Bottles

Abstract

BacT/Alert FAN blood culture bottles have been shown to enhance the recovery of bacteria and yeast from blood compared with standard BacT/Alert bottles. It is well established that standard BacT/Alert blood culture bottles require no more than 5 days of incubation for the detection of routine bacteria and yeast. It is less clear, however, whether FAN bottles also routinely require 5 days of incubation. To address this question, we recently reviewed the results of 17,887 blood culture sets collected in FAN blood culture bottles at Geisinger Medical Center. Of these cultures, 1,780 were positive for bacteria or yeast, yielding a total of 1,242 clinically significant isolates. The numbers of isolates recovered on days 1, 2, 3, 4, and 5 were as follows: (values in parentheses are percentages of total significant isolates): 877 (71%), 269 (22%), 65 (5%), 18 (1%) and, 13 (1%), respectively. In total, 97.5% of all clinically significant isolates were detected in the first 3 days of incubation. Of the 31 significant isolates detected on day 4 or 5 of incubation, 17 were detected in concurrent blood cultures within the first 3 days of incubation. Chart reviews were conducted for the 13 patients with the remaining 14 isolates detected on day 4 or 5 to determine whether therapy was changed due to this blood culture result. Therapy was changed for only 1 patient. These results suggest that it may not be necessary to routinely incubate FAN blood culture bottles for more than 3 days.

When the continuous-monitoring automated blood culture instruments were first introduced, 6- or 7-day incubation periods were generally utilized. It became established over time that no more than 5 days of incubation were required for these systems, and indeed regulatory agencies generally now require 5-day minimum incubation periods.

There have been recent reports that 5 days of incubation may not be required for BacT/Alert FAN blood culture bottles (1, 2) or for other automated blood culture systems (3, 5, 7). The purpose of this study was to determine the time to detection for clinically significant isolates of bacteria and yeast in our laboratory when both a FAN aerobic bottle and FAN anaerobic bottle were used for routine blood cultures.

MATERIALS AND METHODS

BacT/Alert FAN blood culture bottles (Organon Teknika Corp., Durham, N.C.) replaced standard BacT/Alert blood culture bottles for all routine blood cultures at Geisinger Medical Center (Danville, Pa.) in May 1997. With the exception of some pediatric blood cultures collected using Pedi-Bacti bottles and occasional “straggler” standard blood culture bottles, all blood cultures since that date have been collected with FAN bottles.

Except for occasional cultures collected with only a single FAN aerobic bottle, sets were collected with both a FAN aerobic bottle and a FAN anaerobic bottle. Phlebotomists are instructed to collect 10 ml for each bottle from all patients weighing 25 kg or more. Smaller volumes are collected from patients weighing <25 kg according to guidelines published by the laboratory. Volumes were not verified for any of the blood culture sets reviewed in this study.

We retrospectively reviewed the results of 17,887 consecutive blood culture sets collected in FAN bottles. Isolates were judged to be clinically significant or probable contaminants by infection control personnel using Geisinger Medical Center infection control standards, which are based on published, external standards (4). The first positive bottle in a set was used to calculate the time to positivity for that set. When two or more organisms were detected from one bottle, the time that the instrument signaled a positive result was used as the time to positivity for all organisms from that bottle. When a different organism was detected in the paired bottle, the time that the instrument signaled a positive result for the paired bottle was used as the time to positivity for the additional organism.

If a patient had a clinically significant blood culture isolate that was first detected on day 4 or 5 of incubation and the patient did not have a concurrent blood culture with the same isolate detected within the first 3 days of incubation, the patient's medical record was reviewed. All medical record review was done by one of us (J.K.P.), an infectious-diseases clinician. This purpose of this review was to determine whether the patient's antimicrobial therapy or overall management was altered based on a positive blood culture result.

RESULTS

Overall, 1,780 cultures were positive for 1,800 bacterial or yeast isolates, including 1,242 clinically significant isolates and 558 probable contaminants (Table 1). A total of 1,211 of the 1,242 clinically significant isolates were recovered in a least one bottle in a set within the first 3 days of incubation. Of the remaining 31 isolates, 17 were recovered in concurrent blood cultures within the first 3 days of incubation.

TABLE 1.

Time to recovery of microorganisms from blood cultures with BacT/Alert FAN bottles

Organism	No. (%) of organisms recovered on day:					Total
	1	2	3	4	5
Significant isolates
Gram-positive cocci
Staphylococcus aureus	219	71	24	3	4	321
Staphylococcus sp. not S. aureus	76	64	10	1	1	152
Streptococcus pneumoniae	37	0	0	0	0	37
Beta-hemolytic streptococci	46	1	1	0	0	48
Other streptococci	27	7	4	0	0	38
Enterococci	39	8	3	1	2	53
Aerobic gram-positive bacilli	2	1	1	0	0	4
Enterobacteriaceae
Citrobacter freundii	2	0	1	0	0	3
Citrobacter koseri	2	0	0	0	0	2
Enterobacter aerogenes	9	2	0	0	0	11
Enterobacter agglomerans	4	1	0	0	0	5
Enterobacter cloacae	20	0	0	1	0	21
Escherichia coli	200	16	3	2	0	221
Hafnia alvei	2	0	0	0	0	2
Klebsiella oxytoca	12	0	0	0	0	12
Klebsiella pneumoniae	41	6	4	1	1	53
Morganella morganii	5	1	1	0	0	7
Proteus mirabilis	22	6	1	0	0	29
Providencia stuartii	3	0	0	0	0	3
Salmonella sp.	5	3	0	0	0	8
Serratia marcescens	11	3	2	0	0	16
Other GNB
Acinetobacter anitratus	5	0	0	0	0	5
Aeromonas sp.	2	0	0	0	0	2
Cardiobacterium hominis	0	2	1	0	0	3
Chryseomonas luteola	1	0	0	0	0	1
Flavobacterium sp.	2	0	0	0	0	2
Haemophilus influenzae	2	2	0	0	0	4
Neisseria meningitidis	2	0	0	0	0	2
Oxidase-positive GNB	4	1	0	1	0	6
Pasteurella multocida	4	0	0	0	0	4
Pseudomonas aeruginosa	39	4	1	1	0	45
Pseudomonas fluorescens gp.	0	1	0	0	0	1
Stenotrophomonas maltophilia	5	1	0	0	0	6
Yeasts
Candida albicans	7	7	0	0	0	14
Candida glabrata	0	3	2	0	0	5
Candida krusei	0	1	0	0	0	1
Candida parapsilosis	0	10	0	0	0	10
Candida tropicalis	2	5	0	0	0	7
Cryptococcus neoformans	0	0	2	0	1	3
Anaerobes
Bacteroides fragilis group	7	34	3	2	0	46
Fusobacterium nucleatum	0	1	0	0	1	2
Anaerobic GNB	0	3	1	1	0	5
Clostridium sp.	9	1	0	1	2	13
Anaerobic gram-positive cocci	2	3	0	3	1	9
Total	877 (71)	269 (22)	65 (5)	18 (1)	13 (1)	1,242
Probable contaminants
Staphylococcus sp. not S. aureus	185	206	30	9	8	438
Micrococcus sp.	3	15	3	1	4	26
Aerobic gram-positive cocci	22	13	3	4	0	42
Bacillus sp.	15	4	1	0	0	20
Aerobic gram-positive bacilli	0	7	6	1	1	15
Neisseria sp.	0	1	0	0	0	1
Pseudomonas aeruginosa	1	0	0	0	0	1
Bacteroides sp.	0	1	1	0	0	2
Anaerobic gram-positive cocci	0	2	1	0	2	5
Anaerobic gram-positive bacilli	1	0	0	1	6	8
Total	227 (41)	249 (45)	45 (8)	16 (3)	21 (4)	558

Nine significant isolates were recovered on day 4 of incubation. These included single isolates of Pseudomonas aeruginosa, Staphylococcus aureus, Enterobacter cloacae, Klebsiella pneumoniae, an unidentified non-lactose-fermenting gram-negative bacillus (GNB), an unidentified anaerobic GNB, and three anaerobic gram-positive cocci (two from the same patient, recovered from separate cultures). A chart review for each of these eight patients revealed that no changes in antimicrobial therapy were made as a result of any of these positive blood cultures. None of these eight patients expired during this septic episode.

Five significant isolates were recovered on day 5 of incubation. These included single isolates of Cryptococcus neoformans, S. aureus, Enterococcus faecium, K. pneumoniae, and Fusobacterium nucleatum. The patient with the positive blood culture for C. neoformans also had concurrent testing performed on cerebrospinal fluid (CSF) with a positive cryptococcal antigen test result. The CSF culture also grew C. neoformans. Thus, at the time of the positive blood culture report, the patient was already being treated with amphotericin B. The E. faecium isolate was from one set from a cystic fibrosis patient. Ampicillin was added to the patient's therapy based on this culture result; however, it is not clear to us that this was truly a significant isolate. No change in treatment resulted from the positive cultures from any of the other three patients. None of these five patients expired during this septic episode.

The time to recovery of probable contaminants is also listed in Table 1. Ninety-three percent of probable contaminants were recovered in the first 3 days of incubation.

DISCUSSION

When the continuous-monitoring automated blood culture instruments were first introduced, 6- or 7-day incubation periods were generally utilized. Over time, it became established that no more than 5 days of incubation were required for the BacT/Alert blood culture system when standard BacT/Alert blood culture bottles were used (6, 8).

BacT/Alert FAN blood culture bottles, which were introduced at a later time than the standard bottles, were designed to enhance the recovery of fastidious bacteria, bacteria from patients receiving antimicrobial therapy, and yeasts in comparison to the standard BacT/Alert blood culture bottles. We and others have asked whether 5 days of incubation are necessary for the FAN blood culture bottles.

Cornish et al. demonstrated that a BacT/Alert FAN aerobic bottle combined with a BacT/Alert standard anaerobic bottle detected 95% of all clinically significant isolates within 3 days of incubation (1). Importantly, they transiently vented some of the anaerobic bottles in this study, so this study did not utilize the more-traditional aerobic-anaerobic blood culture bottle combination. In a subsequent study, Cornish et al. demonstrated that all clinically significant episodes of bloodstream infections were detected within 4 days of incubation when paired FAN aerobic and FAN anaerobic bottles were utilized (2).

In our study, we recovered 97% of all clinically significant bacterial and fungal isolates within the first 3 days of incubation using FAN blood culture bottles. Indeed, there were a total of only 14 clinically significant isolates from 13 patients that were recovered on day 4 or 5 of incubation that had not been detected by either a paired bottle from the same set or from a concurrent blood culture. For only 1 of these 14 isolates was there any change in antimicrobial therapy based on this positive result.

Two issues raised by Cornish et al. merit discussion (1). They are the relationship of inadequate or inappropriate fill volumes for blood culture bottles to the time to recovery of organisms and the times to detection for certain fastidious organisms, such as Actinobacillus actinomycetemcomitans.

Cornish et al. noted in their study published in 1998 that only blood culture bottles with adequate fill volume (>8 ml) were included in their analysis (1). They raised a concern as to whether bottles with lesser volumes might require longer incubation for detection of significant isolates. In a subsequent study, they addressed this issue by including all bottles (adequate or inadequate fill volume) and still were able to detect all clinically significant episodes of bloodstream infections within 4 days of incubation (2).

In our institution, we publish recommended fill volumes for blood cultures based on the weights of the patients. For adults, the recommended volume is 10 ml in each of the two bottles in a standard set of FAN bottles. However, we do not monitor the fill volume on a routine basis. Based on our experience, the data generated in the study certainly include some bottles with inadequate fill volumes. Nonetheless, 97% of significant isolates in our study were recovered within 3 days of incubation.

The second issue, whether certain fastidious species of bacteria may routinely require longer periods of incubation, is more difficult to definitively address. In our study, 61 species or groups of bacteria and yeast were included in the 1,242 clinically significant isolates. While it is likely that there are other species of bacteria that may require more than 3 days of incubation for routine detection, the fact remains that this study encompassed all FAN blood culture results for an 18-month period in a tertiary-care medical center. When our numbers are combined with those of Cornish et al. (2), the results of over 27,000 FAN blood culture sets can be analyzed. When a 5-day incubation protocol was used, no species of bacteria were recovered in either of these studies, which routinely required the fifth day of incubation.

Similar studies have been performed with other automated blood culture systems to determine whether a 5-day incubation period is required for these systems as well. However, unless paired side-by-side studies are conducted, care must be taken when trying to compare the performance of different blood culture systems. Changes in media, software, and bottles occur on a routine basis. In essence, any study result represents a snapshot in time for a particular blood culture system.

Doern et al. reported that, with the AccuMed ESP blood culture system, only 0.1% of significant isolates would have been missed had a 4-day, rather than a 5-day, incubation period been utilized (3). Han and Truant more recently reported that 95.5% of common pathogens were recovered in the first 3 days of incubation (5). They stated that 3 days of incubation may be all that is routinely required for blood culture bottles using the AccuMed ESP-384 blood culture system.

Reisner and Woods evaluated the time to positivity for blood and sterile-body-fluid cultures with BACTEC 9240 using Plus Aerobic/F and Standard Anaerobic/F bottles (7). They concluded that 4 days of incubation were adequate for bacteria while 6 days were necessary for yeast.

In conclusion, we have demonstrated that 3 days of incubation may be all that is necessary for the detection of routine bacteria and yeast when utilizing BacT/Alert FAN blood culture bottles. Additional studies in other institutions are merited to confirm this observation. Similar observations have been made with other automated blood culture systems, suggesting that 5 days of incubation may not be necessary for these systems as well. In our laboratory, we continue to employ a 5-day blood culture incubation protocol, in part due to regulatory requirements. Beyond that, our current instrumentation has the capacity to hold bottles for 5 days. Should that capacity be exceeded, we would consider shortening the routine incubation period.