Times to Detection of Bacteria and Yeasts in BACTEC 9240 Blood Culture Bottles

Abstract

A 7-day incubation protocol was instituted with the BACTEC 9240 system for a 1-year period to determine the times to detection of clinically relevant organisms. A total of 23,686 blood and 693 sterile body fluid cultures were received; some cultures were held longer by special request. Of 1,609 likely skin contaminants, 42 were recovered on day 5, 34 on day 6, 16 on day 7, and 5 on day 8. Of 2,803 usual pathogens, 34 were recovered on day 5, 24 on day 6, 15 on day 7 and 1 on day 8. Twenty-one of the latter organisms were considered significant laboratory isolates because they were the first isolates from the respective patients. Chart review showed that 10 of 21 were considered clinically significant, but only 3 (all yeasts) affected the treatment of the patient. Our data show that 4 days of incubation were sufficient to recover all clinically relevant bacteria and 6 days were required to recover all clinically relevant yeasts.

Culturing blood specimens for microorganisms is one of the most important functions of the microbiology laboratory. The most technologically advanced systems for culturing blood are the continuously monitored blood culture instruments. There are currently three systems available: ESP (Trek Diagnostic Systems, Westlake, Ohio), BacT/Alert (Organon Teknika Corp., Durham, N.C.), and BACTEC 9240 (Becton Dickinson, Cockeysville, Md.). There does not appear to be any major difference in the performances of the systems, and all have been shown to be highly reliable (7, 9). The primary difference lies in the method used to detect growth. The ESP detects pressure changes in the headspace of the vial, the BacT/Alert detects color changes with a sensor on the bottom of the vial, and the BACTEC 9240 system detects fluorescence. All systems can be programmed by the user to incubate specimens for various time periods; recommendations range from 5 to 7 days (4, 7, 9). With such highly automated systems, very little additional labor is required to incubate cultures for a full 7 days; however, with longer incubation, more organisms considered to be skin contaminants may be isolated and final results of negative cultures will be delayed. Moreover, additional instruments may be required to accommodate the increased number of bottles. Studies with the BACTEC NR660 and NR860 systems (Becton Dickinson), two of the earlier noncontinuously monitored automated blood culture systems, demonstrated that incubation for 5 days is sufficient to identify the vast majority of bloodstream infections (6, 8, 11), and data from one study suggested that 4 days may be sufficient (2). Evaluations of the ESP and BacT/Alert systems also support a 5-day protocol (1, 3, 10, 12). The continuously monitored system used in our laboratory is the BACTEC 9240. Because of the lack of published data regarding the optimal length of incubation for this system, we instituted a 7-day protocol of incubation and, at the end of 1 year, reviewed the times to detection for the organisms recovered to determine the incubation periods necessary to recover clinically significant bacteria and yeasts.

The study was conducted from 6 August 1997 to 5 August 1998 at the University of Texas Medical Branch, a 900-bed tertiary care medical center which includes a burn unit, multiple intensive care units, a children’s hospital, a neonatal nursery, an emergency room, a state prison hospital, and a network of outpatient clinics. Blood cultures were collected primarily by the nursing staff, who were instructed to inoculate 10 ml of blood into a Plus Aerobic/F bottle and 7 ml into a Standard Anaerobic/F bottle. The Peds Plus/F bottle was used for low-volume blood draws ranging from 0.5 to 5 ml. The volume of blood inoculated into the bottles was not monitored by the laboratory. Results from culture bottles containing other types of specimens (pleural, peritoneal, and synovial fluids and bone marrow and fine-needle aspirates) were also included in the study. BACTEC bottles were placed into the 9240 instrument immediately upon receipt in the laboratory and incubated at 35°C for 7 days. Cultures were occasionally incubated for 21 days at the request of the physician. There were approximately two to three requests for extended incubation per month. Organisms recovered from blood or body fluid cultures were identified by using standard methods in the laboratory (4, 5). The times to detection of positive cultures were obtained retrospectively from the BACTEC Vision database by using a custom query developed by the manufacturer for our institution. Selected reviews of medical records were performed to determine the clinical significance of some isolates. Assessment of clinical significance was based on whether the isolate was specifically noted as such in the medical record and/or if antimicrobial therapy was initiated or continued based on the blood culture result.

A total of 23,686 blood cultures and 693 body fluid cultures were received over the course of this study. The number of blood culture sets versus single bottles or pediatric bottles could not be determined from the available retrospective data. The positivity rates were 11.5% for blood specimens and 17.8% for body fluid specimens. The times to detection for organisms recovered are shown in Table 1. All organisms are included in the table regardless of whether a previous bottle in the same culture set was positive. Organisms that are not generally considered to be skin contaminants were classified as “usual pathogens,” and organisms generally considered skin contaminants were classified as “occasional pathogens.” Of the latter group, 42 organisms were recovered on day 5, 34 on day 6, 16 on day 7, and 5 on day 8. Propionibacteria accounted for the majority of these (66%). All Bacillus, Corynebacterium, Micrococcus, and Propionibacterium spp., 13 of the 18 coagulase-negative staphylococci, and 5 of the 7 viridans group streptococci isolated on days 5 or later represented single isolates from the respective patients and were considered clinically insignificant. For the remaining two isolates of viridans group streptococci (both from the same patient) and two of the remaining five isolates of coagulase-negative staphylococci, the same organism was recovered from two or more companion blood cultures in ≤4 days of incubation, suggesting that these isolates may represent significant infections but probably did not contribute to the medical management of the patient. The three remaining isolates of coagulase-negative staphylococci represented the second positive set of cultures for each patient and could potentially indicate a true bloodstream infection; however, review of the patients’ medical records revealed that the positive cultures were not considered clinically significant by the healthcare provider.

TABLE 1.

Organisms recovered from blood culture bottles

Organism	Total no. recovered	No. recovered on day:
		1	2	3	4	5	6	7	8
Usual pathogens
Acinetobacter spp.	20	19	1
Actinomyces spp.	2			1		1
Aeromonas hydrophila	1	1
Alcaligenes xylosoxidans	11	5	5	1
Bacteroides spp.	18	9	4	2	3
Brucella spp.	6				6
Campylobacter jejuni	2				2
Candida spp.a	88	18	36	21	4	4	2	3
Citrobacter spp.	6	6
Clostridium spp.	30	24	1	1		1	3
Corynebacterium jeikeium	2	1	1
Cryptococcus spp.	10				3	3	3	1
Eikenella corrodens	2		2
Enterobacter spp.	95	80	8	3	2	1			1
Enterococcus spp.	255	199	46	6	2		2
Escherichia coli	394	343	25	11	4	3	6	2
Eubacterium lentum	1				1
Flavobacterium indologenes	9	7	2
Fusarium spp.	1					1
Fusobacterium spp.	7	3	2	1	1
Haemophilus spp.	10	9	1
Histoplasma capsulatum	6				2	2		2
Lactobacillus spp.	5	2	1	1		1
Listeria monocytogenes	3	3
Klebsiella spp.	266	236	9	10	6	4		1
Moraxella catharralis	1		1
Morganella morgannii	9	9
Mycobacterium chelonae-M. abscessus	4					1		3
Neisseria spp.	4	4
Ochrobacterium anthropi	1			1
Pasteurella multocida	2	2
Peptostreptococcus spp.	11	2	5		3	1
Porphyromonas asaccharolytica	1		1
Prevotella melanogenica	2	2
Proteus mirabilis	39	36	3
Providencia stuartii	1			1
Pseudomonas aeruginosa	109	88	15	3	2	1
Other Pseudomonas or Burkholderia spp.	9	1	4	1	1	2
Rhodococcus spp.	4	2	2
Salmonella spp.	21	13	7				1
Serratia spp.	45	39	2	2	2
Staphylococcus aureus	1,026	860	110	30	9	7	7	3
Stenotrophomonas maltophilia	34	22	10	2
Streptococcus pneumoniae	119	118	1
Streptococci, beta hemolytic	108	106	2
Streptococcus bovis	1	1
Tisseurella preacuta	1				1
Wollinella spp.	1					1
Total	2,803	2,270	307	98	54	34	24	15	1
Occasional pathogens
Bacillus spp.	39	25	7	3	2	1	1
Corynebacterium spp.	48	6	23	14	2	1	1		1
Micrococcus spp.	12		4	5		2	1
Propionibacterium spp.	80	1	2	3	10	32	18	12	2
Staphylococci, coagulase negative	1,201	630	463	70	20	4	8	4	2
Viridans group streptococci	229	187	26	6	3	2	5
Total	1,609	849	525	101	37	42	34	16	5

^aCandida albicans (n = 47), Candida glabrata (n = 9), Candida parapsilosis (n = 23), and Candida tropicalis (n = 9). 

Of those organisms considered usual pathogens, 34 were recovered on day 5, 24 on day 6, 15 on day 7, and 1 on day 8. Twenty-one of these organisms were considered significant laboratory isolates because they were the first or only organism recovered from a routine blood culture from the respective patients (all were from blood specimens). The medical record of each of these patients was reviewed to determine (i) if the isolate was considered clinically significant by the healthcare provider and (ii) if the culture result influenced the medical management of the patient. These findings are summarized in Table 2. Only 10 of the organisms (8 yeasts and 2 bacteria) recovered on day 5 or greater were considered clinically significant. Of these 10 organisms, detection of 3 yeast isolates resulted in a change in the medical management of the patient (i.e., initiation of antimicrobial therapy). For many of the other clinically significant isolates, appropriate antimicrobial therapy had been started earlier based on a previous laboratory test result.

TABLE 2.

Usual pathogens recovered from single positive blood cultures: days 5 through 7

Day of isolation	Organism	Clinically significant	Therapy affected	Comments
7	Cryptococcus neoformans	Yes	No	Previous cerebrospinal fluid (CSF) antigen test was positive
	Histoplasma capsulatum	Yes	No	Previous fungal culture was positive
6	Cryptococcus neoformans	Yes	Yes
	Candida albicans	No	No	From postmortem culture
	Clostridium spp.	No	No
	Clostridium spp.	No	No
	Clostridium spp.	No	No
	Enterococcus spp.	No	No
	Escherichia coli	Yes	No	Previous urine culture was positive
	Staphylococcus aureus	No	No	Patient was terminally ill; therapy was withdrawn
5	Candida glabrata	Yes	No	Patient deceased
	Candida glabrata	Yes	Yes
	Cryptococcus neoformans	Yes	Yes
	Cryptococcus neoformans	Yes	No	Previous CSF antigen test was positive
	Histoplasma capsulatum	Yes	No	Yeast seen in earlier peripheral smear
	Clostridium spp.	No	No
	Fusarium spp.	No	No	Results not documented in chart
	Klebsiella pneumoniae	No	No	Line related; line previously pulled
	Pseudomonas putida	No	No	Results not documented in chart
	Pseudomonas aeruginosa	Yes	No	Previous urine culture was positive
	Peptostreptococcus magnus	No	No

These data suggest that 4 days of incubation were sufficient to recover all bacteria which were essential for the appropriate medical management of the patient. However, to recover all yeasts in the same category, 6 days of incubation were required. Our conclusions are similar to those of Hardy et al. (3), who found that with the BacT/Alert system, 6 days of incubation were necessary to assure the recovery of yeast. Our findings differ from the conclusions reached in other evaluations of the BacT/Alert or ESP systems (1, 10, 12), at least one of which included a large number of yeasts. In these studies, the vast majority of pathogens were recovered within 5 days, and the authors concluded that the benefit of recovering the rare pathogen on day 6 or 7 was not worth the additional effort and expense associated with the isolation of skin contaminants on these days. In our study, only 34 contaminants (2 to 3 per month) were recovered on day 6. Because this number is relatively low and no additional incubators are required to accommodate the extended incubation protocol, we have instituted a 6-day incubation protocol at our institution, where the incidence of fungemia with Candida spp. and Cryptococcus neoformans is relatively high. Because these parameters will differ from institution to institution, each laboratory should determine the incubation protocol that is most appropriate for them.