An appropriately performed conventional blood culture can facilitate choice of therapy in resource-constrained settings-comparison with BACTEC 9050.

AIMS: Comparison of conventional blood culture with BACTEC 9050 for rate and time to detection of microorganisms. SETTINGS AND
DESIGN: A prospective study was carried out in a multispecialty tertiary care teaching hospital. SUBJECTS AND METHODS: A total of 835 paired specimens (797 blood and 38 nonblood specimens) were collected and processed according to standard microbiological procedures by both conventional method as well as by BACTEC 9050 automated culture system. Clinical details of patients were recorded. Data were analyzed for time to detection and isolation rate by the two systems and compared.
RESULTS: Overall culture positivity for BACTEC 9050 and the conventional system was 32% and 19.88%, respectively. Eighty-five demonstrated concordant growth, 136 specimens were culture positive by BACTEC only, and 38 specimens were culture positive by conventional only. Twelve contaminants in BACTEC and nine contaminants in conventional system were detected. Using BACTEC 9050, higher isolation was observed for Acinetobacter spp., coagulase negative Staphylococcus spp., Streptococcus spp., and Candida spp. A total of 410 patients were on antimicrobial treatment and culture positivity was significantly higher with BACTEC 9050 (P < 0.0001). There was a significant difference in the mean time to detection with BACTEC 9050 recovering 86.8% of isolates within 48 h (P < 0.0001).
CONCLUSIONS: Although BACTEC 9050 demonstrated a significantly higher recovery of microorganisms from blood, an appropriately performed conventional blood culture can facilitate the choice of therapy.

Introduction

Bacteremia or fungemia is associated with high mortality and can be diagnosed early with blood culture.[1] These are usually patients who are seriously ill with sepsis or children and adults with cardiovascular infections and paucibacillary disease. Appropriate antibiotic treatment is based on isolation, identification, and antimicrobial susceptibility pattern of an organism isolated from blood culture. The conventional blood culture methods currently being used are time consuming, labor intensive, and suffer from poor sensitivity due to the lack of use of nutritionally enriching substances, inability to neutralize the inherent antimicrobial components in blood and/or patients receiving antimicrobial treatment.[23]

In patients who are receiving antimicrobials, the detection of bacteremia can be delayed or negated. To prevent this, various approaches have been employed that remove antimicrobials from blood or otherwise render the drugs inactive. BACTEC plus media are reported to effectively remove all penicillins, most cephalosporins, and vancomycin.[4] They utilize ion exchange and nonionic adsorbent resins to remove antimicrobials thereby enhancing the recovery of organisms. Antimicrobial removal formulations are available for aerobic, anaerobic, and pediatric blood cultures. BACTEC 9050 is a fully automated system, which monitors increase in CO2 concentration produced by growing microorganisms by means of a fluorescent sensor present at the bottom of the bottle. Being a fully automated system for detecting growth, it also has the advantage of earlier detection time.[5] Various studies have demonstrated the higher isolation rate of BACTEC automated system over conventional culture methods.[567]

Ours is a large tertiary care multi-specialty teaching hospital where patients are more likely to be on antibiotics at the time of blood collection because patients are either referrals from other public or private health care settings with very few direct walk-ins. The tests are provided free of cost to patients. The cost per patient would be higher for a commercial assay. It is important to assess whether this additional cost is justified. Hence, this study was undertaken to determine the utility of BACTEC 9050 in our setting by comparing its recovery rate and time to detection with the existing conventional blood culture.

Subjects and Methods

A prospective study was conducted in the Department of Microbiology from March 2011 to June 2013 after obtaining permission from Institutional Review Board. All patients in whom blood culture was clinically indicated and who provided written informed consent were included. Indications for blood culture were suspected cases of sepsis, meningitis, osteomyelitis, arthritis, endocarditis, peritonitis, pneumonia, and fever of unknown origin. Age-specific volume of blood was collected aseptically from patients, such that each system was inoculated with 8–10 ml in adults, 3–5 ml in children aged 2–12 years old, 2–3 ml in infants, and toddlers aged 1 month to 2 years and 1 ml in neonates and corresponded to manufacturer's instructions.[8] Conventional blood culture bottles were labeled and marked for appropriate volume to be added. Blood specimens with insufficient volume in any system were excluded. We have also included specimens other than blood such as cerebrospinal fluid, pericardial fluid, ascitic fluid, synovial fluid, pleural fluid, pus, bronchoalveolar lavage, tissue, bone marrow aspirate, and bile. For specimens other than blood, where volume was at least 2 ml, each was divided equally into two and inoculated into Peds Plus BACTEC vials and conventional pediatric blood culture bottle. Specimens other than blood were from different set of patients. Clinical details of the patients were recorded with respect to age, gender, diagnosis at investigation, ward, unit, and present or recent past antibiotic treatment. For each isolate recovered, a dialogue with the clinician was established for its pathogenic potential.

Conventional culture bottles contained either brain heart infusion or Trypticase Soya Broth with volumes of 50 ml in adults, 30 ml in children, and 10 ml in neonates and sodium polyanethol sulfonate at a concentration of 0.025 g/100 ml of broth. These were incubated and processed as per standard protocol.[9] Subcultures were made at the end of 24 h, 72 h, and 7th day of incubation on 5% sheep blood agar. If no growth was detected at the end of 7 days of incubation, it was reported negative. Any growth was phenotypically characterized up to the species level.[9]

BD BACTEC™ 9050 (Becton and Dickinson Microbiology System, Sparks, MD, USA) Blood Culture System is a continuous monitoring automated culture system. Equipment and vials were procured from Becton and Dickinson Company (India). BACTEC vials were processed as per manufacturer's instructions. Aerobic F and Peds Plus were incubated until signal-positive or till the end of day 5, whichever was earlier. When positive signal was observed, bottles were unloaded from instrument and Gram's stain and cultures were performed as per standard microbiological protocol.[9] An isolate was considered significant if the same organism was recovered from both the systems and was not a contaminant or a known pathogen was recovered from either of the systems and for other isolates recovered from only one of the two systems, after clinical consultation. Contaminants for this study included Micrococcus, Bacillus subtilis, Corynebacterium spp. other than Corynebacterium diphtheriae and Corynebacterium jeikeium, Staphylococcus saprophyticus and Staphylococcus citreus. Growth was considered concordant if the same isolate was obtained in both the culture systems and discordant if both systems were culture positive but different isolates were recovered in each system. Isolates were identified up to species level by standard phenotypic procedures.[9] Methicillin resistance in Staphylococcus aureus (MRSA) spp. and vancomycin resistance in Enterococcus spp. was determined as per CLSI standards.[101112]S. aureus 25923 and MRSA 43300 were used as quality control strains respectively.[101112] These strains were procured from Microbiologics supplied by Hi-Media. The time to culture positivity, i.e., time taken from incubation to growth detection for conventional and signal positive for BACTEC 9050, was noted. Mean time to detection (MTTD) is the average of the time to culture positivity for all the isolates in a particular category. Growth was considered false positive when a BACTEC 9050 vial was signal positive by instrument but negative on Gram's stain as well as culture after 72 h of incubation.

Statistics and data analysis

Data were analyzed for time to detection and isolation rate, by comparing both systems, BACTEC 9050 and conventional. The significance of the difference between the two systems was determined by the Chi-square test, and P < 0.05 was considered significant.[13]

Results

From 835 patients, 835 paired specimens were processed of which 797 were blood [Table 1] and 38 were nonblood specimens [Table 2]. A total of 497 aerobic/F and 338 Peds Plus/F vials were used for the study. Totally, 509 males and 326 females were included of which 62% (517) were adults, 22% (181) were children, and 16% (137) were neonates. Percent positivity of BACTEC in adults, children, and neonates was 29.6% (153/517), 32% (58/181), and 38.7% (53/137) respectively whereas in conventional it was 21.7% (112/517), 17.1% (31/181), and 16.8% (23/137), respectively.

Table 1

Distribution of specimens and Comparison of culture positivity (n=835)

Specimens	Processed paired specimens	BACTEC 9050		Conventional culture
		Culture positive (n=264)	Positive (%)	Culture positive (n=166)	Positive (%)
Blood	797	248	31	159	20
CSF (cerebrospinal fluid)	19	8	42	4	21
Pleural fluid	4	0	0	0	0
Ascitic fluid	5	1	20	0	0
Pericardial fluid	3	1	33	0	0
Knee aspirate and synovial fluid	2	1	50	1	50
Miscellaneous	5	5	100	2	40
Total	835	264	32	166	20

Table 2

The details of isolates recovered from specimens other than blood (n=38)

Organisms	BACTEC positive (n=16)	Conventional positive (n=7)
Acinetobacter spp.	1	1
Pseudomonas spp.	1	1
NFGNB*	1	0
E. coli	2	2
CoNS	6	0
S. aureus	1	0
Streptococcus spp.	2	2
Enterococcus spp.	2	1
Total	16	7

*NFGNB: Non-fermenting Gram-negative bacilli, CoNS: Coagulase negative Staphylococcus, E. coli: Escherichia coli, S. aureus: Staphylococcus aureus

Overall, culture positivity for BACTEC 9050 and the conventional system was 32% and 19.88%, respectively. Monomicrobial growth was observed in 91.6% (242/264) and 96% (160/166) of positive BACTEC 9050 and conventional systems, respectively. Seven false positive vials were detected by BACTEC. Using BACTEC 9050, culture positivity was higher both for blood (31%) as well as for other specimens [42%, Table 2] as compared to conventional (19.94% and 18.42%). The difference was found to be statistically significant (P < 0.0001 and P = 0.02463).

Of the cultured paired specimens, 533 specimens were culture negative by both systems, 85 demonstrated concordant growth, 43 demonstrated discordant growth, 136 specimens were culture positive by BACTEC only and 38 specimens were culture positive by conventional only. Twelve contaminants in BACTEC and nine contaminants in conventional system were detected. The same contaminant was recovered once from both systems.

Using BACTEC 9050, significantly higher isolation [Table 3] was observed for Acinetobacter spp., coagulase-negative Staphylococcus (CoNS), Streptococcus spp. and Candida spp. MRSA and CoNS were 67.74% and 74.75%, respectively.

Table 3

Comparison of isolates recovered

Organisms	BACTEC positive n=264		Total	Conventional positive n=166			P
	Mono-microbial	Poly-microbial		Mono-microbial	Poly-microbial	Total
Acinetobacter spp.	27	11	38	18	4	22	0.0354
Pseudomonas spp	16	1	17	15	2	17	1
NFGNB	13	1	14	14	0	14	1
B. cepacia complex	1	0	1	0	0	0	0.3172
E. coli	13	4	17	11	0	11	0.2386
Other Enterobacteriaeceae	5	2	7	4	0	4	0.3641
Klebsiella spp.	24	5	29	20	2	22	0.3195
Salmonella typhi	1	0	1	0	0	0	0.3172
Hemophilus spp.	1	0	1	1	0	1	1
CoNS	57	9	66	35	2	37	0.0031
S. aureus	15	5	20	10	1	11	0.1028
Streptococcus spp.	9	2	11	2	0	2	0.01221
Enterococcus spp.	9	3	12	5	1	6	0.1552
VRE	1	0	1	1	0	1	1
Candida spp. + other yeast	38	3	41	15	0	15	0.0004
Contaminants	12	0	12	9	0	9	0.1533
Total	242	46	288	160	12	172	<0.0001

NFGNB: Non-fermenting Gram-negative bacilli, CoNS: Coagulase negative Staphylococcus, VRE: Vancomycin resistant Enterococcus, S. aureus: Staphylococcus aureus, B. cepacia: Burkholderia cepacia, E. coli: Escherichia coli

A total of 410 patients were on antimicrobial treatment. Culture positivity was 41.95% (172/410) by both methods or either of the methods [Table 4]. 18.29% (75) showed positive growth by BACTEC only and 3.90% (16) by conventional culture only. This difference was statistically significant (P < 0.0000001). Although a higher isolation was observed for most pathogens using BACTEC 9050, none of them were statistically significant except in Candida spp. However, a significantly higher rate of isolation by BACTEC 9050 was observed when all Gram-positive cocci and Gram-negative bacilli were grouped (GPC, 58 vs. 33, P = 0.00544; GNB, 80 vs. 60, P = 0.06343).

Table 4

Comparison of isolates in patients on antimicrobial treatment n=410

	BACTEC Culture positive n=156	Conventional culture positive n=97	P
Acinetobacter spp.	28	18	0.1292
Pseudomonas spp.	10	12	0.6656
NFGNB	7	8	0.7944
E. coli	10	4	0.1058
Other Enterobacteriaeceae	5	5	1
Klebsiella spp.	19	12	0.2003
Hemophilus spp	1	1	1
CoNS	32	19	0.06015
S. aureus	9	7	0.6136
Streptococcus spp.	6	1	0.5771
Enterococcus spp.	10	5	0.1929
VRE	1	1	1
Candida spp. + other yeast	26	6	0.00031
Contaminants	10	4	0.4343
Total	174	103	<0.0001

NFGNB: Non-fermenting Gram-negative bacilli, CoNS: Coagulase negative Staphylococcus, VRE: Vancomycin resistant Enterococcus, S. aureus: Staphylococcus aureus

Of the 43 discordant paired specimens, in 12, growth was polymicrobial by BACTEC and monomicrobial by conventional with recovery of one of the organisms being common to both systems. A contaminant in one of the two systems with a corresponding potential pathogen in other was observed in 4 (3 conventional and one BACTEC). In the remaining 27, different pathogens were recovered by each system.

Using BACTEC, majority (173) of the culture positives [Table 5] were detected on/before day 2 whereas, during the same time, only three were culture positive by conventional method. The difference was statistically significant. P < 0.0001.

Table 5

Comparison of time to detection

Organism	BACTEC positive isolates n=288						Conventional positive isolates n=172
	D1	D2	D3	D4	D5	Total	D2	D3	D4	D5	D6	D7	D8	Total
Acinetobacter spp.	2	23	11		2	38	1	13	3	3	2	0	0	22
Pseudomonas spp.	5	5	6	1		17		9	6	2				17
NFGNB		3	4	3	4	14		4	2	4		3	1	14
B. cepacia complex			1			1							0	0
E. coli	4	12			1	17	2	8			1		0	11
Other Enterobacteriaeceae		3	4			7		3		1				4
Klebsiella spp.	5	19	3	2		29		21	1					22
Salmonella typhi		1				1								0
Hemophilus spp.			1			1				1				1
CoNS	6	35	16	7	2	66		24	3	5	4	1		37
S. aureus	2	10	6	1	1	20		6	3	2				11
Streptococcus spp.	1	7	1	1	1	11		2						2
Enterococcus spp.	1	10	1			12		5		1				6
VRE			1			1		1						1
Candida spp. + other yeast	1	14	17	8	1	41			2	10	2		1	15
Contaminants		4	6	1	1	12		4	1	1	1	1	1	9
Total	27	146	78	24	13	288	3	100	21	30	10	5	3	172
Positive (%)	9	51	27	8	5	100	2	56	13	19	6	3	2	100

NFGNB: Non-fermenting Gram-negative bacilli, CoNS: Coagulase negative Staphylococcus, VRE: Vancomycin resistant Enterococcus, S. aureus: Staphylococcus aureus, B. cepacia: Burkholderia cepacia, E. coli: Escherichia coli

A comparison of MTTD for those organisms that grew concordantly in both systems is depicted in Figure 1. MTTD was different for different bacteria. It ranged from 15 h to 50 h by BACTEC 9050 and 40 h to 106 h by the conventional system. In both systems, it was maximum for other nonfermenting Gram-negative bacilli and Hemophilus spp. Enterobacteriaceae and GPC had the lowest MTTD.

Figure 1

Mean time to detections of concordant and concordant at least one organism similar in both system n = 85

Discussion

There has been a continuous evolution of methodologies and principles to improve and increase the sensitivity for the detection of organisms in bloodstream infections. There is a need for every institute to evaluate the method, which is best suited for their setup. Considering that ours is a tertiary care and referral center where a greater proportion of patients are likely to be on antimicrobials, we compared the fully automated BACTEC 9050 culture system having the capability of neutralizing or removing commonly used antibiotics, with the conventional culture system in terms of recovery rate, species wise significant recovery if any, contamination rate, mean time to detection including patients receiving antibiotics.

As expected, BACTEC 9050 demonstrated a significantly higher rate of isolation overall (32%), in blood (31%), in nonblood specimens (42%), and in patients receiving antibiotic treatment (38.04%) when compared with the conventional culture. However, in the group of patients with discordant growth, the isolation of two different potential pathogens, one in each of the two systems, made clinical correlation difficult and both organisms were reported to the clinician. Maximum concordant growth between the two systems was observed when the isolate was a GNB.

In this study, blood culture negativity was 68% by BACTEC and 80% by conventional system. The reported reasons for culture negativity could be inappropriate timing of blood collection (recommended is 30–90 min before the fever spikes or in practice, as soon after the fever spike), inadequate volume (each milliliter increase in volume of blood cultured increases recovery of organisms by 3%), concurrent administration of antimicrobials, presence of fastidious organisms, delay in transportation and processing, blood broth dilution ratio, number of culture sets, media used, use of anticoagulants (sodium polyanethol sulfonate [SPS]), duration of incubation, temperature of incubation, and use of agents to remove effect of antimicrobials.[2] In this study, adequate volume, immediate transportation and processing, blood broth dilution ratio, use of SPS, media used, and temperature of incubation were as per recommendation. Forty-nine percent of patients had received antibiotics due to suspected diagnosis of sepsis or case mix before blood collection. BACTEC 9050 was used due to its capability to remove antimicrobials most commonly administered.

In this study, which included adults, children, and neonates, culture positivity was higher in the neonatal group (38.7%). This may be because of the magnitude of bacteremias in infants and children which is generally greater than in adults,[14] a high index of suspicion in high-risk neonates and the practice of collecting specimen before starting antimicrobial treatment as a policy. The culture positivity reported in Indian literature in neonatal sepsis ranges from 5.6% to 47%.[151617]

Various studies have highlighted the effectiveness of BACTEC 9050 system in terms of increase in organism yield, decrease in admission time, reduction in inappropriate antibiotic therapy, facilitating an early diagnosis and saving life by reducing morbidity.[5718] In the present study, overall culture positivity for BACTEC 9050 was 32% as compared to 19.88% by conventional system, which is similar to the findings reported in literature.[519] BACTEC 9050 recovered any organism in an additional 136 specimens as compared to the conventional system, which was expected. However in 38 patients, only conventional culture was positive of which significant isolates were recovered in 25, a finding not reported in most of the studies that have compared automated culture system with conventional. Even though conventional system does not have resins to neutralize the antimicrobial effect in specimens, it was still able to detect organisms.

Automated systems are reported to reduce contaminant growth since repeated subculturing is not required. Very few contaminants were recovered by both systems. In this study, contamination rate was 1.4% (12) by BACTEC and 1.07% (9) by conventional. The difference was not statistically significant (P = 0.51). Similar contamination rate has been reported in other studies.[2021] The lower contamination rate in the present study may be attributed to the supply of alcohol-impregnated swabs that were provided along with the blood culture bottles.

In this study, CoNS and Acinetobacter spp. were the predominant isolates in both systems. Isolation of Gram-negative bacteria was more as compared to GPC, which is similar to other studies.[1921] A few studies, however, have reported a higher Gram-positive isolation.[620] This may be attributed to possible geographical variations in antimicrobial spectrum. Individual pathogen comparison between the two systems showed significantly higher recovery of Acinetobacter spp, Streptococcus spp, and CoNS [Table 3] by BACTEC 9050.

A higher recovery by BACTEC 9050 was also observed in patients on antimicrobial treatment. The difference was not statistically significant except for Candida spp.

There was a significant difference in the MTTD with BACTEC 9050 recovering 86.8% of isolates within 48 h. Using BACTEC, the mean detection time was 31.12 h for all isolates as compared to 67.80 h by conventional method. The shortest time to detection by BACTEC 9050 was 4 h for Escherichia coli. Other studies have reported a mean detection time of 19–24 h hours by BACTEC and 5–7 days by conventional method.[6182022] Since BACTEC is a continuous monitoring system, detection occurs earlier and can be reported in hours. However, conventional blood culture, by its very nature of processing, cannot be reported positive before 24 h and that too when a Gram's stained primary smear reveals the presence of microorganisms. This was not a part of the protocol in the present study. Based on culture, the earliest result would be at the end of 48 h. When MTTD of concordant isolates were analyzed, it was observed to be 25.08 h using BACTEC and 67.56 h using conventional. In this study, only one Streptococcus pneumoniae and one Haemophilus spp. were concordantly recovered by both systems. Time to detect these fastidious organisms was 19 h and 50.83 h respectively by BACTEC and 48 h and 96 h respectively by conventional system. Other studies have reported a recovery time from 10 h to 24 h for S. pneumoniae.[202123]

One of the observations of the present study was that in 32 patients with native or prosthetic valve endocarditis, conventional blood culture recovered three isolates (one each of S. pneumoniae, Haemophilus influenzae, and Enterococcus faecalis) whereas BACTEC, in addition, recovered two Candida spp. and two Streptococcus spp.

Of 137 clinically suspected cases of neonatal sepsis, a significantly higher recovery (P < 0.0001) was obtained using BACTEC 9050 (51) as compared to conventional culture (17). In both systems, CoNS, Klebsiella pneumoniae and Candida spp. were the predominant isolates. This finding is similar to other Indian studies.[242526] BACTEC 9050 recovered a significantly higher number of GNB (22 vs. 8, P = 0.0068) and Candida spp. (18 vs. 6, P = 0.0103) as compared to the conventional system.

Acinetobacter spp. and CoNS recovered in blood are not considered clinically significant in some studies.[2728] Of the 38 Acinetobacter spp. isolated by BACTEC 9050, 30 were recovered from ICU patients at a risk of Acinetobacter spp. sepsis and considered clinically relevant. Of the remaining eight, clinical correlation could not be arrived in four cases only. A similar observation was seen with CoNS with only ten that could be considered as clinically nonsignificant. The type of setting, the presence of devices or lines and other risk factors should be considered while defining a contaminant rather than the type of organism.

False positivity in the present study was found to be 0.83%. False positivity rate in other studies ranges from 0.5% to 2.2%.[202930] The cost per test (not including the cost of equipment and manpower) using BACTEC 9050 was Rs. 400 (age unrelated) and using conventional culture (cost of reusable bottles and media) was approximately Rs. 30 and Rs. 20 in adults and children, respectively.

The limitations of this study are as follows. For CoNS to be considered significant in adults and children, the same species should be repeatedly isolated either from specimens collected from two different sites or at two different times which was not possible in the present study. However, only those that had the potential of association after clinical consultation have been classified as significant. CoNS were speciated (Staphylococcus warneri, Staphylococcus lugdenensis, Staphylococcus hemolyticus, Staphylococcus epidermidis) and Staphylococcus saprophyticus has been classified as a contaminant. Furthermore, actual cost benefit of earlier detection of organisms by BACTEC 9050 with respect to treatment, its outcome and length of stay has not been considered in the present study.

Conclusion

By demonstrating a significantly higher recovery including in the group of patients receiving any antimicrobial, in less than half the time as compared to the conventional culture, BACTEC 9050 will be an investment worth the cost. At the same time, recovery by conventional culture even in patients receiving antimicrobials suggests that an appropriately performed blood culture with adequate volumes can provide the much-needed information in resource-constrained settings on the possible etiology and also provide the isolate which can be further tested to facilitate choice of therapy. The provision of skin disinfectant wipes along with blood culture bottles also reduces contamination rate.

The study was sponsored by BD India.

Financial support and sponsorship

The study was supported by Becton, Dickinson and Company, India Ltd.

Conflicts of interest

Consumables and equipment were provided by Becton, Dickinson and Company, India Ltd.