Pediatric Multicenter Evaluation of the Verigene Gram-Negative Blood Culture Test for Rapid Detection of Inpatient Bacteremia Involving Gram-Negative Organisms, Extended-Spectrum Beta-Lactamases, and Carbapenemases

Abstract

We evaluated the investigational use only (IUO) version of the rapid Verigene Gram-negative blood culture test (BC-GN), a microarray that detects 9 genus/species targets (Acinetobacter spp., Citrobacter spp., Enterobacter spp., Escherichia coli/Shigella spp., Klebsiella oxytoca, Klebsiella pneumoniae, Proteus spp., Pseudomonas aeruginosa, and Serratia marcescens) and 6 antimicrobial resistance determinants (bla_CTX-M, bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA) directly from positive blood cultures. BC-GN was performed on positive BacT/Alert Pediatric FAN and Bactec Peds Plus blood cultures with Gram-negative organisms at two tertiary pediatric centers. Vitek MS (bioMérieux, Durham, NC) was used to assign gold standard organism identification. The Check MDR CT-102 microarray (Check Points B.V., Wageningen, Netherlands) was used as an alternative method for detecting resistance determinants. In total, 104 organisms were isolated from 97 clinical blood cultures. BC-GN correctly detected 26/26 cultures with Acinetobacter spp., P. aeruginosa, and S. marcescens, 5/6 with Citrobacter spp., 13/14 with Enterobacter spp., 23/24 with E. coli, 2/3 with K. oxytoca, 16/17 with K. pneumoniae, and 0/1 with Proteus spp. BC-GN appropriately reported negative BC-GN results in 8/13 blood cultures that grew organisms that were not represented on the microarray but failed to detect targets in 3/5 cultures that grew multiple Gram-negative organisms. BC-GN detected 5/5 and 1/1 clinical blood cultures with bla_CTX-M and bla_VIM. All 6 results were corroborated by Check MDR CT-102 microarray testing. The Verigene BC-GN test has the potential to expedite therapeutic decision making in pediatric patients with Gram-negative bacteremia. Sensitivity was satisfactory but may be suboptimal in mixed Gram-negative blood cultures.

INTRODUCTION

Gram-negative bacteremia has been associated with devastating consequences in children as cases tend to occur most frequently in the most vulnerable pediatric populations. These include neonates, children requiring intensive care support, and patients with cancer or other underlying conditions (1, 2). Nosocomial Gram-negative bloodstream infections have been associated with particularly high risks of mortality (1, 2).

Treatment of Gram-negative bacteremia can be complicated by the recovery of organisms with inherently limited therapeutic options (e.g., Acinetobacter spp., Pseudomonas aeruginosa, and Stenotrophomonas maltophilia). Treatment options for bacteremia involving Enterobacteriaceae that produce AmpC cephalosporinase or Ambler class A extended-spectrum beta-lactamases (ESBLs) are also limited. As poor clinical outcomes have been reported with the use of many, if not all, penicillins and cephalosporins, care is required when selecting therapeutic agents in these cases (3–5). Moreover, delayed initiation of effective antimicrobial therapy has been associated with adverse outcomes in ESBL bacteremia (5, 6). Laboratories often require 24 to 48 h to cultivate, identify, and determine the antimicrobial susceptibility profile of an agent of bacteremia. Rapid testing platforms that expedite this process might therefore accelerate clinical decision making and improve clinical outcomes.

The Verigene Gram-negative blood culture nucleic acid test (BC-GN) (Nanosphere, Northbrook, IL) is an automated, FDA-cleared microarray test that detects clinically relevant Gram-negative organisms directly from positive blood cultures in a rapid, 2-hour format and requires only 5 minutes of hands-on time. This study examined the performance of the investigational use only (IUO) version of BC-GN in two tertiary, pediatric populations using the BacT/Alert Pediatric FAN (bioMérieux, Durham, NC) and the Bactec Peds Plus (BD Diagnostics, Sparks, MD) blood culture systems.

MATERIALS AND METHODS

Clinical blood culture specimens.

This study was performed in the clinical microbiology laboratories at The Children's Hospital of Philadelphia (CHOP) and the Cincinnati Children's Hospital Medical Center (CCHMC). All positive clinical BacT/Alert Pediatric FAN blood cultures from the CHOP and all positive clinical Bactec Peds Plus blood cultures from the CCHMC with Gram-negative organisms present on Gram stain were included in this study. Each study site exclusively used one bottle type. The study period was 1 July 2013 to 31 January 2014, inclusive.

Seeded blood cultures.

To further evaluate the performance of BC-GN in detecting targets that were underrepresented in the clinical blood cultures, seeded blood culture bottles were prepared as previously described (7). Clinical strains included Acinetobacter baumannii complex (n = 2), Citrobacter amalonatica (n = 1), Citrobacter koseri (n = 1), Citrobacter youngae (n = 1), Enterobacter cloacae complex (n = 3), Escherichia coli (n = 14), Ewingella americana (n = 1), Klebsiella oxytoca (n = 5), Klebsiella pneumoniae (n = 7), a Pantoea sp. (n = 1), Proteus mirabilis (n = 3), Proteus penneri/vulgaris (n = 1), P. aeruginosa (n = 4), Raoultella planticola (n = 1), Serratia marcescens (n = 1), and Serratia liquefaciens (n = 1). ATCC strains included Acinetobacter anitratus (49137), Acinetobacter baumannii complex (19606), Acinetobacter lwoffii (9957), Escherichia coli (2326), Klebsiella pneumoniae (BAA 1705, 1898, and 2146), Leclercia adecarboxylata (23216), and Proteus vulgaris (13315). The seeded blood cultures included isolates with bla_CTX-M (n = 14), bla_KPC (n = 7), bla_NDM (n = 1), bla_VIM (n = 2), bla_IMP (n = 2), and bla_OXA-48 (n = 1).

Verigene BC-GN testing.

BC-GN was performed on all qualifying blood cultures according to the manufacturer's instructions. This assay uses a microarray format with probes that detect the (i) rpsA, ompA, and mrkC, (ii) gyrB and metB, and (iii) atpD gene targets for genus-level detection of Acinetobacter spp., Citrobacter spp., Enterobacter spp., and Proteus spp., respectively. Escherichia coli, Klebsiella pneumoniae, Klebsiella oxytoca, Pseudomonas aeruginosa, and Serratia marcescens are identified to the species level using (i) oppA, (ii) yggE and dhaM, (iii) ompA, (iv) sodA, and (v) chiC and ampC probes, respectively. Additional probes detect bla_CTX-M, bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA resistance determinants.

Gold standard organism identification and antimicrobial susceptibility testing.

Gold standard genus/species identification of organisms was assigned using Vitek MS matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) (bioMérieux, Durham, NC) at the CCHMC. All study isolates from the CHOP and the CCHMC underwent MALDI-TOF MS testing. All isolates that were represented on the BC-GN microarray were also tested for the presence of bla_CTX-M, bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA using the Check-MDR CT-102 microarray assay (Check Points B.V., Wageningen, Netherlands) at the Laboratory Specialists facilities in Westlake, OH, as an alternative method (8). In the seeded blood cultures, we used clinical isolates that were previously characterized using the Check-MDR CT-102 microarray or real-time PCR assays designed to detect bla_KPC or bla_CTX-M (9, 10).

RESULTS

Results from the clinical blood culture testing are summarized in Table 1. A total of 104 Gram-negative isolates were recovered from 97 positive clinical blood cultures with Gram-negative organisms seen on Gram stain. The sensitivity of BC-GN was variable. BC-GN correctly identified all 26 blood cultures growing Acinetobacter spp., P. aeruginosa, and S. marcescens. BC-GN detected 5/6 blood cultures with Citrobacter spp., 13/14 with Enterobacter spp., 23/24 with E. coli, 2/3 with K. oxytoca, 16/17 with K. pneumoniae, and 0/1 with Proteus spp. All BC-GN targets were appropriately negative in 8/13 blood cultures that grew Gram-negative organisms that were not represented on the BC-GN microarray. All positive BC-GN results in the clinical blood cultures were concordant with the organism(s) grown from the blood culture broth subculture. The specificities for all targets were therefore 100%. In one blood culture, BC-GN tested positive for K. pneumoniae and a Citrobacter sp., while only K. pneumoniae was isolated from subculture. A repeat subculture of the blood culture broth yielded the same result.

TABLE 1.

Performance of Verigene BC-GN assay compared to standard procedures in positive clinical blood cultures for organism identification

Organism(s)	No. of Gram-negative organisms isolated			No. of isolates correctly identified/no. tested (%)a			No. of isolates with no-call result/no. tested (%)b			No. of isolates with target missed/no. tested (%)c
	CCHMCd	CHOPe	Total	CCHMC	CHOP	Total	CCHMC	CHOP	Total	CCHMC	CHOP	Total
Acinetobacter spp.	3	1	4	3/3 (100)	1/1 (100)	4/4 (100)	0	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Citrobacter spp.	6	0	6	5/6 (83)	0	5/6 (83)	0	0 (0)	0 (0)	1/6 (17)	0 (0)	1/6 (17)
Enterobacter spp.	9	5	14	8/9 (89)	5/5 (100)	13/14 (93)	0	0 (0)	0 (0)	1/9 (11)	0 (0)	1/14 (7)
E. coli/Shigella spp.	12	12	24	11/12 (92)	12/12 (100)	23/24 (96)	1/12 (8)f	0 (0)	1/24 (4)	0 (0)	0 (0)	0 (0)
Klebsiella oxytoca	2	1	3	2/2 (100)	0/1 (0)	2/3 (67)	0 (0)	1/1 (100)	1/3 (33)	0 (0)	0 (0)	0 (0)
Klebsiella pneumoniaeg	7	10	17	6/7 (86)	10/10 (100)	16/17 (94)	0 (0)	0 (0)	0 (0)	1/7 (14)	0 (0)	1/17 (6)
Proteus sp.	1	0	1	0/1 (0)	0	0/1 (0)	0 (0)	0 (0)	0 (0)	1/1 (100)	0 (0)	1/1 (100)
Pseudomonas aeruginosa	8	8	16	8/8 (100)	8/8 (100)	16/16 (100)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Serratia marcescens	3	3	6	3/3 (100)	3/3 (100)	6/6 (100)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Other Gram-negative organismsh	6	7	13	6/6 (100)	2/7 (28)	8/13 (62)	0 (0)	5/7 (71)	5/13 (38)	0 (0)	0 (0)	0 (0)
Total	57	47	104	52/57(91)	41/47 (87)	93/104 (89)	1/57 (2)	6/47 (13)	7/104 (7)	4/57 (7)	0 (0)	4/104 (4)

^a“Correctly identified” by BC-GN was defined as concordance between organism identification resulting from standard procedures and BC-GN at the genus level for Acinetobacter spp., Citrobacter spp., Enterobacter spp., and Proteus spp., at the genus and/or species level for Escherichia coli/Shigella spp., and at the species level for Klebsiella oxytoca, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Serratia marcescens.

^bBC-GN results may have been no call due to internal control failure or signal variation.

^c“Target missed” by BC-GN was defined as a BC-GN result that was negative for a microarray target that should have tested positive.

^dCCHMC, Cincinnati Children's Hospital Medical Center.

^eCHOP, Children's Hospital of Philadelphia.

^fMechanical failure during a BC-GN test on a blood culture that grew E. coli. The repeat test resulted in the correct organism identification.

^gIn one blood culture, BC-GN tested positive for K. pneumoniae and Citrobacter spp., while the subculture grew K. pneumoniae only. The blood culture broth was subcultured a second time and again yielded growth of K. pneumoniae only. An unrelated blood culture grew K. pneumoniae, but BC-GN reported negative results for all targets.

^h“Other Gram-negative organisms” included Achromobacter xylosoxidans, a Pandoraea sp., a Pantoea sp., Salmonella spp., Haemophilus spp., Stenotrophomonas maltophilia, Morganella morganii, Chryseobacterium indologenes, and Ochrobactrum anthropi. A “correct” result was defined as a negative result for all microarray targets.

Five blood cultures grew 2 or more Gram-negative organisms represented on the microarray. They are summarized in Table 2. BC-GN failed to detect organisms represented on the microarray in 3 of the 5 cultures.

TABLE 2.

Performance of Verigene BC-GN in mixed Gram-negative cultures

Blood culture	Organisms isolated	BC-GN result(s)	BC-GN error(s)
CCHMC7	C. freundii, S. marcescens, E. cloacae	Citrobacter spp., Serratia spp.	Enterobacter spp. not detected
CCHMC19	E. coli, K. pneumoniae	E. coli, K. pneumoniae	None
CCHMC34	C. freundii, E. coli, P. mirabilis	Citrobacter spp., E. coli	Proteus spp. not detected
CCHMC36	K. pneumoniae, C. youngae	K. pneumoniae	Citrobacter spp. not detected
CCHMC46	C. freundii, S. marcescens	Citrobacter spp., Serratia spp.	None

BC-GN reported invalid (no-call) results in 6% (6/97) of the clinical blood cultures. Five were due to high variabilities in the target or background signals. The sixth was due to inability of the Verigene Reader instrument to identify the imaging controls to locate the array window. An additional test initially failed mechanically. There was sufficient specimen volume to repeat 6 of the tests. Three of the 6 repeat tests yielded correct results, with the remainder reporting a second no-call result.

In the seeded blood cultures, BC-GN tested positive for the appropriate genus and/or species target in 55 of the 57 seeded blood cultures. A blood culture with Leclercia adecarboxylata was reported as an Enterobacter sp., and Raoultella planticola was reported as Klebsiella oxytoca.

Table 3 summarizes the clinical blood cultures that tested positive for a resistance determinant and the blood cultures that were seeded with isolates that were previously characterized for bla_CTX-M, bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA by the Check-MDR CT-102 microarray or real-time PCR. Five clinical blood cultures tested positive for bla_CTX-M by BC-GN (3 K. pneumoniae and 2 E. coli) and one culture grew a strain of P. aeruginosa that tested positive for bla_VIM. The CT-102 microarray results were concordant in all 6 cases. In the 25 blood cultures that were seeded with previously characterized organisms, BC-GN results produced correct results in 24/25 of the cultures. An E. coli isolate with bla_IMP tested negative for all resistance determinants.

TABLE 3.

Clinical and seeded blood cultures with BC-GN resistance determinant targets

Culture type and isolates	Organism(s)	Reference(s)a	BC-GN result(s)
Clinical blood culture
CCHMC 44	E. coli	CTX-M	CTX-M
CHOP3	K. pneumoniae	CTX-M	CTX-M
CHOP4	E. coli	CTX-M	CTX-M
CHOP25	P. aeruginosa	VIM	VIM
CHOP42	K. pneumoniae	CTX-M	CTX-M
CHOP49	K. pneumoniae	CTX-M	CTX-M
Seeded blood culture (clinical isolate)
S5 (CCHMC)	E. coli	CTX-M	CTX-M
S7 (CCHMC)	E. coli	CTX-M	CTX-M
S28 (CCHMC)	E. coli	CTX-M	CTX-M
S34 (CCHMC)	E. coli	CTX-M, OXA	CTX-M, OXA
S37 (CCHMC)	E. coli	CTX-M	CTX-M
S40 (CCHMC)	E. coli	NDM	NDM
S42 (CCHMC)	E. coli	CTX-M	CTX-M
S43 (CCHMC)	E. coli	IMP	Negative targets
S44 (CCHMC)	E. coli	CTX-M	CTX-M
S45 (CHOP)	E. coli	KPC	KPC
S46 (CHOP)	E. coli	CTX-M	CTX-M
S47 (CHOP)	E. coli	CTX-M	CTX-M
S48 (CHOP)	K. pneumoniae	KPC	KPC
S49 (CHOP)	K. pneumoniae	KPC	KPC
S50 (CHOP)	K. pneumoniae	KPC, CTX-M	KPC, CTX-M
S51 (CHOP)	K. pneumoniae	KPC	KPC
S52 (CHOP)	K. pneumoniae	CTX-M	CTX-M
S53 (CHOP)	K. pneumoniae	CTX-M	CTX-M
S54 (CHOP)	P. mirabilis	CTX-M	CTX-M
S55 (CCHMC)	P. aeruginosa	VIM	VIM
S56 (CCHMC)	P. aeruginosa	VIM	VIM
S57 (CCHMC)	P. aeruginosa	IMP	IMP
Seeded blood culture (ATCC strain)b
BAA 1898 (KPC-2)	K. pneumoniae	Not done	KPC
BAA 2326	E. coli	CTX-M	CTX-M
BAA 1705 (KPC-2)	K. pneumoniae	Not done	KPC

^aPresence of CTX-M, KPC, NDM, VIM, IMP, and OXA targets determined by Check-MDR CT-102 microarray assay (Check Points B.V., Wageningen, Netherlands); presence of bla_KPC was determined by real-time PCR (9); and presence of bla_CTX-M was determined by real-time PCR (10).

^bReference testing was not performed on ATCC isolates that were previously characterized.

For organism identification, BC-GN reported correct results in 52/57 (91%) clinical blood cultures at the CCHMC and 41/47 (87%) at the CHOP (P = 0.52).

DISCUSSION

Assays that rapidly and accurately detect Gram-negative agents of bacteremia from positive blood cultures have the potential to improve clinical outcomes. Most recently, studies investigating rapid Gram-negative blood culture testing have focused on the use of MALDI-TOF MS directly on pellets that result from centrifugation of positive blood culture broth. Published data have suggested that this technique may expedite adjustment of antimicrobial therapy in Gram-negative bacteremia, decrease length of hospital stay, and reduce hospital costs (11, 12). While the Verigene BC-GN assay detects a narrower spectrum of Gram-negative organisms, it has the additional advantage of detecting antimicrobial resistance determinants. This multicenter study is the first to evaluate its performance using pediatric blood culture systems.

BC-GN detected 16 of 18 (89%) clinical blood cultures with organisms that are likely to carry the genes that are responsible for AmpC cephalosporinase production (Citrobacter spp., Enterobacter spp., and Serratia marcescens). Because the ampC gene can be expressed following exposure to “inducer” beta-lactams, isolation of such organisms from normally sterile specimens may trigger reconsideration of antimicrobial therapy in favor of using cefepime or a carbapenem (13, 14). In a seeded blood culture, Leclercia adecarboxylata was reported as an Enterobacter sp. Cross-reaction of some strains of Leclercia adecarboxylata with the Enterobacter sp. target is acknowledged in the BC-GN product insert. Such an error may translate into a clinical decision to unnecessarily initiate antibiotic therapy that provides activity against AmpC cephalosporinase-producing organisms.

BC-GN correctly identified 37 of 41 (90%) clinical blood cultures and 34/34 seeded blood cultures with E. coli, K. oxytoca, K. pneumoniae, and Proteus spp. Raoultella planticola was reported as K. oxytoca. Cross-reaction of Raoultella spp. with K. oxytoca is also acknowledged in the BC-GN product insert. Unfortunately, fine-tuning of antimicrobial therapy in these cases without antimicrobial susceptibility data is often challenged by the circulation of cephalosporinase- and carbapenemase-producing strains. In this study, 5 clinical and 15 seeded blood cultures tested positive for bla_CTX-M by BC-GN as well as a second method. Theoretically, expedited reporting of a positive bla_CTX-M target may allow earlier administration of a carbapenem, or possibly, piperacillin-tazobactam (15, 16). Recommended ESBL infection control precautions may also be initiated earlier (17). However, the clinical impact of the BC-GN′s ESBL detection strategy may be limited by a number of factors. First, cocirculation of bla_CTX-M, bla_TEM, and bla_SHV ESBLs is well documented in North America (10). In the Philadelphia area, for example, an adult study from 2009 demonstrated that only 48% of ESBLs from the region tested positive for bla_CTX-M by real-time PCR (10). Without the ability to detect bla_TEM and bla_SHV ESBL determinants, BC-GN would therefore miss a significant proportion of ESBLs if implemented in that region. Second, plasmid-mediated AmpC cephalosporinase production is also well documented in these organisms in North America (18). Unfortunately, BC-GN is not designed to detect AmpC determinants. Accurate, early determination of susceptibility to third- and fourth-generation cephalosporins may therefore be challenging using a platform that detects bla_CTX-M only. Third, the Clinical and Laboratory Standards Institute currently indicates that routine ESBL screening and phenotypic confirmation in Enterobacteriaceae is no longer necessary prior to reporting antimicrobial susceptibility testing results. A major driver behind this recommendation is the inability of this procedure to detect plasmid-mediated AmpC cephalosporinase production. The role and impact of routine genotypic detection of ESBL and AmpC cephalosporinase determinants have yet to be determined.

In contrast, BC-GN detects the carbapenemase determinants bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA, which account for virtually all genotypes reported in the United States (19). We encountered a blood culture with P. aeruginosa that tested positive for bla_VIM by both BC-GN and Check MDR-102. BC-GN correctly reported positive test results for bla_KPC and bla_NDM in all blood cultures that were seeded with KPC- and NDM-producing organisms. Data supporting therapeutic options for treatment of bacteremia with carbapenemase-producing Enterobacteriaceae remain both limited and controversial due to a paucity of clinical data (20, 21). However, early data have suggested that combination therapy may be beneficial in bacteremia caused by KPC-producing K. pneumoniae (21). Rapid detection would therefore allow early examination of antimicrobial options and initiation of recommended infection control precautions to avert institutional outbreaks (22–24).

BC-GN correctly identified 20/20 blood cultures with P. aeruginosa or Acinetobacter spp. Therapeutic decision making is also challenging in these organisms. Drug resistance is mediated by various mechanisms, including outer membrane porin mutations, efflux pumps, and beta-lactamase production. These mechanisms can coexist, affecting in vitro susceptibility status and therapeutic efficacy in concert, making prediction of drug susceptibility difficult (25). The BC-GN bla_OXA target detects OXA-23, OXA-40, OXA-48, and OXA-58, which are all members of the Bush-Jacoby-Medeiros group 2df. These carbapenemases are typically produced by Acinetobacter baumannii complex organisms, but OXA-23 and OXA-48 have been detected in various Enterobacteriaceae (25). We did not encounter any clinical OXA carbapenemase-producing isolates in this study.

Blood cultures that grew multiple Gram-positive organisms represented on the Verigene BC-GP had higher numbers of discordant results (7, 26, 27). In this study, 5% (5/97) of blood cultures analyzed grew 2 or more Gram-negative organisms represented on the microarray. Among blood cultures that yielded valid results on the first BC-GN run, BC-GN missed targets in 3/5 of these blood cultures and 1/85 of the blood cultures that grew a single Gram-negative organism represented on the microarray. It is difficult to make definitive conclusions based on the small number of mixed blood cultures encountered in this study. In pediatrics, polymicrobial bacteremia involving Enterobacteriaceae and non-glucose-fermenting Gram-negative organisms typically occurs in children with underlying illness, including gastrointestinal disorders and cancer (28). Moreover, polymicrobial bacteremia has also been associated with high mortality in adults (29). Given the vulnerability of the patients at risk of polymicrobial Gram-negative bacteremia, until further data are available, a disclaimer along the lines of the following may be useful: “The Verigene BC-GN microarray may have lower sensitivity in detecting mixed Gram-negative bacteremia. Results should be interpreted with caution in patients at high risk of polymicrobial Gram-negative bacteremia (e.g., intra-abdominal sepsis).”

In one blood culture, BC-GN tested positive for K. pneumoniae and a Citrobacter sp., while only K. pneumoniae was isolated from subculture. We hypothesize that this may have been the result of a false-positive BC-GN result. Alternatively, a small quantity of a Citrobacter sp. may have been present and detectable by BC-GN only. We unfortunately did not have the capacity to test the blood culture broth with a second method to determine which was the case. We had 6 invalid results, all of which occurred at the CHOP. Two resolved when BC-GN was repeated. All of the consumables used in these tests originated from the same lot. We did not encounter any further invalid results after starting a new lot.

This study had a number of limitations. First, few blood cultures grew Acinetobacter spp., K. oxytoca, and Proteus spp. Also, only five clinical blood cultures tested positive for a resistance determinant. It is therefore difficult to draw definitive conclusions about the performance of these targets. Second, we encountered only 5 polymicrobial blood cultures. Further investigation of BC-GN′s performance in polymicrobial Gram-negative bloodstream infections is warranted. Finally, in contrast to the Verigene BC-GP assay, BC-GN testing results are not designed to prompt de-escalation of antimicrobial therapy.

In summary, the Verigene BC-GN test has the potential to expedite therapeutic decision making in pediatric patients with Gram-negative bacteremia. BC-GN accurately determined genus and/or species level identification of a broad range of Gram-negative organisms. Further research and development focusing on the addition of other ESBL and possibly AmpC cephalosporinase targets may enhance this assay's ability to aid clinical decision making.