Carbapenem-Resistant Non-Glucose-Fermenting Gram-Negative Bacilli: the Missing Piece to the Puzzle

Abstract

The non-glucose-fermenting Gram-negative bacilli Pseudomonas aeruginosa and Acinetobacter baumannii are increasingly acquiring carbapenem resistance. Given their intrinsic antibiotic resistance, this can cause extremely difficult-to-treat infections. Additionally, resistance gene transfer can occur between Gram-negative species, regardless of their ability to ferment glucose. Thus, the acquisition of carbapenemase genes by these organisms increases the risk of carbapenemase spread in general. Ultimately, infection control practitioners and clinical microbiologists need to work together to determine the risk carried by carbapenem-resistant non-glucose-fermenting Gram-negative bacilli (CR-NF) in their institution and what methods should be considered for surveillance and detection of CR-NF.

INTRODUCTION

The increasing rate of carbapenem resistance in Gram-negative bacteria is a serious global health threat. Carbapenem-resistant Enterobacteriaceae (CRE) have been the focus of a number of studies, guidelines, and infection control efforts. Primarily, this is due to their increasing prevalence, their association with carbapenemase production, and multiple reports of nosocomial outbreaks with high rates of morbidity and mortality (1, 2). However, the often-neglected non-glucose-fermenting Gram-negative bacilli are increasingly acquiring resistance to carbapenem antibiotics as well. Clinically, these organisms can also cause difficult-to-treat life-threatening infections, often in patients with significant comorbidities. As intrinsic antibiotic resistance to multiple antibiotic classes is common with these organisms, carbapenems are frequently used for treatment. If carbapenem resistance becomes more widespread, therapeutic options may become tragically few. Thus, in 2013, these multidrug-resistant (MDR) organisms were identified as a serious public health threat by the Centers for Disease Control and Prevention (CDC) (3).

This minireview will focus on carbapenem-resistant non-glucose-fermenting Gram-negative bacilli (CR-NF) and carbapenemase-producing non-glucose-fermenting Gram-negative bacilli (CP-NF), particularly carbapenem-resistant Pseudomonas aeruginosa and Acinetobacter baumannii. Other CR-NF, such as Stenotrophomonas maltophilia and Burkholderia species, carry chromosomally encoded carbapenemases, which are less likely to be transmitted to other Gram-negative organisms. Together, CRE and CR-NF are referred to here as carbapenem-resistant organisms (CROs). Furthermore, the CROs that produce carbapenemases are referred to as carbapenemase-producing organisms (CPOs) or, more narrowly, carbapenemase-producing Enterobacteriaceae (CPE). Table 1 summarizes the terms used to describe carbapenem resistance mechanisms among Gram-negative bacilli.

TABLE 1.

Summary of terms used to describe carbapenem resistance among Gram-negative bacilli

Term	Abbreviation	Definition	Organisms	Carbapenem mechanisms
Carbapenem-resistant Enterobacteriaceae	CRE	Encompasses organisms from the family Enterobacteriaceae that are resistant to carbapenems regardless of mechanism	Enterobacteriaceae (i.e., Escherichia coli, Klebsiella pneumoniae)	Includes all carbapenem resistance mechanisms (i.e., production of an AmpC/ESBL plus porin mutations or carbapenemase production)
Carbapenemase-producing Enterobacteriaceae	CPE	Encompasses organisms from the family Enterobacteriaceae that are resistant to carbapenems due to the production of a carbapenemase enzyme	Enterobacteriaceae (i.e., E. coli, K. pneumoniae)	Production of a carbapenemase (i.e., KPC, NDM, OXA-48-like)
Carbapenem-resistant organisms	CRO	Encompasses all Gram-negative bacilli, including the Enterobacteriaceae that are resistant to carbapenems regardless of the mechanism	All Gram-negative bacilli, including the Enterobacteriaceae, the non-glucose-fermenting Gram-negative bacilli (i.e., Pseudomonas and Acinetobacter spp.), and others (i.e., Aeromonas spp.)	Includes all carbapenem resistance mechanisms
Carbapenemase-producing organisms	CPO	Encompasses all Gram-negative bacilli, including the Enterobacteriaceae that are resistant to carbapenems due to the production of a carbapenemase enzyme	All Gram-negative bacilli, including the Enterobacteriaceae, the non-glucose-fermenting Gram-negative bacilli (i.e., Pseudomonas and Acinetobacter spp.), and others (i.e., Aeromonas spp.)	Production of a carbapenemase (i.e., KPC, NDM, OXA type, VIM, IMP)
Carbapenem-resistant non-fermenters	CR-NF	Encompasses the non-glucose-fermenting Gram-negative bacilli that are resistant to carbapenems regardless of mechanism	Non-glucose-fermenting Gram-negative bacilli (i.e., Pseudomonas and Acinetobacter spp.)	Includes all carbapenem resistance mechanisms, including carbapenemase production, decreased permeability due to porin mutations, overexpression of efflux pumps, and changes in penicillin-binding proteins
Carbapenemase-producing non-fermenters	CP-NF	Encompasses the non-glucose-fermenting Gram-negative bacilli that are resistant to carbapenems due to the production of a carbapenemase enzyme	Non-glucose-fermenting Gram-negative bacilli (i.e., Pseudomonas and Acinetobacter spp.)	Production of carbapenemases (i.e., VIM and IMP in Pseudomonas spp. and OXA-type in Acinetobacter spp.)

In this review, we discuss the available literature on considerations for screening for CR-NF colonization in patients (who, what, where, and why) and methods for the detection of colonization of CR-NF, as well as detection of carbapenemase production among CR-NF isolates. Ultimately, infection control practitioners and clinical microbiologists need to work together to determine the risk carried by CR-NF in their institution and what methods should be considered for the surveillance and detection of CR-NF/CP-NF. A recent heightened awareness of carbapenem-resistant Enterobacteriaceae (CRE) has occurred, and many institutions have implemented measures to increase the detection of CRE and control their spread. But, CR-NF carry an equal threat, and similar measures need to occur to control their spread. In many ways, they represent the missing piece of the puzzle to control carbapenem resistance among Gram-negative bacilli.

EPIDEMIOLOGY OF CR-NF

The rates of carbapenem resistance in non-glucose-fermenting Gram-negative bacilli have been gradually increasing worldwide over the last 10 years and vary geographically (4, 5). Recent studies from South Korea have reported the proportion of carbapenem resistance in Acinetobacter to be as high as 32 to 56% in hospitalized patients (6, 7). In the United States, resistance rates have been reported from approximately 34% to as high as 62.6% (8–10). Reports from the National Healthcare Safety Network (NHSN) in the United States have demonstrated an increase from 33% carbapenem resistance in 2006 to 2007 to >60% among Acinetobacter species isolates in 2009 to 2010 (10). Studies of antibiotic resistance in Pseudomonas spp. have shown similar trends. Carbapenem resistance rates among Pseudomonas aeruginosa isolates in most countries range from 10 to 50% and have been reported as low as 3.3% in Canada to >50% in Russia, Southwest Asia, and South America (4). A recent study from Iran showed imipenem resistance in 68% of Pseudomonas species isolates from hospitalized burn patients, with phenotypically detectable metallo-β-lactamase (MBL) production in 94% of these isolates (11). In contrast, carbapenem-resistant P. aeruginosa isolates from the United States collected in 2010 as part of the Carbapenem Antimicrobials Pseudomonas Isolate Testing at regional Locations (CAPITAL) study demonstrated the highest rates in the southern United States, with 35%, 26%, and 19% resistance to imipenem, meropenem, and doripenem, respectively (12). These increases in carbapenem resistance among non-glucose-fermenting Gram-negative bacilli can be attributed to multiple factors, such as the increased use of broad-spectrum antibiotics (including the carbapenems), the use of indwelling medical devices, an increase in immunocompromised host populations, and the acquisition of carbapenemases (4, 5). Notably, the highest burden of carbapenem resistance among Gram-negative health care-associated infections in the United States, as reported by the NHSN from 2009 to 2010, was observed among A. baumannii (62.6%) and P. aeruginosa (26.1%) in comparison to CRE, in which carbapenem resistance was highest among Klebsiella pneumoniae, at 12.8%.

MECHANISMS OF CARBAPENEM RESISTANCE AMONG CR-NF

Carbapenem resistance among CR-NF can be mediated by multiple different mechanisms, including carbapenemase production, decreased permeability due to porin mutations, overexpression of efflux pumps, and changes in penicillin-binding proteins (13, 14). Carbapenemase production among CR-NF is the most concerning mechanism due to their increasing prevalence, their easily transmissible nature, as they are carried on mobile genetic elements, and their association with genes conferring resistance to other classes of antimicrobials, leading to multidrug resistance. The Ambler classification scheme is a method of differentiating carbapenemases (and other β-lactamases) based on their amino acid sequences. It does not differentiate resistance mechanisms based on organism type (i.e., glucose nonfermenters and fermenters). All classes of Ambler carbapenemases have been detected in Pseudomonas and Acinetobacter species (15). The most commonly acquired carbapenemases among Acinetobacter spp. are from Ambler class D, in particular, OXA-23, OXA-40, OXA-58, and OXA-143. In addition to the acquired OXA carbapenemases, the A. baumannii-A. calcoaceticus complex possesses chromosomally encoded OXA-51-like enzymes with weak carbapenemase activity. The bla_OXA-51-like genes are not expressed among wild-type strains; however, the insertion of ISAba1 upstream of the genes may lead to their expression (16). In contrast, the most commonly encountered carbapenemases among Pseudomonas spp. are class B MBLs, such as VIM and IMP (15, 17). Of note, other carbapenemases, such as GES, NDM, and KPC, are now also being detected and reported in both Pseudomonas and Acinetobacter species (4, 18). Additionally, most carbapenem resistance among Pseudomonas spp. is due to noncarbapenemase methods, such as the loss of OprD porin expression or upregulation of MexAB-OprM efflux pump (19). The mechanisms mediating carbapenem resistance among the nonfermenters vary geographically (4, 5). Currently, in the United States and Europe, OXA carbapenemase production (specifically that of OXA-23) is the primary resistance mechanism among A. baumannii, whereas the loss of OprD porin expression, without the expression of a carbapenemase, is the primary resistance mechanism among P. aeruginosa species (17). We direct the readers to more in-depth reviews on the mechanisms of carbapenem resistance in CR-NF (4, 5, 13, 14, 19).

DEFINING CARBAPENEM RESISTANCE AMONG CR-NF

CR-NF are defined by nonsusceptibility to at least one of the clinically available carbapenems, excluding ertapenem, by antimicrobial susceptibility testing, according to the Clinical and Laboratory Standards Institutes (CLSI) guidelines. For P. aeruginosa and Acinetobacter spp., the criterion that defines susceptibility is an MIC of ≤2 μg/ml for meropenem, imipenem, and doripenem (20). Among the carbapenems, imipenem appears to be the most sensitive for detecting CR-NF (12). However, further studies are required to determine which carbapenem is most suitable for predicting CR-NF.

DETECTION OF CARBAPENEMASE PRODUCERS AMONG CR-NF

Currently, CLSI does not advocate testing for the detection of carbapenemases in clinical isolates over and above traditional antimicrobial susceptibility testing for patient care. The reasoning is that most carbapenemase producers will be captured within the current resistance breakpoints, and the detection of a resistance mechanism does not provide any additional information for patient management. On the other hand, CLSI recognizes that the detection of carbapenemase producers is important for epidemiological and infection control purposes. Some β-lactamase experts disagree and support resistance testing for both infection control measures and for routine diagnostic testing. One reason cited in support of resistance testing is that carbapenemase production among Gram negatives may lead to elevated carbapenem MICs that do not always cross the resistance threshold based on the current clinical breakpoints (15, 21). Accordingly, EUCAST has proposed the use of epidemiological breakpoints followed by phenotypic carbapenemase detection among the Enterobacteriaceae to prevent the spread of both phenotypically resistant isolates and isolates with MICs below the resistance cutoff carrying transmissible elements that could further spread to susceptible bacteria. CR-NF were not included in this approach. This may be due to the mostly predictable nature of the epidemiology of CR-NF in the United States and Europe (i.e., OprD mutations in Pseudomonas and OXA-23-producing A. baumannii strains) (22). However, a more complex epidemiology may arise with further spread of a variety of carbapenemases in CR-NF (similar to CRE), requiring further consideration of this matter in the future.

CLINICAL SIGNIFICANCE OF CR-NF

Although CR-NF can be recovered from all body sites, they are most commonly associated with health care-acquired pneumonias (23). They are predominantly implicated in infections in patients with previous health care exposure, extensive antibiotic use, indwelling hardware, or immunocompromising conditions (24). Importantly, the identification of CR-NF in hospitalized patients has been associated with increased morbidity and mortality. For instance, mortality in patients with carbapenem-resistant Acinetobacter spp. can be almost three times greater than with carbapenem-susceptible Acinetobacter species (25). Furthermore, mortality rates are especially high when certain carbapenemases are detected, such as the MBLs (26). This is not surprising, since MBLs confer broad-spectrum antibiotic resistance; however, it is important to note that the prevalence of MBL expression in CR-NF varies. In India alone, the rate of reported MBL expression among CR-NF varies from ∼20 to 80% (27).

Most worrisome is the potential for resistance gene transfer from CR-NF to other Gram-negative organisms and vice versa. For instance, GIM-1 was first identified in P. aeruginosa from Dusseldorf, Germany, in 2002. In subsequent years, GIM-1-positive isolates of Enterobacter cloacae, Serratia marcescens, Acinetobacter pittii, and P. aeruginosa were found in that same region in Germany. DNA sequence analysis of these isolates was consistent with horizontal gene transfer between bacterial species. Furthermore, pulsed-field gel electrophoresis of the GIM-1-positive P. aeruginosa isolates showed one clonal cluster, involving isolates that spanned from 2002 to 2012, which was most often found in surgical intensive care unit (ICU) patients (19 of 24 patients) (28). Reports such as these argue that nosocomial spread of resistant organisms may be occurring and that the persistence of carbapenemases, in particular MBLs, in nonfermenting Gram-negative organisms can lead to the transfer of antimicrobial resistance to other Gram-negative bacteria.

Indeed, epidemiologic studies from nosocomial outbreaks of CR-NF have consistently provided evidence that significant monoclonal or oligoclonal outbreaks are occurring. For example, a 2-year study from Algeria described a clonal outbreak of carbapenem-resistant P. aeruginosa strains in hospitalized patients (29). Similarly, a report from Brazil in 2005 described a clonal nosocomial outbreak of MBL-producing P. aeruginosa isolates, which represented 77.1% of the carbapenem-resistant isolates identified during the study period. Given the clonality seen by molecular analysis, we concluded that MBL-producing Pseudomonas was being spread by horizontal transmission between patients (30). Finally, a surveillance study of A. baumannii isolates from multiple hospitals in New York City in 2000 and 2001 reported a rate of carbapenem resistance of approximately 60%. Importantly, 82% of the resistant isolates were made up of only 2 strains. The primary resistance mechanism appeared to be class C cephalosporinase expression (31). Importantly, both the studies in Brazil and Algeria reported a significant increase in mortality for patients who were positive for carbapenem-resistant isolates.

WHO SHOULD BE SCREENED FOR CR-NF COLONIZATION?

Although routine screening is not currently recommended for CR-NF or CP-NF, their increasing prevalence raises the question: what are the costs of not screening? These costs may be greatest in high-risk populations, such as ICU patients or for patient populations at greatest risk of persistent infection, for example, in patients with chronic lung disease. It is well documented among high-risk populations, such as oncology, transplant, and ICU patients, that colonization with CROs may lead to serious infections with these MDR organisms, most notably health care-acquired pneumonia, and it is associated with increased morbidity and mortality (6, 32–36). Understanding CRO and CPO colonization rates among these high-risk patient populations is important for the adoption of appropriate infections control measures to limit the spread in these populations and to allow for early and targeted therapeutic strategies (33).

In patients with chronic lung disease, data have shown that those with sequentially positive respiratory cultures tend to be colonized by clonal strains of multidrug-resistant Pseudomonas spp. and that the risk of colonization increases with disease severity and increased antibiotic use. Unfortunately, chronic lung disease patients may be the most difficult population to treat, as it is difficult to eradicate Pseudomonas species colonization in patients with abnormal airway clearance (37).

Most worrisome are reports of clonal outbreaks of P. aeruginosa between persistently colonized cystic fibrosis patients (38–40). As a result, some centers now spatially segregate patients colonized with clonal strains, which has proven effective at preventing transmission (41).

Currently, there are limited data on the efficacy of widespread CR-NF or CP-NF screening protocols. One surveillance study in Brazil saw an 80% decrease in the percentage of cultures positive for Acinetobacter spp. with carbapenem resistance but no decrease in the rate of carbapenem resistance in P. aeruginosa isolates (42). More studies on the clinical efficacy of screening protocols in areas with MBL-positive and MBL-negative drug-resistant isolates are urgently needed.

WHERE TO SCREEN PATIENTS FOR CR-NF COLONIZATION?

CRE primarily colonize the gastrointestinal tract, so screening is typically performed by rectal swab surveillance cultures. Some screening methods, such as the CDC broth enrichment method, are designed to select for lactose fermenters only, and CR-NF isolates are not routinely identified (43). A variety of additional screening methods have been developed to detect CRE from rectal swabs, most of which can be modified to test for CR-NF (44). However, it is not known if rectal swabs are the best method to screen for colonization for organisms, like Pseudomonas and Acinetobacter spp., since these organisms primarily colonize the respiratory tract. Nonetheless, CR-NF have been shown to colonize the intestinal tract along with CRE, which can promote plasmid transfer between organisms in the health care setting, especially in the presence of selective pressures. Both CP-NF and CPE can serve as reservoirs and vectors for transmission of carbapenemases. Thus, rectal surveillance cultures must be able to detect all CPOs to limit the spread of these pathogens.

A study focused on determining the best site for the detection of MDR A. baumannii carriage found no statistical difference in the detected rates of carriage between testing the nostrils, pharynx, skin, or rectum for colonization; however, the highest overall carriage rate was observed when multiple sites were sampled (45). In contrast, a study evaluating CHROMagar Acinetobacter medium demonstrated a higher detection rate for MDR Acinetobacter species colonization using nasal swabs than that with rectal swabs (18.5% versus 11%, respectively). Similarly, detection rates increased even further (22.7%) when using a combination of nasal and rectal swabs (46). Importantly, though, there is still controversy about which sites to culture and how many sites should be cultured to screen asymptomatic patients for CR-NF colonization. More studies are urgently needed to address this question.

WHY SCREEN FOR CR-NF COLONIZATION?

Most importantly, why should health care facilities consider screening for CR-NF? There are multiple reasons highlighted above, including that they cause significant morbidity and mortality in high-risk patient populations, they cause the highest burden of carbapenem resistance among Gram-negative bacilli, and last, they serve as a reservoir and vectors for carbapenemase transmission in the health care setting. Two different scenarios may present in which screening of CR-NF may be warranted, including screening of high-risk patients for colonization of CR-NF or rectal screening of carbapenemase-producing organisms.

METHODS FOR DETECTING CR-NF AND CP-NF

Based on methodology, assays can be divided into growth-based phenotypic, rapid colorimetric assays, and nonphenotypic-based assays (e.g., molecular methods). Each assay varies in its turnaround time, sensitivity, cost, and information provided.

GROWTH-BASED PHENOTYPIC ASSAYS FOR CR-NF

Growth-based phenotypic assays measure resistance based on growth in the presence of an antibiotic (Table 2). These assays can be further divided into screening methods for detecting CRO colonization and methods for detecting carbapenemase production from cultured isolates.

TABLE 2.

Growth-based phenotypic assays: carbapenem-resistant organism screening and carbapenemase detection methods

Method (reference)	Procedurea	Interpretation	Result	Costb	Notesc
Screening methods for detection of colonization by CR-NF
CDC broth enrichment method	Broth enrichment with a carbapenem disk, then subculture to a MacConkey agar followed by ID and AST	Presence or absence of CRO colonization	Detection of lactose fermenters (modification needed to identify non-lactose-fermenting CROs)	+	False-positive broths (no growth on subculture) and long TAT of up to 4 days
Carbapenem disks on MacConkey agar (47)	Direct plating of rectal swabs on a MacConkey plate and streaked into 4 quadrants. A carbapenem disk is added between the juncture of quadrants 1 and 2 and juncture of quadrants 2 and 3. Two disks are added to maximize the likelihood of obtaining a measurable zone	Growth within ≤24 mm of ertapenem, ≤34 mm of meropenem, or ≤32 mm for imipenem disks	Detection of carbapenem-resistant Gram-negative bacilli	+	TAT, 24–48 h; uses standard laboratory medium without need for chromogenic medium
CHROMagar Acinetobacter and modified CHROMagar Acinetobacter (inclusion of an antimicrobial supplement) (46)	Inoculate media with specimen to detect Acinetobacter spp. or multidrug-resistant Acinetobacter spp. (modified version)	Growth and colony color	Detection of Acinetobacter spp. or MDR Acinetobacter spp., depending on medium	++	TAT, 24–48 h; ID/AST required to confirm organism ID and carbapenem resistance
CHROMagar Pseudomonas (74)	Primary sample plated onto modified carbapenem-containing agar plate	Growth and colony color	Colonies of Pseudomonas are blue; other organisms are inhibited or colorless	++	TAT, 24–48 h; AST profile required to confirm carbapenem resistance
ESBL/CRE chromogenic agar (48)d	Primary sample plated onto cephalosporin (ESBL)- or carbapenem (CRE)-containing chromogenic medium	Growth and colony color	Pseudomonas and Acinetobacter appear colorless or natural colored	++	TAT, 24–48 h; AST profile required to confirm carbapenem resistance
Culture-based methods for detection of carbapenemase producers
Double-disk synergy (49)	EDTA-only disk and EDTA-free meropenem disk placed 10 mm apart on agar streaked with isolate	Synergistic growth inhibition between disks	Specifically detects MBL	+	TAT, 24 h; considered 100% sensitive for MBL
Combined EDTA-disk synergy (75)	EDTA disk with imipenem and EDTA-free imipenem disk placed 20 mm apart on agar streaked with isolate	Relative increase in growth inhibition without EDTA (>7 mm)	Specifically detects MBL	+	TAT, 24 h
Boric acid disk potentiation (51)	APB added to an ertapenem or meropenem disk; also APB-free control disk placed on agar streaked with isolate	Increase in growth inhibition around APB disk (>5 mm) is considered positive	Inhibition of KPC-type carbapenemase by APB	+	TAT, 24 h; can detect KPC-type β-lactamase
MBL Etest (76)	Carbapenem antibiotic gradient strip (with or without EDTA at either end of the strip) applied to agar streaked with isolate	MIC ratio of carbapenem over carbapenem/EDTA or presence of a phantom zone	Detects MBL and MIC of the carbapenem	++	TAT, 24 h
Carbapenem inactivation method (50)	Water/broth inoculation of the organism of interest in the presence of a carbapenem disk for 2 h. The disk is then placed onto a nonselective agar streaked with a carbapenem-susceptible E. coli strain	No zone of inhibition: positive for carbapenemase production; inhibition of growth: negative for carbapenemase production	Detects carbapenemase production	+	TAT, as little as 8 h
Modified Hodge test (49)	Susceptible E. coli strain plated, meropenem disk in center, linear streaks of test isolate away from disk	E. coli growth vs no growth adjacent to the streaked isolate, near the disk	Detects carbapenemase production	+	CLSI does not recommend the MHT for CR-NF; lacks sensitivity for certain MBLs/OXA/SME and produces false-positive results for ESBL/AmpC producers plus porin mutations; TAT, 24 h

^aID, identification; AST, antimicrobial susceptibility testing; APB, 3-aminophenyl boronic acid.

^b+, <$5; ++, $5 to 50; +++, $50 to 100; ++++, >$100.

^cTAT, turnaround time.

^dThere are many non-FDA-cleared chromogenic media available on the market, including CHROMagar KPC (CHROMagar, France), HardyCHROM (Hardy Diagnostics, CA, USA), chromID Carba (bioMérieux, France), chromID ESBL (bioMérieux, France), Colorex KPC (Biomed Diagnostics, OR, USA), RambaChrom KPC (Gibson Bioscience, USA), SpectraCRE (Thermo Diagnostics, USA), Brilliance CRE (Thermo Diagnostics), Colorex KPC (powder-base), and Supercarba agar (patented recipe). These media were recently reviewed by Queenan and Bush (17).

Screening methods for detecting CRO colonization among clinical specimens.

A clinically effective screening test for CROs obtained directly from specimens should be relatively rapid, accurate, and cost-effective. For broad surveillance, the assay must be sensitive enough to detect isolates with various mechanisms of resistance and identify low organism burden, particularly among rectal swabs. If resources do not exist for broad screening methods, surveillance should be considered for the most epidemiologically important and prevalent resistance mechanisms. Screening methods to detect CRO colonization must have a fast enough turnaround time to prevent patient-to-patient and patient-to-environment transmission of highly drug-resistant microorganisms. Alternatively, they need to be rapid enough to remove patients who are preemptively placed on isolation due to predetermined risk factors and who are found not to be colonized with CROs. These methods only detect carbapenem resistance and do not provide mechanistic information. Perhaps the most common and cost-effective approach to screening is the CDC broth enrichment method. Although originally described as a method to isolate only lactose-fermenting CRE, it can be modified to select for all CROs. The major disadvantage to the CDC broth enrichment method is that false-positive broth cultures occur regularly, without isolation of a Gram-negative organism on subculture to solid medium. Another significant disadvantage is the long turnaround time of up to 4 days. Additionally, although the broth enrichment was designed to increase sensitivity, it may in fact hinder the recovery of CROs with low carbapenem MICs or low inocula in the presence of high inocula of competing organisms (15, 81). A second method using standard laboratory medium involves direct inoculation of a MacConkey agar plate and the placement of carbapenem disks between the juncture of quadrants 1 and 2 and juncture of quadrants 2 and 3. Defined zone diameters for screening CROs were established using spiked stool specimens; however, clinical studies demonstrating the utility of this method for the detection of CR-NF from clinical specimens are still required. The method does have the advantage of a 24- to 48-h turnaround time (TAT) compared to the CDC broth enrichment method (47).

Commercial chromogenic agar supplemented with carbapenem antibiotics is widely used for rectal screening for CRE, with decreased turnaround times compared to the CDC method (24 h versus up to 4 days, respectively). These media can also be used for the detection of CR-NF, which appear colorless or natural colored on extended-spectrum β-lactamase (ESBL)/CRE chromogenic agar (examples: CHROMagar KPC [CHROMagar, France], HardyCHROM [Hardy Diagnostics, CA, USA], chromID Carba [bioMérieux, France], chromID ESBL [bioMérieux, France], Colorex KPC [Biomed Diagnostics, OR, USA], and Brilliance CRE [Thermo Diagnostics, USA]).

Specific medium types have been designed to detect both Acinetobacter and Pseudomonas spp. and could be used to detect colonization with these organisms (CHROMagar Acinetobacter, modified CHROMagar Acinetobacter, CHROMagar Pseudomonas, and cetrimide agar). Cetrimide agar is selective for P. aeruginosa. The cetrimide in the medium inhibits normal flora while enhancing P. aeruginosa pigment production. The chromogenic medium with associated color changes facilitates organism identification, even at low colony counts. Some data suggest that directly inoculating chromogenic agar from rectal swabs may compromise the sensitivity for detecting CRE (48). However, other studies have found that broth enrichment is not required and that direct inoculation of rectal swabs on chromogenic medium is comparable or superior to that in the CDC broth enrichment method (15, 81).

Detection of carbapenemases among carbapenem-resistant cultured isolates.

After the recovery of CROs from surveillance or clinical cultures, the next step is to determine if the CROs are carbapenemase producing. Many phenotypic culture-based methods have been described to elucidate the mechanism of carbapenemase resistance. The modified Hodge test (MHT) relies on the ability of CPOs to decrease the local concentration of carbapenems, allowing carbapenem-susceptible Escherichia coli strains to grow. The MHT was the original method used to confirm carbapenemase production among CRE. However, the MHT has low sensitivity for certain carbapenemase types (MBL, some OXA types, and SME) and lacks specificity, since ESBL and AmpC producers with porin mutations can give false-positive results (49). Recently, a method similar to the MHT, the carbapenem inactivation method, was described and involves suspending a loopful of bacteria in water containing a meropenem disk for 2 h. If carbapenemases are present, they hydrolyze meropenem within the disk. The disk is then placed on nonselective agar that had been streaked with carbapenem-susceptible E. coli. Noninhibition of E. coli growth indicates that carbapenem hydrolysis occurred and the organism of interest produces carbapenemases. The carbapenem inactivation method is comparable in terms of cost to the MHT and may provide same-day results with a turnaround time of as little as 8 h (50).

Inhibitor-based methods can be used to differentiate certain carbapenemases from others by applying Ambler class-specific inhibitors. Both the double-disk synergy test and the combined EDTA disk test rely upon decreased MBL activity in the presence of EDTA. Finally, the boric acid disk potentiation assay relies on the sensitivity of KPC carbapenemases to boric acid (51). To our knowledge, there are no data on the performance of the boric acid test using CR-NF isolates. A similar approach is being utilized by Liofilchem Biotechnology (Roseto degli Abruzzi, Italy), which is developing two strip-based assays for KPC and MBL detection using boric acid and EDTA, respectively.

Not all growth-based assays have the same sensitivity for detecting MBLs. In one study from a collection of P. aeruginosa clinical isolates positive for meropenem resistance by the Kirby-Bauer method, 52% were positive by an EDTA-disk synergy test, while only 34% were positive by MHT. Similar results were obtained from isolates of meropenem-resistant Acinetobacter species (52). Consistent with these results, a study from Iraq in 2014 showed that the combined EDTA-disk synergy test had positivity rates roughly equivalent to the double-disk synergy test for detecting carbapenem-resistant Gram-negative bacteria (∼30%). However, the study reported lower positivity rates for the MHT (21%) and Etest (1%) (53). Accordingly, the MHT is endorsed by the CLSI for CRE but not for CR-NF. Finally, there have been reports of false-positive results using EDTA-based MBL assays, in particular, the Etest strips (AB Biodisk, Solna, Sweden) in the setting of certain OXA-type carbapenemases. It has been shown that OXA-10 and OXA-14 exist in highly active dimeric forms and less-active monomeric forms (54). The same phenomenon has been hypothesized to occur among the OXA-23 group of enzymes in A. baumannii (55). Metal ions, such as Zn²⁺ or Ca²⁺, are thought to stabilize the highly active dimeric form of the enzyme. Thus, in the presence of a metal chelator, such as EDTA, it converts the OXA-type enzymes to the less-active monomeric state, resulting in a reduction in carbapenemase activity and a false-positive MBL Etest result. Thus, MBL Etest strip results should be interpreted with caution among organisms that are known to produce OXA-type carbapenemases, such as A. baumannii and P. aeruginosa.

RAPID COLORIMETRIC CARBAPENEMASE ASSAYS

Recently, rapid colorimetric tests have been developed to detect carbapenemases directly from cultured bacterial isolates within 2 h. These tests rely upon the detection of a pH change produced by the hydrolysis of imipenem by carbapenemases using a pH indicator (i.e., phenol red or bromothymol blue). They can distinguish carbapenemase production from other carbapenem resistance mechanisms and are capable of detecting all Ambler classes of carbapenemases, including class A (i.e., KPC and GES), class B MBL (i.e., VIM, IMP, and NDM), and class D OXA-type (i.e., OXA-23, OXA-40, and OXA-48) carbapenemases.

The original colorimetric assay, the Carba NP test, is a manual method that requires frequent reagent preparation because the imipenem-containing solution has a maximum shelf life of 72 h. The Carba NP test is endorsed by both the CLSI and EUCAST for the detection of carbapenemase producers (56). It can be performed on CRE and CR-NF isolates that are nonsusceptible to one or more carbapenems (20). A recent report demonstrated the use of the Carba NP test to detect carbapenemase production in Pseudomonas spp. directly from positive blood cultures, and it had excellent sensitivity and specificity (57). Many modifications to the manual Carba NP assay have been described. The Carba NP test II is similar to the Carba NP test but performs multiple reactions on each isolate. Amber class A, B, and D carbapenemases are identified by their inhibition in the presence of β-lactamase inhibitors, such as tazobactam, EDTA, and dipicolinic acid (DPA), as well as their ability to hydrolyze monobactams (58). Thus, carbapenemases can be correctly classified with clinical sensitivity and specificity comparable to those of molecular methods. The Carba NP test II was shown to be effective for screening CRE and Pseudomonas species. However, to better detect the slower hydrolysis of imipenem by the most frequently encountered class D OXA-type carbapenemase (i.e., OXA-23, OXA-40, OXA-58, and OXA-143) in Acinetobacter spp., the extraction procedure had to be improved. Thus, the CarbAcineto NP test uses higher inocula of bacteria and a lysis buffer of 5 M NaCl, which increased the sensitivity of the assay to 95% for detecting OXA-type carbapenemases. Notably, the assay is unable to detect overexpression of the chromosomally encoded OXA-51-like β-lactamases or GES produced by Acinetobacter species (16). The extraction method was further modified to produce a single extraction reagent that could be used for CRE, Pseudomonas, and Acinetobacter species. The resulting modified Carba NP assay uses 0.02% cetyltrimethylammonium bromide for extraction and a starting pH of 7.5 (previously 7.8) (59). The modified Carba NP assay, performed on a variety of carbapenemase producers (including OXA-23 and OXA-40) among both CPE and CP-NF, has a reported sensitivity and specificity of 100% (59). Last, the Blue-Carba assay is a version of the Carba NP test that utilizes bromothymol blue as a pH indicator, a lower starting pH, and no extraction step. The Blue-Carba assay has demonstrated 100% sensitivity and specificity for CP-NF, including a variety of OXA-producing Acinetobacter species (60).

Since the introduction of the manual Carba NP assay, multiple commercial versions have been introduced: the Neo-Rapid Carb screen (formerly the Rosco Rapid CARB screen; Rosco Diagnostics, Denmark), Rapid CARB Blue kit (Rosco Diagnostics), and the Rapidec Carba NP (marketed by bioMérieux) (61). The commercial non-Food and Drug Administration (FDA)-cleared versions of these assays do not require reagent preparation, since the reagents are lyophilized and resuspended upon testing (the Neo-Rapid Carb/Blue-Carba tests use tablets, and the Rapidec Carba NP test uses wells). The Rapidec Carba NP assay is currently in clinical trials to obtain FDA clearance.

Overall, the rapid colorimetric assays perform well with CP-NF. The main disadvantage of these assays is the poor detection of some OXA-type and chromosomally encoded carbapenemases. Also, interpretation may be subjective in nature, and the color change can vary depending on the carbapenemase type. Furthermore, isolates taken from chromogen-containing growth medium, MacConkey, or Drigalski agar can interfere with interpretation of these colorimetric assays and may require subculture to a nonselective medium prior to testing (62).

NONPHENOTYPIC-BASED METHODS FOR DETECTION OF CR-NF AND CP-NF

Many nonphenotypic-based methods have been developed for the detection of CROs/CPOs, including nucleic amplification methods, microarrays, matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS), UV spectrometry, and next-generation sequencing (Table 3) (63).

TABLE 3.

Non-phenotypic-based methods for detection of CR-NF and CP-NF

Method	Procedure	Product(s)	Costa	Relevance/commentsb
Nucleic acid detection-based molecular assays
PCR	Primer/probes specifically designed to detect resistance genes; may also identify organism	BioFire FilmArray: KPC only, FDA cleared; Cepheid GeneXpert Carba-R: KPC, NDM, OXA-48, IMP-1, VIM, FDA cleared for use on cultured isolates; BD Max (BD Diagnostics, Sparks, MD): KPC, NDM, OXA-48; Check-Points Check-Direct assay: KPC, NDM, OXA-48-like, VIM; laboratory-developed tests (58)	++ to +++	Used directly from specimens and from cultured isolates; targets vary based on the assay; same-day TAT
Loop-mediated isothermal amplification (LAMP) (77)	Isothermal amplification using sequence-specific primers	Laboratory-developed tests	++	Cheaper and less sensitive than other NAAT; better for resource-poor settings
Nanoparticle Probe and microarray (78)	Direct detection of nucleic acids with signal amplification	Verigene BC-GN; KPC, NDM, OXA, IMP, VIM	++	FDA cleared
Oligonucleotide microarray (79)	PCR amplification of extracted DNA with labeled probes, then hybridization to microarray	Check-Points microarray assays	+++	Requires PCR amplification step prior to hybridization
Next-generation sequencing (64)	Massively paralleled target sequencing or de novo assembly, with or without preamplification	Laboratory-developed tests	++++	Sequence information can cover the entire genome or specific targeted sequencing can be applied; cost can be decreased by running multiple samples on a single chip/run; costly instrumentation
MALDI-TOF MS assays
Hydrolysis method (67)	Incubation with carbapenem (∼4 h), followed by mass spectrometry-based analysis of drug and metabolites	Laboratory-developed tests	+	Does not detect resistance due to noncarbapenemase mechanisms; rapid TAT (h)
Peak identification method (70, 71)	Detection of a carbapenemase-bearing plasmid-associated peak	Laboratory-developed tests	+	Simultaneous organism and carbapenemase-bearing plasmid identification within 10 min from isolated colonies or 30 min from spiked blood cultures; this method has only been established for the blaKPC carbapenemase-bearing pKpQIL plasmid
Antigen detection assay
Immunochromatographic (80)	Rapid disposable lateral flow assay with visual readout	Coris BioConcept	+	Antigenic sites on carbapenemase-specific proteins are detected; TAT, 15 min; easy to use; currently, few target proteins detected; no data on CR-NF isolates
Hydrolysis assay
UV spectrophotometry (69)	Detects hydrolysis of imipenem by a cell lysate	Laboratory-developed tests	+	Differentiates carbapenemase from non-carbapenemase producers; no data on CR-NF isolates
Rapid colorimetric assay (59)	Imipenem hydrolysis causes a pH change, causing a color change due to pH indicator	See text for further details	+ to ++	Carba NP endorsed by CLSI and EUCAST; TAT, 2 h; both manual and commercial versions available

^a+, <$5; ++, $5 to 50; +++, $50 to 100; ++++, >$100.

^bTAT, turnaround time; NAAT, nucleic acid amplification test.

Carbapenemase-encoding genes can be detected by a number of nucleic acid-based techniques, both as a screening tool used directly from specimens and from cultured isolates (Table 2). The major advantage of amplification-based methods is their high analytic sensitivity. However, primer-based amplification methods require a priori knowledge of the nucleic acid targets; thus, their clinical sensitivity drops if mutations occur in primer or probe binding regions or if the resistance genes present in an isolate are not targets of the assay. Commercial multiplex assays are available, albeit at a relatively high cost. Similarly, nonamplification-based methods, like the Nanosphere (Chicago, IL, USA) and direct next-generation sequencing, also rely upon the presence of known resistance-defining nucleic acid sequences (64). The greatest disadvantage to using these methods in a widespread screening program is cost; however, they can provide detailed information about the molecular mechanism of resistance and possibly the phylogeny of the isolate.

The use of MALDI-TOF MS to detect antibiotic resistance typically involves incubation of a cultured isolate with an antibiotic and then use of mass spectrometry to identify the presence of the drug as well as drug metabolites if hydrolysis of the agent occurs. MALDI-TOF MS methods have shown lower sensitivity (∼75% sensitivity) due to false-negative results with OXA-48-type carbapenemases. The addition of NH₄HCO₃ to the reaction buffer has been shown to enhance the sensitivity of the assay to 98%. This approach using imipenem has been used to detect carbapenemase activity in Acinetobacter spp. with 100% sensitivity and specificity (65). Similar tests can be performed using liquid chromatography-mass spectrometry, but MALDI-TOF MS has become relatively widespread in clinical laboratories (66). The TAT of a MALDI-TOF MS-based screen can be as short as 1 h, with a cost per sample close to $1. Like most nonphenotypic-based methods, the major disadvantage of this method is that it does not detect resistance due to mechanisms other than carbapenemases (67). However, this method can be applied to carbapenem-resistant Acinetobacter spp. to detect carbapenemase production, and carbapenemase type can be determined by adding class-specific inhibitors (68). Finally, a similar approach has been developed that uses UV spectroscopy to monitor the hydrolysis of imipenem directly in cell lysates (69).

A second MALDI-TOF MS method has been developed for the detection of a carbapenemase-bearing plasmid-associated protein peak. So far, this method has only been established for the bla_KPC carbapenemase-bearing pKpQIL plasmid that was responsible for the CRE outbreak that occurred at the NIH Clinical Center in 2011. The principal advantage of this method is that it is inexpensive (if a MALDI-TOF instrument is already present), and it provides simultaneous organism and carbapenemase-bearing plasmid identification within 10 min from isolated colonies or 30 min from the time of automated detection of growth in spiked blood cultures (70, 71). However, much more work is needed before it can be used generally, since it would need to detect many more carbapenemases. Nevertheless, the ability to track plasmids bearing carbapenemase genes both cost-effectively and rapidly may be sufficient for implementation in an outbreak and for monitoring the presence of the plasmid in the postoutbreak setting.

Immunochromatographic methods are currently under development to detect various carbapenemases and ESBLs from cultured isolates with simple affinity-based colorimetric assays, similar to a home pregnancy test (72). If these products produce good sensitivity and specificity, they may represent an attractive means of testing for specific carbapenemases. The upfront cost to implementation is low, and the tests are relatively easy to use. No studies have described the performance of these assays in CR-NF isolates.

CONCLUSION

Non-glucose-fermenting Gram-negative bacilli are increasingly acquiring carbapenem resistance. Given their intrinsic antibiotic resistance, this can lead to infections that are very difficult to treat. Additionally, CR-NF have been shown to colonize the gastrointestinal tract along with CPE, which can lead to plasmid-mediated resistance gene transfer. Thus, infection control practitioners and clinical microbiologists need to work together to determine the risk that CR-NF carry in their institution and what methods should be considered for the surveillance and detection of CR-NF/CP-NF. As part of this response, clinical microbiology laboratories may need to employ effective surveillance assays that will help guide infection control efforts.

A variety of detection methods have been developed to detect CRE/CPE. Some of these can be or have been further modified for use in non-glucose-fermenting Gram-negative bacilli. These methods include growth-based phenotypic methods, rapid colorimetric methods, nonphenotypic methods, and a combination of these as described above. Cost, turnaround time, and analytical sensitivity are important considerations; however, the pros and cons of each assay also depend on the types of resistance mechanisms encountered, the prevalence of resistant organisms, the patient population, the instruments already present in the laboratory, and the mechanistic information desired from the assay.

Like many drug resistance threats that emerged in previous decades, these CROs/CPOs and their transferrable resistance genes are likely here to stay. Given our current focus on CRE screening, it is troubling to see some studies already demonstrating higher rates of CR-NF than CRE colonization (73). Thus, we need to move beyond a CRE-focused approach to screening and also consider CR-NF in order to further manage the threat posed by CROs.

Biography

Dr. Patricia (Trish) J. Simner is the Director of Medical Bacteriology and Parasitology Laboratories at the Johns Hopkins Medical Institutions and Assistant Professor of Pathology at the Johns Hopkins University, Baltimore, MD. Dr. Simner received her Ph.D. in Medical Microbiology and Infectious Diseases at the University of Manitoba, Winnipeg, Manitoba, Canada. Her Ph.D. thesis was on antimicrobial resistance mechanisms among Gram-negative bacteria, specifically extended-spectrum β-lactamase producers. She completed her postdoctoral training in clinical microbiology at the Mayo School of Graduate Medical Education in Rochester, MN, and is certified by the American Board of Medical Microbiology. Dr. Simner's current research interests focus on understanding the mechanisms of antimicrobial resistance and spread of carbapenemase-producing organisms in the hospital setting and on the evaluation and development of novel diagnostic tools.