Laboratory detection of carbapenemases among Gram-negative organisms

SUMMARY

The carbapenems remain some of the most effective options available for treating patients with serious infections due to Gram-negative bacteria. Carbapenemases are enzymes that hydrolyze carbapenems and are the primary method driving carbapenem resistance globally. Detection of carbapenemases is required for patient management, the rapid implementation of infection prevention and control (IP&C) protocols, and for epidemiologic purposes. Therefore, clinical and public health microbiology laboratories must be able to detect and report carbapenemases among predominant Gram-negative organisms from both cultured isolates and direct from clinical specimens for treatment and surveillance purposes. There is not a “one size fits all” laboratory approach for the detection of bacteria with carbapenemases, and institutions need to determine what fits best with the goals of their antimicrobial stewardship and IP&C programs. Luckily, there are several options and approaches available for clinical laboratories to choose methods that best suits their individual needs. A laboratory approach to detect carbapenemases among bacterial isolates consists of two steps, namely a screening process (e.g., not susceptible to ertapenem, meropenem, and/or imipenem), followed by a confirmation test (i.e., phenotypic, genotypic or proteomic methods) for the presence of a carbapenemase. Direct from specimen testing for the most common carbapenemases generally involves detection via rapid, molecular approaches. The aim of this article is to provide brief overviews on Gram-negative bacteria carbapenem-resistant definitions, types of carbapenemases, global epidemiology, and then describe in detail the laboratory methods for the detection of carbapenemases among Gram-negative bacteria. We will specifically focus on the Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter baumannii complex.

INTRODUCTION

The global spread of antimicrobial-resistant organisms (AROs) is one of the top 10 overall threats to human health during the 21st century (1). The development of antimicrobial resistance (AMR) is an inherent natural process, and AROs are selected over time by the use and misuse of antimicrobial agents (2). Infections with AROs will often not respond to antimicrobial agents to which they were previously susceptible to, making such infections difficult to treat successfully. This leads to infections associated with increased patient mortality and morbidity due to the delayed administration of appropriate agents (3, 4). Patients with ARO infections have prolonged hospital stays with subsequent increases in health-care related costs (5). In 2019, an estimated 4·95 million deaths worldwide were associated with bacterial AMR, including 1·27 million deaths directly attributable to bacterial AMR (6). AMR Gram-negative bacteria, most notably Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Acinetobacter baumannii complex were among the six most common AMR pathogens identified in the 2019 report. There is a global dearth of effective antibiotics for treating Gram-negative ARO infections, and limited financial incentives exist for pharmaceutical companies to invest and develop of new agents (7). There is an urgent need to preserve the world’s remaining antibiotics for future generations, especially those drugs with activities against AROs, that are efficient and cost-effective (8). For this reason, the World Health Organization (WHO) has created a list of critically important antimicrobials requiring judicious use as part of a One Health approach (9).

Carbapenems are some of the most effective options available for treating serious infections caused by Gram-negative AROs (10). Therefore, the emergence and spread of carbapenem resistance are significant public health concerns (11, 12). The WHO identified carbapenem-resistant Gram-negative bacteria (i.e., A. baumannii complex, P. aeruginosa, and Enterobacterales) as critical-priority pathogens (7). Carbapenem-resistant E. coli, K. pneumoniae, P. aeruginosa, and A. baumannii complex were directly responsible for 181,000 deaths and played a role in 910,000 deaths during 2019 (6).

Resistance to the carbapenems involves different underlying mechanisms, and the most clinically relevant in Gram-negative bacteria are summarized as follows (13):

1. Changes in outer membrane permeability that are mediated by the decreased expression or loss of certain outer membrane proteins. For example, in K. pneumoniae, modifications to the porin OmpK36 contributes to carbapenem resistance (14), or the closure of the porin OprD in P. aeruginosa leads to the reduced entry of imipenem into the periplasmic space resulting in not susceptible results (i.e., intermediate or resistant) to this carbapenem.

2. The upregulation of membrane efflux pumps that can “pump out” antimicrobial agents from the bacterial periplasmic space to the external environment. For example, the elevated expression of the MexA-MexB-OprD pump system in P. aeruginosa can diminish the periplasm levels of several antimicrobial agents, including the carbapenems, such as meropenem. In E. coli, the AcrB resistance-nodulation-division (RND) efflux pump plays a role in promoting multidrug resistance, including carbapenem resistance (15).

   Changes in outer membrane permeability (i) and/or upregulation of efflux (ii) leading to carbapenem resistance often occur in conjunction with non-carbapenem hydrolyzing beta-lactamase production, such as extended-spectrum beta-lactamases (ESBL) or AmpC beta-lactamases.

3. The production of β-lactamases, named carbapenemases, that bind to and inactivate the carbapenems through hydrolysis. The production of carbapenemases is the most common and most important underlying carbapenem resistance mechanism among Gram-negative bacteria (16). This is because carbapenemase genes are spread through mobile genetic elements (e.g., insertion sequences, transposons) and can be transferred between different Gram-negative bacteria through conjugative plasmids; this is especially relevant among members of the Enterobacterales (12, 17, 18).

Overall, health-care workers play pivotal roles in limiting and preventing the spread of AMR infections (19). Such measures include the judicious and targeted use of antimicrobial agents and the rapid implementation of infection prevention and control (IP&C) measures. The clinical microbiology laboratory is critical in limiting ARO spread because it functions as an early alerting system, notifying the medical community to new and existing antimicrobial mechanisms present in clinically important bacteria guiding patient management (20). The aggregate data are also used to guide public health initiatives. Therefore, it is essential that clinical laboratories have the capabilities to detect and report carbapenemase-producing Enterobacterales, carbapenemase-producing P. aeruginosa, and carbapenemase-producing Acinetobacter species on cultured bacterial isolates and directly on human specimens (21, 22).

The aim of this article is to provide brief overviews on Gram-negative bacteria carbapenem-resistant definitions, types of carbapenemases, global epidemiology, and then describe in detail the laboratory methods for the detection of carbapenemases among Gram-negative bacteria. We will specifically focus on A. baumannii complex, P. aeruginosa, and Enterobacterales.

DEFINITIONS: CARBAPENEM-RESISTANT AND CARBAPENEMASE-PRODUCING GRAM-NEGATIVE BACTERIA

We will use several acronyms in this manuscript to describe bacteria that are carbapenem resistant (Table 1). CROs are defined by resistance to carbapenems and/or the identification of a CP within a bacterium (23). The carbapenems that are used to define carbapenem resistance among Gram-negative bacteria differ based on whether the organism is known to be intrinsically resistant to ertapenem or not. The Clinical and Laboratory Standards Institute (CLSI) defines intrinsic resistance as antimicrobial resistance that is not acquired, it is inherent or innate which is reflected in wild-type antimicrobial resistance patterns of all or almost all representatives of a species (e.g., >97% of the wild type tests resistant) (24). For organisms that are not intrinsically resistant to ertapenem, such as the Enterobacterales and Aeromonas species, resistance to the following carbapenems (i.e., ertapenem, meropenem, imipenem, and/or doripenem) may be used to define carbapenem resistance. Proteus, Morganella, and Providencia species have intrinsically elevated minimum inhibitory concentrations (MICs) to imipenem by mechanisms other than production of carbapenemases; therefore, resistance to imipenem is not included in the definition for the subset of Enterobacterales (24). For glucose-non-fermenting Gram-negatives, such as P. aeruginosa or A. baumannii complex, carbapenem resistance is defined based on resistance to antipseudomonal carbapenems (i.e., meropenem, imipenem, and/or doripenem; https://arpsp.cdc.gov/profile/arln/crpa). P. aeruginosa is inherently resistant to ertapenem due to the bulky R3 moiety that results in a weak affinity for PBP3 (25). Doripenem remains discontinued within the United States (US) and is rarely used to define CRO within clinical laboratories within the US as it not a common agent on commercial automated antimicrobial susceptibility testing (AST) panels.

TABLE 1.

Terms and definitions used to describe carbapenem-resistant Gram-negative organisms

Term	Definition
Carbapenem-resistant organism (CRO)	An organism that is carbapenem resistant due to any mechanism, most commonly production of a carbapenemase, or a porin mutation or an efflux pump in conjunction with non-carbapenem hydrolyzing beta-lactamase production, such as extended-spectrum beta-lactamases (ESBLs) or AmpC beta-lactamases.
Carbapenem-resistant Enterobacterales (CRE)	Enterobacterales that test carbapenem resistant by any mechanism.
Carbapenem-resistant Pseudomonas aeruginosa (CRPA)	P. aeruginosa that test carbapenem resistant by any mechanism.
Carbapenem-resistant Acinetobacter baumannii (CRAB)	A. baumannii complex that test carbapenem resistant by any mechanism.
Carbapenemase (CP)	Refers to a β-lactamase enzyme that inactivates carbapenems through hydrolysis.
Carbapenemase-producing organism (CPO)	A Gram-negative organism that produces a carbapenemase.
Carbapenemase-producing Enterobacterales (CPE)	Enterobacterales that produces a carbapenemase.
CP-CRO or CP-CRE	A carbapenem-resistant organism (CRO) or carbapenem-resistant Enterobacterales (CRE) that is carbapenem resistant due to production of a carbapenemase.
Non-CP-CRO or Non-CP-CRE	A carbapenem-resistant organism (CRO) or carbapenem-resistant Enterobacterales (CRE) that is carbapenem resistant due to a mechanism other than production of a carbapenemase.

Appropriately defining and detecting CROs require clinical and public health laboratories to apply the current carbapenem clinical breakpoints to interpret AST results as defined by the CLSI and the European Committee on Antimicrobial Susceptibility testing (EUCAST) standards development organizations (SDOs) (24, 26) (Table 2). Most of the carbapenem breakpoints for the Enterobacterales, P. aeruginosa, and A. baumannii complex were revised and lowered by both SDOs in 2012 or earlier (Table 2). The evaluation of pharmacokinetic (PK)/pharmacodynamic (PD) data, limited clinical data and MIC distributions (including for CP producing isolates), indicated that the previous carbapenem breakpoints were set too high. The current, lowered CLSI and EUCAST breakpoints will capture the majority of CP producers as carbapenem MICs will test above the susceptible breakpoint (i.e., intermediate or resistant). Despite, these breakpoint changes occurring over a decade ago, laboratories in both the US and outside of the US continue to apply the older obsolete carbapenem breakpoints (27). The continued use of obsolete carbapenem breakpoints is a patient safety risk that can lead to ineffective carbapenem therapy. Obsolete carbapenem breakpoints also contributes to the spread of CROs due to the lack of IP&C measures (i.e., reporting CROs as carbapenem susceptible when they are actually not susceptible) (27, 28). Application of updated breakpoints to ensure detection of CRO applies to conventional AST methods performed from cultured isolates and for rapid phenotypic AST approaches performed directly from positive blood culture broths (e.g. rapid disk diffusion methods as described by CLSI/EUCUAST or novel methodologies) (24, 26, 29).

TABLE 2.

Current CLSI, FDA and EUCAST carbapenem breakpoints for Gram-negative organisms^a,b

	Ertapenem		Meropenem		Imipenem
	CLSI/FDA S/I/R (µg/mL)	EUCAST S/R (mg/L)	CLSI/FDA S/I/R (µg/mL)	EUCAST S/R (mg/L)	CLSI/FDA S/I/R (µg/mL)	EUCAST S/R (mg/L)
Enterobacterales	≤0.5/1/≥2	≤0.5/>0.5	≤1/2/≥4	≤2/>8c	≤1/2/≥4	≤2/>4d
P. aeruginosa	−	−	≤2/4/≥8	≤2/>8c	≤2/4/≥8	≤0.001/>4
Acinetobacter species	−	−	≤2/4/≥8	≤2/>8c	≤2/4/≥8	≤2/>4
Aeromonas spp.	≤0.5/1/≥2e		≤1/2/≥4e		≤1/2/≥4e
Other non-Enterobacteralesc	−	−	≤4/8/≥16e	−	≤4/8/≥16e	−

^a S: susceptible, I: intermediate, R: resistant, −: not applicable.

^b Breakpoints listed are based on CLSI M100-S34, CLSI M45-ED3:2016, FDA STIC (accessed May, 2024) and EUCAST breakpoints version 14.0 (24, 26, 30).

^c Indications other than meningitis. For meningitis, the breakpoint is S: ≤2 mg/L/R: >2 mg/L.

^d Except the Morganellaceae. For Morganellaceae, the breakpoint is S: ≤0.001 mg/L/R: >4 mg/L.

^e FDA does not recognize the CLSI carbapenem breakpoints for the Other Non-Enterobacterales and Aeromonas species.

The presence of a CP gene or production of CP in the absence of carbapenem resistance also qualifies an organism as a CRO (23). This may occur with bacteria that harbor CP genes but either do not express the gene or expression occurs constitutively at low levels. Alternatively, such genes require induction prior to expression, or the type of CPs (especially the OXA-48-like enzymes) have low hydrolytic activities against certain carbapenems. From a public health standpoint, these cases may serve as silent reservoirs for the spread of CP genes and in some patients can lead to treatment failures with the carbapenems. Therefore, the presence of a CP gene in the absence of carbapenem resistance by phenotypic AST is an important clinical finding.

CLASSIFICATION OF CARBAPENEMASES

Classification schemes

β-lactamases are classified using two different schemes: functional (into groups) and molecular (into classes). The Bush-–Jacoby-Medeiros (BJM) classification divides β-lactamases into groups by their functional properties according to rates of hydrolysis, substrate characteristics, and inhibitor profiles. According to the BJM scheme, β-lactamases fall into functional groups 1 to 4. CPs are part of BJM groups 2f and 3 (Table 3) (31). The group 2 β-lactamases are variably inhibited by active site-directed β-lactamase inhibitors, including tazobactam and clavulanic acid, whereas group 3 β-lactamases (the metallo-β-lactamases or MBLs) are inhibited by metal chelators, such as ethylenediaminetetraacetic acid (EDTA). Commonly recognized CPs in the 2f group include K. pneumoniae carbapenemase (KPC), Guiana extended-spectrum β-lactamase (GES), imipenem-hydrolyzing β-lactamase (IMI), not metalloenzyme carbapenemase (NMC), Serratia marcescens enzyme (SME), whereas the metallo-β-lactamases (MBLs) New Delhi metallo-β-lactamase (NDM), imipenemase (IMP), and Verona integron-encoded metallo-β-lactamase (VIM) are representative examples in group 3 (Table 3). These enzymes are primarily found in the Enterobacterales family and P. aeruginosa. The functional BJM scheme was further updated in 2010 with new functional subgroups as new β-lactamases were identified (32). In this updated scheme, CPs were further classified into functional group 2df, which includes the OXA (oxacillinase)-like carbapenemases, and these enzymes are found primarily in A. baumannii (OXA-23,-24, and -58) and the Enterobacterales (OXA-48-like) family.

TABLE 3.

Classification of carbapenemases by functional (Bush–Jacoby–Medeiros classification) and molecular (Ambler classification) schemes

Functional classification	Molecular classification	Representative enzymes	Most common organisms associated with production	Characteristics
2df	D	OXA-23 OXA-48	Acinetobacter baumannii Enterobacterales	Active site serine Variably inhibited by clavulanic acid and tazobactam
2f	A	KPC-1, SME-1, GES-2	Enterobacterales Serratia marcescens Pseudomonas aeruginosa	Variably inhibited by clavulanic acid and tazobactam Active site serine
3a	B1 B3	NDM-1, IMP-1, VIM-1 CAU-1	Glucose-non-fermenters/Enterobacterales Pseudomonas aeruginosa Pseudomonas aeruginosa Caulobacter crescentus	Inhibited by EDTA Active site zinc
3b	B2	CphA	Aeromonas hydrophyla	Inhibited by EDTA Active site zinc

The more commonly used molecular or Ambler classification system divided β-lactamases into one of four classes (A, B, C, or D) based on amino acid sequence (33). Molecular classes A, C, and D have serine at their active site, whereas molecular class B β-lactamases are MBLs with at least one zinc ion at their active site. CPs fall into molecular classes A, B, and D (Table 3). The class B MBLs are further divided into subclasses B1, B2, and B3 based on functional and structural characteristics (Table 3) (34, 35).

Class A carbapenemases

Molecular class A carbapenemases were discovered in the 1980s in what was then known as the Enterobacteriaceae. These enzymes require an active-site serine for hydrolytic activity. They may be found either on plasmids (e.g., KPC) or be chromosomally encoded (e.g. SME) (36). They are active against penicillins, cephalosporins, carbapenems, and monobactams. Like other class A β-lactamases, this group is typically inhibited by the β-lactamase inhibitors clavulanic acid and tazobactam. The class A carbapenemases fall into functional group 2f (31). The class A carbapenemases fall primarily into five groups, including GES, SME, KPC, IMI, and NMC. KPC is the most commonly described class A carbapenemase with more than 200 variants described (37, 38). Outside of the five groups described above, less commonly reported, class A carbapenemases have also been described (e.g., SFC-1; Serratia fonticola carbapenemase-1) (37, 39, 40). The class A carbapenemases are found primarily in members of the Enterobacterales and P. aeruginosa.

Class B carbapenemases (metallo-β-lactamases)

Molecular class B carbapenemases are also known as the metallo-β-lactamases or MBLs. MBLs differ from Class A and D carbapenemases by the presence of zinc at their active site. They can be further subcategorized into three subclasses—B1, B2, and B3 (Table 3). B1 and B3 enzymes hydrolyze all β-lactams, including penicillins, cephalosporins, carbapenems, except monobactams. B2 enzymes (e.g., CphA associated with Aeromonas species), on the other hand, have a narrower spectrum targeting carbapenems only (41). MBLs are not inhibited by common beta-lactamase inhibitors like clavulanic acid or tazobactam, rather they are inhibited by a metal chelator, such as EDTA. The most recognized MBLs are NDM, IMP, and VIM, which are plasmid-encoded. Numerous variants of these enzymes have been described (42). Other less common MBLs are either plasmid or chromosomally encoded (41). MBLs are prevalent in members of the Enterobacterales, P. aeruginosa, and Acinetobacter.

Class D carbapenemases (OXA carbapenemases)

The molecular class D β-lactamases were originally named due to their ability to hydrolyze oxacillin and cloxacillin (i.e., oxacillinases), hence the name OXA enzymes. While they also require serine at their active site for activity like the class A β-lactamases, they are distinct at a molecular level. The OXA β-lactamases that have carbapenem hydrolyzing activity fall into a different functional subgroup (2df) than those without carbapenem hydrolyzing activity (2d and 2de) (32). The OXA carbapenemases have activity against penicillins, narrow-spectrum cephalosporins and have weak activity against the carbapenems (43). They are variably inhibited by β-lactam inhibitors. Hundreds of OXA carbapenemases have been identified; representative enzymes in this group include OXA-48 (and related types) and OXA-23 (and related types), although 12 different groups exist (44). OXA carbapenemases are most frequently found in Acinetobacter species and represent the primary mechanism of resistance to carbapenems for this genus. OXA carbapenemases, predominantly OXA-48-like variants, are also found in members of the Enterobacterales family and P. aeruginosa (44, 45).

EPIDEMIOLOGY OF CARBAPENEMASE-PRODUCING GRAM-NEGATIVE BACTERIA

The epidemiology of β-lactamase-producing pathogens (including those with carbapenemases) had recently been reviewed (12, 46–50). We also refer the readers to in-depth detailed publications on KPCs (51), OXA-48-like (43, 52), NDMs (53), and OXA-23-like (54) carbapenemases. We will provide a summary on the epidemiology of carbapenemase-producing Enterobacterales, carbapenemase-producing P. aeruginosa, and carbapenemase-producing Acinetobacter species in this article.

Carbapenemase-producing Enterobacterales

The KPCs, NDMs, and OXA-48-like enzymes are the most prevalent carbapenemases among Enterobacterales, especially among K. pneumoniae and E. coli (16, 55–57). Genomic surveillance data from the SMART and INFORM global surveillance programs (2012–2017) showed that 55% of carbapenemase-producing Enterobacterales (n = 1,615) were positive for KPCs, followed by OXA-48-like carbapenemases (27%) and NDMs (26%) (58, 59). Interestingly, the VIMs were the more common carbapenemase among Enterobacter spp. (60, 61) and second most common carbapenemase among Citrobacter spp. (61, 62). Overall, IMPs are rare among members of the Enterobacterales (16, 63).

Enterobacterales with bla_KPCs

KPC-1 (which turned out later to be the same as KPC-2) was initially identified a K. pneumoniae isolate obtained from East coast USA (specifically North Carolina) in the late 1990s as part of the Intensive Care Antimicrobial Resistance Epidemiology surveillance program (64). K. pneumoniae isolates with bla_KPC-2 quickly established themselves in North-East USA, especially in New York City (65) and over time spread to other parts of the USA following an East to West direction (66). K. pneumoniae isolates with bla_KPC-3 were first reported in 2004 (67) and followed the same East–West spreading pattern as KPC-2 (66). KPC-producing K. pneumoniae then spread to other countries such as Israel, Greece, Italy, Colombia, and China (68).

Today, Enterobacterales with bla_KPCs (most often K. pneumoniae isolates) are endemic in North America (i.e., USA), South America (i.e., Colombia, Brazil, Argentina, Ecuador), Europe (i.e., Italy, Greece, Germany, Poland, Portugal), and Asia (i.e., China, South Korea, Taiwan) (46, 47, 50) (Fig. 1A). The frequencies of K. pneumoniae with bla_KPC in Northeastern USA and Israel have declined since the early 2010s due to the implementation of effective infection prevention and control measures (69, 70). However, global nosocomial outbreaks with KPC-producing bacteria remain important causes of nosocomial and long-term care facility (LTCF) outbreaks (71).

Fig 1.

Global distribution of KPC (A), NDM (B), and OXA-48-like (C) carbapenemases. Endemicity was defined applying the CDC definition: “Cases are regularly identified in healthcare facilities across the region, including those in different transfer networks. Cases primarily occur in patients admitted from facilities in the region, suggesting that transmission is sustained without new importations from outside the area.” (72).

Currently, there are over 200 KPC types worldwide, but KPC-2 and KPC-3 remain the most common types of KPC enzymes (38, 42). The presence of bla_KPCs is linked with Tn4401 and the global distribution of certain MDR K. pneumoniae high-risk clones, such as successful clonal complex 258 (CC258), ST258 (clades I and II), ST11, ST147, ST512 and ST307 (12, 73–75).

After the clinical introduction of ceftazidime–avibactam, KPC variants resistant to the combination emerged with greater than 65 resistant variants identified to date (76). Various point mutations, insertions, and/or deletions in bla_KPC have been associated with development of resistance to the combination. These mutations tend to occur at three mutational hot spots within the gene: 1) the Ω-loop region (residues 164–179), 2) loop 237–243, and 3) loop 266–275 (76). K. pneumoniae isolates (among other Enterobacterales) harboring bla_KPC variants that are resistant to ceftazidime–avibactam are increasing in areas of KPC endemicity, such as the US (77, 78), some European countries (i.e., Spain, Italy, and Greece) (76), and South America (79).

Enterobacterales with bla_NDMs

NDM-1 was first reported in 2009 in a K. pneumoniae isolate obtained from a Swedish patient that had previously been admitted to health care facilities in India (80). It soon became apparent that Enterobacterales with bla_NDMs are endemic in the Indian subcontinent (i.e., India, Pakistan), among hospital and community isolates, and such isolates had been around since the mid-2000s (81). Today, NDM-producing bacteria (belonging to 11 bacterial families, including Aeromonadaceae, Alcaligenaceae, Cardiobacteriaceae, Enterobacteriaceae, Moraxellaceae, Morganellaceae, Neisseriaceae, Pseudomonadaceae, Shewanellaceae, Vibrionaceae, and Xanthomonadaceae) are found across the globe with additional endemic regions, such as Bangladesh, Egypt, Romania, Bulgaria, Serbia, and the United Arab Emirates (53) (Fig. 1B). The global spread of NDM-producing bacteria is also linked with travel to endemic regions (53). NDM-producing bacteria are also important causes of nosocomial outbreaks in non-endemic regions, such as Western Europe, Mexico, US, Canada, and certain Asian countries (i.e., Vietnam, Thailand) (47, 82). In the US, NDM is the second most commonly encountered carbapenemase among CP-CRE. Data from the CDC from 2017 to 2019 identified 9% of CP-CRE positive for NDM when screened for the big five carbapenemase genes (83). Despite, recent increases in the prevalence of NDM among CRE within the US, it still does not meet the CDC definition for endemicity (72, 83).

Currently there are more than 60 different NDMs, with NDM-1 and NDM-5 being the most common NDM enzymes globally (38, 42, 57, 84). An intact or truncated ISAba125 is situated upstream of bla_NDM, while ble_MBL, (responsible for bleomycin resistance) is always situated downstream of bla_NDM (53), NDM-positive Enterobacterales are distributed across various species, STs, and found within multiple plasmid platforms (53, 84). K. pneumoniae with bla_NDMs are distributed across multiple clones, but certain STs, especially ST11, ST14, ST15, and ST147 are often described among NDM-positive K. pneumoniae (53, 74, 85). Among global E. coli with carbapenemases, NDM-1 is linked with high-risk clones, such as ST131, while NDM-5 is linked with ST167 and ST410 (84, 86, 87).

Enterobacterales with bla_OXA-48-like

OXA-48-type carbapenemases consist of various enzymes, with OXA-48, OXA-181, and OXA-232 being the most common global OXA-48-like enzymes (43, 52).

OXA-48 is currently the most common global OXA-48-like enzyme and was first described in 2004 from a K. pneumoniae isolate obtained in Turkey (88). During the mid to late 2000s, OXA-48-producing K. pneumoniae and E. coli established themselves as important pathogens in Turkey (89). In the early to mid-2010s, OXA-48 spread at an alarming rate globally, and today Enterobacterales with bla_OXA-48 (most commonly K. pneumoniae), are endemic in the Middle East and North Africa where these bacteria are responsible for nosocomial outbreaks (43) (Fig. 1C). Health care-associated outbreaks, most often due to K. pneumoniae and E. coli with OXA-48, have also been reported in various European countries (e.g., Spain, Belgium, France, The Netherlands), Australia, Mexico, China, Taiwan, and South Africa (43) (Fig. 1C). Based on CDC data from 2017 to 2019, OXA-48-like carbapenemases are the third most commonly encountered CP among CP-CRE. At this time, OXA-48-like is not considered endemic to the US (83). The composite transposon Tn1999 was responsible for the capture of bla_OXA-48 from the aquatic species Shewanella spp. (likely S. xiamenensis) and found its way onto broad-host range IncL-type conjugative plasmids (43). The Tn1999/IncL combination then spread to various members of the Enterobacterales and has been driving the global interspecies spread of bla_OXA-48 (90). High-risk bacterial clones play a minor role in the global dispersion of OXA-48 apart from K. pneumoniae ST11, ST405 in Spain (91) and E. coli ST38 in the UK (92).

OXA-181 is the second most common global OXA-48-like carbapenemase and was first reported among K. pneumoniae and Enterobacter spp. obtained from (93) to 2007 as part of the SENTRY Antimicrobial surveillance program (93). Similar to OXA-48, Enterobacterales with bla_OXA-181 rapidly spread across the world during the mid-2010s, and currently OXA-181-producing K. pneumoniae and E. coli are endemic in the Indian subcontinent (i.e., India, Pakistan, Bangladesh) (94) and certain countries with sub-Saharan Africa (i.e., Nigeria, Angola, Zimbabwe, Mozambique, Tanzania, Kenia, South Africa) (95–98), and the Middle East (i.e., Egypt, Jordan (99, 100) (Fig. 1C). Health care-associated outbreaks, most often due to K. pneumoniae and E. coli with OXA-181, have also been reported in South East Asia (101–103). The insertion element ISEcp1 was responsible for the capture of bla_OXA-181 from the aquatic species Shewanella spp. (likely S. xiamenensis), incorporated into Tn2013 and then found its way onto broad-host range ColE2, IncX3, IncN1, and IncT conjugative plasmids (43). The Tn2013/IncX3 combination then spread to various members of the Enterobacterales and has been driving the global interspecies spread of bla_OXA-181 (43, 84). High-risk clones play an important role in the global dispersion of OXA-181 especially K. pneumoniae ST307, ST147 (74, 104), and E. coli ST410 (86, 87). OXA-232 is a derivative of OXA-181 and share very similar endemic regions and underlying molecular epidemiology features (43).

Carbapenemase-producing P. aeruginosa

The VIM-1 MBL were first reported in 1999 from a P. aeruginosa isolate obtained in Verona, Italy (105). VIM-2, described in 2000 (106), are the most common carbapenemase among global P. aeruginosa and is situated within a class one integron that contain various AMR gene cassettes (57, 107). VIM-2 producing P. aeruginosa is endemic in Europe (especially Greece, Italy, the Netherlands), North America (i.e., Canada [specifically Calgary], Mexico), South America (Colombia, Brazil, Argentina), Africa (South Africa, Kenia), and Asia (China, India) (107–111). ST111 with bla_VIM-2 had been responsible for prolonged global nosocomial outbreaks that were linked to contaminated sinks in a few reports (107, 112). VIM is the most common carbapenemase encountered among CRPA in the US (83). The global distribution of VIM-producing P. aeruginosa is associated with various high-risk clones that includes ST111, ST175, ST233, ST235, ST277, ST357, ST654, and ST733 (113, 114). Other carbapenemases identified among P. aeruginosa are less common than VIM and include NDM, IMP, Adelaide imipenemase (AIM), German imipenemase (GIM), Sao Paulo MBL (SPM), KPC, and GES variants (25, 107, 115, 116).

Carbapenemase-producing A. baumannii complex

The chromosomally encoded class D OXA-type β-lactamases (i.e., OXA-51-like) are intrinsic to A. baumannii complex (44). OXA-51-like carbapenemases are not expressed in their natural state and requires the insertion sequence ISAba1 upstream of bla_OXA-51 to act as a promoter for its expression (117). Acquired OXA-type carbapenemases among A. baumannii complex are divided into three groups that include the following: i) OXA-23-like β-lactamases (examples include OXA-23, OXA-27), ii) OXA-24/40-like β-lactamases (examples include OXA-24/40, OXA-25), iii) OXA-58-like β-lactamases (examples include OXA-58, and OXA-96) (117). These carbapenemases have a global distribution with OXA-23 being the most common acquired OXA-β-lactamase among Acinetobacter spp. (54). There is a strong clonal component among carbapenemase-producing A. baumannii with ST1 and ST2 being the most common high-risk clones (54). Other non-OXA carbapenemases identified among Acinetobacter spp. are less common and include the NDM, VIM, IMP, and KPC variants (54).

Laboratory detection of carbapenemase-producing Gram-negative bacteria obtained from surveillance studies

Surveillance studies provide data regarding the antimicrobial susceptibility profiles of carbapenem-resistant organisms and more specifically carbapenemase producers (48). The identification of carbapenemases among carbapenem-resistant isolates provide important additional information on the types, prevalence, frequencies, global distribution and trends over time of such enzymes (47). These surveys also aid with developing effective IP&C measures, developing novel antimicrobial agents and designing diagnostic platforms to rapidly identify and limit the spread of carbapenemase-producing bacteria within the healthcare facilities as well as the community setting (118). Importantly, depending on the epidemiology, different laboratory algorithms may be applied for detection of CP-CRO to guide patient care.

Phenotypic tests for the detection of carbapenemases in surveys provide limited information on the specific carbapenemase types. Molecular assays (i.e., PCR, LAMP, etc.) or immunochromatographic lateral flow methods are required to identify the different specific carbapenemase types (e.g., NDMs, KPCs, etc.). Such information is especially helpful to determine the associated susceptibilities, frequencies, and global trends over time of different carbapenemase types and the species linked with such enzymes (119).

A recent global genomic survey of carbapenemase-producing E. coli showed that the underlying molecular epidemiology within the same carbapenemase groups (i.e., NDM, OXA-48-like) were very different (84). For example, NDM-1 was linked with multiple STs, while NDM-5 was specifically associated with the E. coli high-risk clones namely ST167 and ST410. Furthermore, the geographical distribution of E. coli with NDM-1 and NDM-5 was also different. NDM-1 isolates showed global distribution, while those isolates with NDM-5 were numerous in specific countries, such as Egypt, Thailand, and Vietnam. Similar differences were noted for E. coli with OXA-48 and those with OXA-181. OXA-48 isolates showed global distribution among various STs. However, OXA-181 isolates showed a high prevalence in the Middle East, North Africa, and Southeast Asia and was linked with the high-risk clone, ST410. This data suggested that future genomic surveys should consider using genomic sequence methodologies that characterize individual carbapenemases to provide in-depth information on the underlying molecular epidemiology of carbapenemase-producing E. coli. Sequencing-based approaches in global surveillance studies have also showed different underlying genomic epidemiology among other members of carbapenemase-producing Enterobacterales [i.e., Enterobacter spp (60) and Citrobacter spp (62)]. As WGS platforms and bioinformatic analysis are morphing into more user-friendly and cost-effective versions, such technologies are likely to become the gold standard for in-depth genomic surveillance studies.

WHY IS IT IMPORTANT FOR THE LABORATORY TO DISTINGUISH CARBAPENEMASE (CP) FROM NON-CARBAPENEMASE (NON-CP-) CRO?

The guidance and necessity for detecting the mechanism mediating resistance to carbapenems among CRO have changed over the years. When the breakpoints were first established for the carbapenems, resistance was rare, and plasmid-mediated carbapenemase genes were uncommon. Thus, initially the breakpoints were set higher, and no mechanism-based testing was necessary. However, as plasmid-mediated carbapenemase genes started to emerge in the 2000s, it became clear that the breakpoints were set too high as carbapenems often tested as susceptible despite the presence of a carbapenemase gene. This led to the lack of detection and subsequent clinical failures (120). To initially address this, the Modified Hodge Test (MHT) was first introduced by CLSI in 2009 to detect carbapenemase producers among isolates with elevated carbapenem MICs (Fig. 2). CLSI guidance was to report all carbapenems as resistant in isolates with a positive MHT. In 2010 and 2012, CLSI lowered the carbapenem breakpoints and no longer required MHT testing as the MICs would likely test not susceptible (i.e., intermediate or resistant), and the carbapenems were reported as tested. The US Food and Drug Administration (FDA) now recognizes the CLSI carbapenem breakpoints for the Enterobacterales, P. aeruginosa, and Acinetobacter species (30).

Fig 2.

Phenotypic and immunochromatographic assays for the detection of carbapenemase-producing carbapenem-resistant organisms. (A) Modified Hodge Test; 1. Klebsiella pneumoniae ATCC BAA-1705, positive result, 2. K. pneumoniae ATCC BAA-1706, negative result, 3. Clinical isolate, positive result. (B) CLSI Carba NP positive result, Tube A. No imipenem added, red, Tube B. Imipenem added, yellow. (C) RAPIDEC CARBA NP (bioMérieux, Inc) positive result, Well d. No imipenem added, red, Well e. Imipenem added, yellow. (D) Neo-Rapid carba screen (Rosco Diagnostica) positive result, Tube 1a no imipenem added, red, Tube 1b imipenem added, yellow (293). (E) Rapid CARB Blue Screen® (Rosco Diagnostica), Tube a, no imipenem added, blue, Tube b, imipenem added, yellow. (F) Modified Carbapenem Inactivation Method (mCIM) positive result; G. mCIM, negative result. (H) mCIM and EDTA-mCIM (eCIM) results; positive for a serine carbapenemase producer as there is no inhibition of carbapenemase activity in the presence of EDTA. (I) mCIM and EDTA-mCIM (eCIM) results; positive for a metallo-beta-lactamase producer as there is inhibition of carbapenemase activity in the presence of EDTA. (J) RESIST-5 O.O.K.N.V. (Coris, Gembloux, Belgium) immunochromatographic lateral flow assay results for a OXA-48-like-producing Escherichia coli. (K) CARBA5 (NG Biotech) immunochromatographic results for a triple carbapenemase-producing (KPC, OXA-48-like and NDM) K. pneumoniae. The mCIM and eCIM results from panel I were used to demonstrate a positive mCIM (panel F) and a negative mCIM result (panel G). Panels A through I are reproduced from reference (121) [panel A originally from reference (122) and panels B and F through I from reference (123), reproduced with permission of the publisher; panel D originally from reference (124)].

Until recently, defining the mechanism of carbapenem resistance among CRO was mostly encouraged for IP&C or for epidemiologic purposes by both CLSI and EUCAST (125, 126). However, with the availability of novel beta-lactam combination agents that have specific niches for coverage of CROs based on the mechanism of resistance and in response to the Infectious Diseases Society of America (IDSA) treatment guidance for multidrug-resistant organisms (MDRO), CLSI modified verbiage throughout the M100 document to reflect that testing should be performed to inform therapeutic management of patients (24, 127, 128).

Beginning in 2022, the IDSA released treatment guidance for ARO including ESBL-producing Enterobacterales, AmpC-producing Enterobacterales, CRE, P. aeruginosa exhibiting difficult-to-treat resistance (DTR P. aeruginosa; inclusive of CR-PA), carbapenem-resistant A. baumannii (CRAB), and Stenotrophomonas maltophilia (116, 117, 129). The IDSA AMR Treatment Guidance is updated annually. The IDSA AMR Guidance encourages clinical microbiology laboratories to perform additional testing on CROs to determine if a carbapenemase is present, as this will inform antibiotic decision making. For example, if a KPC was identified among carbapenemase-positive isolates, one of the novel beta-lactam–beta-lactamase inhibitor combinations (BL–BLI), such as meropenem–vaborbactam, ceftazidime–avibactam, or imipenem–relebactam are suggested as preferred treatment agents. However, if OXA-48 is detected, ceftazidime–avibactam is preferred due to unique coverage of this family of carbapenemases compared with other available BL–BLIs. Finally, if an MBL gene is identified, ceftazidime–avibactam plus aztreonam or cefiderocol is preferred (116). Thus, CP detection and more specifically the type of carbapenemase genotype impacts patient management decisions. Additional beta-lactam combination agents include aztreonam–avibactam, cefepime–taniborbactam, cefepime–zidebactam, cefiderocol–xeruborbactam, and the BLI nacubactam. Fortunately, all of these novel BL–BLI have activity against both serine and MBL carbapenemases. And, except for aztreonam–avibactam, they all have activity against carbapenemases commonly produced by both the Enterobacterales and P. aeruginosa (130, 131). Sulbactam–durlobactam is active against A. baumannii complex isolates producing carbapenem-hydrolyzing OXA variants (e.g., OXA-24, OXA-23, OXA-40) (132). The role of other BL/BLIs in development (e.g., cefepime–zidebactam, cefiderocol–xeruborbactam) are also being investigated against CRAB isolates. Regardless of the mechanism mediating resistance, phenotypic AST should always be performed to inform therapy decisions as the activity of novel agents is not guaranteed, and resistance has emerged even to the novel beta-lactam combination agents (73).

In fact, testing for carbapenemases and the carbapenemase gene family is now important for laboratory reporting practices as well. Data presented to CLSI in 2019 demonstrated that 14% and 23% of KPC-producing Enterobacterales test susceptible (S) or susceptible-dose dependent (SDD) to cefepime by disk diffusion or BD Phoenix Automated system, respectively (133). At that time, there was neither treatment guidance nor in vivo evidence to support that cefepime would not be effective for therapy of KPC producers. Thus, CLSI decided the best approach was to continue to report cefepime as tested and to append a comment to advise caution for using cefepime as a treatment option as there was insufficient evidence to conclude whether therapy of CP-carrying strains would be effective (24). More recently, a murine thigh model infected with CP-CRE and non-CP CRE isolates that tested as S or SDD to cefepime and treated with human-simulated regimens of cefepime found that 1-log bacterial reduction was not attainable. Further blunting of cefepime efficacy was observed among CP-CRE isolates compared with non-CP-CRE. These data indicated that cefepime activity in CRE does not meet the 1-log kill threshold indicative of clinical efficacy, especially for CP-CRE (134). The data were presented to CLSI in June of 2023, and it is now recommended that cefepime S/SDD results should be suppressed (i.e., not reported) or edited and reported as resistant for isolates that demonstrate CP production as reflected in the M100-S34 document (24).

Another example is reporting considerations for meropenem–vaborbactam for OXA-48-like producers. Vaborbactam lacks OXA-48 carbapenemase inhibitory activity as noted in the product label and does not provide any additional activity above meropenem (135). However, the breakpoints for meropenem–vaborbactam (S: ≤4/8 μg/mL; I: 8/8 μg/mL; R ≥ 16/8 μg/mL) are two dilutions higher than meropenem alone (S: ≤1 μg/mL; I: 2 μg/mL; R: ≥4 μg/mL). The differences in breakpoints are mainly driven by the different meropenem dosing practices of 1 g q8 h with standard infusion for meropenem alone versus 2 g q8 h with extended infusion for meropenem–vaborbactam (24). The higher dosing with meropenem–vaborbactam results in treatment of isolates with higher meropenem MICs, which is reflected by the higher breakpoint (24). However, this leads to scenarios where meropenem–vaborbactam may appear susceptible when meropenem tests not susceptible for OXA-48-like producers (despite vaborbactam not adding to the meropenem activity against OXA-48-like isolates). Murine thigh models infected with OXA-48-like producers did not demonstrate ≥1 log killing in most isolates with meropenem–vaborbactam MIC of 2 and 4 μg/mL despite the susceptible phenotype. Thus, CLSI now recommends that if an OXA-48-like producer is detected, suppress meropenem–vaborbactam or report as resistant (24).

Thus, laboratory reporting practices and testing for CP production among CRO for patient care is very important for IP&C practice and it importantly helps with AST reporting to guide therapeutic management of patients infected with CP-CRO isolates.

SCREENING CRITERIA FOR TESTING FOR CARBAPENEMASE (CP) PRODUCTION AMONG GRAM-NEGATIVE BACTERIA?

Testing for CP production among Gram-negative bacteria should routinely be performed by clinical laboratories to aid with guiding patient management. The specific isolates to test vary based on the identification of the Gram-negative bacteria, different national and local guidelines and may vary geographically based on the prevalence and epidemiology of CPOs. Furthermore, laboratories need to balance the risk of missing a CPO with the additional testing required to detect CPOs.

When to and how to test for CP among Enterobacterales?

CLSI states that most CP-producing Enterobacterales will usually test not susceptible to one or more of the carbapenems and usually test resistant to the third-generation cephalosporins when using the current CLSI breakpoints (24). Notable exceptions highlighted include SME, IMI, and OXA-48-like carbapenemase variants that often test susceptible to the third-generation cephalosporins. For CREs, testing not susceptible to ertapenem is often the most sensitive indicator of carbapenemase production, but at the expense of specificity as not-susceptible results often occur with ESBL and AmpC beta-lactamases in combination with permeability defects (136–138).

Thus, testing should be considered for isolates with elevated carbapenem MICs (intermediate or resistant) based on CLSI guidelines. EUCAST highlights that some CP-CRE may have MICs that fall below the EUCAST breakpoint and recommends using the epidemiologic cut-off value (ECOFF) to screen for CP-CRE. The screening cut-off for both ertapenem and meropenem for the Enterobacterales as defined by EUCAST is >0.125 µg/mL. Note the ertapenem screen is one-dilution higher than the ECOFF to mitigate the specificity concerns. EUCAST states that meropenem for CP-CRE screening offers the best compromise between sensitivity and specificity (126). However, many automated susceptibility testing panels do not have doubling-dilutions that go low enough to utilize this screening cutoff.

SENTRY data presented at the (139) CLSI meeting demonstrated differences in screening criteria based on the carbapenemase genotype. Ertapenem not susceptible results (intermediate or resistant) had the best sensitivity for detection of KPC (98.6%), NDM (99.5%), OXA-48-like (97.1%), and IMP (95.5%) producers individually and best overall sensitivity (97.0%) for detecting the big five carbapenemases. VIM producers were best detected applying imipenem not susceptible as a screen for carbapenemase production (139; personal communication Mariana Castanheira; Fig. 3). Note that these data represent the sensitivity of detection; specificity data are not included. The specificity of using a mono-ertapenem-resistant approach to screen for CP-CRE based on US data sets ranged from ~10% to 20% depending on the data set (139).

Fig 3.

Sensitivity of different screening criteria for the detection of carbapenemase-producing Enterobacterales. I: intermediate, R: resistant, CP: carbapenemase.

When to and how to test for CP among P. aeruginosa?

Several phenotypic, genotypic and proteomic tests are available for CP detection in P. aeruginosa, including cost-effective phenotypic approaches (see further details below). Similar to CRE, carbapenem resistance alone does not discriminate CP-CRPA versus CRPA due to non-carbapenemase mechanisms. However, testing among CRPA may have low yield relative to the resources needed to test all CRPA, especially in geographic regions where CP-CRPA is uncommon. For example, only 2% of CRPA in the US are CP-CRPA as non-CP mechanisms predominate (e.g., the loss of OprD porin expression and/or upregulation of MexAB-OprM efflux pumps) (140). Various strategies to identify isolates that are more likely to harbor CPs have been described. A study evaluated a MIC based approach suggesting that a MIC of ≥8 µg/mL for meropenem or imipenem demonstrated >98% sensitivity for detection of CP-CRPA albeit with poor specificity (141). An algorithm including imipenem or meropenem resistance plus ceftazidime not susceptible and cefepime not susceptible results reduced testing of CRPA by 43% but missed many Guiana-extended spectrum (GES) producers. Addition of the criterion of imipenem or meropenem resistance plus ceftolozane–tazobactam not susceptible results decreased the number of CP-CRPA missed by the algorithm and reduced the number of CR-PA isolates tested by 39% (142). Neither CLSI nor EUCAST comments to the appropriate screen nor cutoff to define further testing among CRPA (24, 26).

When to and how to test for CP among A. baumannii complex?

Although CP production is the primary resistance mechanism among CRAB globally due to oxacillinases with carbapenemase activity, commonly employed phenotypic, genotypic and proteomic methods for carbapenemase detection in Enterobacterales and P. aeruginosa perform poorly in their abilities to detect carbapenemases in A. baumannii (see further details below under test methods). As the great majority of CRAB are CP-producers, algorithms to screen for potential carbapenemase producers have not been developed. The CDC defines CRAB as resistant to meropenem, imipenem and/or doripenem (140). Similar to CRPA, neither CLSI nor EUCAST comments to the appropriate screen nor cutoff to define further testing among CRAB (24, 26).

CARBAPENEMASE TESTING METHODS

There are various methods to consider for testing for carbapenemase production among CRO, including phenotypic, molecular, and proteomic-based assays (Fig. 2; Table 4). Selection of a carbapenemase detection test is dependent upon several factors, including local carbapenemase prevalence, regional molecular epidemiology, diagnostic performance characteristics, cost, hands-on-time, and turnaround time (TAT) of the test (143) (Table 4). TAT is important both for therapeutic decision-making and infection control purposes, with same day results being ideal. Other considerations include the organisms to be tested (i.e., Enterobacterales and/or glucose-non-fermenting Gram-negative organisms), ease-of-use, workflow, regulatory status, necessary equipment, and reagent preparation requirements. Unfortunately, there is no one size fits all, and each institution needs to evaluate what fits best with the goals of their antimicrobial stewardship and IP&C programs. Several options are available allowing laboratories to choose a method that best suits their individual needs. Figure 4 provides an algorithm for selecting a confirmatory test for carbapenemase production based on cost and complexity (144).

TABLE 4.

Characteristics of select tests for the detection of carbapenemase production among carbapenem-resistant Gram-negative organisms^a

Test characteristics	Manual Carba NP CLSI and variants	Modified Hodge Test	CIM/mCIM and variants	BD Phoenix CPO Detect	LFA	MALDI-TOF hydrolysis	Molecular methods
Anticipated false negatives and false positives	False-negatives with OXA-48-type and GES, mucoid isolates or isolates with low level expression indicated by low carbapenem MICs	False-positives with ESBL/AmpC and permeability defects; False-negatives with mostly NDM	Rare false-positives with E. cloacae and P. aeruginosa harboring multiple non-carbapenemase mechanisms	False-positives for non-CP-CRO and erroneous classification of carbapenemase class especially among P. aeruginosa and A. baumannii complex	False-negative IMP among Enterobacterales and P. aeruginosa; False-positive IMP for off-label use with A. baumannii	Adding NH4HCO3 or ZnSO4 increases detection of Class D and Class B enzymes, respectfully.	Only detects targeted genes; IMP targets are usually specific to IMP-1 group
Cost per test (USD)	+ (+ + for commercial methods)	+	+	++	++	+	+++
Carbapenem-resistant organisms	Enterobacterales and P. aeruginosa CarbAcineto: Acinetobacter spp	Enterobacterales	Enterobacterales and P. aeruginosa CIMTris: P. aeruginosa and A. baumannii	Enterobacterales, P. aeruginosa, and A. baumannii complex	Assay dependent, Enterobacterales, P. aeruginosa, and A. baumannii	Enterobacterales, P. aeruginosa, and A. baumannii	Assay dependent; many Gram-negative organisms
Equipment	pH meter	Standard laboratory supplies	Standard laboratory supplies	BD Phoenix instrument	All reagents in kit	MALDI-TOF MS instrument	At a minimum a thermocycler; other instrumentation is required for microarray and NGS based methods
Limitations	Color change subjective. Frequent reagents preparation due to short shelf-life of imipenem containing solution (72 h). Commercial methods simplify testing with premade lyophilized reagents	Reading of results can be subjective.	Required initial setup and then plating of disk onto lawn of E. coli	Requires the instrumentation be in place.	Reagents to be used immediately if opened	Requires the instrumentation be in place. Detection range (m/z) of 160–600 (different from bacterial ID); incubation times or lysis steps vary.	Targeted assays have the inability to detect novel genes and variants
Interpretation of results	Color change from red to yellow or alternative color based pH indicator	Enhanced growth of the carbapenem-susceptible strain towards the carbapenem disk along the linear streak of the test isolate in the zone of inhibition	Zone of inhibition	Included with specific BD Phoenix susceptibility panel results	Reaction of antibodies against an antigen (carbapenemase)	Spectra of intact carbapenem and its degradation products	Amplification curves or alternatives; colorimetric detection of hybridization events for microarrays, whole genome sequencing assembly and analysis
TAT	30 min–2 h	1824 h	18 h-24 h	<8 h	5–15 min	30 min–4 h	1–3 h for automated NAT to >48 h for NGS
Regulatory status	LDT; CLSI and EUCAST endorsed FDA cleared: RAPIDEC Carba NP, RUO for other commercial methods	LDT; no longer CLSI endorsed	LDT; mCIM: CLSI endorsed; CIM: EUCAST endorsed	FDA cleared	FDA cleared: NG-Test Carba 5 RUO for other LDTs	LDT	See Table 5 for regulatory status of commercially available molecular methods

^a Modified from Tamma PD et al, 2017 (143); Cost: += <$5.00, ++: ≤$5.00–25.00, +++: >$25.00; Cost can be affected by volumes; RUO: research use only, LDT: laboratory developed test, FDA: Food and Drug Administration, CLSI: Clinical and Laboratory Standards Institute; NGS: Next-generation sequencing: EUCAST: European Committee on Antimicrobial Susceptibility Testing.

Fig 4.

Selecting a carbapenemase confirmation test for suspect clinical isolates based on cost and complexity. Cost: Low= <$5.00, Moderate: ≤$5.00–25.00, High: >$25.00; Cost can be affected by volumes. Low complexity = simple to use. High complexity: difficult to use, requires expertise in technique. LDT: laboratory-developed test.

Phenotypic methods for the detection of CP-CRO

Modified Hodge Test

The Modified Hodge Test (MHT) was the first carbapenemase test to be broadly applied for the detection of carbapenemase producers among Gram-negative bacteria that showed reduced susceptibility to the carbapenems (Fig. 2). It was initially described in 2001 as a screen for metallo-β-lactamase (MBL) producers among P. aeruginosa and A. baumannii complex (145). The test involves swabbing a lawn of a carbapenem-susceptible E. coli indicator strain and adding a carbapenem disk (ertapenem or meropenem) to the center of a Mueller–Hinton (MHA) agar plate. The CRO of interest is then streaked in a line away from the disk to the edge of the plate in such a fashion that three CRO isolates could be evaluated at once on a 100 mm MHA plate. The following day, the zone of inhibition around the carbapenem disk is evaluated at the intersection of the streaked growth of each of the CRO. If the susceptible E. coli strain grew along the streak of the CRO toward the disk, creating an indentation in the zone of inhibition, it indicates a carbapenemase producer, whereas if the zone of inhibition was unchanged (without indentation) along the streak it indicates a non-CP-CRO. The theory of the results lies in the fact that carbapenemase producers would hydrolyze the carbapenem, allowing the susceptible E. coli to grow along the streak of the CP-CRO due to the decreased concentration of the carbapenem in proximity to the CP-CRO growth.

Initial testing of KPC-producing Enterobacteriaceae in the US demonstrated good sensitivity (138). In 2009, CLSI introduced MHT to detect carbapenemase producers among Enterobacterales with elevated carbapenem MICs prior to lowering the carbapenem breakpoints in 2010 and continued to encourage its use (or another carbapenemase test) if the obsolete breakpoints from 2009 were still being applied by laboratories. As the epidemiology of carbapenemases started to evolve with the emergence of OXA-48-like and NDMs, studies demonstrated good sensitivity with OXA-48 producers but poor sensitivity for detection of NDM-producers (50%) (146). The performance issue with NDM was related to the fact that they are zinc-dependent enzymes, and the concentrations of zinc in MHA media vary. As such, adding ZnSO₄ (100 µg/mL) to the MHA media or using MacConkey agar to enhance detection of MBLs was proposed (146, 147). The ultimate demise of the MHT method was largely driven by the subjectivity of interpretating the result and the specificity issues with MBLs that could not be overcome by altering the method. Furthermore, false-positive results were reported due to AmpC or ESBL producers with permeability defects (146, 148–150). Thus, in 2018, CLSI removed the MHT as a CLSI-endorsed test. Due to the availability of several alternative methods with better performance characteristics, laboratories should consider another alternative to the MHT.

The Carba NP test and variants

The Carba NP is a rapid (30 min–2 h) imipenem hydrolysis test that can be visualized by a color change due to the phenol red indicator changing from red to yellow when a carbapenemase producer is present (151) (Fig. 2). Upon hydrolysis of imipenem, a carboxylic derivative is produced that lowers the pH resulting in the color change. The 2-h timeframe is important from a sensitivity and specificity facet. Most carbapenemases will rapidly hydrolyze the imipenem within 2 h, causing a color change, whereas most non-CP-CRO will hydrolyze the imipenem more slowly and will not cause a color change within 2 h. The test was first described by Nordmann and Poirel in 2012, hence the Carba NP, where they described perfect performance characteristics of the test compared with molecular-based techniques for CRE (151).

Multiple subsequent publications reported good performance characteristics with most carbapenemase classes. However, sensitivity issues were noted with OXA-48-like and GES producers, CP-CRE with low level carbapenemase expression (indicated by low carbapenem MICs) and mucoid isolates (143, 152). Furthermore, result interpretation was found to be subjective as slight color changes were difficult to interpret. CLSI and EUCAST endorse the Carba NP for carbapenemase detection among CRE and carbapenem-resistant P. aeruginosa (CLSI only) (24, 26). However, CLSI did introduce an invalid interpretive category for results with slight color changes from red to a dark orange (M100). In one study, invalid results were associated with both CP-CRE and a higher proportion of non-CP-CRE (143). Invalid results may also be associated with reagent deterioration, and the use of fresh reagents is encouraged (24). Poor performance of the CLSI Carba NP method for detection of carbapenemase-producing A. baumannii prompted the removal of the Carba NP as an officially endorsed method for A. baumannii from CLSI guidelines in 2018 (153). The main advantage to the Carba NP test is the capability of providing same day results, although frequent reagent preparation can encumber its implementation due to waste and associated costs if the reagents are not used within the specified shelf-life (72 h for the imipenem containing solution). This was especially problematic in regions where carbapenem resistance is infrequently encountered.

Several modifications to the manual Carba NP have been proposed to overcome limitations of the assay, including changes to extraction reagents, the inoculum, starting pH, pH indicators, and reading times (Yee et al; see Table 4). The Carba NP II was devised to detect carbapenemases and to provide Ambler class differentiation by using class-specific inhibitors (e.g., tazobactam for KPC and EDTA for MBLs) (154). Due to low-level outer membrane permeability and slow hydrolysis of imipenem by Class D OXA-type carbapenemases common to A. baumannii complex, further changes to manual Carba NP tests were required for detection of carbapenemase-producing Acinetobacter spp. One method, the CarbAcineto NP targets carbapenemase production in Acinetobacter spp. was described and reported to have a sensitivity of 89%–95% after increasing inoculum size and using a hyperosmotic solution of 5M NaCl as the lysis solution (22, 155).

In addition to the manual Carba NP and derivatives, several commercial versions were released (Fig. 2). The commercial assays were designed with the aim of simplifying testing by providing premade lyophilized reagents with longer shelf-life to remove the need for reagent preparation (156, 157). These include the RAPIDEC Carba NP (bioMérieux, Marcy L’Etoile, France) (158), which is FDA-cleared for use with the Enterobacterales and P. aeruginosa and the Research Use Only (RUO) Neo-Rapid Carb Screen (159), Rapid Carb Blue Screen (RoscoDiagnostica A/S, Taastrup, Denmark), and the β CARBA NP (Bio-Rad Laboratories N.V., Marnes-la-Coquette, France). The performance characteristics of the commercial methods are comparable to the manual Carba NP methods (22, 143), along with similar limitations as described above. Other hydrolysis methods, such as MALDI-TOF MS based approaches, among other technologies are described in the proteomic methods section below.

The modified carbapenem inactivation method (mCIM) and variants

The mCIM is a modified version of the original CIM method described in 2015 and was endorsed by CLSI for testing Enterobacterales (2016) and P. aeruginosa (2017) isolates with reduced susceptibility to the carbapenems (160). The mCIM functions by placing a meropenem disk in a tryptic soy broth (TSB) suspension with a loopful of CRO for 4 h. The carbapenem in the disk will be rapidly hydrolyzed by CP-CRO within the 4-h, whereas non-CP-CRO will hydrolyze little to none of the meropenem content in the disk within the same time period. To assess the activity of the meropenem disk after the broth incubation step, the disk is removed from the broth and placed onto an MHA plate swabbed with a carbapenem-susceptible indicator strain (E. coli ATCC 25922). After overnight incubation, the disk initially incubated with a CP-CRO will have little to no activity due to the rapid and efficient hydrolysis of meropenem and will yield a zone diameter of 6–15 mm. For non-CP-CRO, the disk will maintain most of the activity and will yield a zone of ≥19 mm.

Modifications made to the original CIM method include the use of a 1 µL inoculum for the Enterobacterales (as opposed to 10 µL; 10 µL is still recommended for P. aeruginosa), use of tryptic soy broth (TSB) over water for incubation of the meropenem disk with the organism suspension, use of an extended incubation of 4 h versus 2 h in the original procedure, and changes in the interpretation of the zone diameter readings. These modifications allowed for improved detection of OXA-48-like and NDM producers and improved the overall sensitivity and specificity for detection of carbapenemase producers to >95% (22, 143). False-negative results have been reported for GES-producing P. aeruginosa (161), and false-positive results have been reported with some inducible AmpC producers, such as Enterobacter cloacae and P. aeruginosa isolates, where a carbapeneamse was not confirmed by WGS (162). Overall, the method is easy to perform and interpret while using affordable and readily available materials (disks and broth), making it easy for most laboratories to implement. The major limitation to the mCIM is the longer turnaround time of 18–24 h. To try to reduce the TAT, some methods have implemented shorter incubation periods with the carbapenem-susceptible indicator strain of 6 h rather than overnight or by directly inoculating the organism on the surface of the meropenem disk on the MHA plate inoculated with the indicator strain, eliminating the 4-h TSB broth incubation step (163–165). Similar to the Carba NP, the mCIM does not have good performance characteristics with A. baumannii, and as such, its use with A. baumannii is not recommended by CLSI. However, a modified version using 0.5 M Tris-HCl buffer for extraction (CIMTris) performs well with Acinetobacter and P. aeruginosa with sensitivities and specificities reported above 90% (166).

Additional modifications to the mCIM were developed to allow differentiation of MBLs from serine carbapenemases by the addition of EDTA, a divalent cation chelator and inhibitor of MBLs. The EDTA mCIM (eCIM) is performed in parallel to the mCIM, and it is only interpreted if the mCIM is positive, indicating carbapenemase production. The CLSI-endorsed method is limited to use with Enterobacterales. The eCIM is positive for a MBL producer if there is a ≥ 5-mm zone diameter difference between the mCIM and eCIM zone diameter results indicating the 5 µM EDTA inhibited the MBL, thereby upholding the activity of the meropenem in the disk in the presence of an MBL-producer. If there is ≤4 mm difference, it is considered positive for a serine carbapenemase that is not effectively inhibited by EDTA. A CLSI study, demonstrated that the eCIM had a sensitivity >95% and specificity of >92% for differentiation of MBLs from serine carbapenemases among Enterobacterales (167). It is important to note, that false-negative eCIM results will occur for dual MBL and serine carbapenemase producers (24). This occurs as serine carbapenemases are not inhibited by EDTA and masks the presence of the MBL. Furthermore, some OXA-type enzymes have also been shown to be inhibited by EDTA because divalent cations are required to stabilize the more active dimeric form of the enzyme. In the presence of EDTA, the OXA-type enzyme is converted to the less-active monomeric form, resulting in reduction in carbapenemase activity and potentially false-positive eCIM result (168, 169). A study evaluating the eCIM among P. aeruginosa found it did not perform well with IMP and SPM producers. A high concentration EDTA (40 µM) eCIM method was described with improved performance for detection of SPM and IMP MBLs among P. aeruginosa without sacrificing the ability to detect other carbapenemases (170, 171). Differentiation of MBLs from serine carbapenemases can be clinically useful when targeting therapy based on genotype for carbapenemase producers, especially as the novel BL–BLI combinations do not have activity against MBL producers. However, in geographic areas where dual serine and MBL producers are common, the eCIM is not recommended.

Carbapenemase class distinguishing antimicrobial susceptibility testing-based assays

Antimicrobial susceptibility testing-based methods, including disks, gradient diffusion, and automated AST methods, are designed to differentiate between different CP classes. They function by testing β-lactam agents with and without a carbapenemase class specific inhibitor to differentiate among the different carbapenemase classes.

The MASTDISCS combi Carba plus disc system D73C (MAST-Carba plus; Mast Group, Bootle, Merseyside, UK) is a five-disk based system aimed to identify carbapenemase types among Enterobacterales by utilizing antibiotic disks with and without carbapenemase class inhibitors. The five disks include faropenem alone (Disk A), combined with MBL inhibitors (Disk B), KPC inhibitors (Disk C), AmpC inhibitors (Disk D), and temocillin combined with MBL inhibitors (Disk E). It can distinguish between KPC, OXA-48-like, and NDM-producers, as well as non-CP-CREs. Studies demonstrate that this method performs the best for detection of OXA-48-like producers, followed by NDM producers and poor performance with KPC producers (sensitivity of 46%–81%) with less-than-ideal specificity (≥77%) (172–174). Furthermore, interpretation of the results can be complex, especially when multiple carbapenemases are produced.

Gradient diffusion methods have been developed, including the combination of a β-lactam agent on one end of the strip and the β-lactam combined with a β-lactamase inhibitor on the other end of the strip. The MICs of the β-lactam individually and in combination with the β-lactamase inhibitor are determined, and a preset ratio of the two or the presence of a phantom zone indicate a positive result. Tests for KPC producers using boronic acid and MBL producers using 2-mercaptopropionic acid (MPA) or EDTA have been developed (175). The MBL imipenem/EDTA strips have reported sensitivities of 78%–84% and specificities 80%–97% among CRE, CR-PA, and CRAB (22, 143). Although these overnight tests are relatively straightforward to use, the major pitfall lies in their cost per test (~$5 USD/strip), which is associated with a narrow-based result for the detection of a defined carbapenemase class per strip. Similarly, dual serine and MBL producers may interfere with result interpretation and cause false-negative results.

Finally, the BD Phoenix CPO Detect (BD Diagnostics Systems, Sparks, MD) is the first automated carbapenemase test that is combined with routine Phoenix antimicrobial susceptibility testing (AST) panel and computer assisted algorithm-based detection (176). Like the ESBL tests on automated systems, the CPO Detect test allows simultaneous AST determination and carbapenemase detection and classification. Initial evaluations of the assay demonstrated high sensitivity of >96% for carbapenemase detection and poor specificity of 69%. It was able to distinguish between the different carbapenemase classes, with a sensitivity of 85% for class A, 72% for class B, and 89% for class D enzymes (176, 177). Subsequent real-world evaluations of the assay confirmed these findings with high sensitivity (>98%) but reported specificity issues and erroneous classification of carbapenemases, especially among glucose-non-fermenting gram-negative bacilli (178, 179). Despite providing same-day carbapenemase detection and classification results at the time of susceptibility testing, the performance issues with the assay with frequent false-positive results require confirmation of positive results. False-positive results can have negative ramifications on infection control and public health policies, procedures, and resources. For example, patients may be placed unnecessarily on contact precautions or surveillance efforts are pursued to detect additional cases of CP-CRO on a particular unit when the organisms aren’t a CP-CRO.

Molecular methods for the detection of carbapenemase genes

In general, several in-house, and commercial molecular methods for detection of carbapenemase genes from bacterial isolates or colonies are available and primarily focused on the detection of the big five carbapenemase genes (i.e., bla_KPC bla_NDM bla_VIM bla_IMP and bla_OXA-48-like) (180). However, the selection of targets may vary based on the type of carbapenem-resistant bacteria. For example, OXA variants that are common among A. baumannii complex may be the focus for molecular tests used on CRAB isolates.

Molecular methods include manual, automated nucleic acid amplifications tests, microarrays, and whole genome sequencing (WGS) applications. Methods, such as in-house PCR, microarrays, and WGS, will require significant molecular expertise, and additional equipment may be required (e.g., thermocycler, micro-array reader, next-generation sequencing instrumentation). Automated, commercial molecular assays have the advantage of simplified workflows and procedures that can be performed with minimal molecular expertise. FDA-cleared panels that contain carbapenemase targets performed from cultured isolates include the Cepheid Xpert Carba-R (Sunnyvale, CA, USA) (181), the Revogene Carba C assay (formerly GenePOC Carba assay; GenePOC, Québec, Canada; now Meridian Bioscience, Cincinnati, OH, USA), and the OpGen Acuitas AMR panel (Table 5). Studies evaluating the Cepheid Xpert Carba-R assay from cultured isolates describe excellent sensitivity and specificity (>97%) (181, 182). A recent meta-analysis included eight studies evaluating the accuracy of the Xpert Carba-R from clinical isolates which found a pooled sensitivity of 100% and specificity of 98% (183). The Revogene Carba C assay has demonstrated excellent sensitivity (93%–100%) and specificity (100%) among CRE, CR-PA, and CRAB isolates (184, 185). In comparison to the Xpert Carba R, it demonstrated similar performance characteristics with enhanced detection of IMP variants (184, 185). The Xpert Carba R assay only detects IMP-1 variants. The OpGen Acuitas AMR panel is a multiplex, real-time PCR containing 28 AMR markers in a 96-well format (186). The multicenter clinical trial for FDA clearance of the assay demonstrated good performance for the big five carbapenemases among Enterobacterales and P. aeruginosa isolates (>95% sensitivity and specificity). However, SPM and GES targets did not meet FDA inclusion criteria as they fell below the acceptable sensitivity threshold with 90% and 92%, respectively (186).

TABLE 5.

Molecular methods for the detection of carbapenemase genes

Test	Panel	Carbapenemase targets	Regulatory	Specimen types	TAT	Performance characteristics for carbapenemase detection	Reference
BD MAX	Check-Points CPO	blaKPC blaNDM blaOXA-48 blaVIM/blaIMP	FDA cleared/CE marked	Rectal ESwabs	2.5 h	PPA: KPC: 88.2%, OXA: 96.2% NPA: ≥99% for each target	(187)
Biofire (bioMérieux) FilmArray	BCID	bla KPC	FDA cleared/CE marked	Positive blood culture	1 h	PPA: 100% NPA: 100%	(188) (189)
	BCID2	blaKPC blaNDM blaVIM blaIMP blaOXA-48	FDA cleared/CE marked	Positive blood culture	1 h	PPA = 100% for each target NPA = 100% for each target Pooled SN: 94.9% Pooled SP: 99.7%	(190) (191) (188) (192)
	Pneumonia	blaKPC blaNDM blaVIM blaIMP blaOXA-48	FDA cleared/CE marked	BAL, mini-BAL, sputum and endotracheal aspirate	1 h	PPA KPC: 100% NPA KPC: 99.3-100% PPA VIM: 100% NPA VIM 99.7% PPA NDM: 0% NPA NDM: 99.3%	(193)
	Joint Infection Panel	blaKPC blaNDM blaVIM blaIMP blaOXA-48	FDA cleared/CE marked	Synovial fluid	1 h	PPA OXA-48: 100% NPA: 100%	(194)
Cepheid Xpert	Carba-R	blaKPC blaNDM blaVIM blaIMP blaOXA-48	FDA cleared/CE marked	Isolates, rectal swabs	1 h	Isolates: Sensitivity and specificity of >97%;	(182, 183)
Check-Points	Check-MDR CT103 XL	blaKPC blaNDM blaVIM blaIMP blaOXA-48blaGES-Carba blaOXA-23-like blaOXA-24-lke blaOXA-48-like blaOXA-58-like blaSPM blaGIM	RUO	Isolates	6 h		(195) (196)
Revogene (GenePOC)	Carba C	blaKPC blaNDM blaVIM blaIMP blaOXA-48	FDA cleared/CE marked	Isolates	70 min	Sensitivity 93-100% and specificity 100%	(184) (184) (185)
ePlex (Roche)	BCID-GN	blaKPC blaNDM blaVIM blaIMP blaOXA-48/blaOXA-23	FDA cleared/CE marked	Positive blood cultures	2 h	PPA KPC: 98.1% PPA NDM: 100% PPA OXA: 94.0% PPA IMP: 100% PPA VIM: 100% NPA ≥ 99.8%	(197) (198)
Diasorin Verigene (Luminex)	BC-GN	blaKPC blaNDM blaVIM blaIMP blaOXA-23blaOXA-24/40 blaOXA-48blaOXA-58	FDA cleared/CE marked	Positive blood culture	2 h	PPA KPC: 100% PPA NDM: 96.2%-100% PPA OXA: 94.3%-100% PPA VIM: 100% PPA IMP: 100% NPA: >99.9%	(199) (200) (201) (202)
OpGen	Acuitas AMR panel	blaKPC blaNDM blaVIM blaIMP blaOXA-48	FDA cleared	Isolates	2.5 h	Sensitivity >95% and specificity of >99%	(186)
Streck	ARM-D beta-lactamase kit	blaKPC blaNDM blaVIM blaIMP blaOXA-48	RUO	Isolates		Isolates 100% sensitivity for KPC and 100% specificity for NDM and KPC	(203)
Unyvero (Curetis)	LRT Panel or Hospitalized Pneumonia Panel	blaKPC blaNDM blaVIM blaOXA-23blaOXA-24 blaOXA-48blaOXA-58	FDA cleared/CE marked	BAL	5 h	Genotypic agreement: 100% for all targets Phenotypic agreement: 100% for all targets except blaOXA(Acinetobacter): 88.9%	(204)
	ITI: Implant and tissue infections	blaKPC blaNDM blaVIM blaOXA-23blaOXA-24 blaOXA-48blaOXA-58	CE marked		5h	Sensitivity of 59% and specificity of 90%	(205)
	BCU: Blood Culture	blaKPC blaNDM blaVIM blaOXA-23blaOXA-24 blaOXA-48blaOXA-58	CE marked		5h	Overall sensitivity of 96.8% and specificity of 99.8%	(206)
	IAI: Intra-Abdominal infection	blaKPC blaNDM blaVIM blaOXA-23blaOXA-24 blaOXA-48blaOXA-58	CE marked		5h	Overall sensitivity of 89.3% and specificity of 99.5%	(207)
T2 Biosystems	T2 Resistance	blaKPC blaNDM blaVIM blaIMP blaOXA	CE marked	EDTA whole blood	5 h	87.5 (KPC) −100% sensitivity	(208)
iCubate	iC-GN	blaKPC blaNDM	FDA cleared, CE marked	Positive blood culture broth	4.5 h	KPC: 100% PPA, >99.2% NPA; NDM: 100% PPA, 100% NPA	(209)
Easyplex system	Superbug CRE	blaKPC blaNDM blaVIM blaOXA-48	CE marked	Bacterial isolate, rectal swabs	30 mins	100% concordance from bacterial isolates, rectal swabs: 100% sensitivity and 59–63% specificity	(210) (211)
Mobidiag (Hologic)	CarbaR+	blaKPC blaNDM blaVIM blaIMP blaOXA-48blaOXA-23blaOXA-24 blaOXA-58	CE marked	Bacterial isolate, rectal swabs	80 mins	100% concordance from bacterial isolates; rectal swabs: 100% sensitivity and 59–63% specificity	(211)

Despite the relatively short turn-around time (TAT) of 1–5 h for commercial automated systems, the main disadvantage is the cost per test compared with phenotypic methods. Another limitation of targeted molecular-based methods is the inability to detect novel genes and variants as they require a priori knowledge of the specific targets for detection. To address this, there have been expanding AMR marker panels to detect AMR more broadly with a wider array of carbapenemase genes past the big five. One example is the Check-Points Check-MDR CT103XL microarray panel that detects additional bla_{OXA-23,-24,-58} variants observed in A. baumannii complex, bla_GES carbapenemase variants, and the rare MBLs bla_GIM and bla_SPM. Evaluations of the Check-Points assay demonstrate good sensitivity (98.1%–100%) and specificity (94.2%–100%) (195). The assay can distinguish between GES extended-spectrum (ESBL) and carbapenemase variants (195). An evaluation of Acinetobacter species isolates found 100% concordance for β-lactamase gene detection when compared to PCR (212). WGS has the advantage of detecting the complete resistome (i.e., the entire set of AMR genes) of the bacterial isolate but is limited by costs, molecular expertise to run the test, bioinformatic expertise required to analyze the sequencing results, and the databases queried for detection of carbapenemase genes. Common broad AMR databases include ResFinder, the Comprehensive Antimicrobial Resistance Database (CARD), Antibiotic Resistance Gene-ANNOTation, and National Center for Biotechnology Information’s (NCBI) National Database of Antibiotic Resistant Organisms (213).

Proteomic methods for the detection of carbapenemases

Immunochromatographic lateral flow methods

Rapid, immunochromatographic methods, such as lateral flow assays have been designed to detect specific carbapenemase enzyme families (i.e., the protein encoded by carbapenemase genes; Fig. 2). These assays use colloidal gold nanoparticles bound to nitrocellulose membrane sensitized with monoclonal antibodies for the capture of epitopes specific to carbapenemase enzymes individually or in a multiplex format within a lateral flow device (K-SeT tests, NDM LFIA, or RESIST-5 O.O.K.N.V.) (214). The FDA-cleared NG-Test Carba 5 LFA targets the five main carbapenemase families (i.e., KPC, NDM, VIM, IMP, and OXA-48-like carbapenemases) and can be performed within 15 min from cultured isolates (215). A multicenter study demonstrated an overall agreement of 100% for the NG-Test Carba 5 compared with a composite reference standard for testing cultured isolates of Enterobacterales and P. aeruginosa recovered on both blood and MacConkey agar (216). The NG-Test Carba5 demonstrates accurate detection of isolates producing multiple carbapenemases (e.g., NDM-type and OXA-48-like carbapenemases), and cross-reactivity was not observed with non-targeted carbapenemases (e.g., GES, SME) among claimed organisms (157).

The NG-Test Carba 5 and RESIST-5 O.O.K.N.V. (Coris, Gembloux, Belgium; two lateral flow devices required) for the big five carbapenemase families serve as simplified, rapid and cost-effective alternatives to molecular approaches. A study comparing the NG-Test Carba 5 and RESIST-5 O.O.K.N.V. among CRE demonstrated similar excellent performance characteristics between the two assays (217). An updated version of the NG-Test Carba 5v2 assay with improved detection of IMP variants that are common among P. aeruginosa has been described (182). Evaluation of this updated version for CRPA isolates demonstrated a good overall agreement of 97% compared with WGS (218). Use of the assay off-label resulted in a high-proportion (95.9%) of false-positive IMP results among A. baumannii complex isolates. The reason for the false-positive results has not been elucidated, but it is thought to be a cross-reacting protein antigen unique to A. baumannii (218). Thus, off-label use of the NG-Test Carba 5 assay for Acinetobacter isolates should be discouraged or interpreted with caution when IMP is detected. The RESIST Acineto (Coris Bioconcept) was developed to detect the major acquired carbapenemases (OXA-23, OXA-40, OXA-58, and NDM) identified in Acinetobacter species, and initial reports demonstrate high accuracy (219, 220). An LFA for the detection of GES-producing Gram-negative organisms has also been described. The assay is unable to distinguish between ESBL and carbapenemase variants but allows for early rapid detection for an under-recognized carbapenemase that is emerging among P. aeruginosa (214). Additional studies have demonstrated the use of LFAs for accurate detection of carbapenemases from positive blood culture broths and direct from rectal swabs (221–224), among other sources (225).

Mass spectrometry methods

Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a proteomic approach that has revolutionized bacterial and fungal identification by providing same day results from growth in culture. Two different approaches have been described to detect carbapenemases by exploiting the technology that has been broadly implemented in clinical and public health microbiology laboratories. The first is a peak-based method by detecting a specific peak associated with carbapenemase on the standard chromatogram used for bacterial identification and the second is a hydrolysis method.

The peak-based method detects a protein marker associated with a carbapenemase-bearing plasmid. To date, this method has only been described for the bla_KPC-bearing pKpQIL plasmid that was associated with a CP-CRE outbreak that occurred at the National Institutes for Health Clinical Center in 2011 (226, 227). Due to the potential benefit of simultaneously identifying the organism and rapidly detecting carbapenemases, Bruker Daltonics further developed a Subtyping Module to detect the ~11,109 m/z peak corresponding to a protein encoded by the p019 gene among the standard spectra range for rapid identification of KPC. Poor sensitivity (42%) of the MALDI Biotyper MBT Subtyping Module for detection of KPC in Klebsiella species was reported (228) when testing a large challenge set, whereas routine clinical use of the assay in an area of KPC endemicity resulted in 75.8% sensitivity (229). Thus, detection may be helpful due to excellent specificity at the time of identification, but the absence of the peak does not rule out the potential of a KPC producer, carbapenem resistance itself, or any other carbapenemase variant. The next generation of mass spectrometry that allows for peptide sequencing (not simply spectra-based fingerprinting), liquid chromatography tandem mass spectrometry (LC-MS/MS), can identify peptides specific to KPC, NDM, OXA-48-like, and VIM from cultured isolates and appears as a promising approach (230–234).

The second MALDI-TOF MS method, is a hydrolysis-based method that involves incubating the CRO of interest with a carbapenem for a defined period, and then the supernatant is analyzed to identify peaks associated with intact carbapenem (negative result; indicating a non-CP-CRO) or a metabolite of the hydrolyzed carbapenem (positive result; indicating a CP-CRO). Overall, the method has demonstrated good performance in several studies (235–237). The addition of NH₄HCO₃ or ZnSO₄ provides improved detection of OXA-48-like and MBL producers, respectively (236, 237). Although it is an inexpensive approach, it requires the setting of the instrument to be altered because the system is optimized for bacterial identification (m/z range of 2,000–20,000) rather than detection of drug metabolites (m/z range 160–600). Hydrolysis products have been detected applying different technologies including UV spectrometry, LC-MS/MS, and electrochemical currents detected via an electrode (238–240).

DIRECT-FROM-SPECIMEN CLINICAL TESTING FOR CARBAPENEMASES

Over the past decade, detection of carbapenemase genes directly from clinical specimens using nucleic acid amplification has become possible. These tests provide the advantage of no longer having to culture the organism prior to phenotypic or molecular carbapenemase detection, significantly improving turnround time. Commercially available, FDA-cleared tests can be used on positive blood cultures demonstrating Gram-negative organisms by direct microscopy, respiratory specimens, synovial fluid, and rectal swabs (for performance characteristics of commercial assays on rectal swabs, see 9.2.2). Detection of carbapenemase genes by direct-from-specimen methods allows escalation and targeting of therapy much more rapidly than when traditional phenotypic detection methods are used. While direct detection of carbapenemases from clinical specimens can provide rapid carbapenemase identification, there are limitations to these methods. First, commercially available assays are limited by their number of carbapenemase targets. All of the commercially available direct-from-specimen carbapenemase gene detection assays detect only a handful of common carbapenemase genes (Table 5) and will miss infrequently encountered carbapenemases or carbapenem resistance conferred by mechanisms other than carbapenemase production. Second, detection of a carbapenemase gene does not always equate to carbapenemase production (241). Third, depending on the assay, detection of the organism producing the carbapenemase may not be available. For example, surveillance panels from fecal specimens, usually only include carbapenemase gene detection. Molecular assays are also associated with high costs, although many laboratories are implementing rapid molecular identification assays anyways, particularly from positive blood cultures, and so AMR gene detection is an added advantage.

Positive blood culture specimens

There are several FDA-cleared assays that detect carbapenemase genes directly from positive blood culture bottles, including the Biofire FilmArray BCID and BCID2 (BioMerieux) assays, the Verigene BC-GN (Luminex), and the ePlex BCID-GC (GenMark). Overall, these platforms perform well for carbapenemase detection, although much of the published data are on simulated/contrived blood cultures. It is evident, however, that these three assays provide carbapenemase detection much more quickly from positive blood cultures than traditional phenotypic methods, allowing rapid therapeutic optimization. Other assays that are not FDA-cleared for detection of carbapenemase genes from positive blood cultures have developed or adapted for use with positive blood cultures (206, 242), but will not be discussed here.

The data are limited regarding the performance of the Biofire FilmArray BCID and BCID 2.0 for detecting carbapenemases. In the most recent version of the BioFire FilmArray blood culture assay, BCID2, 43 targets, including five carbapenemase genes, can be detected in 1 h (Table 5). In published studies, in which the number of carbapenemase producers present in positive blood cultures is low, carbapenemase genes were detected by BCID2 when present (100% PPA) and not detected when absent (100% NPA) (188–191) (Table 5). In a meta-analysis reviewing the performance of the BCID2, the pooled sensitivity and specificity of BCID2 for detecting carbapenemases were 94.9% and 99.7%, respectively (192) (Table 5). The addition of additional carbapenemase gene targets to the BCID2 panel (bla_KPC, bla_NDM, bla_VIM, bla_IMP, bla_OXA-48) compared with the BCID panel (bla_KPC only) allows much better coverage of circulating carbapenemases.

The Verigene BC-GN assay, like the Biofire Filmarray BCID2, detects five carbapenemase genes directly from positive blood cultures as well as eight Gram-negative pathogens (Table 5). The assay has been evaluated in both adult and pediatric patient populations, with excellent performance for detection of the carbapenemase genes present on the panel (199–202). In a study that included a mixture of 179 simulated and positive blood cultures growing organisms that produce carbapenemases, including bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA (including bla_OXA-23, bla_OXA-24/40, bla_OXA-48,bla_OXA-58), the assay correctly identified 100% of the carbapenemase genes (201). In another study that tested fresh, frozen, and simulated blood cultures, the PPA for bla_KPC was 100%, bla_NDM was 96.2%, bla_OXA was 94.3%, bla_VIM was 100%, and bla_IMP was 100% (202). The overall NPA for resistance determinants in this study was >99.9%.

The ePlex BCID-GN assay detects 21 Gram-negative pathogens and five carbapenemase genes (bla_KPC, bla_NDM, bla_VIM, bla_IMP, bla_OXA [bla_OXA-23 and bla_OXA-48]) from positive blood culture bottles (Table 5). Using fresh, frozen and contrived blood culture samples, Wolk et al. evaluated the assay against 238 carbapenemase producers, including 40 bla_IMP, 52 bla_KPC, 54 bla_NDM, 50 bla_OXA, and 42 bla_VIM in a multicenter study (197). The PPA for bla_IMP, bla_NDM, and bla_VIM was 100%, for bla_KPC was 98.1%, and for bla_OXA was 94.0% (197). The NPA was ≥99.8% for all targets (197). Few other studies have evaluated significant numbers of positive blood cultures growing carbapenemase producers using the ePlex assay.

Respiratory and synovial fluid specimens

Molecular assays for detection of carbapenemases beyond surveillance specimens and positive blood cultures are available and FDA-cleared. Diagnosis of lower respiratory tract infections is challenging as the spectrum and type of pathogens causing pneumonia is extensive. Development of multiplex molecular respiratory assays attempts to overcome some of these challenges. The Curetis Unyvero LRT panel covers 90% of the most common causes of bacterial pneumonia, including several resistance markers, with bronchoalveolar lavages being the approved specimen type. The carbapenemase targets on the panel include bla_KPC, bla_NDM, bla_VIM, bla_OXA (bla_OXA-23, bla_OXA-24, bla_OXA-48, bla_OXA-58) (Table 5). While many studies have evaluated the performance of the Unyvero LRT panel against standard of care, many do not discuss performance data for carbapenemase detection or have very few isolates with carbapenemases present. In a multicenter evaluation, Klein et al. found that the assay performed well at detecting carbapenemase genes with 100% genotypic agreement to genome sequencing, albeit only 24 isolates carrying carbapenemase genes were included in the study (204). The Biofire Filmarray Lower Respiratory Tract panel detects 26 pneumonia-causing bacteria and viruses and seven antimicrobial resistance genes, including five carbapenemase genes (bla_KPC, bla_NDM, bla_VIM,bla_IMP, and bla_OXA-48) (Table 5). Specimen types are expanded comparatively to the Unyvero panel and include BALs, mini-BALs, sputa and endotracheal aspirates. In the clinical evaluation of the BioFire Filmarray Pneumonia panel, 846 BAL specimens and 836 sputum specimens were assessed (193). A total of 10 organisms carrying bla_KPC, one organism carrying bla_NDM, and two organisms carrying bla_VIM were detected. While overall carbapenemase prevalence was low, this resulted in a 100% PPA and 99.3%−100% NPA for bla_KPC detection, 100% PPA and 99.7% NPA for bla_VIM detection, and 0% PPA and 99.3% NPA for bla_NDM compared with the comparator method (193). In a further evaluation of a subset of specimens from the above clinical evaluation, BAL specimens were assessed against phenotypic standard of care (culture and AST). Among 53 BAL specimens containing a bacterial target evaluated in a multicenter study, three were reported to have a carbapenemase detected by the Biofire Filmarray Lower Respiratory Tract panel (243). Phenotypically, two (bla_KPC detected) of these three specimens were found to grow an organism in routine culture that was carbapenem resistant. One specimen, however, found to have bla_NDM detected, was found to only contain a carbapenem-susceptible Enterobacter cloacae.

The Biofire Joint Infection Panel is also able to detect carbapenemases directly from clinical specimen. Diagnosing joint infections is highly complex plagued by low sensitivity of Gram staining and challenges with recovering organisms in culture (194). The Biofire Joint Infection Panel attempts to overcome some of these limitations by applying molecular techniques to improve the sensitivity and turnaround time compared with standard methods. The panel contains 39 targets including five carbapenemase genes (Table 5). In a study conducted at 13 different sites in the United Stated and Europe, the panel was evaluated compared with standard of care procedures and PCR/sequencing for AMR genes. The 1,544 synovial fluid specimens were tested (194). A single isolate carrying bla_OXA-48 was detected by the panel and PCR/sequencing, giving a PPA of 100%. The NPA for all carbapenemases was 100%. Although several other studies assessing the performance of this panel have been published, no further data are currently available demonstrating the performance of the assay for detecting carbapenemases genes.

Clinical impact of direct detection of carbapenemases

Rapid molecular diagnostic platforms have allowed for the more rapid detection of CPOs compared to phenotypic antimicrobial susceptibility testing. Alerting physicians to the presence of carbapenemase genes not only allows for contact precautions to be initiated, but also informs early appropriate modification of therapy to agents that cover CPOs. The PRIMERS I and PRIMERS II (Platforms for Rapid Identification of MDR-Gram-negative bacteria and Evaluation of Resistance Studies) studies did the important work of demonstrating that the detection of β-lactamase (bla) genes (including both ESBLs and carbapenemases) correlated to phenotypic susceptibility or resistance in E. coli and K. pneumoniae in >95% of cases (244). The sensitivity of molecular detection of carbapenemases for predicting phenotypic susceptibility or resistance to carbapenems was >90% (244). These studies also showed that molecular detection of carbapenemase genes can transform into an empiric decision-making tool for clinicians and researchers. Real-world, clinical impact studies have shown that direct, rapid detection of carbapenemase genes in clinical specimens decreases time to escalation of antimicrobial therapy (245, 246).

Another important impact of rapid molecular detection of carbapenemases in the prediction of success to newer B-lactam/B-lactamase inhibitors (e.g. meropenem-vaborbactam is effective therapy to class A carbapenemases, but not class B carbapenemases) and other newer antimicrobials (e.g. cefiderocol activity against class B carbapenemases).

The PRIMERS I and II studies also demonstrated limitations of molecular detection of carbapenemases including that not all carbapenem resistance is mediated by carbapenemases, and that the presence of a carbapenemase gene does not always translate to carbapenem resistance (244). Molecular assays that detect carbapenemases are limited to the targets that are probed by the assay. Emerging carbapenemases will not be detected, nor will carbapenem resistance mediated by efflux pumps or porin mutations. To be most effective, direct detection of carbapenemases needs to be coupled with antimicrobial stewardship in order to facilitate appropriate treatment management decisions and have protocols in place to manage discrepancies between genotypic and phenotypic testing (247, 248).

INFECTION PREVENTION AND CONTROL SURVEILLANCE SCREENING FOR CARBAPENEMASE-PRODUCING GRAM-NEGATIVE BACTERIA

What is required for surveillance? Who should be screened?

The basic principles of IP&C remain a foundation for preventing infections in the hospital setting and is the cornerstone in combating the spread of AMR bacteria (249). IP&C measures include cohorting patients that are infected or colonized with AMR bacteria, standard contact precautions, environmental cleaning, and meticulous attention to hand hygiene. Such practices will improve patient outcomes, save lives, and reduce overall health-care costs (51, 118).

Prior rectal colonization with Enterobacterales, P. aeruginosa, and A. baumannii with carbapenemases are predictive for the development of clinical infections with such bacteria (250–252). Therefore, the rapid identification of patients who are colonized with carbapenemase-producing Enterobacterales, P. aeruginosa, and A. baumannii will aid with selecting empiric therapy. If such results are rapidly reported to IP&C, basic principles of infection control can be implemented to prevent nosocomial outbreaks (51, 253).

Ilan and Kjeruf recently reviewed the risk factors associated with acquisition of carbapenemase-producing bacteria and found the following to be significant (254): (i) broad-spectrum antimicrobial therapy, (ii) previous hospitalization (i.e., 7 days in an intensive care setting and 3–4 weeks in general wards), (iii) prior travel history to a MDR endemic region; (iv) several underlying comorbidities; (v) different types of invasive procedures.

Out-of-country hospitalization during the past 6 months is an important risk factor for colonization or infection with carbapenemase-producing Enterobacterales, P. aeruginosa, and A. baumannii complex (68) and intra-hospital spread was responsible for mortality due to nosocomial transmission (255). It is essential to rapidly identify such patients colonized or infected by these bacteria and place them on appropriate infection control precautions.

The clinical laboratory plays a pivotal role in detecting patients infected or colonized with AMR bacteria, including those with carbapenemases (256). The laboratory costs for screening procedures are high, and the clinical laboratory, in collaboration with the IP&C program, should determine which groups of patients to screen (257–259). Minimum screening criteria to consider include (but are not limited to) patients involved in previous or current outbreak situations, admission from other institutions including long-term care facilities, previous travel history to an endemic region and the prior colonization and/or infection with carbapenemase-producing bacteria) (21, 249, 257, 260). National or state or provincial guidelines for carbapenemase screening may override local approaches. It is also imperative to notify receiving facilities of patients colonized with carbapenemase-producing bacteria.

Methods for surveillance screening

The laboratory detection of carbapenemase-producing bacteria on human specimens involves a two-step process, namely inoculating specimens on selective media, followed by the incubation for 18–24 h and then confirming the presence of carbapenemases on bacterial growth (Fig. 5) (257). Molecular or immunogenic tests can be performed directly on patient specimens eliminating the need for time consuming culture (257). This will improve the turn-around times but unfortunately add significant cost. Also, the necessity of detecting all CRO versus only CP-CRO, the requirement to recover an organism or simply the presence of a carbapenemase, may influence the approach. Therefore, clinical laboratories should utilize the most appropriate testing procedures and follow the most cost-effective necessary workflow to suit their settings and budgets (261). Results should be conveyed as soon as possible to clinical staff, antibiotic stewardship and IP&C (258).

Fig 5.

Appearance of carbapenem-resistant organisms on selective chromogenic media.

Chromogenic agars

Carbapenenem-resistant Enterobacterales (CRE)

CRE detection on human specimens can be accomplished using general media (e.g., trypticase soy blood agar, Drigalski agar media, chocolate agar, thioglycolate fluid), and selective media (e.g., MacConkey agar) to which carbapenem disks are added (262, 263). Such agar can also be supplemented with carbapenems (near the respective carbapenem breakpoint concentrations) (262). General and non-chromogenic selective media are readily available in clinical laboratories, and adding the carbapenem disks is cost-effective and easy to implement on a real-time basis. Unfortunately, such approaches are not standardized globally (i.e., which carbapenem disks to use), and they lack sufficient sensitivities for optimal screening procedures (262). When choosing a carbapenem disk, ertapenem provides the highest sensitivity for CRE detection at the expense of specificity (264). For the detection of CRE with KPC enzymes, the sensitivity and specificity of MacConkey agar supplemented with imipenem were superior when compared with MacConkey agar using a combination of different carbapenem disks (265, 266).

Commercial chromogenic-selective media for CRE screening are more expensive, but results are easy to interpret and generally provide higher sensitivities than non-carbapenem resistance media (257, 267). Chromogenic CRE selective media include the following: mSuperCARBA (CHROMagar, Paris, France), chromID CARBA (bioMérieux, Marcy-I ‘Etoile, France), ChromID CARBA SMART (bioMérieux, Marcy-l‘Etoile, France), CHROMagar KPC (CHROMagar, Paris, France), Remel Spectra CRE (Remel, Thermo Fisher, Lenexa, KS, USA), Colorex KPC (Biomed Diagnostics, White City, OR, USA) and RambaCHROM KPC (Gibson Laboratories. Lexington, KY, USA), Brilliance CRE (Thermo Fisher Scientific, Waltham, MA, USA), and Hardy CHROM carbapenemase (Hardy Diagnostics, Santa Maria, CA, USA). Bi-plates are available, such as ChromID CARBA SMART (bioMérieux, Paris, France) that comprised of ChromID CARBA medium and ChromID OXA-48 medium (268).

An advantage of commercial chromogenic agar media is that these methodologies combine the detection of CREs with presumptive organism identification (257, 267). Multiple studies have compared different chromogenic media, and recent reviews on this topic have been published (257, 265, 267, 269). In summary, chromogenic media have high sensitivities and high specificities for CRE detection and several types of specimens (i.e., rectal swabs, stool, respiratory tract, urine, and wound swabs) are suitable. However, reduced sensitivity has been observed for some chromogenic media (i.e., CHROMagar KPC) for the detection of CRE with OXA-48-like carbapenemases (270). The sensitivity for the detection of CRE with OXA-48-like carbapenemases can be increased by using combination methods, such as using a second chromogenic media specific for OXA-48-like carbapenemase (e.g., ChromID OXA-48 medium) (271), or by using biplates (e.g., ChromID CARBA SMART) (272). CHROMagar mSuperCARBA has high sensitivity for the detection of OXA-48-like carbapenemase when used as a single media (273). The presence of carbapenemases needs to be confirmed on bacterial growth when using non-chromogenic or chromogenic media (257). The mere growth of CRE on such media does not indicate that these isolates produce carbapenemases.

Carbapenem-resistant P. aeruginosa

General culture media (e.g., trypticase soy blood agar, Drigalski agar media, chocolate agar, fluid thioglycolate medium, and Campylobacter agar) and non-chromogenic-selective culture media (e.g., MacConkey agar) to which carbapenem disks have been added can be used for the detection of carbapenem-resistant P. aeruginosa from surveillance specimens (e.g., rectal or wound swabs) (274). As with CRE, general and non-chromogenicselective media are readily available in clinical laboratories and adding the carbapenem disks are cost-effective and easy to implement on a real-time basis. Unfortunately, such approaches are not standardized globally (i.e., which carbapenem disks to use), and they lack sufficient sensitivities for optimal screening procedures (265, 274). CRE chromogenic-selective agar that can also be used for the detection of carbapenem resistant P. aeruginosa included CHROMagar KPC, mSuperCARBA, Colorex KPC, and Remel Spectra CRE (275). As with CRE, using general or selective media requires subsequent identification for P. aeruginosa and confirmatory testing for carbapenem production.

Carbapenem-resistant A. baumannii

General culture media (e.g., trypticase soy blood agar, Drigalski agar media, chocolate agar, fluid thioglycolate medium) and and non-chromogenic-selective culture media (e.g., MacConkey agar) to which carbapenem disks have been added can be used for the detection of carbapenem-resistant A. baumannii using rectal or wound surveillance swabs (251, 265). As with CRE, general and non-chromogenic selective media are readily available in clinical laboratories and adding the carbapenem disks are cost-effective and easy to implement on a real-time basis. Unfortunately, such approaches are not standardized globally (i.e., which carbapenem disks to use) and they lack sufficient sensitivities for optimal screening procedures (276). Oxidase negative non-lactose fermenters with consistent colony morphology must be identified and the presence of a carbapenemase must be confirmed (277). An overnight pre-enrichment step in nutrient broth with meropenem can increase the sensitivities of general and non-chromogenic selective media (276).

Selective chromogenic agar specific for carbapenem resistant Acinetobacter spp. includes the CHROMagar Acinetobacter (CHROMagar, Paris, France). Rectal swabs can be plated directly to chromagenic agar (278). CHROMagar Acinetobacter supplemented with CR102 has been used in surveillance prevalence and environmental studies (279). When compared against other media, including CHROMagar KPC and MacConkey supplemented with imipenem, CHROMagar Acinetobacter has demonstrated higher sensitivity and specificity with less breakthrough growth (280, 281). CHROMagar, mSuperCARBA, Colorex KPC, and Remel Spectra CRE can also be used to detect carbapenem-resistant A. baumannii (282).

Molecular assays

Molecular methodologies that include quantitative PCR directly on stool and rectal swab for the detection of carbapenemases have been developed (257, 283, 284). Such approaches have several advantages over phenotypic techniques that include high sensitivities, high specificities, eliminating the need for culture, eliminating carbapenemase confirmation steps and shorter turn-around-times (257). Molecular methodologies do add significant costs that can be offset by using a specimen pooling strategy for universal screening (285).

Currently, there only a few molecular platforms that can be used directly on patient specimens for detection of carbapenemases (257, 286). Molecular platforms for carbapenemase detection generally target bla_KPC, bla_NDM, bla_VIM, bla_IMP and _{blaOXA-48-like} carbapenemase genes (257). The only FDA-cleared assay for the rapid detection of carbapenemase genes is the GeneXpert Carba-R/Xpert MDRO assay (181, 287, 288). Other molecular assays have been approved by the Conformité Européenne In-Vitro Diagnostic Device Regulation (CE-IVD) and include the following: Unyvero P35 (Curetis and OpGen, Germany and USA), Carbaplex IVD PCR (Bruker Daltonics, Bremen, Germany), Amplidiag CarbaR + VRE (Mobidiag, Espoo Finland), Hyplex SuperBug ID (Amplex Biosystem, Gras, Germany), and Check-Points CPO for BD MAX (Checkpoints, Wageningen, The Netherlands). Check-Points CPO can detect KPC, NDM, VIM/IMP and OXA-48; however, it cannot differentiate between VIM from IMP genes. The NucliSENS EasyQ KPC (bioMérieux, Marcy l’Etoile, France) is an automated assay based on nucleic acid sequence-based amplification and real-time detection of the KPC genes (289). Assays that have been evaluated for carbapenemase-producing CRE, P. aeruginosa, and A. baumannii include GeneXpert Carba-R/Xpert MDRO assay Unyvero P35, Hyplex SuperBug ID, and Check-Points CPO for BD MAX (257). Loop-mediated isothermal amplification (LAMP) assays for direct detection of OXA-23 in A. baumannii from sputum, bronchial aspirates, and rectal swabs have been developed (290, 291).

Disadvantages of molecular tests include detection of only select carbapenemase genes, with emerging enzymes or variants going undetected (257). Additionally, the presence of carbapenemase genes may not predict phenotypic resistance, i.e., low level of expression. The cost for commercial nucleic acid amplification assays is higher than that of culture-based assays, and these tests may require specialized equipment and/or expertise. This should be weighed against the benefits of molecular-based assays that include a more rapid turn-around-time and potentially higher sensitivity and specificity compared with conventional methods (257, 283, 284).

Overview of different approaches based on prevalence

Multiple factors need to be considered when selecting the most appropriate methodologies for surveillance cultures. The prevalence and epidemiology of carbapenemase-producing bacteria will aid in deciding the most suitable type of assay. Prevalence studies will provide information on the expected volume of testing. In regions with high carbapenemase prevalence (i.e., >10%) a broader, more cost-effective screening approach should be utilized, while in low prevalence regions (i.e., <10%), screening should be limited to specific high-risk populations. The type of screening samples and the clinical microbiology laboratory capabilities also need to be considered. Screening samples often consist of rectal swabs, but other specimens (i.e., stool, wound, respiratory) are also suitable. Laboratory capabilities include the availability of staff, their expertise in developing, validating, and verifying in-house assays and commercial assays, as well as their experience with different technologies.

There are no hard and fast rules on which surveillance screening methodologies should be used. The clinical laboratory, in collaboration with the IP&C Program, should determine which groups of patients to screen. High-risk patients include those admitted from out of country health facilities, recent travel to an endemic region, prior carbapenemase-positive infection/colonization, admission from long-term care facility, and admission to intensive care unit or bone marrow transplant ward (257). The frequency of screening depending on the local epidemiologic situation (i.e., endemicity, outbreak situations, high prevalence in longterm care facilities etc.) (257).

Specimens should be processed as per the manufacturer instructions. Molecular assays can be used directly on rectal swabs (e.g., GeneXpert Carba-R/Xpert MDRO, Unyvero P35, Carbaplex IVD PCR, Amplidiag CarbaR+, and Check-Points CPO). In addition to rectal swabs, respiratory specimens have been evaluated for use in the GeneXpert Carba-R/Xpert MDRO assay and the Unyvero P35 assay (292). Hyplex SuperBug ID can be used directly on blood cultures and other patient samples (293).

Reporting

Surveillance screening results must be reported to IP&C and the most responsible care provider. Immediate notification by phone, fax or laboratory information systems should occur for positive results as per local policies. IP&C normally are responsible for cohorting patients infected/colonized with carbapenemase-producing bacteria, implementing standard contact precautions, environmental cleaning, and attention to hand hygiene. In some regions, notification of public health officials is also indicated as per local policies.

A final report should be issued for both positive and negative results. A positive result should indicate the detection of a carbapenemase-producing organism. The carbapenemase type should also be included if a molecular or immunoassays were used. A result comment may be appended to indicate the results reflect colonization and not infection and do not require treatment unless the patient develops signs of infection such as sepsis. Susceptibility testing is not normally performed on surveillance specimens. For culture-based methodologies, bacterial isolates should be retained for in case susceptibility testing is later required. Organism identification to species level should be performed for culture-based methodologies. Local policy will determine if isolates should be referred to a reference laboratory for molecular typing.

CONCLUSIONS

The carbapenems remain some of the most effective options available for treating patient with serious infections due to Gram-negative bacteria. Carbapenemases are the primary mechanism driving carbapenem resistance among Gram-negative bacteria. Therefore, clinical microbiology laboratories must be able to detect and report carbapenemases among the Enterobacterales, P. aeruginosa, and Acinetobacter species. This will save lives and prevent the spread of AMR bacteria in hospitals and the community setting. Carbapenemase detection is currently required for patient management, the rapid implementation of IP&C protocols and for epidemiologic purposes. Laboratory testing methodology that identifies the type of carbapenemases is also important to guide therapeutic management of patients infected with such isolates. This is due to the spectrum of activity of the newer β-lactam–β-lactamase inhibitor combinations. Furthermore, carbapenemase detection is important for laboratory reporting practices.

There is not a “one size fits all” laboratory approach for the detection of bacteria with carbapenemases, and institutions need to determine what fits best with the goals of their antimicrobial stewardship and IP&C programs. Luckily, there are several options and approaches available for clinical laboratories to choose methods that best suits their individual needs. A laboratory approach to detect carbapenemases among bacterial isolates (i.e., Enterobacterales, P. aeruginosa, and Acinetobacter species) consists of two steps, namely a screening process (i.e., not susceptibilities to ertapenem, meropenem, and/or imipenem), followed by a confirmation test (i.e., phenotypic, genotypic, or proteomic methods) for the presence of a carbapenemase.

Phenotypic tests, in general terms, are cost-effective, easy to perform and interpret, and can easily be introduced into the workflow of a clinical microbiology laboratory. For the rapid phenotypic confirmation of carbapenemases among the Enterobacterales, the standardized Carba NP test as depicted in the CLSI 2015 guidelines, is a good choice and provides results within minutes. The mCIM also performs adequately but is more time consuming and adds hours to the reporting of the result. Unfortunately, these phenotypic methods are not able to distinguish between the different carbapenemase types. Several modifications to distinguish between MBLs and non-MBLs are available, but they do not perform optimally. Genomic and proteomic confirmation methods have excellent sensitivities and specificities but are unfortunately rather expensive. Molecular tests also have the advantage of being performed directly from clinical specimens, reducing the TAT to result and allow for more rapid initiation of IP&C and ASP initiatives.

Biographies

Dr. Patricia (Trish) Simner, PhD, D(ABMM), is a Professor of Pathology and Infectious Diseases at the Johns Hopkins University School of Medicine and the Director of the Medical Bacteriology and Infectious Diseases Sequencing Laboratories at the Johns Hopkins Hospital. She is widely regarded internationally as an expert in the field of mechanisms of antimicrobial resistance. Her research has focused on understanding the molecular epidemiology and mechanisms of resistance of gram-negative bacteria, in particular those harboring β-lactamases. She also is involved in investigating novel diagnostic tools to rapidly identify infectious pathogens and in developing next-generation sequencing modalities as methodology in clinical microbiology laboratories. She has published greater than 150 peer-reviewed manuscripts in this field. Dr. Simner is an Editor for the Journal of Clinical Microbiology and a Section Editor for the Manual of Clinical Microbiology and Clinical Microbiology Procedures Handbook.

Dr Johann Pitout is a Professor at the Cummings Medical School, University of Calgary and Medical Microbiologist at Alberta Precision Laboratories, Calgary, Alberta, Canada. His main research interests are resistance to antimicrobial agents among Gram-negative bacteria especially the laboratory detection, characterization, molecular epidemiology, and evolution of bacteria with b-lactamases. He has been involved in near patient testing, population-based studies, and the roles of MDR clones and plasmids among bacteria producing the extended-spectrum b-lactamases and carbapenemases. Dr Pitout has been tracking successful MDR clones globally. More recent research projects involve the underlying reasons for the success of such clones and ways to predict the next global MDR clone. He is an Editor for Emerging Infectious Diseases and serves on the Editorial boards of Journal of Clinical Microbiology and Diagnostic Microbiology and Infectious Diseases. He has published over 210 peer reviewed manuscripts in this field.

Dr. Tanis Dingle is a clinical microbiologist at Alberta Precision Laboratories - Public Health Laboratory and oversees antimicrobial resistance surveillance and the mycology laboratory. Dr. Dingle currently holds a primary appointment as Clinical Associate Professor in the Department of Pathology and Laboratory Medicine in the Cumming School of Medicine at the University of Calgary. Dr. Dingle is certified as a diplomate of the American Board of Medical Microbiology and is a Fellow of the Canadian College of Microbiologists. Dr. Dingle’s research interests include antimicrobial susceptibility testing and antimicrobial resistance detection. Other areas of research focus include the development of diagnostic methods and algorithms for fungal infection diagnosis and understanding the changing epidemiology of endemic mycoses.