CARBAPENEMASE

INTRODUCTION Carbapenem antibiotics have an important antibiotic niche in that they

retain activity against the chromosomal cephalosporinases and extended-spectrum beta-
lactamases found in many gram-negative pathogens [1,2]. The emergence of carbapenem-
hydrolyzing beta-lactamases has threatened the clinical utility of this antibiotic class and brings
us a step closer to the challenge of "extreme drug resistance" in gram-negative bacilli [3].
Issues related to carbapenemases will be reviewed here. Penicillinases and cephalosporinases are
discussed in detail separately. (See "Extended-spectrum beta-lactamases".)
CLASSIFICATION Carbapenemases are carbapenem-hydrolyzing beta-lactamases that
confer resistance to a broad spectrum of beta-lactam substrates, including carbapenems. This
mechanism is distinct from other mechanisms of carbapenem resistance such as impaired
permeability due to porin mutations, although the susceptibility patterns for isolates with a
carbapenemase and those with porin mutations can be identical.
The carbapenemases have been organized based on amino acid homology in the Ambler
molecular classification system. Class A, C, and D beta-lactamases all share a serine residue in
the active site, while Class B enzymes require the presence of zinc for activity (and hence are
referred to as metallo-beta-lactamases). Classes A, B, and D are of greatest clinical importance
among nosocomial pathogens.
Class A beta-lactamases Class A beta-lactamases are characterized by their hydrolytic
mechanisms that require an active-site serine at position 70 [4]. These include penicillinases and
cephalosporinases in the TEM, SHV, and CTX-M-type groups (which do not hydrolyze
carbapenems), as well as additional groups that possess beta-lactamase (including
carbapenemase) activity [1,5]. (See "Extended-spectrum beta-lactamases".)
Class A beta-lactamases with carbapenemase activity may be encoded on chromosomes or
plasmids. Chromosomally-encoded enzymes include SME (Serratia marcescens enzyme), NMC
(non-metalloenzyme carbapenemase) and IMI (imipenem-hydrolyzing) beta-lactamases. SME
have been recovered in a small number of S. marcescens isolates, while IMI and NMC have been
identified among Enterobacter isolates [6-8]. Plasmid-encoded enzymes include KPC (Klebsiella
pneumoniae carbapenemase) and GES (Guiana extended spectrum). GES has been described in
P. aeruginosa and K. pneumoniae [5,9-11]. (See 'Klebsiella pneumoniae carbapenemase (KPC)'
below.)
Klebsiella pneumoniae carbapenemase (KPC) The most clinically important of the Class A
carbapenemases is the Klebsiella pneumoniae carbapenemase (KPC) group (table 1). These
enzymes reside on transmissible plasmids and confer resistance to all beta-lactams [9]. Several
different variants of KPC enzymes have been identified. Some of the variants hydrolyze beta-
lactams at varying rates, which may contribute to different susceptibility profiles in KPC-
producing bacteria when tested in vitro [12,13]. KPC can be transmitted from Klebsiella to other
genera, including E. coli, Pseudomonas aeruginosa, Citrobacter, Salmonella, Serratia, and
Enterobacter spp. [14-19].
Class B beta-lactamases Class B beta-lactamases are also known as the metallo-beta-
lactamases (MBLs), which are named for their dependence upon zinc for efficient hydrolysis of
beta-lactams. As a result, MBLs can be inhibited by EDTA (an ion chelator), although they
cannot be inhibited by beta-lactamase inhibitors such as tazobactam, clavulanate, and sulbactam.
The first MBL, IMP-1, was described in Japan in 1991 [20]. Subsequently, additional groups of
acquired MBLs have been identified: IMP, VIM, GIM, SPM, and SIM. There are a number of
variants within each MBL group (for example, there are 19 IMP variants within the IMP group)
[21,22].
There are both naturally occurring and acquired MBLs. Naturally occurring MBLs are
chromosomally encoded and have been described in Aeromonas hydrophilia, Chryseobacterium
spp., and Stenotrophomonas maltophilia [21]. Acquired MBLs consist of genes encoded on
integrons residing on large plasmids that are transferable between both species and genera [4,23-
28]. In a hospital outbreak involving 62 patients (including 40 intensive care unit patients), for
example, an MBL gene (bla IMP-4) spread among seven different gram-negative genera
(Serratia, Klebsiella, Pseudomonas, Escherichia, Acinetobacter, Citrobacter, and Enterobacter)
[23,29].
New Delhi metallo-beta-lactamase (NDM-1) Enterobacteriaceae isolates carrying a novel
MBL gene, the New Delhi metallo-beta-lactamase (NDM-1), were first described in December
2009 in a Swedish patient hospitalized in India with an infection due to Klebsiella pneumoniae
(table 1) [30].
The gene encoding this MBL is located in a very mobile genetic element, and the pattern of
spread appears to be more complex and more unpredictable than that of the gene encoding KPC
[30,31]. Furthermore, the large number of resistance determinants in the isolates studied raise
concern that this gene is an important emerging resistance trait [32]. In general, bacteria
containing NDM-1 have tested susceptible to colistin or tigecycline, though such susceptibility
may be short-lived.
In addition to K. pneumoniae, NDM-1 has also been identified in other Enterobacteriaceae
(including E. coli and Enterobacter cloacae) [33] as well as non-Enterobacteriaceae (including
Acinetobacter) [34].
Class D beta-lactamases Class D beta-lactamases are also referred to as OXA-type enzymes
because of their preferential ability to hydrolyze oxacillin (rather than penicillin) [35]. Enzymes
in this group are variably affected by the beta-lactamase inhibitors clavulanate, sulbactam, or
tazobactam. OXA carbapenemases have been identified in Acinetobacter baumannii [35-44] and
Enterobacteriaceae (especially K. pneumoniae, E. coli, and E. cloacae) [45].
Among the heterogeneous OXA group (which includes more than 100 enzymes), five
subfamilies have been identified with varying degrees of carbapenem-hydrolyzing activity:
OXA-23, OXA-24/OXA40, OXA-48, OXA-58, and OXA-51 (table 2). The first four groups are
carried on transmissible plasmids, while the last group, OXA-51, is chromosomally encoded.
While most isolates of A. baumannii possessing an OXA-23, -24/40, or -58 type carbapenemase
are resistant to carbapenems, Enterobacteriaceae with OXA-48-type enzymes have variable
susceptibility to these agents. Expression of a promoter insertion element (ISAba1) in OXA-23
and OXA-51 likely contributes to carbapenem resistance [38].
EPIDEMIOLOGY
Distribution
Klebsiella pneumoniae carbapenemases The Klebsiella pneumoniae carbapenemase (KPC) is
the most common carbapenemase in the United States (table 1); as of 2010, KPC-production has
been identified in isolates from 36 states [46,47]. It was first described in a clinical isolate of K.
pneumoniae in the late 1990s in North Carolina [9,48]. Shortly thereafter, several reports
described hospital outbreaks in the northeastern United States involving K. pneumoniae carrying
KPC-2, an enzyme later found to be identical to the initially described KPC [49-52].
Subsequently, an outbreak involving K. pneumoniae with KPC-3 (which differs from KPC-
1/KPC-2 by a single amino acid) was described in New York City [53].
In 2011, an outbreak of carbapenem-resistant K. pneumoniae occurred at the US National
Institutes of Health Clinical Center that affected 18 patients, 11 of whom died [54]. The CDC
reported that the proportion of Enterobacteriaceae that were carbapenem resistant increased from
1 to 4 percent between 2001 and 2011; the proportion of carbapenem-resistant Klebsiella
increased from 2 to 10 percent [55]. Los Angeles County noted 675 cases of carbapenem-
resistant K. pneumoniae between June 2010 and May 2011; the incidence was higher in long-
term acute care facilities than in acute care hospitals [56].
KPC-possessing isolates have also been increasingly recovered from other regions of the world,
including Europe [57,58], Asia [14,59,60], and South America [15,61].
Metallo-beta-lactamases Metallo-beta-lactamases (MBLs) were initially described in Japan in
1991 [20]. MBLs have since been described in other parts of Asia, Europe, North America,
South America, and Australia [21,62-66]. The transfer of patients between hospitals and the
increase in international travel may be important factors in the geographical dissemination of
MBL genes [21,27,62,67,68].
The MBL gene, the New Delhi metallo-beta-lactamase (NDM-1) (table 1), was first described in
December 2009 in a K. pneumoniae isolate from a Swedish patient who had been hospitalized in
India [30]. Subsequent reports have included patients who have traveled and undergone
procedures (so called "medical tourism") in India and Pakistan [33], as well as cases reported in
Asia, Europe, North America, and Australia [31,33,47,69,70].
In the United States, between January 2009 and February 2011, seven Enterobacteriaceae
isolates with NDM-1 production were reported to the Centers for Disease Control and Prevention
(CDC) [47]. These were all identified in patients who had traveled to India or Pakistan, the
majority of whom received medical care there.
Class D carbapenemases While A. baumannii carrying OXA-23-, OXA-24/40-, and OXA-58-
type carbapenemases are especially problematic in Europe, they have also been recovered from
medical centers in Eastern Asia, the Middle East, Australia, South America, and the United
States [35]. The first isolate of K. pneumoniae with OXA-48 was identified in Turkey; since
then, hospital outbreaks from that country have been reported [71]. Enterobacteriaceae with
OXA-48-type enzymes have subsequently been recovered in Europe, the Middle East, and
Northern Africa.
Risk factors Carbapenemase-producing organisms can arise from previously carbapenemase-
negative strains by acquisition of genes from other bacteria. Use of broad spectrum
cephalosporins and/or carbapenems is an important risk factor for the development of
colonization or infection with such pathogens [29,67,72]. As an example, in one case-control
study, 86 percent of patients with a KPC-producing Enterobacteriaceae isolate (n = 91) had a
history of cephalosporin use in the past three months, compared with 69 percent of those with
extended-spectrum beta-lactamase-producing isolates and 27 percent of those with fully
susceptible isolates [73].
Although a risk factor, prior receipt of carbapenems is not essential for acquisition of these
strains. Reported carbapenem use among patients prior to the isolation of MBL, for example,
varies from 15 to 75 percent [59,60].
Additional risk factors that have been associated with infection or colonization with a
carbapenemase-producing organism include the following [16,18,23,49,60,61,74-76]:
trauma
diabetes
malignancy
organ transplantation
mechanical ventilation
indwelling urinary or venous catheters
overall poor functional status or severe illness
Clinicians should be also aware of the possibility of NDM-1-producing Enterobacteriaceae in
patients who have received medical care in India and Pakistan [31]. (See 'Metallo-beta-
lactamases' above.)
Transmission Many carbapenemases reside on mobile genetic elements, such as transposons
or plasmids, and have the potential for widespread transmission to other isolates and genera of
bacteria. Furthermore, Enterobacteriaceae, which may harbor carbapenemase-encoding genes,
can spread from person to person.
One particular clone of K. pneumoniae that carries the KPC gene has been reported as the
predominant isolate across several geographic areas, suggesting cross-infection within and
outside of healthcare systems [71].
Limited data using DNA fingerprinting and pulse-field gel electrophoresis also suggest cross-
transmission of bacteria with MBLs within hospitals [28]. This was illustrated in a study of 66
MBL-positive isolates from 54 hospitalized patients in a hospital outbreak [63]. Environmental
screening isolated MBL-producing organisms from sinks and stethoscopes, suggesting these as
possible environmental reservoirs; interestingly, there were no positive cultures from the hands
of the 10 healthcare workers screened. The outbreak was curtailed following intensive
environmental cleaning with hypochlorite, replacement of poorly designed sinks, and
disassembling and cleaning of stethoscopes. In another outbreak, environmental screening
demonstrated a single positive swab from a ventilator; all other environmental swabs of hands,
gowns, medical fluids, air and water samples, and endoscopes were negative [76].
NDM-1-positive bacteria have been identified in public water supplies in India, highlighting the
potential for environmental dissemination, and the importance of environmental surveillance
[77]. In addition, two isolates of Pseudomonas aeruginosa containing an MBL gene (bla VIM)
have been also detected from aquatic sources (one isolate from a river and the second from
sewage), raising the possibility of aquatic reservoirs for these organisms [78].
Patients themselves may also serve as an important reservoir for resistant Enterobacteriaceae, as
intestinal colonization with carbapenemase-producing organisms has been reported
[21,28,79,80].
DETECTION
Clinical laboratory testing Most K. pneumoniae and E. coli without carbapenemases have
minimum inhibitory concentrations (MICs) to imipenem and meropenem that are 0.5 mcg/mL.
Thus, the identification of E. coli or K. pneumoniae with overt resistance to any of the
carbapenems should raise suspicion that it may be harboring a carbapenemase enzyme. Certain
susceptibility patterns in other Enterobacteriaceae may be suggestive of the presence of a
carbapenemase. As an example, recovery of an isolate that is susceptible to third generation
cephalosporins but resistant to imipenem should raise the possibility of an underlying Serratia
marcescens enzyme (SMC) in Serratia species and a non-metalloenzyme carbapenemase (NMC)
or imipenem-hydrolyzing beta-lactamase (IMI) in Enterobacter species [6-8].
Detection of carbapenemase-producing Enterobacteriaceae can be problematic. Some isolates
have MICs that fall just below the traditional breakpoints for susceptibility and therefore may not
be recognized as carbapenemase producers by the clinical laboratory [23,63,81-84]. In one
report, up to 87 percent of K. pneumoniae carbapenemase (KPC)-producing K. pneumoniae were
reported to be susceptible to carbapenems according to breakpoints typically in use prior to 2011
[85]. Detection of Enterobacteriaceae harboring OXA-48-type enzymes can be similarly
challenging [45].
In 2010, the Clinical and Laboratory Standards Institute (CLSI) updated new MIC and disk
diffusion breakpoints for the Enterobacteriaceae based on pharmacokinetic-pharmacodynamic
data of recommended dosage regimens [86-89]. The new MIC breakpoints are one to three
doubling dilutions lower than the original breakpoints, and the new disk diffusion criteria include
larger zone diameters than those in previous guidelines. Thus, many organisms that previously
would have been categorized as susceptible using the former breakpoints may now be considered
intermediate or resistant. In clinical laboratories that have implemented these new breakpoints,
additional tests to detect extended-spectrum beta-lactamases and carbapenemases need not be
routinely performed for clinical management. However, phenotypic or genotypic testing for
carbapenemases may be useful for isolates that test in the susceptible range to carbapenems but
have reduced susceptibility. Additionally, carbapenemase testing may be performed for infection
control purposes, and isolates may be forwarded through state public health laboratories to the
United States Centers for Disease Control and Prevention (CDC) for further characterization
[31]. (See 'Other phenotypic tests' below.)
P. aeruginosa and A. baumannii can become resistant to carbapenems without harboring a
carbapenemase; determining which isolates have a carbapenemase based on susceptibility
profiles can be difficult. In regions where metallo-beta-lactamases (MBL) and OXA enzymes are
endemic, isolates of P. aeruginosa and A. baumannii that are highly resistant to carbapenems,
cephalosporins, and penicillins should be suspected of carrying one of these enzymes. In
addition, presence of an underlying MBL should be considered if an isolate retains susceptibility
to aztreonam [90].
Other phenotypic tests For laboratories that have implemented the updated 2010 CLSI
breakpoints for carbapenem susceptibility, additional tests are no longer routinely recommended
for detection of carbapenemase-producing organisms. For clinical laboratories that have not yet
implemented the updated CLSI breakpoints, additional tests are often employed to improve the
detection of carbapenemases.
For example, the modified Hodge test can be used to detect carbapenemase activity [91,92]. This
test involves streaking a clinical isolate near an imipenem disk that has been placed on an agar
plate containing a susceptible control organism. Growth of the control organism around the
imipenem disk in the region of the streak suggests carbapenemase activity in the clinical isolate.
However, the modified Hodge test has several drawbacks. Assessment of the test result may be
difficult and subject to individual interpretation. Additionally, it has poor sensitivity for metallo-
beta-lactamase (MBL) detection, although this can be improved with the addition of zinc [93].
False positive Hodge tests have also been reported [94].
Other methods of carbapenemase detection include rapid biochemical tests [95], mass
spectrometric detection, and microbiological methods that take advantage of particular
characteristics of carbapenemases, such as their zinc dependence [81]. The particular test
employed varies widely across different laboratories.
Other tests are not always reliable for detecting MBLs in organisms with apparent carbapenem
susceptibility [21,27,96]. These include a combination disc tests using imipenem and EDTA
discs or using two carbapenem discs (including one with EDTA incorporated) [97-100] and the
commercially available MBL E-test.
Direct genotypic identification Identification of specific carbapenemases can be accomplished
utilizing molecular techniques [23,38,83,101-103]. These include multiplex polymerase chain
reaction (PCR) assays and DNA microarrays that can screen at once for several different types of
enzymes, including Klebsiella pneumoniae carbapenemases, specific MBLs, and OXA-type
carbapenemases. Detection of organisms harboring these enzymes will be greatly improved as
these technologies become incorporated into clinical practice.
CLINICAL DISEASE Carbapenemase-producing organisms can cause clinical infections or
asymptomatic colonization [18,59]. Blood stream infections, ventilator-associated pneumonia,
urinary tract infection, and central venous catheter infections have been described [23,28,63,99].
These organisms have been isolated from respiratory tract specimens, abdominal swabs,
catheters, abscesses, urine, and surgical wounds [28,29,59,63,99,104,105]. Sporadic hospital-
acquired infections and outbreaks due to hospital-based clonal spread have been described in
both tertiary and community hospitals [23,29,63,106].
The above clinical syndromes are discussed in more detail in the corresponding topic reviews.
TREATMENT The optimal treatment of infection due to carbapenemase-producing
organisms is uncertain, and antibiotic options are limited. Management of patients with
infections due to carbapenemase-producing organisms should be done in consultation with an
expert in the treatment of multidrug resistant bacteria.
Because the presence of a KPC or metalloenzyme carbapenemase confers resistance to all
penicillins, cephalosporins, and carbapenems, selection of antibiotic therapy should be tailored to
antimicrobial susceptibility results for agents outside the beta-lactam and carbapenem classes.
Additional antibiotic susceptibility testing should be requested for colistin or polymyxin B,
aztreonam, tigecycline, and fosfomycin (particularly for urinary tract isolates) [107,108].
Furthermore, despite limited clinical data, for patients with severe infections (including
bacteremia) due to a carbapenemase-producing gram-negative organism, we suggest using
combination antimicrobial therapy with two or more agents because of the high associated
mortality, the concern for emergence of resistance during monotherapy, and the lack of clearly
effective single drug alternatives. (See 'Therapy for carbapenemase-producing
Enterobacteriaceae' below and 'Therapy for carbapenem-resistant A. baumannii and P.
aeruginosa' below.)
Susceptibility testing The reliability of carbapenem susceptibility testing to guide clinical
management depends on the breakpoints utilized by the clinical laboratory. If the current Clinical
and Laboratory Standards Institute (CLSI) breakpoints are being implemented to determine
carbapenem susceptibility, carbapenemase-producing organisms generally will test intermediate
or resistant, and so the susceptibility testing results can be a reliable guide for antibiotic choice.
However, if older (before 2010) breakpoints are still being utilized, carbapenems and beta-
lactamase inhibitors should not be used alone against organisms that have phenotypic evidence
of a carbapenemase (eg, positive Hodge test) or are otherwise suspected of harboring a
carbapenemase, regardless of susceptibility results [23,27]. Although some carbapenemase-
producing Enterobacteriaceae may have carbapenem minimum inhibitory concentrations (MICs)
within the susceptible range using older breakpoints, the MICs vary with the susceptibility
testing method and can rise as the inoculum increases [104]; thus, carbapenems cannot be relied
upon as single agents in such cases. (See 'Detection' above.)
Current beta-lactamase inhibitors are ineffective against K. pneumoniae carbapenemase (KPC),
metallo-beta-lactamase (MBL), and most OXA-type beta-lactamases. Because the presence of a
KPC or metalloenzyme carbapenemase confers resistance to all penicillins, cephalosporins, and
carbapenems, selection of antibiotic therapy should be tailored to antimicrobial susceptibility
results for agents outside the beta-lactam and carbapenem classes. Unfortunately, resistance
genes for other antibiotics, including aminoglycosides and fluoroquinolones, are frequently
present in carbapenemase-producing strains [21,28]. For KPC-carrying K. pneumoniae,
resistance rates of 98 percent have been reported for the fluoroquinolones, and approximately 50
percent are resistant to gentamicin and amikacin [109].
Additional antibiotic susceptibility testing should be requested for colistin or polymyxin B,
aztreonam, tigecycline, and fosfomycin (particularly for urinary tract isolates).
Therapy for carbapenemase-producing Enterobacteriaceae As above, selection of antibiotic
therapy for infections due to carbapenemase-producing Enterobacteriaceae should be tailored to
antimicrobial susceptibility results for agents outside the beta-lactam and carbapenem classes.
However, options are limited and, because of concurrent resistance genes to other commonly
used antibiotic classes (eg, fluoroquinolones), often only include the polymyxins (eg, colistin),
aztreonam, tigecycline, and, for urinary tract infections, fosfomycin, depending on susceptibility
testing results. Location and severity of infection also influence antibiotic regimen choice.
Isolates causing urinary tract infections can often be successfully treated with an aminoglycoside
or fosfomycin, assuming the isolates retain susceptibility to one of these agents [108,110]. In one
study of a panel of 81 carbapenem-resistant Enterobacteriaceae of various types and species, 60
percent tested susceptible to fosfomycin [108]. In a small prospective study of 11 patients, an
intravenous preparation of fosfomycin was used to successfully treat various carbapenem-
resistant K. pneumoniae infections [111]. Of note, intravenous fosfomycin is not available in
many countries, including the United States and Australia.
Aztreonam may be a useful for patients whose isolates carry a metallo-beta-lactamase and
demonstrate in vitro susceptibility to this agent [112]. However, clinical experience with
aztreonam for treatment of serious infections due to MBL-producing bacteria is very limited.
For all other patients with severe infections (including bacteremia), despite limited data, we
suggest using combination antimicrobial therapy. The use of two or more agents is recommended
because of the high associated mortality, the concern for emergence of resistance during
monotherapy, and the lack of clearly effective single drug alternatives. In the few published
reports, the combination of a polymyxin with tigecycline has been described the most and is thus
a reasonable first choice, assuming the isolate is susceptible to both. For patients who are
critically ill or have deep-seated infections (eg, meningitis), we often add a carbapenem as a third
agent to this combination regimen. If the isolate is resistant to the polymyxins, we generally use
tigecycline with a second agent (most commonly a carbapenem). Conversely, if the isolate is
resistant to tigecycline, we generally use a polymyxin with a second agent (most commonly a
carbapenem or rifampin).
Clinical evidence suggests that treatment with combination therapy may improve mortality [113-
119]. In a retrospective study of 125 patients with bacteremia due to K. pneumoniae confirmed
by polymerase chain reaction to harbor the KPC gene, the overall mortality rate at 30 days was
42 percent [117]. The mortality rate was lower among patients who received combination
therapy with two or more drugs (27 of 79 [34 percent]) compared with monotherapy with
colistin, tigecycline, or gentamicin (25 of 46 [54 percent]). Patients treated with a combination of
a polymyxin plus tigecycline had a mortality rate of 30 percent (7 of 23), while the regimen of
colistin, tigecycline, and extended-infusion meropenem (a dose of 2 grams infused over three or
more hours every eight hours) was associated with the lowest mortality rate (2 of 16 [12.5
percent]). Similar findings were observed in a smaller retrospective study of 41 patients with
KPC-producing K. pneumoniae bacteremia, in which the mortality rate was 13 percent (2 of 15)
with definitive combination therapy (generally a polymyxin or tigecycline with a carbapenem)
compared with 58 percent (11 of 19) with monotherapy [113]. A systematic review of 20
observational studies also concluded that combination therapy may offer a survival advantage in
severely ill patients [119]. Additionally, in vitro studies have demonstrated synergy with
combination regimens [109,120],
Unfortunately, emergence of polymyxin- or tigecycline-resistant, carbapenemase-producing K.
pneumoniae has been reported, and development of resistant K. pneumoniae in patients receiving
polymyxin therapy has been documented [121-124]. Infection with a polymyxin-resistant strain
has been found to be an independent risk factor for mortality [124].
Therapy for carbapenem-resistant A. baumannii and P. aeruginosa Carbapenem-resistant A.
baumannii and P. aeruginosa are typically resistant to all beta-lactams and fluoroquinolones. The
intrinsic resistance of these organisms further limits antibiotic options. Although sulbactam has
been used to treat some infections due to A. baumannii, most multidrug-resistant isolates of A.
baumannii have reduced susceptibility to this agent [39,53,109,125]. Aminoglycosides may be
useful, particularly for urinary tract infections, assuming susceptibility is retained for one of
these antibiotics. Aztreonam may also be considered for MBL-possessing P. aeruginosa, but
supporting clinical data is lacking. P. aeruginosa are intrinsically resistant to tigecycline, the
emergence of tigecycline-resistant A. baumannii has been reported [126]. Most multidrug-
resistant A. baumannii and P. aeruginosa retain susceptibility for the polymyxins. Therefore, the
polymyxins (ie, colistin and polymyxin B) are usually the cornerstones for therapy. Combination
therapy is also advised when polymyxins are used to treat multidrug-resistant A. baumannii and
P. aeruginosa.
Detailed discussion of the management of multi-drug resistant A. baumannii and P. aeruginosa
are found in detail elsewhere. (See "Acinetobacter infection: Treatment and prevention", section
on 'Resistant isolates' and "Principles of antimicrobial therapy of Pseudomonas aeruginosa
infections", section on 'Strategies for multidrug resistant organisms'.)
For ventilator-associated pneumonia due to these organisms, adjunctive use of aerosolized
colistin has been evaluated. This is also discussed in detail elsewhere. (See "Colistin: An
overview", section on 'Inhaled administration' and "Acinetobacter infection: Treatment and
prevention", section on 'Pneumonia' and "Pseudomonas aeruginosa pneumonia", section on
'Inhaled antibiotics'.)
INFECTION CONTROL Hospitalized patients infected or colonized with carbapenemase-
producing bacteria should be placed on contact precautions [28,61,63,79,127]. Other standard
measures, such as hand hygiene, minimizing the use of invasive devices, and antimicrobial
stewardship, are important to infection control in general and likely to limit spread of resistant
organisms.

Screening high-risk patients to detect rectal colonization has been suggested as an important
infection control modality [21,28,79,80]. Several studies have documented reduced transmission
of KPC-producing Klebsiella pneumoniae when comprehensive infection control protocols,
including active surveillance, have been enacted [128-130]. Although the impact of surveillance
itself is difficult to assess, it may be useful in the setting of outbreaks due to carbapenemase-
resistant organisms, as recommended by the United States Centers for Disease Control and
Prevention (CDC), or among patients with recent travel to areas where carbapenemases are more
prevalent. (See "General principles of infection control".)
SUMMARY AND RECOMMENDATIONS
The carbapenem-hydrolyzing beta-lactamase is an important emerging mechanism of
antimicrobial resistance among nosocomial gram-negative pathogens. These enzymes are
classified on the basis of their amino acid homology; Classes A, B, and D are of greatest clinical
importance. (See 'Classification' above.)
The clinically most important of the Class A carbapenemases is the Klebsiella pneumoniae
carbapenemase (KPC) group, which has been implicated in several outbreaks (table 1). (See
'Class A beta-lactamases' above and 'Klebsiella pneumoniae carbapenemase (KPC)' above.)
Class B beta-lactamases are known as the metallo-beta-lactamases (MBLs), which are named
for their dependence upon zinc for efficient hydrolysis of beta-lactams. The New Delhi metallo-
beta-lactamase (NDM-1) is an important emerging carbapenemase in this group (table 1). (See
'Class B beta-lactamases' above and 'New Delhi metallo-beta-lactamase (NDM-1)' above.)
Class D beta-lactamases are referred to as OXA-type enzymes because of their preferential
ability to hydrolyze oxacillin (rather than penicillin). (See 'Class D beta-lactamases' above.)
Implementation of the lower 2010 Clinical and Laboratory Standards Institute (CLSI)
breakpoints for carbapenem susceptibility improves the detection of carbapenemase-producing
Enterobacteriaceae. Identifying strains of P. aeruginosa or A. baumannii with carbapenemases
can be difficult. Isolates resistant to penicillins, cephalosporins, and carbapenems with retained
susceptibility to aztreonam should be suspected of carrying an MBL. (See 'Detection' above.)
Use of broad spectrum cephalosporins and/or carbapenems is an important risk factor for the
development of colonization or infection with carbapenemase-producing organisms, although
prior receipt of carbapenems is not essential for acquisition of these strains. (See 'Risk factors'
above.)
Carbapenemase-producing organisms can cause clinical infections or asymptomatic
colonization. Carbapenemase-producing bacteria have been implicated in a variety of infections,
including bacteremia, ventilator-associated pneumonia, urinary tract infection, and central
venous catheter infection. (See 'Clinical disease' above.)
Selection of antibiotic therapy should be tailored according to the antimicrobial susceptibility
test result. In particular, antibiotic susceptibility testing should be requested for colistin,
aztreonam, tigecycline, and fosfomycin. (See 'Susceptibility testing' above.)
For patients with serious infections, including bacteremia, due to carbapenemase-producing
Enterobacteriaceae, we suggest using combination therapy with two or more antimicrobial agents
instead of monotherapy (Grade 2C). Limited clinical evidence suggests that treatment with a
regimen that combines a polymyxin, tigecycline, or a carbapenem (or all three) is associated with
lower mortality than single-agent regimens for infections with carbapenemase-producing
Enterobacteriaceae. (See 'Therapy for carbapenemase-producing Enterobacteriaceae' above.)
Polymyxins are the cornerstones of therapy for carbapenem-resistant A. baumannii and P.
aeruginosa and are generally used in combination with another agent. (See 'Therapy for
carbapenem-resistant A. baumannii and P. aeruginosa' above and "Acinetobacter infection:
Treatment and prevention", section on 'Resistant isolates' and "Pseudomonas aeruginosa
pneumonia", section on 'Strategies for drug resistant isolates'.)
Patients infected or colonized with carbapenemase-producing bacteria should be placed on
contact precautions. Screening high-risk patients to detect rectal colonization may also be helpful
in controlling transmission. (See 'Infection control' above.)