See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/260396311

Modified CLSI Extended-Spectrum -Lactamase (ESBL) Confirmatory Test for

Phenotypic Detection of ESBLs among Enterobacteriaceae Producing Various
-Lactamases

Article  in  Journal of Clinical Microbiology · February 2014

DOI: 10.1128/JCM.03361-13 · Source: PubMed

CITATIONS READS

119 4,049

7 authors, including:

Georgia Vrioni Vassiliki Koumaki

National and Kapodistrian University of Athens National and Kapodistrian University of Athens
168 PUBLICATIONS   2,527 CITATIONS    6 PUBLICATIONS   121 CITATIONS   

SEE PROFILE SEE PROFILE

Theodoros Pittaras Spyros Pournaras

24 PUBLICATIONS   404 CITATIONS   
National and Kapodistrian University of Athens
237 PUBLICATIONS   8,884 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Antibioitc resistance View project

Epidemiology of MRSA View project

All content following this page was uploaded by Vassiliki Koumaki on 25 May 2017.

The user has requested enhancement of the downloaded file.

Modified CLSI Extended-Spectrum ␤-Lactamase (ESBL) Confirmatory
Test for Phenotypic Detection of ESBLs among Enterobacteriaceae
Producing Various ␤-Lactamases
Aggeliki Poulou,a,b Evgenia Grivakou,a Georgia Vrioni,a Vassiliki Koumaki,a Theodoros Pittaras,a Spyros Pournaras,a
Athanassios Tsakrisa
Department of Microbiology, Medical School, University of Athens, Athens, Greecea; Department of Microbiology, General Hospital of Serres, Serres, Greeceb

The worldwide dissemination of Enterobacteriaceae producing AmpC ␤-lactamases and carbapenemases makes difficult the
phenotypic detection of extended-spectrum ␤-lactamases (ESBLs), as they may be masked by these additional enzymes. A modi-
fication of the CLSI ESBL confirmatory test was developed and evaluated in a comparative study for its ability to successfully
detect ESBLs among Enterobacteriaceae producing various carbapenemases (Klebsiella pneumoniae carbapenemase [KPC],
VIM, NDM, and OXA-48) and plasmidic or derepressed AmpCs. The modified CLSI ESBL confirmatory test was performed with
cefotaxime and ceftazidime disks with and without clavulanate, on which both boronic acid (BA) and EDTA were dispensed. A
total of 162 genotypically confirmed ESBL-positive Enterobacteriaceae isolates (83 carbapenemase/ESBL producers, 25 AmpC/
ESBL producers, and 54 ESBL-only producers) were examined. For comparison, 139 genotypically confirmed ESBL-negative En-
terobacteriaceae isolates (94 of them possessed carbapenemases and 20 possessed AmpCs) were also tested. The standard CLSI
ESBL confirmatory test was positive for 106 of the 162 ESBL producers (sensitivity, 65.4%) and showed false-positive results for
4 of the 139 non-ESBL producers (specificity, 97.1%). The modified CLSI ESBL confirmatory test detected 158 of 162 ESBL pro-
ducers (sensitivity, 97.5%) and showed no false-positive results for non-ESBL producers (specificity, 100%). The findings of the
study demonstrate that the modified CLSI ESBL confirmatory test using antibiotic disks containing both BA and EDTA accu-
rately detects ESBLs in Enterobacteriaceae regardless of the coexistence of additional ␤-lactam resistance mechanisms.

xtended-spectrum ␤-lactamases (ESBLs) are mostly plasmid-

E mediated ␤-lactamases that efficiently hydrolyze oxyimino-
cephalosporins and monobactams, yet are inhibited by ␤-lacta-
been employed to unmask the underlying ESBLs among AmpC-
or KPC-possessing Enterobacteriaceae (16–18). ESBLs, however,
may also coexist with other ␤-lactamase types, such as metallo-␤-
mase inhibitors (1). They were first detected in Enterobacteriaceae, lactamases (MBLs) (e.g., NDM, VIM, and IMP) or both MBLs and
and nowadays various groups of ESBLs are produced by these KPCs (1, 19, 20), which may also interfere with the interpretation
microorganisms, the most common being CTX-M and SHV en- of ESBL detection methods, since they also hydrolyze extended-
zyme types (1, 2). ESBLs are increasingly reported worldwide and spectrum ␤-lactams.
have been linked to successful enterobacterial clones possessing There is a need, therefore, for an alternative method that can
great epidemic potential (1, 3). Plasmids coding for ESBLs may accurately detect ESBLs in Enterobacteriaceae, regardless of a pos-
also carry additional ␤-lactamase genes as well as genes conferring sible coexistence of additional mechanisms of resistance to ␤-lac-
resistance to other antimicrobial classes (2–4). This can limit the tams. EDTA is a chelating agent that inhibits the enzymatic activ-
chemotherapeutic options for ESBL-producing pathogens and fa- ity of MBLs, while BA inhibits the enzymatic activity of both
cilitate the inter- and intraspecies dissemination of ESBLs (3). AmpCs and KPCs (11, 19, 20, 21). We have also previously shown
Therefore, phenotypic detection of ESBLs among Enterobacteria- that the growth-inhibitory zone diameter around a meropenem
ceae species is important for epidemiological purposes as well as (MER) disk with simultaneous addition of BA and EDTA is 5 mm
for limiting the spread of resistance mechanisms. or greater of the growth-inhibitory zone diameter around the disk
The Clinical and Laboratory Standards Institute (CLSI) rec- containing MER alone when a KPC, an MBL, or both KPC and
ommends a phenotypic confirmatory combined-disk test for MBL are coexisting in a clinical isolate (19, 22). In this context, we
ESBL production in Enterobacteriaceae. It consists of measuring further modified the CLSI ESBL confirmatory test by the simulta-
the growth-inhibitory zones around both cefotaxime (CTX) and neous addition of both BA and EDTA on the antibiotic disks con-
ceftazidime (CAZ) disks with or without clavulanate (CA) for taining CTX and CAZ with or without CA and tested its sensitivity
Klebsiella pneumoniae, Klebsiella oxytoca, Escherichia coli, and Pro-
teus mirabilis (5). Different combined-disk and double-disk syn-
ergy tests based on the synergy of CA with various expanded- Received 3 December 2013 Returned for modification 2 January 2014
spectrum cephalosporins and aztreonam have also been proposed Accepted 16 February 2014
(3, 6–9). In addition, a biochemical test based on the in vitro de- Published ahead of print 26 February 2014
tection of cefotaxime hydrolysis that is inhibited by tazobactam Editor: P. Bourbeau
was recently proposed for the rapid detection of ESBLs in Entero- Address correspondence to Athanassios Tsakris, atsakris@med.uoa.gr.
bacteriaceae (10). Moreover, boronic acid (BA) compounds, well- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
known reversible inhibitors of AmpCs and Klebsiella pneumoniae doi:10.1128/JCM.03361-13
carbapenemases (KPCs) (11–15), in combination with CA have

May 2014 Volume 52 Number 5 Journal of Clinical Microbiology p. 1483–1489 jcm.asm.org 1483
Poulou et al.

and specificity among Enterobacteriaceae producing additional CLSI ESBL confirmatory test, an augmentation of ⱖ5 mm in the growth-
␤-lactamases (plasmid-mediated AmpCs or various carbapen- inhibitory zone diameter of either CTX-CA or CAZ-CA in combination
emase types) or overproducing cephalosporinases. with BA and EDTA (CTX-CA-BA-EDTA and CAZ-CA-BA-EDTA, re-
spectively) compared with the zone diameter of CTX or CAZ disks con-
taining BA and EDTA (CTX-BA-EDTA and CAZ-BA-EDTA, respec-
MATERIALS AND METHODS tively) was considered a positive result for ESBL production. It should be
Clinical isolates and antimicrobial susceptibility testing. A total of 301 noted that the concentration of BA and EDTA employed in the present
nonrepetitive (one per patient) clinical isolates of Enterobacteriaceae were study did not show any detectable effect on bacterial growth.
included in the study. The criteria for selection were devised to include a Sensitivity and specificity. The performance of the phenotypic tests
considerable number of bacteria producing various potent ␤-lactamases. for the detection of ESBLs among Enterobacteriaceae producing various
They were recovered during 2007 to 2013 from nine tertiary care Greek ␤-lactamases was evaluated using PCR along with DNA sequencing as the
hospitals. The collection consisted of K. pneumoniae (n ⫽ 174), E. coli gold standard. For each test, the sensitivity was calculated from the num-
(n ⫽ 42), P. mirabilis (n ⫽ 21), Enterobacter aerogenes (n ⫽ 24), Entero- ber of ESBL-possessing organisms that were correctly determined, while
bacter cloacae (n ⫽ 17), Serratia marcescens (n ⫽ 9), Providencia stuartii the specificity was calculated from the number of non-ESBL-possessing
(n ⫽ 9), and K. oxytoca (n ⫽ 5). A sum of 162 of these isolates were organisms that were correctly determined.
genotypically confirmed ESBL positive and the remaining 139 were geno-
typically confirmed ESBL negative. The presence of the ESBL gene was
RESULTS
determined by using previously described oligonucleotide primers and
cycling conditions (23). The identification of all isolates was confirmed by Species distribution and ␤-lactamase content. The species dis-
using the API20E system (bioMérieux, Marcy l’Etoile, France). Detailed tributions of the studied isolates and their ␤-lactamases are pre-
susceptibility analysis was carried out by the agar dilution method follow- sented in Table 1. Phenotypic and molecular testing revealed that
ing the recent CLSI guidelines and interpretative criteria (5). among the 162 ESBL-positive isolates, 83 possessed carbapen-
Molecular testing for ␤-lactamase genes. ␤-Lactamase genes were emases (35 KPC-2 producers, 22 VIM-1 producers, 10 KPC-2/
amplified in single PCRs using a panel of primers for detection of all types VIM-1 producers, 8 NDM-1 producers, and 8 OXA-48 produc-
of ESBL (SHV, TEM, CTX-M, GES. and PER), carbapenemase (KPC, ers), 19 possessed plasmid-mediated AmpCs (15 belonged to the
SME, VIM, NDM, IMP, and OXA-48), and plasmidic AmpC genes (23– clusters MOX-1, MOX-2, CMY-1, and CMY-8 to CMY-11 and 4
27). Among Enterobacter aerogenes and Enterobacter cloacae isolates, total
belonged to the clusters LAT-1 to LAT-4, CMY-2 to CMY-7, and
RNA from logarithmic-phase-grown cultures was extracted with TRI re-
agent (Ambion, Austin, TX), and reverse transcription (RT) of 1 ␮g of BIL-1), 6 hyperproduced chromosomal AmpCs, while the re-
total RNA was performed with the ThermoScript RT-PCR system (Invit- maining 54 possessed only ESBLs (6 of them were ertapenem re-
rogen, Carlsbad, CA). Derepressed AmpC-hyperproducing E. aerogenes sistant due to porin deficiency) (31). Among the 139 ESBL-nega-
and E. cloacae isolates were identified with quantitative real-time PCR tive isolates, 94 possessed carbapenemases (32 KPC-2 producers,
using the Quanti Test SYBR green (Qiagen, Hilden, Germany) and prim- 33 VIM-1 producers, 21 KPC-2/VIM-1 producers, 5 NDM-1 pro-
ers described previously (28, 29). As positive controls, we used previously ducers, and 3 OXA-48 producers), 15 possessed plasmid-medi-
characterized isolates from our collection carrying all types of tested ated AmpCs (12 belonged to the clusters LAT-1 to LAT-4, CMY-2
␤-lactamases. The PCR products were subjected to direct sequencing. to CMY-7, and BIL-1; 2 belonged to the clusters MOX-1, MOX-2,
PCR products were purified using ExoSAP-IT reagent (USB Corporation, CMY-1, and CMY-8 to CMY-11; and 1 produced DHA-1), and 5
Cleveland, OH, USA) and used as the templates for sequencing on both
hyperproduced chromosomal AmpCs, while the remaining 25 did
strands with an ABI Prism 377 sequencer (Applied Biosystems, Foster
City, CA). not contain any expanded-spectrum ␤-lactamase (ESBL, AmpC,
Phenotypic methods to detect ESBLs, AmpCs, and carbapenemases. or carbapenemase).
ESBL production was initially tested with the CLSI confirmatory test using PCR and sequencing analyses showed that among the 162
both CTX (30 mg) and CAZ (30 mg) disks alone and in combination with ESBL-positive isolates, 87 (53.7%) harbored SHV-type ESBLs (61
CA (10 mg) (Becton, Dickinson, Sparks, MD). The test was considered SHV-5 and 26 SHV-12), 65 (41.1%) harbored CTX-M-type
positive when an increase in the growth-inhibitory zone around either the ESBLs (45 CTX-M-15, 19 CTX-M-3, and 1 CTX-M-32), 8 (4.9%)
CTX or the CAZ disk with CA was 5 mm or greater of the diameter around harbored both SHV and CTX-M ESBLs (4 SHV-5 plus CTX-M-3,
the disk containing CTX or CAZ alone (5). 3 SHV-12 plus CTX-M-15, and 1 SHV-5 plus CTX-M-15), and 2
For detecting and differentiating the production of MBL, KPC, or (1.2%) harbored GES-7 ESBL (Table 1). Moreover, 104 (64.2%)
both MBL and KPC carbapenemases, a phenotypic method was applied
of the ESBL producers and 76 (54.7%) of the non-ESBL producers
using disks of MER (10 ␮g) alone and with 400 ␮g of phenylboronic acid
or 292 ␮g of EDTA or both 400 ␮g of phenylboronic acid and 292 ␮g of harbored the broad-spectrum TEM-1 ␤-lactamase.
EDTA (19). Phenotypic detection of AmpC production was carried out by Antimicrobial susceptibilities. The susceptibility data for
using disks of cefotetan without and with BA (11) and Etest strips (bio- ESBL-producing and non-ESBL-producing isolates of the study
Mérieux), which contain cefotetan without or with cloxacillin. are summarized in Table 2. Aztreonam MIC50s, MIC90s, ranges of
Phenotypic detection of ESBLs using the modified CLSI ESBL con- MICs, and resistance rates were considerably higher among ESBL
firmatory test. The modification of the CLSI ESBL confirmatory test was producers than among non-ESBL producers. Nevertheless, re-
performed employing disks of CTX and CAZ with or without CA, on garding the remaining ␤-lactam antibiotics, the above parameters
which both BA and EDTA were dispensed. The stock of BA solution was did not differ considerably among ESBL and non-ESBL produc-
prepared by dissolving phenylboronic acid (benzeneboronic acid) at a ers, due to the presence of additional ␤-lactamases.
concentration of 40 mg/ml (11, 30). From this solution, 10 ␮l (containing
Comparative phenotypic testing for ESBLs. Table 3 summa-
400 ␮g of BA) was dispensed onto commercially available antibiotic disks
containing CTX (30 ␮g) or CAZ (30 ␮g) with or without CA (10 ␮g). rizes results of the phenotypic tests for ESBL production and their
Additionally, 10 ␮l of 0.1 M EDTA (containing 292 ␮g of EDTA) was performance characteristics for the 162 genotypically ESBL-posi-
dispensed onto the same antibiotic disks. The test was performed by in- tive and the 139 genotypically ESBL-negative clinical isolates.
oculating a Mueller-Hinton agar plate with a sample of the tested strain. (i) CLSI ESBL confirmatory test. By employing the standard
The agar plates were incubated at 37␱ C for 18 h. Similar to the standard CLSI ESBL confirmatory test, we found that 106 of the 162 ESBL

1484 jcm.asm.org Journal of Clinical Microbiology

 Evaluation of a Modified CLSI ESBL Confirmatory Test

TABLE 1 Distribution of expanded-spectrum ␤-lactamase genes among ESBL-producing (n ⫽ 162) and non-ESBL producing (n ⫽ 139) isolates
used for the evaluation of the modified CLSI ESBL confirmatory test
No. of isolates
Klebsiella Klebsiella Escherichia Enterobacter Enterobacter Proteus Providencia Serratia
pneumoniae oxytoca coli aerogenes cloacae mirabilis stuartii marcescens Total
Strain group and genotype(s) (n ⫽ 174) (n ⫽ 5) (n ⫽ 42) (n ⫽ 24) (n ⫽ 17) (n ⫽ 21) (n ⫽ 9) (n ⫽ 9) (n ⫽ 301)
ESBL-producing isolates (n ⫽ 162)a
blaCTX-M-3 6 5 1 4 16
blaCTX-M-15 5 1 5 11
blaCTX-M-15 ⫹ porin deficient 6 6
blaCTX-M-32 1 1
blaSHV-5 6 5 3 14
blaGES-7 2 2
blaSHV-5 ⫹ blaCTX-M-3 4 4
blaCMY-1-like ⫹ blaSHV-5 13 2 15
blaCMY-2-like ⫹ blaCTX-M-15 2 1 3
blaCMY-2-like ⫹ blaSHV-5⫹blaCTX-M-15 1 1

blaSHV-5 ⫹ AmpC hyperproducers 3 3 6

blaKPC-2 ⫹ blaCTX-M-15 4 4
blaKPC-2 ⫹ blaSHV-12 24 1 1 26
blaKPC-2 ⫹ blaSHV-12 ⫹ blaCTX-M-15 3 3
blaKPC-2 ⫹ blaSHV-5 2 2
blaKPC-2 ⫹ blaVIM-1 ⫹ blaSHV-5 9 9
blaKPC-2 ⫹ blaVIM-1 ⫹ blaCTX-M-15 1 1
blaVIM-1 ⫹ blaCTX-M-3 3 3
blaVIM-1 ⫹ blaCTX-M-15 4 4
blaVIM-1 ⫹ blaSHV-5 8 3 4 15
blaNDM-1 ⫹ blaCTX-M-15 8 8
blaOXA-48 ⫹ blaCTX-M-15 8 8
Subtotal 111 2 26 12 3 4 4 162

Non-ESBL- producing isolates (n ⫽ 139)b

Non-expanded-spectrum ␤-lactamase 4 2 4 2 3 4 3 3 25
producers
blaCMY-1-like 2 2
blaCMY-2-like 1 5 6 12
blaDHA-1 1 1
AmpC hyperproducers 2 3 5
blaKPC-2 18 2 6 6 32
blaKPC-2 ⫹ blaVIM-1 21 21
blaVIM-1 11 3 2 8 7 2 33
blaNDM-1 5 5
blaOXA-48 3 3
Subtotal 63 3 16 12 14 17 5 9 139
a
The blaTEM-1 gene was detected in 104 of the ESBL-producing isolates.
b
The blaTEM-1 gene was detected in 76 of the non-ESBL-producing isolates.

PCR-positive isolates had a ⱖ5-mm increase in the growth-inhib- either CTX-CA-BA-EDTA or CAZ-CA-BA-EDTA (sensitivity,
itory zone diameter around either CTX-CA or CAZ-CA and were 97.5%). In more detail, 106 (65.4%) of the 162 ESBL PCR-positive
considered phenotypically positive for ESBL production (sensitiv- isolates showed a ⱖ5-mm increase in the growth-inhibitory zone
ity, 65.4%). In more detail, 52 (32.1%) of the 162 isolates had a diameter around both CTX-CA-BA-EDTA and CAZ-CA-BA-
ⱖ5-mm increase in the growth-inhibitory zone diameter around EDTA, 35 (21.6%) isolates showed a ⱖ5-mm increase in the
both CTX-CA and CAZ-CA, 43 (26.5%) isolates had a ⱖ5-mm growth-inhibitory zone diameter only around CAZ-CA-BA-
increase in the growth-inhibitory zone diameter only around EDTA, and 17 (10.5%) isolates showed a ⱖ5-mm increase in the
CAZ-CA, and 11 (6.8%) isolates had a ⱖ5-mm increase in the growth-inhibitory zone diameter only around CTX-CA-BA-
growth-inhibitory zone diameter only around CTX-CA (Fig. 1). It EDTA (Fig. 1). In contrast to the CLSI ESBL confirmatory test, the
is of note that the test detected ESBLs only in 4 (13.3%) of the 30 modified test detected ESBLs among all KPC/ESBL or AmpC/
MBL/ESBL producers and in none of the 10 KPC/MBL/ESBL pro- ESBL producers, as well as among 29 (96.7%) of the 30 MBL/ESBL
ducers. The test was negative for all but 4 of the 139 non-ESBL- producers and 9 (90%) of the 10 KPC/VIM/ESBL producers.
producing isolates. In these isolates (2 MBL producers and 2 Using the modified test, we found that none of the 132 non-ESBL-
AmpC producers) the test showed a ⱖ5-mm increase in the zone producing isolates showed a ⱖ5-mm increase in the zone diame-
diameter around either the CTX-CA or CAZ-CA disks (specific- ter around either the CTX-CA-BA-EDTA or the CAZ-CA-BA-
ity, 97.1%) (Table 3). EDTA disks (specificity, 100%) (Table 3).
(ii) Modified CLSI ESBL confirmatory test with CA, BA, and (iii) Increases in the inhibition zone diameters using the two
EDTA. By employing the modified CLSI ESBL confirmatory test, phenotypic tests for ESBL detection. Among genotypically
we found that 158 of the 162 ESBL PCR-positive isolates showed a ESBL-positive isolates, the modified CLSI ESBL confirmatory test
ⱖ5-mm increase in the growth-inhibitory zone diameter around in comparison with the CLSI ESBL confirmatory test showed

May 2014 Volume 52 Number 5 jcm.asm.org 1485

Poulou et al.

TABLE 2 Antimicrobial susceptibilities to ␤-lactam antibiotics for the higher increases in the inhibition zone diameters of disks contain-
162 ESBL-producing isolates and 139 non-ESBL-producing isolates ing either CAZ or CTX. This was more obvious among ESBL-
MIC values (␮g/ml) positive isolates possessing carbapenemases (KPC, VIM, NDM,
Strain group and % of isolates
and KPC/VIM) or AmpCs (Table 4). Moreover, using the modi-
antimicrobial Range MIC50 MIC90 resistant
fied test, the average increases in the inhibition zone diameters
ESBL-producing isolates among ESBL producers harboring KPC, KPC/VIM, or AmpC
(n ⫽ 162)
were higher for the CAZ-CA-BA-EDTA disk than for the CTX-
Aztreonam 8 to ⬎256 256 ⬎256 80.2
Cefepime 1 to 128 32 128 71.6
CA-BA-EDTA disk (Table 4), since the majority of these isolates
Cefoxitin 1 to 256 32 256 64.8 carried SHV-type ESBLs.
Cefotaxime 1 to ⬎128 64 ⬎128 91.4
Ceftazidime 2 to ⬎256 128 ⬎256 86.4 DISCUSSION
Ertapenem 0.125 to 128 16 64 56.8 ESBLs have emerged gradually during the last decades in species of
Imipenem 0.250 to 128 2 32 48.8 Enterobacteriaceae and their prevalences reach alarming rates (1,
Meropenem 0.125 to 64 4 32 51.9 3, 32). Infections caused by such pathogens often limit therapeutic
Piperacillin-tazobactam 2 to ⬎256 128 ⬎256 67.3 options and cause treatment failures (3, 32, 33). Thus, in order to
successfully detect and treat infections due to ESBL-producing
Non-ESBL-producing
isolates (n ⫽ 139)
Enterobacteriaceae, the CLSI has recommended a phenotypic con-
Aztreonam 1 to ⬎256 4 256 40.3 firmatory test for ESBL production. This test can accurately detect
Cefepime 0.5 to 128 32 64 59.7 ESBLs among enterobacterial species when no other potent ␤-lac-
Cefoxitin 0.5 to 256 64 256 79.1 tamases are coproduced (2, 5, 6, 17).
Cefotaxime 0.250 to 128 64 128 76.9 However, this confirmatory method needs to be adjusted, as
Ceftazidime 0.5 to 256 128 256 78.4 multiple mechanisms of resistance to ␤-lactam antibiotics may be
Ertapenem 0.125 to 128 32 128 67.6 present in a single ESBL-producing isolate (1, 6, 34, 35). Hence,
Imipenem 0.250 to 64 16 64 64.7 the coexistence of ESBLs with derepressed chromosomal cepha-
Meropenem 0.125 to 64 16 32 63.3 losporinases, plasmid-mediated AmpCs, and carbapenemases
Piperacillin-tazobactam 2 to ⬎256 128 ⬎256 76.9
may complicate their phenotypic detection (6, 7, 16–18). The ex-

TABLE 3 Phenotypic detection of ESBLs among genotypically ESBL-positive (n ⫽ 162) and ESBL-negative (n ⫽ 139) clinical isolates using the
CLSI ESBL confirmatory test and the modified CLSI ESBL confirmatory test
Test performance (%)a
ESBL screening method No. (%) of isolates confirmed by PCR to have the indicated phenotype Sensitivity Specificity PPV NPV
KPC- and ESBL-producing isolates KPC- and non-ESBL-producing
(n ⫽ 35) isolates (n ⫽ 32)
CLSI ESBL confirmatory testb 23 (65.7) 0(0) 65.7 100 100 72.7
Modified CLSI ESBL confirmatory testc 35 (100) 0(0) 100 100 100 100

MBL- and ESBL-producing isolates MBL- and non-ESBL-producing

(n ⫽ 30) isolates (n ⫽ 38)
b
CLSI ESBL confirmatory test 4 (13.3) 2 (0) 13.3 94.7 66.7 60
Modified CLSI ESBL confirmatory testc 29 (96.7) 0 (0) 96.7 100 100 97.4

KPC-, MBL-, and ESBL-producing KPC-, MBL-, and non-ESBL-

isolates (n ⫽ 10) producing isolates (n ⫽ 21)
CLSI ESBL confirmatory testb 0 (0) 0 (0) 0 100 0 67.7
Modified CLSI ESBL confirmatory testc 9 (90) 0 (0) 90 100 100 95.5

AmpC- and ESBL-producing isolates AmpC- and non-ESBL-producing

(n ⫽ 25) isolates (n ⫽ 20)
b
CLSI ESBL confirmatory test 18 (72) 2 (0) 72 90 90 72
Modified CLSI ESBL confirmatory testc 25 (100) 0 (0) 100 100 100 100

OXA-48- and ESBL-producing isolates OXA-48- and non-ESBL-producing

(n ⫽ 8) isolates (n ⫽ 3)
b
CLSI ESBL confirmatory test 8 (100) 0 (0) 100 100 100 100
Modified CLSI ESBL confirmatory testc 8 (100) 0 (0) 100 100 100 100

Only ESBL-producing isolates (n ⫽ 54) Non-ESBL-producing isolates (n ⫽ 25)

CLSI ESBL confirmatory testb 53 (98.1) 0 (0) 98.1 100 100 96.2
Modified CLSI ESBL confirmatory testc 52 (96.3) 0 (0) 96.3 100 100 92.6

Total ESBL-producing isolates Total non-ESBL-producing isolates

(n ⫽ 162) (n ⫽ 139)
CLSI ESBL confirmatory testb 106 (65.4) 4(0) 65.4 97.1 96.4 70.7
Modified CLSI ESBL confirmatory testc 158 (97.5) 0(0) 97.5 100 100 97.2
a
PPV, positive predictive value; NPV, negative predictive value.
b
CTX-CA versus CTX and/or CAZ-CA versus CAZ.
c
CTX-CA-BA-EDTA versus CTX-BA-EDTA and/or CAZ-CA-BA-EDTA versus CAZ-BA-EDTA.

1486 jcm.asm.org Journal of Clinical Microbiology

 Evaluation of a Modified CLSI ESBL Confirmatory Test

pression of the latter ␤-lactamases can mask the presence of

ESBLs, so that, in terms of phenotypic screening, the prevalences
of ESBLs may be underestimated (8, 11, 17, 18). Thus, although
the detection of ESBLs among Enterobacteriaceae producing other
potent ␤-lactamases is overlooked more often than the detection
of ESBLs among Enterobacteriaceae producing only ESBLs, a
number of alternative phenotypic tests have been proposed to
improve detection of ESBLs among strains with derepressed chro-
mosomal AmpCs (6–9), plasmid-mediated AmpCs (8, 9, 16, 17),
or KPCs (18). It should also be mentioned that plasmid-mediated
AmpC and carbapenemase genes are characterized by high mobil-
ity and may be cotransferred with ESBLs among various species (1,
20, 32, 33, 34, 35). Moreover, ESBL genes are often cotransferred
with plasmid-mediated fluoroquinolone and aminoglycoside re-
sistance genes, thus contributing to the dissemination of multi-
drug resistance mechanisms (2, 4, 33). Therefore, an accurate
method for the phenotypic detection of ESBLs among Enterobac-
teriaceae irrespective of the presence of other ␤-lactamases is es-
sential in order to successfully address surveillance studies as well
as for infection control issues (16, 36). The accurate detection of
ESBLs might also guide therapeutic options for infections caused
by multidrug-resistant pathogens possessing AmpCs, OXA-48, or
MBLs; absence of ESBLs among such pathogens may allow the use
of cefepime, oxyimino-cephalosporins, and aztreonam, respec-
tively (25, 32, 36, 37).
In the present study, the standard CLSI ESBL confirmatory test
was found unable to detect the vast majority of ESBL producers
when MBLs, or KPCs and MBLs, were coproduced. Also, it was
negative in several isolates coproducing ESBLs with KPCs, plas-
mid-mediated AmpCs, or derepressed chromosomal cephalospo-
rinases. It should be noted, however, that the test sufficiently de-
tected ESBLs among OXA-48-possessing isolates, since OXA-48
derivatives only weakly hydrolyze cephalosporins (25, 37). Also, as
expected, among the set of isolates producing only ESBL, the con-
ventional test adequately detected ESBL production.
FIG 1 Representative results of the CLSI ESBL confirmatory test (second Furthermore, in order to improve the sensitivity of the CLSI
column) and the proposed modification (third column) using antibiotic disks
containing EDTA and BA for representative isolates producing KPC, VIM, and ESBL confirmatory test, we evaluated a modification of this test
ESBL (A), KPC and ESBL (B), VIM and ESBL (C), and VIM (D) ␤-lactamases. using antibiotic disks containing both BA and EDTA, well-known
The first column represents results from phenotypic testing for the detection inhibitors of KPCs/AmpCs and MBLs, respectively. The design of
and differentiation of carbapenemases in Enterobacteriaceae using disks of our modified method was based on the hypothesis that the inhib-
meropenem (MEM) without and with EDTA, BA, or EDTA plus BA (19).
itory activity of BA and EDTA will enlarge differences in zone

TABLE 4 Average increases in the inhibition zone diameters of CAZ-CA versus those of CAZ and of CTX-CA versus those of CTX by the CLSI
ESBL confirmatory test and average increases in the inhibition zone diameters of CAZ-CA-BA-EDTA versus those of CAZ-BA-EDTA and of CTX-
CA-BA-EDTA versus those of CTX-BA-EDTA by the modified CLSI ESBL confirmatory test
Average increases (mm) in inhibition zone diameter with:
CLSI ESBL
confirmatory test Modified CLSI ESBL confirmatory test
CAZ-CA/ CTX-CA/ CAZ-CA-BA-EDTA/ CTX-CA-BA-EDTA/
Strain group (genotypes) CAZ CTX CAZ-BA-EDTA CTX-BA-EDTA
KPC/ESBL producers (28 SHV type, 4 CTX-M type, 3 SHV ⫹ CTX-M type) 3.9 3.3 7.3 5.8
VIM/ESBL producers (15 SHV type, 7 CTX-M type) 3.8 1.0 8.6 8.3
NDM/ESBL producer (8 CTX-M type) 3.1 2.2 8.3 10.2
KPC/VIM/ESBL producers (9 SHV type, 1 CTX-M type) 2.6 1.3 7.4 4.0
AmpC/ESBL producers (21 SHV type, 3 CTX-M type, 1 SHV ⫹ CTX-M type) 8.9 3.7 9.3 5.8
OXA-48/ESBL producers (8 CTX-M type) 5.8 6.9 6.5 6.4
Only ESBL producers (14 SHV type, 34 CTX-M type, 2 IBC type, 4 SHV ⫹ 8.9 12.0 8.5 10.9
CTX-M type)

May 2014 Volume 52 Number 5 jcm.asm.org 1487

Poulou et al.

diameters between CTX and CTX-CA disks as well as between 8. Garrec H, Drieux-Rouzet L, Golmard JL, Jarlier V, Robert J. 2011.
CAZ and CAZ-CA disks in the phenotypic detection of ESBLs Comparison of nine phenotypic methods for detection of extended-
spectrum ␤-lactamase production by Enterobacteriaceae. J. Clin. Micro-
among Enterobacteriaceae expressing various carbapenemases or biol. 49:1048 –1057. http://dx.doi.org/10.1128/JCM.02130-10.
AmpC ␤-lactamases. 9. Stuart JC, Diederen B, Al Naiemi N, Fluit A, Arents N, Thijsen S,
The modified method was found accurate for detecting ESBLs Vlaminckx B, Mouton JW, Leverstein-van Hall M. 2011. Method for
not only among Enterobacteriaceae producing various potent phenotypic detection of extended-spectrum ␤-lactamases in enterobacter
␤-lactamase genes but also among those producing only ESBLs. It species in the routine clinical setting. J. Clin. Microbiol. 49:2711–2713.
http://dx.doi.org/10.1128/JCM.00864-11.
identified almost all genotypically ESBL-positive isolates and did 10. Nordmann P, Dortet L, Poirel L. 2012. Rapid detection of extended-
not give false-positive results for any of the ESBL-negative isolates. spectrum-␤-lactamase-producing Enterobacteriaceae. J. Clin. Microbiol.
In contrast to the CLSI ESBL confirmatory test, the modified test 50:3016 –3022. http://dx.doi.org/10.1128/JCM.00859-12.
detected ESBLs among all KPC, NDM, and AmpC producers, as 11. Coudron PE. 2005. Inhibitor-based methods for detection of plasmid-
mediated AmpC ␤-lactamases in Klebsiella spp., Escherichia coli, and Pro-
well as the vast majority of VIM and KPC/VIM producers. More-
teus mirabilis. J. Clin. Microbiol. 43:4163– 4167. http://dx.doi.org/10.1128
over, the modified test detected almost all of the study isolates /JCM.43.8.4163-4167.2005.
producing ESBL only, including those that exhibited ertapenem 12. Doi Y, Potoski BA, Adams-Haduch JM, Sidjabat HE, Pasculle AW,
resistance due to porin deficiency. It is also of note that although Paterson DL. 2008. Simple disk-based method for detection of Klebsiella
the CLSI ESBL confirmatory test showed false-positive results pneumoniae carbapenemase-type ␤-lactamase by use of a boronic acid
compound. J. Clin. Microbiol. 46:4083-4086. http://dx.doi.org/10.1128
among four non-ESBL-producing isolates, the modified test was /JCM.01408-08.
negative among all non-ESBL producers of the study. Accord- 13. Pitout JD, Le PG, Moore KL, Church DL, Gregson DB. 2010. Detection
ingly, previous surveys have shown that the CLSI ESBL confirma- of AmpC ␤-lactamases in Escherichia coli, Klebsiella spp., Salmonella spp.
tory test may give a few false-positive results among non-ESBL-, and Proteus mirabilis in a regional clinical microbiology laboratory. Clin.
AmpC-producing Enterobacteriaceae, while a modification of the Microbiol. Infect. 16:165–170. http://dx.doi.org/10.1111/j.1469-0691
.2009.02756.x.
test using BA did not give any false-positive results in this bacterial 14. Tsakris A, Kristo I, Poulou A, Markou F, Ikonomidis A, Pournaras S.
population (16, 38). In addition, among isolates coproducing 2008. First occurrence of KPC-2-possessing Klebsiella pneumoniae in a
AmpCs or carbapenems, the proposed modified test provided Greek hospital and recommendation for detection with boronic acid disc
considerably higher increases in the inhibitory zone diameters tests. J. Antimicrob. Chemother. 62:1257–1260. http://dx.doi.org/10.1093
/jac/dkn364.
around disks containing CAZ or CTX, allowing an easy and 15. Giske CG, Gezelius L, Samuelsen Ø Warner M, Sundsfjord A, Wood-
straightforward interpretation of the phenotypic test. ford N. 2011. A sensitive and specific phenotypic assay for detection of
In conclusion, the modification of the CLSI ESBL confirma- metallo-␤-lactamases and KPC in Klebsiella pneumoniae with the use of
tory test was found to accurately detect ESBL-producing Entero- meropenem disks supplemented with aminophenylboronic acid, dipico-
bacteriaceae regardless of the underlying ␤-lactam resistance linic acid and cloxacillin. Clin. Microbiol. Infect. 17:552–556. http://dx
.doi.org/10.1111/j.1469-0691.2010.03294.x.
mechanism. It is understandable that molecular assays may pro- 16. Jeong SH, Song W, Park MJ, Kim JS, Kim HS, Bae IK, Lee KM. 2008.
vide accurate results in the identification of ESBL genes (3), but Boronic acid disk tests for identification of extended-spectrum ␤-lacta-
their accessibility is often limited; nevertheless, they are expensive. mase production in clinical isolates of Enterobacteriaceae producing chro-
The proposed phenotypic method is a highly sensitive and specific mosomal AmpC ␤-lactamases. Int. J. Antimicrob. Agents 31:467– 471.
http://dx.doi.org/10.1016/j.ijantimicag.2007.12.014.
method that can be easily performed and interpreted in ⬍24 h 17. Song W, Bae IK, Lee YN, Lee CH, Lee SH, Jeong SH. 2007. Detection of
without requiring the previous knowledge of the presence of other extended-spectrum ␤-lactamases by using boronic acid as an AmpC
␤-lactamases. ␤-lactamase inhibitor in clinical isolates of Klebsiella spp and Escherichia
coli. J. Clin. Microbiol. 45:1180 –1184. http://dx.doi.org/10.1128/JCM
.02322-06.
REFERENCES 18. Tsakris A, Poulou A, Themeli-Digalaki K, Voulgari E, Pittaras T,
1. Bush K, Fisher JF. 2011. Epidemiological expansion, structural studies, Sofianou D, Pournaras S, Petropoulou D. 2009. Use of boronic acid disk
and clinical challenges of new ␤-lactamases from Gram-negative bacteria. tests to detect extended-spectrum ␤-lactamases in clinical isolates of KPC
Annu. Rev. Microbiol. 65:455– 478. http://dx.doi.org/10.1146/annurev carbapenemase-possessing Enterobacteriaceae. J. Clin. Microbiol. 47:
-micro-090110-102911. 3420 –3426. http://dx.doi.org/10.1128/JCM.01314-09.
2. Pitout JD, Laupland KB. 2008. Extended-spectrum ␤-lactamase-producing 19. Tsakris A, Poulou A, Pournaras S, Voulgari E, Vrioni G, Themeli-
Enterobacteriaceae: an emerging public-health concern. Lancet Infect. Dis. Digalaki K, Petropoulou D, Sofianou D. 2010. A simple phenotypic
8:159 –166. http://dx.doi.org/10.1016/S1473-3099(08)70041-0. method for the differentiation of metallo-␤-lactamases and class A KPC
3. Zahar JR, Lortholary O, Martin C, Potel G, Plesiat P, Nordmann P. carbapenemases in Enterobacteriaceae clinical isolates. J. Antimicrob. Che-
2009. Addressing the challenge of extended-spectrum ␤-lactamases. Curr. mother. 65:1664 –1671. http://dx.doi.org/10.1093/jac/dkq210.
Opin. Investig. Drugs 10:172–180. 20. Nordmann P, Gniadkowski M, Giske CG, Poirel L, Woodford N,
4. Carattoli A. 2009. Resistance plasmid families in Enterobacteriaceae. An- Miriagou V; European Network on Carbapenemases. 2012. Identifica-
timicrob. Agents Chemother. 53:2227–2238. http://dx.doi.org/10.1128 tion and screening of carbapenemase-producing Enterobacteriaceae. Clin.
/AAC.01707-08. Microbiol. Infect. 18:432– 438. http://dx.doi.org/10.1111/j.1469-0691
5. Clinical and Laboratory Standards Institute. 2012. Performance stan- .2012.03815.x.
dards for antimicrobial susceptibility testing. Twenty second informa- 21. Franklin C, Liolios L, Peleg AY. 2006. Phenotypic detection of carbap-
tional supplement update. CLSI document M100-S22 U. Clinical and enem-susceptible metallo-␤-lactamase-producing gram-negative bacilli
Laboratory Standards Institute, Wayne, PA. in the clinical laboratory. J. Clin. Microbiol. 44:3139 –3144. http://dx.doi
6. Drieux L, Brossier F, Sougakoff W, Jarlier V. 2008. Phenotypic detection .org/10.1128/JCM.00879-06.
of extended spectrum ␤-lactamase production in Enterobacteriaceae: re- 22. Pournaras S, Poulou A, Tsakris A. 2010. Inhibitor-based methods for the
view and bench guide. Clin. Microbiol. Infect. 14(Suppl 1):90 –103. http: detection of KPC carbapenemase-producing Enterobacteriaceae in clinical
//dx.doi.org/10.1111/j.1469-0691.2007.01846.x. practice by using boronic acid compounds. J. Antimicrob. Chemother.
7. Tzelepi E, Giakkoupi P, Sofianou D, Loukova V, Kemeroglou A, Tsakris 65:1319 –1321. http://dx.doi.org/10.1093/jac/dkq124.
A. 2000. Detection of extended-spectrum ␤-lactamases in clinical isolates 23. Tzelepi E, Magana C, Platsouka E, Sofianou D, Paniara O, Legakis NJ,
of Enterobacter cloacae and Enterobacter aerogenes. J. Clin. Microbiol. 38: Vatopoulos AC, Tzouvelekis LS. 2003. Extended-spectrum ␤-lactamase
542–546. types in Klebsiella pneumoniae and Escherichia coli in two Greek hospitals.

1488 jcm.asm.org Journal of Clinical Microbiology

 Evaluation of a Modified CLSI ESBL Confirmatory Test

Intern. J. Antimicrob. Agents 21:285–288. http://dx.doi.org/10.1016 31. Poulou A, Voulgari E, Vrioni G, Koumaki V, Xidopoulos G, Chatzi-
/S0924-8579(02)00361-8. pantazi V, Markou F, Tsakris A. 2013. Outbreak caused by an ertap-
24. Smith Moland E, Hanson ND, Herrera VL, Black JA, Lockhart TJ, enem-resistant, CTX-M-15-producing Klebsiella pneumoniae sequence
Hossain A, Johnson JA, Goering RV, Thomson KS. 2003. Plasmid- type 101 clone carrying an OmpK36 porin variant. J. Clin. Microbiol.
mediated, carbapenem-hydrolysing ␤-lactamase, KPC-2, in Klebsiella 51:3176 –3182. http://dx.doi.org/10.1128/JCM.01244-13.
pneumoniae isolates. J. Antimicrob. Chemother. 51:711–714. http://dx 32. .Rossolini GM, Mantengoli E, Docquier JD, Musmanno RA, Coratza G.
.doi.org/10.1093/jac/dkg124. 2007. Epidemiology of infections caused by multiresistant gram-
25. Poirel L, Héritier C, Tolün V, Nordmann P. 2004. Emergence of oxa- negatives: ESBLs, MBLs, panresistant strains. New Microbiol. 30:332–339.
cillinase-mediated resistance to imipenem in Klebsiella pneumoniae. An- 33. Livermore DM. 2009. Has the era of untreatable infections arrived? J. Anti-
timicrob. Agents Chemother. 48:15–22. http://dx.doi.org/10.1128/AAC microb. Chemother. 64:29 –36. http://dx.doi.org/10.1093/jac/dkp255.
.48.1.15-22.2004. 34. Endimiani A, Hujer AM, Perez F, Bethel CR, Hujer KM, Kroeger J,
26. Agodi A, Voulgari E, Barchitta M, Politi L, Koumaki V, Spanakis N, Oethinger M, Paterson DL, Adams MD, Jacobs MR, Diekema DJ, Hall
Giaquinta L, Valenti G, Romeo MA, Tsakris A. 2011. Containment of an GS, Jenkins SG, Rice LB, Tenover FC, Bonomo RA. 2009. Character-
outbreak of KPC-3-producing Klebsiella pneumoniae in Italy. J. Clin. Mi- ization of blaKPC-containing Klebsiella pneumoniae isolates detected in
crobiol. 49:3986 –3989. http://dx.doi.org/10.1128/JCM.01242-11.
different institutions in the Eastern USA. J. Antimicrob. Chemother. 63:
27. Pérez-Pérez FJ, Hanson ND. 2002. Detection of plasmid-mediated AmpC ␤-lac-
427– 437. http://dx.doi.org/10.1093/jac/dkn547.
tamase genes in clinical isolates by using multiplex PCR. J. Clin. Microbiol. 40:
35. Pournaras S, Poulou A, Voulgari E, Vrioni G, Kristo I, Tsakris A. 2010.
2153–2162. http://dx.doi.org/10.1128/JCM.40.6.2153-2162.2002.
28. Preston KE, Radomski CC, Venezia RA. 2000. Nucleotide sequence of Detection of the new metallo-␤-lactamase VIM-19 along with KPC-2,
the chromosomal ampC gene of Enterobacter aerogenes. Antimicrob. CMY-2 and CTX-M-15 in Klebsiella pneumoniae. J. Antimicrob. Che-
Agents Chemother. 44:3158 –3162. http://dx.doi.org/10.1128/AAC.44.11 mother. 65:1604 –1607. http://dx.doi.org/10.1093/jac/dkq190.
.3158-3162.2000. 36. Livermore DM, Andrews JM, Hawkey PM, Ho PL, Keness Y, Doi Y,
29. Barnaud G, Labia R, Raskine L, Sanson-Le Pors MJ, Philippon A, Arlet Paterson D, Woodford N. 2012. Are susceptibility tests enough, or should
G. 2001. Extension of resistance to cefepime and cefpirome associated to a laboratories still seek ESBLs and carbapenemases directly? J. Antimicrob.
six amino acid deletion in the H-10 helix of the cephalosporinase of an Chemother. 67:1569-1577. http://dx.doi.org/10.1093/jac/dks088.
Enterobacter cloacae clinical isolate. FEMS Microbiol. Lett. 195:185–190. 37. Poirel L, Potron A, Nordmann P. 2012. OXA-48-like carbapenemases:
http://dx.doi.org/10.1111/j.1574-6968.2001.tb10519.x. the phantom menace. J. Antimicrob. Chemother. 67:1597–1606. http://dx
30. Tsakris A, Kristo I, Poulou A, Themeli-Digalaki K, Ikonomidis A, .doi.org/10.1093/jac/dks121.
Petropoulou D, Pournaras S, Sofianou D. 2009. Evaluation of boronic 38. Robberts FJ, Kohner PC, Patel R. 2009. Unreliable extended-spectrum
acid disk tests for differentiating KPC-possessing Klebsiella pneumoniae ␤-lactamase detection in the presence of plasmid-mediated AmpC in
isolates in the clinical laboratory. J. Clin. Microbiol. 47:362–367. http://dx Escherichia coli clinical isolates. J. Clin. Microbiol. 47:358 –361. http://dx
.doi.org/10.1128/JCM.01922-08. .doi.org/10.1128/JCM.01687-08.

May 2014 Volume 52 Number 5 jcm.asm.org 1489

View publication stats