JOURNAL OF CLINICAL MICROBIOLOGY, Mar. 2011, p. 1143–1147
Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Vol. 49, No. 3
Detection of Klebsiella pneumoniae Carbapenemase (KPC) Production
in Non-Klebsiella pneumoniae Enterobacteriaceae Isolates by Use
of the Phoenix, Vitek 2, and Disk Diffusion Methods?
Christopher D. Doern,1* W. Michael Dunne, Jr.,2and Carey-Ann D. Burnham2
The Department of Pathology, University of Texas Southwestern Medical Center, Dallas, Texas 75235,1and Department of
Pathology and Departments of Immunology and Pediatrics, Washington University School of
Medicine, Saint Louis, Missouri 631102
Received 25 October 2010/Returned for modification 23 November 2010/Accepted 28 December 2010
In this study, we tested the abilities of the Vitek 2, BD Phoenix, and Kirby Bauer disk diffusion tests to detect
carbapenemase production in a collection of 14 Klebsiella pneumoniae carbapenemase (KPC)-producing non-
Klebsiella pneumoniae isolates. In addition, we evaluated 13 KPC-positive K. pneumoniae isolates by using each
of these methods and applied both 2009 and 2010 CLSI carbapenem interpretive guidelines.
Klebsiella pneumoniae carbapenemases (KPC) are Ambler
class A plasmid-encoded enzymes that are capable of hydro-
lyzing all beta-lactam antibiotics, including monobactams, ex-
tended-spectrum cephalosporins, and carbapenems. The KPC
enzyme was originally described in 2001 in a Klebsiella pneu-
moniae isolate from North Carolina (18). Since then, KPC has
disseminated worldwide and, although harbored predomi-
nantly by Klebsiella pneumoniae, has also been identified in
numerous other genera of the Enterobacteriaceae (1, 8, 15, 17).
Many clinical microbiology laboratories utilize automated in
vitro susceptibility testing systems, but some of these have
difficulty detecting carbapenemase production (2, 17). Due to
these limitations, current guidelines recommend that KPC pro-
duction be confirmed by an alternative method (5). However,
confirmatory testing options are limited to the modified Hodge
test (MHT), which is subjective and suffers from false-positive
results, and PCR-based methods, which may not be available in
all labs (3). The issue of confirmation has led others to evaluate
boronic acid inhibition for KPC confirmation. However, these
methods also inhibit AmpC activity and are, therefore, not
specific to KPC (11, 13, 14).
While the performance of automated systems for the detec-
tion of KPC-producing K. pneumoniae has been evaluated,
little is known about the ability of these systems to detect KPC
production in other genera of the Enterobacteriaceae. The pur-
pose of this study was to evaluate KPC detection in this pop-
ulation of organisms. A collection of KPC-positive and -nega-
tive organisms isolated from patient specimens at Barnes
Jewish Hospital and St. Louis Children’s Hospital in St. Louis,
MO, were evaluated in this study. Fourteen KPC-producing
non-K. pneumoniae (non-KPNE?) isolates of members of the
Enterobacteriaceae and13 KPC-positive
(KPNE?) isolates were analyzed using the Phoenix and Vitek
2 instruments. KPC-positive strains included Citrobacter freun-
dii (n ? 2), Proteus mirabilis (n ? 2), Klebsiella pneumoniae
(n ? 13), Escherichia coli (n ? 3), Enterobacter aerogenes (n ?
2), Enterobacter cloacae (n ? 4), and Klebsiella oxytoca (n ? 1).
The Clinical and Laboratory Standards Institute (CLSI) re-
cently published a revision to the carbapenem breakpoints
(Table 1) (5, 6), resulting in a reduction of the MIC by two
doubling dilutions per interpretive category for doripenem,
meropenem, and imipenem and by three doubling dilutions for
ertapenem. For ertapenem, disk diffusion zone sizes were in-
creased by 4 mm per interpretative category. Imipenem and
meropenem zone sizes were increased by 7 mm for the sus-
ceptible category and 6 mm for the resistant category. Inter-
pretive criteria for doripenem were also included in the up-
dated guidelines. Our results were analyzed using both 2009
and 2010 CLSI carbapenem breakpoints to understand the
impact of these changes on KPC detection and interpretation
of carbapenem resistance among genera of the Enterobacteri-
KPC-producing organisms were identified and confirmed via
a two-step process. Gram-negative organisms for which sus-
ceptibilities were deemed necessary were tested using a 12-
drug disk diffusion panel, which included meropenem (4). Or-
ganisms having a reduced meropenem zone size (defined as
?25 mm) were analyzed for blaKPCby PCR. The TaqMan
real-time KPC PCR assay uses primers and probes which de-
tect all currently described KPC variants. The sequences were
as follows: for the KPC forward primer, 5?-GCG GAA CCA
TTC GCT AAA CTC GAA-3?; for the KPC reverse primer,
5?-AGA AAG CCC TTG AAT GAG CTG CAC-3?; and for
the KPC probe, 5?-/6-FAM/ATA CCG GCT CAG GCG CAA
CTG TAA GTT A/6-TAMSp/-3? (where 6-FAM represents
6-carboxyfluorescein and 6-TAM represents 6-carboxytetra-
methylrhodamine). DNA was purified using a bead-beating
extraction process (BD GeneOhm lysis kit; BD, Franklin
Lakes, NJ) and analyzed on a Cepheid SmartCycler (Sunny-
vale, CA). PCR-positive isolates with reduced meropenem
zone sizes were considered to be KPC positive.
Susceptibility testing for the test strains was performed using
three different methods; the BD Phoenix (Becton Dickinson,
Franklin Lakes, NJ) was used in conjunction with the NMIC/
ID-134 panel, the Vitek 2 (bioMerie ´ux, Marcy l’Etoile, France)
* Corresponding author. Mailing address: 1935 Medical District,
Drive Mail code B1.06, Dallas, TX 75235. Phone: (214) 456-1344. Fax:
(214) 456-4713. E-mail: firstname.lastname@example.org.
?Published ahead of print on 5 January 2011.
was used in conjunction with the EXN9 and GN24 cards, and
disk diffusion susceptibilities for ertapenem, meropenem, and
imipenem were determined using 2009 and 2010 CLSI guide-
The Vitek 2 system uses two cards in tandem to test four
carbapenems (ertapenem, meropenem, doripenem, and imi-
penem). The GN24 card can be used alone and includes
ertapenem and imipenem. The “extension” EXN9 card cannot
be used alone and must be coupled with the GN24 card. By
using the 2010 CLSI guidelines for carbapenem interpretation,
the Vitek 2 predicted 92.9 and 100% of KPC-positive non-K.
pneumoniae (non-KPNE?) and K. pneumoniae (KPNE?) iso-
lates, respectively, as nonsusceptible (i.e., intermediate or re-
sistant) to doripenem (Fig. 1). It performed less well for mero-
penem and imipenem, where only 64.3 and 84.6% of
non-KPNE?and KPNE?isolates, respectively, were nonsus-
TABLE 1. 2009 and 2010 CLSI carbapenem interpretive guidelinesa
MIC (mg/liter)Zone size (mm)
2009 2010 2009 20102009 2010 20092010 20092010 2009 2010
a2009 CLSI breakpoints and 2010 proposed CLSI breakpoints are given. S, susceptible; I, intermediate; R, resistant.
b2009 FDA breakpoint for doripenem.
FIG. 1. Percent susceptible and nonsusceptible KPC-producing Klebsiella pneumoniae and non-Klebsiella pneumoniae isolates, as tested on
Vitek 2. (A) EXN9 card. (B) GN24 card. **, the GN24 card does not possess low enough dilutions to categorize isolates as susceptible to
ertapenem by the 2010 guidelines.
1144 NOTESJ. CLIN. MICROBIOL.
ceptible to meropenem. For imipenem, 92.9% of non-KPNE?
isolates and 84.6% of KPNE?isolates were reported as non-
susceptible (Fig. 1). All KPC-producing isolates were classified
as nonsusceptible to ertapenem. However, with the implemen-
tation of the new CLSI breakpoints, the Vitek cards lack dilu-
tions less than 0.5 mg/liter for ertapenem and as a result cannot
classify organisms as susceptible (Fig. 1).
When 2009 CLSI guidelines were applied, the Vitek 2 clas-
sified only 21.4 and 30.8% of non-KPNE?and KPNE?iso-
lates, respectively, as nonsusceptible to meropenem (Fig. 1).
For imipenem, only 64.3 and 15% of non-KPNE?and KPNE?
isolates were classified as nonsusceptible, respectively. As ex-
pected, ertapenem demonstrated superior sensitivity for KPC
detection, with 78.6 and 100% of non-KPNE?and KPNE?
isolates testing as nonsusceptible, respectively.
The problem with ertapenem is not unique to the Vitek 2, as
the BD Phoenix NMIC/ID-134 panel also lacks the dilutions
necessary to classify isolates as susceptible with the revised
(2010) breakpoints. Without that function, the BD Phoenix
NMIC/ID-134 is left to rely on meropenem as the sole carba-
penem representative in the “susceptible” category. For mero-
penem, 71.4 and 100% of non-KPNE?and KPNE?isolates,
respectively, were classified as nonsusceptible (Fig. 2).
When 2009 breakpoints were used, ertapenem interpreta-
tions classified 85.7 and 100% of non-KPNE?and KPNE?
isolates as nonsusceptible, respectively. For meropenem,
57.1% of non-KPNE?isolates were classified as nonsuscep-
tible, while 100% of KPNE?isolates were reported as nonsus-
ceptible (Fig. 2).
We also evaluated disk diffusion using ertapenem, mero-
penem, and imipenem and CLSI 2010 zone size criteria. For
ertapenem and meropenem, all KPC-positive isolates were
classified as nonsusceptible. Imipenem classified 100% of
KPNE?isolates and 90.9% of non-KPNE?isolates as nonsus-
No differences in categorization were observed for erta-
penem by using 2009 CLSI interpretations, as 100% of isolates
were categorized as intermediate or resistant. A single KPC-
positive isolate each of E. coli and K. pneumoniae was misclas-
sified as meropenem susceptible by using the old breakpoints.
The CLSI breakpoint revisions had the greatest impact on
classification of KPC-positive isolates when imipenem was
used. With the 2009 criteria, 36.4% of non-KPNE?isolates
and 25% of KPNE?isolates were categorized as susceptible to
imipenem (Fig. 3).
Although revision to MIC and zone size interpretive guide-
lines should improve the ability of these methodologies to
detect KPC-producing Enterobacteriaceae, the improved sensi-
tivity may come with decreased specificity. To better under-
stand how these changes will impact the classification of
non-KPC-producing Enterobacteriaceae, we evaluated 22 KPC-
negative isolates (6 K. pneumoniae isolates and 16 isolates of
other genera of the Enterobacteriaceae). Because both the
Vitek 2 and the BD Phoenix are currently unable to classify
organisms as susceptible to ertapenem, ertapenem was ex-
cluded from these analyses. For this set of organisms, the Vitek
2 reported 3 (14%) isolates as “intermediate” to at least one
carbapenem. Two of these were Enterobacter cloacae isolates
and were also resistant to extended-spectrum cephalosporins.
We believe that these strains combined a derepressed ampC
along with a porin mutation to generate increased resistance.
The BD Phoenix categorized one Enterobacter aerogenes iso-
late as resistant to meropenem and one E. cloacae isolate as
intermediate (9.1%). Based on disk diffusion, 8 isolates (36%)
were interpreted as resistant to ertapenem and 1 (4.5%) was
interpreted as intermediate. Fifteen isolates (5 K. pneumoniae
isolates and 10 isolates of other genera) were available for
meropenem and imipenem disk diffusion testing. For mero-
penem, 2 (13%) of the E. cloacae isolates were intermediate,
and 1 (6.5%) E. aerogenes isolate was resistant. The same E.
aerogenes isolate that was interpreted as resistant to mero-
penem was also resistant to imipenem (6.5%) (data not
2009 CLSI disk diffusion zone sizes classified 100% of K.
pneumoniae isolates as susceptible to ertapenem, meropenem,
and imipenem. For the other genera, nearly half were catego-
FIG. 2. Percent susceptible and nonsusceptible KPC-producing Klebsiella pneumoniae and non-Klebsiella pneumoniae isolates, as tested on the
BD Phoenix NMIC/ID-134 panel. **, the NMIC/ID-134 panel does not possess low enough dilutions to categorize isolates as susceptible to
ertapenem by the 2010 guidelines.
VOL. 49, 2011NOTES1145
rized as intermediate or resistant to ertapenem, while for imi-
penem and meropenem, 88.9% of isolates were susceptible.
Only 11.1% were resistant to imipenem and meropenem,
which represented a single isolate of Enterobacter aerogenes
(data not shown).
While the KPC enzyme is still found most commonly in
association with K. pneumoniae, the blaKPCgene has now been
identified in numerous other members of the Enterobacteria-
ceae (7, 9, 12). Given the infection control implications and the
high morbidity and mortality associated with infections caused
by these organisms, it is essential that susceptibility methods
identify carbapenem resistance rapidly and reliably (10, 16,
19). Our data show that the new CLSI breakpoints have in-
creased the likelihood that these systems will classify a KPC-
producing organism as nonsusceptible to carbapenem antibi-
For disk diffusion, our data show that the 2010 breakpoints
will have the greatest impact on imipenem categorical inter-
pretation, as 35% of all isolates were classified as susceptible
by 2009 criteria whereas all but one were reported as resistant
by 2010 criteria (Fig. 3).
Although each system misclassified several KPC-positive
strains as susceptible to a carbapenem, most KPC-positive
strains were reported as having an unusual susceptibility pat-
tern. In addition, most KPC-positive isolates tested resistant to
extended-spectrum cephalosporins and cefepime. One concern
is the performance of the Phoenix for KPC-positive Proteus
mirabilis isolates. Both strains (confirmed as KPC positive by
PCR) were reported to be susceptible to meropenem, ceftazi-
dime, cefotaxime, and cefoxitin, with no result generated for
cefepime. None of our other KPC-positive study isolates dem-
onstrated across-the-board susceptibility to the carbapenems
and extended-spectrum cephalosporins and cefepime with ei-
ther the Phoenix or the Vitek 2.
One limitation of the 2010 breakpoints for automated sys-
tems such as Phoenix and Vitek 2 is that the lowest ertapenem
dilution currently available on either panel is 0.5 mg/liter. With
the new breakpoints, ertapenem susceptibility is defined as
?0.25 mg/liter, and both systems fail to include dilutions within
the actual susceptible range.
In conclusion, accurate and rapid detection of Enterobac-
teriaceae harboring the KPC enzyme is of clinical importance
FIG. 3. Percent susceptible and nonsusceptible KPC-producing Klebsiella pneumoniae and non-Klebsiella pneumoniae isolates, as tested by disk
diffusion. (A) 2009 CLSI zone sizes. (B) 2010 CLSI zone sizes.
1146 NOTESJ. CLIN. MICROBIOL.
in ensuring that the correct antimicrobial therapy is given to
patients infected with these organisms and that appropriate
infection control measures are initiated. Historically, auto-
mated systems have struggled to accurately identify KPC-pro-
ducing organisms. Our analysis indicates that while the non-
KPNE?Enterobacteriaceae isolates are more likely to be
falsely classified as susceptible, this is largely rectified by the
2010 CLSI update to the carbapenem breakpoints.
1. Bratu, S., D. Landman, M. Alam, E. Tolentino, and J. Quale. 2005. Detec-
tion of KPC carbapenem-hydrolyzing enzymes in Enterobacter spp. from
Brooklyn, New York. Antimicrob. Agents Chemother. 49:776–778.
2. Bulik, C. C., et al. 2010. Comparison of meropenem MICs and susceptibil-
ities for carbapenemase-producing Klebsiella pneumoniae isolates by vari-
ous testing methods. J. Clin. Microbiol. 48:2402–2406.
3. Carvalhaes, C. G., R. C. Picao, A. G. Nicoletti, D. E. Xavier, and A. C. Gales.
2010. Cloverleaf test (modified Hodge test) for detecting carbapenemase
production in Klebsiella pneumoniae: be aware of false positive results. J.
Antimicrob. Chemother. 65:249–251.
4. Clinical and Laboratory Standards Institute. 2009. Performance standards
for antimicrobial susceptibility testing. Nineteenth informational supple-
ment. Document M100-S19. CLSI, Wayne, PA.
5. Clinical and Laboratory Standards Institute. 2010. Performance standards
for antimicrobial susceptibility testing. Twentieth informational supplement.
Document M100-S20. CLSI, Wayne, PA.
6. Clinical and Laboratory Standards Institute. 2010. Performance standards
for antimicrobial susceptibility testing. Twentieth informational supplement
update. Document M100-S20 U. CLSI, Wayne, PA.
7. D’Alincourt Carvalho-Assef, A. P., et al. 2010. Escherichia coli producing
KPC-2 carbapenemase: first report in Brazil. Diagn. Microbiol. Infect. Dis.
8. Fisher, M. A., et al. 2009. Performance of the Phoenix bacterial identification
system compared with disc diffusion methods for identifying extended-spec-
trum beta-lactamase, AmpC and KPC producers. J. Med. Microbiol. 58:774–
9. Livermore, D. M., R. Hope, G. Brick, M. Lillie, and R. Reynolds. 2008.
Non-susceptibility trends among Enterobacteriaceae from bacteraemias in
the UK and Ireland, 2001-06. J. Antimicrob. Chemother. 62(Suppl. 2):ii41–
10. Marchaim, D., S. Navon-Venezia, M. J. Schwaber, and Y. Carmeli. 2008.
Isolation of imipenem-resistant Enterobacter species: emergence of KPC-2
carbapenemase, molecular characterization, epidemiology, and outcomes.
Antimicrob. Agents Chemother. 52:1413–1418.
11. Pournaras, S., et al. 2010. Detection of the new metallo-beta-lactamase
VIM-19 along with KPC-2, CMY-2 and CTX-M-15 in Klebsiella pneu-
moniae. J. Antimicrob. Chemother. 65:1604–1607.
12. Tibbetts, R., J. G. Frye, J. Marschall, D. Warren, and W. Dunne. 2008.
Detection of KPC-2 in a clinical isolate of Proteus mirabilis and first re-
ported description of carbapenemase resistance caused by a KPC beta-
lactamase in P. mirabilis. J. Clin. Microbiol. 46:3080–3083.
13. Tsakris, A., et al. 2009. Evaluation of boronic acid disk tests for differenti-
ating KPC-possessing Klebsiella pneumoniae isolates in the clinical labora-
tory. J. Clin. Microbiol. 47:362–367.
14. Tsakris, A., et al. 2009. Use of boronic acid disk tests to detect extended-
spectrum beta-lactamases in clinical isolates of KPC carbapenemase-pos-
sessing enterobacteriaceae. J. Clin. Microbiol. 47:3420–3426.
15. Vading, M., O. Samuelsen, B. Haldorsen, A. S. Sundsfjord, and C. G. Giske.
23 July 2010. Comparison of disk diffusion, Etest and VITEK2 for detection
of carbapenemase-producing Klebsiella pneumoniae with EUCAST and
CLSI breakpoint systems. Clin. Microbiol. Infect. [Epub ahead of print.]
16. Weisenberg, S. A., D. J. Morgan, R. Espinal-Witter, and D. H. Larone. 2009.
Clinical outcomes of patients with Klebsiella pneumoniae carbapenemase-
producing K. pneumoniae after treatment with imipenem or meropenem.
Diagn. Microbiol. Infect. Dis. 64:233–235.
17. Woodford, N., et al. 2010. Comparison of BD Phoenix, Vitek 2, and
MicroScan automated systems for detection and inference of mechanisms
responsible for carbapenem resistance in Enterobacteriaceae. J. Clin.
18. Yigit, H., et al. 2001. Novel carbapenem-hydrolyzing beta-lactamase, KPC-1,
from a carbapenem-resistant strain of Klebsiella pneumoniae. Antimicrob.
Agents Chemother. 45:1151–1161.
19. Zarkotou, O., et al. 2010. Risk factors and outcomes associated with acqui-
sition of colistin-resistant KPC-producing Klebsiella pneumoniae: a matched
case-control study. J. Clin. Microbiol. 48:2271–2274.
VOL. 49, 2011 NOTES1147