International Journal of Antimicrobial Agents 28 (2006) 578–581

Short communication

Occurrence of plasmidic AmpC type ␤-lactamase-mediated

resistance in Escherichia coli: report from the SENTRY
Antimicrobial Surveillance Program (North America, 2004)
Lalitagauri M. Deshpande a,∗ , Ronald N. Jones a,b , Thomas R. Fritsche a , Helio S. Sader a
a JMI Laboratories, 345 Beaver Kreek Centre, Suite A, North Liberty, IA 52317, USA
b Tufts University School of Medicine, Boston, MA, USA

Received 14 April 2006; accepted 31 July 2006

Abstract

Among 1429 Escherichia coli isolates collected as part of the SENTRY Antimicrobial Surveillance Program (2004) from 30 North American
medical centres, 65 (4.5%) were screen-positive for an extended-spectrum ␤-lactamase (ESBL). Among the strains with a negative ESBL
confirmatory test (n = 26; clavulanic acid inhibition), a CMY-2 enzyme was detected in 13 isolates (50.0%), FOX-5 in 3 isolates (11.5%)
and DHA-1 in 1 isolate (3.8%). These AmpC-producing E. coli were cephamycin (cefoxitin)-resistant but susceptible to cefepime (minimum
inhibitory concentration (MIC) ≤0.12–4 mg/L). Clearly, the ESBL tests recommended by the Clinical and Laboratory Standards Institute
identify only a fraction of E. coli with elevated ␤-lactam MIC values as ESBL-producing strains; the majority of the remaining strains would
be potentially responsive to some other ␤-lactams, directed by accurately performed and interpreted susceptibility methods.
© 2006 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

Keywords: AmpC; CMY-2; ESBL; Escherichia coli; SENTRY Program

1. Introduction confer broad-spectrum resistance to most ␤-lactams (other

Escherichia coli rank highest among Gram-negative
pathogens causing significant infections worldwide, but they
generally remain susceptible to many older ␤-lactams and
cephalosporins. Whilst E. coli does carry a chromosomal
AmpC gene, its expression is only at a basal level and thus
does not confer resistance to ampicillin or cephalosporins.
Mutations in the promoter and/or the attenuator regions of
this gene can, however, occasionally result in hyperproduc-
tion of the AmpC enzyme. In recent years, some E. coli strains
carrying plasmid-mediated AmpC genes ‘imported’ from
other Enterobacteriaceae have been identified with promoter
modifications responsible for high-level expression [1–3].
Plasmid-mediated AmpC determinants are clinically signif-
icant as they have the ability to spread ␤-lactam resistance
further among Gram-negative bacilli [4]. These genes can
∗ Corresponding author. Tel.: +1 319 665 3370; fax: +1 319 665 3371.
E-mail address: gauri-deshpande@jmilabs.com (L.M. Deshpande).
than cefepime and carbapenems) and hence pose major ther-
apeutic challenges [2,3].
According to the criteria of the Clinical and Laboratory
Standards Institute (CLSI) [5], E. coli isolates with minimum
inhibitory concentrations (MICs) ≥2 mg/L for ceftazidime,
ceftriaxone, cefotaxime or aztreonam are potential extended-
spectrum ␤-lactamase (ESBL) producers. The CLSI recom-
mends that a confirmatory clavulanic acid inhibition test be
performed on these ‘screen test’-positive clinical isolates
before reporting susceptibility results for cephalosporins and
the monobactam aztreonam [5]. Enterobacteriaceae isolates
with an ESBL phenotype but a negative ESBL confirma-
tory test are potential candidates for production of an AmpC
enzyme, either mediated by chromosomal derepression or
through mobilisation on a plasmid. Many clinical microbi-
ologists appear to be unaware of the presence of plasmid-
mediated AmpC ␤-lactamase enzymes in resistant isolates
because phenotypic detection can be difficult and the strains
may be misidentified as ESBL producers [6,7]. It is important

0924-8579/$ – see front matter © 2006 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.
doi:10.1016/j.ijantimicag.2006.07.025
 L.M. Deshpande et al. / International Journal of Antimicrobial Agents 28 (2006) 578–581 579

to differentiate AmpC and ESBL enzymes as mechanisms of
resistance in order to direct physicians in prescribing the most
appropriate ␤-lactam agents, to assist in hospital infection
control investigations and to decrease antimicrobial selective
pressure towards resistance. The prevalence of the AmpC
type resistance mechanisms among E. coli isolates in the
USA and Canada was systematically assessed by the SEN-
TRY Antimicrobial Surveillance Program (2004).
2. Materials and methods
2.1. Bacterial isolates
As part of the SENTRY Program, a total of 1429 E. coli
isolates were processed from 30 medical centres in North
America. Isolates were collected consecutively from blood-
stream infections, skin and soft tissue infections, infections
of paediatric patients in the intensive care unit and pneu-
monias in hospitalised patients; all were submitted to the
monitoring laboratory (JMI Laboratories, North Liberty, IA)
for further processing. Only non-duplicate isolates respon-
sible for significant infection were included in the study.
Species identification was confirmed at the monitoring lab-
oratory using standard biochemical tests and Vitek system
cards (bioMérieux, Hazelwood, MO).
2.2. Susceptibility testing
All isolates were tested for susceptibility (Table 1) using
the broth microdilution method as described by the CLSI
[8] in validated panels manufactured by TREK Diagnostics
(Cleveland, OH). Interpretations of susceptibility results for
all antimicrobials tested were those published by the CLSI
[5].
2.3. ESBL and AmpC phenotypic tests
A confirmatory test was performed on all isolates meeting
the CLSI criteria for potential ESBL production [5]. Briefly,
an inoculum suspension equal to a 0.5 McFarland standard
was made from fresh culture and was used to inoculate the
surface of a Mueller–Hinton agar plate. A disk containing
amoxicillin/clavulanic acid (Remel, Lenexa, KS) or clavu-
lanic acid alone (10 ␮g) was placed in the centre of the
plate and disks containing ␤-lactam substrates, namely cef-
tazidime, ceftriaxone and aztreonam (Remel), were placed
15 mm from the disk (edge to edge) containing clavulanic
acid. Widening of the zone of inhibition or a ‘phantom zone’
effect between any or all of the substrate/inhibitor combi-
nations was considered a positive ESBL confirmatory test.
This was confirmed by the Etest (AB Biodisk, Solna, Swe-
den) ESBL strip using ceftazidime and cefotaxime with and
without clavulanic acid.
Isolates with an ESBL phenotype but a negative ESBL
confirmatory test were considered potential AmpC producers
and were screened for production of AmpC by the three-
dimensional (3D) screening test described by Coudron et al.
[9], using cefoxitin as a substrate.
2.4. Multiplex polymerase chain reaction (PCR) for
characterisation of blaAmpC
Those E. coli isolates with a positive 3D AmpC test were
screened for nine different families of acquired AmpC genes
(MOX-1, -2; CMY-1 to -11; LAT-1 to -4; BIL-1; DHA-
1, -2; ACC; MIR-1; ACT-1; FOX-1 to -5) by a multiplex
PCR screening procedure as described previously [7,10].
Tentative enzyme type designations were assigned based
on the size of the PCR amplicon and were further con-
firmed by amplification with blaAmpC -specific primers [11]
followed by sequencing. The PCR products were processed
using the Qiagen (Hilden, Germany) PCR purification kit.
Sequencing was performed using the Sanger-based dideoxy
sequencing strategy involving incorporation of fluorescent
dye-labelled terminators into the sequencing reaction prod-
ucts. Sequences obtained were compared with the publicly
available sequences via a National Centre for Biotechnol-
ogy Information BLAST search to determine the type of
␤-lactamase.

Table 1
In vitro activity of nine selected antimicrobial agents tested against Escherichia coli (n = 1429) isolated in North American medical centres of the SENTRY
Program (2004)
Antimicrobial agent MIC (mg/L) Category (%) % with ESBL phenotype (MIC ≥ 2 mg/L)

MIC50 MIC90 Susceptible Resistant

Aztreonam ≤0.12 0.25 97.4 1.8 4.1a
Ceftazidime ≤1 ≤1 96.9 1.5 4.3a
Ceftriaxone ≤0.25 ≤0.25 97.1 2.1 3.8a
Cefoxitin 4 8 91.2 3.4 –
Cefepime ≤0.12 ≤0.12 98.5 1.3 –
Piperacillin/tazobactam 2 4 96.4 2.0 –
Imipenem ≤0.12 ≤0.12 100.0 0.0 –
Ciprofloxacin ≤0.03 >4 85.2 14.9 –
Gentamicin ≤2 4 91.3 8.2 –
MIC, minimum inhibitory concentration; ESBL, extended-spectrum ␤-lactamase.
a A total of 4.5% of isolates had an elevated MIC (≥2 mg/L) for at least one of these ␤-lactam substrates [5].
580 L.M. Deshpande et al. / International Journal of Antimicrobial Agents 28 (2006) 578–581
3. Results and discussion
Third-generation cephalosporins were highly active
against the 1429 E. coli isolates (Table 1), with cef-
tazidime, ceftriaxone and aztreonam exhibiting similar
potency (96.9–97.4% susceptible). However, only 91.2% of
the isolates were susceptible to cefoxitin, although the resis-
tance rate was only 3.4%. Cefepime activity was greatest
(98.5% susceptible) among the cephalosporins tested. All iso-
lates were inhibited by 4 mg/L imipenem; only three isolates
(0.21%) showed a MIC ≥0.5 mg/L. Fluoroquinolone resis-
tance among the E. coli isolates was significantly elevated,
despite the generally susceptible nature of this species in
North America. Resistance to ciprofloxacin rose from 14.9%
in the overall population of E. coli isolates to 46.2% among
those strains expressing AmpC ␤-lactamases.
A total of 65 isolates (4.5%) met the CLSI screening crite-
ria for ESBL production, however only 39 (2.7%) were con-
firmed as ESBL producers. Nineteen (63.3%) of the 30 med-
ical centres participating had ESBL-positive E. coli isolates.
PCR screens for AmpC genes followed by sequencing of the
non-ESBL-producing isolates (n = 26) revealed CMY-2 in 13
isolates (50.0%), FOX-5 in 3 isolates (11.5%) and DHA-1 in
1 isolate (3.8%) [4]. All CMY-2- and FOX-5-producing E.
coli were resistant to cefoxitin; however, these isolates were
susceptible to cefepime (MIC ≤ 0.12–4 mg/L). The DHA-1-
producing isolate was susceptible to all cephalosporins [5]
tested and the highest MIC values were documented for
ceftazidime and cefoxitin (8 mg/L). This was the only E.
coli isolate in the presented sample that coincidently had a
polymyxin B MIC > 8 mg/L (data not shown).
One isolate from a New York City medical centre showed
elevated MIC values for imipenem and meropenem (2 mg/L
and 4 mg/L, respectively) in addition to cephalosporin resis-
tance and was found to produce a KPC-3 serine carbapene-
mase [12]. Five E. coli isolates showing a positive 3D AmpC
test and a negative ESBL confirmatory test failed to gener-
ate PCR products. These isolates may produce enzymes not
covered by the selected PCR primer sets or may have addi-
tional mutations affecting porin channels for antimicrobial
uptake [13]. The three remaining strains had a negative 3D
AmpC test and only an isolated elevation of the ceftazidime
MIC (2 mg/L), the latter indicating a lack of specificity at this
ESBL screening concentration [5].
The geographic distribution of AmpC-producing isolates
is listed in Table 2. CMY-2 and FOX-5 were prevalent among
␤-lactam-resistant E. coli from the New York area. Eight
(47.1%) of 17 AmpC producers and one KPC-3 carbapen-
emase producer were isolated from three medical centres in
New York state. Widely discordant antibiograms were present
in strains having the same AmpC enzyme in the same medi-
cal centre of this state. The number of antimicrobial classes
with different susceptibilities (susceptible versus resistant or
vice versa) was two to six among 34 tested agents.
No ESBL- or AmpC-producing E. coli were isolated
from SENTRY Program medical centres in Canada, although
Table 2
Distribution of mobile AmpC enzyme types detected in Escherichia coli
strains isolated from various US medical centresa
Location of medical ␤-Lactamase type (no. of isolates)
centre (state)
CMY-2 FOX-5 DHA-1
New York 5 (A, B)b 3 (A, C)b –
Indiana 2 – –
Hawaii 1 – –
Kentucky 1 – –
Massachusetts 1 – –
Nebraska 1 – –
Texas 1 – –
Wisconsin 1 – –
Utah – – 1
a No strains were detected in Canada.
b
A, B and C are designations for three different medical centres in New
York state. The five CMY-2 and three FOX-5 enzyme-producing strains had
widely varying antibiograms.
occurrences of CMY-2 have been described previously
among epidemiologically diverse E. coli strains in that coun-
try [14]. CMY-2 originated from the chromosomal AmpC
gene of Citrobacter freundii and has now become more preva-
lent among Salmonella spp. obtained from food animals and
human patients [4,15]. Third-generation cephalosporin resis-
tance can be used as a potential marker for CMY-2-producing
Salmonella spp. [16]; however, ceftriaxone MIC values of
CMY-2- and FOX-5-producing E. coli isolates in the SEN-
TRY Program (2004) ranged from 8 mg/L (susceptible) to
>32 mg/L and from 2 mg/L to 8 mg/L (susceptible), respec-
tively. On the other hand, cefoxitin MIC values consistently
displayed high-level resistance (>32 mg/L). Thus, cefoxitin
resistance in combination with positive ESBL screening crite-
ria (≥2 mg/L) usually serves as a sensitive indicator of AmpC
production by E. coli isolates in this region [5].
Clearly, plasmidic AmpC enzymes among E. coli are
responsible for a significant number of isolates present-
ing with elevated MIC results for newer cephalosporins,
cephamycins and other ␤-lactams [1–3,7,9,10,14,17].
Differentiation between these enzymes and the ESBLs
requires supplemental phenotypic or molecular testing
[5] if the current ␤-lactam interpretive breakpoints of the
CLSI are utilised. Therefore, one must question why the
susceptibility breakpoint criteria of the CLSI are appropriate
for a CMY-2-like enzyme-producing isolate with a MIC to
ceftriaxone (example only) of 2, 4 or 8 mg/L, but are not
so for an ESBL-producer having the same MIC for that
third-generation cephem or an even more active fourth-
generation product. Thoughtfully selected susceptibility
breakpoints founded upon the principles of pharmacody-
namics (with Monte Carlo simulations and target attainment
assessments) and wild-type MIC population distributions
have produced candidate values that would be predictive
of clinical therapeutic success regardless of the resistance
mechanism harboured by the organism, and would relegate
the use of ESBL screening/confirmatory methods to guide
infection control interventions [18; http://www.escmid.org/
 L.M. Deshpande et al. / International Journal of Antimicrobial Agents 28 (2006) 578–581
sites/index f.aspx?par=2.4] and leave the classification
of mobile AmpC enzymes to structured surveillance
programmes.
References
[1] Alvarez M, Tran JH, Chow N, Jacoby GA. Epidemiology of conjugative
plasmid-mediated AmpC beta-lactamases in the United States. Antimi-
crob Agents Chemother 2004;48:533–7.
[2] Odeh R, Kelkar S, Hujer AM, Bonomo RA, Schreckenberger PC, Quinn
JP. Broad resistance due to plasmid-mediated AmpC beta-lactamases
in clinical isolates of Escherichia coli. Clin Infect Dis 2002;35:140–5.
[3] Philippon A, Arlet G, Jacoby GA. Plasmid-determined AmpC-type
beta-lactamases. Antimicrob Agents Chemother 2002;46:1–11.
[4] Liebana E, Batchelor M, Clifton-Hadley FA, Davies RH, Hopkins KL,
Threlfall EJ. First report of Salmonella isolates with the DHA-1 AmpC
beta-lactamase in the United Kingdom. Antimicrob Agents Chemother
2004;48:4492.
[5] Clinical and Laboratory Standards Institute. Performance standards for
antimicrobial susceptibility testing. 16th Informational Supplement.
M100-S16. Wayne, PA: CLSI; 2006.
[6] College of American Pathologists. Surveys Critique D19 (2005).
Northfield, IL: College of American Pathologists. http://www.cap.org
(accessed 28 September 2006).
[7] Hanson ND. AmpC beta-lactamases: what do we need to know for the
future? J Antimicrob Chemother 2003;52:2–4.
[8] Clinical and Laboratory Standards Institute. Methods for dilution
antimicrobial susceptibility tests for bacteria that grow aerobically. 7th
ed. Approved Standard M7-A7. Wayne, PA: CLSI; 2006.
[9] Coudron PE, Moland ES, Thomson KS. Occurrence and detection of
AmpC beta-lactamases among Escherichia coli, Klebsiella pneumo-
niae, and Proteus mirabilis isolates at a veterans medical center. J Clin
Microbiol 2000;38:1791–6.
581
[10] Perez-Perez FJ, Hanson ND. Detection of plasmid-mediated AmpC
beta-lactamase genes in clinical isolates by using multiplex PCR. J
Clin Microbiol 2002;40:2153–62.
[11] Carattoli A, Tosini F, Giles WP, et al. Characterization of plasmids
carrying CMY-2 from expanded-spectrum cephalosporin-resistant
Salmonella strains isolated in the United States between 1996 and 1998.
Antimicrob Agents Chemother 2002;46:1269–72.
[12] Bratu S, Landman D, Haag R, et al. Rapid spread of carbapenem-
resistant Klebsiella pneumoniae in New York City: a new threat
to our antibiotic armamentarium. Arch Intern Med 2005;165:
1430–5.
[13] Martinez-Martinez L, Hernandez-Alles S, Alberti S, Tomas JM,
Benedi VJ, Jacoby GA. In vivo selection of porin-deficient mutants
of Klebsiella pneumoniae with increased resistance to cefoxitin and
expanded-spectrum-cephalosporins. Antimicrob Agents Chemother
1996;40:342–8.
[14] Mulvey MR, Bryce E, Boyd DA, et al. Molecular characterization of
cefoxitin-resistant Escherichia coli from Canadian hospitals. Antimi-
crob Agents Chemother 2005;49:358–65.
[15] Yan JJ, Ko WC, Chiu CH, Tsai SH, Wu HM, Wu JJ. Emergence of
ceftriaxone-resistant Salmonella isolates and rapid spread of plasmid-
encoded CMY-2-like cephalosporinase, Taiwan. Emerg Infect Dis
2003;9:323–8.
[16] Fey PD, Safranek TJ, Rupp ME, et al. Ceftriaxone-resistant
Salmonella infection acquired by a child from cattle. N Engl J Med
2000;342:1242–9.
[17] Lee K, Lee M, Shin JH, et al. Prevalence of plasmid-mediated Amp
C beta-lactamases in Escherichia coli and Klebsiella pneumoniae in
Korea. Microb Drug Resist 2006;12:44–9.
[18] Jones RN, Craig WA, Ambrose PG, Dudley MN, Pottumarthy S. Re-
evaluation of Enterobacteriaceae MIC/disk diffusion zone diameter
regression scattergrams for 9 beta-lactams: adjustments of breakpoints
for strains producing extended spectrum beta-lactamases. Diagn Micro-
biol Infect Dis 2005;52:235–46.