EDITORIAL COMMENTARY
Bloodstream Infection Caused by Extended-Spectrum
β-Lactamase–Producing Gram-Negative Bacteria: How
to Define the Best Treatment Regimen?
Federico Perez1,2,3 and Robert A. Bonomo1,2,3,4,5
1
Medicine and 2Research Services, Louis Stokes Cleveland Department of Vetarans Affairs Medical Center, and Departments of 3Medicine, 4Pharmacology, and 5Molecular
Biology and Microbiology, Case Western Reserve University School of Medicine, Cleveland, Ohio
(See the Major Article by Tamma et al on pages 1319–25.)
Keywords. extended-spectrum β-lactamases; β-lactam/β-lactamase inhibitor combinations; carbapenems; bloodstream infection.
Extended-spectrum β-lactamases (ESBLs),
found chiefly among Escherichia coli,
Klebsiella pneumoniae, Klebsiella oxytoca,
and Proteus species, are a diverse group of
bacterial enzymes that share the ability to
hydrolyze third-generation cephalospo-
rins and aztreonam, yet are inhibited by
the commercially available β-lactamase
inhibitors (clavulanic acid, sulbactam,
and tazobactam) [1]. Since the discovery
of the first ESBL-producing K. pneumo-
niae in Germany nearly 30 years ago,
gram-negative bacteria expressing ESBLs
have posed a very serious therapeutic
challenge [2]. We have learned, from the
cost of human life, the tremendous clini-
cal significance carried by the presence of
ESBLs.
The molecular epidemiology of ESBL-
producing gram-negative bacteria has
evolved dramatically in the past decades.
After an early dominance of ESBLs belong-
Received 23 December 2014; accepted 25 December 2014;
electronically published 13 January 2015.
Correspondence: Robert A. Bonomo, MD, Louis Stokes
Cleveland Department of Veterans Affairs Medical Center,
10701 East Blvd, Cleveland, OH 44106 (robert.bonomo@va.
gov).
Clinical Infectious Diseases® 2015;60(9):1326–9
Published by Oxford University Press on behalf of the Infectious
Diseases Society of America 2015. This work is written by (a) US
Government employee(s) and is in the public domain in the US.
DOI: 10.1093/cid/civ007
1326 • CID 2015:60 (1 May) • EDITORIAL COMMENTARY
ing to the TEM and SHV β-lactamases,
we have seen in the past decade the global
emergence of CTX-M–type enzymes
[3]. Presently, in E. coli the CTX-M
β-lactamases far outnumber the TEM
and SHV types, and are linked with the
dissemination of the “high-risk clone” se-
quence type (ST) 131, associated with
increased virulence [4]. In addition, novel
β-lactamases of the PER, OXA, VEB, GES,
BES, TLA, and SFO families have only
added more complexity to this challenging
area (http://www.lahey.org/Studies/). We
know from systematic molecular sur-
veys that genes that encode for ESBLs
are usually found on large plasmids ac-
companied by genetic determinants of
resistance against other antibiotics, such
as aminoglycosides, sulfa drugs, and fluo-
roquinolones. We have also become
aware that patients who develop coloniza-
tion or infection with these pathogens
often are previously treated with fluoro-
quinolones and third-generation cepha-
losporins, usually are seriously ill, and
are exposed heavily to nosocomial inter-
ventions and environment [5]. It is inter-
esting that despite our longstanding
knowledge of these risk factors, the corre-
lation between cephalosporin use and the
emergence of ESBLs persists. Additional-
ly, there are many ongoing issues related
Downloaded from https://academic.oup.com/cid/article/60/9/1326/404309 by guest on 06 January 2022
to the detection of ESBLs in clinical mi-
crobiology laboratories, including a re-
cent shift from phenotypic confirmation
with clavulanic acid to lowering cephalo-
sporin breakpoints [6]. These difficulties
are highlighted by pathogens such as K.
oxytoca where detection of ESBL produc-
tion in the laboratory is problematic, as
the ESBL phenotype is mediated by the
hyperproduction of a chromosomal en-
zyme called K1 that hydrolyzes some
cephalosporins but not others [7].
The prolonged time that transpires
between the initial detection of bacterial
infection and the eventual determination
of minimum inhibitory concentrations
(MICs) and detection of ESBLs in the
laboratory dictates a period of “empiric
therapy.” Thus, defining appropriate ther-
apy for suspected or confirmed infections
caused by ESBL-producing organisms re-
mains an elusive goal. Nevertheless, there
are some certainties. To start, studies from
some time ago show that carbapenems
should be regarded as the drugs of choice
for serious infections caused by ESBL-
producing organisms; clearly, oxyimino
cephalosporins are a poor choice against
these pathogens [8].
The role of other agents is not as well
defined. Despite its superior ability to
readily penetrate gram-negative cells,
cefepime (an aminothiazolylacetamido
cephalosporin) is rapidly hydrolyzed by
ESBLs such as CTX-M-10 and TEM-10.
Although in vitro susceptibility may ap-
pear encouraging, cefepime does not
achieve the appropriate pharmacokinetic
and pharmacodynamic (PK/PD) param-
eters that are required to treat serious
infections caused by ESBL-producing or-
ganisms, especially with MICs >2 µg/mL,
and may be subject to the inoculum effect
[9, 10]. Importantly, we know from sever-
al clinical studies that cefepime performs
less satisfactorily than carbapenems in
the treatment of ESBL-producing or-
ganisms responsible for bloodstream in-
fection and nosocomial pneumonia
[11–13]. This is in contrast with infec-
tions caused by Enterobacteriaceae that
constitutively express AmpC β-lactamases;
cefepime, unlike other cephalosporins,
has poor affinity for these enzymes
and stands as a good option for the treat-
ment of infections caused by AmpC-
producing organisms, provided ESBLs
are not in the background [14, 15].
What about β-lactam/β-lactamase in-
hibitor combinations? There is clear in
vitro demonstration of the ability of β-
lactamase inhibitors to inactivate ESBLs
[16]. There are, however, also misgivings
given the potential role of the inoculum
effect, more apparent with piperacillin-
tazobactam than with amoxicillin–
clavulanic acid [17]. PK/PD models may
support the use of piperacillin/tazobactam
for ESBL-producing E. coli, especially if
resorting to high doses and prolonged in-
fusions, but are much less encouraging
for K. pneumoniae [18]. In fact, several
observational studies ranging from case
descriptions to multivariate analyses re-
port treatment failures when piperacil-
lin-tazobactam is used to treat infections
caused by ESBL-producing K. pneumo-
niae [19, 20].
In this issue of Clinical Infectious Dis-
eases, Tamma et al present the results of a
retrospective observational study com-
paring empiric treatment of bloodstream
infection caused by ESBL-producing
E. coli, Klebsiella species, and Proteus spe-
cies with piperacillin-tazobactam vs carba-
penems; the authors found that empiric
treatment with piperacillin-tazobactam
was associated with increased risk of mor-
tality at 14 days (adjusted hazard ratio,
1.92 [95% confidence interval, 1.07–
3.45]). However, we also know from other
retrospective observational studies by
Rodríguez-Baño et al that piperacillin-
tazobactam and amoxicillin–clavulanic
acid are equivalent to carbapenems for
the empiric and definite treatment of
bloodstream infection caused by ESBL-
producing E. coli [21].
Why the discrepancy? What are the
real differences between the analyses?
Certainly, both groups of drugs should be
effective. Carbapenems, by a process that
is poorly appreciated, are “trapped” in the
active site of β-lactamases for prolonged
periods of time and act as inhibitors, in
a similar fashion to tazobactam and clav-
ulanic acid. Additionally, carbapenems
and piperacillin or amoxicillin are all ex-
cellent β-lactams that readily inactivate
penicillin-binding proteins. Mechanisti-
cally, at least, there are no clear differences
between the 2 classes of drugs. Similarly,
both studies demonstrate methodological
soundness, notwithstanding their obser-
vational design. In both instances, the in-
vestigators chose to match the treatment
and comparator groups according to a
propensity score derived from the proba-
bility of receiving the treatment, which is
the current gold standard among the
strategies to analyze observational data.
This attempts to overcome the selection
bias inherent to observational studies,
where treatment is chosen by clinicians
and not randomly allocated as in experi-
mental studies, resulting in imbalances
between the treatment groups.
Perhaps conflicting observations arise
because the objects that are being contem-
plated are so different in terms of (1) source
of infection, (2) genetic background of
the pathogen, and (3) regional molecular
epidemiology. In studies supporting the
efficacy of piperacillin-tazobactam for
EDITORIAL COMMENTARY
the treatment of ESBL-producing gram-
negative bacteria from Spain and beyond,
we see a concentration of attention on
E. coli predominantly arising from the
urinary and biliary tracts [21, 22]. In the
Tamma et al study, we see a collection
of strains that include not only E. coli,
but also K. pneumoniae, K. oxytoca, and
Proteus mirabilis, while central lines and
pneumonia are the most common sourc-
es of bacteremia. These contrasting distri-
butions may especially matter, because
respiratory tract infections imply a high
inoculum of bacteria in a compartment
Downloaded from https://academic.oup.com/cid/article/60/9/1326/404309 by guest on 06 January 2022
where penetration of antibiotics may be
impaired, whereas urinary tract infec-
tions have a more moderate inoculum
and β-lactams easily concentrate in the
urine. Furthermore, there may be differ-
ences in permeability, virulence determi-
nants, and host predilection between the
different bacterial species, as well as vari-
ations in the type, quantities, and expres-
sion of β-lactamases. A recent survey of
ESBLs in the United States found that
CTX-M-15 and CTX-M-14 are the most
common ESBLs both in E. coli and K.
pneumoniae, and that K. pneumoniae fre-
quently co-harbors several types of SHV-
like ESBLs. Of note, OXA-1/30, which
can confer resistance to both piperacillin
and tazobactam, is frequently associated
with CTX-M-15 both in E. coli and K.
pneumoniae [23]. Surveys in Spain have
also revealed the increasing importance
of this mechanism in E. coli ST131, but
also demonstrate its association with bac-
teremia from non–urinary tract and non–
biliary tract sources [24]. The presence of
OXA-1/30 and CTX-M-15 in the emer-
gent E. coli ST131 may complicate the
therapeutic role of β-lactam/β-lactamase
inhibitor combinations in the future.
How do we reconcile these conflicting
observations and decide treatment for
our patients? Based on the available evi-
dence, carbapenems may be the best “em-
piric therapy” for patients with serious
bloodstream infections caused by ESBL-
producing bacteria with a “complex genet-
ic background,” including K. pneumoniae
• CID 2015:60 (1 May) • 1327
and multiple β-lactamases. Combina-
tions of β-lactams with β-lactamase in-
hibitors remain an effective treatment
option for bloodstream infections caused
by ESBL-producing E. coli originating
from the urinary tract, or with low MICs
(≤2 µg/mL in the case of piperacillin-
tazobactam). Clues that may guide us
in this decision can be gleaned from a
thoughtful consideration of the clinical
presentation, risk factors, and antecedent
microbiology, with careful inspection of
the results of the antibiogram. When in-
corporated into the clinic, the results of
rapid molecular screening may reveal just
how complex the β-lactamase background
really is. Even then, we may not be able to
anticipate the role of efflux pumps, loss of
porin channels, and alterations in penicillin-
binding proteins that conspire to thwart
our best diagnostic and therapeutic efforts.
This landscape will change, not only as
the result of the adoption of rapid molec-
ular diagnostics, but also as new therapies
enter the arena. The ability to rapidly de-
termine the identity of the causative agent
of bacteremia and its resistance mecha-
nisms promises to shorten the period of
empiric therapy from 3.5 days, as in the
present study, to 1 day or less. Novel β-
lactamase inhibitors with activity against
ESBLs, such as avibactam and relebac-
tam, are in advanced stages of clinical de-
velopment. Ceftolozane-tazobactam, a
new cephalosporin combined with an es-
tablished β-lactamase inhibitor, recently
received approval by the US Food and
Drug Administration for the treatment
of urinary tract and intrabdominal in-
fections. In vitro data and PK/PD sim-
ulations indicate that the addition of
tazobactam extends the activity of cefto-
lozane to include “most” ESBL producers,
with improved efficacy over piperacillin-
tazobactam. This drug could certainly
position itself as a carbapenem-sparing
agent in the treatment of infections caused
by ESBL-producing bacteria. However, in
some rare strains with high resistance, the
reduction in MICs afforded by the inhib-
itor is insufficient to attain susceptibility
1328 • CID 2015:60 (1 May) • EDITORIAL COMMENTARY
[25]. A randomized controlled trial to
compare piperacillin-tazobactam to mer-
openem for the definitive treatment of
bloodstream infection caused by ceftriaxone
non-susceptible E. coli and Klebsiella sp.
is underway (MERINO trial, Professor
D. L. Paterson, NCT02176122). This and
other studies conducted on isolates whose
genetic background has been defined pre-
cisely, will determine the future role of
β-lactam/β-lactamase inhibitor combina-
tions vs carbapenems in the treatment of se-
rious infections caused by ESBL-producing
organisms, and help preserve and enhance
the utility of our antibiotic armamentarium.
Notes
Disclaimer. Funding organizations were not
involved in the preparation, review, or approval of
the manuscript. The content is solely the responsi-
bility of the authors and does not represent the of-
ficial views of the National Institutes of Health
(NIH) nor the Department of Veterans Affairs.
Financial support. This work was supported
by the Clinical and Translational Science Collab-
orative of Cleveland (UL1TR000439) from the
National Center for Advancing Translational Sci-
ences component of the NIH and NIH Roadmap
for Medical Research, the Veterans Affairs Merit
Review Program (award number 1I01BX001974),
the NIH (grant number AI072219-05 and
AI063517-07), and the Geriatric Research Educa-
tion and Clinical Center VISN 10.
Potential conflict of interest. R. A. B. has re-
ceived grant support from AstraZeneca, Melinda,
and Steris. F. P. has no potential conflicts to report.
Both authors have submitted the ICMJE Form
for Disclosure of Potential Conflicts of Interest.
Conflicts that the editors consider relevant to the
content of the manuscript have been disclosed.
References
1. Bush K, Jacoby GA. Updated functional clas-
sification of beta-lactamases. Antimicrob
Agents Chemother 2010; 54:969–76.
2. Knothe H, Shah P, Krcmery V, Antal M,
Mitsuhashi S. Transferable resistance to
cefotaxime, cefoxitin, cefamandole and ce-
furoxime in clinical isolates of Klebsiella
pneumoniae and Serratia marcescens. Infec-
tion 1983; 11:315–7.
3. Pitout JD, Laupland KB. Extended-spectrum
beta-lactamase-producing Enterobacteria-
ceae: an emerging public-health concern.
Lancet Infect Dis 2008; 8:159–66.
4. Petty NK, Ben Zakour NL, Stanton-Cook M,
et al. Global dissemination of a multidrug re-
sistant Escherichia coli clone. Proc Natl Acad
Sci U S A 2014; 111:5694–9.
5. Paterson DL, Bonomo RA. Extended-
spectrum beta-lactamases: a clinical update.
Clin Microbiol Rev 2005; 18:657–86.
6. Huang Y, Carroll KC, Cosgrove SE, Tamma
PD. Determining the optimal ceftriaxone
MIC for triggering extended-spectrum beta-
lactamase confirmatory testing. J Clin Micro-
biol 2014; 52:2228–30.
7. Gheorghiu R, Yuan M, Hall LM, Livermore
DM. Bases of variation in resistance to
beta-lactams in Klebsiella oxytoca isolates hy-
perproducing K1 beta-lactamase. J Antimi-
crob Chemother 1997; 40:533–41.
8. Paterson DL, Ko WC, Von Gottberg A, et al.
International prospective study of Klebsiella
pneumoniae bacteremia: implications of ex-
tended-spectrum beta-lactamase production
in nosocomial infections. Ann Intern Med
Downloaded from https://academic.oup.com/cid/article/60/9/1326/404309 by guest on 06 January 2022
2004; 140:26–32.
9. Thomson KS, Moland ES. Cefepime, pipera-
cillin-tazobactam, and the inoculum effect
in tests with extended-spectrum beta-lacta-
mase-producing Enterobacteriaceae. Antimi-
crob Agents Chemother 2001; 45:3548–54.
10. Tam VH, McKinnon PS, Akins RL, Rybak MJ,
Drusano GL. Pharmacodynamics of cefepime
in patients with gram-negative infections. J
Antimicrob Chemother 2002; 50:425–8.
11. Lee NY, Lee CC, Huang WH, Tsui KC,
Hsueh PR, Ko WC. Cefepime therapy for
monomicrobial bacteremia caused by cefe-
pime-susceptible extended-spectrum beta-
lactamase-producing Enterobacteriaceae: MIC
matters. Clin Infect Dis 2013; 56:488–95.
12. Chopra T, Marchaim D, Veltman J, et al.
Impact of cefepime therapy on mortality
among patients with bloodstream infections
caused by extended-spectrum-beta-lactamase-
producing Klebsiella pneumoniae and
Escherichia coli. Antimicrob Agents Chemo-
ther 2012; 56:3936–42.
13. Zanetti G, Bally F, Greub G, et al. Cefepime
versus imipenem-cilastatin for treatment of
nosocomial pneumonia in intensive care
unit patients: a multicenter, evaluator-blind,
prospective, randomized study. Antimicrob
Agents Chemother 2003; 47:3442–7.
14. Hilty M, Sendi P, Seiffert SN, et al. Character-
isation and clinical features of Enterobacter
cloacae bloodstream infections occurring at
a tertiary care university hospital in Switzer-
land: is cefepime adequate therapy? Int J
Antimicrob Agents 2013; 41:236–49.
15. Tamma PD, Girdwood SC, Gopaul R, et al.
The use of cefepime for treating AmpC
beta-lactamase-producing Enterobacteria-
ceae. Clin Infect Dis 2013; 57:781–8.
16. Drawz SM, Bonomo RA. Three decades of
beta-lactamase inhibitors. Clin Microbiol
Rev 2010; 23:160–201.
17. Lopez-Cerero L, Picon E, Morillo C, et al.
Comparative assessment of inoculum effects
on the antimicrobial activity of amoxycillin-
clavulanate and piperacillin-tazobactam with
extended-spectrum beta-lactamase-produc-
ing and extended-spectrum beta-lactamase-
 non-producing Escherichia coli isolates. Clin
Microbiol Infect 2010; 16:132–6.
18. Ambrose PG, Bhavnani SM, Jones RN. Phar-
macokinetics-pharmacodynamics of cefe-
pime and piperacillin-tazobactam against
Escherichia coli and Klebsiella pneumoniae
strains producing extended-spectrum beta-
lactamases: report from the ARREST pro-
gram. Antimicrob Agents Chemother 2003;
47:1643–6.
19. Zimhony O, Chmelnitsky I, Bardenstein R,
et al. Endocarditis caused by extended-
spectrum-beta-lactamase-producing Klebsi-
ella pneumoniae: emergence of resistance
to ciprofloxacin and piperacillin-tazobactam
during treatment despite initial susceptibil-
ity. Antimicrob Agents Chemother 2006;
50:3179–82.
20. Tumbarello M, Sanguinetti M, Montuori E,
et al. Predictors of mortality in patients with
bloodstream infections caused by extended-
spectrum-beta-lactamase-producing Entero-
bacteriaceae: importance of inadequate initial
antimicrobial treatment. Antimicrob Agents
Chemother 2007; 51:1987–94.
21. Rodríguez-Baño J, Navarro MD, Retamar P,
Picon E, Pascual A. Beta-lactam/beta-lactam
inhibitor combinations for the treatment of
bacteremia due to extended-spectrum beta-
lactamase-producing Escherichia coli: a post
hoc analysis of prospective cohorts. Clin
Infect Dis 2012; 54:167–74.
22. Gutierrez-Gutierrez BS, Salamanca E, Perez-
Galera S, et al. Assessment of β-lactam/β-
lactam inhibitor (BL/BLI) combinations for
the treatment of bacteremia due to extend-
ed-spectrum β-lactamase (ESBL)-producing
Enterobacteriaceae: The INCREMENT
project. Abstract K 1041. In: Interscience
Conference on Antmicrobial Agents and
Chemotherapy. Washington, DC: American
Society for Microbiology, 2014.
EDITORIAL COMMENTARY
23. Castanheira M, Farrell SE, Deshpande LM,
Mendes RE, Jones RN. Prevalence of beta-
lactamase-encoding genes among Enterobac-
teriaceae bacteremia isolates collected in 26 U.
S. hospitals: report from the SENTRY Antimi-
crobial Surveillance Program (2010). Antimic-
rob Agents Chemother 2013; 57:3012–20.
24. Rodriguez-Bano J, Mingorance J, Fernandez-
Romero N, Serrano L, Lopez-Cerero L,
Pascual A. Virulence profiles of bacteremic ex-
tended-spectrum beta-lactamase-producing
Escherichia coli: association with epidemio-
logical and clinical features. PloS One 2012;
7:e44238.
25. Bulik CC, Tessier PR, Keel RA, Sutherland CA,
Nicolau DP. In vivo comparison of CXA-101
(FR264205) with and without tazobactam
versus piperacillin-tazobactam using human
Downloaded from https://academic.oup.com/cid/article/60/9/1326/404309 by guest on 06 January 2022
simulated exposures against phenotypically
diverse gram-negative organisms. Antimi-
crob Agents Chemother 2012; 56:544–9.
• CID 2015:60 (1 May) • 1329