Download Assessment of Antimicrobial Combinations for Klebsiella

MAJOR ARTICLE
Assessment of Antimicrobial Combinations for
Klebsiella pneumoniae Carbapenemase–
Producing K. pneumoniae
Elizabeth B. Hirsch,1,2 Beining Guo,1,4 Kai-Tai Chang,1 Henry Cao,1 Kimberly R. Ledesma,1 Manisha Singh,3 and
Vincent H. Tam1
1
University of Houston, Texas; 2Northeastern University, Boston, Massachusetts; 3SUNY-Downstate Medical Center, Brooklyn, New York; and 4Institute
of Antibiotics, Huashan Hospital, Fudan University, Shanghai, China
Background. The prevalence of blaKPC among gram-negative bacteria continues to increase worldwide.
Limited treatment options exist for this multidrug-resistant phenotype, often necessitating combination therapy.
We investigated the in vitro and in vivo efficacy of multiple antimicrobial combinations.
Methods. Two clinical strains of Klebsiella pneumoniae carbapenemase (KPC)–producing K. pneumoniae
were studied. The killing activities of six 2-agent combinations of amikacin, doripenem, levofloxacin, and rifampin
were quantitatively assessed using a validated mathematical model. Combination time-kill studies were conducted
using clinically relevant concentrations; observed bacterial burdens were modeled using 3-dimensional response
surfaces. Selected combinations were further validated in a neutropenic murine pneumonia model, using humanlike dosing exposures.
Results. The most enhanced killing effect in time-kill studies was seen with amikacin plus doripenem. Compared with placebo controls, this combination resulted in significant reduction of the bacterial burden in tissue at
24 hours, along with prolonged animal survival. In contrast, amikacin plus levofloxacin was found to be antagonistic in time-kill studies, showing inferior animal survival, as predicted.
Conclusions. Our modeling approach appeared to be robust in assessing the effectiveness of various combinations for KPC-producing isolates. Amikacin plus doripenem was the most effective combination in both in vitro
and in vivo infection models. Empirical selection of combinations against KPCs may result in antagonism and
should be avoided.
Keywords. beta-lactamases; combined killing; pharmacodynamics; synergism; AmpC.
The frequency of antimicrobial resistance mediated by
Klebsiellapneumoniaecarbapenemase(KPC)–producing
bacteria is increasing at an alarming rate [1, 2]. Since
the first outbreak reports surfaced from the northeastern region of the United States, KPC-harboring isolates have become fairly widespread and have been
identified in at least 36 states [3, 4]. In the United
Received 22 May 2012; accepted 11 October 2012; electronically published 13
December 2012.
Correspondence: Vincent H. Tam, PharmD, Department of Clinical Sciences and
Administration, University of Houston College of Pharmacy, 1441 Moursund St,
Houston, TX 77030 ([email protected]).
The Journal of Infectious Diseases 2013;207:786–93
© The Author 2012. Published by Oxford University Press on behalf of the Infectious
Diseases Society of America. All rights reserved. For Permissions, please e-mail:
[email protected].
DOI: 10.1093/infdis/jis766
786
•
JID 2013:207 (1 March)
•
Hirsch et al
States, it appears that a single dominant clone, ST 258,
accounts for almost 70% of KPC-producing K. pneumoniae isolates in the Center for Disease Control’s
pulsed-field gel electrophoresis database [5]. International dissemination rates are also on the rise, with
endemic and epidemic situations reported in several
countries, including Israel, China, and Greece [1].
KPC-producing isolates represent a significant treatment problem because they are able to hydrolyze a
broad spectrum of β-lactam antibiotics. Additional
resistance mechanisms (ie, multiple enzymes and
porin changes) may also be co-carried in these isolates, conferring cross-resistance to multiple antimicrobial classes [6, 7]. Large susceptibility studies
have demonstrated that a great number of KPCharboring isolates are also resistant to the fluoroquinolones, the tetracyclines, selected aminoglycosides, and
aztreonam, in addition to most β-lactam agents [8–10]. In
general, susceptibility to a few antimicrobials, such as polymyxins and tigecycline, may only be retained [11]. As a result
of broad antimicrobial resistance, treatment options are very
limited. Further compounding the severity of the problem,
high mortality rates have been associated with infection due to
KPC-producing organisms, ranging up to 57% in some outbreaks [4, 12].
It has been noted that combination therapy may often be
necessary for successful patient outcomes, but data in humans
are still lacking and are often limited by retrospective and
noncomparative study designs [11]. Several investigations have
explored the use of various combination regimens for KPCproducing Enterobacteriaceae, but these investigations often lack
fluctuating dosing exposures or in vivo validation [13–17]. It
remains unknown which combinations of antimicrobial
agents/classes are most effective for the treatment of infection
due to KPC-producing organisms. Thus, in an effort to
provide a rationale for the selection of effective combination
therapy, we chose 4 agents (doripenem, amikacin, levofloxacin, and rifampin) representing various classes and mechanisms of action for this study. In view of the relative stability
of doripenem against hydrolysis by KPC, we chose doripenem
as our representative carbapenem [18]. The other agents were
chosen in part for their favorable rates of in vitro susceptibility
or their previously reported synergism against gram-negative
organisms [8, 9, 19]. We used these 4 agents to investigate the
in vitro and in vivo efficacy of six 2-agent combinations
against 2 clinical KPC-producing isolates.
METHODS
Bacteria
One KPC-2–producing isolate (KPVM9) obtained from
Brooklyn, New York, and 1 KPC-3–producing isolate
(KP6153) obtained from Houston, Texas, were used in the
studies. KPVM9 minimum inhibitory concentrations (MICs)
for doripenem, amikacin, rifampin, and levofloxacin were 16
mg/L, 64 mg/L, >64 mg/L, and 128 mg/L, respectively. This
isolate also possessed SHV-11 and TEM-1 β-lactamases, along
with reduced expression of ompK35 [20]. KP6153 MICs for
doripenem, amikacin, rifampin, and levofloxacin were 32 mg/L,
32 mg/L, >256 mg/L, and 8 mg/L, respectively. KP6153 also
possessed TEM-1 and a TEM-1–like enzyme with a single
mutation (E26K), in addition to an SHV-1–like enzyme with a
novel mutation (T114A). Furthermore, expression of ompK35
and ompK36 were both downregulated in this isolate. Spectrophotometric studies that used crude cell lysates from each
isolate demonstrated effective hydrolysis of imipenem [21].
Real-time reverse transcription polymerase chain reaction
studies of KP6153 revealed that KPC transcription was 3 times
that of KPVM9.
Antimicrobial Agents
For in vitro experiments, doripenem, amikacin, rifampin, and
levofloxacin were obtained from Peninsula Pharmaceuticals
(Alameda, CA), LKT Laboratories (St. Paul, MN), SigmaAldrich (St. Louis, MO), and Waterstone Technologies
(Carmel, IN), respectively. Stock solutions of each agent were
prepared in sterile water, aliquoted, and stored at −70°C. For
in vivo investigations, doripenem, levofloxacin, and amikacin
were manufactured by Shionogi (Osaka, Japan), Janssen Pharmaceuticals (Beerse, Belgium), and Sicor Pharmaceuticals
(Irvine, CA), respectively. Fresh solutions of each agent were
made according to the appropriate concentrations, frozen at
−70°C, and thawed immediately prior to administration.
Time-Kill Studies and Modeling
Single-agent and combination time-kill studies, using approximately 1–5 × 105 colony-forming units (CFU)/mL were conducted in triplicate as previously described [22, 23]. The
concentration ranges examined for combination time-kill
studies were determined on the basis of optimal sampling and
were constrained to those clinically achievable in human
serum: amikacin, 4–80 mg/L; doripenem, 4–32 mg/L; levofloxacin, 0.5–10 mg/L; and rifampin, 0.25–6 mg/L. Total bacterial
burden at 24 hours (measured in triplicate) was mathematically modeled using a 3-dimensional response surface as described in detail elsewhere (Mathematica 5.2; Wolfram
Research, Champaign, IL) [23]. Each experiment was conducted at least twice, on separate days. Volumes under the plane
(VUP) of the observed (VUPobserved) and expected
(VUPexpected) surfaces were computed by interpolation and
double integration, respectively. Confidence intervals (CIs) of
VUPobserved were computed with mean data points (±1.96
SDs). Synergy and antagonism were defined as interaction
indices (VUPobserved/VUPpredicted) of <1 and >1, respectively.
Animals
Female Swiss Webster mice weighing 20–25 g (Harlan, Indianapolis, IN) were used in the study. The mice were housed in
negative-pressure-ventilated microisolator cages to decrease
the risk of infection from extraneous pathogens. Animals were
allowed food and water ad libitum. The experimental protocol
was approved by the Institutional Animal Care and Use Committee of the University of Houston.
Experimental Pneumonia Model
Selected combinations from the in vitro model were validated
in a neutropenic murine pneumonia model. Details of this
model are described elsewhere [24, 25]. Briefly, transient neutropenia was induced in the mice by administering 2 intraperitoneal doses of cyclophosphamide (150 mg/kg 4 days prior to
infection and 100 mg/kg 1 day prior to infection) [26]. One
intraperitoneal injection of uranyl nitrate (5 mg/kg 2 days
KPC Combination Therapy
•
JID 2013:207 (1 March)
•
787
prior to infection) was also administered, to induce transient
nephrotoxicity [27]. On the day of infection, the bacterial inoculum was prepared by inoculating a flask containing prewarmed cations adjusted Mueller Hinton broth (Ca-MHB)
and placed in a shaker water bath at 35°C until reaching logphase growth. This bacterial suspension was then diluted with
Ca-MHB according to absorbance (630 nm), washed with
sterile saline, and inoculated (10 µL) into the trachea of anesthetized animals, with laryngoscopic guidance.
To adjust for virulence differences between the 2 isolates,
lethal inoculum studies were performed before pneumonia
studies. Here, mice were infected as described above. A total
of 39 mice were randomly divided into 3 groups of 13 mice
each. Three different inocula, ranging from 4 × 107 to 5 × 109
CFU/mL, were used for each bacterium. The inoculum resulting in 50%–100% mortality between 24 and 72 hours was
used in the subsequent studies. This approach was chosen in
an effort to mimic a window of opportunity in which pharmacologic intervention (clinically) may have an impact on
patient outcomes.
Immune Response Studies
Immune response was studied in 12 animals. Six animals were
infected as detailed above, while 6 control animals received
sterile saline intratracheally. After 24 hours, serum tumor necrosis factor α (TNF-α) and interleukin 6 (IL-6) levels were measured by enzyme-linked immunosorbent assay (OptEIA kit; BD
Biosciences, San Diego, CA) according to the manufacturer’s instructions and then compared between the 2 groups.
Pharmacokinetic Studies
Pharmacokinetic studies to achieve clinically relevant dosing
exposures for amikacin (1500 mg every 24 hours) and levofloxacin (750 mg every 24 hours) were previously conducted
by our group [24]. To establish human-like dosing exposures
for doripenem, 54 mice were infected as described above and
divided into 3 different dosing groups: 20, 80, and 200 mg/kg.
Two hours after infection, the mice were administered a 0.2-mL
(20 and 80 mg/kg) or 0.4-mL (200 mg/kg) intraperitoneal injection of doripenem. For each dosing group, 3 mice were sacrificed by CO2 asphyxiation, and blood was collected via
cardiac puncture serially at various time points over 6 hours.
After collection, the blood was allowed to clot on ice, and the
serum was isolated after centrifugation. Serum was immediately frozen at −70°C (for up to 2 months) until analysis. Drug
concentrations were assayed by the validated method described below. Serum doripenem concentrations were then analyzed by fitting a 1-compartment model with first-order
absorption to the data. Best-fit model parameter estimates
were used to determine the clinically relevant dose exposure.
788
•
JID 2013:207 (1 March)
•
Hirsch et al
Drug Assay
Doripenem concentrations in mouse serum were assayed
using a validated ultraperformance liquid chromatography
(UPLC) method after protein precipitation. To a 50-μL aliquot
of sample, 20 μL (40 mg/L ertapenem) was added as internal
standard, followed by 100 μL acetonitrile. The sample mixtures were mixed and precipitates were removed by centrifugation at 15 000 × g for 15 minutes. The supernatants were
recovered and evaporated to dryness at 40°C under air. The
dry residue was then reconstituted in 200 μL of 25% acetonitrile (vol/vol) and centrifuged at 15 000 × g for an additional
10 minutes. A 10-μL aliquot of the resulting supernatant solution was injected onto the UPLC system for analysis.
The UPLC system consisted of a Waters Acquity UPLC and
BEH C18 column (50 mm × 2.1 mm × 1.8 μm, Waters,
Milford, MA). Gradient elution was performed using various
proportions of mobile phases A (0.1% [vol/vol] formic acid)
and B (100% [vol/vol] acetonitrile) at 0.45 mL/min. Doripenem and ertapenem peaks were detected at 305 nm; retention
times for doripenem and ertapenem were 1.54 and 1.84
minutes, respectively. The linear range for the assay was 5–80
mg/L, and the intraday and interday variability (calculated as
coefficients of variation) were <10% and <16%, respectively.
Samples expected to be above the linear range were diluted
5–10-fold with blank mouse serum prior to assay; those expected to be below the range were concentrated 2–3-fold in
the final reconstitution step.
Assessment of Efficacy
Two hours after infection, 16 mice were randomly divided into
treatment groups. From each group, 3 mice were sacrificed by
CO2 asphyxiation prior drug treatment, to ascertain lung tissue
bacterial burden. Antibiotic doses were given by intraperitoneal
injections (0.2 mL) as follows: amikacin, 80 mg/kg every 24
hours; doripenem, 120 mg/kg every 8 hours; or levofloxacin,
150 mg/kg every 24 hours. Lung tissues were collected aseptically and homogenized in 10 mL sterile saline; the resulting pellets
were resuspended in normal saline to 10 × their original
volumes, to minimize the drug carryover effect. After 10-fold
serial dilutions, they were then quantitatively cultured on
Mueller-Hinton agar plates and incubated for 24 hours, and the
bacterial burden in lung tissues were calculated after visual inspection of colony growth. Bacterial burden in lung tissues was
assessed for each group at baseline (n = 3), after 24 hours of
treatment (n = 3), and at death or the end of the experiment
(96 hours). Survival was assessed in 10 mice per treatment
group. The mice were examined every 8 hours; at each inspection time, any animals that appeared to be moribund were humanely euthanized. Death was recorded as it occurred, at the
next inspection time. Any mice remaining at the end of the 96hour experiment were euthanized by CO2 asphyxiation.
Statistical Analysis
Lung tissue bacterial burden and serum cytokine levels were
analyzed using the Student t test. Survival was evaluated with
the Kaplan-Meier survival analysis and log-rank test. A
P value of < .05 was considered statistically significant.
RESULTS
Combination Time-Kill Studies
Selected graphical representation of combined killing effects
against KPVM9 are shown in Figure 1. The most enhanced
killing effect was seen with doripenem plus amikacin (KPVM9,
interactive index = 0.60; 95% CI, .56–.65) against both isolates
in combination time-kill studies. In contrast, amikacin plus levofloxacin was found to be most antagonistic (interactive
index = 1.11; 95% CI, 1.06–1.16) against KPVM9 (Figure 1C).
The assessment and ranking of combined killing for all combinations against KPVM9 are shown in Table 1. Interactive
indices were not calculated for KP6153, because of an inherent
limitation of the model. Reduction in bacterial burden was observed with high (but clinically relevant) drug concentrations.
As a result, the predicted combined effect of 2 agents was unrealistic (negative bacterial burden). Consequently, the agent combinations were ranked on the basis of the effect observed with
the highest drug concentration combinations (data not shown).
Two combination regimens were chosen for in vivo validation:
amikacin plus doripenem, because of the synergistic activity
seen in vitro, and amikacin plus levofloxacin, because of the
interesting antagonistic activity seen.
Experimental Pneumonia Validation
In mice infected with approximately 2 × 109 CFU/mL of
KPVM9 and 6 × 109 CFU/mL of KP6153, mortality was 50%
at 72 hours after inoculation. These inocula were used in the
subsequent experiments. Serum TNF-α and IL-6 levels were
significantly higher in infected mice, compared with control
mice; mean TNF-α concentrations (±SD) were 397.3 ± 70.4
pg/mL, compared with 14.7 ± 4.3 pg/mL (P < .001), and mean
IL-6 concentrations (±SD) were 25795.4 ± 3348.6 pg/mL,
compared with 87.2 ± 66.5 pg/mL (P < .001). These findings
were consistent with those observed in acute pneumonia in
humans [28].
Figure 1. Graphical representation of the combined killing activity of
various antimicrobial combinations against KPVM9. The solid dots
denote observed bacterial burdens seen in combination time-kill studies,
while the red mesh represents expected bacterial burden of the antimicrobial combination. When a dot (grey) is below the red mesh surface,
observed killing is greater than expected killing (ie, synergism). Conversely, when a dot (blue) is above the red mesh surface, observed killing is
less than expected killing (ie, antagonism). For each combination, an interactive index was computed by calculating the ratio of the volume
under the plane (VUP) between the observed and expected surfaces
(ie, VUPobserved/VUPexpected). Synergy and antagonism were defined as
interactive indices of <1 and >1, respectively. As shown above, synergism
Pharmacokinetics
The observed and best-fit pharmacokinetic parameters for
doripenem in infected animals are shown in Figure 2. The
Figure 1 Continued. was observed with doripenem plus amikacin (A; interactive index = 0.60; 95% confidence interval [CI], 0.56–0.65), while the
combination of levofloxacin plus doripenem (B) was viewed to be additive
(interactive index = 0.98; 95% CI, 0.98–1.00). On the other hand, the combination of levofloxacin plus amikacin (C) was antagonistic (interactive
index = 1.11; 95% CI, 1.06–1.16).
KPC Combination Therapy
•
JID 2013:207 (1 March)
•
789
Table 1. Assessment and
combinations against KPVM9
ranking
of
antimicrobial
Antimicrobial combination
Interactive index
(95% CI)
Interpretation
Doripenem plus amikacin
Doripenem plus rifampin
0.60 (.56–.65)
0.99 (.98–1.00)
Synergism
Additivity
Levofloxacin plus rifampin
0.99 (.98–1.00)
Additivity
Doripenem plus levofloxacin
Amikacin plus rifampin
0.98 (.98–1.00)
1.05 (1.03–1.06)
Additivity
Antagonism
Levofloxacin plus amikacin
1.11 (1.06–1.16)
Antagonism
Synergy and antagonism are defined as an interactive index of <1 and >1,
respectively.
Abbreviation: CI, confidence interval.
model fit to the observed data was reasonable (r 2 > 0.95). On
the basis of best-fit model parameter estimates, doripenem was
administered at a dose of 120 mg/kg intraperitoneally 3 times
daily. This dosing regimen was selected to simulate a clinically
relevant pharmacodynamic exposure (a doripenem concentration of ≥1 mg/L 5 hours after the start of an infusion), using a
human dose of 500 mg of doripenem over 1 hour.
Bacterial Lung Clearance
Bacterial burdens in mouse lung tissue observed at 24 hours
after infection with either isolate (n = 8 for baseline; n = 3 per
treatment group) are shown in Figure 3. In addition to the
combination regimens, monotherapy regimens with amikacin
and doripenem were also performed for 24 hours, for comparison. As expected, the bacterial burden increased by at least 1
log CFU/g at 24 hours in the absence of treatment, which was
comparable to doripenem monotherapy against both isolates.
Amikacin monotherapy resulted in increased bacterial burden
against KPVM9 but a slight reduction in burden against
KP6153. Compared with placebo controls, there were statistically significant decreases in bacterial burden at 24 hours for
amikacin plus doripenem for both isolates (P < .01).
Animal Survival
Figure 4 shows the survival of infected animals after various
treatment regimens (n = 10 per treatment group). Bacterial
burdens observed in dead animals were significantly higher
than those observed at baseline, suggesting pneumonia was
likely the primary cause of death (data not shown). Survival
was significantly prolonged for doripenem plus amikacin,
compared with placebo controls (P < .01), for both isolates. As
predicted by the mathematical model, there was an inferior
trend in animal survival seen with the amikacin plus levofloxacin combination (Figure 4A; P = .31, compared with placebo
control).
DISCUSSION
Infections caused by KPC-producing organisms are on the
rise worldwide and have developed into endemic and epidemic situations in multiple countries [3]. As a result of resistance
to multiple classes of antimicrobials, including carbapenems,
the treatment of infections due to KPC-producing organisms
poses a particular challenge to clinicians [29, 30]. Only limited
clinical data are available, and the optimal treatment for such
infections is unknown [11]. Although there are experimental
compounds in development against KPC-producing bacteria,
Figure 2. Single-dose pharmacokinetics of doripenem in infected animals. The open symbols represent experimental observations, while dotted lines
represent the best-fit models. The best-fit model parameter estimates were an elimination rate constant (h−1) of 2.547, a volume of distribution (L/kg) of
0.08, and an absorption rate constant (h−1) of 1.47.
790
•
JID 2013:207 (1 March)
•
Hirsch et al
Figure 3. Change in tissue bacterial burden from baseline after 24
hours of therapy. Data are shown as mean values (±SD). Abbreviation:
CFU, colony-forming units.
they may not be commercially available for some time. For
example, NXL104 and MK-7655 are both novel β-lactamase
inhibitors that have shown in vitro activity in suppressing
class A and C enzymes, including KPC-type carbapenemases
[31–33]. However, until these agents are available for clinical
use, effective treatment options are desperately needed.
Previous attempts at pharmacokinetic/pharmacodynamic
optimization for antimicrobials against KPC-producing isolates have met with limited success [14, 16, 34, 35]. Prolonged
infusion of a high dose of meropenem (one 2-g dose every 8
hours, with each dose administered over 3 hours) was found
to be ineffective against 9 of 11 clinical KPC-producing
K. pneumoniae isolates in an in vitro pharmacodynamic
model [35]. Although this regimen achieved a >3 log CFU/mL
reduction against all isolates by 6 hours, regrowth occurred for
most of the isolates tested. Similarly, the efficacy of doripenem
(human-simulated doses of 1 and 2 g every 8 hours as 4-hour
infusions) were evaluated in a murine thigh infection model
against 7 KPC-producing K. pneumoniae clinical isolates [34].
A decrease of <1 log CFU was observed in infected mice. In
the third approach, doripenem and ertapenem were used in
combination against 1 KPC-producing isolate [14]. In these
experiments, doripenem 2 g every 8 hours (as 3-hour infusions) plus ertapenem 1 g every 24 hours were simulated. In
the in vitro model, there was a modest decrease in bacterial
density, but regrowth was seen at the end of the 24-hour experiment. In the thigh infection model, reduction in the bacterial burden in mice treated with doripenem monotherapy
(mean bacterial load [±SD], 0.47 ± 0.16 CFU/g) was similar to
that observed in untreated immunocompetent mice
(0.57 ± 0.33 log CFU/g). Finally, combinations of tigecycline
with either meropenem or rifampin were evaluated against 5
clinical KPC-2– or KPC-3–producing K. pneumoniae isolates
in an in vitro pharmacodynamic model. Tigecycline (50 mg
Figure 4.
Survival of animals infected with KPVM9 (A) and KP6153 (B).
every 12 hours), meropenem (one 2-g dose every 8 hours,
each administered over 3 hours), and rifampin (600 mg every
12 hours) doses simulated steady-state epithelial lining fluid
concentrations for each agent. Tigecycline plus meropenem
was the most effective combination, but regrowth was seen by
24 hours against the majority of tested isolates.
KPC-producing isolates are particularly resilient against antimicrobial therapy. Consequently, our group sought to screen a
larger selection of 2-agent combinations possessing varied
mechanisms of action. Conventional methods used to study the
in vitro interaction between 2 antimicrobial agents, such as the
fractional inhibitory concentration index and the checkerboard
method, are associated with limitations [36, 37]. In particular,
these studies lack prospective validation and are often not
useful for predicting clinical outcomes [38]. In the past, our
group has performed investigations using a similar mathematical model, followed by in vitro or in vivo validation [22–24].
Our work here is an extension of the previous model used to
identify effective combinations against KPC-producing isolates.
Here, we studied 2 clinical isolates producing the most prevalent variants of blaKPC, KPC-2 and KPC-3 [1]. These isolates
harbor multiple resistance mechanisms, including β-lactamases and outer membrane porin deletions. As predicted by our
quantitative method, doripenem plus amikacin was the most
effective combination against both isolates. At clinically relevant drug concentrations for both amikacin and doripenem, a
considerable synergistic effect was observed after 24-hour
time-kill studies (Figure 1A). When neutropenic, infected
mice were treated with human-like doses of this combination,
KPC Combination Therapy
•
JID 2013:207 (1 March)
•
791
prolonged animal survival also resulted (Figure 4). To our
knowledge, this is one of the first polymyxin-sparing combinations found to have a considerable killing effect against
KPC-producing isolates. Although the doripenem and amikacin MICs of both isolates were considered non-susceptible,
the combination of these two agents resulted in significant
activity. In areas where KPC is endemic, use of similar combinations may be helpful, to avoid overuse of polymyxins in an
effort to retain their activity against non–KPC-producing
multidrug-resistant isolates. Although the mechanistic explanation for the observed synergy between doripenem and
amikacin is not entirely known, we hypothesize that this interaction may be due to KPC’s weak hydrolytic activity against
doripenem [18].
In contrast to the synergistic activity found, the mathematical model identified the combination of amikacin plus levofloxacin to be antagonistic (Figure 1C) against KPVM9.
Animals treated with this combination experienced corresponding inferior survival rates, again validating our mathematical model assessments. This underscores the importance
of avoiding the empirical selection of antimicrobial combinations, particularly for infections involving KPC-producing organisms in which high mortality may already be likely [39].
One limitation of the mathematical model was seen in this
set of experiments. For KP6153, interactive indices were not
calculated. Because significant reductions in bacterial burden
were observed with clinically relevant, high single-drug concentrations, the predicted effect of several combinations was
unrealistic and resulted in negative bacterial burdens. As a
result, the surface response plots were not as helpful in identifying the most promising combinations against KP6153.
Second, although favorable results have been shown in our research laboratory, this model is challenging to apply clinically.
In the future, the possibility of similar testing may only be applicable in a clinical laboratory, using a scaled-down version
that is reasonable to conduct with limited laboratory personnel and/or some level of automation. Furthermore, we chose
only 1 representative agent to study for the entire class of antibiotics having the same mechanistic activity; therefore, it is
not known whether other combinations of a carbapenem plus
aminoglycoside would result in identical results seen with doripenem and amikacin. Last, it is unknown whether the synergistic/antagonistic activities observed against the most
prevalent variants (KPC-2/KPC-3) would be identical against
other KPC variants or against other isolates with alternative
mechanisms of resistance. The future direction of our work
will aim to answer several of these questions.
In closing, our model appears robust at identifying a promising combination for the treatment of multidrug-resistant
KPC-producing isolates. The present study also underscores
the importance of avoiding empirical antimicrobial combinations, given that antagonistic effects may result. Exploration of
792
•
JID 2013:207 (1 March)
•
Hirsch et al
additional combinations and validation against other KPC
variants are warranted.
Notes
Acknowledgments.We thank Drs David Landman and John Quale
(SUNY-Downstate Medical Center, Brooklyn, NY) for the gift of KPVM9
and for their assistance with the molecular characterization of KP6153.
Financial support. This work was supported in part by an unrestricted
grant from Ortho McNeil-Janssen Pharmaceuticals.
Potential conflicts of interest. E. B. H. has received an unrestricted
research grant from Pfizer. V. H. T. has received unrestricted research
grants from AstraZeneca and Merck and is on the speakers bureau of
Merck. All other authors report no potential conflicts.
All 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. Nordmann P, Cuzon G, Naas T. The real threat of Klebsiella pneumoniae carbapenemase-producing bacteria. Lancet Infect Dis 2009;
9:228–36.
2. Srinivasan A, Patel JB. Klebsiella pneumoniae carbapenemase-producing organisms: an ounce of prevention really is worth a pound of
cure. Infect Control Hosp Epidemiol 2008; 29:1107–9.
3. Gupta N, Limbago BM, Patel JB, Kallen AJ. Carbapenem-resistant Enterobacteriaceae: epidemiology and prevention. Clin Infect Dis 2011;
53:60–7.
4. Bratu S, Landman D, Haag R, et al. Rapid spread of carbapenemresistant Klebsiella pneumoniae in New York City: a new threat to our
antibiotic armamentarium. Arch Intern Med 2005; 165:1430–5.
5. Kitchel B, Rasheed JK, Patel JB, et al. Molecular epidemiology of
KPC-producing Klebsiella pneumoniae isolates in the United States:
clonal expansion of multilocus sequence type 258. Antimicrob Agents
Chemother 2009; 53:3365–70.
6. Yigit H, Queenan AM, Anderson GJ, et al. Novel carbapenemhydrolyzing beta-lactamase, KPC-1, from a carbapenem-resistant
strain of Klebsiella pneumoniae. Antimicrob Agents Chemother 2001;
45:1151–61.
7. Endimiani A, Hujer AM, Perez F, et al. Characterization of blaKPCcontaining Klebsiella pneumoniae isolates detected in different institutions in the Eastern USA. J Antimicrob Chemother 2009; 63:427–37.
8. Castanheira M, Sader HS, Deshpande LM, Fritsche TR, Jones RN.
Antimicrobial activities of tigecycline and other broad-spectrum antimicrobials tested against serine carbapenemase- and metallo-betalactamase-producing Enterobacteriaceae: report from the SENTRY
Antimicrobial Surveillance Program. Antimicrob Agents Chemother
2008; 52:570–3.
9. Bratu S, Mooty M, Nichani S, et al. Emergence of KPC-possessing
Klebsiella pneumoniae in Brooklyn, New York: epidemiology and recommendations for detection. Antimicrob Agents Chemother 2005;
49:3018–20.
10. Bratu S, Tolaney P, Karumudi U, et al. Carbapenemase-producing
Klebsiella pneumoniae in Brooklyn, NY: molecular epidemiology and
in vitro activity of polymyxin B and other agents. J Antimicrob Chemother 2005; 56:128–32.
11. Hirsch EB, Tam VH. Detection and treatment options for Klebsiella
pneumoniae carbapenemases (KPCs): an emerging cause of multidrug-resistant infection. J Antimicrob Chemother 2010; 65:1119–25.
12. Woodford N, Tierno PM Jr., Young K, et al. Outbreak of Klebsiella
pneumoniae producing a new carbapenem-hydrolyzing class A betalactamase, KPC-3, in a New York Medical Center. Antimicrob Agents
Chemother 2004; 48:4793–9.
13. Le J, McKee B, Srisupha-Olarn W, Burgess DS. In vitro activity of
carbapenems alone and in combination with amikacin against KPCproducing Klebsiella pneumoniae. J Clin Med Res 2011; 3:106–10.
14. Bulik CC, Nicolau DP. Double-carbapenem therapy for carbapenemase-producing Klebsiella pneumoniae. Antimicrob Agents Chemother 2011; 55:3002–4.
15. Souli M, Galani I, Boukovalas S, et al. In vitro interactions of antimicrobial combinations with fosfomycin against KPC-2-producing Klebsiella pneumoniae and protection of resistance development.
Antimicrob Agents Chemother 2011; 55:2395–7.
16. Wiskirchen DE, Koomanachai P, Nicasio AM, Nicolau DP, Kuti JL. In
vitro pharmacodynamics of simulated pulmonary exposures of tigecycline alone and in combination against Klebsiella pneumoniae isolates
producing a KPC carbapenemase. Antimicrob Agents Chemother
2011; 55:1420–7.
17. Pournaras S, Vrioni G, Neou E, et al. Activity of tigecycline alone and
in combination with colistin and meropenem against Klebsiella pneumoniae carbapenemase (KPC)-producing Enterobacteriaceae strains
by time-kill assay. Int J Antimicrob Agents 2011; 37:244–7.
18. Queenan AM, Shang W, Flamm R, Bush K. Hydrolysis and inhibition
profiles of {beta}-lactamases from molecular classes A to D with doripenem, imipenem and meropenem. Antimicrob Agents Chemother
2010; 54:565–9.
19. Urban C, Mariano N, Rahal JJ. In vitro double and triple bactericidal
activities of doripenem, polymyxin B, and rifampin against multidrugresistant Acinetobacter baumannii, Pseudomonas aeruginosa, Klebsiella
pneumoniae, and Escherichia coli. Antimicrob Agents Chemother
2010; 54:2732–4.
20. Landman D, Bratu S, Quale J. Contribution of OmpK36 to carbapenem susceptibility in KPC-producing Klebsiella pneumoniae. J Med
Microbiol 2009; 58:1303–8.
21. Hirsch EB, Chang KT, Lasco TM, Caeiro JP, Tam VH. Emergence of
KPC-producing Klebsiella pneumoniae in Texas. Diagn Microbiol
Infect Dis 2011; 69:234–5.
22. Lim TP, Ledesma KR, Chang KT, et al. Quantitative assessment of
combination antimicrobial therapy against multidrug-resistant
Acinetobacter baumannii. Antimicrob Agents Chemother 2008; 52:
2898–904.
23. Tam VH, Schilling AN, Lewis RE, Melnick DA, Boucher AN. Novel
approach to characterization of combined pharmacodynamic effects
of antimicrobial agents. Antimicrob Agents Chemother 2004; 48:
4315–21.
24. Yuan Z, Ledesma KR, Singh R, Hou J, Prince RA, Tam VH. Quantitative assessment of combination antimicrobial therapy against multidrug-resistant bacteria in a murine pneumonia model. J Infect Dis
2010; 201:889–97.
25. Tam VH, Ledesma KR, Schilling AN, et al. In vivo dynamics of carbapenem-resistant Pseudomonas aeruginosa selection after suboptimal
dosing. Diagn Microbiol Infect Dis 2009; 64:427–33.
26. Craig WA, Redington J, Ebert SC. Pharmacodynamics of amikacin in
vitro and in mouse thigh and lung infections. J Antimicrob Chemother 1991; 27 (Suppl C):29–40.
27. Andes D, Craig WA. In vivo activities of amoxicillin and amoxicillinclavulanate against Streptococcus pneumoniae: application to breakpoint
determinations. Antimicrob Agents Chemother 1998; 42:2375–9.
28. Monton C, Torres A, El-Ebiary M, Filella X, Xaubet A, de la Bellacasa
JP. Cytokine expression in severe pneumonia: a bronchoalveolar
lavage study. Critical Care Medicine 1999; 27:1745–53.
29. Davies TA, Marie Queenan A, Morrow BJ, et al. Longitudinal survey
of carbapenem resistance and resistance mechanisms in Enterobacteriaceae and non-fermenters from the USA in 2007–09. J Antimicrob
Chemother 2011; 66:2298–307.
30. Landman D, Salamera J, Singh M, Quale J. Accuracy of carbapenem
nonsusceptibility for identification of KPC-possessing Enterobacteriaceae by use of the revised CLSI breakpoints. J Clin Microbiol 2011;
49:3931–3.
31. Endimiani A, Choudhary Y, Bonomo RA. In vitro activity of NXL104
in combination with beta-lactams against Klebsiella pneumoniae isolates producing KPC carbapenemases. Antimicrob Agents Chemother
2009; 53:3599–601.
32. Hirsch EB, Ledesma KR, Chang KT, Schwartz MS, Motyl MR, Tam
VH. In vitro activity of MK-7655, a novel beta-lactamase inhibitor, in
combination with imipenem against carbapenem-resistant Gram-negative bacteria. Antimicrob Agents Chemother 2012; 56:3753–7.
33. Young K, Raghoobar SL, Hairston NN, et al. In vitro activity of the
class A and C β-lactamase inhibitor MK-7655. 50th Interscience Conference on Antimicrobial Agents and Chemotherapy. September 12 - 15,
2010. Boston, MA. Abstract F1-2139.
34. Bulik CC, Nicolau DP. In vivo efficacy of simulated human dosing
regimens of prolonged-infusion doripenem against carbapenemaseproducing Klebsiella pneumoniae. Antimicrob Agents Chemother
2010; 54:4112–5.
35. Bulik CC, Christensen H, Li P, Sutherland CA, Nicolau DP, Kuti JL.
Comparison of the activity of a human simulated, high-dose, prolonged infusion of meropenem against Klebsiella pneumoniae producing the KPC carbapenemase versus that against Pseudomonas
aeruginosa in an in vitro pharmacodynamic model. Antimicrob
Agents Chemother 2010; 54:804–10.
36. King TC, Schlessinger D, Krogstad DJ. The assessment of antimicrobial combinations. Rev Infect Dis 1981; 3:627–33.
37. Rand KH, Houck HJ, Brown P, Bennett D. Reproducibility of the microdilution checkerboard method for antibiotic synergy. Antimicrob
Agents Chemother 1993; 37:613–5.
38. Tan TY, Lim TP, Lee WH, Sasikala S, Hsu LY, Kwa AL. In vitro antibiotic synergy in extensively drug-resistant Acinetobacter baumannii:
the effect of testing by time-kill, checkerboard, and Etest methods.
Antimicrob Agents Chemother 2011; 55:436–8.
39. Gasink LB, Edelstein PH, Lautenbach E, Synnestvedt M, Fishman NO.
Risk factors and clinical impact of Klebsiella pneumoniae carbapenemase-producing K. pneumoniae. Infect Control Hosp Epidemiol 2009;
30:1180–5.
KPC Combination Therapy
•
JID 2013:207 (1 March)
•
793