J Infect Chemother xxx (2018) 1e6

Contents lists available at ScienceDirect

Journal of Infection and Chemotherapy

journal homepage: http://www.elsevier.com/locate/jic

Original Article

Impact of carbapenem versus non-carbapenem treatment on the rates

of superinfection: A meta-analysis of randomized controlled trials*
Khalid Eljaaly a, b, c, *, 1, Mushira A. Enani d, Jaffar A. Al-Tawfiq e, f
a
Department of Clinical Pharmacy, College of Pharmacy, King Abdulaziz University, Jeddah, Saudi Arabia
b
Department of Infectious Diseases, King Abdulaziz University Hospital, Jeddah, Saudi Arabia
c
College of Pharmacy, University of Arizona, Tucson, AZ, USA
d
Medical Specialties Department, Section of Infectious Diseases, King Fahad Medical City, Riyadh, Saudi Arabia
e
Department of Specialty Internal Medicine, Johns Hopkins Aramco Healthcare, Dhahran, Saudi Arabia
f
Department of Medicine, Indiana University School of Medicine, Indianapolis, IN, USA

a r t i c l e i n f o a b s t r a c t

Article history: Imipenem and meropenem are the recommended antipseudomonal carbapenems for nosocomial
Received 22 June 2018 pneumonia per clinical practice guidelines. However, these agents have a relatively broader spectrum of
Received in revised form activity than other antibiotics and need to be reserved. Carbapenems might cause higher rate of su-
29 July 2018
perinfection. The aim of this study was to compare the rate of superinfection between patients who
Accepted 9 August 2018
received imipenem or meropenem versus those who received non-carbapenem treatment. PubMed,
Available online xxx
EMBASE, Cochrane Library databases and two trial registries were searched for relevant randomized
controlled trials of hospitalized adults with pneumonia through February 24, 2017 without date or
Keywords:
Carbapenem
language restrictions. Risk ratios (RRs) with 95% confidence intervals (CIs) were estimated using random-
Imipenem effects models. The primary outcome was based on the intention-to-treat analysis while clinically
Opportunistic evaluable patients were analyzed as secondary outcome. Eight RCTs were included in this meta-analysis.
Hospital infection A statistically higher risk of superinfection with low heterogeneity (RR ¼ 1.690, 95% CI 1.247e2.291,
Antibiotic p ¼ 0.001, I2 ¼ 0%) was associated with the two carbapenems compared to non-carbapenems. However,
in comparison with non-carbapenems, superinfection with imipenem was significantly higher
(RR ¼ 1.694, 95% CI 1.234e2.325, p ¼ 0.001, I2 ¼ 0%), while it was non-significant with meropenem
(RR ¼ 1.647, 95% CI 0.552e4.919, p ¼ 0.371, I2 ¼ 0%). Superinfection was statistically higher in both
double-blind and open-label studies and when carbapenems were compared to other antipseudomonal
beta-lactams. This meta-analysis identified significantly higher superinfection with imipenem compared
to non-carbapenems. The findings confirm the theory of higher superinfections with broader spectrum
agents and provide additional support for reserving carbapenems for the treatment of infections caused
by multidrug-resistant organisms.
© 2018 Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases.
Published by Elsevier Ltd. All rights reserved.

1. Introduction recommended monitoring evidence of adverse events related to

The 2016 Infectious Diseases Society of America (IDSA) and the
Society for Healthcare Epidemiology of America (SHEA) guidelines
for implementing antibiotic stewardship program (ASP)
*
All authors meet the ICMJE authorship criteria.
* Corresponding author. 1-College of Pharmacy, University of Arizona, Drachman
Hall e B306, 1295 N Martin Ave, P.O.Box 210202, Tucson, AZ, USA.
E-mail address: keljaaly@kau.edu.sa (K. Eljaaly).
1
Department of Clinical Pharmacy, King Abdulaziz University P.O. Box 80200,
Jeddah, postal code 21441, Saudi Arabia ext: 20675.
antibiotics and considering these factors in antibiotic selection
[1,2]. Moreover, the guidelines recommended specific ASP in-
terventions designed to decrease the use of antibiotics that are
associated with a high risk of Clostridium difficile infection (CDI).
The development of CDI is one type of superinfection but the
guidelines have not addressed other superinfections likely due to
limited data available on antibiotic-induced superinfections.
Broad-spectrum antibiotic therapy is a risk factor for increased
rates of superinfection [3]. In addition, longer antibiotic therapy
duration for ventilator-associated pneumonia (VAP) is associated
with higher rates of susceptible and multidrug-resistant

https://doi.org/10.1016/j.jiac.2018.08.004
1341-321X/© 2018 Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Eljaaly K, et al., Impact of carbapenem versus non-carbapenem treatment on the rates of superinfection: A
meta-analysis of randomized controlled trials, J Infect Chemother (2018), https://doi.org/10.1016/j.jiac.2018.08.004
2 K. Eljaaly et al. / J Infect Chemother xxx (2018) 1e6
superinfection [4]. Imipenem and meropenem are the recom-
mended antipseudomonal carbapenems for nosocomial pneu-
monia per the current clinical practice guidelines but they need to
be reserved [4]. These agents might cause higher rate of superin-
fection because of having a relatively broader spectrum of activity
than other antibiotics. Randomized controlled trials (RCTs) reduce
selection bias but unfortunately do not consistently report rates of
superinfection. It is very unlikely that one study would be
adequately powered to detect a statistically significant difference of
superinfection rate. Therefore, the objective of this meta-analysis is
to compare the rate of superinfection between patients who
received imipenem or meropenem versus those who received non-
carbapenem antibiotics for treatment of pneumonia.
2. Materials and methods
This meta-analysis was conducted according to the Preferred
Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) guidelines (appendix. eTable 2). This study is registered
with PROSPERO, number CRD42017058062.
2.1. Search strategy and selection criteria
The full search strategy is provided in the appendix. Two authors
(K.E. and M.E.) independently extracted data after searching
PubMed, Embase, and Cochrane Library databases without date or
language restrictions through February 25, 2017. The reference lists
of included articles were hand searched for additional trials. In
addition, the ClinicalTrials.gov, and ClinicalTrialsRegister.eu web-
sites were searched for unpublished trials through February 25, 2017.
RCTs of hospitalized adult patients with pneumonia that
compared imipenem or meropenem to non-carbapenems were
included if they reported rates of superinfection using these two
regimens. Superinfection was defined as isolation of a new path-
ogen, after starting study antibiotic therapy, and at least one of the
following, to reduce the likelihood of colonization: symptoms and
signs of infection and requiring treatment. We excluded studies that
provided no or different definition for superinfection. Any
disagreement between the authors was resolved through discussion.
2.2. Outcomes
To be conservative and reduce selection bias, the primary outcome
was the rate of superinfection based on the intention-to-treat (ITT)
analysis. If the study only reported results from clinically evaluable
(CE) patients, the ITT analysis was used to recalculate the results if they
were not provided as ITT. Secondary outcome was the superinfection
rate among CE patients. Then, subgroup analyses were performed for
each carbapenem and pathogen according to the ITT principle and of
blind versus unblind studies that compared any carbapenem with
non-carbapenems. To further examine the robustness of the indicated
primary outcome, we analyzed superinfection rates of carbapenems
in comparison to other antipseudomonal beta-lactams.
2.3. Risk of bias assessment and data analysis
Heterogeneity (I [2]) was assessed by using Cochran's chi-
squared test. The risk ratios (RRs) with 95% confidence intervals
(CIs) were estimated using random-effects models. We assessed the
quality of studies by using the Cochrane risk of bias tool for RCTs
(low, unclear or high) [5]. Funnel plot was used to visually evaluate
publication bias. The Egger test was used to test for funnel plot
asymmetry. All analyses were conducted using Comprehensive
Meta- Analysis Version 3 software (Biostat, Englewood, NJ, USA).
Please cite this article in press as: Eljaaly K, et al., Impact of carbapenem versus non-carbapenem treatment on the rates of superinfection: A
meta-analysis of randomized controlled trials, J Infect Chemother (2018), https://doi.org/10.1016/j.jiac.2018.08.004
3. Results
The search process identified 431 articles (PubMed 122; Embase
142; Cochrane Library 167) [Fig. 1]. After removal of duplicates, a
total of 253 articles were screened. Twenty-four full-text articles
were assessed for eligibility after screening by title and/or abstract
and eight RCTs (with 1874 patients) were included [6e13]. A total of
814 ITT-patients treated with carbapenems (imipenem, 641; mer-
openem, 173) were compared to 855 patients treated with non-
carbapenems [6e13]. On the other hand, a total of 427 CE-
patients treated with carbapenems were compared to 421 CE-
patients treated with non-carbapenems [6e8,12,13].
3.1. Study characteristics
The characteristics of the eight included studies are summarized
in Table 1 [6e13], while the risk of bias assessment table is provided
in the appendix. eTable 3. They were conducted between 1990 and
2011 and all were in English language. All studies were multicentre
and one study was multicontinental [10]; seven studies were su-
periority studies in design [6e9,11e13], and one was a non-
inferiority study [10]; three were double-blind [7,8,10] while five
were open-label [6,9,11e13]. All the studies were funded by drug
manufacturers: the six imipenem studies were funded by the
comparative drug manufacturers [7e11,13], while the two mer-
openem studies [6,12] were funded by the meropenem manufac-
turer. The mean subject age in all studies was 65 years. All studies
included patients with nosocomial pneumonia, except one study
included patients with community-acquired pneumonia [6] who
represented the minority of study population. Two studies included
only patients in intensive-care unit [6,13] and two studies included
only patients with VAP [6,11]. Carbapenems were compared to
beta-lactams in five studies [6,8,10,12,13], to fluoroquinolones in
two studies [7,11], and to tigecycline in one study [10]. The average
duration of antibiotic therapy did not exceed 10 days. The assess-
ments of bias risk are summarized in the appendix. eFigure 3.
3.2. Study outcomes
Based on ITT-analysis, the mean of superinfection was 11.79%
(range, 2.88e30.51%) in the carbapenem group vs. 6.67% (range,
0e17.46%) in the non-carbapenem group. A statistically higher risk
of superinfection (RR ¼ 1.690, 95% CI 1.247e2.291, p < 0.001,
I2 ¼ 0%; Q ¼ 2.895) was associated with the two carbapenems
compared to non-carbapenems [Fig. 2]. The asymmetry in the
funnel plot [appendix. eFigure 2] was not statistically significant
(p ¼ 0.273; Egger test).
In comparison with non-carbapenems, subgroup analysis
showed that superinfection with imipenem was significantly
higher (RR ¼ 1.694 [95% CI 1.234e2.325]; p < 0.001; I2 ¼ 0%;
Q ¼ 2.885), while it was non-significant with meropenem
(RR ¼ 1.647 [95% CI 0.552e4.919]; p ¼ 0.371; I2 ¼ 0%; Q ¼ 0.008)
[Fig. 2]. In addition, superinfection was statistically higher with
carbapenems versus non-carbapenems in both double-blind
(RR ¼ 1.627 [95% CI 1.105e2.397]; p ¼ 0.014; I2 ¼ 0%; Q ¼ 1.231)
and open-label studies (RR ¼ 1.796 [95% CI 1.100e2.934]; p ¼ 0.019;
I2 ¼ 0%; Q ¼ 1.569). Sensitivity analysis identified higher risk of
superinfection when carbapenem was compared to other anti-
pseudomonal beta-lactams (RR ¼ 1.739 [95% CI 1.107e2.732],
p ¼ 0.016; I2 ¼ 0%; Q ¼ 0.775) [appendix. eFigure 2].
For CE-patients, 13.11% (range, 2.97e28.04%) in the carbapenem
group versus 7.84% (range, 2.78e18.37%) in the non-carbapenem
group had superinfection. The difference was statistically signifi-
cant (RR ¼ 1.610 [95% CI 1.084e2.392]; p ¼ 0.018; I2 ¼ 0%;
Q ¼ 0.740) [Fig. 3].
 K. Eljaaly et al. / J Infect Chemother xxx (2018) 1e6 3

Fig. 1. Flowchart of the process of literature search and extraction of studies meeting the inclusion criteria.

Only three studies reported the organisms causing superinfec-
tion (5,6,10) [appendix. eFigure 1]. Pseudomonas aeruginosa caused
a statistically higher superinfection in the carbapenem group
versus the fluoroquinolones group (RR ¼ 3.638 [95% CI
1.382e9.580]; p ¼ 0.047; I2 ¼ 0%; Q ¼ 0.009). There was no sig-
nificant difference with other pathogens. However, it was not re-
ported in any of included studies whether the bacteria causing
superinfection were susceptible or resistant to the study antibiotic.
4. Discussion
The emergence and development of superinfection during
therapy with broad spectrum antibiotics are major concerns for
the use of these agents [3]. Antibiotics alter the normal protective
microflora and its ecological balance in the host, resulting in
overgrowth of opportunistic pathogens and superinfections [14].
We attempted to shed light on the development of superinfection
when using carbapenems and non-carbapenems therapy in
pneumonia patients based on data from RCTs. We specifically
examined the rate of superinfection in patients who received
imipenem or meropenem. We did not examine the mortality,
morbidity or other outcomes of the included patients. Carbape-
nems increased risk of superinfection by approximately 70%
compared to non-carbapenems. To our knowledge, this is the first
meta-analysis of RCTs showing higher risk of superinfection with
carbapenems, especially imipenem. This was found in ITT and CE
patients, irrespective of whether studies were blind or not, and
when carbapenem was compared to other antipseudomonal beta-
lactams. The results were consistent among studies with low
heterogeneity.
Risk of superinfection is related to the duration of therapy and
previously it was stated that use of carbapenems for 3 days does not
lead to superinfection [15]. Longer duration of therapy and
immunosuppression may lead to superinfection. In this meta-
analysis, the average duration of therapy was less than 10 days;
however, longer duration of therapy is frequently used and
immunosuppression are not rare in clinical practice and might
cause even more superinfection episodes. The rate of superinfec-
tion is variable dependent on the study and the antibiotics used. In
a meta-analysis of RCTs, superinfection was lower in the beta lac-
tam monotherapy group than the beta lactam-aminoglycoside
combination therapy (odds ratio, 0.62; 95% CI, 0.42e0.93) [16].
In sensitivity analysis, beta-lactams were the most common
comparators allowing higher statistical power to test differences
unlike other comparators. Thus, it is a common clinical question
whether using carbapenem cause more superinfection than other
beta-lactam alternatives. In the current analysis, sensitivity anal-
ysis identified higher risk of superinfection when carbapenem was

Please cite this article in press as: Eljaaly K, et al., Impact of carbapenem versus non-carbapenem treatment on the rates of superinfection: A
meta-analysis of randomized controlled trials, J Infect Chemother (2018), https://doi.org/10.1016/j.jiac.2018.08.004
 4
meta-analysis of randomized controlled trials, J Infect Chemother (2018), https://doi.org/10.1016/j.jiac.2018.08.004
Please cite this article in press as: Eljaaly K, et al., Impact of carbapenem versus non-carbapenem treatment on the rates of superinfection: A

Table 1
Characteristics of Included Studies. Results are presented as carbapenem compared to non-carbapenem. Abbreviations: NA: not available or not applicable; RCT, randomized clinical trial; ITT, intention to treat; CE, clinically
evaluable; VAP, ventilator-associated pneumonia; PNA, pneumonia; MV, mechanical ventilation; PO, orally; IV, intravenously.

Study Study Design Location Funding source Enrolled patients; Mean age Patient Carbapenem vs Additional antibiotics Duration of
period (manufactured drug) ITT; CE (years) characteristics non-carbapenem therapy allowed therapy (days)

Alvarez Lerma NA Superiority, open- 14 sites in Spain Industry 140; 62 VAP in ICU; IV meropenem 1 g q8h vs. NA 9.3 vs. 8.3
et al., 2001 label, RCT (meropenem) 69 vs. 71; APACHE II score: IV ceftazidime 2 g
57 vs. 59 16.5 vs. 16.6 q8h þ amikacin 7.5 mg/
kg q12h
Fink et al., 1994 1990e1992 Superiority, 20 sites in U.S. Industry 405; 60 Most had IV imipenem/cilastatin metronidazole (with NA
double-blind, RCT (ciprofloxacin) 200 vs. 202; nosocomial PNA, 1000 mg q8h or 500 mg ciprofloxacin), IV/PO
98 vs 107 were in ICU, and q6h for highly susceptible vancomycin,
required MV. pathogens vs. IV antifungals, antivirals
APACHE II score: ciprofloxacin 400 mg q8h
17.6 vs. 17.7 or 400 mg q12h for highly

K. Eljaaly et al. / J Infect Chemother xxx (2018) 1e6

susceptible pathogens
Joshi et al., 2006 1997e2001 Superiority, 66 sites in U.S. Industry (piperacillin/ 449; 53 Nosocomial PNA; IV imipenem/cilastatin NA 9.7 vs. 9.4
double-blind, RCT and Canada tazobactam) 215 vs. 222; 99 vs. APACHE II score: 13 500 mg
98 vs. 13.9 q6h þ tobramycin 5 mg/
kg/day (or amikacin) vs.
IV piperacillin-
tazobactam 4.5 g
q6h þ tobramycin 5 mg/
kg/day
Polk et al., 1997 1991e1993 Superiority, open- 6 sites in U.S. Industry 122; NA Trauma patients IV imipenem/cilastatin NA NA
label, RCT (aztreonam) 63 vs. 59; with nosocomial 500 mg q6h vs. IV
NA PNA who required aztreonam 2 g q8h þ IV
MV vancomycin 1 g q12h
Ramirez 2008e2011 Non-inferiority, 75 sites in Industry 114; 65 vs. 61 HAP/VAP; IV imipenem/cilastatin IV ceftazidime (with 8.6 vs. 8.2
et al., 2013 double-blind, RCT Europe, Asia, (tigecycline) 34 vs. 71; APACHE II score: 1 g q8h vs. IV tigecycline tigecycline), IV
Latin America, 24 vs. 43 13.8 150 mg, followed by vancomycin (with
U.S., Canada 75 mg q12h or 200 mg, imipenem), IV
and Australia followed by 100 mg q12h tobramycin or amikacin
Shorr et al., 2005 1997e2001 Superiority, open- 67 sites in U.S. Industry 222; 53 VAP; IV imipenem/cilastatin IV aminoglycoside NA
label, RCT and Canada (levofloxacin) 111 vs. 111; APACHE II score: 500e1000 mg q6-8h, (with imipenem), IV
NA 15.1 vs. 14.8 followed by PO non-carbapenem beta-
ciprofloxacin 750 mg lactam (with
q12h vs. IV levofloxacin levofloxacin), IV
750 mg q24h þ IV vancomycin
vancomycin 1 g q12h
Sieger et al., 1997 NA Superiority, open- 22 sites in U.S. Industry 211; 55 vs. 54 Nosocomial PNA; IV meropenem 1 g q8h vs. NA 7.8 vs. 7.4
label, RCT (meropenem) 104 vs. 107; most required MV IV ceftazidime 2 g
63 vs. 58 q8h þ tobramycin 1 mg/
kg q8h
Zanetti et al., 2003 1997e1999 Superiority, open- 13 sites in 6 Industry 281; 54 Nosocomial PNA in IV imipenem/cilastatin NA 9.4 vs. 9.1
label, RCT European (cefepime) 138 vs. 132; ICU; 66% with VAP; 500 mg q6h vs. IV
countries 101 vs. 108 APACHE II score: cefepime 2 g q8h
14.8 vs. 15.4
meta-analysis of randomized controlled trials, J Infect Chemother (2018), https://doi.org/10.1016/j.jiac.2018.08.004
Please cite this article in press as: Eljaaly K, et al., Impact of carbapenem versus non-carbapenem treatment on the rates of superinfection: A

K. Eljaaly et al. / J Infect Chemother xxx (2018) 1e6

Fig. 2. Forest plot showing the risk ratios of superinfection based on intent to treat analysis for patients receiving carbapenem versus non-carbapenem, including subgroup analysis according to type of carbapenem. Vertical line, "no
difference" point between the 2 groups; horizontal line, 95% confidence interval; squares, risk ratios; diamonds, pooled risk ratios.

Fig. 3. Forest plot showing the risk ratios of superinfection based on analysis of clinically-evaluable patients receiving carbapenem versus non-carbapenem. Vertical line, "no difference" point between the 2 groups; horizontal line, 95%
confidence interval; squares, risk ratios; diamonds, pooled risk ratios.

5
6 K. Eljaaly et al. / J Infect Chemother xxx (2018) 1e6
compared to other antipseudomonal beta-lactams (RR ¼ 1.739).
Thus, the result supports the use of other antipseudomonal beta-
lactams and reserve carbapenem for cases when it is really
needed. One limitation of this meta-analysis is that superinfection
was not characterized in all included studies and thus superin-
fection may refer to fungal infection or the development of drug-
resistant bacteria. Only three studies reported the organisms
causing superinfection [6,7,11]. In these studies, P. aeruginosa
caused a statistically higher superinfection in the carbapenem
group versus the fluoroquinolones group (RR ¼ 3.638). Since the
susceptibility results were not included, it was not possible to
tease out if the organism was already resistant to the included
antibiotic or there was already high rate of resistance in the cen-
ters to the used antibiotic. This is particularly important since
fluoroquinolone resistance is a major concern in many countries.
In a systematic review, most frequent superinfection was caused
by Candida spp. (42.3%), Enterococcus spp. (18.8%), enterobacteria
(13.8%), Staphylococcus spp. (9.5%), and P. aeruginosa (6.6%).16 The
most frequent implicated antibiotics were: ciprofloxacin (38.1%),
cefotaxime (23.3%), imipenem (12%), meropenem (10.2%), and
cefepime (6.1%) [17]. Systematic collection of data regarding su-
perinfection is needed based on specific definitions. Superinfec-
tion had been defined as growth of the specific antibiotic resistant
organism from any clinical specimen on days 4e30 after the start
of that antibiotic or a positive C. difficile in stool [18]. Other meta-
analyses defined superinfection as the isolation of a pathogen
responsible for a subsequent infection and of a species different
from the initially isolated pathogen [3]. In this meta-analysis, the
definition of superinfection was a clinical definition that took into
account the development of symptoms and/or signs of infection
and the need for treatment. This was done to reduce the possibility
of colonization, which is another dimension of antibiotic use.
Including the definition of superinfection as having new pathogen
during therapy with or without symptoms or the requirement of
antibiotic therapy will encompass a larger number of studies but
will include those who had colonization.
The small number of the included studies might limit the
generalization of the finding in this meta-analysis. This is partic-
ularly true in the case of meropenem as only few studies were
included. Based on the superinfection rates of 4.624% and 2.809%
in meropenem and non-carbapenem group, respectively, sample
sizes of 173 in the meropenem group and 178 in the non-
carbapenem group have only a 14.5% power to detect a statisti-
cally significant difference with a significance level (alpha) of 0.05
(two-tailed). Moreover, the pooled superinfection rate was not
precise because the confidence interval was wide. Another draw-
back of the studies is the fact that the imipenem studies were
funded by the comparative drug manufacturers [7e11,13], and the
two meropenem studies [6,12] were funded by the meropenem
manufacturer. Thus, the presence of bias in reporting could not be
entirely excluded. Another limitation of this meta-analysis is re-
striction to studies of pneumonia patients. However, including any
infection might have increased the heterogeneity between the
studies and despite this restriction, adequate statistical power
allowed finding significantly higher superinfection rates with
carbapenems.
In conclusion, a meta-analysis of pneumonia data of RCTs showed
significantly higher superinfection with imipenem compared to non-
carbapenems. Larger sample size is likely needed to determine if the
same results apply to meropenem. This additional adverse outcome
of carbapenem use provides an added evidence to support reserving
these valuable agents for the treatment of pneumonia caused by
multidrug-resistant organisms. Antibiotic stewardship program
should seek to reduce unnecessary use of carbapenems.
Please cite this article in press as: Eljaaly K, et al., Impact of carbapenem versus non-carbapenem treatment on the rates of superinfection: A
meta-analysis of randomized controlled trials, J Infect Chemother (2018), https://doi.org/10.1016/j.jiac.2018.08.004
Conflicts of interest
None.
Acknowledgements
None.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.jiac.2018.08.004.
References
[1] Barlam TF, Cosgrove SE, Abbo LM, MacDougall C, Schuetz AN, Schuetz AN,
Septimus EJ, et al. Implementing an antibiotic stewardship program: guidelines
by the infectious Diseases society of America and the society for Healthcare
Epidemiology of America. Clin Infect Dis 2016;62:e51e77.
[2] Eljaaly K, Elarabi S, Alshehri S, Nix DE. Impact of requiring re-authorization
of restricted antibiotics on day three of therapy. J Antimicrob Chemother
2018;73:527e30.
[3] Bliziotis IA, Samonis G, Vardakas KZ, Chrysanthopoulou S, Falagas ME. Effect of
aminoglycoside and beta-lactam combination therapy versus beta-lactam
monotherapy on the emergence of antimicrobial resistance: a meta-analysis
of randomized, controlled trials. Clin Infect Dis 2005;41(2):149e58.
[4] Kalil AC, Metersky ML, Klompas M, Muscedere J, Sweeney DA, Palmer LB, et al.
Management of adults with hospital-acquired and ventilator-associated pneu-
monia: 2016 clinical practice guidelines by the infectious Diseases society of
America and the American thoracic society. Clin Infect Dis 2016;63(5):e61e111.
[5] Higgins J, Altman D, Sterne J. Chapter 8: assessing risk of bias in included
studies. In: Higgins JPT, Green S, editors. Cochrane handbook for systematic
reviews of interventions version 5.1.0. The Cochrane Collaboration; 2011.
Available from: http://www.cochrane-handbook.org.
[6] Lerma FA, Group OB of TSI. Efficacy of meropenem as monotherapy in the
treatment of ventilator-associated pneumonia. J Chemother 2001;13(1):70e81.
[7] Fink MP, Snydman DR, Niederman MS, Leeper Jr KV, Johnson RH, Heard SO,
et al. Treatment of severe pneumonia in hospitalized patients: results of a
multicenter, randomized, double-blind trial comparing intravenous ciproflox-
acin with imipenem-cilastatin. The Severe Pneumonia Study Group. Anti-
microb Agents Chemother 1994;38(3):547e57.
[8] Joshi M, Metzler M, McCarthy M, Olvey S, Kassira W, Cooper A. Comparison of
piperacillin/tazobactam and imipenem/cilastatin, both in combination with
tobramycin, administered every 6 h for treatment of nosocomial pneumonia.
Respir Med 2006;100(9):1554e65.
[9] Polk HC, Livingston DH, Fry DE, Malangoni MA, Fabian T, Trachtenberg LS,
et al. Treatment of pneumonia in mechanically ventilated trauma patients.
Results of a prospective trial. Arch Surg 1997;132(10):1086e92.
[10] Ramirez J, Dartois N, Gandjini H, Yan JL, Korth-Bradley J, McGovern PC. Ran-
domized phase 2 trial to evaluate the clinical efficacy of two high-dosage
tigecycline regimens versus imipenem-cilastatin for treatment of hospital-
acquired pneumonia. Antimicrob Agents Chemother 2013;57(4):1756e62.
[11] Shorr AF, Zadeikis N, Jackson WL, Ramage AS, Wu SC, Tennenberg AM, et al.
Levofloxacin for treatment of ventilator-associated pneumonia: a subgroup
analysis from a randomized trial. Clin Infect Dis 2005;40(Suppl 2):S123e9.
[12] Sieger B, Berman SJ, Geckler RW, Farkas SA. Empiric treatment of hospital-
acquired lower respiratory tract infections with meropenem or ceftazidime
with tobramycin: a randomized study. Meropenem Lower Respiratory Infec-
tion Group. Crit Care Med 1997;25(10):1663e70.
[13] Zanetti G, Bally F, Greub G, Garbino J, Kinge T, Lew D, 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(11):3442e7.
[14] Sullivan A, Edlund C, Nord CE. Effect of antimicrobial agents on the ecological
balance of human microflora. Lancet Infect Dis 2001;1(2):101e14.
[15] Masterton RG. Carbapenems in serious infections: maximizing the power,
minimizing resistance. Clin Microbiol Infect 1999;5(s3):12e4.
[16] Paul M, Benuri-Silbiger I, Soares-Weiser K, Leibovici L. Beta lactam mono-
therapy versus beta lactam-aminoglycoside combination therapy for sepsis in
immunocompetent patients: systematic review and meta-analysis of rando-
mised trials. BMJ 2004;328(7441):668.
[17] Alvarez C, Ramos JM, San Juan R, Lumbreras C, Aguado JM. Risk of superinfec-
tion related to antibiotic use. Are all antibiotics the same? Rev Esp Quimioter
2005;18(1):39e44.
[18] Harris PNA, Peleg AY, Iredell J, Ingram PR, Miyakis S, Stewardson AJ, et al.
Meropenem versus piperacillin-tazobactam for definitive treatment of
bloodstream infections due to ceftriaxone non-susceptible Escherichia coli
and Klebsiella spp (the MERINO trial): study protocol for a randomised
controlled trial. Trials 2015;16(1):24.