Article Text

Original research
Efficacy and safety of ceftazidime–avibactam versus polymyxins in the treatment of carbapenem-resistant Enterobacteriaceae infection: a systematic review and meta-analysis
  1. Ping Yang1,2,
  2. Yinyan Li1,
  3. Xiaojuan Wang1,
  4. Na Chen1,
  5. Xiaoyang Lu1
  1. 1Department of Clinical Pharmacy, First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, Zhejiang, China
  2. 2Zhejiang Provincial Key Laboratory for Drug Evaluation and Clinical Research, Hangzhou, People's Republic of China
  1. Correspondence to Dr Xiaoyang Lu; 183937719{at}


Objectives Carbapenem-resistant Enterobacteriaceae is increasingly recognised as a significant public health concern. Ceftazidime-avibactam (CAZ-AVI) and polymyxins are considered as the last therapeutic options worldwide. This is the first meta-analysis of recently published data to compare the clinical efficacy and safety of CAZ-AVI with polymyxins in the treatment of carbapenem-resistant Enterobacteriaceae infections.

Design Systematic review and meta-analysis.

Data sources PubMed, Embase and the Cochrane Library were systematically searched, for publications in any language, from database inception to February 2023.

Eligibility criteria for selecting studies Studies comparing the clinical efficacy and safety of CAZ-AVI with polymyxins were included. Mortality, clinical success, microbiological eradication and nephrotoxicity were assessed as the main outcomes.

Data extraction and synthesis Literature screening, data extraction and the quality evaluation of studies were conducted by two researchers independently, with disagreements resolved by another researcher. The Newcastle–Ottawa Scale was used to assess the bias risk for the included studies. Review Manager V.5.3 was employed for the meta-analysis.

Results The meta-analysis included seven retrospective and four prospective cohort studies with 1111 patients enrolled. The CAZ-AVI groups demonstrated a lower 30-day mortality (risk ratio (RR)=0.48, 95% CI of 0.37 to 0.63, I2=10%, p<0.0001) in nine studies with 766 patients; higher clinical success (RR=1.71, 95% CI 1.33 to 2.20, I2=35%, p<0.0001) in four studies with 463 patients; and lower nephrotoxicity in seven studies with 696 patients (RR=0.42, 95% CI 0.23 to 0.77, I2=35%, p<0.05). However, no significant difference in microbiological eradication rates was observed in 249 patients from two studies (RR=1.16, 95% CI 0.97 to 1.39, I2=0, p>0.05).

Conclusion Available evidence suggested that CAZ-AVI treatment held a dominant position with respect to efficacy and safety compared with polymyxins in carbapenem-resistant Enterobacteriaceae infections. However, the analysis included only observational studies, and high-quality, large-scale, multicentre, double-blind randomised controlled trials are needed to confirm the advantage of CAZ-AVI.

  • GENERAL MEDICINE (see Internal Medicine)
  • Infection control

Data availability statement

No data are available.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Rigorous reviewing methods were used in this systematic review and meta-analysis, including a comprehensive search strategy, explicit eligibility criteria and the selection of studies by two independent reviewers.

  • No restrictions were implemented on the type of articles, study design, language or publication year, and the Newcastle–Ottawa Scale was used to evaluate the quality of the studies.

  • Subgroup analyses of diseases and strains were not performed due to the lack of some necessary information.

  • The included studies were all observational studies with limited sample sizes; high-quality and large-scale multicentre randomised controlled trials are needed to confirm our findings.


Globally, bacterial resistance is becoming an increasingly serious problem owing to the use of antibacterial drugs. Carbapenem-resistant Enterobacteriaceae (CRE) infections have emerged throughout the world, posing a global public health threat and a formidable challenge to antimicrobial therapy.1 Nonetheless, increased carbapenem use could lead to carbapenem resistance in gram-negative bacteria, which is the chief reason of antimicrobial resistance.2 CRE infections have presented a particularly grave threat worldwide. In the past decades, tigecycline has been considered as one of the last lines of defence against severe CRE infections. However, suboptimal concentrations of tigecycline have been found in both serum and pulmonary epithelial lining fluid, and this observation has prompted many physicians to use either combination therapy or high-dose tigecycline to treat CRE infections. Severe coagulopathy with hypofibrinogenaemia and diarrhoea have recently been reported to be associated with high-dose tigecycline.3 4 Therefore, the Food and Drug Administration (FDA) has warned against the off-label use of tigecycline to treat nosocomial pneumonia because of the increased mortality risk indicated in randomised trials.5 6 Recently, some new antibiotics also show great advantages in the treatment of CRE infection, such as cefiderocol, meropenem/vaborbactam and imipenem/cilastatin/relebactam.7–9 However, given the accessibility of drugs, the optimal treatment for CRE infection in China still involves either ceftazidime–avibactam (CAZ-AVI) or polymyxins-based regimens.

CAZ-AVI was approved by the European Medicines Agency for infections caused by gram-negative aerobic bacteria in adults with limited treatment options, including Klebsiella pneumoniae, Escherichia coli and Pseudomonas aeruginosa.10 Recently, polymyxins were reintroduced into medical practice as one of the last resorts for treating extensively drug-resistant gram-negative bacteria. However, acute kidney injury was frequently experienced after conventional doses of polymyxins.11

Recent meta-analysis demonstrated that CAZ-AVI had a favourable pharmacological profile and might be an option for empirical therapy of CRE infection.12 However, large-scale clinical trials comparing the safety and efficacy of CAZ-AVI with polymyxins in the treatment of CRE infection are still lacking.

Naturally, systematically assessing the results of the previous researches in multiple centres could provide important information. Therefore, we conducted a meta-analysis to further explore the efficacy and safety profile of CAZ-AVI relative to polymyxins in the treatment of CRE infection.


Literature search

The PubMed, Embase and Cochrane Central Register of Controlled Trials electronic databases were independently searched by two authors, starting from the database inception to February 2023. The full search strategy aimed to include any clinical studies performed on patients with CRE infection treated with CAZ-AVI versus polymyxins. The PubMed search strategy was (‘ceftazidime-avibactam’) AND (‘polymyxin’ OR ‘polymyxins’ OR ‘colistin’) AND (‘carbapenem-resistant klebsiella pneumoniae’ OR ‘carbapenem resistant Enterobacter*’ OR ‘carbapenem resistant gram-negative bacteria’ OR ‘carbapenem resistant organism’ OR ‘multidrug-resistant gram-negative bacteria’) as applied on both the medical subject heading and free text. This search strategy was subsequently modified for searching in Embase and the Cochrane Central Register of Controlled Trials (online supplemental file 1). Previously published systematic reviews were also checked to identify any additional studies that might have been overlooked in our search strategy.

Study selection

Two authors independently screened the eligibility of the literature by examining the titles, abstracts and full-text of the retrieved articles. Eligible studies included all available published studies that compared CAZ-AVI with polymyxins for the treatment of CRE infections in adult patients (>18 years). CAZ-AVI or polymyxins-based combined administration schemes were allowed. Studies lacking quantitative or qualitative target outcome results were excluded. Disagreements between reviewers were settled by another researcher.

Data extraction

Data extraction was conducted by two authors independently. Each study was reviewed for the following information: (1) study author, publication year and the study regions; (2) study design and sample size; (3) patients’ characteristics (age, sex, drug administration regimen, infection site and causative pathogen); and (4) outcomes, such as mortality, clinical efficacy, bacterial eradication and adverse reactions.

Quality assessment

The risk of bias of the included studies in terms of patient selection, comparability between groups, outcome and exposure factors assessment was evaluated by the Newcastle–Ottawa Scale (NOS) or a modified NOS.13 NOS scores ranged from 0 to 9, and the studies were then classified according to quality as poor (0–4), moderate (5–6) or high-quality (7–9) research. A consensus was reached between reviewers to resolve differences.

Statistical analysis

Review Manager V.5.3 was used to conduct the statistical analysis. Based on a random-effects model, dichotomous outcomes were represented as a risk ratio (RR) with a 95% CI. The Cochrane I2 statistic was used to measure heterogeneity. Note that heterogeneity was significant if the I2 value exceeded 50%. The model was considered robust if no significant difference was observed in the p value of the corresponding combined effect size. Moreover, publication biases were visually assessed using funnel plots.

Patient and public involvement



Description of included studies

The retrieval strategy initially retrieved 580 articles. From those, 482 potentially useful studies were obtained after removing duplicates. The full texts of the remaining articles were evaluated for eligibility based on the inclusion and exclusion criteria. Finally, 11 cohort studies were included in the meta-analysis.14–24 The detailed process was displayed in figure 1. table 1, 1a summarised the details of the included studies, including study year, author, region, design and participant information .

Table 1

Characteristics of the included studies

Figure 1

Flow chart of literature search process and review based on eligible criteria.

The included studies involved 467 patients for the CAZ-AVI group and 644 patients for polymyxins group. All study designs were available, and the eligible articles consisted of seven retrospective15 16 18–22 and four prospective cohort studies.14 17 23 24 One of them is a single-centre study,19 while the other 10 are multicentre studies.14–18 20–24 Three studies originated from the USA,14 19 23 two studies were conducted in Saudi Arabia,20 22 four studies were performed in China,15 16 18 24 and two studies were conducted in Italy and Greece, respectively.17 21 Among the 11 studies, 6 articles reported the patient age, with the mean age of 57–66 years in the CAV-AVI group and 49–67.5 years in the polymyxins group.14 15 18–20 22 Further, six articles reported participant sex, with 31.7%–67.6% male patients. Based on our inclusion criteria, monotherapy or combination therapy was allowed. CAZ-AVI was provided as monotherapy in two studies19 23 and polymyxins were provided as monotherapy in one study.23 Only one study was conducted according to the combination of the above two drugs,15 and all others were based on monotherapy or combination regimens (online supplemental table 1). The most common infection site in the enrolling studies was the bloodstream, followed by respiratory, abdominal and urinary tract. The studied strain included CRE14 16 17 20 22–24 and carbapenem-resistant K. pneumoniae.15 18 19 21 The outcomes involved 30-day mortality in nine studies,14 16–19 21–24 clinical response in four studies,16 19 20 22 microbiological response in two studies20 22 and nephrotoxicity in seven studies.14 17–20 22 23

Assessment of study quality

As depicted in online supplemental table 2, all studies were evaluated by NOS. The NOS scores of all assessed studies were ≥7. Thus, all studies were considered to have low risk of bias.



The 30-day mortality was reported in nine studies, including 766 patients.14 16–19 21–24 As displayed in figure 2, the CAZ-AVI group showed a lower 30-day mortality rate compared with polymyxins group when random effects were employed (RR=0.48, 95% CI 0.37 to 0.63, I2=10%, p<0.0001). Funnel plots of included studies showed that all plots exhibited roughly symmetrically inverted funnel shapes, indicating no publication bias (online supplemental figure 1).

Figure 2

The 30-day mortality of the CAZ-AVI regimens compared with polymyxins regimens. CAZ-AVI, ceftazidime–avibactam.

Clinical success

Clinical success was reported in four studies, including 463 patients.16 19 20 22 Compared with those in the polymyxins group, patients in the CAZ-AVI group had a significantly higher clinical cure rate as illustrated in figure 3 (RR=1.71, 95% CI 1.33 to 2.20, I2=35%, p<0.0001).

Figure 3

Clinical success of the CAZ-AVI regimens compared with polymyxins regimens. CAZ-AVI, ceftazidime–avibactam.

Bacterial eradication

Two studies with 249 patients reported data on microbiological response.20 22 As shown in figure 4, a pooled analysis with random-effects models revealed comparable potencies of CAZ-AVI with polymyxins in microbiological eradication abilities (RR=1.16, 95% CI 0.97 to 1.39, I2=0, p>0.05).

Figure 4

Bacterial eradication of the CAZ-AVI regimens compared with polymyxins regimens. CAZ-AVI, ceftazidime–avibactam.


Seven studies with 696 patients reported nephrotoxicity.14 17–20 22 23 As shown in figure 5, pooled results from the included studies indicated a lower nephrotoxicity rate in the CAZ-AVI group relative to the polymyxins counterpart (RR=0.42, 95% CI 0.23 to 0.77; I2=35%, p<0.05).

Figure 5

Nephrotoxicity of the CAZ-AVI regimens compared with polymyxins regimens. CAZ-AVI, ceftazidime–avibactam.


In this meta-analysis, seven retrospective and four prospective cohort studies with 1111 patients were included to compare the efficacy and safety of CAZ-AVI with polymyxins regimens in patients with CRE infection. Polymyxins-containing regimens had an increased risk of mortality, clinical failure and nephrotoxicity compared with CAZ-AVI. However, CAZ-AVI did not exhibit superior bacterial eradication ability over polymyxins. To date, this work remains the first meta-analysis to compare the efficacy and safety of CAZ-AVI with polymyxins in treating CRE infection.

CRE pathogens have spread alarmingly in recent years, showing marked correlations with a high risk of morbidity, mortality and considerable economic burden. Inactivated enzyme production, enhanced efflux activity and reduced cell permeability have been considered as the most frequent and important mechanisms of CRE prevalence. Until now, current antimicrobial therapy options for CRE infections are still very limited, including polymyxins and novel β-compound preparations of lactamase inhibitors, such as ceftazidime/avibactam, aztreonam/avibactam, meropenem/vaborbactam and imipenem/cilastatin/relebactam.25 However, given the availability of such drugs, CAZ-AVI and polymyxins-containing regimens have been highly recommended as the frontline agents in the treatment of multidrug-resistant gram-negative bacterial infections in China. Polymyxins exert their antibacterial effect by increasing the permeability of the bacterial outer membrane through interaction with lipopolysaccharides in the outer membrane of gram-negative bacteria.26 As a novel β-lactam/β-lactamase inhibitor combination, CAZ-AVI exhibits activity against various clinically important β-lactam-resistant bacteria producing class A and K. pneumoniae carbapenemases and class C and certain class D enzymes, but not against the metallo-β-lactamases (MBL) of class B enzymes.27 Consequently, the distinct action mechanisms of the two drugs contribute different clinical effectiveness.

Early meta-analysis including three randomised controlled trials observed that CAZ-AVI was similar to carbapenem for the treatment of Enterobacteriaceae infections and could provide an alternative to carbapenem.28 However, a recent meta-analysis demonstrated an advantage of CAZ-AVI to treat CRE bloodstream infections on efficacy and safety, and subgroup analysis revealed that the CAZ-AVI group had a significantly lower 30-day mortality than colistin-based regimens.12 Moreover, previous studies indicated the incidence rate of polymyxins-related nephrotoxicity ranged from 11.8% to 50.6%.29 Therefore, the safety advantage of CAZ-AVI compared with polymyxins was not surprising.

In our meta-analysis, most studies applied combination therapy, which was consistent with the current recommended guidelines, mainly because most CRE infections involved multidrug-resistant mechanisms. Therefore, most of the included studies were based on the combined administration of CAZ-AVI or polymyxins, and the combined administration plan also included other activity antibiotics. Given the heterogeneously resistant nature of polymyxins, the combination administration is recommended as the first-line treatment. However, a recent meta-analysis did not observe an advantage of polymyxins combination therapy on multidrug-resistant gram-negative bacterial infections.30 In addition, another meta-analysis suggested that CAZ-AVI in monotherapy or combination therapy for CRE infections demonstrated similar effect on mortality and microbiological cure rates.31 Therefore, both combined and single-drug regimens were included in our meta-analysis.

CAZ-AVI is the first new type β-lactamase inhibitor for the CRE infection treatment approved by FDA.32 Moreover, CAZ-AVI has been approved for infections without additional therapeutic options, such as for complicated intra-abdominal infections, complicated urinary tract infections and hospital-acquired pneumonia/ventilator-associated pneumonia. CAZ-AVI shows low plasma protein binding and steady-state distribution volume, allowing it to maintain an adequate trough concentration to achieve the bactericidal effect. Additionally, CAZ-AVI is excreted almost exclusively through renal excretion, resulting in high urinary drug concentrations,33 and its good blood concentration in the bronchial epithelial lining fluid endows CAZ-AVI with a favourable pharmacological profile for bloodstream, abdominal, respiratory and urinary tract infection. These infection sites were the most common in our meta-analysis. Regarding the causative pathogens, CAZ-AVI sustains excellent ability in the treatment of CRE infection except for the MBL-producing Enterobacteriaceae. Interestingly, the CAZ-AVI and aztreonam combination offers a therapeutic advantage on patients with bloodstream infection caused by MBL-producing Enterobacterales.17 Although the types of bacteria and diseases were not limited in our studies, little heterogeneity was detected in our work. Considering the well-characterised pharmacokinetic parameters and pharmacological activity of CAZ-AVI, all CRE infections were included in our meta-analysis.

This review had some limitations that should be acknowledged. First, all included studies were observational in design, and the seven retrospective and four prospective cohort studies included only 1111 patients, resulting in a small sample size. Second, most studies did not provide antimicrobial resistance or enzyme production information; moreover, changes in the pathogen’s resistance to antibiotic exposure and specific information of various diseases were not available in all included studies. Thus, subgroup analyses of diseases and strains were not performed given the lower heterogeneity. Finally, among these included studies, colistin was taken as a control group in six studies, while polymyxin B in four studies, and both two drugs were used as a control group in the other article, but the dosage was not reported in detail. Therefore, further subgroup analysis was not performed for the two drugs as a control group.


Overall, CAZ-AVI showed advantages over polymyxins in terms of mortality, clinical success and safety. In addition, similar bacterial eradication profiles were observed for CAZ-AVI and polymyxins groups. To the best of our knowledge, this work is the first meta-analysis to be conducted on this topic. However, given the small sample size and the limitations of the studies included in the present work, large, high-quality, multicentre, randomised, double-blind, controlled trials should be conducted to establish the dominance of CAZ-ZVI over polymyxins in treating CRE infections.

Data availability statement

No data are available.

Ethics statements

Patient consent for publication

Ethics approval

Ethics approval was not required as the study was based on existing publicly available data.


The authors gratefully acknowledge the suggestions given by Rong-Rong Wang on the manuscript.


Supplementary materials


  • Contributors YY-L and XJ-W are responsible for literature retrieval and data extraction. NC is responsible for handling disagreements and data analysis. PY is responsible for paper writing. XY-L is responsible for overall design and final approval of the manuscript.

  • Funding This work was funded by Zhejiang Pharmaceutical Association hospital pharmacy special scientific research funding project (2019ZYY11), Zhejiang Medical Association clinical research funding project (2020ZYC-A107), Zhejiang medical and health science technology project (2019RC168) and the Nature Science Foundation of Zhejiang province (LQ20H300003, LYY21H300004).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.