Ceftazidime-Avibactam in the Treatment of Patients with Bacteremia or Nosocomial Pneumonia: A Systematic Review and Meta-analysis

Abstract

Introduction

Ceftazidime-avibactam (CAZ-AVI) is a combination of the third-generation cephalosporin ceftazidime and the novel, non-β-lactam β-lactamase inhibitor avibactam that is approved for the treatment of pediatric (≥ 3 months) and adult patients with complicated infections including hospital-acquired and ventilator-associated pneumonia (HAP/VAP), and bacteremia. This systematic literature review and meta-analysis (PROSPERO registration: CRD42022362856) aimed to provide a quantitative and qualitative synthesis to evaluate the effectiveness of CAZ-AVI in treating adult patients with bacteremia or nosocomial pneumonia caused by carbapenem-resistant Enterobacterales (non metallo-β-lactamase-producing strains) and multi-drug resistant (MDR) Pseudomonas aeruginosa infections.

Methods

The databases included in the search, until November 7, 2022, were Embase and PubMed. A total of 24 studies (retrospective: 22, prospective: 2) with separate outcomes for patients with bacteremia or pneumonia were included.

Results

The outcomes assessed were all-cause mortality, clinical cure, and microbiological cure. Qualitative (24 studies) and quantitative (8/24 studies) syntheses were performed. The quality of the studies was assessed using the MINORS checklist and the overall risk of bias was moderate to high.

Conclusions

In studies included in the meta-analysis, lower all-cause mortality for patients with bacteremia (OR = 0.30, 95% CI 0.19–0.46) and improved rates of clinical cure for patients with bacteremia (OR = 4.90, 95% CI 2.60–9.23) and nosocomial pneumonia (OR = 3.20, 95% CI 1.55–6.60) was observed in the CAZ-AVI group compared with the comparator group. Data provided here may be considered while using CAZ-AVI for the treatment of patients with difficult-to-treat infections.

Systematic Review Registration

PROSPERO CRD42022362856.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40121-024-00999-y.

Key Summary Points

Ceftazidime-avibactam (CAZ-AVI) in bacteremia and nosocomial pneumonia (NP); pathogens carbapenem-resistant Enterobacterales (CRE), multidrug resistant (MDR) P. aeruginosa.

Bacteremia treated with CAZ-AVI was associated with lower mortality and improved clinical cure.

Patients with NP who were treated with CAZ-AVI demonstrated improved clinical cure.

CAZ-AVI is a treatment option for patients with bacteremia and NP with MDR-gram-negative bacteria (GNB).

Introduction

Gram-negative bacilli, Enterobacterales (such as Klebsiella pneumoniae, Escherichia coli, and Enterobacter spp.) and Pseudomonas aeruginosa are common human pathogens with increasing antibacterial resistance [1]. Carbapenems have been an effective treatment for most infections caused by Gram-negative bacteria; however, increased use clinically has resulted in carbapenem-resistant Enterobacterales (CRE) and P. aeruginosa (CRPA) [2]. Based on World Health Organization (WHO) classification, these are critical-priority pathogens [3]. CRE is one of the three most urgent antimicrobial-resistance threats identified by the Centers for Disease Control and Prevention (CDC) [4]. Both CRE and CRPA cause severe and often fatal infections including bacteremia, hospital-acquired pneumonia (HAP), and ventilator-associated pneumonia (VAP), and may further develop resistance to several antimicrobials [5, 6]. Progressive accumulation of antibiotic resistance mechanisms manifests in multidrug-resistance (MDR; defined as non-susceptible to ≥ 1 agent in ≥ 3 categories of antimicrobials). Invasive infections (e.g., bacteremia, nosocomial pneumonia) caused by MDR Enterobacterales and P. aeruginosa, are associated with mortality rates of ~ 34–56% [4, 7–11].

Ceftazidime-avibactam (CAZ-AVI) is a combination of the third-generation cephalosporin ceftazidime and the novel, non-β-lactam β-lactamase inhibitor avibactam. The combination exhibits in vitro activity against a broad range of Gram-negative bacteria, including against isolates harboring extended spectrum β-lactamase (ESBL)-, AmpC-, OXA-48, and serine carbapenemases (K. pneumoniae carbapenemase [KPC]) as well as MDR P. aeruginosa isolates; but not against metallo-β-lactamase (MBL)-producing strains [12]. CAZ-AVI is a frontline agent for the treatment of CRE infections and was approved by the U.S. Food and Drug Administration (FDA) in 2015 and the European Medicines Agency (EMA) in 2016 for the treatment of adult patients, and later pediatric patients (≥ 3 months), with complicated intraabdominal infections, complicated UTIs, hospital-acquired pneumonia (HAP) including VAP (and in EU for bacteremia in adults associated with any of these infections) [13–15]. In the EU specifically, CAZ-AVI is also approved for the treatment of patients with infections caused by Gram-negative organisms with limited treatment options.

A growing body of real-world evidence (RWE) has evaluated the effectiveness of CAZ-AVI in patients with bacteremia or nosocomial pneumonia. In the context of increasing antimicrobial resistance, it is important to assess the usefulness of newer-generation antimicrobials such as CAZ-AVI in treating these severe infections and use them judiciously to ensure continued effectiveness. Previous systematic literature reviews (SLRs) and meta-analyses evaluating the real-world use of CAZ-AVI have generally focused on any infection caused by Gram-negative organisms [16–18]. Such reviews have not elucidated outcomes for specific infection types due to the high variability in reported outcomes as well as the potential for bias in outcome assessments as a result of combining infection types with disparate outcomes. Bacteremia and nosocomial pneumonia cause disproportionately worse outcomes than other infection types [1, 5]. This SLR and meta-analysis is a comprehensive, quantitative, and qualitative synthesis of the clinical and microbiological outcomes of treating adult patients with CAZ-AVI, including those with bacteremia or nosocomial pneumonia caused by CRE (non-MBL producing) and MDR P. aeruginosa infections.

Methods

This SLR followed the preferred reporting items for systematic reviews and meta-analysis (PRISMA) statement and guidance [19]. This SLR protocol was registered on the international prospective register of systematic reviews (PROSPERO: CRD42022362856) on October 6, 2022 [20] and was updated to include a meta-analysis due to the availability of quantitative data on bacteremia and nosocomial pneumonia. This article is based on published literature and does not contain any previously unreported studies with human participants or animals.

Search Strategy

Systematic Embase and PubMed searches were conducted to include articles from 2015 onwards to November 7, 2022. Search terms used were “ceftazidime avibactam”, “hospital-acquired pneumonia”, “ventilator-associated pneumonia”, “bacteremia”, “pneumonia”, “bloodstream infection”, P. aeruginosa, Enterobacterales, Klebsiella pneumoniae, E. coli, Enterobacter cloacae and other synonyms (Supplemental Table 1). Additionally, similar reviews were also scanned manually for relevant articles that may not have been captured in the database search.

Study Selection

We included real-world studies (both prospective and retrospective), published in the English language. After search results were extracted from the selected databases, duplicates were removed using citation management software (Endnote V20.2 Clarivate Analytics, PA, USA), and a list of all eligible studies was created. The process of screening and inclusion and exclusion criteria are presented in Fig. 1 (PRISMA flow diagram).

Fig. 1.

PRISMA flow diagram of the search process and study selection. *Studies included patients with multiple infection types and outcomes could not be differentiated by indications and such studies were excluded. AMR antimicrobial resistance, AMS antimicrobial stewardship, PK pharmacokinetics, PD pharmacodynamics, RCTs randomized controlled trials

Screening, Data Extraction, and Quality Assessment

A two-step process involving two reviewers and a third independent reviewer was followed for screening and data extraction. Relevant studies were identified based on a title-abstract screening, followed by a full-text review of the shortlisted studies. Any disagreements on inclusion were resolved by discussion with a third reviewer. Data from eligible studies were extracted in an Excel spreadsheet (Microsoft Corp., Redmond, WA, USA). The quality of relevant studies was assessed using the methodological index for non-randomized studies (MINORS) checklist [21] with cross-review.

Inclusion and Exclusion Criteria

The following article types were excluded, with reasons for exclusion documented in each case: clinical trial data for any of the primary indications for CAZ-AVI, pediatric data, in vitro/animal/modeling simulation studies, studies including microbiological surveillance or population PK and PK/PD modeling, conference materials, review articles, guidelines, commentaries/opinion pieces, and editorials not reporting original outcomes data for patients treated with CAZ-AVI. We also excluded case reports and case series, studies without extractable data for the indication of interest (bacteremia and nosocomial pneumonia) and articles primarily reporting outcomes for bacteria with exclusive intrinsic resistant (MBL, New Delhi metallo-β-lactamases [NDM], Verona integron encoded metallo-β-lactamase [VIM], and imipenemase metallo-β-lactamase [IMP] producing organisms) to CAZ-AVI. Meta-analyses of published literature were also excluded from the analysis to avoid duplication of data; however, the source data/references within such articles were reviewed to ensure all relevant primary data were included.

Statistical Analysis

Meta-analysis was performed for synthesizing information and pooled estimates using package 'meta' on R version 1.1.453. τ² and I² (< 50%: no heterogeneity, ≥ 50%: significant heterogeneity) statistics were used to quantify heterogeneity among included studies. A random effects model was applied in case of significant heterogeneity; otherwise, a common/fixed-effect model was applied. Meta-analytical methods used were the Mantel–Haenszel method, restricted maximum-likelihood estimator for τ², and Q-profile method for the confidence interval (CI) of τ² and τ.

Separate analysis was conducted for each outcome and plots with ≥ 2 studies were included. CAZ-AVI therapy given as either monotherapy, combination therapy, or a mix of monotherapy and combination therapy was considered as one treatment group. Similarly, data from different comparators such as best available therapy (BAT) were grouped and presented in a combined manner as part of a single comparator group.

Results

Literature Search

A total of 1006 publications were identified. After further screening and exclusion, 24 relevant full-length articles were included in the qualitative synthesis and 8/24 articles in the quantitative synthesis (meta-analysis) (Fig. 1). Included publications reported data for a total of 1114 patients (bacteremia: 946, nosocomial pneumonia: 168) treated with CAZ-AVI either as a part of monotherapy, combination therapy, or both, and 640 patients (bacteremia: 539, nosocomial pneumonia: 101) treated with alternative or comparator antibiotics. The majority of patients included had bacteremia or pneumonia due to Enterobacterales except for two studies where patients with P. aeruginosa (bacteremia: CAZ-AVI—24, other antibiotic—37) [22, 23] infections were included.

For completeness, data from five relevant case series with aggregate data reported for bacteremia and nosocomial pneumonia is presented separately in Supplemental Tables 2 and 3.

Study Characteristics

Publications comprised 12 retrospective cohort studies [23–34], ten retrospective comparative/case–control studies [22, 35–43], and two prospective registry studies [44, 45] (Table 1, Fig. 2A). Most studies were 2020 onwards and were from China [22, 26, 30, 31, 33, 39, 41, 43] and Spain [23, 25, 32, 36, 42, 45] (Fig. 2B and C).

Table 1.

Summary of studies included in quantitative and qualitative synthesis

Author, year (ref) (study type)	Baseline population	Country	No. of patients		Pathogens isolated from patients treated with CAZ-AVI (resistance mechanism)	Combination drugs with CAZ-AVI	Time from first culture to start of specific therapy	Treatment durationa (median), days
			CAZ-AVI with indication of interest	Other treatment
Studies with bacteremia
Falcone M, 2020 [35] (Retrospective comparative/case control)	Patients with KPC-Kp BSI hospitalized in ICU	Italy	Bacteremia: 13	Colistin: 61, other: 17	KPC-Kp (100% KPC)	Monotherapy or combination (aminoglycosides or fosfomycin)	Median (IQR): 8.5 (1–36) h in survivors vs. 48 (5–108) h in non-survivors	NR
Tsolaki V, 2020 [38] (Retrospective comparative/case control)	Critical ill, MV patients with CRE infections	Greece	Bacteremia: 22	BAT: 28	CRE infections (100% KPC)	Monotherapy or combination (aminoglycosides, colistin, tigecycline, trimethoprim-sulfamethoxazole, fosfomycin)	Mean: CAZ-AVI: 1.82 ± 0.57 Comparator: 0.96 ± 0.26 days	10.5 (9.75–14.25)
Chen L, 2021 [43] (Retrospective comparative/case control)	Hospitalized patients with CRE- BSI	China	Bacteremia: 35	Other antibiotics: 152	CRE infections (NR)	Monotherapy or combination (tigecycline and PB sulfate)	Median (IQR): 6–7 (4–12.75) days	10.0 (7.5–13.5)
Lima O, 2022 [36] (Retrospective comparative/case control)	CRKp-OXA-48-type bacteremia	Spain	Bacteremia: 33	BAT: 43	K. pneumoniae (100% OXA-48)	Monotherapy: 30/33; combination: 3/33 (colistin and fosfomycin)	NR	Mean (IQR): 12 (9–15)
Shields RK, 2017 [37] (Retrospective comparative/case control)	Carbapenem-resistant K. pneumoniae (CR-Kp) bacteremia	United States	Bacteremia: 13	(CB + AG) + (CB + COL) + others = 96	CR-Kp (KPC-2: 76%; KPC-3: 24%)	Monotherapy: 8/13; combination: 5/13 (gentamicin)	Median (IQR): 55.7 (25–67) h	13 (5–23)
Castón JJ, 2017 [40] (Retrospective comparative/case control)	Hematologic malignancies or aplastic anemia patients with CPE bacteremia	3 centers in Spain and 1 in Israel	Bacteremia: 8	OAA: 23	CPE (OXA-48: 61.3%; KPC: 38.7%)	Combination: all patients (aminoglycoside, carbapenems, fosfomycin, tigecycline, or colistin)	Median: 2 days	8 days
Chen J, 2022 [22] (Retrospective comparative/case control)	Patients with CRPA infection	China	Bacteremia: 12	Polymyxin B with carbapenems and amikacin: 37	Monomicrobial infection: CRPA infection (NR)	Monotherapy: all patients	Median (IQR): 6–7 (4–12.75) days	10 (6–17) days
Micozzi A 2021 [34] (Retrospective)	Patients with high-risk hematological malignancies with BSI	Italy	Bacteremia: 12	–	K. pneumoniae (100% KPC)	Monotherapy: 1/12; combination: 11/12 (tigecycline and/or gentamycin)	4 h	NR
Sousa A, 2018 [45] (Prospective)	Any infection produced by OXA-48-producing CPE	Spain	Bacteremia: 26	–	OXA-48-producing K. pneumoniae (100% OXA-48)	Monotherapy: 46; combination: 11 (IV colistin, inhaled colistin, tigecycline, amikacin, and imipenem)	Median (IQR); Monotherapy = 2 (0–15); Combination = 4 (2–17)	13
De la Calle C, 2019 [25] (Retrospective)	Patients with OXA-48 CPE and OXA-48 -like infections	Spain	Bacteremia: 8 episodes	–	OXA-48 CPE and OXA-48 -like infections (OXA-48: 95.8%, CTX-M15: 4.17%)	Monotherapy: 14 episodes; Combination: 10 episodes (amikacin, colistin, tigecycline or a combination of 2 of these drugs)	Mean (SD): 2.5 (1.7) days	14 (8.5–30)
Meng H, 2022 [26] (Retrospective)	Hospitalized hematological malignancy patients with CRKP BSI	China	Bacteremia: 7	–	CRKP (NR)	Monotherapy	NR	NR
Tumbarello M, 2019 [29] (Retrospective)	KPC-Kp bacteremia	Italy	Bacteremia: 104	–	KPC-producing K. pneumoniae (100% KPC)	Monotherapy and combination (gentamicin, tigecycline, colistin, fosfomycin, and other drugs)	Median (IQR): 7 (3–9) days	14 (3–28)
Tumbarello M, 2021 [28] (Retrospective)	KPC-Kp infections	Italy	Bacteremia: 391	–	K. pneumoniae (100% KPC)	Monotherapy: 113/391 (28.9%) Combination therapy: 278/391 (71.1%) (fosfomycin, tigecycline, gentamicin, or meropenem, colistin, amikacin, others)	NR	12 days (9–16)
Chen Y, 2022 [31] (Retrospective)	Patients with BSI K. pneumoniae	China	Bacteremia: 17	–	CRKP (NR)	Monotherapy	NR	> 48 h
Corbella L, 2022 [23] (Retrospective)	Patients with MDR/XDR P. aeruginosa infections	Spain	Bacteremia: 12 (includes 3 vascular CR- bacteremia)	–	MDR/XDR P. aeruginosa (NR)	Combination: all patients (colistin, aztreonam, and amikacin)	Median (IQR): 7 (3.5–12) days	10.0 (6.0–14.0)
Shen L, 2021 [33] (Retrospective)	Patients with SOT with BSIs due to CRKP	China	Bacteremia: 9	–	CRKP (NR)	Monotherapy: 4; combination: 5 (PB, meropenem, tigecycline)	Time until appropriate therapy = 1 (0–3) day	NR
Studies with pneumonia
Shi Y, 2021 [39] (Retrospective comparative/case control)	Critically ill ICU patients with CRKP-induced HAP or VAP	China	VAP: 33	Tigecycline (TGC): 42	CRKP- CRE (NR)	Monotherapy and combination (Carbapenem and others [aztreonam, fosfomycin, amikacin, and PB])	Median (IQR): CAZ-AVI- 4 (1–8), TGC- 2 (1–5.5) days	≥ 72 h
Chen W, 2020 [30] (Retrospective)	Lung transplant recipients with infections caused by XDR-GNB	China	Pneumonia: 10	–	CRKP: 9, CRPA: 1; (KPC-2: 100% and other [CTX-M-64 and 65, SHV-11 and 64, TEM-1, DHA-1, OXA-10])	Monotherapy: 2; combination: 8 (tigecycline and PB)	NR	15.7 (9–22)
Studies with bacteremia and pneumonia
Castón JJ, 2022 [42] (Retrospective comparative/case control)	Hospitalized patients with CPE infection	Spain	Bacteremia: 72, pneumonia: 23	BAT: Bacteremia: 39; pneumonia: 16	CPE (OXA-48: 73.5%; KPC: 26.5%)	Monotherapy and combination (amikacin, tigecycline, colistin, gentamicin, fosfomycin, tobramycin, and aztreonam)	Median (IQR): 2 (1–4) days	11 (8–15)
Fang J, 2021 [41] (Retrospective comparative/case control)	CRKP infection	China	Bacteremia: 14, pneumonia: 23	Polymyxin B: bacteremia: 44; pneumonia: 43	CRKP (NR)	Monotherapy and combination (most common- carbapenems, aminoglycosides, tigecycline; Others: quinolones, fosfomycin, cephalosporin, SMZ)	NR	15 (9–18)
Vena A, 2020 [24] (Retrospective)	Patients hospitalized for infections due to MDR-GNB other than CRE	Italy	Bacteremia: 7; nosocomial pneumonia: 20	–	P. aeruginosa and ESBL-producing Enterobacterales (NR)	Monotherapy and combination (colistin, aminoglycosides, carbapenems, or fosfomycin)	11 days to switch to CAZ-AVI when given as secondary therapy	13 (3–49)
King M, 2017 [27] (Retrospective)	CRE infection	United States	Bacteremia: 13, pneumonia: 8	–	CRE (K. pneumoniae, E. coli, Enterobacter spp. Providencia stuartii, Serratia marcescens, K. oxytoca) (NR)	Monotherapy: 33/60 (55%); Combination: 27/60 (45%) (aminoglycosides, polymyxin, and tigecycline)	Hospital day CAZAVI was started: median (IQR)- Monotherapy:7.5 (4.8–21.5) Combination: 7 (4–7)	NR
Castón JJ, 2020 [32] (Retrospective)	Hospitalized patients with KPC-producing Klebsiella pneumoniae (KPC-Kp) infections	Spain	Bacteremia: 23; pneumonia: 14	–	KPC-Kp (100% KPC)	Monotherapy and combination (gentamicin, tigecycline, colistin, tigecycline + gentamicin, fosfomycin + gentamicin)	Median (IQR): 3 (1–5) days	14 (9–17)
Karaiskos I, 2021 [44] (Prospective)	Patients with KPC- or OXA-48-Kp infections	Greece	Bacteremia: 95 VAP/HAP: 37	–	Carbapenemase-producing K. pneumoniae (KPC: 95%, OXA-48: 5%)	Monotherapy: 68; combination: 79 (with at least another active agent [colistin, aminoglycosides, colistin plus tigecycline, tigecycline, trimethoprim/sulfamethoxazole, fosfomycin, ciprofloxacin, colistin plus aminoglycoside and tigecycline plus fosfomycin])	NR	13 (5–50)

Dose adjustment in case of renal impairment according to the manufacturer's instructions was mentioned in Sousa 2018, Tumbarello 2019, 2021 and Caston 2017 studies. Chen W 2020 reported that renal adjustment for CAZ-AVI occurred in 4 (40%) patients

BAT best available therapy, BSI bloodstream infection, CAZ-AVI ceftazidime-avibactam, CB + AG carbapenem + aminoglycosides, CB + COL carbapenem + colistin, Ccr creatinine clearance, CPE carbapenemase-producing Enterobacteriaceae, CRE carbapenem-resistant Enterobacterales, CRKP carbapenem-resistant Klebsiella pneumoniae, CRPA carbapenem-resistant Pseudomonas aeruginosa, ESBL extended-spectrum beta-lactamase, GNB Gram-negative bacteria, HAP hospital-acquired pneumonia, ICU intensive care unit, IQR interquartile range, IV intravenously, KPC Klebsiella pneumoniae carbapenemase, KPC-Kp KPC-producing K. pneumoniae, q8h every 8 h, MDR multidrug resistance, MV mechanical ventilation, NA not available, NR not reported, OXA-48 oxacillinase-48, SOT solid organ transplantation, VAP ventilator-associated pneumonia, VC-related vascular catheter-related, XDR extensive drug-resistant

^aReported for all patients included in the study

Fig. 2.

Pie charts for the study characteristics (A) study types, (B) study period, and (C) region/country

Of a total of 24 studies, 13 included patients with infection caused by K. pneumoniae, comprising six studies focused on CR-K. pneumoniae (CRKP, mechanism of resistance not reported), five studies focused on KPC- K. pneumoniae, and two studies focused on OXA-48K. pneumoniae. Six studies included patients with infections caused by more than one Enterobacterales comprising three studies focused on carbapenemase-producing Enterobacterales (CPE) and three studies focused on CRE. Two studies included patients with infection caused by P. aeruginosa (MDR/extensive drug-resistant, CR), and three studies with mixed pathogens.

Outcomes were not segregated based on the type of CAZ-AVI treatment (mono/combination) for bacteremia and nosocomial pneumonia in the included studies, hence outcomes extracted for the CAZ-AVI group include data from patients treated with either monotherapy, combination therapy, or both. We could extract separate outcomes for patients with bacteremia in 16/24 studies, patients with nosocomial pneumonia in 2/24 studies, and for both bacteremia and nosocomial pneumonia in 6/24 studies.

Patients with Bacteremia

Qualitative Synthesis

Mortality was reported in 18/22 studies which enrolled patients with bacteremia (Table 2). Overall, 30-day mortality was 24% (n = 217/904) in patients treated with CAZ-AVI and 40.12% (n = 199/496) in patients treated with other antibiotics.

Table 2.

Outcomes of studies involving ceftazidime-avibactam for the treatment of patients with bacteremia or nosocomial pneumonia

Publication	Patient population characteristics Age, median (range), years/male (%)		Baseline severityc	Reported outcomes (%)				Adverse events/safety	Predictors-mortality/survival
	Ceftazidime-avibactam	Comparator		Clinical cure	30-day mortalityd	90-day mortality	Microbiological cure
Studies with bacteremia
Micozzi A 2021	50.8 (3–68); 37%	–	Shock: 7/16 (44%); Neutropenia: < 1000 neutrophils/cmm, 16/16 (100%); < 100 neutrophils/cmm, 15/16 (94%)	NR	9.09	NR	NR	NR	KPC-K. pneumoniae BSI developing during inactive antibiotic treatment, initial inactive antibiotic treatment/NR
Sousa A, 2018	64 (26–86); 77%	–	Sepsis/Septic shock, n/N (%) 31/57 (54); vasopressor use 20/57 (35)	NR	15.38c	NR	NR	Acute kidney injury (n = 2); decrease in the level of consciousness (n = 1); status epilepticus (n = 1); NR with CAZ-AVI	INCREMENT-CPE score > 7/NR
De la Calle C, 2019	Mean (SD) = 58.85 (16.03); 82.6%	–	Septic shock: 4/24 episodes (16.7%)	62.5	NA	25	NR	CAZ-AVI-related AE = 4 (16.7%)- Diarrhea (mild, self-limited)-1, Death with VAP showed progressive thrombocytopenia and cholestasis-1, myoclonus-1 and encephalopathy-1 (2 deaths associated with renal impairment)	NR/NR
Meng H, 2022	46.1 ± 13.8; 65.9%	–	Septic shock: 86/129 (66.7%)	NR	14.2	NR	NR	NR	NR/NR
Falcone Mb, 2020	64 (53–74); 63.7%	–	Septic shock: 40/102 (39.2%); AKI: 15; SOFA score median: 5; CCI median: 2; KPC-Kp intestinal colonization: 52/102; APACHE II score, median: 15	NR	CAZAVI: 23.1, COL: 44.3; other: 41.2	NR	NR	Endocarditis death: 1	Age, SOFA score, time from blood culture collection to appropriate therapy, CVD, primary bacteremia/NR
Tumbarello M, 2019	61 (27–79); 65.4%a	–	Septic shock: 34/104 (32.7) (Present when CAZ-AVI treatment was started); Pitt score, median: 4 (0–7); High-risk pneumonia: 43: 64 (61.5)	NR	36.5	NR	NR	NR	Charlson Comorbidity Index ≥ 3, neutropenia, septic shock (at start of salvage therapy), recent MV/Receipt of CAZ-AVI
Tumbarello M, 2021	65 (56–75); 70.8%a	–	Severity of illness at infection onset: INCREMENT-CPE score ≥ 8: 109/391 (27.8); Septic shock: 70/391 (17.9)	NR	26.3	NR	NR	Adverse reactions: 13/391 (3.3%) patients	INCREMENT score ≥ 8 at infection onset, neutropenia at infection onset, septic shock at infection onset, LRTI, CAZ-AVI dose adjustment for renal function/ administration of CAZ-AVI via prolonged infusion (lasting ≥ 3 h)
Lima O, 2022	61 (46–76); 67%	70 (59–82); 74%	CAZ-AVI vs. BAT: INCREMENT-CPE score ≥ 7 n (%) = 16 (49) vs. 18 (42); Pitt's Index > 2: 17 (51) vs. 19 (44); Septic shock: 13 (39) vs. 19 (44)	CAZ-AVI: 97 BAT: 74	CAZ-AVI: 12; BAT: 26	NR	NR	CAZ-AVI: 1 (3%)—Encephalopathy related with CAZ-AVI; BAT 8 (19%)- colistin-related acute renal injures (5 patients, 12%) and neurologic symptoms (2 due to colistin and 1 secondary to imipenem)	Age/NR
Shields RK, 2017	66 (32–91); 54%	CB + AG = 57 (32–87); CB + COL = 59 (26–84); Others = 62 (25–90); CB + AG = 64%; CB + COL = 60%; Others = 51%	ICU at the time of bacteremia 6/13 (46%), RRT 2/13 (15%), Pitt bacteremia score 4 (1–6) and APACHE II score 20 (16–33)	CAZ-AVI: 85	CAZ-AVI: 7.7; Comparators: 31	CAZ-AVI: 7.7; Comparators: 44.8	NR	Drug toxicity: AKI at 48 h, 7 days, EOT: 1(9), 1(9), 2 (18)	NR/NR
Tsolaki V, 2020	Mean ± SEM 62.82 ± 3.24; 68.2%a	57.07 ± 3.04; 67.85%a	Septic shock at infection onset n (%), CAZ-AVI vs. control: 15 (68.2) vs. 19 (67.9); Pitt score mean ± SEM: 5.90 ± 0.51 vs. 5.78 ± 0.45	CAZ-AVI: 81.8; comparator: 53.57	CAZ-AVI: 18.2; comparator: 42.9	NR	NR	CAZ-AVI: Non-significant decrease in creatinine levels compared to the control group	NR/treatment with CAZ-AVI containing regime, lower SOFA score at infection onset
Chen Yb, 2022	Mean (SD) = 53.36 (18.92); 74.8%	–	NR	NR	17.6	NR	NR	NR	Urinary catheterization/NR
Castón JJ, 2017	61 ± 42; 50%	59 ± 59; 65.2%	For CAZ-AVI: Sepsis 3 (37.5%), severe sepsis 2 (25%), septic shock 3 (37.5%); Charlson Comorbidity Index, median (IQR) 2.5 ± 7; Pitt Index 3 ± 10; neutropenia 5 (62.5%)	CAZ-AVI: 75 Other Ab: 34.8	CAZ-AVI: 25 Other Ab: 52.2	NR	NR	NR	Pitt Index score, unresolved neutropenia, septic shock/NR
Chen L, 2021	67.0 ± 14.5; 61.5% (for CAZ-AVI + comparator)	NR	Total patients: APACHE II score 12.0 (9.0, 15.0), Pitt score 2.0 (1.0, 3.0)	NR	CAZ-AVI: 17.1; comparator: 47.4	NR	NR	NR	Pitt bacteremia score, immunocompromised status/Source control of infection, appropriate empirical therapy
Corbella L, 2022	65.1 ± 15.9; 68.9%	–	Sepsis 31 (50.8%); acute respiratory failure 18 (29.5); acute renal failure 22 (36.1); hypoperfusion with no requirement of inotropic support 25 (41.0); inotropic support 6 (9.8)	Pneumonia: 43: 58.33;	16.67	NR	NR	Diarrhea = 1 (1.6); rash and/or drug hypersensitivity = 1 (1.6); nephrotoxicity = 1 (1.6)	Previous CVD, acute renal failure, acute respiratory failure and hypoperfusion/ NR
Shen L, 2021	Median 50; 66.67%	–	Severe sepsis or septic shock 51/ 89 (57.3%); Pitt bacteremia score 3 (1–6); INCREMENT-CPE mortality score, points 8 (6–12); High INCREMENT-CPE mortality score (≥ 8) 58 (65.2%)	77.78	22.22	NR	88.89	NR	Pitt bacteremia score, hematological malignancy/NR
Chen J, 2022	65 (58–70); 72.5%	58 (48–69); 74.1%	CAZ-AVI vs. Polymyxin B: CCI 3 (2–4) vs. 3 (2–4); aCCI 5 (3–6) vs. 4 (3–6); APACHE II score 15 (12–19) vs. 18 (12–22), SOFA score 5 (4–7) vs. 7 (4–8); septic shock: 16 (31.4) vs. 41 (48.2)	NR	NR	NR	NR	NR	Age/CAZ-AVI therapy, central venous catheterization
Studies with pneumonia
Chen Wb, 2020	Mean (range) 51 (31–68); 100%	–	Median simplified acute physiology score II (SAPS II): 9.2 (range, 1–20), mean sequential organ failure assessment (SOFA) score: 5.8 (range, 1–13)	NR	0	10	90	CAZ-AVI-related AEs: Blood urea and creatinine increased = 3/10; ALT and AST increased = 2/10; ALP, GGT, and total bilirubin increased = 3/10; thrombocytosis = 1/10; no severe AEs reported	NR/NR
Shi Y, 2021	59.2 ± 19.4; 88.4%	64.1 ± 17.0; 71.0%	CAZ-AVI vs. TGC: sepsis: 6 (14.0%) vs. 6 (9.7%); Charlson’s score: 2 (0–4) vs. 2 (0–4), SOFA score (at infection onset): 7 (3–10) vs. 6 (4–8), APACHE II scores (at infection onset): 12 (10–16) vs. 13 (9–16.75)	CAZ-AVI: 51.2; TGC: 29	CAZ-AVI: 30.2; TGC: 33.9	NR	CAZ-AVI: 74.4; TGC: 33.9	CAZ-AVI vs. TGC: ALT (U/l) 34.7 (20.48, 59.25) vs. 35.70 (20.60, 76.60); TBil (lmol/l) 23.55 (13.13, 44.70) vs. 14.70 (8.00, 55.81); Scr (lmol/l) 69.7 (41.5, 107.68) vs. 87.7 (41.5, 170.40); Fib (g/l) 1.84 (1.51, 2.25) vs. 3.47 (3.04, 4.21); APTT (s) 43.20 (34.48, 53.78) vs. 36.20 (29.60, 41.40); No mention of CAZAVI-related or not	NR/NR
Studies with bacteremia and pneumonia
Vena A, 2020	62 (41–70); 68.3%	–	ICU 17/41 (41.5); sepsis 17 (41.5); septic shock 7 (17.1)	Bacteremia: 86: nosocomial pneumonia: 90	NA	NR	NR	NR	Receiving CRRT during CAZ-AVI treatment/NR
King M, 2017	59 (51–69); 51%	59 (50–69); 70%	Charlson Comorbidity Index (median, IQR)— Monotherapy = 4.5 (2.8–7); Combination = 4.5 (2–7); Pitt bacteremia score: 2 (0–5) for mono and combi both	Bacteremia: 60.87; pneumonia: 56.25	39.13 56.25	NR	82.61 43.75	NR	High Pitt bacteremia score, ICU admission, neutropenia (absolute neutrophil count < 500 at discharge)/NR
Castón JJ, 2020	70 (54–79); 61.7%	–	Septic shock: 25/47 (53.2); SOFA score, median (IQR): 3 (2–6); APACHE II: 14 (9–19); INCREMENT-CPE score > 7, n (%): 27 (57.4%)	Bacteremia: 54.2; pneumonia: 35.7	Bacteremia: 34.78; pneumonia: 35.7	NR	Bacteremia: NR pneumonia: 64.29 (Microbiological response)	Renal failure, n (%) = 2 (4.2); Clostridium difficile colitis = 1 (2.1)	INCREMENT-CPE score > 7, pneumonia/NR
Fang J, 2021	62.5 (52.5–70.5); 67.6%	63 (51–72); Female = 27 (34.6%); Male = 51 (65.4%)	PB vs. CAZ-AVI: Charlson Comorbidity Index 1 (0–2) vs. 2 (1–4), APACHE II score 14.1 ± 5.51 vs. 13.68 ± 5.56; Health care Interventions: Vasoactive Drugs 55 (70.5%) vs. 28 (75.7%); ECMO 1 (1.3%) vs. 0	CAZ-AVI: NR; Comparator: NR	NR	NR	NR	NR	Charlson Comorbidity Index (≥ 3), prior antibiotic use within 90 days/ CAZ-AVI based regimen, length of hospital stay after CRKP infection, creatinine clearance
Castón JJ, 2022	Bacteremia and pneumonia: 64 (55–78) and 70.5 (61–74); 70.8% and 69.5%	Bacteremia and pneumonia: 69 (60–83) and 77 (57–84); 53.8% and 75.0%	CAZ-AVI vs. BAT, n (%): Septic shock [bacteremia—16 (22.2) vs. 6 (15.3); pneumonia- 6 (26.1) vs. 1 (6.3)], SOFA score, median IQR: bacteremia = 4 (2–6) vs. 3 (2–6); pneumonia = 4 (1–9) vs. 3.5 (1.5–5); APACHE II score: bacteremia = 14 (10–19) vs. 12 (9–20); pneumonia = 15 (9–18) vs. 16 (10.5–20); INCREMENT-CPE score > 7: bacteremia = 32 (44.4) vs. 16 (41.0); pneumonia = 14 (60.9) vs. 10 (62.5); Pitt score: bacteremia = 1 (0–3) vs. 3 (0–6); pneumonia = 2 (0–7) vs. 1 (0–3); Charlson Comorbidity Index > 3 [38 (52.8) vs. 28 (71.8); 11 (47.8) vs. 8 (50.0)]	Bacteremia- CAZ-AVI: 90.3, BAT: 76.9; pneumonia: CAZ-AVI: 91.3 BAT: 56.2	Bacteremia- CAZ-AVI: 13.9, BAT: 30.8; pneumonia: CAZ-AVI: 21.7 BAT: 37.5	NR	Bacteremia- CAZ-AVI: 58.3, BAT: 35.9; pneumonia: CAZ-AVI: 43.5 BAT: 18.7	All AEs: 41 (12.1%) patients; treatment with CAZAVI was less related to AEs than BAT (11/189, 5.8% vs. 30/150, 20%; P,0.001); most frequent, CAZ-AVI vs. BAT: renal failure (18/41, 43.9%) = 3/189 (1.6%) vs. 15/150 (10%); diarrhea (9/41, 21.9%) = 5/11 (45.4%) vs. 4/30 (13.3%)	INCREMENT-CPE score > 7, SOFA score at the diagnosis of infection/ CAZ-AVI treatment
Karaiskos I, 2021	Mean (SD) 60.9 (17.1); 74.1%	–	Sepsis 97/147 (66.0); septic shock 50/147 (34.0)	NR	Bacteremia: 17.9; VAP/HAP: 37.8	NR	NR	NR	Charlson Comorbidity Index ≥ 2, rapidly fatal diseases/treatment with CAZ-AVI containing regimens

Caston JJ et al. 2017: ‘Other antibiotic’ comparator group is not defined in the article

Sousa A et al.: 14-day mortality is reported. Chen W et al.: includes one patient with bacteremia. Castón JJ et al. 2022: clinical cure is 21-day clinical response; Caston JJ et al. 2020: % clinical cure is calculated by subtracting 14-day clinical failure from 100. Shields R et al. 2017: mortality is calculated by subtracting survival from total patients

LRTI lower respiratory tract infection, NR not reported

^aReported for patients with mentioned indication, not for the total study population

^bNR specifically for treatment and comparators, presented for total patients

^cReported for total patients

^dIncludes 28-day or 30-day mortality

Clinical cure was evaluated in 11/22 studies. Overall, clinical cure rates were 80.33% (n = 196/244) in patients treated with CAZ-AVI and 54.15% (n = 124/229) in patients treated with other antibiotics.

Microbiological cure occurred in 58.33% (n = 42/72) of patients in the CAZ-AVI group and 35.90% (n = 14/39) of patients in the comparator group, which was in a single study (Table 2) [42]. Resistance to CAZ-AVI was observed in 2.89% (n = 19/656, repeat susceptibility data not available) patients during the therapy, reported across five studies [24, 28, 29, 36, 44].

Sources of bacteremia were reported in six studies, but outcomes could not be differentiated by source. Overall, 10.80–99.30% of patients had bacteremia secondary to a pulmonary source sites [22, 33, 35, 36, 40, 43].

Quantitative Synthesis (Meta-analysis)

Of the total eight studies included in the quantitative synthesis, there was no statistical heterogeneity among included studies (I² = 0%; P = 0.72–0.99), and therefore a common effect model was applied for all outcomes.

Mortality

All-cause 30-day mortality was evaluated in all eight studies (Fig. 3A). Overall, observed 30-day mortality was 15.38% (n = 32/208) in the CAZ-AVI group and 40.12% (n = 199/496) in the comparator group. Study heterogeneity was low (I² = 0%, P = 0.99). Lower odds of 30-day mortality were evident among patients treated with CAZ-AVI versus comparators (Odds ratio [OR] = 0.30, 95% CI 0.19–0.46) (Fig. 3A) [22, 35–38, 40, 42, 43].

Fig. 3.

Meta-analysis of studies reporting 30-day mortality and clinical cure outcomes in patients with bacteremia. I² significance of heterogeneity, CI confidence interval, OR odds ratio, Te number of events observed in the treatment group, TN total number of patients in the treatment group, Ce number of events observed in the comparator group, CN total number of patients in the comparator group

Clinical Cure

Clinical cure was evaluated in five studies (Fig. 3B) [36–38, 40, 42]. Overall, the clinical cure rates were 89.19% (n = 132/148) in the CAZ-AVI group and 54.15% (n = 124/229) in the comparator group (Fig. 3B). Study heterogeneity was low (I² = 0%, P = 0.72). Higher clinical cure rates were evident among patients treated with CAZ-AVI vs. comparators (OR = 4.90, 95% CI 2.60–9.23).

Patients with Nosocomial Pneumonia

Qualitative Analysis

Of a total of eight studies that reported outcomes of patients with nosocomial pneumonia, mortality was reported in 6/8 studies (Table 2). Overall, all-cause 30-day mortality rates were 32.17% (n = 46/143) in patients treated with CAZ-AVI based therapy and 34.62% (n = 27/78) in patients treated with other antibiotics. Clinical cure was evaluated in 5/8 studies. Overall, clinical cure rates were 61.16% (n = 63/103) in patients treated with CAZ-AVI and 34.62% (n = 27/78) in patients treated with other antibiotics.

Microbiological cure was evaluated in five studies with overall rates of 63.21% (n = 67/106) in patients treated with CAZ-AVI and 30.77% (n = 24/78) in the patients treated with other antibiotics. Resistance was not reported in any of the studies.

Quantitative Synthesis (Meta-analysis)

Mortality and Clinical Cure

All cause 30-day mortality and clinical cure was evaluated in two studies (Fig. 4) [39, 42]. The 30-day mortality was 27.27% (n = 18/66) in the CAZ-AVI and 34.62% (n = 27/78) in the comparator group (Table 2). Study heterogeneity was low (I² = 0%, P = 0.47). However, since data was available from two studies with high confidence interval (OR = 0.73, 95% CI 0.35–1.49), no definitive conclusions can be made (Fig. 4A). Clinical cure rates were 65.15% (n = 43/66) in the CAZ-AVI and 34.62% (n = 27/78) in the comparator group. Study heterogeneity was low (I² = 28%, P = 0.24). Higher clinical cure rates were evident among patients treated with CAZ-AVI versus comparators (OR = 3.20, 95% CI 1.55–6.60) (Fig. 4B).

Fig. 4.

Meta-analysis of studies reporting 30-day mortality, clinical cure, and microbiological cure in patients with nosocomial pneumonia. I² significance of heterogeneity, CI confidence interval, OR odds ratio, Te number of events observed in the treatment group, TN total number of patients in the treatment group, Ce number of events observed in the comparator group, CN total number of patients in the comparator group

Overall, in patients with nosocomial pneumonia higher rates of clinical cure were noted among patients treated with CAZ-AVI versus the comparator group; however, no differences were noted for mortality.

Microbiological Cure

Microbiological cure was 63.63% (n = 42/66) in the CAZ-AVI and 30.76% (n = 24/78) in the comparator group, also evaluated in the above two studies [39, 42]. Study heterogeneity was low (I² = 0%, P = 0.55). Higher microbiological cure rates were evident among patients treated with CAZ-AVI versus comparators (OR = 4.95, 95% CI 2.34–10.46) (Fig. 4C).

Renal Failure at Baseline

Renal failure at baseline in patients with bacteremia was present in 26.50% of patients with CAZ-AVI (n = 159/600, four studies) and 19.51% in patients received other antibiotics (n = 16/82, two studies [28, 29, 36, 42]. The 30-day mortality in these studies was 22.17% in the CAZ-AVI and 28.40% in other antibiotics group; clinical cure was 93.65% in the CAZ-AVI and 75.45% for other antibiotics group. For nosocomial pneumonia, renal failure at baseline was reported in a single study (CAZ-AVI: 26.10%, other antibiotics: 12.5%) [42]; 30-day mortality was 21.70% in the CAZ-AVI and 37.50% in other antibiotics group; clinical cure was 91.30% in the CAZ-AVI and 56.20% for other antibiotics group.

Mortality Predictors

Mortality predictors were reported in 19 studies, of which 17 used multivariate analysis [22–24, 26, 28, 29, 31–36, 38, 41–43, 45] and one each used univariate [44] and bivariate analysis [40]. Mortality predictors reported in ≥ 3 studies were Pitt score (median: 3.0–4.0) [26, 33, 40, 43], INCREMENT-CPE score (> 7) [28, 32, 42, 45], neutropenia (absolute neutrophil count < 500) [26, 28, 29, 40], Charlson Comorbidity Index (≥ 2) [29, 41, 44], age [22, 35, 36], and septic shock [28, 29, 40]. Other mortality predictors reported in ≤ 2 studies are presented in Supplemental Fig. 1 and listed in Table 2 in detail. Thirteen mortality predictors, including time from blood culture collection to appropriate therapy, were individually reported in one study each and presented under the ‘other’ category.

Survival Predictors

Of a total of nine studies that reported survival predictors [22, 28, 29, 31, 33, 38, 41, 43, 44], eight reported CAZ-AVI as one of the survival predictors comprising two studies with CAZ-AVI monotherapy, five studies with CAZ-AVI combination therapy, and one with administration of CAZ-AVI via prolonged infusion (lasting ≥ 3 h) [22, 28, 29, 31, 33, 38, 41, 44]. Lower sequential organ failure assessment (SOFA) score at infection onset [38], source control of infection, appropriate empirical therapy [43], length of hospital stay after CRKP infection, creatinine clearance [41], and central venous catheterization [22] were reported in one study each and are presented under ‘other’ category (Supplemental Fig. 1, Table 2).

Adverse Events

Adverse events (AEs) were reported across the infection types and were not separated by type of indication for patients with bacteremia or nosocomial pneumonia. AEs were reported in 12/24 studies (Table 2). CAZ-AVI related AEs were mild diarrhea [25]; encephalopathy [36]; mild blood urea and creatinine increase, ALT and AST increase, ALP, GGT and total bilirubin increase; and thrombocytosis (no severe AEs) [30]; and decrease in creatinine levels [38]. Renal failure unrelated to CAZ-AVI was reported in two and three patients in two studies, respectively [32, 42] and nephrotoxicity in one patient in one study [23]. Significantly fewer AEs were reported in CAZ-AVI group than comparator (CAZ-AVI: 11/189, 5.8% vs. BAT: 30/150, 20%; P = 0.001) [42].

Renal adjustment for CAZ-AVI was reported in four (40%) patients in Chen et al., 2020 study [30]. King et al., 2017 and Sousa et al., 2018, also reported renal adjustment of CAZ-AVI in 33/60 (14/33 of those patients received RRT) and 20/57 (no treatment interruption) of total patients across infection type, respectively [27, 45]. In King et al., 2017, patients trended towards high in-hospital mortality (42% vs. 19% without renal adjustment, P = 0.057) with no drug-related AEs [27].

Quality Assessment

The quality assessment of included studies is presented in Supplemental Table 4. Ten studies assessed treatment with CAZ-AVI and a comparator scored 16–20/24 [22, 35–43] and 14 studies assessed treatment with CAZ-AVI without a comparator scored 8–14/16 [23–34, 44, 45]. The overall risk of bias in the studies was moderate to high. No exclusions were made based on the quality score.

Discussion

This SLR and meta-analysis provides comprehensive insights into the real-world use of CAZ-AVI for the treatment of patients with bacteremia or nosocomial pneumonia. From an analysis of 24 real-world studies (retrospective: 22, prospective: 2), compared to non-CAZ-AVI treatment regimens, CAZ-AVI or CAZ-AVI based regimens demonstrated lower all-cause 30-day mortality and improved clinical cure rates among patients with bacteremia and improved clinical cure rates among patients with nosocomial pneumonia.

The majority of studies that were assessed included patients with severe infections and concomitant conditions/diseases including hematological malignancies, solid organ transplantation, COVID-19, etc. Key infection-causing pathogens were CRE, including CPE, and MDR P. aeruginosa; however, the data for patients infected with P. aeruginosa is limited. A large proportion of reported cases treated with CAZ-AVI also included combinations with other antibiotics.

We performed a meta-analysis to synthesize results from studies with a comparator arm that evaluated patients with bacteremia or nosocomial pneumonia. Our analysis suggested that treatment with CAZ-AVI resulted in lower mortality and improved cure rates compared to non-CAZ-AVI regimens in patients with bacteremia specifically. Furthermore, since the chance of fatality is high in bacteremia [1, 5], receipt of appropriate initial therapy is critical, and thus, patients at high risk for infections due to CRE or MDR P. aeruginosa may be considered candidates for empiric therapy with CAZ-AVI. Clinical cure rates were also improved in our meta-analysis for patients with bacteremia. These findings are in line with prior meta-analyses showing that CAZ-AVI improved 30-day mortality and clinical cure rate when compared with other antibiotics (carbapenems, colistin, etc.); however, those analyses were pathogen-focused rather than indication-focused [16, 17, 46, 47]. Similarly, improved clinical cure rates were also observed in patients with nosocomial pneumonia; however, since data were available from few studies, results should be interpreted with caution. Additionally, a few studies were available to assess other outcomes, including mortality and microbiologic relapse/resistance.

The IDSA panel recommends CAZ-AVI as a preferred treatment option along with other antibiotics for the treatment of complicated urinary tract infection, pyelonephritis, and infections outside of the urinary tract caused by CRE [48]. In addition, if an OXA-48 enzyme is identified, CAZ-AVI is preferred. The current analysis is supportive of IDSA panel recommendation and suggests that CAZ-AVI may be considered as one of the initial treatment options in patients with known or suspected CRE bacteremia.

This review provides information on the real-world use of CAZ-AVI, and this quantitative and qualitative synthesis of available evidence may be useful for clinicians and caregivers in treating such patients. Our results should be considered in light of some limitations. Most included studies (22/24) were retrospective, which are subject to confounding and bias given their nature. Additionally, the outcomes assessed by different studies were not always uniform. While some studies reported mixed data for monotherapy and combination therapy, few reported clinical cure, and others clinical success. Considering the heterogeneity of studies, direct comparisons across studies (except studies included in the meta-analysis) were not possible due to variability in definitions, methods, and outcomes analyzed. Moreover, individual patient-level data were not available, which limits our ability to control for confounders and draw more definitive conclusions. Our analysis was also limited by grouping together patients who received CAZ-AVI alone or in combination with other agents. Moreover, our comparator group included various agents (Table 1), which reflects real-world practice. Finally, outcomes for P. aeruginosa were limited due to low number of studies and therefore the generalizability of these findings is limited.

This meta-analysis also has a number of strengths. First, a focused literature search strategy was applied to identify studies with data on bacteremia and nosocomial pneumonia, and a sufficient number of studies was included. Second, the outcomes used (mortality and clinical cure) can be considered significant in clinical practice and are useful in evaluating the effectiveness of an antibiotic regimen in the treatment of an infection. Third, no between-study heterogeneity was observed for meta-analysis. Heterogeneity is a potential problem in interpreting the results of any meta-analysis and, the absence of heterogeneity demonstrates the reliability of the quantitative results obtained.

Taken together, there are several key insights from this review. Information on the effectiveness of CAZ-AVI in the treatment of severe infections, which are not typically represented in homogenous populations in clinical trials was synthesized. Appropriate use of CAZ-AVI was one of the survival predictors reported in eight studies and on the other hand Pitt score, INCREMENT score (> 7), and neutropenia were prominent mortality predictors. RWE is especially useful to evaluate the changing rates of drug resistance to advanced combination antibiotics such as CAZ-AVI. Ideally these data should be evaluated systematically across studies and determine the impact of antimicrobial stewardship efforts to minimize rates of treatment-emergent resistance. Some studies have used CAZ-AVI as empiric treatment while others were reserving its use for targeted treatment. Choosing to treat aggressively early for better outcomes in patients rather than waiting for culture tests and susceptibility results must be balanced, however, waiting for the targeted treatment might result in worsening clinical outcomes in patients at high risk of death. In vitro surveillance of antimicrobial resistance trends is being monitored by various databases such as ATLAS, however, they are not based on patient outcomes. Clinical outcomes from RWE can provide useful indicators to prescribe appropriate antibiotics. Real-world data are required to fill in this gap, though, there is high heterogeneity in the use of different comparators and differences in outcomes based on which drug or drug combinations CAZ-AVI are being compared with and the conclusion should be interpreted with caution. Future studies are needed to compare CAZ-AVI with other novel β-lactam/β-lactamase inhibitors used to treat CRE and/or MDR P. aeruginosa.

Conclusions

This SLR and meta-analysis synthesized clinical outcome data and presents quantitative and qualitative evidence that may be considered while using CAZ-AVI for the treatment of patients with bacteremia and nosocomial pneumonia caused by non-MBL producing CRE and/or MDR P. aeruginosa. Nevertheless, there is a need for continued robust evidence generation to further fortify the role CAZ-AVI play in improving patients’ outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Ryan K. Shields, Shweta Kamat, Paurus M. Irani, Margaret Tawadrous and Tobias Welte: Conceptualization and investigation; Shweta Kamat: data curation, funding, project administration, resources; Shweta Kamat and Tobias Welte: supervision; Ryan K. Shields, Shweta Kamat, Paurus M. Irani, Margaret Tawadrous, Tobias Welte and Juan P. Horcajada: Methodology; Shweta Kamat, Tobias Welte and Juan P. Horcajada: validation; Ryan K. Shields, Tobias Welte and Juan P. Horcajada: visualization; Ryan K. Shields and Margaret Tawadrous: writing original draft; All authors: Writing-critical review and editing; approval of final version of this manuscript.

Funding

Sponsorship for this study and the journal’s Rapid Service Fee was funded by Pfizer Inc.

Data Availability

This article is based on published literature. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Shweta Kamat and Paurus M. Irani are current employees of Pfizer and may hold stock/stock options with Pfizer. Margaret Tawadrous was an employee of Pfizer when this analysis was conducted. Ryan K. Shields is an employee of University of Pittsburgh and has served as a consultant for Allergan, Cidara, Entasis, GlaxoSmithKline, Melinta, Menarini, Merck, Pfizer, Shionogi, Utility, and Venatorx, and has received investigator-initiated funding from Merck, Melinta, Roche, Shionogi, and Venatorx; Juan P. Horcajada is an employee of Hospital del Mar and University of Pompeu Fabra, Barcelona, Spain and has received honoraria for advisory activities and served as a speaker for Pharma companies; Tobias Welte is an employee of Hannover School of Medicine and has received fees for lectures and served in advisory board for AstraZeneca and Pfizer.

Ethical Approval

This article is based on published literature and does not contain any previously unreported studies with human participants or animals.