Introduction

Antimicrobial resistance among Gram-negative bacteria is a long-standing problem that needs to be monitored in an effort to preserve the efficacy of current antimicrobial agents. Pseudomonas aeruginosa and species of Enterobacterales are common causative pathogens of respiratory infections such as hospital-acquired bacterial pneumonia (HABP) and ventilator-associated bacterial pneumonia (VABP), and respiratory infections caused by antimicrobial-resistant Enterobacterales and P. aeruginosa are associated with higher patient mortality [1, 2]. Among isolates of Enterobacterales and P. aeruginosa, antimicrobial resistance can be mediated by the production of β-lactamases, for example, extended-spectrum β-lactamases (ESBLs), serine carbapenemases and metallo-β-lactamases (MBLs) [3].

Ceftazidime-avibactam is a combination antimicrobial agent comprising ceftazidime, a third-generation cephalosporin, and avibactam, a non-β-lactam β-lactamase-inhibitor. Avibactam inhibits Ambler class A, class C, and certain class D OXA-type β-lactamases, but not MBLs. Therefore ceftazidime-avibactam is active against ESBL- and serine carbapenemase-positive Gram-negative isolates, but not against MBL-positive isolates [4,5,6]. Ceftazidime-avibactam has been approved for the treatment of HABP and VABP, as well as complicated intra-abdominal infection (in combination with metronidazole) and complicated urinary tract infection (including pyelonephritis) [7, 8]. In Europe, ceftazidime-avibactam is also indicated for the treatment of infections due to aerobic Gram-negative organisms in adult patients with limited treatment options [8].

The aim of this study is to report in vitro antimicrobial activity and susceptibility data for a panel of antimicrobial agents against isolates of Enterobacterales and P. aeruginosa collected from respiratory specimens as part of the Antimicrobial Testing Leadership and Surveillance (ATLAS) program (2016–2018). Data on rates of resistant isolates will also be presented. The geographical regions of collection included in the analysis are Africa/Middle East, Asia/South Pacific, Europe and Latin America.

Methods

Isolates from respiratory specimens were collected from hospitalised patients from participating study centers between 2016 and 2018 in four geographical regions (Africa/Middle East, Asia/South Pacific, Europe and Latin America). Only non-duplicate isolates of the organism considered to be the potential causative pathogen of the infection were included in the study. Demographic information recorded for each isolate included specimen source, patient age and sex, and type of hospital setting.

Isolates were collected and identified at the participating center, and pure cultures were shipped to the central laboratory (International Health Management Associates [IHMA] Inc., Schaumburg, IL, USA). The central laboratory then re-identified and confirmed bacterial species using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Biotyper; Bruker Daltonics, Billerica, MA, USA), and performed antimicrobial susceptibility testing using self-manufactured frozen broth microdilution panels [9]. Ceftazidime-avibactam was tested with a fixed concentration of 4 mg/L avibactam with doubling dilutions of ceftazidime. All minimum inhibitory concentrations (MICs) were interpreted using version 10.0 of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) clinical breakpoint tables [10]. Among the Enterobacterales, isolates of Morganella spp., Proteus spp., Providencia spp. or Serratia spp. were excluded from the analysis for colistin activity, due to intrinsic resistance. All isolates of Enterobacterales were included in the analysis for tigecycline activity, but tigecycline EUCAST breakpoints are only available for isolates of Citrobacter koseri and Escherichia coli. For isolates of P. aeruginosa, EUCAST have revised the susceptible breakpoints at the standard dosing regimen (S) for piperacillin-tazobactam, aztreonam, ceftazidime, cefepime, imipenem and levofloxacin [10], therefore isolates tested against these antimicrobial agents are categorized as either susceptible at increased exposure (I), or as resistant (R). For all antimicrobial agents tested against P. aeruginosa, the combined susceptibility rates (susceptible at standard dosing regimen plus susceptible at increased exposure) are presented here.

A multidrug-resistant (MDR) phenotype among isolates of P. aeruginosa was defined as resistance to one or more antimicrobial agent (given in parentheses) from three or more of the following antimicrobial classes: aminoglycosides (amikacin), carbapenems (imipenem or meropenem), cephalosporins (ceftazidime or cefepime), fluoroquinolones (levofloxacin) and β-lactam/β-lactamase inhibitor combinations (piperacillin-tazobactam).

Enterobacterales isolates with MIC values of ≥2 mg/L to meropenem were screened for genes encoding clinically-relevant β-lactamases (ESBLs: SHV, TEM, CTX-M, VEB, PER and GES; plasmid-mediated AmpC β-lactamases: ACC, ACT, CMY, DHA, FOX, MIR and MOX; serine carbapenemases: GES, KPC and OXA-48-like; and MBLs: NDM, IMP, VIM, SPM and GIM) using published multiplex PCR assays [11]. Additionally, isolates of E. coli, Klebsiella pneumoniae, Klebsiella oxytoca and Proteus mirabilis with MIC values of ≥2 mg/L to ceftazidime or aztreonam were screened for the same genes. P. aeruginosa isolates with meropenem MIC values of ≥4 mg/L were screened for genes encoding MBLs (IMP, VIM, NDM, GIM and SPM) and serine carbapenemases (KPC and GES) using published multiplex PCR assays [12]. All detected carbapenemase genes were amplified using flanking primers and sequenced, and sequences were compared against publicly available databases.

The following subsets of resistant isolates are presented: ESBL-positive Enterobacterales; ESBL-positive and carbapenemase-positive Enterobacterales (hereafter described as ESBL/carbapenemase-positive); carbapenemase-positive and MBL-negative Enterobacterales or P. aeruginosa; carbapenemase-positive and MBL-positive Enterobacterales or P. aeruginosa (hereafter described as MBL-positive) and MDR P. aeruginosa.

Results

Isolates collected

A total of 15,460 isolates from respiratory specimens (10,128 Enterobacterales and 5332 P. aeruginosa) were collected from a total of 51 countries in four geographical regions between 2016 and 2018. The number of centers in each participating country and the number of isolates collected by each center are presented in Supplementary Table 1. More than half of isolates were collected in Europe (58.6%; n = 9055), followed by Asia/South Pacific (19.7%; n = 3040); Latin America (13.3%; n = 2061) and Africa/Middle East (8.4%; n = 1304).

Respiratory specimen sources were: sputum, 45.8% (n = 7088); endotracheal aspirate, 27.1% (n = 4190); bronchoalveolar lavage, 13.9% (n = 2146); bronchus, 8.2% (n = 1269). Less than 3% of isolates were classified as unspecified, thoracentesis fluid, lungs, trachea or aspirate.

The majority of isolates from respiratory specimens were collected from male patients (64.5%; n = 9972) and half were from patients aged 65 years and older (50.0%; n = 7723). The percentage of isolates collected from respiratory specimens on general wards (47.9%; n = 7398) was similar to intensive care units (43.4%; n = 6716).

Enterobacterales

Antimicrobial activity and susceptibility data for a panel of antimicrobial agents against the collection of Enterobacterales isolates from respiratory sources are shown by region in Table 1. In each region, the highest rates of susceptibility among the collection of Enterobacterales were to ceftazidime-avibactam, meropenem, colistin and amikacin (≥94.4%). Tigecycline susceptibility in all regions was ≥97.9% among isolates of C. koseri and E. coli, the only species within the Enterobacterales collection to which EUCAST breakpoints apply.

Table 1 Antimicrobial activity against respiratory Enterobacterales isolates by region (ATLAS 2016–2018)
Full size table

Among ESBL-positive Enterobacterales, susceptibility was highest to ceftazidime-avibactam and colistin (≥93.6%) (Table 2). Susceptibility to tigecycline was high among ESBL-positive C. koseri and E. coli (98.8%). The susceptibility among ESBL/carbapenemase-positive Enterobacterales was highest to colistin and ceftazidime-avibactam (≥64.7%). In this subset, 92.9% of C. koseri and E. coli were susceptible to tigecycline. Among carbapenemase-positive/MBL-negative Enterobacterales, susceptibility was highest to ceftazidime-avibactam and colistin (≥74.1%). All carbapenemase-positive/MBL-negative C. koseri and E. coli were susceptible to tigecycline. For MBL-positive Enterobacterales, susceptibility was highest to colistin, and the susceptibility of MBL-positive C. koseri and E. coli to tigecycline was 90.0%. No isolates of MBL-positive Enterobacterales were susceptible to amoxicillin-clavulanate, ceftazidime, ceftazidime-avibactam or cefepime.

Table 2 Antimicrobial activity against resistant respiratory Enterobacterales isolates by subset (ATLAS 2016–2018)
Full size table

By region, the 2016–2018 rate of ESBL-positive Enterobacterales was lowest in Europe and highest in Africa/Middle East (Fig. 1). Fewer than 5% of Enterobacterales isolates collected in each region were ESBL/carbapenemase-positive, carbapenemase-positive/MBL-negative or MBL-positive.

Fig. 1
figure 1

Rates of resistant respiratory Enterobacterales isolates by subset/region (ATLAS 2016–2018). aTotal isolate numbers (2016–2018) by region were: Africa/Middle East, 833; Asia/South Pacific, 2033; Europe, 6006; Latin America, 1256. bIsolates in this subset are carbapenemase-positive and MBL-positive

Full size image

Pseudomonas aeruginosa

Antimicrobial activity and susceptibility data for a panel of antimicrobial agents against the collection of P. aeruginosa isolates from respiratory sources are shown by region in Table 3. The combined susceptibility (susceptible at standard dosing regimen plus susceptible at increased exposure) among the collection of P. aeruginosa in each region was highest to colistin, ceftazidime-avibactam and amikacin (≥82.4%). Rates of susceptibility to ceftazidime-avibactam, meropenem and amikacin were lower in Latin America than in Africa/Middle East, Asia/South Pacific and Europe.

Table 3 Antimicrobial activity against respiratory Pseudomonas aeruginosa isolates by region (ATLAS 2016–2018)
Full size table

Among MDR isolates of P. aeruginosa, susceptibility was highest to colistin, ceftazidime-avibactam and amikacin (≥61.3%) (Table 4). Susceptibility among carbapenemase-positive/MBL-negative P. aeruginosa was highest to colistin and ceftazidime-avibactam (≥71.7%). No isolates of carbapenemase-positive/MBL-negative P. aeruginosa were susceptible to imipenem. All MBL-positive P. aeruginosa isolates were susceptible to colistin, with ≤19.1% susceptible to amikacin or ceftazidime-avibactam.

Table 4 Antimicrobial activity against resistant respiratory Pseudomonas aeruginosa isolates by subset (ATLAS 2016–2018)
Full size table

By region, the 2016–2018 rate of MDR P. aeruginosa was lowest in Asia/South Pacific and highest in Latin America (Fig. 2). The 2016–2018 rate of carbapenemase-positive/MBL-negative P. aeruginosa in each region was ≤3%, with only one such isolate collected in Africa/Middle East and three isolates in Asia/South Pacific. The 2016–2018 rate of MBL-positive P. aeruginosa was lowest in Asia/South Pacific and highest in Latin America.

Fig. 2
figure 2

Rates of resistant respiratory Pseudomonas aeruginosa isolates by subset/region (ATLAS 2016–2018). aAfrica/Middle East, 471; Asia/South Pacific, 1007; Europe, 3049; Latin America, 805. bIsolates in this subset are carbapenemase-positive and MBL-positive

Full size image

Discussion

This study presents in vitro antimicrobial activity and susceptibility data for a panel of antimicrobial agents against respiratory isolates of Enterobacterales and P. aeruginosa, collected as part of the ATLAS program (2016–2018), as well as data on subsets of resistant isolates.

Among the Enterobacterales isolates, rates of antimicrobial susceptibility to amikacin, ceftazidime-avibactam, colistin, meropenem and tigecycline were similar, irrespective of the geographical region of collection. For P. aeruginosa isolates, however, rates of susceptibility to ceftazidime-avibactam, colistin and amikacin were lower in Latin America, when compared with the other regions. In a phase 3 trial of hospitalized adults with HABP or VABP due to Gram-negative pathogens, the overall ceftazidime-avibactam MIC90 against isolates of Enterobacterales (n = 317) was 0.5 mg/L, and against isolates of P. aeruginosa (n = 101) was 8 mg/L [13]. These data are comparable with the MIC90 values for ceftazidime-avibactam in the current study, with the exception of isolates from Latin America, where the ceftazidime-avibactam MIC90 against the collection of P. aeruginosa was 32 mg/L.

The rates of all subsets of resistant P. aeruginosa isolates presented here were higher in Latin America than the other regions, which may explain the lower ceftazidime-avibactam activity and susceptibility seen in Latin America. The rate of MDR P. aeruginosa in Latin America (28.7%) was similar to a 2015–2016 global antimicrobial surveillance study (34.6%), where the rate in Latin America was the highest among seven geographical regions [14]. The 2015–2016 global study included aztreonam and colistin in their MDR study definition [14], whereas the current study definition of MDR P. aeruginosa omitted aztreonam and colistin, based on guidance from the Belgian High Council of Health [15] and the combinations of pathogens and antimicrobial agents under continued European Antimicrobial Resistance Surveillance Network (EARS-Net) surveillance [16]. A 2012–2015 study of clinical P. aeruginosa isolates from Latin America reported that 35.8% (643/1794) of P. aeruginosa isolates were meropenem-nonsusceptible (% intermediate plus % resistant) [17]; similar to the current study (37.0% [298/805]).

In the present study, susceptibility to ceftazidime-avibactam among carbapenemase-positive/MBL-negative Enterobacterales isolates remained high (98.9%). This was comparable to the 2012–2015 rate of ceftazidime-avibactam susceptibility reported among isolates of Enterobacterales, collected from the same regions as the current study, that were OXA-48-positive and MBL-negative (99.2%) [6]. The lack of ceftazidime-avibactam activity against MBL-positive isolates (due to the hydrolysis of both ceftazidime and avibactam by the MBL class of β-lactamases [18]) is clearly demonstrated in the present study. In addition, ceftazidime-avibactam susceptibility was notably lower among ESBL/carbapenemase-positive Enterobacterales (64.7%), compared with ESBL-positive isolates (93.6%). A total of 124 isolates in the ESBL/carbapenemase-positive subset were ceftazidime-avibactam-resistant, of which 123 were MBL-positive and the remaining isolate was carbapenemase-positive/MBL-negative (data not shown). The single carbapenemase-positive/MBL-negative isolate had a ceftazidime-avibactam MIC of 32 mg/L (EUCAST resistance breakpoint, > 8 mg/L). Isolates of Enterobacterales have been found to coproduce MBLs and Ambler class A β-lactamases, such as ESBLs [19].

The limitations of this study were the predefined number of isolates per species, as well as the variability in center and country participation between the study years, meaning that these results cannot be interpreted as epidemiology findings. The details on the type or size of study centers are not recorded at the time of isolate collection, which could limit the clinical relevance of the data. Despite this, the data reported here highlight the presence of antimicrobial-resistant respiratory Gram-negative pathogens in the four geographical regions presented. Ceftazidime-avibactam was active against resistant isolates of Enterobacterales and P. aeruginosa, with the exception of those organisms that were MBL-positive; among this subset of Enterobacterales, susceptibility was highest to colistin. This study provides valuable information to clinicians on the susceptibility of these resistant isolates to antimicrobial agents in current use. Continued monitoring of the antimicrobial susceptibility profiles of respiratory Enterobacterales and P. aeruginosa isolates is necessary to identify the most difficult-to-treat respiratory pathogens and improve patient outcomes.

Conclusions

Among the collection of respiratory Enterobacterales isolates, rates of susceptibility to ceftazidime-avibactam, meropenem, colistin and amikacin were 94.4–99.6% in each region. For the subsets of resistant Enterobacterales, rates of susceptibility to ceftazidime-avibactam, meropenem and amikacin were lowest among MBL-positive isolates of Enterobacterales (0.0–42.7%), whereas colistin susceptibility was 89.0%. At least 90.0% of all Citrobacter koseri and Escherichia coli isolates were susceptible to tigecycline, including all subsets of resistant isolates.

For the collection of respiratory P. aeruginosa isolates, ceftazidime-avibactam, colistin and amikacin susceptibility was 82.4–99.8%. Among all resistant isolates of P. aeruginosa, colistin susceptibility remained ≥98.1%. For MBL-positive P. aeruginosa, antimicrobial susceptibility was lowest to all agents except colistin and aztreonam (0.0–19.1%).

For the majority of agents, antimicrobial susceptibility was reduced among the subsets of resistant Gram-negative isolates; most notably the MBL-positive subset. Monitoring of antimicrobial susceptibility is therefore necessary to help physicians to choose appropriate treatment options and improve the treatment outcomes for respiratory Gram-negative infections.