In vitro activity of ceftazidime/avibactam against clinical isolates of Enterobacterales and Pseudomonas aeruginosa from Middle Eastern and African countries: ATLAS global surveillance programme 2015–18

Abstract Objectives To assess the in vitro activity of ceftazidime/avibactam against a recent, 2015–18, collection of clinical isolates of Gram-negative bacilli from Middle Eastern and African countries with a focus on isolates from ICUs and with MDR and difficult-to-treat resistance (DTR) phenotypes. Methods Antimicrobial susceptibility testing of 4608 isolates of Enterobacterales (997 isolates from ICU patients) and 1358 isolates of Pseudomonas aeruginosa (374 isolates from ICU patients) was performed by CLSI broth microdilution methodology in a central laboratory. MICs were interpreted using both CLSI (2020) and EUCAST (2020) MIC breakpoints. Results Most isolates of Enterobacterales (Middle East: ICU, 99.1% susceptible, non-ICU, 99.1%; Africa: ICU, 96.9% susceptible, non-ICU, 98.3%) and P. aeruginosa (Middle East: ICU, 93.4%, non-ICU, 92.1%; Africa: ICU, 89.8%; non-ICU, 94.1%) were susceptible to ceftazidime/avibactam. Applying CLSI and EUCAST breakpoints, MDR rates were similar for Enterobacterales (27.8%–36.0% of isolates) and P. aeruginosa (25.0%–36.4%) while DTR rates were lower for Enterobacterales (1.6%–1.8%) than for P. aeruginosa (5.2%–7.4%). Percentage susceptible rates for ceftazidime/avibactam for MDR Enterobacterales were 96.8%–97.5% (Middle East) and 92.5%–94.3% (Africa) while rates for P. aeruginosa were 70.1%–80.0% (Middle East) and 69.5%–78.2% (Africa). 60.5%–65.8% (Middle East) and 38.9%–52.2% (Africa) of isolates of Enterobacterales with DTR phenotypes were ceftazidime/avibactam susceptible as were 29.2%–31.1% (Middle East) and 28.2%–35.8% (Africa) of DTR P. aeruginosa. Conclusions Overall, the isolates of Enterobacterales and P. aeruginosa tested from Middle Eastern and African countries were highly susceptible to ceftazidime/avibactam. Most MDR and many DTR isolates of Enterobacterales and P. aeruginosa were susceptible to ceftazidime/avibactam.


Introduction
It is important to identify clinical isolates of Gram-negative bacilli (GNB) with resistance determinants and MDR phenotypes that limit empirical and first-line therapeutic options, particularly in ICUs. MDR is frequently defined using criteria established by Magiorakos et al., 1 that is, isolates non-susceptible (intermediate or resistant) to at least one agent in three or more antimicrobial categories. More recently, Kadri et al. 2 identified a more stringent phenotypic category termed difficult-to-treat resistance (DTR) that focuses on treatment-limiting non-susceptibility (intermediate or resistant) to all first-line agents (all b-lactams, including carbapenems, and fluoroquinolones). DTR has been associated with increased patient mortality/treatment failure and requires clinicians to use other potentially less effective or more toxic agents such as aminoglycosides, tigecycline and polymyxins. 2 To date, only two surveillance studies have published regionspecific data describing GNB isolates from Middle Eastern and African countries tested against ceftazidime/avibactam. 3,4 Both of these studies grouped Middle Eastern and African countries together and did not provide information describing the activity of ceftazidime/avibactam against ICU isolates or isolates with MDR or DTR phenotypes. Other publications from Middle Eastern and African countries describing the activity of ceftazidime/avibactam against GNB isolates only include case reports (seven cases in total). 5,6 One of these reports, a case series from a tertiary-care centre in Saudi Arabia, reported that five of six patients infected with carbapenem-resistant Enterobacterales or Pseudomonas aeruginosa achieved both clinical and microbiological cure when treated with ceftazidime/avibactam. 5 The current study intended to evaluate the in vitro activity of ceftazidime/avibactam against Enterobacterales and P. aeruginosa isolates, gathered in 2015-18, from Middle Eastern and African countries, with a focus on ICU and non-ICU patient isolates with MDR and DTR phenotypes to assess its potential benefit against these resistant isolate subsets.

Bacterial isolates
Bacterial isolates tested in the current study were collected as a part of the ATLAS global surveillance programme by laboratories in 12 medical centres in four Middle Eastern countries (six in Israel, two in Jordan, three in Kuwait, one in Saudi Arabia) and 13 medical centres in three countries in Africa (four in Morocco, three in Nigeria, six in South Africa) from 2015 to 2018. Isolates were from bloodstream, intra-abdominal, respiratory tract, skin and soft tissue and urinary tract infection specimen sources and comprised 4608 isolates of Enterobacterales (997 ICU isolates, 3611 non-ICU isolates) and 1358 isolates of P. aeruginosa (374 ICU isolates, 984 non-ICU isolates) (

Statistical analysis
The v 2 statistic with Yates correction (XLSTAT version 2019.1.3) was used to establish statistical significance (P , 0.05) between categorical variables.

Ethics
Ethical approval was not required.
Percentages of isolates of individual species from ICU and non-ICU patients were largely similar between sites in the Middle East and Africa (Table S1). However, ICU and non-ICU isolates of individual species of Enterobacterales and P. aeruginosa from Middle Eastern and African countries demonstrated significant differences in percentage susceptibility for agents other than ceftazidime/ avibactam by both CLSI or EUCAST breakpoints (Tables S3 and S4).
CLSI breakpoints identified more isolates of Enterobacterales and P. aeruginosa as MDR and DTR than EUCAST breakpoints with the notable exception of P. aeruginosa from the Middle East for which both CLSI and EUCAST breakpoints identified 38 isolates as DTR (Table 2). Ceftazidime/avibactam inhibited most isolates (92.5%-97.5%) of MDR Enterobacterales and 69.5% to 80.0% of MDR P. aeruginosa from Middle Eastern and African countries at its susceptible MIC breakpoint (MIC 8 mg/L). Many isolates of DTR Enterobacterales (38.9%-65.8%) and DTR P. aeruginosa (28.2%-35.8%) were also susceptible to ceftazidime/avibactam.
Using CLSI MIC breakpoints, MDR rates among species of GNB from both Middle Eastern and African countries combined ranged from 8.2% for Serratia marcescens to 63.2% for M. morganii and DTR rates ranged from 0.1% for Escherichia coli to 7.4% for P. aeruginosa ( Figure S1). Using EUCAST MIC breakpoints, MDR rates ranged from 7.5% for S. marcescens to 44.2% for Providencia stuartii and DTR rates ranged from 0.1% for E. coli to 5.2% for P. aeruginosa. For all isolates of Enterobacterales, rates of MDR were 20 times greater than DTR using CLSI MIC breakpoints and 17 times greater than DTR using EUCAST MIC breakpoints. For P. aeruginosa, rates of MDR were 5 times greater than DTR using both CLSI and EUCAST MIC breakpoints.
For Enterobacterales, using CLSI MIC breakpoints, MDR phenotypes were 14 times more common than DTR phenotypes in ICU isolates and 23 times more common in non-ICU isolates ( Figure  S2). For Enterobacterales, using EUCAST MIC breakpoints, MDR phenotypes were 13 times more common than DTR phenotypes in ICU isolates and 19 times more common in non-ICU isolates. For Enterobacterales both MDR phenotypes and DTR phenotypes were significantly more common (P , 0.05) for ICU than non-ICU isolates using both CLSI and EUCAST MIC breakpoints. For P. aeruginosa, MDR phenotypes were 5 times more common than DTR phenotypes in both ICU and non-ICU isolates using both CLSI and EUCAST MIC breakpoints. For P. aeruginosa, the differences in percentage of MDR or DTR phenotypes among ICU and non-ICU isolates were not significant (P . 0.05) using either CLSI or EUCAST MIC breakpoints.
For Enterobacterales, using CLSI MIC breakpoints, MDR phenotypes were 12 times (blood) to 33 times (urinary tract) more common than DTR phenotypes ( Figure S3). For Enterobacterales, using EUCAST MIC breakpoints, MDR phenotypes were 11 times (blood) to 25 times (urinary tract) more common than DTR phenotypes. For Enterobacterales, the percentage of isolates with MDR phenotypes and DTR phenotypes were both significantly different (P , 0.05) among specimen sources using both CLSI and EUCAST MIC breakpoints. Blood isolates had the highest percentage of isolates with both MDR and DTR phenotypes (P , 0.05).

Karlowsky et al.
For P. aeruginosa, MDR phenotypes were 4 to 8 times more common than DTR phenotypes across the five specimen sources using both CLSI and EUCAST MIC breakpoints. Differences in the percentage of MDR or DTR phenotypes of P. aeruginosa isolates across the five specimen sources were not significantly different (P . 0.05) using either CLSI or EUCAST MIC breakpoints. For P. aeruginosa, blood isolates had the lowest percentage of isolates that were MDR and DTR.

Discussion
The Using CLSI or EUCAST MIC breakpoints, MDR rates were up to 250 times higher than the corresponding DTR rates for the same collections of isolates (e.g. E. coli, Figure S1). This observation suggests that many MDR phenotypes identified for Enterobacterales include antimicrobial agents not considered firstline agents (i.e. b-lactams and fluoroquinolones) and may be of less importance in terms of impact on patient care, treatment options or public health. Published studies describing DTR isolates are currently limited and describe primarily bacteraemia isolates. 2,10-12 Rates of DTR have ranged from ,1% to 1.4% for Enterobacterales and from 2.3% to 9.0% P. aeruginosa in studies published by investigators in the United States, Italy and Korea, 2,10-12 and are comparable with the rates observed in the current study.
Avibactam, a non-b-lactam diazabicyclooctane inhibitor of Ambler class A b-lactamases, including ESBLs and KPCs, class C (AmpC) b-lactamases and some class D (OXA-48) b-lactamases, restores activity to ceftazidime in most isolates of Enterobacterales and P. aeruginosa that carry these b-lactamases. 3,[13][14][15] Ceftazidime/avibactam also inhibits clinical isolates of P. aeruginosa that are carbapenem resistant because of a combination of porin loss or upregulated antimicrobial agent efflux and elevated production of Pseudomonas-derived cephalosporinase (PDC; intrinsic AmpC). 14 Region-specific prevalence of carbapenem resistance mechanisms should be considered when evaluating empirical treatment options. Previous studies reported that among carbapenem-resistant Enterobacterales, KPC was uncommon in Middle Eastern countries, except Israel, and that carbapenem-resistant Enterobacterales commonly carry NDM and OXA-48-like carbapenemases. 6,16,17 Carbapenemase-producing Enterobacterales in Saudi Arabia have been mainly associated with acquisition of NDM and OXA-48-like carbapenemases and rarely with KPCs. 16 In conclusion, Enterobacterales with DTR phenotypes were uncommon (1.6%-1.8% of isolates) in Middle Eastern and African countries in 2015-18 while MDR isolates were frequently identified (27.8%-36.0% of isolates). MDR P. aeruginosa (25.0%-36.4%) were also commonly observed. A DTR phenotype was three to four times more common among P. aeruginosa (5.2%-7.4%) than Enterobacterales. Ceftazidime/avibactam retained in vitro activity against the majority of MDR and many DTR isolates of Ceftazidime/avibactam surveillance in the Middle East and Africa JAR Enterobacterales and P. aeruginosa. Ceftazidime/avibactam is an important treatment option for infections caused by resistant GNB that do not carry metallo-b-lactamases, particularly Enterobacterales. Increases in infections caused by DTR isolates of GNB will pose major treatment challenges.