Ceftazidime – avibactam and ceftolozane tazobactam susceptibility of multidrug resistant Pseudomonas aeruginosa strains in Hungary

Our objective was to compare the activity ceftazidime-avibactam (C/A) and ceftolozane – tazobactam (C/T) against multidrug (including carbapenem) resistant Pseudomonas aeruginosa clinical isolates collected from six diagnostic centers in Hungary and to reveal the genetic background of their carbapenem resistance. Two hundred and fifty consecutive, non-duplicate, carbapenem-resistant multidrug resistant (MDR) P. aeruginosa isolates were collected in 2017. Minimal inhibitory concentration values of ceftazidime, cefepime, piperacillin/tazobactam, C/A and C/T were determined by broth microdilution method and gradient diffusion test. Carbapenem inactivation method (CIM) test was performed on all isolates. Carbapenemase-encoding bla VIM , bla IMP , bla KPC , bla OXA-48-like and bla NDM genes were identi ﬁ ed by multiplex PCR. Of the isolates tested, 33.6% and 32.4% showed resistance to C/A and C/T, respectively. According to the CIM test results, 26% of the isolates were classified as carbapenemase producers. The susceptibility of P. aeruginosa isolates to C/A and C/T without carbapenemase production was 89% and 91%, respectively. Of the CIM-positive isolates, 80% were positive for bla VIM and 11% for bla NDM . The prevalence of Verona integron-encoded metallo-beta-lactamase (VIM)-type carbapenemase was 20.8%. NDM was present in 2.8% of the isolates. Although the rate of carbapenemase-producing P. aeruginosa strains is high, a negative CIM result indicates that either C/A or C/T could be effective even if carbapenem resistance has been observed.


INTRODUCTION
Pseudomonas aeruginosa is a Gram-negative bacterial pathogen that is an important cause of multidrug-resistant healthcare-associated infections [1]. P. aeruginosa has recently shown an increasing resistance rate to carbapenems, making treatment more challenging [2][3][4]. To treat carbapenem-resistant P. aeruginosa, two new combinations of b-lactams with b-lactamase-inhibitors have recently become commercially available. Ceftazidime (CAZ), an older third generation cephalosporin, has been re-issued in combination with the new broad-spectrum b-lactamase inhibitor avibactam in 2015 [5,6]. Ceftolozane, a novel fifth generation cephalosporin, was issued in combination with the older b-lactamase inhibitor tazobactam in late 2014 [7]. Both ceftazidime-avibactam (C/A) and ceftolozane-tazobactam (C/T) have been shown to successfully treat most carbapenem-resistant P. aeruginosa infections [8][9][10].
There are several advantages of the novel drugs avibactam and ceftolozane compared to their older counterparts. Both tazobactam and avibactam inhibit serin blactamases. While tazobactam is ineffective against many known b-lactamases and carbapenemases, avibactam is active against ESBL and AmpC b-lactamases as well as the Klebsiella pneumoniae carbapenemase (KPC) and OXA-48 carbapenemases [11]. Additionally, avibactam binds to blactamases reversibly, allowing it to be recycled and inhibit additional b-lactamases [11]. Likewise, when comparing ceftolozane to CAZ, the newer drug ceftolozane has further advantages including higher penetration despite porin downregulation, resistance to efflux pump mechanisms, and greater stability to AmpC b-lactamases [11][12][13].
Regardless of the advantages of the new drugs avibactam and ceftolozane, neither are effective against most P. aeruginosa strains that produce carbapenemases. [3,13,14]. The metallo-b-lactamase (MBL)-type carbapenemase P. aeruginosa strains are particularly concerning due to their strong capacity to hydrolyze b-lactams, co-resistance with other antimicrobial classes, and their accelerating spread worldwide [3,[13][14][15][16][17]. Early laboratory detection of carbapenemase production can be critical to provide optimal antimicrobial treatment. Carbapenemase production can be evaluated by the carbapenem inactivation method (CIM), a simple and cost-effective test [18]. Multiplex PCR is used to detect various carbapenemase-encoding genes, including the carbapenem-hydrolyzing class D b-lactamase OXA-48, Ambler class A KPC, and the MBL Verona integron-encoded metallo-beta-lactamase (VIM), IMP, and NDM [19].
The purpose of the study was threefold. First, we evaluated the susceptibility of carbapenem-resistant multidrug resistant (MDR) P. aeruginosa isolates to C/A and C/T. Second, we examined the genotype of carbapenemase-producing P. aeruginosa isolates and how carbapenemase production influences C/A and C/T resistance. Third, we sought to determine the clinical value of the CIM test as a tool for determining antibiotic susceptibility.

MATERIALS AND METHODS
A total of 250 consecutive, non-duplicate, multidrug (including carbapenem) resistant P. aeruginosa isolates from six diagnostic centers in Hungary (representing the country) were collected in 2017. The criteria of collection were multidrug resistance (non-susceptible to at least one agent in ≥3 antimicrobial groups: aminoglycosides (testing gentamycin, tobramycin, amikacin), fluoroquinolones (testing ciprofloxacin and levofloxacin) and b-lactams (testing CAZ, cefepime (FEP), piperacillin/tazobactam (P/T)), including both meropenem and imipenem resistance, based on disk diffusion method according to EUCAST guidelines [20]. Bacterial identification was performed by MALDI-TOF MS.
Susceptibility patterns of CAZ, FEP and P/T were also tested by broth microdilution method [21]. Minimal inhibitory concentration (MIC) values of C/A and C/T were determined by gradient diffusion test (Liofilchem). Escherichia coli ATCC 25922 and P. aeruginosa American Tissue Type Collection (ATCC) 27853 were used as control strains.
CIM test was performed on all isolates in the manner described by Zwalul [18]. Briefly, a full 10 mL inoculation loop of bacteria was suspended in 400 mL water, then a disk containing 10 mg meropenem was immersed in the suspension and incubated for 3 hours at 35 8C. After incubation, the disk was removed from the suspension and placed on a Mueller-Hinton agar plate that had already been inoculated with ATCC 29522 E. coli indicator strain. The plate was subsequently incubated at 35 8C overnight. Isolates showing an inhibition zone ≤15 mm around the meropenem disk were defined as carbapenemase producers [2]. The carbapenemase-encoding genes bla VIM , bla IMP , bla KPC , bla OXA-48-like and bla NDM were tested among CIM-positive isolates by multiplex PCR according to Poirel et al. [19].

RESULTS
Results of antimicrobial susceptibility testing are summarized in Table 1. Of the 250 tested isolates, 33.6% (n 5 84) were resistant to C/A and 32.4% (n 5 81) were resistant to C/T. Most of the isolates (97%, n 5 243) showed categorical agreement to C/A and C/T. Two isolates showed susceptibility to C/A but resistance to C/T, and five isolates showed resistance to C/A yet susceptibility to C/T. All strains with discordant susceptibility results were retested to minimize methodical error. Distribution of MIC values of C/A and C/ T are shown in Fig. 1. Only 17 of the 250 isolates (7%) were susceptible to CAZ without avibactam. Sixty-four percent of the CAZ-resistant isolates (149/233) retained susceptibility to C/A. Only seven (2.8%) strains showed susceptibility to P/T. All of these susceptible isolates also showed susceptibility to C/T. Of the 243 P/T resistant isolates, 162 (66%) showed susceptibility to C/T.
Sixty-five of the 250 isolates were classified as carbapenemase producers according to the CIM test results. All but one CIM-positive isolate showed resistance to both C/A and C/T. CIM positivity showed strong correlation with resistance to both tested cephalosporin/b-lactamase inhibitor combinations. For both C/A and C/T resistance, the CIM test showed a positive predictive value (PPV) of 98% and a specificity of 99%. For C/A resistance, the CIM test showed a sensitivity of 76% and negative predictive value (NPV) of 89%. For C/T resistance alone, the CIM test showed a sensitivity of 79% and NPV of 91%.
Of the 65 isolates that tested positive in the CIM test, 80% (n 5 52) were positive for the VIM gene and 11% (n 5 7) were positive for the NDM gene. All seven isolates found to contain NDM genes were resistant to C/A and C/T, with minimal inhibitory concentrations ≥256 mg/L. All VIM-positive isolates conferred resistance to C/A and C/T with MIC values 16 to ≥256 mg/L and ≥256 mg/L, respectively. The remaining six CIM-positive isolates did not contain any of the genes tested.

DISCUSSION
In this study the susceptibility rates of carbapenem-resistant MDR P. aeruginosa to both C/A and C/T are comparable to one another, at 66.4% and 67.6% respectively. Of note, other studies reported higher susceptibility to C/T than C/A: Grupper et al. observed 91% C/T susceptibility and 81% C/A susceptibility among 290 meropenem-nonsusceptible isolates from 34 hospitals in the United States in 2013 and 2014 [22]. Humphries et al. observed 61.8% susceptibility to C/A and 72.5% susceptibility to C/T in a similar study of b-lactam resistant P. aeruginosa, with 36.4% of C/A-resistant isolates still susceptible to C/T [7]. Ceftolozane is believed to have several advantages over CAZ regarding resistance: greater affinity for penicillinbinding proteins produced by P. aeruginosa, better membrane permeability, greater stability against AmpC b-lactamases, and more potency against P. aeruginosa isolates with up-regulated efflux pumps and loss of porins [9,12]. Our results support only modest advantages of C/T over C/A. Of the tested 250 isolates just five were susceptible to C/T but not to C/A. Additionally, two isolates were susceptible to C/A but not to C/T. The previously-mentioned advantageous mechanisms of ceftolozane may be responsible for the strains that were susceptible to C/T but resistant to C/A. For the strains that were susceptible to C/ A but resistant to C/T, differences between the carbapenemase inhibitor potency of avibactam and tazobactam may have an impact. The susceptibility to CAZ among our isolates increased from 7% to 66% when avibactam is included, indicating strong potency of avibactam as a blactamase inhibitor. We do not have a similar comparison to make with ceftolozane and tazobactam, but P/T showed an even higher rate of resistance than that was seen in CAZ without a b-lactamase inhibitor. Prior use of tazobactam as a b-lactamase inhibitor has allowed several decades for resistance mechanisms to develop, and this may be a contributing factor to resistance to C/T but not C/A.
The potency of C/A and C/T in P. aeruginosa isolates that do not produce some form of carbapenemase is consistent with previous studies [8]. For strains that did not produce carbapenemase (tested negative in the CIM test), susceptibility to C/A and to C/T was 89% and 91%, respectively. This susceptibility rate among CIM-negative isolates is much higher than the overall susceptibility rates of 66-68%. Buehrle et al. reported similar patterns of 92% susceptibility to both C/A and C/T among 38 meropenemresistant P. aeruginosa isolates without carbapenemase production [8]. This indicates that the CIM test could be valuable in routine clinical antibiotic susceptibility testing. CIM is a simple and low-cost test that can be completed in less than one day [2,18]. A negative CIM result indicates that either C/A or C/T would likely be effective treatments, even if carbapenem resistance has already been established.
The rate of spread of the VIM-type MBL is alarming, appearing in 20.8% of the total number of isolates. The presence of the NDM in 2.8% of our P. aeruginosa isolates is also a cause for concern, since the first bla NDM-1 positive P. aeruginosa in Hungary was reported in 2019 [23]. While the sample size is small (n 5 7), the isolates expressing NDM showed the strongest possible resistance to all blactam antibiotics that were tested. The rapid spread of VIM and NDM carbapenemases among many species of Gram-negative bacteria over the last decade could soon end many of the currently-available antimicrobial therapeutic options, including the newer treatments C/A and C/T. Only more toxic antimicrobials such as colistin appear to maintain potency against these multidrug or extensively drug-resistant P. aeruginosa isolates, although resistance against even these antimicrobial classes is beginning to develop [15,17]. Although the rate of carbapenemase-producing P. aeruginosa strains is high according to our study, a negative CIM result indicates that either C/A or C/T could be effective even if carbapenem resistance has been observed. As P. aeruginosa is known to quickly gain antibiotic resistance via several mechanisms besides carbapenemase production, further testing of susceptibility patterns to C/A and C/T should be performed regularly. As the rate of resistance rises, there is an urgent need to develop new and safe classes of antimicrobial therapy. Until novel agents have been developed, strong infection control measurements are essential to protect patients from MDR P. aeruginosa infections.
Ethics approval: Not required.
Funding: The C/A gradient diffusion tests were provided by Pfizer. The C/T gradient diffusion tests were provided by MSD.
Authors' contributions: KK, EJ conceived the experiments. DO, EJ, KK performed the testes, analyzed the data. DO wrote the manuscript. DO, EJ, KK corrected the manuscript. AT, KK, EU, SzM, JSz collected the strains according to certain criteria. All authors read and approved the final manuscript.
Competing interests: None declared.

Consent for publication: Not applicable.
Availability of data and material: The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.