Evaluation of Etests and the disk diffusion method for assessment of the activity of ceftazidime-avibactam against Enterobacterales and Pseudomonas aeruginosa in China


 Background: Gram-negative bacilli, particularly Enterobacterales and Pseudomonas aeruginosa, often acquire antimicrobial resistance. Ceftazidime-avibactam was approved for use in China in 2019. However, currently available commercial antimicrobial susceptibility test kits have not yet been developed. Here, we evaluated the Etest and disk diffusion method for assessment of the efficacy of ceftazidime-avibactam against Enterobacterales and P. aeruginosa in China. Results: In total, 194 Enterobacterales and 77 P. aeruginosa isolates, which were divided into a random selection group (140 Enterobacterales and 54 P. aeruginosa isolates) and a stock group (46 Enterobacterales and 31 P. aeruginosa isolates), were assessed by the Etest, disk diffusion, and broth microdilution (BMD) methods. Minimum inhibitory concentrations (MICs) and zone diameters were interpreted according to the CLSI M100 30th edition. For all 271 Enterobacterales and P. aeruginosa isolates, no very major errors were found using Etests. The overall categorical agreement rates (CA%) of Etests for Enterobacterales and P. aeruginosa were 99.5% (193/194) and 96.1% (74/77), respectively. The overall essential agreement rates (EA%) of Etests for Enterobacterales and P. aeruginosa were 95.9% (186/194) and 94.8% (73/77), respectively. In both the random selection and stock groups, EA% and CA% values of Etests exceeded 90%. Overall CA% values of the disk diffusion method for Enterobacterales and P. aeruginosa were 98.5% (191/194) and 93.5% (71/77), respectively. There was no linear relationship between zone diameter and BMD MIC. Conclusions: For Enterobacterales and P. aeruginosa, Etests and the disk diffusion method could have better performance as alternative methods to meet the needs of clinical treatment interpretation. Application of the disk diffusion method in Enterobacterales was superior to that in P. aeruginosa.


Background
Gram-negative bacilli, particularly carbapenem-resistant Enterobacterales (CRE) and Pseudomonas aeruginosa, exhibit major antimicrobial resistance worldwide, including in European countries, the United States of America (USA), and China [1][2][3]. The approval of ceftazidime-avibactam for clinical use in Europe and the USA has brought new treatment options to CRE-infected patients, particularly those with serine-carbapenemase resistance mechanisms [4,5]. Ceftazidime-avibactam was approved for use in China in 2019. However, currently available commercial antimicrobial susceptibility test kits have not yet been developed for analysis of ceftazidime-avibactam resistance in China, and in patients with infections caused by CRE, which are resistant to multiple antimicrobials, the results of ceftazidime-avibactam susceptibility tests are urgently needed to facilitate appropriate targeted treatment. Although ceftazidime-avibactam has excellent in vitro activity against carbapenem-resistant Klebsiella pneumoniae (CRKP) and Pseudomonas aeruginosa strains, a small number of resistant strains will still appear during treatment [6][7][8]. Therefore, susceptibility test results for ceftazidime-avibactam are even more critical.
In most of the laboratories in China, performing the standard broth microdilution method (BMD) is challenging. Therefore, other rapid, simple methods are required as alternatives to determine ceftazidime-avibactam susceptibility. In this study, we evaluated two antimicrobial susceptibility test methods for ceftazidime-avibactam, i.e., the Etest method and the disk diffusion method with the standard BMD, to evaluate whether these easy-to-use methods could replace standard methods when required in the clinical setting.

Materials And Methods Groups
Isolates were divided into two groups, i.e., the random selection group and the stock group. For the random selection group, we randomly selected 140 Enterobacterales and 46 Pseudomonas aeruginosa strains from clinical nonrepeated isolates obtained at Peking University People's Hospital. three Providencia stuartii, two Providencia rettgeri, two K. aerogenes, and two Citrobacter koseri strains. For the stock group, we selected 54 strains of Enterobacterales from 15 hospitals in the CRE China-Network from January 2015 to October 2019 and requested that the minimum inhibitory concentration (MIC) of ceftazidime-avibactam be between 2 and 256 µg/mL. Among these strains, six strains (11.1%) showed MICs for ceftazidime-avibactam of between 8 and 16 µg/mL, and 15 strains (27.8%) showed MIC between 4 and 32 µg/mL. The carbapenem-resistance genes present in these strains were elucidated in previous studies [9]. The 54 strains of Enterobacterales used in this study included 29 K. pneumoniae (18 strains having the bla KPC gene and 10 strains having the bla NDM gene), 12 Escherichia coli (two strains having the bla KPC gene and seven strains having the bla NDM gene), eight Enterobacter cloacae (one strain having the bla KPC gene, five strains having the bla NDM gene, one strain having the bla IMP gene, and one strain having bla VIM gene), three K. oxytocai (two strains having the bla IMP gene and one strain having the bla NDM gene), and two Citrobacter freundii (one strain having the bla IMP gene and one strain having the bla NDM gene).
We selected 31 strains of Pseudomonas aeruginosa from the eight hospitals involved in the CARES 2018 project as stock group strains and required the MIC of ceftazidime-avibactam to be between 2 and 256 µg/mL. Among these strains, 12 strains (38.7%) had MICs for ceftazidime-avibactam between 8 and 16 µg/mL, and 25 strains (80.6%) had MICs between 4 and 32 µg/mL. All strains were removed from a -80 °C ultra-low temperature refrigerator and transferred to Columbia blood agar twice before antimicrobial susceptibility testing.

Antimicrobial susceptibility testing
For the disk diffusion method, ceftazidime-avibactam disks were obtained from Oxoid (Thermo Scientific, China) The content of ceftazidime-avibactam in each disk was 30 µg/20 µg. The operation process was strictly tested following the requirements of the CLSI standardized method [10].
The MH agar used for both the disk diffusion method and the Etest gradient diffusion method for antimicrobial susceptibility testing was obtained from Oxoid (Thermo Scientific Inc).
The BMD was performed strictly following CLSI guidelines [11]. Ceftazidime and avibactam powers were obtained from Pfizer (USA). The concentration ranged from the standard double dilution of 0.016 µg/mL to 256 µg/mL. The MICs and zone diameters of ceftazidime-avibactam for Enterobacterales and Pseudomonas aeruginosa were interpreted according to CLSI M100 30th edition [11]. Briefly, MICs of less than or equal to 8/4 µg/mL or a zone diameter of greater than or equal to 21 mm indicated that the strain was susceptible, whereas MICs of greater than or equal to 16/4 µg/mL or a zone diameter of less than or equal to 20 mm indicated that the strain was resistant.
Essential agreement (EA) indicated that the difference between the MIC value measured by Etest and the BMD did not exceed one dilution factor. Categorical agreement (CA) indicated that the BMD method was consistent with the classification results from the disk diffusion method and Etest method based on the same CLSI breakpoints.
Very major errors (VMEs) indicated that the strain was susceptible by Etest or the disk diffusion method but resistant by the BMD. Major errors (MEs) indicated that the strain was susceptible by the BMD but resistant by Etest or the disk diffusion method.

Result
Etest versus the BMD the overall CA rate was 99.5%, and the overall EA rate was 95.9%. The CA rate of the stock group was 90.7%, and the CA rate in the random selection group was 97.9%. When comparing Etest results with BMD results, we found that the MICs of eight strains exceeded the two dilution factors. As shown in Fig. 1, the Etest MICs of 102 (52.5%) strains were consistent with the BMD MICs. The Etest MICs of 71 (36.6%) strains were a dilution multiple higher than the BMD MICs. Only 12 (6.2%) strains had Etest MICs that were a dilution multiple lower than the BMD MICs. For 77 Pseudomonas aeruginosa strains, no VMEs were found using the Etest method. As shown in Table 1, the overall CA rate was 96.1%, and the overall EA rate was 94.8%. The CA rate in the stock group was 90.3%, and the CA rate in the random selection group was 100%. As shown in Fig. 2, the Etest MICs of 45 (58.4%) strains were consistent with the MICs obtained by the BMD. The Etest MICs of 22 (28.6%) strains were one dilution higher than those obtained by the BMD, whereas those of six (7.8%) strains were one dilution lower than the MICs obtained by the BMD. For one strain, the Etest when the Etest method was used.

Disk diffusion method versus BMD
A comparison of the disk diffusion method and BMD results for 194 Enterobacterales strains (Fig. 3) showed that there were no linear relationships between zone diameter and MIC. The overall CA rate in the 194 Enterobacterales strains was 98.5%. Two VMEs and one ME were found using the disk diffusion method in the stock group; all were for K. pneumoniae strains. No VMEs or MEs were found using the disk diffusion method in the random selection group. There were 22 strains of Enterobacterales with zone diameters of 2 mm (19-22 mm susceptibity breakpoint). Forty-one of the 45 resistant strains obtained by BMD had zone diameters in the range of 13-20 mm.
In a comparison of the results of the disk diffusion method and BMD for 77 Pseudomonas aeruginosa strains, the number of MEs was four, and the number of VMEs was 1. As shown in Table 1, the overall CA rate was 93.5%. Moreover, as shown in Fig. 4, when the zone diameter was 20 mm, the MICs obtained by the BMD method were 4, 8, 16, or 64 µg/mL.

Discussion
In the past ten years, the incidence of CRE, particularly CRKP, has increased significantly in China.
The latest China Antimicrobial Resistance Surveillance System (CARSS) data show that the incidence of CRKP nationwide is as high as 10.1% (http://www.carss.cn/Report/Details?aId=648). However, few active antibacterial agents, such as tigecycline and colistin, are available to treat CRKP in the clinical setting [12,13], resulting in high mortality worldwide. Previous molecular epidemiological data show that more than 70% of CRE isolated in China from 2012 to 2016 produce KPC-type carbapenemases [9]. Ceftazidime-avibactam, a drug with potent antibacterial activity against serine-carbapenemase, was approved for use in China in 2019. Recent retrospective studies have shown that ceftazidimeavibactam treatment of CRKP and Pseudomonas aeruginosa infection in patients who have undergone solid organ transplantation improves clinical success rates [14,15]. Despite these promising findings, ceftazidime-avibactam antimicrobial susceptibility test results are essential for clinical use of this treatment. However, no ceftazidime-avibactam combination kits are available for assessing resistance in China.
To the best of our knowledge, this is the first study comparing Etests, the disk diffusion method, and the BMD with regard to detection of ceftazidime-avibactam susceptibility in China. Compared with the standard BMD method, no VMEs were found using the Etest method. The results of Etest MICs and BMD MICs were reasonably well correlated for both Enterobacterales and Pseudomonas aeruginosa.
The overall of EA% of 271 tested strains was 95.6%. Compared with the BMD method, the Etest method exhibited an excellent linear correlation, supporting the use of this approach as an alternative to the standard clinical method without considering economic costs. This result was similar to the previous research findings of other authors [16]. The disk diffusion method is easy to implement in the clinical setting from an economic standpoint. The CA% values of the ceftazidime-avibactam disk diffusion method against Enterobacterales and Pseudomonas aeruginosa were 98.5% and 93.5%, respectively, similar to the findings of Shields and other scholars [16,17]. Indeed, our results showed that application of the disk diffusion method was more appropriate for Enterobacterales than for Pseudomonas aeruginosa. Notably, however, the disk diffusion method did not have an excellent linear correlation with BMD. Therefore, the zone diameters of strains near the breakpoints (± 2 mm) should be checked by the BMD carefully, consistent with other research results [17,18]. This should also be considered when users refer to the new version of CLSI M100, which suggests using the disk diffusion method [11].