Performance Evaluation of BD Phoenix NMIC-413 Antimicrobial Susceptibility Testing Panel for Imipenem, Meropenem, and Ertapenem Against Clinical Carbapenem-Resistant and Carbapenem-Susceptible Enterobacterales

Purpose: The infection of carbapenem-resistant Enterobacterales (CRE) has become a major clinical and healthcare problem worldwide. The screening methods of CRE have been extensively developed but still need improving [e.g., tests with accurate and simple minimum inhibitory (MICs)]. In this study, the performance of the BD Phoenix NMIC-413 AST panel was evaluated against clinical CRE and carbapenem-susceptible Enterobacterales (CSE) in China. The panel was first evaluated in the Chinese clinical lab. Methods: Antimicrobial susceptibility testing of 303 clinical Enterobacterales isolates were conducted by broth microdilution (BMD), Phoenix NMIC-413 AST panel, and disk diffusion method for imipenem, ertapenem, and meropenem. Considering BMD is a gold standard, essential agreement (EA), categorical agreement (CA), minor error (MIE), major error (ME), and very major error (VME) were determined according to CLSI guidelines. CA and EA > 90%, ME <3%, and VME <1.5% were considered as acceptable criteria. Polymerase chain reaction and sanger sequencing were performed to determine the β-lactamase genotypes of CRE isolates. Results: Three hundred and three isolates included 195 CREs and 108 CSEs were enrolled according to the BMD-MIC values of three carbapenems. Tested CREs showing 100 blaKPC−2-positive organisms, 31 blaIMP-positive organisms, 28 blaNDM-positive organisms, 5 blaVIM-positive organisms, 2 both blaIMP and blaVIM-positive organisms, 2 blaOXA−48-positive organisms, and 27 isolates without carbapenemase genes. For the Phoenix NMIC-413 method, CA and EA rates >93%, MIE rates <5%, ME rates <1.75%, and VME rates were 0%, across the three drugs. For the disk diffusion method, the CA rates for three drugs were all >93%, while the MIE and ME rates were all <5 and <3%, respectively. VME rate was 3.28% for imipenem, exceeded the cut-off value specified by CLSI M52, 0 and 0.56% for ertapenem and meropenem, separately. Conclusion: Based on the genomic data, the detection of CRE and CSE was more reliable using the BD Phoenix NMIC-413 panel compared to the BMD and disk approaches. Therefore, our study supports the use of BD Phoenix NMIC-413 panel as a suitable alternative to BMD for the detection of carbapenem resistant isolates in a clinical setting.


INTRODUCTION
Carbapenem-resistant Enterobacterales (CRE) is a major clinical and public health issue worldwide, which can cause infections associated with high mortality and have limited treatment options (1,2). CREs are generally resistant to all βlactams, including carbapenems such as imipenem, meropenem, ertapenem, doripenem (3), and other antibiotics such as cephalosporins, quinolones, and aminoglycosides, which further restrict the choice of antibiotic treatment. After the initial report of KPC-1 (Klebsiella pneumoniae carbapenemase-1) from a strain of K. pneumoniae discovered in North Carolina in 2001 (4), CRE has been widely reported in almost every state (1). In China, the incidence of carbapenem-resistant Escherichia coli and K. pneumoniae increased from 0 and 0.7% in 2004 to 1.0 and 13.4% in 2014 (5,6).
The mechanism of Carbapenem resistance can be divided into two types: carbapenemase factor and carbapenemase-nonproducing factor. Carbapenem resistance in Enterobacterales is mainly mediated by the horizontal transfer of genes encoding carbapenemases, although porin mutations or overexpression of efflux pumps can lead to carbapenem resistance, especially in combination with the hyperproduction of β-lactamase (7). Carbapenemases consisted of different molecular classes: A, B, and D of the Ambler classification (8). The clinically most important and frequent carbapenemases in Enterobacterales are class A (KPCs), class B metallo-β-lactamases (VIM, IMP, and NDM), and class D (OXA-48) subgroups and their variants (9)(10)(11)(12).
Recently, different methods were developed to detect CRE, such as the disk diffusion method, Brilliance TM CRE Agar, chromID Carba, and molecular methods (13,14). However, phenotypic antimicrobial susceptibility assay which can accurately determine minimum inhibitory concentrations (MICs) is still the key method to guide clinical medication quickly and precisely. Most of the products in the market used to measure the MICs were based on the broth microdilution (BMD) or improved BMD method, such as BioMérieux VITEK 2, Beckman Coulter MicroScan WalkAway, and BD Phoenix. BD Phoenix NMIC-413 panel is a new panel that has recently been marketed, and covered the main cephalosporins and carbapenems such as imipenem, meropenem, and ertapenem. However, its performance to detect carbapenem susceptibility was not well-evaluated yet. In this study, the performance of the BD Phoenix NMIC-413 panel were evaluated using carbapenem-resistant and carbapenem-susceptible clinical isolates in China.

BD Phoenix System
the BD Phoenix NMIC-413 panel (BD Catalog Number: 448442) was used to determined MICs of imipenem (range: 0.25-8 mg/L), meropenem (range: 0.125-8 mg/L), and ertapenem (range: 0.25-2 mg/L) according to the manufacturer's instructions. In short, the identification broth was regulated with bacterial colonies from Blood Agar Plates to 0.5 McFarland standard using a spectrophotometric device. Transferred 25 µL 0.5 McFarland identification broth suspension to the Phoenix Antibiotic susceptibility testing (AST) broth, which was supplemented with 50 µL of the Phoenix AST indicator for the organism growth detecting before added to the panels. The panels were loaded into the Phoenix device (M50). The results were analyzed using Epicenter data management software version 6.61A (BD Diagnostic Systems) after 16 h of incubation (15,16).

BMD Method
The susceptibility of strains to imipenem (range: 0.12-128 mg/L), meropenem (range: 0.12-128 mg/L), and ertapenem (range: 0.12-128 mg/L) were tested and analyzed. Three antimicrobial powders were obtained from National Institutes for Food and Drug Control (Beijing, China). A 0.5 McFarland standard suspension was prepared and used to inoculate the reference BMD panel according to CLSI M100 (17). Incubated these panels at 35 • C for 16-20 h. Clinical carbapenem breakpoints for susceptibility/resistance were ≤1/≥4 mg/L for imipenem and meropenem and ≤0.5/≥2 mg/L for ertapenem.

Disk Diffusion Method
Disk diffusion tests for imipenem, meropenem, and ertapenem (Oxoid, Basingstoke, United Kingdom) were carried out according to CLSI M2 (18). The content of three antimicrobials in each disk was 10 µg. The disk was pasted to the MH agar plate using sterile tweezers and inoculated with 0.5 McFarland standard suspension. Incubated these plates at 35 • C for 16-20 h, the diameter of the inhibition zone was measured with a vernier caliper. The zone diameter ≥23 mm indicated that the strain was susceptible to imipenem and meropenem and ≥22 mm to ertapenem, whereas the zone diameter ≤19 mm indicated that the strain was resistant to imipenem and meropenem and ≤18 mm to ertapenem.

Screening of Carbapenemase Genes
Polymerase chain reaction (PCR) and Sanger sequencing were used to screen out carbapenemase genes, including KPC, NDM, VIM, IMP, and OXA-48. The oligonucleotide sequences of the primers were listed in Supplementary Table 2 (9,19,20). The QIAquick PCR Purification Kit was used to purify the PCR product, #REF is 28104. The PCR products were sequenced and analyzed using BLAST (http://www.ncbi.nlm.nih.gov/BLAST). Our BLAST cut-off is that the percent identity must be 100%.

Data Analysis
Using BMD as a gold stand, the categorical agreement (CA), essential agreement (EA), minor error (MIE), major error (ME), and very major error (VME) were calculated (21,22). Results were considered CA when isolates had the same susceptible, intermediate, susceptible-dose dependent, and resistant category with the BMD method category result. Results were considered EA when the MIC obtained with the BD Phoenix NMIC-413 panel was within one doubling dilution step (two-fold serial) of the MIC value established by the BMD method. Results were considered ME when the BMD method result was susceptible and the BD Phoenix NMIC-413 panel is resistant. Results were considered MIE when one result was intermediate and the other was susceptible or resistant. Results were considered VME when the BMD method result was resistant and the BD Phoenix NMIC-413 panel was susceptible. The calculation formulas of related parameters were shown in Supplementary Table 3. The Spearman correlation coefficients (P-value) were calculated by SPSS 26.0. The linear regression curve was performed using GraphPad Prism8, and the R 2 -value was obtained at same time.

Genotype Determination
The genotypes determined by PCR and Sanger sequencing demonstrated different resistance mechanisms which were shown in Table 1. The highest proportion was the KPC-2 carbapenemase gene (100/195), followed by IMP (33/195), NDM (28/195), VIM (7/195), and OXA-48(2/195). Two isolates produced both IMP-1 and VIM-1. A total of 27 CRE isolates were carbapenemase gene negative in this study. KPC-2 was the most prevalent carbapenemase gene in K. pneumoniae strains, while IMP and NDM accounted for the majority gene type in E. cloacae and C. freundii. All carbapenem-susceptible isolates (N = 108) were proved no carbapenemase genes by PCR method.  Table 2 showed CAs, EAs, and the number of MIE, ME, and VME for 303 clinical Enterobacterales isolates. CA and EA were all above 90%, and the CA of the three antibiotics in the BD Phoenix NMIC-413 was greater than the disk diffusion.

Performance of BD NMIC-413 Panel vs. Disk Diffusion Method
For imipenem, CA and MIE of the BD Phoenix NMIC-413 were similar with the disk diffusion. However, the VME of BD Phoenix NMIC-413 was 0%, while the disk diffusion was 3.28%. The ME rate of disk diffusion was close to 3%, while the BD Phoenix NMIC-413 was only 0.99%. For meropenem, the CA, MIE, ME, VME of BD

Performance Evaluation Against CRE Strains With Different Genotypes
For imipenem, the BD NMIC-413 and disk diffusion method showed a CA rate of 62.96 and 74.07% separately against carbapenemase-non-producing CREs. The disk diffusion performed not well in IMP-producing CREs, with CA of 72.73% and VME of 9.09%. There was a KPC-producing K. pneumoniae, an OXA-48-producing Es. coli, and an IMPproducing K. aerogenes in the BD NMIC-413 that belong to MIE, the MIC detected by BMD and BD NMIC-413 were 4 and 2 mg/L, respectively. For ertapenem, two methods performed well in the detection of different enzyme types, only disk diffusion had a slightly lower CA (93.94%) in the detection of IMP type isolates. Ertapenem was the best to detect CRE among the three antibiotics.
For meropenem, the EA of IMP-producing CREs was <90% (87.88%) when detected by the BD Phoenix NMIC-413. When testing the IMP/VIM producing isolates or carbapenemasenon-producing isolates, the disk diffusion method showed that the CA was <90%, while the VME of IMP-producing isolates was >3% (3.13%). There was one KPC-producing Es. coli, one carbapenemase-non-producing E. cloacae, and one carbapenemase-non-producing K. pneumoniae in BD NMIC-413 belong to MIE. These results of BMD were resistant (MIC was 8, 64, and 4 mg/L, respectively), while BD NMIC-413 were intermediate (MIC all was 2 mg/L). Meanwhile, there was also one IMP-producing K. pneumoniae and one OXA-48-producing   Table 3 showed the rates of CAs, EAs, and the number of MIE and VME for 195 resistant isolates with the three antibiotics.

DISCUSSION
In prior studies, the BD Phoenix NMIC-500 and NMIC-203 panels have been reported for the performance evaluation of negative bacilli, and PMIC-84 panel has reported for positive cocci (15,23), including identification and antimicrobial susceptibility testing evaluation (24). In this study, the performance of the BD Phoenix NMIC-413 system used for CRE testing was compared with the traditional disk diffusion method, while BMD was used as the reference method. Compared with BMD and disk diffusion, the BD Phoenix NMIC-413 panel had higher CA/EA and lower ME/MIE/VME. For carbapenemase genes negative CRE isolates, the CA rates of imipenem were <75%, while the other two antibiotics were all >90%. Imipenem was often degraded by the enzyme dehydropeptidase-1 (DHP-1), hence the co-administration with a DHP-1 inhibitor such as cilastatin was required (25). Later meropenem and ertapenem demonstrated increased stability to DHP-1 and were administered without a DHP-1 inhibitor (25). So meropenem and ertapenem are more stable than Imipenem. For intermediate isolates, the MICs obtained by BD Phoenix system NMIC-413 panel were probably higher than BMD. Considering that the BD Phoenix system is a broth-based microdilution method that not only measures turbidity but also detects the bacterial growth utilizing the redox indicator (26), enabling it to detect resistant bacteria with high sensitivity. In addition, the BD Phoenix system had fewer manual operations, which was simpler and more convenient compared with the BMD method. The advantages and disadvantages of the three methods were shown in Supplementary Table 4.
Haffler et al. (27) have evaluated one panel of the BD Phoenix system and showed that the CA of ertapenem and meropenem were 94 and 50%. Hogan et al. (28) have evaluated the AST of VITEK 2 (bioMérieux, France) with Gram-negative bacteria by blood culture and reported that the CA and EA were 86.5 and 84.6, 96.2, and 96.2% for meropenem and ertapenem, respectively. In our study, the CA and EA of the BD Phoenix system NMIC-413 panel were ranged from 93 to 98%.
However, the BD Phoenix system NMIC-413 panel can only obtain MICs of the designated antibiotics, so it cannot be used for taxonomic identification of the selected isolates. However, the BD Phoenix CPO panel can detect the MIC and classification of CRE at the same time. It can divide CRE into three categories: A (KPC), B (NDM, IMP, and VIM), and D (OXA-48). Saad Albichr et al. (29) evaluated the performance of the automated BD Phoenix CPO Detect-test for detection and Ambler classification of carbapenemases in Enterobacterales, P. aeruginosa, and A. baumannii complex, the overall sensitivity and specificity were 89.7 and 83.5%, respectively, 68.9 and 62.1% for P. aeruginosa, respectively. Although the BD Phoenix system can detect the classification of CRE expediently, the accuracy needed to improve, it also needs other methods to verify.
The BD phoenix NMIC-413 panel also had a notable limitation in this study. There are not enough bla VIM and bla OXA producing strains among the isolates tested in this study. Mainly because these two carbapenemase types are very rare in China (32,33). From 2010 to 2019, we isolated only 7 bla VIM and 2 bla OXA producing strains in our laboratory. Now we do not have enough VIM and OXA-48 to make up for this limitation. In the next work, we will deliberately save these two types of strains to facilitate future research.

CONCLUSION
In conclusion, for the BD Phoenix NMIC-413 panel, CA and EA rates > 93%, MIE rates < 5%, ME rates < 1.75%, and VME rates were 0%, across the three drugs. It showed acceptable performance as alternatives to the BMD method for clinical treatment explanation.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The protocol has been reviewed by the Human Research Ethics Committee of the Institutional Review Board (IRB) of the Peking Union Medical College Hospital and since all bacterial strains were from residual samples used in clinical diagnosis or strains from their subcultures, it met the criteria for exemption. This project did not involve any patient information nor did it affect the normal diagnosis and treatment of patients, and after consultation with the IRB, formal ethical approval was reviewed and waived; written patient consent was not required (Ethics Approval Number: S-K677).

AUTHOR CONTRIBUTIONS
QY and YX conceived and designed the experiments. JZ, GZ, PJ, and YZ performed the experiments, analyzed the data, and wrote the paper. All authors approved the final version of the manuscript.