Diagnostic Performance of BD Phoenix CPO Detect Panels for Detection and Classification of Carbapenemase-Producing Gram-Negative Bacteria

ABSTRACT BD Phoenix CPO Detect panels can identify and classify carbapenemase-producing organisms (CPOs) simultaneously with antimicrobial susceptibility testing (AST) for Gram-negative bacteria. Detection and classification of carbapenemase producers were performed using the BD Phoenix CPO Detect panels NMIC/ID-441 for Enterobacterales, NMIC/ID-442 for nonfermenting bacteria, and NMIC-440 for both. The results were compared with those obtained using comparator methods. A total of 133 strains (32 Klebsiella pneumoniae, 37 Enterobacter cloacae complex, 33 Pseudomonas aeruginosa, and 31 Acinetobacter baumannii complex strains), including 60 carbapenemase producers (54 imipenemases [IMPs] and 6 OXA type), were analyzed. Using panels NMIC-440 and NMIC/ID-441 or NMIC/ID-442, all 54 IMP producers were accurately identified as CPOs (positive percent agreement [PPA], 100.0%; 54/54). Among the 54 IMP producers identified as CPOs using panels NMIC-440 and NMIC/ID-441, 12 and 14 Enterobacterales were not resistant to carbapenem, respectively. Among all 54 IMP producers, 48 (88.9%; 48/54) were correctly classified as Ambler class B using panel NMIC-440. Using panels NMIC-440 and NMIC/ID-442, all four OXA-23-like carbapenemase-producing A. baumannii complex strains (100.0%, 4/4) were correctly identified as CPOs, and three (75.0%, 3/4) were precisely classified as class D using panel NMIC-440. Both carbapenemase producers harboring the blaISAba1-OXA-51-like gene were incorrectly identified as non-CPOs using panels NMIC-440 and NMIC/ID-442. For detecting carbapenemase producers, the overall PPA and negative percent agreement (NPA) between panel NMIC-440 and the comparator methods were 96.7% (58/60) and 71.2% (52/73), respectively, and the PPA and NPA between panels NMIC/ID-441 or NMIC/ID-442 and the comparator methods were 96.7% (58/60) and 74.0% (54/73), respectively. BD Phoenix CPO Detect panels can successfully screen carbapenemase producers, particularly IMP producers, regardless of the presence of carbapenem resistance and can be beneficial in routine AST workflows. IMPORTANCE Simple and efficient screening methods of detecting carbapenemase producers are required. BD Phoenix CPO Detect panels effectively screened carbapenemase producers, particularly IMP producers, with a high overall PPA. As the panels enable automatic screening for carbapenemase producers simultaneously with AST, the workflow from AST to confirmatory testing for carbapenemase production can be shortened. In addition, because carbapenem resistance varies among carbapenemase producers, the BD Phoenix CPO Detect panels, which can screen carbapenemase producers regardless of carbapenem susceptibility, can contribute to the accurate detection of carbapenemase producers. Our results report that these panels can help streamline the AST workflow before confirmatory testing for carbapenemase production in routine microbiological tests.

The detection of carbapenemase producers is crucial for appropriate antimicrobial therapy and infection control. However, it is difficult to identify them using antimicrobial susceptibility testing (AST) alone because carbapenemase producers may show various levels of MICs for carbapenems, depending on the gene expression, efficacy of carbapenem hydrolysis, and presence of other resistance mechanisms, such as permeability (1,8). We previously reported that carbapenemase-producing K. pneumoniae and E. cloacae complex strains could be overlooked if screened based on carbapenem resistance detected using the broth microdilution method (9). Therefore, simple and efficient screening methods that are independent of the MIC results are required in daily practice.
In this study, we evaluated BD Phoenix CPO Detect panels, which enable the detection and classification of carbapenemase-producing organisms (CPOs) simultaneously with AST for Gram-negative bacteria, and discuss their usefulness in a routine microbiological testing workflow.

RESULTS
A total of 133 (32 K. pneumoniae, 37 E. cloacae complex, 33 P. aeruginosa, and 31 Acinetobacter baumannii complex) strains were included in this study, among which, 60 strains (15, 20, 14, and 11, respectively) were identified as true carbapenemase producers using the comparator methods. All carbapenemases observed in K. pneumoniae, E. cloacae complex, and P. aeruginosa strains belonged to group IMP-1, including the IMP-6 observed in only one P. aeruginosa strain. Five bla IMP-1 group, four bla OXA-23-like , and two bla ISAba1-OXA-51-like genes were included in the 11 carbapenemase-producing A. baumannii complex strains. Tables 1 to 3 show the results of comparisons between panel NMIC-440 and the comparator methods for detecting carbapenemase producers. Using panel NMIC-440, all 54 true IMP producers (15 K. pneumoniae, 20 E. cloacae complex, 14 P. aeruginosa, and five A. baumannii complex strains) were identified as CPOs. Among these 54, 42 and 12 were carbapenem-resistant and nonresistant strains, respectively. Furthermore, using panel NMIC-440, 48 (88.9%, 48/54) were correctly classified as class B, and 6 (11.1%, 6/54) were unclassified. In the A. baumannii complex strains not harboring the bla IMP-1 group, four true OXA-23-like carbapenemase producers were identified as CPOs using panel NMIC-440; among these, three were classified as class D, and one was incorrectly classified as class B. Using panel NMIC-440, two true carbapenemase producers harboring the bla ISAba1-OXA-51-like gene were not identified as CPOs. Meanwhile, among the 73 true non-carbapenemase producers, panel NMIC-440 incorrectly detected 21 (5 K. pneumoniae, 7 P. aeruginosa, and 9 A. baumannii complex) strains as CPOs, among which 18 were incorrectly classified as belonging to several classes (class A, 5 P. aeruginosa strains; class B, 3 K. pneumoniae strains and 1 P. aeruginosa strain; class D, 9 A. baumannii complex strains), whereas three remained unclassified. With respect to carbapenem resistance, among the 58 true carbapenemase producers with correct detection using the panel, 12 (2 K. pneumoniae and 10 E. cloacae complex strains) were not resistant to carbapenem. Conversely, among the 52 true non-carbapenemase producers correctly determined as non-CPOs using the panel, four (two E. cloacae complex and two P. aeruginosa) strains exhibited carbapenem resistance.
The detection results of carbapenemase producers using panels NMIC/ID-441 or NMIC/ID-442 and the comparator methods are shown in Tables 3, 5, and 6. The results using panel NMIC/ID-441 for K. pneumoniae and E. cloacae complex strains were fully concordant with those using panel NMIC-440. Using panel NMIC/ID-442, all 19 true IMP producers (14 P. aeruginosa and 5 A. baumannii complex strains) were identified as CPOs. Four true OXA-23-like carbapenemase-producing A. baumannii complex strains  were accurately detected as CPOs. Both true carbapenemase producers harboring the bla ISAba1-OXA-51-like gene were incorrectly identified as non-CPOs using panel NMIC/ID-442. In contrast, 14 true non-carbapenemase producers (five P. aeruginosa and nine A. baumannii complex strains) were incorrectly identified as CPOs using panel NMIC/ID-442. Among the 58 carbapenemase producers that were detected successfully using panels NMIC/ID-441 or NMIC/ID-442, 14 strains (two K. pneumoniae and 12 E. cloacae complex strains) were carbapenem nonresistant. Among the 54 non-carbapenemase producers with correct determination using panels NMIC/ID-441 or NMIC/ID-442, six (two E. cloacae complex, three P. aeruginosa, and one A. baumannii complex) strains showed carbapenem resistance. The concordance in the detection of carbapenemase producers between panels NMIC/ ID-441 or NMIC/ID-442 and the comparator methods is presented in Table 7. Based on the full concordance in the detection of carbapenemase producers in K. pneumoniae and E. cloacae complex strains between panels NMIC-440 and NMIC/ID-441, the performance was the same between these two panels (Tables 4 and 7). The detection results of panels NMIC-440 and NMIC/ID-442 for carbapenemase producers in nonfermenting bacteria were different only for the two P. aeruginosa and two A. baumannii complex strains. Consequently, the performance was similar between the two panels in the PPA (100.0%; 14/14) for the P. aeruginosa strains and in the PPA (81.8%; 9/11) and NPA (55.0%; 11/20) for the A. baumannii complex strains. The NPA between panel NMIC/ID-442 and the comparator methods was 73.7% (14/19) for P. aeruginosa.
Collectively, based on their detection performance with high PPAs compared to the comparator methods, panels NMIC-440, NMIC/ID-441, and NMIC/ID-442 can be used to detect CPOs, particularly IMP producers, without omission. Meanwhile, we should report the results to clinicians with regard to carbapenemase production and classification for strains detected as CPOs using these panels after performing confirmatory tests, in consideration of NPAs. However, confirmatory tests are usually performed after determining antimicrobial resistance using AST and require additional procedures and time. Therefore, rapid and highly sensitive screening methods are desirable to determine whether additional confirmatory tests are required. In this regard, the BD Phoenix CPO Detect panels, which screen IMP producers simultaneously with AST, with high PPAs in comparison with comparator methods, might be beneficial in the daily AST workflow. In addition, carbapenemase producers may be overlooked if confirmatory tests are performed based on AST results, because not all carbapenemase producers exhibit carbapenem resistance. Therefore, BD Phoenix CPO Detect panels, which can screen for carbapenemase producers regardless of carbapenem susceptibility, can contribute to their detection without omission, thereby preventing their proliferation in medical settings. In particular, the panels might be helpful for regions with high IMP prevalence based on the findings of this study.
There are a few limitations of this study. Because we mainly analyzed IMP producers as carbapenemase producers in this study, we could not fully evaluate other types of carbapenemases. Additionally, we used a modified carbapenem inactivation method (mCIM) as a comparator method for detecting carbapenemase-producing A. baumannii complex strains; however, other methods might be appropriate in some cases. Finally, if IMP is not the dominant carbapenemase, the performance data reported here, such as the PPA and NPA, might change.
In conclusion, we evaluated the performance of BD Phoenix CPO Detect panels (NMIC-440, NMIC/ID-441, and NMIC/ID-442) for the detection and classification of carbapenemase producers. These panels were successfully used to screen for carbapenemase producers, particularly IMP producers, and could be beneficial in routine AST workflows. Bacterial strains and comparator methods. We used bacterial strains isolated between 2009 and 2017 at Nagasaki University Hospital and stored at 280°C. Bacteria were cultured on Nissui separated plate sheep blood agar/chocolate agar EXII (Nissui Pharmaceutical Co., Ltd.) in 5% CO 2 at 35 6 2°C for 18 to 24 h. None of the groups contained duplicate isolates from a single patient. Carbapenemase genes were examined using a multiplex PCR assay kit, the Cica Geneus carbapenemase genotype detection kit 2 (Kanto Chemical Co., Inc.), which can detect bla IMP-1 group, bla IMP-6 , bla VIM group, bla NDM group, bla OXA-48 group, bla KPC group, and bla GES group carbapenemases, according to the manufacturer's instructions. Amplification was performed under the following conditions: 1 min at 94°C; 30 cycles of 15 s at 94°C, 15 s at 63°C, and 40 s at 72°C. In addition, bla OXA-23-like and bla ISAba1-OXA-51-like genes were examined using PCR, and the following primers were used (13,14): OXA-23-like forward, 59-GATCGGATTGGAGAACCAGA-39; OXA-23-like reverse, 59-ATTTCTGACCGCATTTCCAT-39; ISAba1 forward, 59-CACGAATGCAGAAGTTG-39; and OXA-51-like reverse, 59-TGGATTGCACTTCATCTTGG-39. Amplification was conducted under the following conditions: 5 min at 94°C; 30 cycles of 25 s at 94°C, 40 s at 52°C or 58°C, and 50 s at 72°C; and 6 min at 72°C for the final extension.
Carbapenemase production was examined using a modified carbapenem inactivation method (mCIM), according to Clinical and Laboratory Standards Institute (CLSI) guideline M100-ED32, which indicates that the method can be used for Enterobacterales and P. aeruginosa. To determine carbapenemase production in A. baumannii complex strains, we preliminarily compared the performance of mCIM and CIMTris, which is another modified CIM (15), for identifying A. baumannii complex strains carrying carbapenemase genes as carbapenemase producers. Among the 16 A. baumannii complex strains carrying bla IMP-1 group, bla OXA-like , or bla ISAba1-OXA-51-like carbapenemase genes (5, 4, and 7 isolates, respectively), 11 (68.8%; 11/16) and 7 (43.8%; 7/16) tested positive using mCIM and CIMTris, respectively. Using both methods, all 20 A. baumannii strains not carrying those genes were determined as negative. Because mCIM was found to be more sensitive, we used mCIM to detect carbapenemase production in A. baumannii complex strains in this study. True carbapenemase producers were defined as strains that tested positive for both carbapenemase genes and mCIM, whereas true non-carbapenemase producers were defined as strains that tested negative for both carbapenemase genes and mCIM. We excluded strains for which the results were not concordant between PCR and mCIM.
Assays using the panels were performed according to the manufacturer's instructions. Briefly, colonies were suspended in a BD Phoenix ID broth, and the suspension was adjusted to a McFarland standard of approximately 0.25. After the BD Phoenix AST indicator solution was added to the BD Phoenix AST broth, 50 mL of the prepared bacterial suspension was transferred to the broth. After the mixture was injected, the panels were loaded into the BD Phoenix M50 instrument, which is an automated bacterial identification and AST system. The system automatically performs assays, determines the results of MICs, detects CPOs, and provides Ambler classification results based on the specifications of each panel.
Resistance to imipenem and meropenem was determined according to CLSI definitions (MICs of $4 mg/mL for K. pneumoniae and E. cloacae complex strains; MICs of $8 mg/mL for P. aeruginosa and A. baumannii complex strains). Carbapenem-resistant strains were defined as strains exhibiting resistance to either imipenem or meropenem. Susceptible and intermediate strains were classified as not resistant in this study.
Comparisons between BD Phoenix CPO Detect panels and the comparator methods. The results of BD Phoenix CPO Detect panels for carbapenemase producers were compared to those of the comparator methods, and positive and negative percent agreements (PPA and NPA) were calculated as follows: PPA, the number of concordant positive results in the detection of carbapenemase producers divided by the number of all positive results in it obtained using comparator methods; NPA, the number of concordant negative results in the detection of carbapenemase producers divided by the number of all negative results in it obtained using comparator methods (16). The 95% confidence intervals were calculated using R version 4.2.2 (17).

ACKNOWLEDGMENTS
Reagents, instrumentation, and funding were provided by the Nippon Becton Funding was provided by the Nippon Becton, Dickinson Company, Ltd. This study was partly supported by a grant for the research and development of diagnostic methods and therapies for antimicrobial-resistant bacteria from the Japan Agency for Medical Research and Development (AMED) (JP22fk0108133).