|Home | About | Journals | Submit | Contact Us | Français|
CTX-M extended-spectrum beta-lactamases (ESBLs) have emerged as the most common type of ESBL globally, their incidence easily surpassing those of SHV and TEM ESBLs in most locales. This study compared the performance of two MicroScan dried panels with CLSI reference broth microdilution and disk diffusion methods on a collection of genetically characterized ESBL-producing isolates. These included 64 Enterobacteriaceae isolates that produced CTX-M8, -14, -15, or -16 according to PCR and sequencing of the bla gene, 17 isolates that produced a SHV or TEM ESBL, and 19 that produced both CTX-M and SHV ESBLs. Each isolate was tested by a frozen reference microdilution panel, the MicroScan ESβL plus confirmation panel, and a routine dried panel containing streamlined ESBL confirmation dilutions (MicroScan Neg MIC panel type 32) that included cefotaxime and ceftazidime tested alone or with a fixed concentration of 4 μg/ml of clavulanate. Each isolate was also tested by the standard CLSI double-disk confirmation tests. The disk diffusion method detected all ESBL-producing isolates, the frozen reference panel detected 90% of isolates (10 out of 100 could not be analyzed because of off-scale MICs that exceeded the clavulanate combination concentrations in the panel), the ESβL plus panel detected 98% (1 missed and 1 off scale), and the streamlined ESBL panel detected 95% (5 off scale). Very high MICs for a few strains that produced SHV or both CTX-M and SHV ESBLs precluded noting the required three twofold-dilution differences with clavulanate needed to confirm an ESBL primarily in the reference panel and the Neg type 32 panel.
Extended-spectrum beta-lactamases (ESBLs) have historically arisen from mutations in the genes encoding the very common plasmid-mediated TEM-1 and SHV-1 enzymes (6). These “classic” ESBLs are enzymes with hydrolytic activities against newer penicillins, cephalosporins, and aztreonam. At present, 175 different TEM enzymes and 127 SHV different enzymes have been described (www.lahey.org/studies/). However, a newer group of ESBLs have arisen and spread quickly on a global basis. The CTX-M family of enzymes appears to have arisen by initial transfer of the chromosomal beta-lactamase gene from Kluyvera spp. to conjugative plasmids that have readily spread among different members of the Enterobacteriaceae and certain other gram-negative bacteria (2). Mutations in these genes have resulted in the described enzymes CTX-M1 through CTX-M91, more than 30 of which have been recognized in 2008 to 2009 (www.lahey.org/studies/). These enzymes have become the most common type of ESBL in our health care system (the University Health System in San Antonio, TX; 8). Many of the CTX-M enzymes hydrolyze cefotaxime more efficiently than ceftazidime, the opposite of the case for many of the more familiar TEM and SHV enzymes (2, 6). Like the “classic” ESBLs, the CTX-M ESBLs have often been found in Escherichia coli and Klebsiella spp. but have also been found in Proteus, Enterobacter, and Morganella isolates (2, 6, 8). The disk and MIC screening tests using cefpodoxime and the confirmatory tests recommended by the CLSI which incorporate cefotaxime and ceftazidime tested alone and combined with clavulanate (3) were developed to detect the classic ESBLs. However, it appears that the disk tests work equally well in detection of CTX-M enzymes as long as cefotaxime is used for the confirmatory testing (8). Less information is available regarding the performance of broth microdilution tests that incorporate clavulanate combinations. The purpose of this study was to assess the performances of two commercial dried MIC panels (the MicroScan ESβL plus panel and the routine Neg MIC panel type 32 containing streamlined ESBL dilutions) relative to those of both a reference CLSI frozen broth microdilution panel and the CLSI disk methods on a collection of genetically characterized CTX-M-producing strains.
A group of previously characterized ESBL-producing isolates recovered between 2000 and 2007 from patients from the University Health System, San Antonio, TX, were used in this study. Their ESBLs were characterized by PCR amplification using primer sets for CTX-M, SHV, and TEM genes, followed by sequencing of the PCR products (8). Most isolates produced a single ESBL, while some produced a second ESBL. Additional strains that produced only SHV or TEM ESBLs were also tested for comparative purposes. For the purposes of this study, only E. coli, Klebsiella spp., and Proteus mirabilis isolates, those for which CLSI has described confirmatory ESBL tests, were included.
The standard CLSI disk tests were performed on Mueller-Hinton agar (BD Microbiology Systems, Cockeysville, MD) and incubated for 16 to 18 h at 35°C. Disks (BD Microbiology) included 30 μg cefotaxime, 30 μg cefotaxime plus 10 μg clavulanate, 30 μg ceftazidime, and 30 μg ceftazidime plus 10 μg clavulanate. A test was considered positive if there was an increase of at least 5 mm in the zone diameter from the clavulanate combination test compared to that from the test with cephalosporin substrate alone (3).
Broth microdilution panels were prepared according to CLSI guidelines by using cation-adjusted Mueller-Hinton broth and frozen at −70°C until used for testing. Panels were inoculated with the standard 5 × 105-CFU/ml density and incubated for 16 to 20 h at 35°C. The panels included cefpodoxime (0.12 to 16 μg/ml), cefotaxime (0.06 to 128 μg/ml), cefotaxime (0.06 to 16 μg/ml) plus clavulanate (4 μg/ml), ceftazidime (0.06 to 128 μg/ml), and ceftazidime (0.06 to 16 μg/ml) plus clavulanate (4 μg/ml). A confirmatory test was considered positive if there was a decrease in the MIC of at least 3 twofold dilutions for the clavulanate combination relative to the cephalosporin MIC when tested alone (3).
Both the MicroScan ESβL plus and MicroScan Neg type 32 dried panels were included in the study. The two dried panels and the frozen reference panels were inoculated using the same 0.5 McFarland standard organism suspensions. The manufacturer's intermediary inoculum dilution tubes and inoculum devices were used to inoculate the dried panels to achieve the same goal of a 5 × 105-CFU/ml initial inoculum density, followed by incubation for 16 to 20 h at 35°C. The ESβL plus panel included cefpodoxime (0.5 to 64 μg/ml), cefotaxime (0.5 to 128 μg/ml), cefotaxime plus clavulanate (0.12/4 to 16/4 μg/ml), ceftazidime (0.5 to 128 μg/ml), and ceftazidime plus clavulanate (0.12/4 to 16/4 μg/ml). The Neg type 32 panel included cefotaxime (2, 8, 16, and 32 μg/ml), cefotaxime plus clavulanate (0.5/4 to 4/4 μg/ml), ceftazidime (1, 4, 8, and 16 μg/ml), and ceftazidime plus clavulanate (0.25/4 to 2/4 μg/ml). The MicroScan dried panels and the frozen reference panels were read visually using the same mirror reading stand.
E. coli ATCC 25922 and Klebsiella pneumoniae ATCC 700603 were used as control strains on each day of testing (3).
The achievements of the required ≥5-mm zone diameter differences and the ≥3 twofold-dilution MIC differences were assessed for each strain and each test method (3). The performance of each method was determined for the ESBL type (i.e., CTX-M, SHV, TEM, or an ESBL combination) and each species tested. A method was considered to have accurately detected an ESBL if the criteria described above were achieved for either substrate pair, i.e., either the cefotaxime test pair or the ceftazidime test pair. If the specified criteria were not met with one substrate pair and the other substrate pair provided equivocal results due to off-scale (high) MICs, the method was considered to yield an equivocal result.
A group of 100 Enterobacteriaceae isolates that produced ESBLs previously characterized using PCR and gene sequencing was examined by the four phenotypic susceptibility methods described above. The species included in the test group and the ESBL enzymes they produced are indicated in Table Table1.1. These included predominantly E. coli isolates. A total of 83 isolates produced a CTX-M ESBL alone or in combination with a SHV ESBL. Seventeen additional isolates that produced only a SHV or TEM ESBL were included for comparative purposes. All isolates yielded cefpodoxime MICs ≥ 2 μg/ml in the frozen reference and ESβL plus panels (data not shown).
As indicated in Table Table2,2, the disk diffusion method detected all ESBL-producing isolates, the frozen reference panel unequivocally detected 90% of all isolates, the ESβL plus panel detected 98%, and the Neg 32 panel detected 95%. All of the methods performed well with the 64 isolates that produced a CTX-M ESBL. One missed isolate and the isolates that were associated with equivocal test results by one or more of the dilution methods are detailed in Table Table3.3. Only one isolate was not detected by the ESβL plus panel, an E. coli isolate that produced CTX-M15 (Table (Table3).3). Equivocal results were obtained with 10 isolates in the frozen reference panel, while 1 and 5 isolates had equivocal results with the ESβL plus and Neg type 32 panels, respectively. These equivocal results were due to high off-scale substrate-clavulanate MICs for a few strains that produced SHV or both CTX-M and SHV ESBLs that precluded our being able to observe the required three twofold-dilution differences with clavulanate needed to confirm an ESBL (Table (Table33).
The types and frequencies of extended-spectrum beta-lactamases produced by commonly encountered members of the Enterobacteriaceae have changed significantly since the early part of this decade. SHV- and TEM-derived ESBLs have arisen from mutations in the very common genes that encode TEM-1 and SHV-1 enzymes (6). Most of these “classic” ESBLs have caused nosocomial infections in hospitalized patients or perhaps those residing in long-term-care facilities (6). However, a new group of ESBLs, the CTX-M family, has rapidly emerged and become the predominant ESBL type in most parts of the world (2, 10, 12, 13, 14, 16). Infections caused by CTX-M-producing gram-negative bacteria are often community acquired or community onset but health care associated (2, 4, 11, 14, 15). There is a preponderance of urinary infections due to E. coli strains that produce a CTX-M ESBL, such as CTX-M15 (1, 2, 5, 8, 11, 15). The CTX-M ESBLs may have spread rapidly in the international clone of virulent and fluoroquinolone-resistant E. coli, ST131 (also known as O25:H4) (1, 7). There is some evidence that there is a food-borne source that has facilitated the spread of CTX-M-producing organisms in the community (4, 9, 17). It is necessary for clinical microbiology laboratories to be prepared to detect the CTX-M ESBLs as part of routine susceptibility testing. The CLSI screening method incorporating cefpodoxime followed by confirmation with the double-disk method using clavulanic acid inhibition of the plasmid-mediated ESBLs seems to be reliable for routine detection of the CTX-M enzymes (3). This study has assessed the ability of dried microdilution panels from one manufacturer to detect these newer CTX-M ESBLs.
The CLSI double-disk diffusion method using cefotaxime and ceftazidime in combinations with clavulanate detected all ESBL-producing isolates included in this study of genetically characterized ESBL-producing isolates, in part because zone diameters could always be measured and the clavulanate effect could always be assessed. Most ESBL producers were effectively detected by using a frozen reference broth microdilution panel and the two MicroScan dried panels. However, occasional high, off-scale MICs with the cephalosporin-clavulanate combinations obtained with a few potent enzyme producers limited the ability to fully assess the inhibition by clavulanate. This occurred primarily with some K. pneumoniae isolates that produced both CTX-M15 and SHV-12 ESBLs. We regarded such results as equivocal, suggesting but not confirming the presence of an ESBL. Interestingly, this occurred more often with the frozen reference panels than the dried MicroScan ESβL plus panels, which contained similar concentration ranges of the drugs of interest. It is not clear if this represented lesser stability of the clavulanate in the frozen panels, although that was not reflected in tests of the quality control strain K. pneumoniae ATCC 700603. Another possibility might be a slightly higher inoculum density in the frozen reference panels, although both dried and frozen panels had the same target inoculum density and were prepared from the same organism suspension. Off-scale cephalosporin-clavulanate MICs were an occasional problem for the MicroScan Neg type 32 panels, since only limited combination wells were included.
This study has shown that the standard CLSI ESBL confirmatory disk diffusion and broth microdilution tests and two MicroScan dried MIC panels detected CTX-M ESBLs very efficiently. The rapid emergence of CTX-M ESBLs, the fact that many related infections occur among outpatients, and the occurrence of related urinary infections place great emphasis on the use of accurate and convenient testing methods for routine detection of CTX-M-producing strains. Clinical microbiology laboratories should continue to detect and report ESBLs routinely and to expand testing to include testing for outpatient urinary infections if not already done so as part of their standard protocol.
This study was supported in part by Siemens Healthcare Diagnostics.
Published ahead of print on 4 November 2009.