Evaluation of Boronic Acid Disk Tests for Differentiating KPC-Possessing Klebsiella pneumoniae Isolates in the Clinical Laboratory

Clinical isolates.The boronic acid disk tests for the detection of KPC were evaluated with 57 genotypically confirmed KPC-possessing K. pneumoniae isolates. The isolates were recovered from separate patients who were hospitalized in five tertiary-care hospitals located in four distinct Greek regions (two hospitals in the broad region of Athens and one hospital each in Thessaloniki, Larissa, and Serres). The presence of bla_KPC was determined using previously described oligonucleotide primers and cycling conditions (19). Reverse transcription-PCR using these primers and a previously described protocol (22) was performed to demonstrate the activity of the KPC enzyme. To measure the specificity of methods to detect KPC-mediated resistance, 89 K. pneumoniae isolates were chosen for testing. All 89 isolates were negative for bla_KPC by PCR. Non-KPC-possessing isolates were randomly selected from isolates that were considered to have reduced susceptibility to either cefoxitin (MIC, >8 μg/ml), expanded-spectrum cephalosporins (cefotaxime or ceftazidime; MIC, >8 μg/ml), or carbapenems (imipenem or meropenem; MIC, >4 μg/ml). In addition, 17 KPC-negative Escherichia coli isolates with the same characteristics were included. These 106 isolates possessed a variety of plasmid-mediated AmpC types, ESBLs, and MBLs and came from collections held at the clinical laboratories providing the KPC-positive isolates for the present study. The identification of all isolates was confirmed by using the API 20E system (bioMerieux, Marcy l'Etoile, France).

Antimicrobial susceptibility testing and phenotypic screening.Detailed susceptibility analysis was carried out by the agar dilution method according to CLSI guidelines and interpretative criteria (6). For tigecycline, the U.S. Food and Drug Administration recommendation was used (susceptibility, MIC of ≤2 μg/ml; resistance, MIC of ≥8 μg/ml), and for colistin, the CLSI recommendation for Acinetobacter spp. was used (susceptibility, MIC of ≤2 μg/ml; resistance, MIC of ≥4 μg/ml). Screening for the presence of a carbapenemase was performed with the modified Hodge test (16). The MBL Etest (AB Biodisk, Solna, Sweden) and the combined-disk test with imipenem and EDTA (10) were used to screen for the production of class B carbapenemases. ESBL production was tested using the CLSI confirmatory test and a modification, which uses clavulanate in combination with boronic acid (26). Isolates were phenotypically tested for the presence of AmpC β-lactamase by determining cefoxitin and imipenem MICs in agar with and without 200 μg/ml cloxacillin (23) and by using the AmpC detection Etest strips (AB Biodisk), which contain cefotetan with or without cloxacillin.

Molecular testing for β-lactamase genes.β-Lactamase genes were amplified by using a panel of primers for the detection of all types of MBLs (12), KPCs (19), plasmid-mediated AmpC enzymes (with single PCRs for each gene) (21), and ESBLs, including SHV, TEM, CTX-M, and GES/IBC enzymes (29). The primers for KPC are located from nucleotide positions −39 to +68 of the total, 882-bp bla_KPC gene (GenBank accession no. EU176014) and produce a 989-bp amplicon. As positive controls we used previously characterized isolates from our collection carrying all types of tested β-lactamases. In some isolates, PCR products were purified using ExoSAP-IT reagent (USB Corporation, Cleveland, OH) and used as templates for sequencing on both strands with an ABI Prism 377 DNA sequencer (Applied Biosystems, Foster City, CA).

Boronic acid disk tests.The phenotypic detection of KPC-possessing K. pneumoniae isolates was evaluated with boronic acid disk tests. The stock solution was prepared as previously recommended (7) by dissolving phenylboronic acid (benzeneboronic acid; Sigma-Aldrich, Steinheim, Germany) in dimethyl sulfoxide at a concentration of 20 mg/ml. From this solution, 20 μl (containing 400 μg of boronic acid) was dispensed onto commercially available antibiotic disks. The disks were then dried and used within 60 min. The tests were performed by inoculating Mueller-Hinton agar by the standard diffusion method (6) and placing disks containing eight different β-lactams (imipenem, meropenem, ertapenem, cefepime, cefoxitin, cefotetan, cefotaxime, and ceftazidime) with or without boronic acid onto the agar. The agar plates were incubated at 37°C overnight. The diameter of the growth-inhibitory zone around a β-lactam disk with boronic acid was compared with that around the corresponding β-lactam disk without boronic acid. The test was considered positive for the detection of KPC enzyme production when the diameter of the growth-inhibitory zone around a β-lactam disk with boronic acid was ≥5 mm larger than that around a disk containing the β-lactam substrate alone. It should be noted that the concentration of boronic acid employed in the present study did not show any detectable effect on bacterial growth, since the boronic acid MIC exceeded 2,500 μg/ml.

Sensitivity and specificity.The performances of the various boronic acid disk tests for the detection of KPC enzymes were evaluated using PCR as the “gold standard.” For each test, sensitivity was calculated from the number of KPC-possessing organisms that were correctly determined, while specificity was calculated from the number of non-KPC-possessing organisms that were correctly determined.

Clinical isolates.Of the 57 KPC-possessing K. pneumoniae isolates, 19 were from blood (33.3%), 18 from wound swabs (31.6%), 11 from urine (19.3%), 6 from respiratory tract specimens (10.5%), and 3 from vascular catheter tips (5.3%). Of the 106 non-KPC-possessing isolates, 32 were from blood (30.2%), 29 from urine (27.4%), 21 from respiratory tract specimens (19.8%), 19 from wound swabs (17.9%), 4 from vascular catheter tips (3.8%), and 1 from cerebrospinal fluid (0.9%).

Antimicrobial susceptibilities.Carbapenem MICs differed substantially among the 57 KPC-positive isolates, ranging from 4 to 64 μg/ml, 2 to 64 μg/ml, and 4 to >128 μg/ml for imipenem, meropenem, and ertapenem, respectively. Ten of the KPC producers were susceptible to meropenem (MICs, 2 to 4 μg/ml), susceptible or intermediate to imipenem (MICs, 4 to 8 μg/ml), and intermediate or resistant to ertapenem (MICs, 4 to 16 μg/ml), while the remaining 47 isolates were either intermediate or resistant to all three carbapenems (Table 1). All KPC-positive isolates exhibited resistance to penicillins (ampicillin and piperacillin), β-lactam-inhibitor combinations (amoxicillin-clavulanate, piperacillin-tazobactam), cefoxitin, expanded-spectrum cephalosporins (ceftazidime, cefotaxime, cefepime), and aztreonam. Additionally, all KPC producers showed resistance to ciprofloxacin, while 51 showed resistance to trimethoprim, 23 to amikacin, 6 to colistin, and 4 to gentamicin, but none were resistant to tigecycline (MICs, 0.5 to 2 μg/ml).

Among the 106 KPC-negative isolates, 61 were nonsusceptible (imipenem and meropenem MICs, >4 μg/ml; ertapenem MIC, >2 μg/ml) to at least one of the three carbapenems (Table 1). Additionally, as many as 93, 100, and 99 were nonsusceptible (MIC, >8 μg/ml) to cefoxitin, ceftazidime, and cefotaxime, respectively.

Phenotypic and molecular screening.All 57 KPC-positive isolates had positive results by the modified Hodge test, and reverse transcription-PCR experiments confirmed the expression of the KPC gene in all of them. Sequencing results for 10 randomly selected isolates identified the KPC-2 variant of the gene. Phenotypic testing for ESBL production in combination with molecular testing revealed that as many as 48 of the 57 KPC-bearing isolates were also ESBL producers. In all of these cases, an SHV-type ESBL was detected. Sequencing analysis was performed for all these amplicons; it detected SHV-11 in 2 cases and SHV-12 in the remaining 46 cases. PCR testing for other groups of ESBL genes (CTX-M, TEM, GES/IBC), as well as for MBL genes and all known clusters of plasmid-mediated AmpC genes, was consistently negative. In accordance with the latter results, all isolates gave negative results on the phenotypic tests for MBL detection and the cloxacillin inhibition tests, further indicating the absence of MBLs and plasmid-mediated AmpC enzymes. It is also noteworthy that 26 of the 57 KPC producers harbored the TEM-1 β-lactamase.

Molecular testing in combination with phenotypic testing of the 106 non-KPC-possessing isolates showed that 53 harbored VIM-type MBLs, 56 harbored ESBLs (33 harbored SHV-type ESBLs, 19 harbored CTX-M-type ESBLs, 2 harbored SHV- and TEM-type ESBLs, and 2 harbored GES/IBC-type ESBLs), and 40 harbored AmpC β-lactamases, which belonged to two of the six plasmid-mediated AmpC enzyme clusters (24 belonged to the cluster comprising MOX-1, MOX-2, CMY-1, and CMY-8 to CMY-11, and 16 belonged to the cluster comprising LAT-1 to LAT-4, CMY-2 to CMY-7, and BIL-1).

Boronic acid disk tests.Detailed results of the boronic acid disk tests are shown in Table 2, while their sensitivities, specificities, positive predictive values, and negative predictive values are shown in Table 3. All 57 KPC-possessing isolates showed increases of 5 mm or more in the zone diameters of the combined disks over those for cefotetan or cefepime alone (sensitivity, 100%). Boronic acid also enhanced remarkably (by ≥7 mm) the activities of all three carbapenems (imipenem, meropenem, and ertapenem) for all 57 KPC producers (sensitivity, 100%) irrespective of the carabapenem MICs, clearly indicating the inhibitory activity of boronic acid against KPC enzymes. Figure 1a and b show the growth patterns of a representative KPC-possessing K. pneumoniae isolate without or with boronic acid and with meropenem, ertapenem, cefotetan, and cefepime as antibiotic substrates. This method showed lower sensitivity when cefoxitin or cefotaxime was used as a substrate (96.5% or 66.7%, respectively). The inhibitor affected the activity of ceftazidime against the 9 ESBL-negative isolates as well as the 2 SHV-11 producers but not against the remaining 46 SHV-12 producers (sensitivity, 19.3%), most likely due to the presence of the SHV-12 ESBL, which is not restrained by boronic acid.

• Open in new tab
• Download powerpoint

FIG. 1.

Representative results using the boronic acid-based method without or with boronic acid (bor) for isolates possessing KPC/ESBL (a and b), VIM (c and d), AmpC/ESBL (e and f), or ESBL (g and h). MER, meropenem; ETP, ertapenem; CTT, cefotetan; FEP, cefepime.

None of the 106 isolates negative for KPC by PCR gave positive results in boronic acid tests using disks of cefepime, imipenem, or meropenem (specificity, 100%) (Tables 2 and 3). When disks containing ertapenem were used, all isolates negative for KPC by PCR were correctly identified as KPC negative except for five AmpC producers that gave false-positive results (specificity, 95.3%) (Tables 2 and 3). Boronic acid tests using cefotetan, cefoxitin, cefotaxime, and ceftazidime as antibiotic substrates were positive for 42, 32, 19, and 13 isolates, respectively (specificities for KPC detection ranged from 60.4% to 87.7%) (Table 3). Figure 1c to h show the growth patterns of three representative non-KPC-possessing K. pneumoniae isolates (harboring a VIM enzyme, plasmid-mediated AmpC, and ESBL, respectively) without or with boronic acid and with meropenem, ertapenem, cefotetan, and cefepime as antibiotic substrates. All these results were consistent for two different batches of Mueller-Hinton agar.

It is noteworthy that among the three antibiotic compounds (imipenem, meropenem, and cefepime) that correctly differentiated all 163 isolates in the study, meropenem demonstrated the largest differences in inhibition zone diameters between KPC producers and non-KPC producers (Fig. 2).

• Open in new tab
• Download powerpoint

FIG. 2.

Increases in the inhibition zone diameters of meropenem (in millimeters) in the presence of boronic acid for 57 isolates positive for KPC by PCR and 106 isolates negative for KPC by PCR.