ORIGINAL ARTICLE BACTERIOLOGY A sensitive and specific phenotypic assay for detection of metallo-blactamases and KPC in Klebsiella pneumoniae with the use of meropenem disks supplemented with aminophenylboronic acid, dipicolinic acid and cloxacillin C. G. Giske 1,2, L. Gezelius 2, Ø. Samuelsen 3, M. Warner 4, A. Sundsfjord 3,5 and N. Woodford 4 1) Clinical Microbiology, MTC, Karolinska Institutet, Karolinska University Hospital, 2) Swedish Institute for Infectious Disease Control, Stockholm, Sweden, 3) Reference Centre for Detection of Antimicrobial Resistance, Department of Microbiology and Infection Control, University Hospital of North Norway, Tromsø, Norway, 4) Antibiotic Resistance Monitoring and Reference Laboratory, Centre for Infections, Health Protection Agency, London, UK and 5) Department of Medical Biology, Faculty of Health Sciences, University of Tromsø, Tromsø, Norway Abstract Enterobacteriaceae producing carbapenemases, such as KPC or metallo-b-lactamases (MBLs), have emerged on several continents. Phenotypic tests are urgently needed for their rapid and accurate detection. A novel carbapenemase detection test, comprising a meropenem disk, and meropenem disks supplemented with 730 lg of EDTA, 1000 lg of dipicolinic acid (DPA), 600 lg of aminophenylboronic acid (APBA), or 750 lg of cloxacillin, was evaluated against Klebsiella pneumoniae isolates with KPC (n = 34), VIM (n = 21), IMP (n =4)or OXA-48 (n = 9) carbapenemases, and carbapenem-resistant Enterobacteriaceae with porin loss in combination with an extended-spectrum b-lactamase (ESBL) (n = 9) or AmpC hyperproduction (n = 5). Commercially available diagnostics tablets from Rosco containing meropenem and the same inhibitors as described above (except EDTA) were also evaluated. An d meropenem inhibition zone was sought in the presence of each added b-lactamase inhibitor. APBA had excellent sensitivity for detecting K. pneumoniae with KPC enzymes. Isolates with combined AmpC hyperproduction and porin loss were also positive in the APBA test but, unlike KPC producers, showed cloxacillin synergy. Both DPA and EDTA had excellent sensitivity for detection of MBL-producing K. pneumoniae. However, EDTA showed poor specificity, with positive results noted for 1/9 ESBL-producing isolates, for 4/34 KPC-producing isolates, and for 4/9 OXA-48-producing isolates, whereas all of these were negative when DPA was used. The in-house test distinguished accurately between several different mechanisms mediating reduced susceptibility to carbapenems in Enterobacteriaceae. The commercial combination tablets from Rosco performed similarly to the in-house test, with the exception of one false-positive MBL result and one falsepositive KPC result among the OXA-48 producers. Keywords: ESBL CARBA, IMP, KPC, OXA-48, VIM Original Submission: 11 April 2010; Revised Submission: 26 May 2010; Accepted: 29 May 2010 Editor: R. Cantón Article published online: 28 June 2010 Clin Microbiol Infect 2011; 17: 552 556 10.1111/j.1469-0691.2010.03294.x Corresponding author: C. G. Giske, Clinical Microbiology L2:02, Karolinska Institutet-MTC, Karolinska University Hospital Solna, Stockholm SE-17176, Sweden, UK E-mail: christian.giske@karolinska.se Introduction The rapid international spread of Klebsiella pneumoniae strains that produce KPC class A carbapenemases, recently also classified as extended-spectrum b-lactamase (ESBL) CARBA-A [1], is a major concern. Treatment options are severely limited, owing to linked multiresistance, and the association of both KPC-2 and KPC-3 with the successful ST258 K. pneumoniae clone [2] is a great challenge for infection control. Several tests have been described for phenotypic detection of KPC b-lactamases. A modified version of the cloverleaf (Hodge) test, originally established for detection of penicillinases [3], accurately detects KPC b-lactamases [4], but is not able to discriminate between these and other carbapenemases. Moreover, performance of this test requires experience, as the interpretation may not be straightforward with all carbapenemase-producing Enterobacteriaceae (A. Vatopoulos, Clinical Microbiology and Infection ª2010 European Society of Clinical Microbiology and Infectious Diseases
CMI Giske et al. Carbapenemase detection in Klebsiella pneumoniae 553 personal communication). Recently, a boronic acid disk test for the detection of KPC-producing K. pneumoniae was described [5], but it was not evaluated against Enterobacteriaceae with carbapenem resistance arising from a combination of porin loss and hyperproduction of chromosomal AmpC/ plasmid-mediated AmpC, which might also be predicted to give positive results. Thus, the specificity of this test has not been fully evaluated. Moreover, the emergence of metallo-b-lactamases (MBLs) in K. pneumoniae [6] and other Enterobacteriaceae, and their predominance among carbapenemase producers in, for example, Greece, has created a need for a reliable phenotypic test that can detect and distinguish clinically important carbapenemases. The rapid emergence of KPC-producing Enterobacteriaceae on several continents [7] emphasizes that such an assay is urgently needed from an infection control perspective. The aim of this study was to determine whether meropenem disks supplemented with four b-lactamase inhibitors (dipicolinic acid (DPA), EDTA, aminophenylboronic acid (APBA) or cloxacillin) would be able to discriminate between various carbapenemase-producing Enterobacteriaceae, and carbapenem-non-susceptible isolates with ESBL/AmpC hyperproduction in combination with porin changes. Also, we evaluated the performance of commercially available diagnostic tablets containing meropenem in combination with DPA, APBA and cloxacillin. Materials and Methods Bacterial isolates Well-characterized strains from several countries producing various b-lactamases were included in this study. The panel consisted of clinical isolates of KPC-2-producing (n = 9) and KPC-3-producing (n = 12) K. pneumoniae from the CDC [8], KPC-producing ST258 K. pneumoniae from the Health Protection Agency (HPA), London (n = 6) [9], and KPC-2-producing (n = 6) and KPC-3-producing (n = 1) ST258 K. pneumoniae from Norway and Sweden [2]. Isolates of VIM-1-producing K. pneumoniae from Greece (n = 12), VIM- 1-producing K. pneumoniae from Norway and Sweden (n = 6), VIM-1-producing (n = 3) and IMP-producing (n =3) K. pneumoniae and Escherichia coli (n = 1) from the HPA represented MBL-producing Enterobacteriaceae. K. pneumoniae isolates producing OXA-48 carbapenemase were obtained from the HPA (n = 9) (Zhang, 19th ECCMID, 2009, Abstract L249). Carbapenem-resistant K. pneumoniae isolates with CTX-M-15-type and/or SHV-type ESBLs in combination with loss of porins OmpK35 and/or OmpK36 (n = 9), as well as Enterobacter cloacae and Enterobacter aerogenes isolates hyperproducing AmpC and with complete or partial loss of OmpF and/or OmpC (n = 5) [10], were included to investigate the specificity of the various tests. Finally, three isolates of E. coli and one isolate of Enterobacter cloacae with a combination of plasmid-mediated (CMY-2) or chromosomal AmpC, ESBLs (CTX-M-15 or SHV-12) and porin loss were included. Preparation of combination disks Meropenem disks (Oxoid, Basingstoke, UK) were supplemented with 10 ll of three different b-lactamase inhibitors: 100 mg/ml DPA (Sigma, St Louis, MO, USA) (modified from Kimura et al.) [11], 0.2 M EDTA (Sigma) [12], 60 mg/ml APBA (Sigma) and 75 mg/ml cloxacillin (Sigma). Hence, the final amounts of b-lactamase inhibitor in the disks were 1000 lg of DPA, 730 lg of EDTA, 600 lg of APBA and 750 lg of cloxacillin. DPA was dissolved in dimethylsulphoxide (Sigma), whereas EDTA, APBA and cloxacillin were dissolved in sterile water. Disks were left to dry at room temperature for 30 min before they were used. Test procedure and interpretation of the combination disk assay A 0.5 McFarland inoculum was prepared and spread on cation-adjusted Mueller Hinton II agar plates (Becton-Dickinson, Cockeysville, MD, USA). Five disks were placed on each plate: meropenem 10 lg, meropenem 10 lg + DPA, meropenem 10 lg + EDTA, meropenem 10 lg + APBA and meropenem 10 lg + cloxacillin. An 5 mm in zone diameter around disks containing b-lactamase inhibitors, as compared with the disk with meropenem alone, was considered to be a positive result for DPA, EDTA and cloxacillin, whereas an 4 mm was considered to be a positive result for APBA. An 4 mm for APBA was selected because 4/34 KPC-producing K. pneumoniae isolates only had an of 4 mm and were thus not detected with a cutoff 5 mm. Commercial diagnostic tablets from Rosco (Rosco Diagnostica A/S, Taastrup, Denmark) were evaluated against the same isolates. Four tablets were placed on each plate: meropenem, meropenem + DPA, meropenem + APBA and meropenem + cloxacillin. The procedure was identical to the one used for applying in-house disks. An 5mmin zone diameter around tablets containing b-lactamase inhibitors, as compared with the tablet with meropenem alone, was considered a positive result, according to the manufacturer s instruction. Modified cloverleaf (Hodge) test) The modified cloverleaf test was performed according to Anderson et al. [4]. In brief, cation-adjusted Mueller-Hinton II
554 Clinical Microbiology and Infection, Volume 17 Number 4, April 2011 CMI agar plates (Becton-Dickinson) were inoculated with a 1 : 10 dilution of a 0.5 McFarland suspension of E. coli ATCC 25922 and streaked for confluent growth with a swab. A 10-lg imipenem disk (Oxoid) was placed in the centre, and each test isolate was streaked from the disk to the edge of the plate. The presence of a distorted inhibition zone after overnight incubation was interpreted as a positive result. Sensitivity and specificity The performance of the various disks for detection of different b-lactamases was determined with genotypically defined carbapenem resistance mechanisms as the reference standard. For each test and category of b-lactamase, sensitivity was calculated from the number of true-positive bacteria, whereas specificity was calculated from the number of truenegative bacteria. Results and Discussion The results are displayed in Table 1. Synergy with EDTA was observed in all 25 MBL producers, but also in 4/34 isolates of KPC-producing K. pneumoniae, in 1/9 isolates of K. pneumoniae with CTX-M-15 and porin loss, and in 4/9 isolates of OXA-48-producing K. pneumoniae. An association between OXA carbapenemase production and a false-positive EDTA test result has been described previously [13], but not for the OXA-48 enzyme. In comparison, all MBL producers were detected with disks containing DPA, and no isolates featuring other mechanisms of resistance displayed any synergy with DPA (Table 1). APBA synergy detected all isolates of KPC-producing K. pneumoniae. Many isolates with a combination of AmpC hyperproduction and porin loss were also positive in the APBA test, but these isolates additionally showed cloxacillin synergy, in contrast to the isolates with KPC enzymes. One OXA-48-producing K. pneumoniae isolate was positive in the APBA test, but not with cloxacillin. None of the other isolates in this strain collection was positive with APBA or cloxacillin. In summary, the sensitivity of the APBA test for detection of KPC was 100%, and the specificity was 98% if the additional criterion of a negative cloxacillin result was included (Table 2). The commercial tablets from Rosco performed almost as well as the in-house disks for detection of MBL and KPC b-lactamase (Table 3). The sensitivity was 100% in both cases, whereas one false-positive OXA-48-producing isolate was detected both with tablets containing DPA and with tablets containing APBA. The tablet containing cloxacillin identified only 2/5 isolates with combined AmpC hyperproduction and TABLE 1. Inhibitory activity of dipicolinic acid (DPA), aminophenylboronic acid (APBA) and cloxacillin against Enterobacteriaceae producing various b-lactamases Increase in zone diameter around meropenem disks containing b-lactamase inhibitors b and number of isolates positive/negative in each test Modified cloverleaf test (imipenem) DPA APBA Cloxacillin Positive Negative Positive Negative Positive Negative Positive Negative Porin loss a b-lactamase VIM (n = 21) NT 5 15 9 21 0 )1 0 0 0 21 0 2 0 0 21 21 0 IMP (n = 4) NT 8 14 10 4 0 )1 2 0 0 4 0 3 1 0 4 4 0 KPC (n = 34) NT )1 3 0 0 34 4 16 7 34 0 0 2 1 0 34 34 0 OXA-48 (n = 9) NT )4 4 0 0 9 )1 4 1 1 8 0 2 1 0 9 9 0 ESBL (n = 9) Yes )3 2 )1 0 9 )2 3 0 0 9 )1 0 0 0 9 0 9 AmpC (n = 9) Yes )3 1 )1 0 9 1 7 5 4 1 0 7 5 4 1 3 2 ESBL, extended-spectrum b-lactamase; NT, not tested. Bold type denotes expected synergy. a Loss of OmpK35 and/or OmpK36 in Klebsiella pneumoniae, or loss of OmpC and/or OmpF in Escherichia coli and Enterobacter cloacae. b A significant in zone diameter was defined as 5 mm for DPA, 4 mm for APBA and 5 mm for cloxacillin.
CMI Giske et al. Carbapenemase detection in Klebsiella pneumoniae 555 TABLE 2. Sensitivities and specificities of dipicolinic acid (DPA), aminophenylboronic acid (APBA), cloxacillin and the modified cloverleaf test for detecting various b-lactamases Non-wild type* to 1 carbapenem in K. pneumoniae Test b-lactamases sought by test(s) Sensitivity Specificity Synergy with APBA but not cloxacillin Synergy with APBA and cloxacillin Synergy with DPA only APBA-positive, cloxacillin-negative KPC 100 98 APBA-positive, cloxacillin-positive AmpC a 80 100 EDTA-positive MBL 100 88 DPA-positive MBL 100 100 Positive cloverleaf test result KPC, MBL, OXA-48 100 78 MBL, metallo-b-lactamase. a Combination of AmpC hyperproduction and porin loss. TABLE 3. Sensitivities and specificities of commercial disks from Rosco for detecting various b-lactamases Test b-lactamases sought by test(s) Sensitivity Specificity APBA-positive, cloxacillin-negative KPC 100 98 APBA-positive, cloxacillin-positive AmpC a 40 100 DPA-positive MBL 100 98 APBA, aminophenylboronic acid; DPA, dipicolinic acid; MBL, metallo-b-lactamase. a Combination of AmpC hyperproduction and porin loss. porin loss. More importantly, these two isolates were also the only ones that were positive with the tablet containing APBA, for which reason none of them was misidentified as a KPC-producing isolate. The modified cloverleaf test has been suggested as a useful method for detecting carbapenemases in Enterobacteriaceae. In our study, the sensitivity of this method was excellent, but the specificity was only 78%, the reason being that four isolates with the combination of AmpC hyperproduction and porin loss were positive. It should be noted that all of these isolates were Enterobacter spp., and the question of false-positive reactions in other species of Enterobacteriaceae, including those with acquired AmpC variants, remains unanswered. Furthermore, the subjective character of the test may cause problems of interpretation in laboratories lacking experience or with few carbapenem-resistant Enterobacteriaceae. Recently, Pasteran et al. and Carvalhaes et al. [14,15] have both reported similar findings, with poor specificity of the modified cloverleaf test. As a consequence of the poor performance of the test, Pasteran et al. [14] suggested yet another modification of the test. In the last modification of the cloverleaf test, inhibitors are added to the disk placed in the centre of the agar plate. KPC (or other class A carbapenemase) AmpC and porin loss MBL FIG. 1. Algorithm for interpretation of results with b-lactamase inhibitors. APBA, aminophenylboronic acid; DPA, dipicolinic acid; MBL, metallo-b-lactamase. * MIC above the epidemiological cut-off values defined by EUCAST (European Committee on Antimicrobial Susceptibility Testing). One limitation of the study is the large proportion of ST258 KPC-producing isolates, as other clones may theoretically be more difficult to detect with these methods. A certain level of clonal diversity was observed between the KPC producers [8], and strains that were not K. pneumoniae ST258 responded equally well to the applied b-lactamase inhibitors. Also, the recent emergence in Greece of K. pneumoniae isolates featuring both VIM and KPC b-lactamases simultaneously [16] poses another challenge to this protocol. Probably, such isolates can be identified with addition of both DPA and APBA to meropenem disks, but owing to the lack of control isolates with this mechanism of resistance, this remains speculative. In conclusion, synergy with APBA and cloxacillin was seen in isolates with porin loss and hyperproduction of AmpC, whereas synergy with APBA only had excellent specificity for KPC-producing isolates (Figure 1). DPA had excellent sensitivity and specificity for the detection of MBL-producing K. pneumoniae, whereas the poor specificity of EDTA questions the usefulness of this inhibitor. OXA-48-producing K. pneumoniae may become positive both in the DPA and in the APBA assay. It should also be noted that concomitant production of OXA-48 and the novel MBL NDM-1 has now been described (Karthikeyan et al., 20th ECCMID, 2010, Abstract P742), and such isolates will most probably be positive in the DPA assay. OXA-48 production should be considered for isolates that cannot be confirmed to produce a class A or class B carbapenemase, and investigated by PCR of bla OXA-48 in such cases. The main advantages of this novel assay over the modified cloverleaf test are the more objective analysis and the possibility of discriminating between various classes of carbapenemases. The addition of cloxacillin enabled discrimination between KPC producers and isolates with combined porin loss and
556 Clinical Microbiology and Infection, Volume 17 Number 4, April 2011 CMI AmpC hyperproduction. The current data are, perhaps, not in support of interpretative reading of test results based on the finding of a carbapenemase, at least for VIM-producing K. pneumoniae [17]. However, detection of carbapenemases in Enterobacteriaceae is still very important from an infection control and epidemiological point of view, and may impact greatly on patient management at institutions experiencing increasing rates of carbapenem resistance in Enterobacteriaceae. Acknowledgements We thank J. B. Patel, CDC, for providing clinical isolates of K. pneumoniae producing KPC b-lactamases. We would also like to thank A. Vatopoulos, Department of Microbiology, National School of Public Health, Athens, Greece, for providing clinical isolates of K. pneumoniae producing VIM-1 b- lactamase. Ø. Samuelsen is supported by a grant from the Northern Norway Regional Health Authority Medical Research Programme. Transparency Declaration C. G. Giske has received conference support from Calixa Therapeutics Inc. N. Woodford has received grants and conference support from major pharmaceutical companies, including all those that manufacture and supply carbapenems internationally (AstraZeneca, MSD, Johnson and Johnson/ Janssen-Cilag). The other authors have no conflicts of interest to declare. References 1. Giske CG, Sundsfjord AS, Kahlmeter G et al. Redefining extendedspectrum b-lactamases: balancing science and clinical need. J Antimicrob Chemother 2009; 63: 1 4. 2. Samuelsen Ø, Naseer U, Tofteland S et al. Emergence of clonally related Klebsiella pneumoniae isolates of sequence type 258 producing plasmid-mediated KPC carbapenemase in Norway and Sweden. J Antimicrob Chemother 2009; 63: 654 658. 3. Ørstavik I, Ødegaard K. A simple test for penicillinase production in Staphylococcus aureus. Acta Pathol Microbiol Scand B Microbiol Immunol 1971; 79: 855 856. 4. Anderson KF, Lonsway DR, Rasheed JK et al. Evaluation of methods to identify the Klebsiella pneumoniae carbapenemase in Enterobacteriaceae. J Clin Microbiol 2007; 45: 2723 2725. 5. Tsakris A, Kristo I, Poulou A et al. Evaluation of boronic acid disk tests for differentiating KPC-possessing Klebsiella pneumoniae isolates in the clinical laboratory. J Clin Microbiol 2009; 47: 362 367. 6. Vatopoulos A. High rates of metallo-b-lactamase-producing Klebsiella pneumoniae in Greece a review of the current evidence. Euro Surveill 2008; 13: 1 6. 7. Nordmann P, Cuzon G, Naas T. The real threat of Klebsiella pneumoniae carbapenemase-producing bacteria. Lancet Infect Dis 2009; 9: 228 236. 8. Kitchel B, Rasheed JK, Patel JB et al. Molecular epidemiology of KPCproducing Klebsiella pneumoniae in the United States: clonal expansion of MLST sequence type 258. Antimicrob Agents Chemother 2009; 53: 3365 3370. 9. Woodford N, Zhang J, Warner M et al. Arrival of Klebsiella pneumoniae producing KPC carbapenemase in the United Kingdom. J Antimicrob Chemother 2008; 62: 1261 1264. 10. Doumith M, Ellington MJ, Livermore DM et al. Molecular mechanisms disrupting porin expression in ertapenem-resistant Klebsiella and Enterobacter spp. clinical isolates from the UK. J Antimicrob Chemother 2009; 63: 659 667. 11. Kimura S, Ishii Y, Yamaguchi K. Evaluation of dipicolinic acid for detection of IMP- or VIM-type metallo-beta-lactamase-producing Pseudomonas aeruginosa clinical isolates. Diagn Microbiol Infect Dis 2005; 53: 241 244. 12. Galani I, Rekatsina PD, Hatzaki D et al. Evaluation of different laboratory tests for the detection of metallo-b-lactamase production in Enterobacteriaceae. J Antimicrob Chemother 2008; 61: 548 553. 13. Segal H, Elisha BG. Use of Etest MBL strips for the detection of carbapenemases in Acinetobacter baumannii. J Antimicrob Chemother 2005; 56: 598. 14. Pasteran F, Mendez T, Rapoport M, Guerriero L, Corso A. Controlling the false positive results of the Hodge and Masuda assays for class A carbapenemase detection in species of Enterobacteriaceae. J Clin Microbiol 2010; 48: 1323 1332. 15. Carvalhaes CG, Picão RC, Nicoletti AG, Xavier DE, Gales AC. Cloverleaf test (modified Hodge test) for detecting carbapenemase production in Klebsiella pneumoniae: be aware of false positive results. J Antimicrob Chemother 2010; 65: 249 251. 16. Giakkoupi P, Pappa O, Polemis M et al. Emerging Klebsiella pneumoniae isolates coproducing KPC-2 and VIM-1 carbapenemases. Antimicrob Agents Chemother 2009; 53: 4048 4050. 17. Daikos GL, Petrikkos P, Psichogiou M et al. Prospective observational study of the impact of VIM-1 metallo-b-lactamase on the outcome of patients with Klebsiella pneumoniae bloodstream infections. Antimicrob Agents Chemother 2009; 53: 1868 1873.