Detection and characterization of extended-spectrum b-lactamases among bloodstream isolates of Enterobacter spp. in Hong Kong,

Journal of Antimicrobial Chemotherapy (2005) 55, 326 332 doi:10.1093/jac/dki010 Advance Access publication 28 January 2005 Detection and characterization of extended-spectrum b-lactamases among bloodstream isolates of Enterobacter spp. in Hong Kong, 2000 2002 P. L. Ho 1 *, Ricky H. L. Shek 2, K. H. Chow 1, R. S. Duan 1, Gannon C. Mak 1, Eileen L. Lai 1, W. C. Yam 1, Kenneth W. Tsang 3 and W. M. Lai 2 JAC 1 Division of Infectious Diseases, Department of Microbiology and Centre of Infection and 2 Department of Clinical Pathology, United Christian Hospital; 3 Department of Medicine, Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China Received 19 August 2004; returned 23 October 2004; revised 23 November 2004; accepted 6 December 2004 Objectives: A total of 139 consecutive and non-duplicate bloodstream isolates of Enterobacter spp. collected from inpatients in Hong Kong during 2000 2002 were studied for production of extendedspectrum b-lactamases (ESBLs). Methods: All isolates were evaluated by the modified double-disc synergy test (m-ddst), the combined disc method (CDM) and the three-dimensional (3D) test. The m-ddst and CDM were modified by the use of cefepime discs. b-lactamases were characterized by isoelectric focusing and PCR sequencing using specific primers. Results: ESBLs were identified in nine isolates (overall 6.5%), including seven of 39 (17.9%) Enterobacter hormaechei, one of 27 (3.7%) Enterobacter aerogenes and the only Enterobacter intermedius strain. The E. intermedius strain was positive only in the 3D test but not in the other two tests. The other eight strains were positive in all three tests. No ESBL was detected in the other species, including non-hormaechei members of the Enterobacter cloacae complex (n 5 61), Enterobacter agglomerans (n 5 7), Enterobacter gergoviae (n 5 4) and Enterobacter sakazakii (n 5 1). The ESBL content included five different CTX-M enzymes (CTX-M-9, CTX-M-13, CTX-M-14, CTX-M-24 and a novel CTX-M-2-like b-lactamase), SHV-12 (n 5 2) and unidentifiable ESBLs with a pi of 7.7 or 7.9 in two strains. The seven ESBL-producing E. hormaechei were genotyped by pulsed-field gel electrophoresis and were found to be unrelated to each other. In three of the CTX-M-producing strains, ISEcp1-like elements, including promoters for the b-lactamase gene, were found. Conclusions: Our data underscore the diversity of CTX-M enzymes among Enterobacter spp. in Hong Kong. Keywords: cephalosporin resistance, ESBLs, CTX-M enzymes Introduction Among Enterobacteriaceae, production of extended-spectrum b-lactamases (ESBLs) is increasingly recognized as a cause of resistance to expanded spectrum cephalosporins. In large hospitals, this resistance mechanism is found in 10% 40% of Klebsiella spp. and Escherichia coli. 1 Less is known about the incidence of ESBL in Enterobacter because ESBL phenotypic screening of Enterobacteriaceae is commonly performed in the clinical laboratories only for E. coli and Klebsiella spp. Detection of ESBL in Enterobacter spp. is more complicated because high-level AmpC production interferes with tests that rely on synergism between clavulanic acid and a third-generation cephalosporin. Thus, recent studies have addressed this by using cefepime to replace the third-generation cephalosporins that were conventionally used. 2,3 With this modification, Tzelepi et al. 3... *Correspondence address. Division of Infectious Diseases, Department of Microbiology and Centre of Infection, The University of Hong Kong, Queen Mary Hospital, Pokfulam Road, Pokfulam, Hong Kong SAR, China. Tel: +852-2855-4897; Fax: +852-2855-1241; E-mail: plho@hkucc.hku.hk... 326 JAC vol.55 no.3 q The British Society for Antimicrobial Chemotherapy 2005; all rights reserved.

ESBLs among Enterobacter spp. in Hong Kong reported ESBL rates of 25% and 58% in a consecutive collection of Enterobacter cloacae and Enterobacter aerogenes, respectively, obtained in a general hospital in Greece during 1998 1999. Similarly, Pai et al. 4 tested blood isolates of Enterobacter spp. in a Korean tertiary hospital during 1994 2001 and reported an overall ESBL prevalence of 43%. These findings highlight the potential for overlooking high ESBL rates in Enterobacter spp. In this study, we examined the production of AmpC and ESBL in blood isolates of Enterobacter spp. collected over a 3 year period at two regional hospitals in Hong Kong. Materials and methods Bacterial strains, identification, susceptibility testing and patient data Blood isolates of Enterobacter isolated during 2000 2002 in two general regional hospitals (A and B with 1300 and 1100 beds, respectively) were evaluated in this study. Both hospitals provide acute care for all the major specialties. Hospital A is also a teaching hospital and is the major bone marrow and liver transplant centre in Hong Kong. Duplicate isolates were excluded by the first isolate per patient method. 5 The VITEK GNI system (biomérieux Vitek Inc., Hazelwood, MO, USA) was used for bacterial identification. For accurate identification to the species level, the glucose oxidation test was carried out as described. 6,7 Antibiotic susceptibilities to amoxicillin clavulanic acid, piperacillin tazobactam, cefoperazone sulbactam, cefoxitin, ceftazidime, ceftriaxone, cefepime, imipenem, gentamicin, amikacin, ciprofloxacin, chloramphenicol, sulfamethoxazole, trimethoprim, co-trimoxazole and tetracycline were tested by the disc diffusion method using Mueller Hinton agar (Oxoid, Basingstoke, UK) and interpreted according to the NCCLS. 8 All antibiotic discs were obtained commercially (BBL, Becton Dickinson, Cockeysville, MD, USA). For the ESBL-producing strains, the patient records were reviewed for the following data and risk factors: date of isolation, date of admission, age, gender, old-age home residence, type of infection, co-morbidity and previous hospitalization. Co-morbidity was defined as any underlying illness that might increase the risk of infection. Healthcare-associated ESBL cases were empirically defined as patients with (1) an infection identified after 48 h of admission to a hospital; or (2) a history of recent hospitalization in the preceding 6 weeks, or residence in an old-age home. Cases that had none of the above were classified as community-associated. Detection of ESBL and AmpC Three methods were used to detect ESBL production. In the modified double-disc synergy test (m-ddst), cefepime discs (30 mg) were placed adjacent to an amoxicillin clavulanic acid disc (20 10 mg) at inter-disc distances (centre to centre) of 20, 25 and 30 mm. 2 A clear extension of the edge of the cefepime disc inhibition zone towards the disc containing clavulanic acid was interpreted as positive for ESBL production. In the combined disc method (CDM), we modified an NCCLS procedure, in that a cefepime disc (30 mg) and a cefepime clavulanic acid disc (30 10 mg) were used. A >_ 5 mm increase in zone diameter for the cefepime clavulanic acid disc versus the cefepime disc was interpreted as production of ESBL. In the three-dimensional extract test (3D test), enzyme extracts from the test strains were examined for their abilities to hydrolyse ceftriaxone, ceftazidime and cefoxitin as described. 9 For extraction of b-lactamase, the Lindberg Normark lysozyme method was used. 10 Lysozyme was obtained from Sigma Chemical Co. (St Louis, MO, USA). E. coli ATCC 25922 was used as the surface organism. To determine the effect of b-lactamase inhibitors on hydrolysis of the b-lactams, enzyme extracts were mixed with inhibitor(s) (cloxacillin 1 mm, clavulanic acid 4 mg/l or both) and pre-incubated at 37 8C for 15 min before testing with the 3D-test procedure. The criteria for ESBL production were hydrolysis of ceftriaxone and/or ceftazidime (when enzyme extracts were tested in the presence of the AmpC inhibitor, cloxacillin), and susceptible to inhibition by clavulanic acid. The criteria for AmpC production were hydrolysis of cefoxitin that was resistant to inhibition by clavulanic acid but susceptible to inhibition by cloxacillin. Clavulanic acid and cloxacillin powders were obtained from GlaxoSmith- Kline (Hong Kong) and Sigma Chemical Co., respectively. Since production of ESBL could affect interpretive reading of the antibiogram, results of the 3D test were taken into consideration in evaluating the expression of the chromosomal AmpC cephalosporinase, as follows: basal expression if the strain was susceptible to cefoxitin and there was no hydrolysis of cefoxitin; inducible expression if the strain was resistant to cefoxitin and there was no hydrolysis of cefoxitin; de-repressed (i.e. hyper-expression) if the strain was resistant to cefoxitin and ceftriaxone, and there was hydrolysis of cefoxitin. 11 Isoelectric focusing (IEF) Analytic IEF was performed in ampholine gel (ph 3.0 to 10.0; Pharmacia, Hong Kong, China) using lysozyme-based enzyme extract on an HE 950 isoelectric focusing apparatus (Hoefer Scientific Instruments, CA, USA). 10,12 Isolates that had an ESBL phenotype in the screening tests, but in which the enzyme type remained unidentified by the PCR and sequencing strategies (see below) were examined. The pi value of each enzyme was determined by spreading nitrocefin on the gel surface. 13 Preparations from standard strains known to harbour TEM-1, TEM-8, SHV-4 and SHV-5 were used as markers. 14 ESBL activity was detected by the substrate gel overlay method using cefotaxime at 1 mg/l and E. coli ATCC 25922 as an indicator. Mating experiments Filter conjugation was carried out in filters with E. coli J53Az r as the recipient. 12 Cultures of the donor and recipient were collected on sterilized filters (Whatman International Ltd, Maidstone, UK) and the filter was incubated on the surface of an LB agar plate overnight. Transconjugants were selected on trypticase soy agar plates containing sodium azide (150 mg/l; Sigma Chemical Co.) for counterselection and cefotaxime (0.5 or 1 mg/l; Sigma Chemical Co.) to select for plasmid-encoded resistance. ESBLs in the transconjugants were tested by the double-disc synergy test using cefotaxime, amoxicillin clavulanic acid and ceftazidime at inter-disc distances (centre to centre) of 20 and 25 mm. Molecular analysis Template DNA preparation and PCR amplifications were carried out as described previously 15 17 in a final volume of 50 ml with 2 ml of template DNA and 0.4 mm of each primer. PCR amplification was performed in a Perkin-Elmer thermal cycler 9600. Amplification parameters involved 30 cycles with 1 min at 95 8C, 45 s at 55 8C and 45 s at 72 8C. A final elongation was carried out at 72 8C for 10 min. The amplicons generated by the primers specific for bla TEM, bla SHV or bla CTX-M were sequenced directly after purification with the High Pure PCR Product Purification kit, as directed by the manufacturer 327

P. L. Ho et al. (Roche Diagnostics, Mannheim, Germany). The primers TEM-F 5 0 -TTC TTG AAG ACG AAA GGG C-3 0 (position 882 900, accession AF126482) and TEM-R 5 0 -ACG CTC AGT GGA ACG AAA AC-3 0 (position 2089 2070, accession AF126482) were used to amplify a 1207 bp fragment containing the entire bla TEM gene and the upstream promoter. The nomenclature used for the bla TEM variants and promoters was as previously proposed. 18 The primers SHV-1017F 5 0 -GCC CGG GTT ATT CTT ATT TGT CGC-3 0 (position 130 151, accession AF124984) and SHV-1017B 5 0 -TCT TTC CGA TGC CGC CGC CAG TCA-3 0 (position 1143 1120, accession AF124984) were used for the entire bla SHV gene. The CTX-M consensus primers MA1 forward 5 0 -SCV ATG TGC AGY ACC AGT AA-3 0 (position 270 289, accession 92506) and MA2 backward 5 0 -CCG CRA TAT GRT TGG TGG TR-3 0 (position 794 813, accession X92506) were used as an initial screen. Subsequently, other primers specific for each CTX-M cluster were used to amplify the entire bla CTX-M gene. The CTX-M-9 group primers were M9U forward 5 0 -ATG GTG ACA AAG AGA GTG CA-3 0 (position 112 131, accession D89862) and M9L backward 5 0 -CCC TTC GGC GAT GAT TCT C-3 0 (position 957 975, D89862). For the CTX-M-positive strains, the region immediately upstream from bla CTX-M was explored by direct sequencing of amplicons obtained with forward primer ISEcp1 U1 5 0 -AAA AAT GAT TGA AAG GTG GT-3 0 (position 6285 6304, accession AF458080) and the consensus reverse MA2. 17 All amplicons were generated and sequenced at least twice by automated PCR cycle sequencing with the ByeDye dideoxynucleotide chain termination method in a ABI PRISM 377 Genetic Analyzer (Perkin-Elmer, Foster City, CA, USA). Amino acids in the b-lactamases were numbered according to Ambler et al. 19 Positive (E. coli J53 Azi r with encoded bla TEM-1, bla SHV-1 and bla CTX-M-1, 2 AND 9 ) and negative (E. coli J53 Azi r and water) controls were used in all reactions. 14 The subset of ESBL-producing Enterobacter hormaechei was examined further by pulsed-field gel electrophoresis (PFGE). Xba I (Amersham Pharmacia Biotech, Little Chalfont, UK) was used for macrorestriction digestion of genomic DNA. The fragments were resolved in 1% gel in a CHEP Mapper XA (BioRad) for 22 h at 6 V/cm gradient and an included angle of 1208. The switching time ramped from an initial 2.2 s to a final 54.2 s. Results were interpreted as described previously. 20 Results Bacterial identification A total of 182 Enterobacter spp. were isolated from blood cultures in the two hospitals. Forty-three duplicate isolates were excluded by the first isolate per patient approach, thus leaving 139 isolates (96 from hospital A and 43 from hospital B) for phenotypic testing. This collection comprised 99 E. cloacae complex, 27 E. aerogenes, seven Enterobacter agglomerans, four Enterobacter gergoviae, one E. intermedius and one Enterobacter sakazakii. According to the VITEK system, the 99 members of the E. cloacae complex were identified as E. cloacae (n = 96), Enterobacter asburiae (n = 2) and E. hormaechei (n = 1). Thirty-seven of the 96 (38.5%) E. cloacae strains and the initial E. hormaechei gave positive results in the glucose oxidation test and thus these 38 isolates were designated as E. hormaechei. Frequency of ESBL production and AmpC expression Of the 139 isolates tested, the distribution of ESBL phenotypes and AmpC phenotypes is summarized in Table 1. Overall, an ESBL phenotype was detected in nine strains (6.5%), including 7/39 (17.9%) E. hormaechei, 1/27 (3.7%) E. aerogenes and 1/1 (100%) E. intermedius. Eight of the nine strains had an ESBL phenotype in all three ESBL screening tests (Table 2). The remaining E. intermedius strain gave a positive result in the 3D test but not the m-ddst and the CDM screening. For the m-ddst, sensitivity for ESBL detection was higher at 20 mm inter-disc distance (88.8%) than at 25 (55.5%) or 30 mm (33.3%). In the CDM test, the zone augmentations with cefepime clavulanic acid for the non-esbl producers were 3 mm or less. Except for the E. intermedius strain 201I9, zone augmentations for the ESBL producers were 5 mm or greater (range 5 18 mm). In the 3D test, nine strains had clavulanic acid-susceptible hydrolysis of ceftriaxone, compared with three strains only for that of ceftazidime. Among the 139 strains, the expression of AmpC was basal in 10 (7.2%), inducible in 96 (69.1%) and de-repressed in 33 (23.7%). No association between ESBL production and type of AmpC expression was found (P = 0.2). Statistical analysis The x 2, Fisher s exact or Student s t-test was used for statistical analysis. A P value of < 0.05 was considered significant. For the purpose of analysis, resistant included both intermediate and resistant categories. The inhibition zone diameter cutoff for cefepime as an initial ESBL screen was determined by the receiver operating characteristic (ROC) curve. Nucleotide sequence accession number The sequence data for the ISEcp1B segment has been submitted to the GenBank nucleotide sequence database under accession number AY830894 (E. hormaechei 3701), AY830895 (E. hormaechei 02C9), AY830896 (E. aerogenes 255B4). The sequence for the CTX-M-2-like gene in strain Enterobacter intermedius 201I9 was assigned under accession number AY835399. Table 1. Distribution of ESBL- and AmpC-phenotypes among 139 Enterobacter spp. Species (n) ESBL basal No. of isolates AmpC inducible De-repressed Enterobacter cloacae complex (99) E. cloacae (59) 0 2 40 17 E. hormaechei (38) 7 1 28 9 E. asburiae (2) 0 0 0 2 E. aerogenes (27) 1 2 20 5 E. agglomerans (7) 0 3 4 0 E. gergoviae (4) 0 1 3 0 E. sakazakii (1) 0 0 1 0 E. intermedius (1) 1 1 0 0 Total with phenotype 9 10 96 33 328

ESBLs among Enterobacter spp. in Hong Kong Table 2. b-lactamase detection results for nine strains of ESBL-producing Enterobacter spp. m-ddst a 3D test Species Strain 20 25 30 CDM b CRO CAZ AmpC E. hormaechei 254G9 + + + + + inducible E. hormaechei 4952 + + + inducible E. hormaechei 192C3 + + + + + + inducible E. hormaechei 2662 + + + + + + inducible E. hormaechei 3701 + + + + + de-repressed E. hormaechei 3806 + + + inducible E. hormaechei 02C9 + + + inducible E. aerogenes 255B4 + + + + basal E. intermedius 201I9 + basal a Cefepime discs (30 mg) were placed adjacent to an amoxicillin clavulanic acid disc (20 10 mg) at the indicated interdisc distances (centre to centre) of 20 mm, 25 mm and 30 mm. b Cefepime (30 mg) and cefepime clavulanic acid (30 10 mg) discs were used. 3D test, three-dimensional extract test; CAZ, ceftazidime; CDM, modified combined-disc method; CRO, ceftriaxone; m-ddst, modified double-disc synergy test;, negative; +, positive. Antimicrobial susceptibility Strains with an ESBL phenotype were more likely than those without to have resistance to gentamicin (5/9 versus 1/130, P < 0.001), amikacin (3/9 versus 0/130, P < 0.001) ciprofloxacin (4/9 versus 8/130, P < 0.001) and co-trimoxazole (5/9 versus 4/130, P < 0.001). In contrast, resistance rates for chloramphenicol (3/9 versus 28/130, P = 0.6) and tetracycline (4/9 versus 23/130, P = 0.1) were similar for the two groups. Imipenem was the only agent tested that was uniformly active against all the isolates. Based on the inhibition zones, 29.5% (41/139) were resistant to ceftazidime and/or ceftriaxone. Overall, resistance to piperacillin tazobactam was 21.6% (30/139), being higher in the de-repressed AmpC group (69.7%) than the inducible (6.3%) and basal (0%) AmpC groups (P < 0.01). On the basis of the inhibition zone diameters alone, only two of the nine ESBL-producing isolates were resistant to cefepime at standard breakpoints ( <_ 17 mm). If a separate ESBL screening test had not been performed, these ESBL-producing isolates would have been missed by the routine disc diffusion test. The mean ( ± S.D.) cefepime disc inhibition zone diameters for the isolates without ESBL were larger than those with ESBL (32 ± 4.1 mm versus 20 ± 5.9 mm, P < 0.01). Antimicrobial disc tests for initial ESBL screen In order to evaluate the effectiveness of extending the NCCLS E. coli and Klebsiella spp. ESBL breakpoints for initial screening of Enterobacter spp., interpretation of the disc results for ceftriaxone and ceftazidime were tabulated against the ESBL production status (Table 3). The data showed that both agents lacked specificity, although the sensitivity was 100% for ceftriaxone. To determine the utility of using a revised cefepime disc cutoff for the initial ESBL screen, the inhibition zone diameters were analysed against the ESBL production status in a scatter plot (data not shown) and the result was evaluated by the ROC method. At cutoff <_ 25 mm, the best discrimination was achieved with sensitivity and specificity of 89.9% and 93.8%. Characteristics of the ESBL-producing strains The nine ESBL-producing strains represented two communityacquired infections and seven healthcare-associated infections. The sources of bacteraemia were as follows: urinary (n = 4), neutropenic fever (n = 3), biliary (n = 1) and wound (n = 1). The b- lactamase-specific PCR and sequencing revealed five different CTX-M-type enzymes, including CTX-M-9, -13, -14 and -24 that belong to the CTX-M-9 cluster and a novel CTX-M-2-like enzyme (Table 4). Sequencing of the gene included the region covering the mature peptide except for the terminal nine amino acids. The novel CTX-M-2-like b-lactamase differs from Table 3. Distribution of results by extending the NCCLS E. coli and Klebsiella spp. ESBL initial screen method to 139 Enterobacter spp. Category (no. of isolates) Agent and NCCLS ESBL initial screen result a ESBL Non-ESBL Sensitivity/ specificity b Ceftriaxone 100%/63.1% Positive 9 48 c Negative 0 82 Ceftazidime 77.8%/72.3% Positive 7 36 d Negative 2 94 a The NCCLS disc inhibition zone diameter breakpoints for ESBL initial screen of E. coli and Klebsiella spp. were: <_ 25 mm for ceftriaxone (30 mg) and <_ 22 mm for ceftazidime (30 mg). The equivalent MIC breakpoints were <_ 1 mg/l for both agents. b Sensitivity was defined as the proportion of ESBL producers with a positive initial ESBL screen result. Specificity was defined as the proportion of non- ESBL producers with a negative initial ESBL screen result. c AmpC expression was inducible in 16 isolates and de-repressed in 32 isolates. d AmpC expression was basal in one isolate, inducible in five isolates and derepressed in 30 isolates. 329

P. L. Ho et al. Table 4. Characteristics of ESBLs in nine Enterobacter blood isolates Species Strain Associated resistance markers b-lactamase content Conjugation frequency Resistance co-transferred E. hormaechei 254G9 AMK, CIP, CHL, GEN, SMX, SHV-11, pi 7.9 TET, TMP E. hormaechei 4952 AMK, CIP, CHL, GEN, SMX, TEM-1b, pi 7.7 10 7 AMK, CHL, GEN TET, TMP E. hormaechei 192C3 CIP, CHL, GEN, SMX, TET, SHV-12 TMP E. hormaechei 2662 GEN, SMX, TET, TMP SHV-12, TEM-1b E. hormaechei 3701 AMK, GEN, SMX, TMP CTX-M-13 10 3 none E. hormaechei 3806 CIP. SMX, TMP CTX-M-9 10 3 TMP E. hormaechei 02C9 none CTX-M-14, SHV-12 10 3 none E. aerogenes 255B4 none CTX-M-24, SHV-1 10 3 none E. intermedius 201I9 none CTX-M-2 like AMK, amikacin; CIP, ciprofloxacin; CHL, chloramphenicol; GEN, gentamicin; SMX, sulfamethoxazole; TET, tetracycline; TMP, trimethoprim. KLUA-9 (accession CAD20590) in Kluyvera ascorbata by two amino acids: A26T and D269E. SHV-12 was found in three strains and TEM-1b in two. In the two strains with bla TEM-1B, the weak P3 promoter was identified. Two strains had unidentifiable ESBLs with a pi of 7.7 or 7.9. The seven strains of E. hormaechei were genotyped by PFGE. The findings revealed that they were unrelated to each other. In conjugation experiments, the ESBL determinant could be transferred in five of the nine strains. Three of the strains could be amplified with primers ISEcp U1 and MA2. Sequencing of the PCR products revealed a right inverted repeat (IRR) and a 10 putative promoter region in all three strains. In E. hormaechei 02C9 and E. aerogenes 255B4, there was a 42 bp interval between the IRR and the start codon of bla CTX-M. The sequence for E. hormaechei 3701 differs in having an additional 120 bp insertion to the right of the IRR. The 120 bp additional sequence shares 99% identity with those upstream of the chromosomal bla KLUY-2 of a Kluyvera species (accession AY623935). Discussion This study showed that the incidence of ESBL among certain bacteraemic Enterobacter isolates in Hong Kong is high. The relevance of careful speciation is highlighted by the finding that ESBL carriage was higher in E. hormaechei than among other species. The nomenspecies E. cloacae as a heterogeneous group is well known and the taxonomical status of closely related members previously put under this nomenspecies is evolving. In the most recent edition of Manual of Clinical of Clinical Microbiology, E. hormaechei is listed as a distinct species, differing from E. cloacae by not producing acid from D-sorbitol and melibiose. 21 However, strains of E. hormaechei that were positive for both have been reported. 22 By the D-glucose oxidation test, 6 we found that many strains identified as E. cloacae in VITEK GNI actually belong to the species E. hormaechei. In general, previous studies that evaluated the prevalence and diversity of ESBL in E. cloacae did not carefully segregate E. hormaechei from the other non-hormaechei members. Thus, it was difficult to relate the species-specific rates in our sample to those of others. For E. cloacae, the reported prevalence of ESBL among clinical isolates from hospitals have ranged from 1% (two of 201 isolates) for the Kinki region in Japan, 23 6% (three of 50 isolates) for Beijing in China, 24 14% (six of 44 isolates) for Buenos Aires in Argentina 25 and up to 43% (24 of 56 isolates) for Thessaloniki in Greece. 3 The present study found that the simple application of NCCLS criteria designed for ESBL detection in E. coli and Klebsiella spp. would lead to many false-positive results in Enterobacter spp., in contrast to specificities of 97% 99% reported for screening of E. coli and Klebsiella spp. 26 In our collection of Enterobacter spp., our findings demonstrated that it may be possible to increase the test specificity by including cefepime in initial screening. In agreement with other studies, 3,25 our data showed that both CDM based on cefepime and the m-ddst are valuable for confirming ESBL in Enterobacter spp. According to Quinteros et al., 25 ESBL confirmatory testing with the two pairs of discs containing ceftazidime and cefotaxime alone, and in combination with clavulanic acid, lacks sensitivity in E. cloacae. The high diversity of ESBLs among Enterobacter spp. in Hong Kong is intriguing. In addition to having CTX-M and SHV classes of ESBL, each b-lactamase combination in the nine strains was distinct. For the SHV class of ESBL, only SHV-12 was present among the strains. Elsewhere, this ESBL is known to exist among E. aerogenes in Italy and among E. cloacae in Beijing, 24 Taiwan, 27 Korea 28 and Thailand. 29 In agreement with previous studies, 30 we demonstrated that four of the five isolates carrying bla CTX-M genes were able to transfer the ESBL phenotype in mating experiments, suggesting the spread of bla CTX-M genes by conjugative plasmids. Interestingly, similar b-lactamase combinations involving SHV-12, CTX-M-9, CTX-M-13 and CTX-M-14 were known to exist among Enterobacteriaceae collected from a hospital in Guangzhou. 31 Millions of passengers are travelling between Guangzhou and Hong Kong annually, due to the linkages between Hong Kong and mainland China. The finding thus indicates a probable cross-boundary spread of ESBL-producing organisms or their determinants. As suggested by Poirel et al., 32 the ISEcp1-like element observed upstream of the bla CTX-M genes in the present study may be involved in the translocation and dissemination of these b-lactamase genes. In addition, the ISEcp1-like element provides the promoters that drive expression of the b-lactamase. 32 At 330

ESBLs among Enterobacter spp. in Hong Kong present, five major clusters of CTX enzymes are known. 30 Based upon analysis of the sequence upstream of some bla CTX-M subtypes, Kluyvera georgiana and K. ascorbata have been suggested as the progenitors of the CTX-M-8 and CTX-M-2 clusters, respectively. 17,33,34 In this study, the bla CTX-M-13 and a 120 bp sequence upstream fragment in an E. hormaechei strain were found to be similar to the bla KLUY-2 and the chromosomal sequence upstream of the b-lactamase gene in an unspecified Kluyvera species recently deposited in the GenBank (Olson et al. unpublished data, accession AY623935). This provides evidence for the probable progenitor of the CTX-M-9 cluster. Interestingly, the 42 bp non-coding sequence between ISEcp1 and bla CTX-M in two of our isolates appears to be signatory of members in the CTX-M-9 cluster. 30 In a blastn search, an identical interval has been reported upstream of CTX-M-9 (AF174129), -14 (AJ416341), -16 (AY033516), -17 (AF454633), -19 (AF458080), -21 (AJ416346) and Toho-2 (D89862). None was found in those that belong to the other clusters. The isolates containing the 42 bp interval have diverse geographic origins including China, France, Japan, Spain, Japan and Vietnam. 35 In conclusion, this study found that the prevalence of ESBL among Enterobacter spp. in Hong Kong is high and that their b- lactamase content is diverse. Our finding adds to the increasing recognition of CTX-M enzymes in the Far East and further emphasizes the need for screening ESBL in clinical isolates of Enterobacter spp. Acknowledgements Part of the work was submitted by Ricky H. L. Shek to the University of Hong Kong in partial fulfilment of the requirement for a Masters in Medical Sciences. The work is supported by research grants from the Research Fund for the Control of Infectious Diseases of the Health, Welfare and Food Bureau of the Hong Kong SAR Government and from the UDF Project- Research Centre of Emerging Infection Diseases. References 1. Winokur, P. L., Canton, R., Casellas, J. M. et al. (2001). Variations in the prevalence of strains expressing an extendedspectrum b-lactamase phenotype and characterization of isolates from Europe, the Americas, and the Western Pacific region. Clinical Infectious Diseases 32, Suppl. 2, S94 103. 2. Pitout, J. D. D., Reisbig, M. D., Venter, E. C. et al. (2003). Modification of the double-disk test for detection of Enterobacteriaceae producing extended-spectrum and AmpC b-lactamases. Journal of Clinical Microbiology 41, 3933 5. 3. Tzelepi, E., Giakkoupi, P., Sofianou, D. et al. (2000). Detection of extended-spectrum b-lactamases in clinical isolates of Enterobacter cloacae and Enterobacter aerogenes. Journal of Clinical Microbiology 38, 542 6. 4. Pai, H., Hong, J. Y., Byeon, J. H. et al. (2004). High prevalence of extended-spectrum b-lactamase-producing strains among blood isolates of Enterobacter spp. collected in a tertiary hospital during an 8- year period and their antimicrobial susceptibility patterns. Antimicrobial Agents and Chemotherapy 48, 3159 61. 5. National Committee for Clinical Laboratory Standards. (2002). Analysis and Presentation of Cumulative Antimicrobial Susceptibility Test Data: Approved Guideline M39-A. NCCLS, Wayne, PA, USA. 6. Bouvet, O. M., Lenormand, P., Ageron, E. et al. (1995). Taxonomic diversity of anaerobic glycerol dissimilation in the Enterobacteriaceae. Research in Microbiology 146, 279 90. 7. Stock, I., Gruger, T. & Wiedemann, B. (2001). Natural antibiotic susceptibility of strains of the Enterobacter cloacae complex. International Journal of Antimicrobial Agents 18, 537 45. 8. National Committee for Clinical Laboratory Standards. (2002). Performance Standards for Antimicrobial Susceptibility Testing: Twelfth Informational Supplement M100-S12. NCCLS, Wayne, PA, USA. 9. Vercauteren, E., Descheemaeker, P., Ieven, M. et al. (1997). Comparison of screening methods for detection of extended-spectrum b-lactamases and their prevalence among blood isolates of Escherichia coli and Klebsiella spp. in a Belgian teaching hospital. Journal of Clinical Microbiology 35, 2191 7. 10. Lindberg, F., Westman, L. & Normark, S. (1985). Regulatory components in Citrobacter freundii ampc b-lactamase induction. Proceedings of the National Academy of Sciences, USA 82, 4620 4. 11. Hoffmann, H. & Roggenkamp, A. (2003). Population genetics of the nomenspecies Enterobacter cloacae. Applied and Environmental Microbiology 69, 5306 18. 12. Siu, L. K., Lo, J. Y.C., Yuen, K. Y. et al. (2000). b-lactamases in Shigella flexneri isolates from Hong Kong and Shanghai and a novel OXA-1-like b-lactamase, OXA-30. Antimicrobial Agents and Chemotherapy 44, 2034 8. 13. Ho, P. L., Cheung, T. K., Yam, W. C. et al. (2002). Characterization of a laboratory-generated variant of BPS b-lactamase from Burkholderia pseudomallei that hydrolyses ceftazidime. Journal of Antimicrobial Chemotherapy 50, 723 6. 14. Ho, P. L., Chow, K. H., Yuen, K. Y. et al. (1998). Comparison of a novel, inhibitor-potentiated disc-diffusion test with other methods for the detection of extended-spectrum b-lactamases in Escherichia coli and Klebsiella pneumoniae. Journal of Antimicrobial Chemotherapy 42, 49 54. 15. Brinas, L., Zarazaga, M., Saenz, Y. et al. (2002). b-lactamases in ampicillin-resistant Escherichia coli isolates from foods, humans, and healthy animals. Antimicrobial Agents and Chemotherapy 46, 3156 63. 16. Gruteke, P., Goessens, W., Van Gils, J. et al. (2003). Patterns of resistance associated with integrons, the extended-spectrum b- lactamase SHV-5 gene, and a multidrug efflux pump of Klebsiella pneumoniae causing a nosocomial outbreak. Journal of Clinical Microbiology 41, 1161 6. 17. Saladin, M., Cao, V. T., Lambert, T. et al. (2002). Diversity of CTX-M b-lactamases and their promoter regions from Enterobacteriaceae isolated in three Parisian hospitals. FEMS Microbiology Letters 209, 161 8. 18. Leflon-Guibout, V., Heym, B. & Nicolas-Chanoine, M. H. (2000). Updated sequence information and proposed nomenclature for bla TEM genes and their promoters. Antimicrobial Agents and Chemotherapy 44, 3232 4. 19. Ambler, R. P., Coulson, A. F., Frere, J. M. et al. (1991). A standard numbering scheme for the class A b-lactamases. Biochemical Journal 276, 269 70. 20. Tenover, F. C., Arbeit, R. D., Goering, R. V. et al. (1995). Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. Journal of Clinical Microbiology 33, 2233 9. 21. Abbott, S. L. (2003). Klebsiella, Enterobacter, Citrobacter, Serratia, Plesiomonas, and other Enterobacteriaceae. In Manual of Clinical Microbiology, 8th edn. (eds Murray, P. R., Baron, E. J., Jorgensen, J. H. et al.), pp. 684 96. ASM Press, Washington, DC, USA. 22. Davin-Regli, A., Bosi, C., Charrel, R. et al. (1997). A nosocomial outbreak due to Enterobacter cloacae strains with the E. hormaechei genotype in patients treated with fluoroquinolones. Journal of Clinical Microbiology 35, 1008 10. 23. Yamasaki, K., Komatsu, M., Yamashita, T. et al. (2003). Production of CTX-M-3 extended-spectrum b-lactamase and IMP-1 331

P. L. Ho et al. metallo b-lactamase by five Gram-negative bacilli: survey of clinical isolates from seven laboratories collected in 1998 and 2000, in the Kinki region of Japan. Journal of Antimicrobial Chemotherapy 51, 631 8. 24. Wang, H., Kelkar, S., Wu, W. et al. (2003). Clinical isolates of Enterobacteriaceae producing extended-spectrum b-lactamases: prevalence of CTX-M-3 at a hospital in China. Antimicrobial Agents and Chemotherapy 47, 790 3. 25. Quinteros, M., Radice, M., Gardella, N. et al. (2003). Extendedspectrum b-lactamases in Enterobacteriaceae in Buenos Aires, Argentina, public hospitals. Antimicrobial Agents and Chemotherapy 47, 2864 7. 26. Katz, O., Peled, N. & Yagupsky, P. (2003). Evaluation of the current NCCLS recommendations for screening and confirming ESBL production. In Program and Abstracts of the One-hundred and third American Society for Microbiology General Meeting, Washington, DC, 2003. American Society for Microbiology, Washington, DC, USA. 27. Wu, T. L., Chia, J. H., Su, L. H. et al. (2003). Dissemination of extended-spectrum b-lactamase-producing Enterobacteriaceae in pediatric intensive care units. Journal of Clinical Microbiology 41, 4836 8. 28. Lee, S. H., Kim, J. Y., Shin, S. H. et al. (2003). Dissemination of SHV-12 and characterization of new AmpC-type b-lactamase genes among clinical isolates of Enterobacter species in Korea. Journal of Clinical Microbiology 41, 2477 82. 29. Chanawong, A., M Zali, F. H., Heritage, J. et al. (2001). SHV-12, SHV-5, SHV-2a and VEB-1 extended-spectrum b-lactamases in Gramnegative bacteria isolated in a university hospital in Thailand. Journal of Antimicrobial Chemotherapy 48, 839 52. 30. Bonnet, R. (2004). Growing group of extended-spectrum b-lactamases: the CTX-M enzymes. Antimicrobial Agents and Chemotherapy 48, 1 14. 31. Chanawong, A., M Zali, F. H., Heritage, J. et al. (2002). Three cefotaximases, CTX-M-9, CTX-M-13, and CTX-M-14, among Enterobacteriaceae in the People s Republic of China. Antimicrobial Agents and Chemotherapy 46, 630 7. 32. Poirel, L., Decousser, J. W. & Nordmann, P. (2003). Insertion sequence ISEcp1B is involved in expression and mobilization of a bla CTX-M b-lactamase gene. Antimicrobial Agents and Chemotherapy 47, 2938 45. 33. Humeniuk, C., Arlet, G., Gautier, V. et al. (2002). b- Lactamases of Kluyvera ascorbata, probable progenitors of some plasmid-encoded CTX-M types. Antimicrobial Agents and Chemotherapy 46, 3045 9. 34. Poirel, L., Kampfer, P. & Nordmann, P. (2002). Chromosomeencoded Ambler class A b-lactamase of Kluyvera georgiana, a probable progenitor of a subgroup of CTX-M extendedspectrum b-lactamases. Antimicrobial Agents and Chemotherapy 46, 4038 40. 35. Lartigue, M. F., Poirel, L. & Nordmann, P. (2004). Diversity of genetic environment of bla CTX-M genes. FEMS Microbiology Letters 234, 201 7. 332