TEM-24 extended-spectrum b-lactamase-producing Enterobacter aerogenes: long-term clonal dissemination in French hospitals

Transcription

1 ORIGINAL ARTICLE TEM-24 extended-spectrum b-lactamase-producing Enterobacter aerogenes: long-term clonal dissemination in French hospitals J.-O. Galdbart, F. Lëmann, D. Ainouz, P. Fëron, N. Lambert-Zechovsky and C. Branger Service de Microbiologie, Hoª pital Beaujon, Clichy, France Objective To investigate interstrain relatedness of TEM-24 -producing Enterobacter aerogenes clinical strains isolated between 1993 and 1998 in 10 French hospitals from nine areas by pulsed- eld gel electrophoresis (PFGE) and plasmid patterns. Methods FifteenTEM-24-producing strains and a set of 16 control strains having various other antibiotic resistance phenotypes were genotyped by PFGE. Plasmid DNA fromtem-24-producing strains and transconjugants was analyzed. Results Analysis of XbaI macrorestriction patterns revealed only minor variations, and showed that all15 TEM-24-producing strains were closely related. Some isolates originating from distant areas had indistinguishable patterns. According to their clustering correlation coe cients, they were also genomically distant from the control strains.two plasmid patterns were observed intem-24 -producing strains, one of them in13 of the strains. Large plasmids of 85 kb encodingtem-24 b-lactamase were present in all isolates and, in all except one strain, could be transferred with high frequency by conjugation. Conclusions Theseresultscon rmthatthespreadofthetem-24extended-spectrumb-lactamase in France was essentially due to the dissemination of a single clone. Keywords Enterobacter aerogenes,tem-24, ESBL, dissemination Accepted 15March2000 Clin Microbiol Infect 2000: 6: 316^323 INTRODUCTION Corresponding author and reprint requests: C. Branger, Laboratoire de Microbiologie, Hoª pital Beaujon, 100 Boulevard du Gëne ral Leclerc, 92110, Clichy, France Tel: Fax: catherine.branger@bjn.ap-hop-paris.fr Enterobacter aerogenes has recently emerged as an important hospital pathogen [1,2]. In this species, as in other b-lactamaseinducible Gram-negative bacilli, resistance to broad-spectrum cephalosporins is due mostly to stable de-repressed mutants overproducing their chromosomal enzyme; this may be selected during therapy, and can spread within the hospital environment [3,4]. Other mechanisms of b-lactam resistance havebeendescribed in Enterobacter aerogenes; they include plasmid-mediated b-lactamases, and the diminution of membrane permeability due to lack of a porin combined with high-level constitutive b-lactamase production leading to imipenem resistance [5^7]. Detection of extended-spectrum b-lactamase (ESBL) production in Enterobacter aerogenes can be di cult when chromosomal cephalosporinase is overproduced, so rendering the organisms resistant to all b-lactam antibiotics except the carbapenems and cefepime or cefpirome [8]. Outbreaks of multiresistant Enterobacter aerogenes have been reported in several countries, especially among patients in intensive care units [3,9,10]. These organisms have acquired large plasmids encoding a variety of ESBLs, including TEM- 3, TEM-10, TEM-12, TEM-24 and TEM-26 [10^12]. In France, Enterobacter aerogenes strains producing TEM-24 ESBL have been found to be responsible for several outbreaks of nosocomial infections [13,14]. Our own hospital has experienced two outbreaks caused by such strains [15]. The epidemiology of ESBL-producing Enterobacteriaceae is complex; it involves the spread of self-transferable plasmids and the transposition of resistance genes among di erent plasmids, as well as clonal strain spread [16]. In France, for example,tem-3 b-lactamase became widespread in the late 1980s, mainly because of the spread of a plasmid carrying its gene [17, 18]. Inter-hospital transmission of ESBL-producing Enterobacteriaceae has also been seen in the USA [19,20]. The shortcomings of phenotypically based typing methods have led to the development of methods based on the microbial genotype, so minimizing problems with typing and reproducibility [21]. = 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases

2 Galdbart et al Extended-spectrum b-lactamase-producing Enterobacter aerogenes â"æ Among them, PCR-mediated procedures such as RAPD or Rep-PCR, which can be performed rapidly and easily, have been used successfully for identifying the short-term spread of bacterial strains [9,22]. However, standardization and interpretation are not always easy.with these techniques,tem-24 - producing Enterobacter aerogenes strains isolated in France during a 6-month period showed similar characteristics [23]. Pulsed- eld gel electrophoresis (PFGE) of genomic macrorestriction fragments, with its high reproducibility and discriminatory power, has been widely applied for subtyping various pathogens. It allows interstrain relatedness to be assessed, a characteristic which is useful for tracing the evolution of epidemic strains over extended periods [16,24]. This study was designed to use macrorestriction ngerprinting of genomic DNA and plasmid typing to investigate the clonal relatedness of TEM-24 -producing Enterobacter aerogenes strains isolated in hospitals from various areas of France and to assess the genomic evolution of strains obtained over a 5-year period. They were also compared to a set of control strains of other antibiotic resistance phenotypes. MATERIALS AND METHODS Bacterial strains We analyzed 15 clinical isolates of Enterobacter aerogenes producing TEM-24 b-lactamase, all of which were isolated in France between 1993 and 1998 (Table1). They were selected to be representative of a broad geographic area. Four were isolated in the Paris area (Clichy and Pontoise) and 11 were isolated from eight other areas (Bordeaux, Colmar, Dieppe, Dijon, Marseille, Saint Calais, Saint Etienne and Strasbourg). One isolate was studied from each of seven hospitals; from the three other hospitals, two to three isolates were studied. A control group of 16 Enterobacter aerogenes isolates was also included for study by PFGE. These were collected during the same period from four areas (Clichy, Saint Calais, Dieppe and Brest). Five produced a putative SHV- 4 ESBL (SHV phenotype), six had high-level production of chromosomal cephalosporinase (Case phenotype), and ve were of wild phenotype. All the isolates were identi ed with the API 20E system (biome rieux, Marcy l'e toile, France). Antibiotic susceptibility testing Susceptibility to antimicrobial agents was tested by the disk di usion method on Mueller^Hinton agar (Sano Diagnostics Pasteur, Marnes la Coquette, France) with commercially available disks according to the guidelines of the Antibiogram Committee of the French Society for Microbiology [25]. These agents were amoxycillin, amoxycillin^clavulanate, ticarcillin, ticarcillin^clavulanate, piperacillin, piperacillin^ tazobactam, cefotaxime, ceftazidime, cefpirome, imipenem, aztreonam, gentamicin, tobramycin, kanamycin, streptomycin, tetracycline, netilmicin, amikacin, spectinomycin, chloramphenicol, trimethoprim^sulfamethoxazole, and pe oxacin. Disks of apramycin, 2 0 -N-ethyl-netilmicin and N-ethyl-netilmicin (Schering-Plough Research Institute, Kenilworth, NJ, USA) were also tested. ESBLs were detected Table 1 Relevant characteristics of the 15 Enterobacter aerogenes strains producing TEM-24 b-lactamase Strain Date of isolation (month/year) City Resistance co-transferred with TEM-24 b-lactamase pi of ESBL Plasmid pattern XbaI pattern 4 9/93 Bordeaux Sm Sp Km Tm Nt Am Tc Tp 6.5 P2 1d 8 10/93 Bordeaux Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1g 15 10/93 Dijon No transconjugant 6.5 P2 1i 13 1/96 Clichy Sm Sp Km Tm Nt Am Cm Tp 6.5 P2 1k 10 2/96 Clichy Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1i 12 3/96 Pontoise Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1i 6 5/96 Clichy Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1f 3 8/96 Marseille Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1c 5 1/97 St Calais Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1e 7 5/97 St Etienne Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P1 1f 9 7/97 St Etienne Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1 h 1 9/97 St Etienne Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P1 1a 14 12/97 Dieppe Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1j 2 2/98 Colmar Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1b 11 3/98 Strasbourg Sm Sp Km Tm Nt Am Tc Cm Tp 6.5 P2 1i Sm, streptomycin; Sp, spectinomycin; Km, kanamycin; Tm, tobramycin; Nt, netilmicin; Am, amikacin; Tc, tetracycline; Cm, chloramphenicol; Tp, trimethoprim.

3 â"ð Clinical Microbiology and Infection, Volume 6 Number 6, June 2000 by the double-disk synergy test with a 30-mm spacing between the disks as previously described [17], and also with a 20-mm spacing. The synergy test was also carried out on Mueller^Hinton agar containing 500 mg/l of cloxacillin [26]. A disk of ticarcillin plus clavulanic acid was used with disks of ceftazidime, cefotaxime, aztreonam, ceftriaxone, cefpirome and cefepime. Minimal inhibitory concentrations (MICs) of b-lactams were determined by the Etest technique (AB Biodisk, Solna, Sweden) as recommended by the manufacturer. MICs of cefotaxime and ceftazidime were also determined in the presence of 500 mg/l of cloxacillin [26]. Iso-electric focusing of b-lactamases Analytical iso-electric focusing was performed with crude cell sonic extracts onph3.5^9.5 polyacrylamide gel (Pharmacia Biotech, Orsay, France) [8,27]. b-lactamase activity was detected by an iodometric method with benzylpenicillin (100 mg/l) or ceftriaxone (200 mg/l gel) as substrates [27]. b-lactamases with known pis (TEM-1, 5.4;TEM-3, 6.3; SHV-1, 7.6; SHV- 4, 7.8; SHV- 5, 8.2) were run in parallel with the extracts. PFGE of macrorestriction fragments and comparative analysis of banding patterns The preparation of DNA and the digestion with XbaIwereas described previously [29]. Fragments in 0.5 TBE bu er, ph 8.0, at 12 C were separated by a contour-clamped homogeneous electric eld (CHEF-DRII system, BioRad) at 200 V for 18 hwithpulsetimesof6^15s for 5 h and 23^40 s for 13 h. Concatameric bacteriophage lambda DNA molecules (48.5 kb; BioRad) were used as a size standard. To estimate the degree of relatedness between isolates, a similarity matrix was created using the band-based Dice similarity coe cient [30]. Clustering correlation coe cients were calculated by the unweighted pair-group method using average linkages (UPGMA) determined with GelCompar software (Applied Maths, Kortrijk, Belgium). Transfer of resistance to oxyimino-b-lactams Escherichia coli K-12 J53-2 resistant to rifampicin was used as a recipient for conjugation of the plasmids encoding the TEM- 24 ESBL. Transconjugants were selected on plates containing Mueller^Hinton agar supplemented with ceftazidime (4 mg/ L) and rifampicin (250 mg/l). The frequency of transfer was expressed relative to the number of donor cells [31]. TEM-24 PCR analysis and sequencing Total cellular DNA was extracted with Instagen (BioRad, Ivry sur Seine, France), as recommended by the manufacturer. Primer 1 (5 0 -ATGAGTATTCAACATTTGCG-3 0 ) and primer 2 (5 0 -CTGACAGTTACCAATGCTTA-3 0 ) described by Rasheed et al [28] were used for the ampli cation of the entire bla TEM gene. Primer A (5 0 -GGGCAAGAGCAACTCGGT-3 0 ) and primer B (5 0 -AGACCCACGCTTACCGGT-3 0 ) described by Neuwirth et al [13] were used for speci c ampli- cation of the bla TEM-24 gene. Ampli cation reactions were performed in a total volume of 50 ml containing 10 ml of DNA template, 200 mm (each) deoxynucloside triphosphate, 0.5 mm primer and 1U oftaq polymerase in 1 PCR bu er (Roche Molecular Biochemicals, Meylan, France). The ampli- cation program started with denaturation for 5 minat94 C. The 35 subsequent cycles of ampli cation consisted of denaturation for 1minat94 C, annealing for 1minat57 Cand extension for 1min at 72 C. Ten microliters of PCR products was electrophoresed in a 1% agarose gel and stained with ethidium bromide (0.5 mg/l). PCR products, generated with primers 1 and 2 from DNA extracts of strains 1, 6 and 13, were used for direct sequencing. This was performed with an automatic sequencer (377, ABI Prism, Perkin-Elmer, Roissy, France) by Genome Express, Paris, France. Plasmid analysis Plasmid DNA from bacterial strains and transconjugant Escherichia coli was extracted and puri ed with ion exchange cartridges (Nucleobond AX, Macherey-Nagel, Hoerdt, France) according to the manufacturer's instruction, and DNAwas analyzed after separation on 0.7% agarose gel. Plasmid paff2 with a molecular size of 85^90 kb from Escherichia coli BM 694 was used as standard [32]. Puri ed plasmid DNA from transconjugant Escherichia coli for isolates 2 and 12 was digested with restriction enzyme PstI by using the bu er and reaction conditions recommended by the manufacturer (Roche Molecular Biochemicals) and electrophoresed on a 0.8% agarose gel. RESULTS Antimicrobial susceptibility By disk agar di usion, all the TEM-24 -producing ESBL Enterobacter aerogenes isolates were resistant to all b-lactams tested except imipenem. The MICs of ceftazidime were higher than those of cefotaxime (Table 2). In the presence of 500 mg/ L of cloxacillin, which inhibits cephalosporinase activity, the MICs of these two cephalosporins were decreased (Table 2). Although the presence of clavulanic acid lowered the MICs of ticarcillin and piperacillin, susceptibility of the isolates to

4 Galdbart et al Extended-spectrum b-lactamase-producing Enterobacter aerogenes â"ñ Table 2 Resistance pattern to b-lactam antibiotics of the Enterobacter aerogenes strains studied MIC range (mg/l) b-lactams TEM-24 a Case b Wild c Cefotaxime 4±32 2± ±0.25 Cefotaxime cloxacillin (500 mg/l) 2± < Ceftazidime 64 to >256 8± ±0.5 Ceftazidime cloxacillin (500 mg/l) 32± ±2 0.03±0.5 Aztreonam 32± ± ±0.125 Cefepime 2±8 0.19± ±0.047 Cefpirome 2±8 0.25± ±0.125 Ticarcillin > to >128 2±8 Ticarcillin clavulanic acid 64± to >128 2±8 Piperacillin 128± to >128 4±8 Piperacillin clavulanic acid 32± to >128 2±8 a TEM-24, TEM-24-producing isolates. b Case, cephalosporinase-producing isolates. c Wild phenotype isolates. these antibiotics was not fully restored. These results showed that in association with the production of the TEM-24 ESBL, the chromosomal Bush group 1 b-lactamase was overproduced. When disks were placed 30 mm apart, a synergistic e ect was observed between ticarcillin^clavulanic acid and cefepime or cefpirome, but not between ticarcillin^clavulanic acid and ceftazidime, cefotaxime or aztreonam. For these last three antibiotics, a synergistic e ect with ticarcillin^clavulanic acid was obtained when tested on Mueller^Hinton agar containing 500 mg/l cloxacillin, or when disks were placed 20mm apart. In addition to the b-lactam resistance, all the TEM-24 -producing isolates were resistant to pe oxacin, tetracycline, chloramphenicol, trimethoprim and all the aminoglycosides tested except gentamicin, apramycin and 6 0 -Nethyl-netilmicin. This aminoglycoside spectrum of activity suggests that all the isolates produced two types of modifying enzyme: an AAC(6 0 )-1 (resistance to tobramycin, netilmicin, 2 0 -N-ethyl-netilmicin, amikacin) and an ANT (3 00 )-1 (resistance to streptomycin and spectinomycin). Iso-electric focusing All 15 isolates of Enterobacter aerogenes producing a putative TEM-24 ESBL and the ve control isolates producing a putative SHV- 4 ESBL were analyzed for their b-lactamase content. With ceftriaxone as substrate, all the isolates had two bands of b-lactamase activity. The rst band was of pi 8.3, corresponding to de-repressed chromosomal cephalosporinase. The second band was for 15 isolates of pi 6.5 (putative TEM- 24) and for ve isolates of pi 7.8 (putative SHV-4). No additional band of activity was detected with benzylpenicillin as substrate. Identi cation of encoding gene by TEM-24-speci c PCR and DNA sequencing By using TEM-24 -speci c primers, a DNA fragment of 476 bp was ampli ed from total DNA samples of all the isolates producing an enzyme of pi 6.5. No ampli cation fragment was present in the total DNA sample of the ve control isolates producing an enzyme of pi 7.8, or in the six control isolates with a Case phenotype. Nucleotide sequencing of the bla TEM PCR products of strains 1, 6 and 13 producing an enzyme of pi 6.5 showed that they represent ampli ed genes coding for the TEM-24enzyme as describedby Chanalet al [33], except that one silent mutation (T C) was found. PFGE analysis After digestion with XbaI, Enterobacter aerogenes DNA gave 13^ 19 fragments of 30^850 kb (Figure 1). The patterns obtained for thetem-24 -producing isolates di ered by only 1^5 bands and had similarity coe cients >88%. They clustered into the same major pattern, number 1 (Figure 2). Another 15 separate major patterns were found for the 16 control isolates. All the major patterns were distinct, showing no more than 38^70% similarities (Figure 2). Indistinguishable patterns were observed for several TEM-24 -producing isolates originating from distant areas: pattern 1f for isolates 6 and 7, and pattern 1i for isolates 10, 11, 12 and 15. The two isolates responsible for two episodes of outbreak in our hospital di ered by three bands (patterns 1i and 1f). All the control isolates, except two of SHV phenotype, had di erent patterns. These two isolates, although isolated from distant areas, were of the same major pattern, 4, and di ered by only one band (patterns 4a and 4b) (Figure 2).

5 âáò Clinical Microbiology and Infection, Volume 6 Number 6, June 2000 Figure 1 PFGE of XbaI-digested DNA from 20 Enterobacter aerogenes isolates. Lanes 1, 5, 13 and 20: pattern 1f. Lane 2: pattern 1i. Lane 3: pattern 1j. Lane 4: pattern 1b. Lane 6: pattern 1a. Lane 7: pattern 1e. Lane 8: pattern 1 h. Lane 9: pattern 1k. Lane 10: pattern 1c. Lane 11: pattern 1d. Lane 12: pattern 1g. Lane 14: pattern 2. Lane 15: pattern 7. Lane 16: pattern 4b. Lane 17: pattern 8. Lane 18: pattern 16. Lane 19: pattern 5. Lane 21: bacteriophage lambda DNA concatemers (BioRad). Transfer of oxyimino-b-lactam resistance Conjugation experiments were carried out on the Enterobacter aerogenes isolates producingtem-24 ESBL. Oxyimino-b-lactam antibiotic resistance was transferred to Escherichia coli K-12 with high frequency ( ) in all isolates except one (isolate 15). Some resistance markers were also co-transferred (Table 1): aminoglycosides and trimethoprim in all isolates; tetracycline and chloramphenicol in 13 isolates (Table 1). Plasmid analysis Two di erent plasmid patterns were observed among the TEM-24 -producing isolates (Figure 3,Table 1). Pattern P2 was common to 13 isolates and corresponded to the presence of two plasmids. In the two other isolates, in addition to the large plasmid, two plasmid bands, with low molecular weight, were detected. In Escherichia coli transconjugants, acquisition of resistance to oxyimino-b-lactam antibiotics was correlated with the transfer of a large plasmid whose molecular mass was estimated tobe 85^90 kbafter restrictionwith PstI (Figure 4). Figure 2 Classi cation and schematic representation of the XbaI patterns of the Enterobacter aerogenes isolates. The b- lactam phenotypes corresponding to the XbaI patterns are indicated: production of ESBLs (SHV or TEM-24); over-production of the chromosomal cephalosporinase alone (Case); or wild phenotype (Wild). DISCUSSION Nosocomial infections with Enterobacteriaceae expressing ESBLs have become widespread in clinical medicine [34, >35]. It is important to detect Enterobacter strains producing ESBLs in a clinical laboratory, and to di erentiate them from derepressed mutants, because they can be responsible for the spread of resistance genes in hospitals. However, di culties have been noted in trying to detect ESBL production in Enterobacter over-expressing chromosomally mediated cephalosporinase [36]. In our study, the easiest detection conditions by double disk synergy were with disks of cefepime and amoxycillin^clavulanic acid placed 30 mm apart.

6 Galdbart et al Extended-spectrum b-lactamase-producing Enterobacter aerogenes âá" Figure 3 Plasmid pattern of TEM-24-producing Enterobacter aerogenes and transconjugant. Lane 1: molecular mass marker X (Roche Molecular Biochemicals). Lane 2: paff2 (SHV-5). Lane 3: representative transconjugant. Lanes 4, 5 and 7: pattern P2. Lane 6: pattern P1. Figure 4 PstI ngerprint of plasmids encoding TEM-24 b-lactamase isolated from the Escherichia coli transconjugants. Lane 1: plasmid from Enterobacter aerogenes 2. Lane 2: plasmid from Enterobacter aerogenes 12. Lane 3: molecular weight marker X (Roche Molecular Biochemicals). Lane 4: phage lambda digested by HindIII (Roche Molecular Biochemicals). Strains of Enterobacter aerogenes producing TEM-24 ESBL have been reported in several areas of France since 1993, and have been responsible for several outbreaks [13^15]. Using AP-PCR as a typing method, Bosi et al suggested that such strains were closely related [23]. Genetic typing showed that strains which undergo short-term spread during an outbreak accumulate a limited number of genetic variations [24]. Therefore, it would be expected that in a nationwide spread of bacterial strains over a longer period, an increasing amount of polymorphism would be detected. To evaluate the polymorphism of TEM-24-producing strains of Enterobacter aerogenes collected over 5 years from hospitals in distant areas in France, we chose PFGE as a molecular tool because it has good discriminatory power and allows a quantitative assessment of genomic relatedness to be made. In the present study, PFGE gave consistent results. Enterobacter aerogenes control isolates, with diverse antibiotic resistance pro les (wild phenotype, case phenotype, SHV phenotype), exhibited a high degree of polymorphism, so con rming heterogeneity within the Enterobacter aerogenes species. All the 15 TEM-24-producing isolates had very similar XbaI patterns (pattern 1), di ering by no more than ve fragments. Using computer analysis, all these patterns clustered at greater than 88% similarity.they were likely to represent genotypic variants of the same clone which, over time, underwent minor rearrangements. Four isolates originating from four distant areas collected 5 years apart had indistinguishable XbaIpatterns. Nopatternotherthanpattern1was found among the TEM-24-producing isolates, and this pattern was not seen among the control isolates. Even more surprising was the fact that the TEM-24 -encoding plasmid was found to be transferable at high frequency during conjugation. These observations suggest that strains of pattern 1are strongly associated with thetem-24 -encoding plasmid. The epidemiology of ESBL is complex, and DNA ngerprinting methods for analyzing both the bacterial genome and the R plasmid are required.the DNA composition of antibiotic resistance plasmids may change through various transfers, or because of the acquisition or loss of resistance-encoding transposons. As a result, epidemiologically related isolates may exhibit di ering plasmid pro les which have evolved under antibiotic pressure.variations of plasmid pro les during an outbreak or during longitudinal analysis have been documented [37]. In our study it was remarkable that all but two of the TEM-24-producing isolates had the same plasmid pro le. A plasmid of about 85 kb encoding TEM-24 b-lactamase was easily transferred to Escherichia coli in all except one of the 15 isolates. The plasmid co-transferred the same resistance markers, except for two isolates in which tetracycline or chloramphenicol resistance was not transferred. Antibiotic resistance patterns are in uenced by mutation, which may be selected under antibiotic pressure. Further experiments using restric-

7 âáá Clinical Microbiology and Infection, Volume 6 Number 6, June 2000 tion endonuclease analysis would be useful to compare the stability of the plasmid among the isolates. Di erent epidemiologic patterns in the dissemination of ESBL genes have been observed. Prodinger et al found that several SHV- 5-producing Klebsiella pneumoniae outbreak strains were linked by the same R plasmid [38]. Similarly, molecular analysis has demonstrated that the spread of the TEM-3 bla gene among Enterobacteriaceae in France involved a single epidemic plasmid that underwent minor modi cation [18]. By contrast, the dissemination of the SHV-4 bla gene in French hospitals occurred via a single K. pneumoniae strain [37]. Likewise,TEM-24 b-lactamase among isolates of Enterobacter aerogenes in France has been mainly the result of the dissemination of a single strain. PFGE measured only limited or no genetic variations of this strain over-time, suggesting that its genome is stable. It has been demonstrated that acquisition of virulence factors, antibiotic resistance or resistance to environmental stress favors the development and spread of pathogenic strains [1,39]. This observation makes it hard to di erentiate strains during epidemiologic studies. In our study, it was di cult to decide if the second outbreak in our hospital was due to the reintroduction of a TEM-24 -producing strain, or to the rst epidemic strain having persisted in an undetected reservoir. Epidemiologic surveillance or the investigation of bacterial outbreaks should take into account the extent of geographic spread. Because the hospitals from which the isolates originated are located so far apart in France, only the transfer of patients between hospitals can account for the spread of TEM-24 -producing Enterobacter aerogenes strains at a national level. This emphasizes the importance of following guidelines for infection control measures, such as early detection of colonized patients and the use of standard contact isolation precautions. ACKNOWLEDGMENT We thank C. Arpin, G. Berthelot, A. L. Lesimple, P. Martres, M. Morillon, B. Picard,Y. Piemont, A. Ros and E. Siebor for providing strains of Enterobacter aerogenes. The English text was corrected by G.Watts. REFERENCES 1. Sanders WE Jr, Sanders CC. Enterobacter spp.: pathogens poised to ourish at the turn of the century. Clin Microbiol Rev1997;10: 220^ Jarvis WR, Martone WJ. Predominant pathogens in hospital infections. JAntimicrob Chemother1992; 29(suppl): S19^ Arpin C, Coze C, Rogues AM, Gachie JP, Bebear C, Quentin C. Epidemiological study of an outbreak due to multidrug-resistant Enterobacteraerogenes in a medical intensive care unit. JClin Microbiol 1996;34: 2163^9. 4. Davin-Regli A, Monnet D, Saux P et al. Molecular epidemiology of Enterobacter aerogenes acquisition: one-year prospective study in two intensive care units. J Clin Microbiol1996;34:1474^ De Gueldre Y, Maes N, Rost F et al. Molecular epidemiology of an outbreak of multidrug-resistant Enterobacter aerogenes infections and in vivo emergence of imipenem resistance. J Clin Microbiol 1997; 35: 152^ De Champs C, Henquell C, Guelon D, Sirot D, Gazuy N, Sirot J. Clinical and bacteriological study of nosocomial infections due to Enterobacter aerogenes resistant to imipenem. J Clin Microbiol 1993; 31: 123^7. 7. De Champs C, Sirot D, Chanal C, Poupart MC, Dumas MP, Sirot J. Concomitant dissemination of three extended-spectrum b-lactamases among di erent Enterobacteriaceae isolated in a French hospital. JAntimicrob Chemother1991;27: 441^ Pitout JDD, Moland ES, Sanders CC,Thomson KS, Fitzsimmons SR. b-lactamases and detection of b-lactam resistance in Enterobacter spp. Antimicrob Agents Chemother1997; 41: 35^9. 9. Georghiou PR, Hamill RJ,Wright CE et al. Molecular epidemiology of infections due to Enterobacteraerogenes: identi cation of hospital outbreak-associated strains by molecular techniques. Clin Infect Dis1995; 20: 84^ Rice LB,Wileey SH, Papanicolaou GA et al. Outbreak of ceftazidime resistance caused by extended-spectrum b-lactamases at a Massachusetts chronic-care facility. Antimicrob Agents Chemother 1990; 34: 2193^ Pitout JDD, Thomson KS, Hanson ND, Ehrhardt AF, Coudron P, Sanders CC. Plasmid-mediated resistance to expanded-spectrum cephalosporins among Enterobacteraerogenes strains. Antimicrob Agents Chemother1998;42: 596^ De Champs C, Sauvant MP, Chanal C et al. Prospective surveyof colonization and infection caused by expanded-spectrum b-lactamase-producing members of the family Enterobacteriaceae in an intensive care unit. J Clin Microbiol1989;27: 2887^ Neuwirth C, Siebor E, Lopez J, Pechinot A, Kazmierczak A. Outbreak of TEM-24-producing Enterobacter aerogenes in an intensive care unit and dissemination of the extended-spectrum b-lactamase to other members of the family Enterobacteriaceae. J Clin Microbiol1996;34: 76^ Morillon M, De Pina JJ, Garnotel E et al. Epidemiologie moleculaire des infections nosocomiales a Enterobacter aerogenes producteurs de TEM-24 survenues en un an dans le meª me hoª pital [abstract 288]. In: Program and abstracts of the 15e me Re union Interdisciplinaire de Chimiothërapie Anti-Infectieuse, Paris. Clermont, Ferrand: Aoric, 1995: Galdbart JO, Le mann F, Berry P, Lambert-Zechovsky N. Typage mole culaire de souches d'enterobacter aerogenes productrices de b- lactamase a spectre e tendu TEM-24 isole es au cours de deux e pide mies [abstract 188]. In: Program and abstracts of the 17e me Re union Interdisciplinaire de Chimiothërapie Anti-Infectieuse, Paris. Clermont Ferrand: Aoric, 1997: Branger C, Lesimple AL, Bruneau B, Berry P, Lambert- Zechovsky N. Long-term investigation of the clonal dissemination of Klebsiella pneumoniae isolates producing extended-spectrum b-lactamases in a university hospital. J Med Microbiol 1998; 38: 201^ Jarlier V, Nicolas MH, Fournier G, Philippon A. Extended broad-spectrum b-lactamases conferring transferable resistance to newer b-lactam agents in Enterobacteriaceae: hospital prevalence and susceptibility patterns. Rev Infect Dis1988;10: 867^78.

8 Galdbart et al Extended-spectrum b-lactamase-producing Enterobacter aerogenes âáâ 18. Petit A, Gerbaud G, Sirot D, Courvalin P, Sirot J. Molecular epidemiology of TEM-3 (CTX-1) b-lactamase. Antimicrob Agents Chemother1990; 34: 219^ Jaroux E, Collet L, Bellon O et al. Emergence of Enterobacter aerogenes strains with an extended-spectrum betalactamase in Provence [abstract MCB6]. In: 3rd International Conference on the Prevention of Infection, Nice. 1994: Sader HS, Pfaller MA, Jones RN. Prevalence of important pathogens and the antimicrobial activity of parenteral drugs at numerous medical centers in the United States. II. Study of the intra- and inter-laboratory dissemination of extended-spectrum b-lactamase-producing Enterobacteriaceae. Diagn Microbiol Infect Dis 1994; 21: 203^ Jalaluddin S, Devaster JM, Scheen R, Gerard M, Butzler JP. Molecular epidemiological study of nosocomial Enterobacter aerogenes isolates in a Belgian hospital. J Clin Microbiol1998; 36: 1846^ Grattard F, Pozetto B, Tabard L, Petit M, Ros A, Gaudin OG. Characterization of nosocomial strains of Enterobacter aerogenes by arbitrarily primed-pcr analysis and ribotyping. Infect Control Hosp Epidemiol1995;16: 224^ Bosi C, Davin-Regli A, Bornet C, Mallea M, Pages JM, Bollet C. Most Enterobacteraerogenes strains in France belong to a prevalent clone. J Clin Microbiol1999; 37: 2165^ van Leeuwen W, van Belkum A, Kreiswirth B, Verbrugh H. Genetic diversi cation of methicillin-resistant Staphylococcus aureus as a function of prolonged geographic dissemination and as measured by binary typing and other genotyping methods. Res Microbiol1998; 149: 497^ Comite l'antibiogramme de la Socie te Franµ aise de Microbiologie. Communique. Extrait Pathol Biol1998; 46(8): I^XVI. 26. Talamin A, Cavallo JD, Buisson Y, Meyran M. Mëthode simple d'identi cation des me canismes de rësistance aux beª ta-lactamines chez Pseudomonas aeruginosa. [abstract 226]. In: Abstracts of the 14th Re union Interdisciplinaire de Chimiothërapie Anti-Infectieuse, Paris. Clermont Ferrand: Aoric, 1994: Matthew M, Harris AM, Marshall MJ, Ross GW. The use of analytical isoelectric focusing for detection and identi cation of betalactamases. J Gen Microbiol1975;88:169^ Rasheed JK, Jay C, Metchock B et al. Evolution of extended-spectrum b-lactam resistance (SHV8) in a strain of Escherichia coli during multiple episodes of bacteremia. Antimicrob Agents Chemother 1997; 41: 647^ Branger C, Bruneau B, Lesimple AL et al. Epidemiological typing of extended-spectrum b-lactamase-producing Klebsiella pneumoniae isolates responsible for ve outbreaks in a university hospital. J Hosp Infect1997; 36: 23^ Dice LR. Measures of the amount of ecological association between species. Ecology1945; 26: 297^ Courvalin P, Goldstein F, Philippon A, Sirot J. In: l'antibiogramme. Paris: MCP-VIDEOM,1985: 267^ Gutmann L, Ferrë B, Goldstein FW et al. SHV-5, a novel SHVtype b-lactamase that hydrolyses broad-spectrum cephalosporins and monobactams. Antimicrob Agents Chemother1989;33: 951^ Chanal C, Poupart MC, Sirot D, Labia R, Sirot J, Cluzel R. Nucleotide sequences of CAZ-2, CAZ- 6, and CAZ-7 b-lactamase genes. Antimicrob Agents Chemother1992; 36: 1817^ Philippon A, Labia R, Jacoby G. Extended-spectrum b-lactamases. Antimicrob Agents Chemother1989; 33: 1131^ Shah PM, Stille W. Escherichia coli and Klebsiella pneumoniae more susceptible to cefoxitin than third generation cephalosporins. J Antimicrob Chemother1983; 11: 597^ Huber TW, Thomas JS. Detection of resistance due to inducible b-lactamase in Enterobacter aerogenes and Enterobacter cloacae. J Clin Microbiol1994; 32: 2481^ Arlet G, Rouveau M, Casin I, Bouvet PJM, Lagrange PH, Philippon A. Molecular epidemiology of Klebsiella pneumoniae strains that produce SHV- 4 b-lactamase and which were isolated in 14 French hospitals. J Clin Microbiol1994; 2: 2553^ Prodinger WM, Fille M, Bauernfeind A et al. Molecular epidemiology of Klebsiella pneumoniae producing SHV-5 b-lactamase: parallel outbreaks due to multiple plasmid transfer. J Clin Microbiol 1996;34: 564^ Vernet V, Madoulet C, Chippaux C, Philippon A. Incidence of two virulence factors (aerobactin and mucoid phenotype) among 190 clinical isolates of Klebsiella pneumoniae producing extendedspectrum b-lactamase. FEMS Microbiol Lett1992; 96: 1^6.