| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, September 2007, p. 3440-3444, Vol. 51, No. 9
0066-4804/07/$08.00+0 doi:10.1128/AAC.01431-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

CNRS UMR 5234, Université de Bordeaux 2, 146 rue Léo Saignat, 33076 Bordeaux Cedex,1 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 33000 Bordeaux,2 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 33390 Blaye,3 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 64000 Bayonne,4 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 33210 Langon,5 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 64000 Pau,6 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 33140 Villenave d'Ornon,7 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 16000 Angoulême,8 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 40000 Mont-de-Marsan,9 Laboratoires d'Analyses Médicales du Réseau Aquitaine, 24100 Bergerac, France,10
Received 16 November 2006/ Returned for modification 26 January 2007/ Accepted 1 June 2007
| ABSTRACT |
|---|
|
|
|---|
| TEXT |
|---|
|
|
|---|
In France, antibiotic resistance is regularly monitored in the hospital environment (20, 28) but is less well documented in private health care centers, which can be surveyed only by private laboratories in charge of their microbiological tests (23). In 1998, we founded a network of private laboratories to monitor antibiotic resistance in the extrahospital practices of the Aquitaine region (in southwestern France). In 1999, a survey of the Aquitaine network revealed that ESBL-producing enterobacteria were present in private health care centers at rates similar to those found in hospitals (23) and were occasionally responsible for authentic outbreaks in these facilities (2). The aim of this study was to perform a similar survey focused on private health care centers in order to analyze the incidence of ESBL-producing enterobacteria according to the type of institution and to examine the evolution over the five preceding years.
From January to June 2004, 1,570 clinically relevant, consecutive, and nonredundant strains of enterobacteria were collected from institutionalized patients by the Aquitaine network. ESBLs were detected by the double-disk synergy test (13) with or without the presence of cloxacillin, which inhibits cephalosporinases (26). ESBLs were found in 49 strains, mainly isolated from the urinary tract (41 samples) of 45 institutionalized patients (total, 1,398 patients) residing in 21 of the 69 examined centers (7/19 clinics; 4/8 "follow-up" centers, including 2/6 convalescence and 2/2 rehabilitation centers, 9/40 nursing homes, and 1/2 local hospitals). The global proportion of private institutions accommodating patients infected with ESBL-positive enterobacteria (omitting the local hospitals absent from the 1999 survey) was similar (29.4% versus 29.8% in 1999; P < 0.005). The overall incidence of ESBL producers in private health care centers slightly decreased compared to 1999 (2.8% versus 3.3%), but not significantly (P = 0.51). The variations observed between the two surveys can be ascribed, at least in part, to the evolution of the network (16 laboratories in 2004 versus 8 in 1999). The percentages of ESBL producers in private facilities increased from the clinics (1.5%) and convalescence centers (1.7%) to rehabilitation centers (5.3%) and nursing homes (10.9%). This rank order correlates with a longer stay (e.g., mean duration, 7 days in clinics versus 58 days in rehabilitation centers). Nursing homes are known to be an important reservoir of ESBL-producing enterobacteria (19). The high frequency of ESBL producers in local hospitals (11.1%) shows that these public institutions, too small to have their own laboratories and excluded from public statistics, can be heavily contaminated. In addition, plasmid-mediated cephalosporinases were found in two strains recovered from the urinary samples of two patients residing in CLI-2/RC-1, one of which also expressed an ESBL (Table 1). Thus, a total of 50 strains were analyzed, including 49 ESBL-expressing isolates and one plasmidic AmpC-expressing isolate.
|
ESBL producers were multidrug resistant, as shown by the agar diffusion method using 27 disks (27) (Table 1). Indeed, most of the 50 strains were resistant to aminoglycosides, including 24 that exhibited a TNt(A) phenotype (tobramycin, netilmicin, and low-level amikacin resistance) associated with an aac(6')-I gene, 1 strain with a GTNt phenotype (gentamicin, tobramycin, and netilmicin resistance) related to an aac(3)-II gene, and 12 strains with the combined GTNtA phenotype and both enzyme-encoding genes, as verified by PCR amplifications (2, 3, 8, 9). The strains also exhibited resistances to sulfonamides (78%), trimethoprim (74%), chloramphenicol (64%), and tetracycline (37%) (Table 1). Although most ESBL-positive strains were resistant to nalidixic acid (94%) and ofloxacin (86%), no qnr genes could be found using a previously published method (12, 24). However, eight E. coli produced the aac(6')-Ib-cr variant, as demonstrated after gene sequencing. Only carbapenems were active against all 50 strains (Table 2).
|
Besides antibiotyping, the epidemiological relationship between strains belonging to the same species has been investigated by pulsed-field gel electrophoresis using a CHEF-DRIII system (Bio-Rad), and the restriction endonucleases XbaI (E. coli, E. aerogenes, and K. pneumoniae), SpeI (E. cloacae), and SfiI (P. mirabilis). The obtained patterns were interpreted according to the criteria of Tenover et al. (29) (Table 1). Resistances cotransferred with ESBLs and AmpC enzymes were studied by a filter-mating assay using a nalidixic acid- and rifampin-resistant mutant of E. coli K-12 or an azide-resistant E. coli C600 as a recipient (2, 3). Restriction plasmid profiles were determined after plasmid DNA extraction (25) by EcoRI or HpaI digestion (Promega, Charbonnière-les-Bains, France).
The combined analysis of ß-lactamase content and typing methods demonstrated that among the 18 E. aerogenes strains, 16 belonged to the molecular type Ea1 and produced the TEM-24b enzyme. The Ea1 clone was highly similar to that previously epidemic in our region and prevalent in France and bordering countries at the end of the 1990s (1-3). This clone was responsible for an outbreak in CLI-2 in 1999 and still persisted in the clinic in 2004, albeit at a lower rate (4.7% versus 8.7%). However, it had unexpectedly evolved toward increased susceptibility, consistent with a genetic shift associated with its progressive decline. Indeed, not only were none of the Ea1 isolates resistant to imipenem (versus 12% in 1999), but the chromosomal AmpC ß-lactamase was not inducible in six of them (28% versus 0% in 1999), leading to atypical amoxicillin-clavulanate and cefoxitin susceptibilities (7) (Table 2). Moreover, four E. aerogenes strains showing only a decreased tobramycin susceptibility [(T) phenotype] possessed an aac(6')-I gene governed by a weak promoter, as shown by sequencing (data not shown). As reported in 1999, the conjugative plasmid encoding the TEM-24b enzyme was found in eight strains belonging to four other enterobacterial species, and it occasionally underwent limited variations, sometimes associated with the loss or gain of cotransferred resistances (2). Similarly, a highly transferable TEM-21-encoding plasmid, widespread in the Bordeaux area (2, 8, 30) and responsible for an outbreak in NH-1 in 1999, was still present in that institution in 2004 (identical EcoRI patterns) (data not shown). The TEM-3-producing clone of E. aerogenes Ea2 and the TEM-15-producing clone of K. pneumoniae Kp2 also persisted in the Southern "Pays Basque" part of our region, particularly in CLI-5.
However, one of the major differences from the 1999 survey was the increase in CTX-M-expressing E. coli strains (24.5% versus 2.6%). Among the 17 ESBL-producing E. coli strains, 12 expressed a CTX-M enzyme, including 9 that elaborated CTX-M-15. Of these nine strains, scattered among five institutions, five exhibited the molecular type Ec1, as did two control strains (TN03 and DOS) collected in northern French hospitals (11, 14). Moreover, the five Ec1 E. coli strains showed the same HpaI-restricted plasmid profile (pEc1-A), antibiotic resistance patterns, and ß-lactam MICs (Table 1), providing evidence of the dissemination of a clonal strain in the local hospital and a nearby nursing home. The pEc1-A plasmid was identical to pTN03, a plasmid extracted from the control strain, TN03, which carries an ISEcp1-like element upstream from blaCTX-M-15 (10). pTN03 possesses the same blaCTX-M-15 environment as Pc15-1a, an epidemic plasmid originally described in Canada (5). Among the four other CTX-M-15-producing E. coli strains, three were Ec2 or Ec2-a types (possibly related, according to the criteria of Tenover et al.), like the MTPB5 control strain from a southwestern French hospital (15). The Ec2/Ec2-a strains presented different HpaI-restricted plasmid patterns, cotransferred resistances, and insertion sites of IS26 within the ISEcp1-like elements (data not shown). In particular, this insertion in the pEC1-D plasmid generated a weak promoter, translating into low cefoxatime MICs (Table 2). CTX-M-15 is currently the predominant CTX-M enzyme worldwide, and its propagation has previously been attributed to its presence in an epidemic clone of E. coli belonging to the phylogenetic group B2, which expresses multiple virulence factors (16, 17, 22). The remaining strains of E. coli, including three other CTX-M producers, were unrelated, except for two (profile Ec7) that were isolated from the same patient but were considered nonduplicates because of their different ß-lactam susceptibilities. These differences were due to the presence or absence of IS26, providing a strong promoter upstream from the blaSHV-12 gene (Tables 1 and 2). Analysis of all collected enterobacteria, including the 184 redundant ones, underscored prolonged ESBL carriage (6 days to 3 months), in vivo interspecific transfer in the same patient, ESBL acquisition by new patients in known contaminated facilities, and interinstitution transfer of contaminated patients.
The other main difference from the 1999 survey was the detection of the two plasmid-mediated cephalosporinase-producing K. pneumoniae strains. One strain (Kp4) expressed the CMY-4 enzyme associated with the SHV-12 ESBL, while the other (Kp5) produced the ACC-1 enzyme and exhibited the same molecular pattern as SLK 54, a control strain involved in a nosocomial outbreak in the north of France several years before (18). Our study is the first description of the emergence of plasmid-mediated cephalosporinases in the French extrahospital setting.
In conclusion, almost one-third of French private health care centers accommodated patients infected with ESBL-producing strains. The frequencies of these organisms varied according to the type of private institution and was highest in nursing homes and rehabilitation centers. Compared to a survey 5 years before, ESBL producers were isolated at a globally similar rate but showed a drastic change in species and enzyme distributions. Indeed, the previously epidemic TEM- or SHV-producing E. aerogenes or K. pneumoniae strains tended to be supplanted by CTX-M-expressing E. coli strains. Unusual phenotypes and variable levels of expression of the same ESBL in relation to promoter modifications were observed. Owing to the complex epidemiology of ESBL producers, regular surveys in private health care institutions are warranted.
Nucleotide sequence accession number. The partial nucleotide sequence of the rpoB gene from E. aerogenes strain Ea2822 has been assigned the GenBank/EMBL database accession number EF108305.
| ACKNOWLEDGMENTS |
|---|
This work was supported in part by a grant from the Ministère de l'Education Nationale et de la Recherche.
| FOOTNOTES |
|---|
Published ahead of print on 25 June 2007. ![]()
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |