Nationwide Survey of CTX-M-Type Extended-Spectrum β-Lactamases among Klebsiella pneumoniae Isolates in Slovenian Hospitals

First reported in the late 1980s and sporadically found in different hosts for a decade, CTX-M-type enzymes have now become the most prevalent extended-spectrum β-lactamases (ESBLs) in much of the world, the most widespread variant being the CTX-M-15 enzyme (4, 9). Their species distribution differs from that of TEM- and SHV-derived ESBLs, occurring mainly in Escherichia coli but also in Klebsiella pneumoniae (2, 9). E. coli strains with CTX-M ESBLs have proliferated in the community in some regions, although it remains unclear to what extent direct transmission is occurring, insofar as many “community” cases involve a history of recent hospital exposure (3, 11). Because of the extensive use of antibiotics, hospitals remain a perfect environment for selecting resistance genes (1) and it is plausible that K. pneumoniae, a major source organism for ESBLs, increasingly including CTX-M types, is acting as a vector for their transfer into E. coli and other species (1, 9).

To date, only SHV-type ESBLs have been reported in K. pneumoniae in Slovenia, and then only in one hospital (7), with no record of CTX-M ESBLs in any species. The aim of the present nationwide survey was therefore to investigate whether CTX-M-type ESBLs have now appeared in K. pneumoniae in Slovenian hospitals.

(These results were presented in part at the 17th European Congress of Clinical Microbiology and Infectious Diseases and 25th International Congress of Chemotherapy, Munich, Germany, 2007, poster P676.)

We analyzed nonreplicate isolates of ESBL-producing K. pneumoniae from 177 patients with infections and/or colonization collected in 11 Slovenian hospitals from January 2005 to May 2006, covering almost all of the Slovenian patient population; just one microbiology department serving a small regional hospital did not participate. Identification and initial susceptibility testing results were confirmed with the Vitek 2 system (bioMérieux, Marcy l'Étoile, France) with the ID-GN and the AST-N041 test cards. ESBL production was confirmed by ESBL Etests (AB Biodisk, Solna, Sweden) with ceftazidime, cefotaxime, and cefepime strips, each with and without clavulanate. MICs were determined by Etest, with the results interpreted in accordance with CLSI guidelines (6).

CTX-M alleles were sought by a multiplex PCR assay which can detect all five major CTX-M lineages (14). bla_CTX-M genes were sequenced on a CEQ 8000 (Beckmann-Coulter, High Wycombe, United Kingdom) with primers for ISEcp1 (12), with additional internal sequencing primers designed for this study (CTX-M-G1-IF [5′-AAA CTC TGC GGA ATC TGA CG-3′] and CTX-M-G1-IR [5′-TCG GTT CGC TTT CAC TTT TC-3′]).

Clonal relationships were studied by pulsed-field gel electrophoresis (PFGE) of XbaI-digested genomic DNA with a CHEF DRII apparatus (Bio-Rad Laboratories, Hemel Hempstead, United Kingdom) (13). Banding patterns were analyzed with BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium) with isolate clustering performed with Dice's coefficient in combination with the unweighted-pair group method using average linkages. Strains were defined as having PFGE profiles with ≥80% similarity.

To investigate whether plasmid transfer contributed to the spread of CTX-M ESBLs, conjugative transfer of bla_CTX-M genes was studied by mating experiments on nutrient agar plates with single representatives of the major strains as plasmid donors and E. coli J53-2 pro Rif^r as the recipient. Transconjugants were selected on nutrient agar plates supplemented with 2 μg/ml cefotaxime and 200 μg/ml rifampin (Sigma-Aldrich, Poole, United Kingdom) (8). Hybridization studies were performed on the same plasmids with a digoxigenin-labeled probe derived from the bla_CTX-M gene sequence with the primer pair MA1/MA2 (8, 13).

Among the K. pneumoniae isolates collected in 2005 and in the first few months of 2006, we noted a statistically significant increase in the prevalence of ESBL production (P = 0.005) (Table 1). CTX-M-type β-lactamase genes were detected in 60 (34%) of the 177 ESBL-producing isolates analyzed, thus currently still representing a minority of the ESBL-producing K. pneumoniae strains in Slovenia. As in Italy and Poland (9, 10), only group 1 CTX-M enzymes were found. These were identified in 8 of the 11 participating hospitals, with a countrywide distribution (Table 1). The large difference in the prevalence of CTX-M enzymes among the eight affected hospitals (Table 1) may suggest that Slovenia is in an early stage of the dissemination of these ESBLs, as recently reported in Italy (10), or may reflect variation in the success of infection control at particular sites.

In contrast to our neighboring country Italy, where CTX-M-15, CTX-M-1, and, less often, CTX-M-32 enzymes were detected (10), we identified only CTX-M-15. Each bla_CTX-M-15 gene was linked to an upstream ISEcp1-like element, shown previously to play a key role in their mobilization (5).

All isolates with bla_CTX-M genes showed substantial resistance to both cefotaxime and ceftazidime (Table 2) with uniform susceptibility to imipenem. Most were multiresistant, including to ciprofloxacin and gentamicin. Variable susceptibility to amoxicillin-clavulanic acid and piperacillin-tazobactam may have reflected the presence or absence of the OXA-1 inhibitor-resistant penicillinase, which is often associated with CTX-M-15 enzymes (8). Further analysis of resistance genes encoded by CTX-M-carrying plasmids is in progress.

Comparison by PFGE showed that the isolates with CTX-M ESBLs included 11 lineages (Fig. 1). Several clusters of related isolates were observed, some of them including isolates from more than one center. The largest cluster (strain A) comprised 26 clonally related producers of CTX-M group 1 ESBLs, 24 from one hospital (H9), along with 2 from further hospitals (H10 and H11), indicating possible spread with patients transferred from one hospital to another, reflecting common patient transfer patterns. Three other substantial clusters were noted (strains B, C, and D) (Fig. 1). The most complex epidemiology was found in larger hospitals (H9, H10, and H11); in hospital H11, strains A to E were all recognized (Table 1), with clusters of isolates belonging to strains B, C, and D. These clusters were from different clinical wards; e.g., the strain C isolates were from the urine of patients treated on urology wards, as well as on a nephrology ward and the infection ward, reflecting patient transfer patterns at the site. Smaller hospitals were mainly affected by only single clones with CTX-M ESBLs (e.g., H7 by strain B), which were always found also in larger centers (Table 1).

Clearly, the clonal spread of K. pneumoniae strains producing CTX-M enzymes is an important route of dissemination of these enzymes in our hospitals; in addition, the bla_CTX-M-carrying plasmids from the three major outbreak strains, A, B, and C, were highly conjugative, with transfer frequencies of 3.7 × 10⁻⁷, 2.8 × 10⁻³, and 1.1 × 10⁻⁵ per donor cell, respectively. Hybridization studies showed that bla_CTX-M-15 was located on ∼150-kb plasmids. The resistance phenotypes conferred by the transferred plasmids are shown in Table 3. Thus, spread of bla_CTX-M-carrying plasmids to other strains of K. pneumoniae, and perhaps also to E. coli, seems likely, as reported for an epidemic CTX-M-3-encoding plasmid in Poland (1, 9).

We have shown a complex epidemiology with endemic K. pneumoniae strains producing CTX-M-15 enzymes in multiple Slovenian hospitals and causing smaller clonal outbreaks in several. The ready transferability of the plasmids from the main outbreak strains increases the risk of their transfer to other bacterial species, like E. coli, and into the community.

• Open in new tab
• Download powerpoint

FIG. 1.

Dendrogram based on PFGE typing of 60 bla_CTX-M-positive K. pneumoniae isolates, in comparison with 27 bla_CTX-M-negative ESBL-producing isolates, from 11 Slovenian hospitals to illustrate their relatedness. The vertical line indicates the ≥80% similarity score adopted to assign isolates to the same strain (unweighted-pair group method using average linkages, Dice method).

• Copyright © 2009 American Society for Microbiology