Molecular epidemiological typing of Enterobacter cloacae isolates from a neonatal intensive care unit: three-year prospective study

Journal of Hospital Infection (2001) 49: 173–182 doi:10.1053/jhin. 2001.1053, available online at http://www.idealibrary.com on

Molecular epidemiological typing of Enterobacter cloacae isolates from a neonatal intensive care unit: three-year prospective study V. Fernández-Baca*, F. Ballesteros*||, J. A. Hervás†, P. Villalón*, M. A. Domínguez‡, V. J. Benedí§ and S. Albertí§¶ *Servicio de Microbiología, †Servicio de Pediatría and ¶Unidad de Investigación, Hospital Universitario Son Dureta, Palma de Mallorca, ‡Servicio de Microbiología, Hospital de Bellvitge, Barcelona, and §Área de Microbiología, Departamento de Biología, Universidad de las Islas Baleares and IMEDEA (CSIC-UIB), Palma de Mallorca, Spain

Summary: Since 1992, there has been an increase in the incidence of Enterobacter sepsis in the neonatal intensive care unit (NICU) of the authors’ hospital. From 1995 to 1997, a prospective molecular epidemiological survey of the colonizing and infecting strains isolated from neonates was conducted. Enterobacter cloacae was the most frequent cause of neonatal sepsis, accounting for 19.2% of all neonatal infections, reaching a peak incidence of 2.2/1000 during 1996. Fifty isolates from the NICU and four epidemiologically unrelated strains were characterized by pulse-field gel electrophoresis (PFGE), ribotyping, enterobacterial repetitive intergenic consensus (ERIC)-PCR and plasmid profiling. PFGE was the most discriminatory technique and identified 13 types (two of them classified into two and three subtypes) compared with ERIC-PCR, plasmid profiling and ribotyping that identified 11, 11 and seven types, respectively. A good correlation was found between all techniques. Five different clones caused 15 cases of sepsis. Clones A and B were prevalent in 1995 and 1996, but they were not isolated in 1997. An outbreak caused by clone G in 1997 was controlled by cohort nursing and hygienic measures, without changing the antibiotic policy. Strains were characterized by their antibiotic resistance pattern and divided into three groups. Group I correlated with PFGE types A, B1 and B2, which hyperproduced Bush type 1 chromosomal ␤-lactamase and expressed extended-spectrum ␤-lactamases (ESBLs). Group II only hyperproduced Bush type 1 chromosomal ␤-lactamase and correlated with PFGE-types D1, D2, D3 and I. Finally, Group III, with inducible ␤-lactamases, correlated with the rest of PFGE types. The sudden disappearance of E. cloacae after reinforcement of hygienic measures confirms the importance of patient-to-patient transmission. © 2001 The Hospital Infection Society

Keywords: Nosocomial infections; Enterobacter cloacae; neonate.

Received 1 February 2001; revised manuscript accepted 10 July 2001; published online 10 September 2001. || Present address: Área de Microbiologia, Departamento de Biologia, Universidad de las Islas Baleares, Palma de Mallorca, Spain. Author for correspondence: Dr S. Alberti, Unidad de Investigación. Hospital Son Dureta, Andrea Doria 55, Edificio D, 1a planta, Palma de Mallorca 07014, Spain. Fax: ;34-971-175228; E-mail: [email protected]

0195-6701/01/030173;10 $35.00/0

Introduction Enterobacter cloacae is a saprophytic bacterium of the normal human gastrointestinal tract that has emerged as an important opportunistic pathogen in hospitalized patients.1–5 Its intrinsic resistance to ampicillin and narrow-spectrum cephalosporins, as © 2001 The Hospital Infection Society

174

well as the frequent selection of mutants resistant to extended- and broad-spectrum cephalosporins or to aminoglycosides, has contributed to the increase in the prevalence of this species, particularly in the intensive care units where these antibiotics are used frequently.3 The origin of the majority of the nosocomial infections caused by E. cloacae is an endogenous source such as the gastrointestinal tract, especially in debilitated patients that previously received antibiotic therapy.3,6–9 However, when nosocomial outbreaks caused by this micro-organism occur, they are usually subsequent either to the administration of contaminated pharmaceutical products or patient-to-patient cross-infection via hospital materials and/or the hands of personnel.3,10,11 Traditional typing methods used to compare strains of E. cloacae are based on phenotypic characteristics, phage typing, bacteriocin typing, antibiotic resistance phenotype, serotyping and biotyping systems. These techniques cannot be applied to all strains because their discriminatory power is insufficient.12 However, in the last few years, genotypic methods have replaced the phenotypic methods because of their increased discriminatory power, and have been applied successfully to a large number of epidemiological studies.4,5,10,11 Genotypic techniques include pulse-field gel electrophoresis (PFGE), ribotyping, plasmid pattern analysis and PCR-based typing methods. All of the methods detect DNA polymorphisms to varying degrees. However, PFGE and ribotyping are probably the most reliable and reproducible typing procedures. Furthermore, PFGE allows the detection of a high degree of DNA polymorphism. On the other hand, PCR-based typing methods are more rapid and simple than PFGE and ribotyping, but less reproducible because of the low level of stringency inherent in the procedure. As described in a previous study,13 a resurgence of infections caused by Enterobacter species was observed in the authors’ neonatal intensive care unit (NICU) during the last decade. Thus, in the 1992–1998 period, Enterobacter became the most frequent pathogen causing nosocomial infections in neonates. When this trend was initially observed in 1994, the authors conducted a three-year prospective epidemiological survey of the strains of E. cloacae isolated between 1995 and 1997 from neonates in order to determine their relatedness, and also the reservoir of the infections and the mode of transmission. The majority of these isolates were resistant to multiple antibiotics, in

V. Fernández-Baca et al.

particular to ␤-lactams, because of the chromosomal derepressed cephalosporinase alone or in association with ESBLs. The results of this epidemiological investigation are reported. Methods Setting The NICU at Son Dureta University Hospital is the only tertiary centre in the Balearic Islands (Spain). This NICU is an eight-bed unit that cares for approximately 160 neonates per year. Records of bacteriologically proven sepsis and information on the state of bacterial colonization are currently maintained in the unit. A case of bloodstream infection was defined as bacteraemia using the criteria established by the Centres for Disease Control and Prevention. Blood cultures positive for Enterobacter species obtained from a central venous or arterial catheter were only considered as bacteraemia if this was corroborated by an additional positive blood culture from a peripheral vessel. For the analysis of incidence of neonatal sepsis, only patients born at this hospital were considered. Colonization was defined as isolation from one or more superficial sites in the absence of bloodstream infection. Microbiology In order to determine the evolution of nosocomial infection and colonization of neonates by E. cloacae, all infants admitted to the NICU during the threeyear prospective study were routinely screened for bacterial colonization on arrival in the unit and at regular weekly intervals thereafter, and also on suspicion of sepsis. The E. cloacae strains used in this study were from colonized and infected neonates, and also from environmental samples collected from the NICU. For comparison, some unrelated strains of E. cloacae isolated from an adult intensive care unit located in other building at the same hospital were also used. Enterobacter cloacae isolates were primarily identified in the microbiology laboratory by standard culture techniques. Subsequently, the identities of the strains were confirmed by the API 20E system (bioMérieux). Ribotyping analysis Ten micrograms of purified genomic DNA, isolated following standard procedures,14 was digested

Molecular epidemiology of Enterobacter cloacae

with Bgl I according to the specifications of the manufacturer (Pharmacia). DNA fragments were separated on a 1% agarose gel. Southern blot analysis and probe labelling and detection were carried out using the ECL kit (Amersham). Probe was obtained by PCR amplification of the 16SRNA gene using chromosomal DNA from E. cloacae ATCC13047 as a template and primers 616V (5⬘-AGAGGTTTGATYMTGCTGGCTCAG-3⬘) and 630R (5⬘-CAKAAAGGAGGTGATCC-3⬘) annealing to conserved regions of the 16SRNA gene. PCR amplifications were performed in a Thermoline Amplitron 1 thermal cycler using Taq polymerase (Pharmacia) with 30 cycles of amplification (1 min at 94⬚C, 1 min at 55⬚C, 1 min at 72⬚C). Plasmid profile analysis Plasmid DNA was isolated by the alkaline lysis method using the protocol and the reagents from a commercial kit (Wizard Miniprep kit, Promega). Plasmids were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. ERIC-PCR analysis Amplification reactions were carried out in a Thermoline Amplitron 1 thermal cycler using primers enterobacterial repetitive intergenic consensus (ERIC)-1R (5⬘-ATGTAAGCTCCTGG-GGATTCAC-3⬘) and ERIC-2 (5⬘-AAGTAAGTGACTGGGGTGAGCG-3⬘), which represents extragenic sequences, highly conserved and dispersed in many eubacterial species. Reactions were performed in a total volume of 50 ␮ l, containing 500 ng of DNA as the template, 500 pmol of each primer, 200 ␮ M nucleotides mix (Pharmacia), 10 mM Tris-HCl (pH 8.3), 5 mM KCl, 1.5 mM MgCl2 and 2.5 U of Taq polymerase (Pharmacia). Reactions were started after a preheat time of 5 min at 94⬚C. The basic cycling protocol consisted of 35 cycles of amplification. Cycling parameters were denaturation at 94⬚C for 1 min, annealing at 45⬚C for 1 min, and extension at 72⬚C for 1 min. Finally, samples were incubated at 72⬚C for 10 min to complete extension of the PCR products. PFGE analysis Analysis of chromosomal DNA was carried out by PFGE following standard procedures.15 Chromosomal DNA was digested with XbaI (NewEngland

175

Biolabs) and fragments were separated in a CHEFDRIII apparatus (Bio Rad). Electrophoretic pulses were linearly distributed from 1 to 30 s, for a run time of 23 h. The voltage was 6v/cm and the temperature of the electrophoresis chamber was kept at 14⬚C. The gels were stained with ethidium bromide and photographed. Differences in banding patterns were documented by visual examination and indexed by capital lettering. Interpretation of restriction fragment patterns was performed in accordance with recent consensus publications.16,17 Isolates were considered indistinguishable if there were no chromosomal band differences, related if they differed by one to three bands, and unrelated if they differed by seven or more bands. Isolates with banding patterns that differed in more than three bands and less than seven were considered as possibly related isolates.16 Antibiotic resistance testing The susceptibility to piperacillin, cefotaxime, ceftazidime, aztreonam and gentamicin were determined by broth microdilution according to the National Committee for Clinical Laboratory Standards recommendations with cation-adjusted Mueller-Hinton broth.18 Antibiotic-standard powders with stated potencies supplied by the drug manufacturers were used.

␤-Lactamase extraction and analytical isoelectric focusing ␤-Lactamases were released by ultrasonic treatment, and their pIs were determined by isoelectric focusing in polyacrylamide gels as described by Matthew et al.19 To distinguish between chromosome and plasmid-mediated ␤-lactamases, the authors used the methodology described by Sanders et al.20 ␤-Lactamase bands were visualized after isoelectric focusing with nitrocefin. Screening test for ESBL-producing strains Enterobacter cloacae plasmids were transferred to Escherichia coli by conjugation, and then the presence of ESBL activity was determined by placing disks containing cefotaxime, ceftazidime, or aztreonam near a disk containing a ␤-lactamase inhibitor (amoxicillin-clavulanic acid). Extension of the zone of inhibition toward the disk containing clavulanate indicated the presence of an ESBL.21

176

Results General epidemiology From January 1995 to December 1997, there were 456 neonates admitted to the NICU. Three hundred and forty-eight cases (76.3%) were neonates born in the hospital, and 108 (23.7%) were transferred from other hospitals. Three hundred and fifty patients (76.7%) were premature. The mean occupancy rate of the unit was 84% and the average length of stay of the patients was 16 days. During this period, there were 78 cases of neonatal sepsis, which accounted for a total incidence rate of 7.5 cases per 1000 live births. During 1995 and 1996, E. cloacae was the most frequent cause of nosocomial infection with an incidence of 0.8/1000 and 2.2/1000, respectively. In 1997, the most common organisms were E. coli (2.8/1000), E. cloacae (1.3/1000) and Staphylococcus epidermidis (1.3/1000). Eight of the 78 patients with sepsis died (10.2%, 0.8/1000), and two of the deaths were associated with E. cloacae sepsis (2/15, 13%, 0.2/1000). From 1995 to 1997, 175 E. cloacae isolates were recovered from 98 colonized neonates, and there were 16 isolates that caused bacteraemia, two in the same patient. This study included all blood isolates (16) and 31 isolates selected randomly from colonized patients. Three isolates from the NICU environment and four epidemiologically unrelated strains collected from the adults’ intensive care unit were also included. Sources and dates of isolation of the E. cloacae strains of this epidemiological study are shown in Table I.

The 1997 outbreak and the infection control measures In June 1997, three cases of sepsis caused by E. cloacae occurred in a single week and one of the patients died. Due to this outbreak, the NICU was temporarily closed, and it was thoroughly cleaned and disinfected. In an attempt to identify a primary source of infection, surveillance cultures of babies (eye, throat or tracheal aspirate, stool), staff (hands), the environment and instruments (sinks, handbasins, incubators, respirators, blood-gas analyser and milk) were carried out. Enterobacter cloacae was not recovered from the hands of personnel nor from the NICU instruments. The strain isolated from a sink proved later to be an unrelated strain.

V. Fernández-Baca et al.

Infection control measures on the unit were reviewed and personnel were urged to be stricter in their use. Cohort nursing of infected or colonized babies and proper handwashing was implemented. As a new measure, the use of new gloves when handling each infant was required. No change in the antibiotic policy was made. After these procedures, one additional case of sepsis by E. cloacae occurred in August 1997. However, no case of infection by this organism was observed in the following two years. Molecular typing of E. cloacae isolates All E. cloacae isolates from the NICU and four unrelated strains were typeable by ribotyping, ERIC-PCR, PFGE and plasmid profiling (Table I). Ribotyping with genomic DNA from NICU isolates digested with BglI and probed with an internal fragment of the 16SRNA gene led to seven different patterns, which are illustrated in Figure 1. Patterns with one discordant band were considered different. Each strain was tested at least twice and with four different enzymes (BglI, SalI, EcoRI and HindIII); however, BglI-digested fragments were more easily interpretable than the SalI, EcoRI and HindIII-digested fragments (data not shown). Type A and type B were the most common isolates (A: 17 isolates, 34%; B: eight isolates, 16%) followed by types C and D (seven isolates, 14% for each). After ERIC-PCR analysis (Figure 2), the E. cloacae NICU isolates could be divided into 11 types. Each strain was assayed at least three times and the profiles of these types were reproducible. With this technique, the prevalent clone was type A (17 isolates, 34%) followed by types, C and B (six and seven isolates, 14% and 12%, respectively). Representative PFGE profiles of the NICU isolates are shown in Figure 3. Thirteen distinct PFGE types were identified, two (B and D) of them divided in two (B1 and B2) and three (D1, D2 and D3) subtypes, among 50 isolates from NICU. As occurred with the other techniques, type A was the most frequent isolate (17 isolates, 34%) followed by type C (six isolates, 12%), G (five isolates, 10%) and B (five isolates, 10%). Plasmids were isolated in all E. cloacae isolates and established 11 types. Successive plasmid DNA preparations were found to be identical in the same isolate over time. Most of the strains harbored more than four plasmids of different sizes. All strains presented a high molecular weight plasmid (923 Kb) and plasmids ranging in size from 2 to 8 Kb.

Molecular epidemiology of Enterobacter cloacae

177

Table I

Characteristics of Enterobacter cloacae isolates from the neonatal intensive care unit (NICU)

Isolate no.

Patient no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

1 2 3 4 1 5 6 7 8 1 9 10 11 9 12 13 13 14 12 15 16 17 18 19 5 20 21 22 23 23 24 24 25 26 27 28 29 30 31 30 32 31 – – – 33 33 34 35 36

Source

Date of isolation

Ribotype

ERIC type

PFGE type

Plasmid profile

Catheter Blood culture Catheter Catheter Blood culture BAL Catheter Eye Catheter BAL Eye Catheter Eye Blood culture Catheter Wound infection Peritoneal exudate Catheter Catheter Blood culture Blood culture Blood culture Blood culture BAL BAL Eye Eye Blood culture Blood culture Blood culture Blood culture Catheter Eye Blood culture Catheter Eye Throat swab Tracheal aspirate Catheter Blood culture Blood culture Blood culture Sink of NICU Sink of NICU Breast milk Throat swab Urine Blood culture Throat swab Eye

13/01/95 13/01/95 26/01/95 31/01/95 03/02/95 06/02/95 07/02/95 14/02/95 28/02/95 01/03/95 13/03/95 13/03/95 15/03/95 27/03/95 10/04/95 25/04/95 28/04/95 19/05/95 03/06/95 13/02/96 26/02/96 14/03/96 15/03/96 16/03/96 18/04/96 01/07/96 01/07/96 11/07/96 17/07/96 31/07/96 01/08/96 03/08/96 06/08/96 05/11/96 03/02/97 06/02/97 16/06/97 16/06/97 17/06/97 18/06/97 18/06/97 24/06/97 25/06/97 10/07/97 10/07/97 28/07/97 30/07/97 14/08/97 29/08/97 29/08/97

A B B B A C C A D A A A ND B A B B D C A A A A A D A D A A A B B A D E E C G G G G G D D C C H C E E

A B B B A C C A J A A A F B A B B D C A A A A A D A D A A A I I A D E E C G G G G G K K C C H C E E

A B2 L B1 A C C A J A A A F B1 A B1 B1 D1 C A A A A A D2 A D1 A A A I I A D3 E E C G G G G G K K C C H M E E

A A B A A C C A F A A A F A A A A D C A A A A A D A D A A A I I A D E E C G G G G G K F C C H M E E

01/04/96 20/05/96 22/03/96 01/05/97

P R S W

P R S W

ND R S ND

ND ND ND ND

Epidemiologically unrelated isolates 51 37 AICU 52 38 AICU 53 39 AICU 54 40 AICU

BAL, bronchoalveolar lavage; AICU, adults intensive care unit; ND, not determined. Dates are given as day/month/year.

178

V. Fernández-Baca et al.

Figure 1 Representative ribotypes of Enterobacter cloacae neonatal intensive care unit isolates generated by BglI. Bacterial genomic DNA was isolated and digested with BglI. DNA fragments were resolved by electrophoresis, transferred to nylon and probed with a 16SRNA gene fragment. Capital letters on the lanes refer to types listed in Table I and numbers to the number of isolates of each type. Molecular size markers (in Kb) are indicated on the left.

Figure 2 Representative patterns of Enterobacter cloacae neonatal intensive care unit isolates generated by ERIC-PCR. Bacterial genomic DNA was subjected to PCR amplification using primers ERIC-1R and ERIC-2. PCR products were separated by electrophoresis, stained with ethidium bromide and visualized under ultraviolet light. Capital letters on the lanes refer to types listed in Table I and numbers to the number of isolates of each type. Molecular size markers (in Kb) are indicated on the left.

All molecular techniques used in this study were highly reproducible. All molecular techniques were applied at least three times to all E. cloacae isolates included in the study, obtaining identical results in all experiments. In addition, molecular typing patterns of isolates from the same patient on different days were consistent except for patients 5, 9, 12 and 33 that were infected by two different clones of E. cloacae. Furthermore, all epidemiologically unrelated strains had distinct PFGE, ribotype and ERIC-PCR profiles compared with the E. cloacae NICU isolates. There was a good correlation between the results obtained by PFGE, ribotyping, ERIC-PCR and

plasmid profiling. However, PFGE was more discriminatory than ERIC-PCR and ribotyping (Table II). For example, type D obtained by ribotyping could be divided into types D, J and K by ERIC-PCR typing and into types D1, D2, D3, J and K by PFGE. In summary, results obtained by molecular typing techniques indicate that clones A and B were prevalent during 1995 and 1996, representing together almost 75% of the isolates in these two years. Interestingly, neither clone was detected in 1997. Clone ribotype C was detected in 1995 and 1997 but not in 1996. In 1997, five new clones were identified: E, G, K, H and M. Clones E and G were prevalent in

Molecular epidemiology of Enterobacter cloacae

179

1997 representing together 56% of the isolates that year. Screening of environmental specimens detected only two isolates of E. cloacae in the sink of the NICU, and one in breast milk (isolates 43, 44

and 45, respectively). Although both isolates from the sink presented the same ribotyping profile (type D) as other isolates obtained from patients, PFGE and ERIC-PCR typing clearly demonstrated that they were different clones. However, strain 45 isolated from breast milk belonged to type C either by ribotyping, PFGE or ERIC-PCR typing, suggesting the existence of cross-infection since this clone was isolated in patients hospitalized during 1995–1997. No E. cloacae was found in any other environmental specimens. Antibiotyping

Figure 3 Representative profiles of Enterobacter cloacae neonatal intensive care unit isolates generated by PFGE types. Bacterial cells were embedded in agarose and lysed. Genomic DNA was digested with XbaI and fragments were separated by pulsed-field electrophoresis. Capital letters on the lanes refer to types listed in Table I and numbers to the number of isolates of each type. Molecular size markers (in Kb) are indicated on the left.

Table II

The 50 E. cloacae isolates were divided into three groups according to their resistance to piperacillin, cefotaxime, ceftazidime, aztreonam and gentamicin (Table II). The first group included 22 isolates, which were highly resistant to all the antibiotics tested. The second group included six strains, which were susceptible to gentamicin but highly resistant to the rest of the antibiotics. The third group consisted of 22 isolates, which were susceptible to all the antibiotics tested. Isolates from group I correlated with PFGE types A, B1 and B2. Isoelectrofocusing analysis of extracts from strains of group I identified the presence of the Bush type 1 chromosomal ␤-lactamase (pI ≈ 7.9), and revealed the existence of a plasmid ␤-lactamase

Molecular typing correlation and antibiotic resistance pattern of Enterobacter cloacae isolates

Ribotype

ERIC-PCR type

PFGE type

PRL*

CTX*

CAZ*

AZT*

GM*

␤-Lactamase type Antibiotyping group

A

A

A

512

128–256

128

16–32

64–128

HYP;ESBL

I

B1,B2

512

512

256

64–128

256

HYP;ESBL

I

L I

1 256

0.125 256

0.5 128

0.06 32

0.125 0.25–0.5

IND HYP

III II

8–4

0.5

:0.125

:0.125

0.25

IND

III

512

128

256

64–128

0.25

HYP

II

4

0.5

:0.125

:0.125

0.5

IND

III

B

B

B

I

C

C

C M D D

D J

D1,D2,D3 J

D

K

K

E

E

E

4

0.5

:0.125

:0.125

0.5

IND

III

ND

F

F

4

0.5

1

:0.125

0.5

IND

III

G

G

G

1–2

:0.125

:0.125

:0.125

0.25

IND

III

H

H

H

8

0.5

0.5

:0.125

0.5

IND

III

*

MICs (␮g/ml) for piperacillin (PRL), cefotaxime (CTX), ceftazidime (CAZ), aztreonam (AZT) and gentamicin (GM). HYP, hyperproduced Bush type 1 chromosomal ␤-lactamase; ESBL, extended-spectrum ␤-lactamase; IND, inducible ␤-lactamase.

180

(pI:5.5). Conjugation experiments and analysis of the susceptibilities of the resulting transconjugants indicated that the plasmid ␤-lactamase was an ESBL. Thus, strains of antibiotyping group I were resistant to the ␤-lactams (piperacillin, cefotaxime, ceftazidime and aztreonam). Furthermore, conjugation experiments indicated that gentamicin resistance was due to additional plasmid encoded resistance genes. Group II isolates hyperproduced Bush type 1 chromosomal ␤-lactamase and correlated with PFGE types I, D1, D2 and D3. Isoelectrofocusing analysis identified two chromosomal ␤-lactamases. PFGE type I strains presented a ␤-lactamase with a pI of approximately 7.9, while PFGE type D1, D2 and D3 strains encoded a ␤-lactamase with a pI of approximately 8.1. The authors did not detect plasmid ␤lactamases in group II isolates. These strains were resistant to ␤-lactams due to the chromosomal ␤-lactamase, but they were sensitive to gentamicin because they lacked the plasmid/s encoding for gentamicin resistance presented in the isolates of group I. All the other isolates included in antibiotyping group III presented inducible ␤-lactamases. For this reason, they were susceptible to all antibiotics tested. Discussion In 1992, Enterobacter became the most frequent cause of nosocomial infection in the authors’ NICU and its resurgence was confirmed in the following years. As shown in previous studies on the epidemiology of neonatal infections,1,13 from 1992 to 1996, Enterobacter replaced other organisms such as S. epidermidis and Klebsiella pneumoniae that had been the most prevalent in different times during the previous 15 years. From 1995 to 1997, sepsis caused by E. cloacae accounted for 19.2% of all neonatal infections and reached a maximal incidence of 2.2 cases per 1000 live births during 1996. These figures were clearly higher than ever in the authors’ NICU. In a situation like this, it is important to establish whether there was a general increase of infection in the unit or if a single clonal strain was causing a prolonged outbreak. The molecular studies demonstrated that Enterobacter had increased and that several different clones were present in this unit. Molecular typing techniques are very useful today for epidemiological investigations, especially of outbreaks occurring in NICUs. Phenotypic markers, including serological, phage typing, bacteriocin typing and biotyping systems,

V. Fernández-Baca et al.

are not sufficiently discriminatory to distinguish different strains of Enterobacter spp.3,6,12 On the contrary, some molecular typing methods have been applied successfully for epidemiological typing of E. cloacae.4,5,22 This study compares four different molecular techniques that have been described separately in different reports, i.e., ribotyping, ERICPCR, PFGE and plasmid profile analysis.4,22,23 All techniques provided discriminatory profiles that allowed differentiation between epidemiologically related and unrelated strains of E. cloacae. However, PFGE is the most powerful method for the analysis of nosocomial isolates because of its high reproducibility and discriminatory power.15,22,24 In this study, PFGE identified 13 types (two of them were subclassified into two and three subtypes) compared with ERIC-PCR, the second most-discriminatory technique used in this study (identified 11 types). ERIC-PCR is a more rapid and less laborious method than PFGE and provides stable profiles that allows discrimination between epidemiologically related and unrelated strains with almost the same discriminatory power obtained with PFGE. In contrast, although ribotyping has been found very useful for discriminating among strains of E. cloacae,6,12 in this study, ribotyping only identified seven types and was the least discriminatory method. Plasmid profile analysis has been used widely for comparing isolates of E. cloacae because it is a simple and rapid molecular technique. In this study, most of the epidemiologically unrelated strains were found to be different by this technique. However, instability of profiles caused by the acquisition or loss of plasmids represents a disadvantage illustrated previously, and detected in this study in patient 9. In addition, results obtained with this technique are difficult to interpret, particularly when plasmids are not linearized using restriction endonucleases or when the number of plasmids is low. In summary, despite the fact that a good correlation was found between all molecular techniques, these results indicate that PFGE and ERIC-PCR are the best methods to compare E. cloacae isolates in an epidemiological study. The molecular epidemiology has shown that the majority of E. cloacae strains isolated during the three years studied were from sporadic colonizations or infections. However, in June 1997, there was an outbreak caused by clone G, which resulted in the death of an infant. This outbreak was brought under control by cohort nursing and reinforcement of hygienic measures without limiting the use of

Molecular epidemiology of Enterobacter cloacae

cefotaxime. After a single case of sepsis caused by a different clone (clone C) that occurred two months later, no other case of sepsis by this micro-organism was observed in the following two years. From 1995 to 1997, five different clones caused sepsis in the authors’ NICU. Clones A and B accounted for the majority of the colonizing or infecting strains isolated from January 1995 to November 1996, but they were not isolated in 1997. The origin and the cause of the disappearance of the predominant clonal strains A and B was not known, although stricter control measures may have been relevant. Cross-contamination of patients by E. cloacae strains was proved in the authors’ NICU during the study period, but the environmental isolate obtained from the sink was not related to the 1997 outbreak. Although hand transfer of enterobacteria to patients by healthcare personnel has been implicated in outbreaks of nosocomial infection,2,23 the authors were unable to isolate E. cloacae from the hands of the studied personnel. However, the sudden disappearance of this micro-organism after the reinforcement of hygienic measures suggests that patient-to-patient transmission was most likely. Antibiotyping, although not discriminatory for the typing of different bacteria including E. cloacae, is very useful in epidemiological investigations. Thus, multiresistant E. cloacae strains isolated during 1995–1996 but not 1997, correlated with PFGE types A and B which hyperproduced identical Bush type I chromosomal B-lactamases (pI ≈ 7.9) and encoded the identical ESBLs with a pI of 5.5. The selection of multiresistant Enterobacter strains has been related to the use of third-generation cephalosporins.2,25 In fact, during the last decade, there has been a greater use of third-generation cephalosporins in the authors’ NICU, mainly cefotaxime, for empirical therapy of suspected nosocomial infections, due to a remarkable increase in the number of very low birthweight premature infants. However, the multiresistant strains isolated from the authors’ NICU during the first two years of study spontaneously disappeared in spite of maintaining the same antibiotic pressure on the neonate’s microflora. In other words, using the same antibiotic policy, some strains developed multiresistance and others did not. The authors suggest that the antibiotic pressure may not be the only factor in the selection of multiresistant strains of Enterobacter, and unknown bacterial features can possibly facilitate the acquisition of multiresistance.

181

A certain level of use of antibiotic may also be required to trigger the mechanisms that cause multiresistance in E. cloacae. According to the authors’ data, multiresistance may not be related with a higher virulence of the strain. For example, clone G (which caused the June 1997 outbreak) was especially virulent but it was an inducible nonmultiresistant strain. In summary, this study emphasizes the importance of cross-transmission of E. cloacae in the NICU (most probably by the hands of healthcare personnel) and the value of molecular typing methods for the investigation of nosocomial outbreaks. Frequent infection screenings of the neonates for early detection of enterobacteria and knowledge of their antibiotic resistance pattern is recommendable. This hygienic reinforcement alone, without any change in the antibiotic policy, has been enough to eradicate Enterobacter from the authors’ NICU for two years. Acknowledgements This work was supported by grants from Fondo de Investigaciones Sanitarias (FIS) and Comisión Interministerial de Ciencia y Tecnología (CICYT). References 1. Hervás JA, Alomar A, Salva F, Reina J, Benedí VJ. Neonatal sepsis and meningitis in Mallorca (Spain), 1977–1991. Clin Infec Dis 1993; 16: 719–724. 2. Acolet D, Ahmet Z, Houang E, Hurley R, Kaufmann ME. Enterobacter cloacae in a neonatal intensive care unit: account of an outbreak and its relationship to use of third generation cephalosporins. J Hosp Infect 1994; 28: 273–286. 3. Gaston MA. Enterobacter, an emerging nosocomial pathogen. J Hosp Infect 1988; 11: 197–208. 4. Grattard F, Pozzetto B, Berthelot P et al. Arbitrarily primed PCR, ribotyping, and plasmid pattern analysis applied to investigation of a nosocomial outbreak due to Enterobacter cloacae in a neonatal intensive care unit. J Clin Microbiol 1994; 32: 596–602. 5. van Nierop WH, Duse AG, Stewart RR, Bilgeri YR, Koornhof HJ. Molecular epidemiology of an outbreak of Enterobacter cloacae in the neonatal intensive care unit of a provincial hospital in Gauteng, South Africa. J Clin Microbiol 1998; 36: 3055–3087. 6. Bingen E, Denamur E, Lambert-Zechovsky N, Brahimi N, El Lakany M, Elion J. Rapid genotyping shows the absence of cross-contamination in Enterobacter cloacae nosocomial infection. J Hosp Infect 1992; 21: 95–101. 7. Falkiner FR. Enterobacter in hospital. J Hosp Infect 1992; 20: 137–140.

182

8. Flynn DM, Weinstein RA, Nathan C, Gaston MA, Kabins SA. Patients endogenous flora as the source of nosocomial Enterobacter in cardiac surgery. J Infect Dis 1987; 156: 363–368. 9. Lambert-Zechovsky N, Bingne E, Denamur E et al. Molecular analysis provides evidence for the endogenous origin of bacteremia and meningitis due to Enterobacter cloacae in an infant. Clin Infect Dis 1992; 15: 30–32. 10. Haertl R, Bandlow G. Epidemiological fingerprinting of Enterobacter cloacae by small-fragment restriction endonuclease analysis and pulsefield gel electrophoresis of genome restriction fragments. J Clin Microbiol 1993; 31: 128–133. 11. Wand CC, Chu ML, Ho LJ, Hwang RC. Analysis of plasmid pattern in paediatric intensive care unit outbreaks of nosocomial infection due to Enterobacter cloacae. J Hosp Infect 1991; 19: 33–40. 12. Garaizar J, Kaufmann ME, Pitt TL. Comparison of ribotyping with conventional methods for the type identification of Enterobacter cloacae. J Clin Microbiol 1991; 29: 1303–1307. 13. Hervás JA, Ballesteros F, Alomar A, Gil J, Benedí VJ, Alberti S. Increase of Enterobacter in neonatal sepsis: a 22-year study. Pediatric Infec Dis J 2001; 20: 134–140. 14. Ausubel FM, Brent R, Kingston RE et al. Current Protocols in Molecular Biology. New York: Greene Publishing and Wiley Interscience 1997. 15. Maslow JN, Slutsky AM, Arbeit RD. Application of pulse-field gel electrophoresis to molecular epidemiology. In: Persing DH et al. eds. Diagnostic Molecular Micro-biology; 1993; 563–571. 16. Tenover RC, Arbeit RD, Goering RV et al. Interpreting chromosomal DNA restriction patterns by pulsed-field electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995; 33: 2233–2239. 17. Struelens MJ, Baurnfeind A, van Belkum A et al. Consensus guidelines for appropiate use and evaluation

V. Fernández-Baca et al.

18.

19.

20.

21. 22.

23.

24.

25.

of microbial epidemiologic typing systems. Clin Microbiol Infect 1996; 2: 2–11. National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7-A4. Wayne, PA: National Committee for Clinical Laboratory Standards 1997. Matthew M, Harris AM, Smith SM, Williams DS. The use of analytical isoelectric focusing for detection and identification of ␤-lactamases. J Gen Microbiol 1975; 88: 169–178. Sanders CC, Sanders WE Jr, Moland ES. Characterization of ␤-lactamases in situ on polyacrylamide gels. Antimicrob Agents Chemother 1986; 30: 951–952. Philippon A, Labia R, Jacoby G. Extended-spectrum ␤-lactamases. Antimicrob Agents Chemother 1989; 33: 1131–1136. Shi Z, Yuk-Fong-Liu P, Lau Y, Lin Y, Hu B. Epidemiological typing of isolates from an outbreak of infection with multidrug-resistant Enterobacter cloacae by repetitive extragenic palindromic unit b1-primed PCR and pulse-field gel electrophoresis. J Clin Microbiol 1996; 34: 2784–2790. 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 Microbiol 1996; 34: 1474–1480. Jalaluddin S, Devaster JM, Scheen R, Gerard M, Butzler JP. Molecular epidemiology study of nosocomial Enterobacter aerogenes isolates in a Belgian Hospital. J Clin Microbiol 1998; 36: 1846–1852. Finnström O, Isaksson B, Hæggman S, Burman LG. Control of an outbreak of a highly beta-lactamresistant Enterobacter cloacae strain in a neonatal special care unit. Acta Pædiatr 1998; 87: 1070–1074.