Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1237

Epidemiology of Enterococci withAcquired Resistance toAntibiotics in Sweden

Special emphasis on Ampicillin and Vancomycin

BY

ERIK TORELL

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2003

Contents

List of papers ..................................................................................................1

Introduction and background ..........................................................................2 Genus description and habitat ....................................................................3 Clinical significance of enterococci. ..........................................................4

Epidemiology..................................................................................................5 Prevalence of ARE and VRE .....................................................................5 Epidemiology of ARE................................................................................6 Epidemiology of VRE................................................................................8

Nosocomial spread........................................................................................11 Spread in hospitals ...................................................................................11 Hands, thermometers and environmental surfaces...................................12

Antimicrobial resistance in enterococci........................................................14 Betalactams. .............................................................................................14 Glycopeptides...........................................................................................15 Fluoroquinolones......................................................................................18 Aminoglycosides......................................................................................19

Impact of antibiotic use on antimicrobial resistance in enterococci .............21 Selection of resistant strains.....................................................................21 Can a change of antibiotic strategies reduce the selection of resistant enterococci?..............................................................................................22 Are resistant strains a greater threat to the patients than susceptible ones?..................................................................................................................23

Typing methods. ...........................................................................................24

Virulence factors in enterococci ...................................................................26

Aims of this study .........................................................................................27

Materials and Methods..................................................................................28 Sample size and origin of enterococcal isolates.......................................28 Study designs. ..........................................................................................30 Microbiology............................................................................................33

Sampling and culture techniques.........................................................33 Identification of genus and species......................................................35 Antimicrobial susceptibility testing.....................................................36 Identification of gyrA, parC and esp genes. ........................................36 Typing..................................................................................................37

Statistics ...................................................................................................38 Ehtics........................................................................................................38

Results and discussion ..................................................................................39 Prevalence of VRE and ARE in Sweden (I, II)........................................39

The first hospital outbreak of VRE in Sweden (I) ...............................39 Fecal carriage rates of VRE and ARE in hospitals (II) .......................41

Risk factors for colonization and infection with ARE (II, III, IV)...........46 Antibiotics (II, III, V) ..........................................................................46 Other risk-factors for acquiring ARE (II, III, V) .................................49 Clinical impact of ARE infection (V)..................................................51

Shedding and spread of ARE (I, II, III, V)...............................................52 Shedding of ARE from colonized patients (III) ..................................52 Spread of VRE and ARE between patients in hospitals (I, II, III, V)..53

Is there a clonal origin of E. faecium in Sweden?(IV) .............................56 Presence and relation to invasiveness of esp in E. faecium (V). ..............62

General conclusions......................................................................................64

Final comment ..............................................................................................66

Acknowledgements.......................................................................................67

References.....................................................................................................69

Abbreviations

APACHE II Scoring system for severity of illness AME Aminoglycoside Modifying Enzyme ARE Ampicillin Resistant Enterococcus ASE Ampicillin Susceptible Entrococcus Cfu Colony forming unit CDC Centers for Diseases Control and Prevention CRP C-reactive protein DNA Deoxyribonucleic Acid HLR High Level Resistance HSCT Hematopoietic Stem Cell Transplantation ICU Intensive Care Unit PBP Penicillin Binding Protein PCR Polymerase Chain Reaction PFGE Pulsed Field Gel Electrophoresis MIC Minimal Inhibitory Concentration MLST Multi-Locus Sequence Typing QRDR Quinolone Resistance Determining Region RNA Ribonucleic Acid rRNA Ribosomal Ribonucleic Acid SMI Swedish Institute for Infectious Disease Control SRGA Swedish Reference Group for Antibiotics UUH Uppsala University Hospital VRE a Vancomycin Resistant Enterococcus VSE Vancomycin Susceptible Enterococcus

a In the U.S.A. where teicoplanin is not used, glycopetide resistant enterococci are referred to as VRE. In Europe where teicoplanin is widely used all glycopeptide resistant strains are referred to as GRE (glycopeptide resistant enterococci). In Sweden, the term VRE has become more established and in the following, strains with acquired resistance to glycopeptides will be referred to as VRE and glycopeptide resistant enterococci in general as GRE.

1

List of papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals: I Torell Erik, Fredlund Hans, Törnkvist Eva, Myhre Erling, Sjöberg

Lennart, Sundsfjord Arnfinn. (1997). Intrahospital spread of vancomycin-resistant Enterococcus faecium in Sweden.. Scandinavian Journal of Infectious Diseases 29(3): 259-263

II Torell Erik, Hoffman Britt-Marie, Olsson-Liljequist Barbro, Cars

Otto, Burman Lars G. (1999). Near absence of VRE but high carriage rates of quinolone-resistant ampicillin-resistant enterococci among hospitalized patients and non-hospitalized individuals in Sweden. J Clin Microbiol. 37(11): 3509-3513

III Torell Erik, Cars Otto, Hambraeus Anna. (2001). Ampicillin-

resistant enterococci in a Swedish University Hospital; nosocomial spread and risk-factors for infection. Scandinavian Journal of Infectious Diseases 33(3): 182-187

IV Torell Erik, Olsson-Liljequist Barbro, Hoffman Britt-Marie ,

Haeggman Sara, Lindahl Cecilia, Burman Lars G . Clonality among ampicillin-resistant Enterococcus faecium in Sweden and relation to ciprofloxacin resistance. Clinical Microbiology and Infection (Accepted for publication)

V Torell Erik, Lagerqvist-Widh Angela, Johansson Lotta, Gillman

Anna, Hjalmarsson Sandra , Gustavsson Ingegerd, Cars Otto, Hambraeus Anna. Risk-factors for ampicillin-resistant Enterococcus faecium bacteremia in a hematology unit and presence of the variant esp gene in isolates. (Manuscript)

2

Introduction and background

The name Enterococcus first appeared in a paper written by Thiercelin in C.R. Soc. Biol. 1899, “Sur un diplocoque saprophyte de l`intestine susceptible de devenir pathogene” (Thiercelin 1899). The pathogenicity of enterococci was recognized the same year by MacCallum and Hastings who isolated an organism which they named Micrococcus zymogenes with properties consistent with those of enterococci from a case of acute endocarditis (McCallum and Hastings 1899). During the next three decades, these Gram-positive bacteria originating from the human intestine were more intensively studied. They seemed associated with endocarditis, wound infection, urinary tract infection, puerperal sepsis and even food poisoning, and the name Streptococcus faecalis was used more frequently by various authors (Murray 1990). In the 1930s, Lancefield suggested the division of streptococci into four groups based on their reaction with A-G antisera and enterococci were classified as group D streptococci. In 1970, a separate genus named Enterococcus including Streptococcus faecalis and Streptococcus faecium, was suggested (Kalina 1970). The genus was not accepted, however, until genetic evidence fully distinguished the enterococci from streptococci in 1984 (Schliefer and Killper-Baltz 1984). Until now, DNA-DNA, DNA-rRNA hybridizations and 16S rRNA sequencing studies have so far resulted in 23 distinct species included in the genus Enterococcus. (Table 1).

3

Table 1. Species proposed for inclusion in the genus Enterococcus.

Species Year of

description

Reference

E. faecalis 1984 (Schliefer and Killper-Baltz 1984) E. faecium 1984 (Schliefer and Killper-Baltz 1984) E. avium 1984 (Collins 1984) E. casseliflavusa 1984 (Collins 1984) E.gallinarium 1984 (Collins 1984) E. durans 1984 (Collins 1984) E. malodoratus 1984 (Collins 1984) E. hirae 1985 (Farrow and Collins 1985) E. mundtii 1986 (Collins 1986) E. pseudoavium 1989 (Collins 1989) E. raffinosus 1989 (Collins 1989) E. cecorum 1989 (Williams 1989) E. saccharolyticus 1990 (Rodrigues and Collins 1990) E. columbae 1990 (Devriese 1990) E. dispar 1991 (Collins 1991) E. sulfureus 1991 (Martinez-Murcia and Collins 1991) E. flavescensa 1992 (Pompei 1992) E. asini 1998 (de Vaux 1998) E. ratti 2001 (Teixeira 2001) E. porcinusb 2001 (Teixeira 2001) E. villorumb 2001 (Vancanneyt 2001) E. haemoperoxidus 2001 (Svec 2001) E. moraviensis 2001 (Svec 2001) E. pallens 2002 (Tyrrell 2002) E. gilvus 2002 (Tyrrell 2002)

aDNA reassociation studies indicate these to be the same species bDNA homology studies indicate these two to be the same specie.

Genus description and habitat Enterococci are facultative anaerobic Gram positive cocci that appear singly, in pairs and in short chains. The optimum growth temperature is 35ºC and most strains grow at 10 and 45ºC. All strains grow in broth containing 6.5% NaCl, produce leucine aminopeptidase (LAP) and hydrolyze esculin in the presence of 40% bile salts, which kills most other microorganisms. Enterococci are usually catalase negative but strains occasionally produce pseudocatalase. Most species hydrolyze pyrrolidonyl-ß-naphtylamide (PYR) and produce a cell wall antigen that is identified as the streptococcal group D antigen. They are considered strict fermenters because they lack a Kreb´s cycle respiratory chain.

4

Enterococci are widespread in nature, can grow and persist in harsh environments and have been detected in the fecal flora of most animals, from insects to mammals. They are also readily recovered from foods such as milk and meat products, from various environmental sources and in waste and surface water (Aarestrup 2002). Enterococci are part of the normal bacterial flora of the human bowel and E. faecalis have been one of the most common bacteria isolated from feces of healthy individuals. It is also the dominating species among enterococci isolated from infected sites (approximately 80%), and with E. faecium being isolated from most of the rest. Since strains with acquired resistance to ampicillin and vancomycin have been predominantly E. faecium, this species has increased among hospital enterococcal isolates during the last 15 years (Iwen 1997).

Clinical significance of enterococci In community acquired infections, enterococci are commonly associated with urinary tract and wound infections, most often caused by E. faecalis. In infected wounds, such as diabetic foot wounds, E. faecalis is regularly isolated as part of a polymicrobial flora and the clinical significance of enterococci in such cultures is often ambiguous. It was very early recognized that enterococci were able to cause bacteremia and endocarditis and it has been estimated that 5-20% of all endocarditis cases are caused by enterococci (Murray 1990). Since the late 1980s, enterococci, and mainly E. faecium, have emerged as important nosocomial pathogens with the ability to acquire resistance to almost all known classes of antibiotics. In a point-prevalence study on nosocomial urinary tract infection in 228 European hospitals during 1999, enterococci were the second most commonly isolated microorganisms (15.8%) (Bouza 2001). Emerging nosocomial enterococcal infections include bacteremia, surgical site and intra-abdominal infections, and more rarely CNS, neonatal and pulmonary infections (Richards 2000) (Sitges-Serra 2002) (Dettenkofer 1999) (Nachman 1995).

In spite of the wide range of infections associated with enterococci they are generally considered less virulent than streptococci and S. aureus. It has been noted that some enterococcal infections, even blood-stream infections, resolve without specific therapy (Quale 1996) (Dougherty 1984). The underlying condition of the patient seems to play an important role for the outcome of enterococcal infections. Patients with hematological malignancies, a history of transplantation or severe burns have been more readily colonized with multi-resistant strains and have also been more likely to experience bacteremia and subsequent serious outcome than non-immunocompromised patients (Papanicolaou 1996, Bach 2002) (Still 2001) (Roghmann 1997) (Montecalvo 1996) (Linden 1996).

5

Epidemiology

The study of the epidemiology of enterococci with acquired antimicrobial resistance is interesting but complex. There are contrasting differences between continents and sometimes even between individual countries, depending on the resistance phenotype and genotype studied. Factors associated with these contrasting findings are associated with differences in the use of antimicrobial agents among humans and animals as well as differeces asssociated with spread and colonization of individuals in different countries.

Prevalence of ARE and VRE

North-America According to the National Nosocomial Infection Surveillance System of the CDC, the proportion of VRE of the total number of enterococcal blood isolates increased from 13% to 26% between 1995 and 2000. Data from the TNS database in 1999, collecting information from more than 100 clinical U.S. laboratories, showed resistance to ampicillin and the glycopeptides to be rare in E. faecalis but very common in E. faecium, 83% and 52%, respectively (Huycke 1998) (Anonymous 2000).

Europe Data from the European Antimicrobial Resistance Surveillance System (EARSS 2001) on invasive enterococcal isolates from 24 countries revealed considerable differences in resistance rates between countries. In 2001, the highest proportions of VRE among E. faecium were reported from Greece, Italy and Portugal (14%, 15% and 21%, respectively) but most countries reported rates of 0-5%. Ampicillin resistance among invasive strains of E. faecium was frequently found (mean 71%, range 40-86%) indicating an increasing prevalence of ARE strains in most European countries.

6

Nordic countries In Sweden 2001, 73% of E. faecium (n=196) but no E. faecalis (n=479) blood-isolates were ampicillin resistant. No VRE were identified (EARSS 2001). Isolation and clinical identification of VRE (E. faecium and E. faecalis) were made notifiable under the Swedish Communicable Diseases Act in 2000 and during the following two years, not more than 20 isolates were reported each year. VRE were rare in the other Scandinavian countries as well. Iceland reported no cases and so far only sporadic human VRE cases have been found in Denmark and Norway (personal communication). Finland reported a major VRE outbreak in the Helsinki area during 1996-97 (Suppola 1999).

Epidemiology of ARE

North-America

The first report of high-level beta-lactam resistance in enterococci dates back to the late 1980s when a ß-lactamase producing aminoglycoside resistant E. faecalis strain spread rapidly on an infant-toddler ward (Rhinehart 1990) Some years later, Barbara E. Murray and coworkers suggested clonal spread of a single ß-lactamase producing E. faecalis strain to six hospitals in five US states. They based their suggestion on PFGE analyses of isolates obtained from the hospitals during a seven-year period. The study also indicated the strain to be genetically stable for several years (Murray 1991). Since then, ß-lactamase producing strains have very rarely been reported.

Instead, an increase of ampicillin-resistant E. faecium strains with overproduction of PBPs with low affinity to penicillin (see antimicrobial resistance section) was noted, mainly in the U.S. but also in Europe. Such ARE strains increased sevenfold in a university hospital in Michigan during the years 1986-88. Imipenem use was found to be independently associated with ARE infections in the hospital, also when adjusting for length of hospital stay (Boyce 1992). Risk factors for colonization with ARE were further prospectively studied in a medical unit and an ICU in a veterans affairs medical center in California. Previous treatment with cephalosporins, urinary tract catheterization and being bedridden were significantly associated with ARE colonization in univariate analysis (Chirurgi 1992). Advanced age and clindamycin use were independently associated with acquiring ARE in another study at a teaching hospital in North Carolina and ARE were found to be endemic and of multi-focal origin in this hospital (Sexton 1993). In an outbreak of vancomycin-ampicillin and aminoglycoside resistant E. faecium in an adult oncology unit in New York state 1994, all

7

isolates carried the vanA gene and genetic fingerprinting with PFGE analysis suggested a clonal outbreak. Five patients had VRE septicemia and four of these died. Logistic regression analysis showed antibiotic treatment to be a significant risk factor for infection when infected and non-infected stool carries were compared (Montecalvo 1994). The same year, a Canadian study found risk factors for acquisition of ARE (colonization or infection) to be long duration of hospitalization, prior antimicrobial therapy and combined therapy for at least 7 days (McCarthy 1994). Acquisition of ARE was associated with a higher mortality compared to ASE controls and ARE infection was suggested to be a marker of poor outcome.

Finally, a large prospective US study on the prevalence, incidence and factors associated with colonization with resistant enterococci in a university hospital in North Carolina found prior hospitalization, treatment with more than three antibiotics, empiric use of antibiotics, use of third generation of cephalosporins and the use of enteral feeding tubes all to be factors associated with nosocomial acquisition of ARE (Weinstein 1996).

Europe ARE have been studied to some extent in the Nordic countries, including the papers in this thesis, but there are few studies from other Europen countries.

In Finland, vanA and vanB were incorporated into two endemic ARE clones that spread among patients in the Helsinki area (Suppola 1999).

In Norway, the first outbreak of ARE at a university hospital was described in 2000. Even in this outbreak, a majority of the ARE isolates were clonally related. Four VRE (vanB) isolates were also found and these were closely related to the ARE outbreak clone. Independent risk factors for ARE infection were underlying neurological or gastrointestinal disease, use of antimicrobial agents, use of cephalosporins, and length of hospital stay (Harthug 2000b). Recently another Norwegian study on the carriage rates of ARE and high-level gentamicin resistant enterococci in 10 major hospitals showed the mean prevalence of ARE carriage to be 6.9%, and also showed 48% of the ARE isolates to be clonally related according to PFGE. (Harthug 2002).

These Nordic experiences demonstrate the capability of ARE clones to spread in hospitals and also the risk of incorporation of additional resistance genes into such clones.

8

Epidemiology of VRE

North America The first outbreak of VRE in a British renal unit in the end of the 1980s (Uttley 1989) was soon followed by reports of VRE outbreaks in U.S. hospitals. Initially, VRE was assumed to be a strictly hospital-based problem usually affecting immunocompromised hosts in specialist units. When the proportion of vancomycin resistant enterococci among all U.S. enterococcal isolates between 1989 and 1993 increased more than 20-fold, it became clear that VRE were becoming endemic.

A large number of investigations on risk-factors for VRE colonization and infection in hospitals as well as of clonality and the genetics of VRE were carried out in the US during the 1990s. The overall impression of all these studies was that severely ill patients often with immunosuppression and/or renal impairment, prolonged hospital stay and heavy antibiotic consumption (especially vancomycin, cephalosporins and anti-anaerobic antibiotics) were readily colonized and infected with VRE (Livornese 1992, Frieden 1993, Handwerger 1993, Morris 1995). Clonal spread of VRE in and between hospitals was common but colonization among non-hospitalized persons was not found (Frieden 1993, Morris 1995, Coque 1996) (Silverman 1998).

In contrast to Europe (see below), glycopeptides were never licensed for use for other purposes than in human medicine in the US. On the other hand, the use of vancomycin in US hospitals increased dramatically during the 1990s (Kirst 1998). In 1995, a CDC advisory committee formulated guidelines for the preventing the spread of vancomycin resistance. These included limitations in the use of vancomycin, improved infection-control measures, education efforts and use of surveillance cultures (HICPAC 1995). The guidelines have been successful in controlling VRE in regions but have not hindered the continuous emergence of VRE nation-wide, indicating problems of adherence (Ostrowsky 2001).

In summary, the heavy increase of VRE in US hospitals seems to be due to serious problems with both antibiotic overuse and infection control practices but there were no indications of input of VRE to hospitals from reservoirs in the community.

9

Europe European studies revealed much lower VRE carriage rates in hospitals compared to the US, although a heavy increase of VRE in hospitals was noted in England and Wales lately (Austin and Anderson 1999). European outbreaks of VRE have been sporadic and usually with few serious infections, but mainly of the vanA genotype.

In contrast to the situation in European hospitals, VRE was found to be a common colonizer of farm animals and healthy citizens in central European countries (van den Bogaard 1997) (Aarestrup 1996) (Bates 1994). One small but interesting study found high carriage rates of VRE among meat eating elderly persons in Holland but absence of VRE among elderly vegetarians, supporting the hypothesis of a link between animal food products and VRE carriage among healthy individuals (van den Braak 1997).

It was soon suspected that the widespread use of avoparcin for growth promoting purposes in farm animals might select for VRE among farm animals (Klare 1999). Denmark, where 24 000 kg of avoparcin but only 24 kg of vancomycin was used in 1994, may illustrate the heavy selection pressure among farm animals compared to humans during those years (Wegener 1998). Avoparcin confers cross-resistance to vancomycin of the vanA resistance genotype and this genotype has been dominant among both human and animal VRE isolates in Europe, whereas vanA and vanB have been equally prevalent in US hospitals. Moreover, strains with identical transposons Tn1546 have been found in Europe among both animals and humans (Jensen 1998) (Willems 1999). Thus, there are several strong indications of the spread not only of resistant bacterial strains but also of glycopeptide resistance genes from animals and via the food-chain to humans in European countries.

The above discussed differences in the use of glycopeptides in hospitals and in animal husbandry in the US and Europe have been proposed as an explanation for the different epidemiology of VRE on the two continents. The excessive and inappropriate use of glycopeptides among livestock and the evidence of transmission of antibiotic resistance genes as well as bacteria from animals via meat products to man, was one direct cause for the 1998 EU ban of the use of avoparcin, as well as a number of additional antimicrobial agents, as growth promoters to livestock within the union.

In Sweden, a national feedstuffs act prohibited all administration of antimicrobials (including avoparcin) for growth promoting purposes to livestock in January 1986. Thus, Sweden is one of the rare countries in the world totally lacking this major risk factor for emergence of antimicrobial drug resistance and may be regarded as a “negative control country” in this

10

respect. Previous to the period of the present study, only one human VRE case was reported in Sweden (Melhus and Tjernberg 1996).

Figure 1. Swedish calf with no need for growth promoting antibiotics in the feed.

11

Nosocomial spread

Spread in hospitals In the previously mentioned outbreak of a ß-lactamase E. faecalis strain among infants and staff on an infant-toddler surgical ward, the strain was found in eight of 33 employees and 78 of 1028 children admitted during 1998 (Rhinehart 1990). A nurse carried the resistant strain both in her feces and on her hands and care by this nurse was an independent risk factor for colonization of patients (odds ratio 11.6).

Since then several studies have demonstrated the ability of resistant strains of enterococci to prevail and spread in hospitals. Enterococci are able to colonize not only the gut but also the skin, oral cavity and lungs of hospitalized patients (Beezhold 1997) (Bonten 1996). Long duration of stay in hospitals, stay in units with high proportions of colonized patients, use of electronic thermometers, and diarrhea have all been factors associated with spread of resistant strains (Morris 1995, Brooks 1998) (Bonten 1998b) (Boyce 1994).

All these experiences make obvious the importance of having efficient infection-control routines in hospitals. Use of surveillance, isolation and barrier precautions have been successful in controlling minor outbreaks in non-endemic situations (Boyce 1994, Handwerger 1993). Recently it was demonstrated that VRE could be controlled in a large region with 32 health care facilities, provided that consequent and active infection-control measures including surveillance cultures and isolation of all infected and colonized patients were carried out. In that region, the overall prevalence of VRE decreased from 2.2% to 0.5% within 2 years (Ostrowsky 2001). On the other hand, if smaller outbreaks are not controlled the situation may become complex and even endemic. Bonten and coworkers studied VRE epidemiology in an ICU and introduced the term “colonization pressure” meaning the proportion of colonized patients in a setting during a given period. They found that if the colonization pressure was >50% all other measures had only little impact on the time to acquisition of VRE. In an endemic setting measures such as cohorting of patients, protection of high-risk groups, reduction of the total antibiotic pressure and education of staff

12

may be more important than surveillance and isolation of every patient (Bonten 1998b) (Hayden 2000).

Figure 2 Enterococcus faecium on human skin and hair follicle.

Image copyright Dennis Kunkel Microscopy, Inc. www.denniskunkel.com

Hands, thermometers and environmental surfaces That the hands of health care workers are efficient tools for spreading microorganisms between patients in hospitals is true also for resistant enterococci (Bonilla 1997, Boyce 1995). The impact of hand-washing has not been uniformly positive in US studies, possibly because use of gloves sometimes seems to have replaced hand-washing. Half a minute of wash with 60% alcohol solutions was more effective in eradicating VRE from the hands than was soap and water (Wade 1991). Thus, the Swedish

13

recommendation of hand disinfectant use as the first choice for cleaning of hands without visible contamination, may be a wise one. E. faecium isolates survive for 7 days on counter tops, 24 hours on bedrails, 60 minutes on telephones and 30 minutes on stethoscopes (Noskin 1995). Rectal thermometers and even blood pressure cuffs have been involved in the spread of VRE (Livornese 1992, Rhinehart 1990, Bonten 1996). One recent study showed the environment to be contaminated around VRE colonized patients, another that the hands of medical staff may be equally easy contaminated with VRE by environmental contacts and colonized patients (Donskey 2000a) (Bonten 2001). These findings illustrate the importance of proper cleaning of devices shared between patients and the use of hand disinfectants both before and after patient contact when trying to control spread of resistant bacteria in hospitals.

14

Antimicrobial resistance in enterococci

Enterococci are intrinsically resistant to many antibiotics, including aminoglycosides, beta-lactam antibiotics (high-level to cephalosporins), clindamycin and in vivo to trimetoprim-sulphamethoxazole (Murray 1990). The following presents mechanisms of resistance to betalactams, vancomycin, fluoroquinolones and the aminoglycosides only as they apply to enterococci.

Betalactams.

Action of beta-lactamantibiotics and intrinsic resistance Gram-positive bacteria have a thick cross-linked outer membrane peptidoglycan layer. Beta-lactam antibiotics act by inhibiting cell wall synthesis by binding to penicillin binding proteins (PBPs) involved in the synthesis of bacterial peptidoglycan. All enterococci are naturally high-level resistant to cephalosporins, the semisyntetic penicillinase resistant penicillins and low-level resistant to other penicillins due to bacterial production of a low-affinity penicillin- binding protein (PBP 5) (Fontana 1983). In addition, enterococci also have the ability to achieve tolerance to the bactericidal effect of penicillins and vancomycin (Murray 1990). Ampicillin is the drug of choice for enterococcal infections and ampicillin MICs for E. faecalis generally are 1-4 mg/L and for E. faecium 4-8 mg/L.

Acquired resistance Beta-lactamase production is the first mechanism of acquired resistance and is only described in E. faecalis. It was first recovered from a patient in 1981 in Houston, Texas (Murray and Mederski-Samaroj 1983). Strains producing beta-lactamase have since been found only on the American continent. Recently a survey found no such E. faecalis in Swedish ICUs (Hallgren 2001). Beta-lactamases hydrolyze the beta-lactam ring and thereby inactivate the drug. The enterococcal penicillinase gene is identical to the gene

15

encoding staphylococcal type A penicillinase and is often found on a transferable plasmid that also encodes high-level resistance to gentamicin. The activity of the beta-lactamase of E. faecalis is reversed by the beta-lactamase inhibitors clavulanate, sulbactam and tazobactam.

The second mechanism of acquired resistance to beta-lactam antibiotics in enterococci is caused by a mutations in chromosomal DNA (pbp5 gene) resulting in overproduction of a modified PBP5 with low affinity to beta-lactam antibiotics (Fontana 1994) (Ligozzi 1996). These strains are often referred to as ARE and the vast majority are E. faecium. A number of point mutations in the penicillin binding region of the pbp5 gene confers different levels of resistance to all beta-lactam antibiotics including imipenem. ARE isolates typically have ampicillin MICs of 8– 64 mg/L, but may have MICs of >128 (and 3rd generation cephalosporin MICs of >10000). This type of beta-lactam resistance is not reversible with the beta-lactam inhibitors mentioned above. Recently, pbp5 has been described as transferable on a plasmid (Raze 1998).

Ampicillin resistance may also be linked to vancomycin resistance of the vanB-type and may be transferred between strains of E. faecium on the mobile element Tn5382. This genetic package was found in unrelated E. faecium strains from several states in the U.S.A. suggesting horizontal dissemination of these genes among enterococci in that region (Carias 1998)

Glycopeptides

Action of the glycopeptides and intrinsic resistance Like beta-lactam antibiotics, glycopeptides are inhibitors of cell wall synthesis, but through a different mechanism which does not interact with the enzymes involved in cell wall synthesis. The glycopeptides are very large hydrophobic molecules that bind to the peptidyl-D-alanyl-D-alanine termini of the peptidoglycan precursors at the cell surface. The mechanism of action is thought to be as simple as steric inhibition of further cell wall synthesis by the presence of these large molecules at the surface of the cytoplasmic membrane alone (Figure 3).

Intrinsic resistance Glycopeptide resistance is due to bacterial synthesis of modified peptidoglycan precursors with reduced affinity for the glycopeptides (Figure Courvalin). Enterococcus gallinarium harbor vanC1 and Enterococcus flavescens and Enterococcus casseliflavus the vanC2 gene and all display the

16

VanC phenotype characterized by intrinsic resistance to vancomycin (MICs, 4-32 mg/L) but susceptibility to teicoplanin (Table 2).

Acquired resistance The biochemical mechanism of regulation and expression of acquired resistance to the glycopeptides in enterococci is the most sophisticated and perfect example of the genetic adaptation of bacteria ever described (Leclercq and Courvalin 1997). There are four phenotypes of acquired resistance to the glycopeptides, VanA, VanB, VanD and VanE (Murray 2000). The relation of these phenotypes to different enterococcal species, MIC levels and transferability of resistance genes is presented in Table 2.

Table 2. glycopeptide MIC levels and presence of transferable resistance in the 5 different phenotypes of glycopeptide-resistant enterococci.

Phenotype and species

Vancomycin MIC (mg/L)

Teicoplanin MIC (mg/L)

Transferable resistance

VAN A: E. faecium E. faecalis E. avium E. durans

64 - >1000

16 – 512

Yes

VAN B: E. faecium E. faecalis

4 - 1024

0.25 – 2

Yes

VAN C: E. gallinarium E. caseliflavus E. flavescens

2 - 32

0.12 - 2

No

VAN D: E. faecium E. faecalis

16 - 64

2 – 4

ND

VAN E: E. faecalis

16

0.5

No

N.D. Not determined. The VanA phenotype is characterized by high-level resistance to both vancomycin and teicoplanin. Resistance is mediated by seven genes on a mobile genetic element Tn1546 (Arthur 1993). This transposon has the

17

ability to direct its own transfer from the chromosome of one enterococcal strain to another (Handwerger and Skoble 1995) The genes encode seven polypeptides (VanR, VanS, VanH, VanA, VanX, VanY and VanZ) that act cooperatively to confer glycopeptide resistance. The two first, VanR and VanS, regulate the expression of resistance genes, the next three, VanH, VanA and VanX, confer resistance to glycopeptides by translation of a modified cell wall precursor ending in D-alanyl-D-lactate (D-Ala-D-Ala-Lac) instead of the normal D-Ala-D-Ala. The last two (VanY and VanZ) are accessory proteins that contribute to resistance by inhibition of the normal pathway for peptidoglycan synthesis. A schematic representation of the pathways for peptidoglycan synthesis in susceptible and resistant Enterococcus species is outlined in Figure 3. Figure 3. Schematic representation of pathways for peptidoglycan synthesis in glycopeptide-susceptible (right) and resistant (left) enterococci. In resistant strains, theVanH dehydrogenase synthesizes D-lactate from pyruvate and VanA ligase catalyzes the formation of D-Ala-D-Lac which is then branched to form a pentadepsipeptide. Vancomycin has a greatly reduced affinity to D-Ala-D-Lac compared to D-Ala-D-Ala. VanX and VanY have important functions in inhibiting the normal pathway for synthesis of D-Ala-D-Ala. The figure is modified from (Leclercq and Courvalin 1997).

A. Susceptible enterococcus

Cytoplasmic

membrane

2 D-alaninePyruvate+ +

D-Ala-D-Ala

Tripeptide

Pentapeptide

D-Ala-D-Lac

Addingenzyme

Addingenzyme

Tripeptide

Pentadepsipeptide

Ligase VanA

VanXVanH

No growth of peptidoglycan layer

VV

VanY

Vancomycin binding site No binding of vancomycin- normalgrowth of the petidoglycan layer

B. Resistant enterococcus

18

The VanB phenotype is characterized by low to high-level resistance (MICs 4 mg/L to >1000 mg/L) to vancomycin but normally retained susceptibility to teicoplanin. A few isolates with resistance also to teicoplanin have been described (Murray 2000). Resistance is induced by vancomycin but not by teicoplanin. The vanB gene cluster is also associated with a mobile genetic element Tn1547. Another such element, Tn5382 was recently found to be inserted immediately downstream of a pbp5 gene explaining the close association between and VanB and ampicillin resistance, discussed earlier. Genetically, the vanA and vanB resistance genes are quite similar, only a single amino acid in VanS differs between the two genotypes. VanD and VanE are the most recently described phenotypes (Perichon 1997) (Patino 2002). They are characterized by low to moderate resistance to vancomycin and low-level resistance to teicoplanin. The genes encoding this type of resistance seem to be located on the chromosome and transfer of these to other enterococci has so far not been demonstrated.

Fluoroquinolones

Action of fluoroquinolones and intrinsic resistance In all Gram-positive bacteria, two proteins, DNA-gyrase and topoisomerase IV, are considered to be the main targets for the fluoroquinolones. DNA-gyrase is a tetrameric enzyme with two subunits, encoded by the gyrA and gyrB genes respectively, that catalyses the negative supercoiling of DNA. Negative supercoils are important for initiation of DNA replication. Topoisomerase IV acts by separating interlocked DNA strands allowing the forming of daughter chromosomes into daughter cells. Topoisomerase IV also has two subunits, encoded by the parC and parE genes respectively (Drlica and Zhao 1997). Different fluoroquinolones have different levels of action against the two enzymes. Topoisomerase IV seems to be more sensitive and is often regarded as the primary target of fluoroquinolones in Gram-positive bacteria (Hooper 2002).

Intrinsic and acquired resistance Two main groups of chromosomal mutations cause two mechanisms of resistance in enterococci. The first causes low-level resistance by reduced drug accumulation either by decreasing the uptake or increasing the efflux of the drug. Endogenous efflux pumps seem to be widespread among wild-type strains of enterococci and might be the explanation for the intrinsic low-level resistance of most enterococci to the fluoroquinolones (Lynch 1997).

19

High-level resistance to fluoroquinolones is due to mutations in regions encoding subunits of DNA gyrase and topoisomerase IV (gyrA, gyrB, parC and parE). Fluoroquinolones interact with complexes of each enzyme in DNA by trapping the complex and hinder further DNA replication (Hooper 2002). This leads to cell death by yet poorly defined mechanisms. In enterococci mutations in gyrA at positions 83 and 87 and parC at position 80 are more extensively studied than mutations in gyrB and parE (Brisse 1999) (el Amin 1999). Ciprofloxacin resistance seems to be more widespread in E. Faecium, maybe due to clonal spread of such strains. It should be emphasized that ciprofloxacin-resistance in enterococci also confers cross-resistance to newer quinolones with better Gram-positive activity and that superinfections in patients treated with fluoroquinolones have been reported. (Hooper 2002)

Aminoglycosides.

Action and intrinsic resistance of the aminoglycoside Aminoglycosides bind to the 30S subunit of the bacterial ribosome and interfere with protein synthesis. The spectrum of activity includes mainly Gram-negative but also some Gram-positive bacteria such as S. aureus. Aminoglycosides are bactericidal with increased killing at higher concentrations. The combination of an aminoglycoside with a beta-lactam antibiotic results in synergistic efficacy and has long been the golden standard in enterococcal endocarditis (Wilson 1995).

Low-level resistance Low-level aminoglycoside resistance is an intrinsic characteristic of all enterococci. It is mediated by reduced uptake of the aminoglycoside through the cell membrane resulting in MICs to strepto/kanamycin of 250 and to genta/tobramycin of 8-64 mg/L. In E. faecium, tobramycin, netilmicin, kanamycin and sisomicin are inactivated by a chromosomally encoded enzyme, aminoglycoside- 6´-acetyltransferase (AAC-6´) resulting in MICs of typically 128-500 mg/ml. Gentamicin and streptomycin are not inactivated by AAC-6` (Moellering 1973) and gentamicin is therefore the aminoglycoside of choice in serious infections with E. faecium.

Acquired resistance High level resistance to gentamicin (HLGR) in enterococci (MICs ≥1000 mg/ml) was described in E. faecalis in 1979 and E. faecium in 1988. The

20

gene encoding HLGR was probably the result of the fusion of two genes encoding two aminoglycoside modifying enzymes (AMEs) (aac-6´and aph-2´´) into one gene encoding a very powerful bi-functional AME, the Aac(6´)-Ie (Ferretti 1986) This enzyme complex confers resistance to all clinically useful aminoglycosides (gentamicin, tobramycin, amikacin and netilmicin) and also confers resistance to synergism between cell-wall active agents and the aminoglycoside to which the strain is resistant (Murray 1990). Transposons and plasmids have spread this gene world-wide in both staphylococci and enterococci (Simjee and Gill 1997). Recently E. faecium clusters harboring this gene and the transposon Tn5281 were reported in isolates from several Swedish ICUs (Hällgren 2003). Unfortunately, several additional genes also confering HLGR were recently reported (Chow 2000).

21

Impact of antibiotic use on antimicrobial resistance in enterococci

Selection of resistant strains Antibiotics may theoretically promote intestinal colonization and infection with enterococci, and particularly E. faecium, by several mechanisms. First, most of the commonly used antibiotics in hospitals (cephalosporins fluoroquinolones, extended spectrum penicillins, aminoglycosides) have little or no activity against enterococcal strains, and even less so resistant strains. Moreover, E. faecium generally expresses higher MICs to beta-lactam antibiotics than E. faecalis and therefore has advantages in an environment where these agents are widely used. Second, colonization can be promoted by antibiotic inhibition of other bacteria (such as intestinal anaerobes) that compete with enterococci for colonization niches. As mentioned before, an association between antibiotic use and colonization and infection with resistant enterococci has been supported by a large number of studies over the years. Long duration of antibiotic therapy, use of multiple antibiotics and single use of vancomycin, third-generation cephalosporins, imipenem and antianaerobic antibiotics have all been found to be risk factors. Oral vancomycin and particularly teicoplanin administration strongly selected for VRE in the fecal flora of healthy volunteers (Van der Auwera 1996). Vancomycin use has frequently been pointed out as the most important risk factor for the emergence of VRE in hospitals (HICPAC 1995).

However, to blame the use of one single antibiotic class for the emergence of antimicrobial resistance is probably a simplification In 2001, use of both vancomycin and third generation cephalosporins were reported to be independently associated with increased prevalence of VRE in 126 U.S. intensive care units (Fridkin 2001). Recently, a meta-analysis of U.S. studies, performed before 1996, and reporting an association between vancomycin use and VRE infection and colonization was carried out. When adjusted for publication bias, confounding by length of stay and the selection of wrong control groups, vancomycin was no longer significantly associated

22

with VRE (Carmeli 1999). Studies involving oral vancomycin use were not included in this analysis.

Can a change of antibiotic strategies reduce the selection of resistant enterococci? Most physicians agree on the need to minimize unnecessary antibiotic use in order to avoid the emergence of resistant bacteria. However, in some units in hospitals, such as ICUs and hematology units, there is still a need for extensive use of antibiotics. An interesting question then arises: can a switch from one class of antibiotics to another for major use in a unit influence the risk of acquiring resistant enterococcal strains?

In the U.S., E. faecium strains with high MICs to ampicillin (>128 mg/L) and very high MICs to extended spectrum cephalosporins (>10000 mg/L) but lower to piperacillin (256-1024 mg/L) have emerged recently. During ordinary dosing, the concentrations of ceftriaxone in human bile are < 5000 mg/L but piperacillin reaches >1000 mg/L. Consequently, almost all bacteria are killed during piperacillin treatment, and almost all except E. faecium are also killed during ceftriaxone therapy. Thus, E. faecium would be able to colonize the gut of patients treated with ceftriaxone.

This hypothesis was confirmed in a mouse model where ceftriaxone was found to promote VRE colonization in the mice and piperacillin to protect from such colonization. In another mouse model, the same authors found that antianaerobic antibiotics promoted persistence of VRE colonization, probably by inhibiting the anaerobic flora in the gut competing with VRE for a colonization niche (Donskey 2000b, Donskey 1999). In humans, antianaerobic antibiotics were found to promote high-density VRE colonization in 40 of 42 antianaerobic regimens in 33 VRE colonized patients (Donskey 2000a) and was also found to be the only significiant for VRE bacteremia in a Baltimore hospital (Lucas 1998).

The effect of piperacillin was recently investigated in a hematology unit in London. A significant decrease of VRE colonization rates was noted (from 29% to 8%) among patients after a switch of recommended first-line therapy from ceftazidime to piperaillin-tazobactam. VRE colonization rates increased significantly after a switch back to ceftazidime. Rigorous infection-control measures were kept unchanged during the investigation (Bradley 1999).

In a Veterans Affairs medical Center in Brooklyn, NY, VRE colonization rates increased steadily to reach almost 50% of all hospitalized patients in 1995, despite implementation of infection-control practices as recommended (HICPAC 1995). In an attempt to control this situation, the use of cephalosporins and clindamycin was restricted and use of betalactamase

23

inhibiting antibiotics was recommended to replace third-generation cephalosporins. After 6 months, VRE colonization rates had decreased from 47% to 15% (p<.001) and the use of cephalosporins and clindamycin had decreased significantly. Notably, the decline in vancomycin use was modest during this period (Quale 1996).

All the above experiences show that altering antibiotic policies can gain benefits in the struggle against antimicrobial resistance. Studies in this field will clearly be even more important in the future. Unfortunately, such studies must be prospective and will require personnel and economical resources that may be hard to generate in the strained economy of many hospitals today.

Are resistant strains a greater threat to the patients than susceptible ones? The mortality associated with enterococcal bacteremia is generally high (34 to 46%), regardless of antimicrobial susceptibility pattern of the infecting strain. Some studies showed attributable mortality rates in VRE bacteremia of 20-50 % (Edmond 1996) (Bhavnani 2000). Others indicated the higher mortality rates in VRE infections to be associated with the often severe underlying conditions of the patients rather than the susceptibility of the infecting enterococcal strain. One carefully conducted study found no difference in death rates among patients infected with VRE compared to VSE if the patients were controlled for severity of illness and gender (Shay 1995). Another study, using multivariate analysis showed increasing APACHE II score to be the major risk factor for death and the resistant properties of the infecting enterococcal strain to be of minor importance (Lucas 1998). In liver transplant recipients vancomycin resistance was shown be an independent predictor of enterococcus related mortality (Linden 1996). Excess mortality due to ARE infection is not widely studied but in a Norwegian study of an outbreak in a university hospital, the death rate of patients infected with ARE was 18.7% compared to 8.9% for controls, corresponding to an attributable mortality of 9.8% (Harthug 2000b). Thus, resistant enterococcal strains clearly pose a threat to immunocompromised patients but whether an infection with such a strain is more dangerous in non-immunocompromised patients is still under debate.

24

Typing methods.

Systems for typing of microorganisms can be divided into genotypic and phenotypic:

Genotypic methods The rapid development of the field of genetics has been invaluable for our understanding of the molecular epidemiology of bacteria. Genetic methods, such as pulsed-field gel electrophoresis (PFGE), including preparation of chromosomal DNA, cleavage with restriction enzymes and PFGE have been the methods of choice when investigating clonal relationships among enterococci. The guidelines for interpretation of relatedness and clonality (<7 band differences between strains) as proposed by Fred Tenover have been considered the golden standard in such investigations (Maslow 1993) (Tenover 1995). PFGE offers high reproducibility and has a high discriminatory power which is useful when investigating local outbreaks during shorter time periods (e.g. six months) but can be a disadvantage when investigating clonal relationships over longer time periods. One single deletion or insertion of a base pair in the genome of the bacteria can result in three band differences and insertion of a transposon (in vitro) more than 6 band differences in the same strain (Thal 1997). However, investigations of VRE strains in long-term colonized patients over periods up to 160 days have shown strains to be genetically stable, suggesting that large changes in the genome occur infrequently among clinical isolates in nature (Bonten 1998a). Another drawback of PFGE is that it is a labor-intensive method and other methods may be more suitable when typing large numbers of isolates.

To further overcome the shortcomings of PFGE, even more sophisticated genetic typing systems such as amplified fragment length polymorphism (AFLP) and multilocus sequence typing (MLST) have been developed (Willems 2000) (Homan 2002). These are even more labor intensive and expensive methods that include PCR of genomic restriction fragments (AFLP) or direct sequencing of defined sections of house keeping genes of the bacteria (MLST). A power computer then constructs dendrograms based on the PCR and sequence results. The methods are suitable for studies of clonal relations in an evolutionary sense rather than clonal spread in an outbreak situation. Advantages are lack of biased results, easy interpretation

25

and possibilities of exchange of data via the internet (MLST). Recently, an MLST scheme was developed for E. faecium and typing results suggest that epidemic lineages of E. faecium emerged worldwide and that certain such lineages have the ability to persist and colonize patients in hospitals (Homan 2002). These lineages (only VRE) have almost uniformly harbored a variant esp gene, suggested to be associated with colonization and possibly with increased virulence in these bacteria (Willems 2001).

Phenotypic methods These methods are based on the phenotypic expression of genes rather than the sequences of these and are cheaper and often simpler to perform but not as sensitive as the genetic methods. The growing interest in ecological investigations, when often large number of isolates need to be typed, has resulted in a need for faster and cheaper typing techniques for bacteria. The PhenePlateTM RF (PhP-RF) system is a recently developed phenotypic method, based on a 96 well microplate containing 8 sets of eleven dehydrated reagents, selected to have a high discriminatory power among enterococcal isolates. The kinetics of each reaction is evaluated by measuring the absorbance value of each well three times during 64 hours, and a biochemical fingerprint is calculated as the mean value for each reagent over the three readings. The PhP-RF method was shown to be highly reproducible, even when results from different laboratories were compared, and the discriminatory power, measured as Simpson’s diversity index, was as high as 0.96 for all enterococci (Kuhn 1995).

26

Virulence factors in enterococci

Despite the increasing significance of E. faecium in human infection, virulence factors and the genetic determinants encoding such factors remain poorly characterized. Hemolysin, aggregation substance and gelatinase/proteinase are all well established as virulence factors in E. faecalis but have not been found in E. faecium. The enterococcal surface protein (Esp) is another virulence factor in E. faecalis that has been strongly associated with adherence to urinary epithelium in mice and clearly seems associated with colonization of the urinary tract (Shankar 1999, Shankar 2001). Moreover, there is convincing evidence that this large protein is involved in biofilm formation. The presence of Esp in a strain would clearly be an advantage for colonizing patients with indwelling devices (Toledo-Arana 2001). Recently, esp was associated also with epidemic spread of vancomycin-resistant E. faecium in hospitals in the U.S., Australia and Europe (Willems 2001). The relation of this gene to infection with E. faecium has not been studied so far.

27

Aims of this study

The aims were to: • investigate the carriage rates of ARE and VRE in hospitalized patients

and non-hospitalized individuals in Sweden. • determine clonal relationships among ARE strains in Sweden by use of

phenotypic and genetic methods, and to study the relation of mutations in resistant-determining genetic regions to clonality in such strains.

• evaluate risk-factors for colonization and infection with ARE among

hospitalized patients, and the clinical impact of ARE infection. • investigate the molecular epidemiology of ARE in a large hospital. • identify shedders of ARE and risk-factors for such shedding to the

environment. • investigate the presence of a virulence gene, esp in ARE isolates from

infected and colonized patients in a hematology unit and the significance of this gene in colonization and infection with ARE.

28

Materials and Methods

Sample size and origin of enterococcal isolates Enterococcal isolates from patients in 27 hospitals and non-hospitalized individuals in cities all over Sweden were used. The numbers of the collections are referred to further on in the result section. For an overview of isolates and corresponding papers, see Table 3.

Isolates from hospitalized patients and non-hospitalized individuals in Sweden 1. 181 ARE, 9 VRE and 169 ASE isolates from fecal samples of patients at

27 major hospitals in Sweden 1997 (Figure Y). 40 ARE fecal isolates from non-hospitalized individuals in 20 cities nation-wide collected in 1998 (II, IV).

2. Five ARE isolates submitted to SMI in 1990-1992 when an increase of

ARE was first observed in the country. Three of these were clinical isolates from the University Hospital in Linköping, 1991 (LI84, LI134, and LI135) and one (HA1) from Halmstad County Hospital representing the first clonal outbreak of ARE in Sweden 1992. The oldest fecal ARE strain (JO10) was found in 1990 during a screening program of healthy individuals (IV).

3. 29 fecal ASE isolates from healthy individuals obtained during the

above screening program 1990 (IV). 4. Four VRE isolates (and additional surveillance isolates) from 4 patients

and an outbreak of VRE in Örebro during 1996-97 (I). 5. 38 ARE isolates from infectious sites of 38 patients on different wards at

UUH during 1994-95. The ARE isolates from the infectious sites and perineal and environmental cultures (total n=95) were used (III).

29

6. 29 isolates from patients with ARE bacteremia in a hematology unit during 1994-2001 and isolates from colonized sites of patients in this and other units at UUH (n=92) (IV).

Control strains • E. faecalis ATCC 29212, E. coli ATCC 25922 and S. aureus ATCC

29213 were used for antibiotic susceptibility testing and E. faecium BM4147 vanA and E. faecalis V583 vanB, kindly provided by P Courvalin, Institute Pasteur, Paris for control PCR.

• An ampicillin and ciprofloxacin susceptible E. faecium, ATCC 19434T

was used for sequencing (IV). • Two VRE control strains, E0060 (esp negative) and E0300 (esp positive)

were kindly provided by Rob Willems, Bilthoven, Netherlands (V).

30

Table 3. Numbers and sources of enterococcal isolates used and corresponding papers with aims and main methods used.

Paper No.

No of isolates and resistance type

Sites of culture

Source of isolates (collection1)

Aim of study Main methods used

I 4 2 VRE Urine, feces, Wound CAPD fluid

Patients at Örebro County Hospital (4)

Study of a VRE outbreak

Outbreak investigation, PCR, PFGE

II 181 ARE 9 VRE

Feces Patients in 27 major hospitals nation-wide (1)

II 40 ARE Feces Non- hospitalized individuals (1)

Carriage rates and risk factors for VRE and ARE

Point-prevalence and case-control study, PCR, PFGE

III 95 ARE

Clinical and perineal

Patients at UUH (5)

Risk factors for infection and shedding of ARE

Case-control study, PFGE

IV 224 ARE 198 ASE

Fecal and historical clinical and fecal

Same as paper II and SMI isolates from 1990-92 (1-3)

Clonal relationships

Php and PFGE typing, sequencing

V 92 ARE

Blood Fecal/ perineal

Hematological and other patients at UUH (6)

Risk factors for bacteremia /virulence

Case-control study, PFGE, PCR

1 See Materials and Methods 2 In addition to the 4 index VRE isolates, a number of fecal samples were taken from the cases and other patients.

Study designs.

The first hospital outbreak of VRE in Sweden (I) During a 17-week period 1995-96, 3 patients (cases) in a renal unit at Örebro Medical Center Hospital (ÖMCH) and a fourth case in a surgical ward at

31

Lindesberg hospital were found to be infected with VRE. The epidemiology of this VRE outbreak was investigated with molecular fingerprinting and PCR for van and ddl E. faecium genes of the VRE isolates. Rectal swabs were taken from all patients and all personnel in the renal unit. Follow-up fecal samples were taken from all cases until negative for VRE.

Fecal carriage rates of VRE and ARE in hospitals (II) Study A (II): A multi-center nation-wide point-prevalence study on the fecal carriage rates of ARE and VRE among patients on two wards in each of 27 major hospitals in Sweden was performed in 1997 (Figure 4). Data on gender, age, travel outside Scandinavia, previous hospitalizations and antibiotic treatment were recorded (Tables 5 and 6). The study was coordinated from SMI. Nurses and physicians of a National Enterococcal Study Group, created for this purpose, carried out data recording and fecal sampling at each hospital. The members of the study group in each hospital selected two wards, with high and low consumption of antibiotics, respectively. The records of drug deliveries to each ward during 1997 were obtained from hospital pharmacies and the number of patient bed days for each ward from hospital administrations. Each sampled patients record was reviewed. Age, gender, period of hospitalization, prior hospitalizations within 3 and travel outside Scandinavia within 12 months and antibiotic treatment during the previous two weeks were recorded. When 9 VRE isolates were found in a renal unit during the study and since only 3 renal units were included, we considered the risk of selection bias. Therefore 5 additional renal units at 5 other university hospitals were asked to submit fecal samples from admitted patients. No further characteristics of these patients were recorded. Study B (II): To investigate the prevalence of VRE among healthy citizens, a second phase of the study was performed. Individuals visiting 20 infectious diseases outpatient units were asked to submit fecal samples for ARE and VRE cultures. Individuals with previous hospitalization 30 days prior to attending the outpatient clinics were not included in the study. A data set similar to the one of study A, but with some modifications for outpatients was used in study B.

Risk factors for infection and spread of ARE at UUH (III) Between 1991 and 1995 the percentage of ARE among enterococcal isolates in clinical samples from patients at UUH increased from 0.5% to 8.1%. Risk factors for infection with ARE in 38 ARE infected case patients, vs. 38 controls patients infected with ASE were studied prospectively during seven

32

months. All cases and controls were visited on the wards and their records were reviewed for the origin of specimen, age, sex, ward, underlying disease and immunosuppressive chemotherapy. Antimicrobial therapy, any history of surgery or endoscopy, the use of urinary or central line catheters and previous hospital admissions during the preceding 3 months and body temperature, leukocyte counts and CRP levels during one week before the ARE/ASE findings were recorded. To evaluate the physicians´ opinions of the significance of the ARE and ASE culture findings (infection or colonization), a second review of each case and control was performed four weeks after the first one. Any antimicrobial therapy initiated due to the ARE or ASE finding was then registered. Colonization was studied by sampling with a perineal swab from each case. Presence of ARE in this culture was defined as skin carriage. Shedding of ARE from the cases to the near environment was studied using settle plates and VPP pads. Isolates from infectious sites and environmental isolates were subjected to PFGE of genomic DNA (collection 5). Shedding was defined as growth of ≥1 cfu of ARE on the settle plates and/or the bedside table in the room of the case. Cases with growth of >5 cfu of ARE on the settle plates were defined as heavy shedders and the others as low-grade shedders.

Risk-factors for bacteremia with ARE in a hematology unit during seven years (V) For each ARE bacteremia case during the years 1994-2001, two control patients were randomly picked among other patients hospitalized in the unit during the same month as the case until a matched pair of the same gender and age (+\- 5 years) was found. The records of cases and controls were reviewed for diagnoses, treatment with HSCT, colonization with ARE, antibiotic treatment and length of stay. Cases and controls were classified as “colonized” if ARE were isolated in fecal, perianal or urinary samples previous to or at admission, “not colonized” if such cultures had been ARE negative and as “unknown colonization status” if this information was missing. The length of stay in the unit was defined as the number of days from admission to growth of ARE in blood for cases and the total number of days in the unit for controls. Deaths during the stay of cases and controls were also recorded.

Clonality among ampicillin resistant Enterococcus faecium isolates in Sweden and relation to ciprofloxacin resistance (IV) The PhP typing method was used to compare 180 fecal ARE from patients in 27 hospitals, collected in 1998 to 169 fecal ASE isolates from patients in 23 of these hospitals. The 39 fecal ARE isolates from non-hospitalized

33

individuals 1998, and 5 ARE and 29 ASE isolates from the early 1990s were also analyzed (collections 1-3). In addition, representative ARE and ASE isolates were subjected to PFGE analysis of genomic DNA and sequencing of the regions encoding the fluoroquinolone targets of the enzymes GyrA and ParC in order to see how mutations in these regions correlated with the typing results.

Presence of the variant esp gene in ARE isolates collected in a hematology unit (HU) during 7 years and relation of this gene to invasive ARE infection (V) ARE isolates (collection 6) in blood from patients in the HU and other infected and colonized sites from patients in the HU and other units at UUH were analyzed for the presence of the variant esp gene. 40 ARE isolates from outpatients and 11 E. faecium isolates from healthy individuals in Dalarna were also analysed. Blood-isolates (n=29) from all patients with ARE septicemia over a 7 year period were fingerprinted by PFGE. The objectives were to study the prevalence of this gene in ARE isolates, to assess its relation to invasive ARE disease and to investigate the molecular epidemiology of invasive ARE isolates in the HU during 7 years.

Microbiology

Sampling and culture techniques

Fecal and perineal sampling (I, II, III, V) Fecal cutures: Swabs (II) and regular feces (I, V) were used Swabs were inserted in the rectum of study patients, inspected visually for brown color change when withdrawn, and transferred to transport medium. Regular fecal sampled were collected according to standard hospital routines. Perineal sampling was carried out by rubbing a cotton swab firmly to the skin in the perineal area (around the anus) (III, V).

34

Culturing

Paper I For VRE screening, fecal samples were cultured on BBL agar plates and typical colonies growing in the inhibition zone around a 30 µg vancomycin disc were suspected to be VRE.

Paper II

Swabs were put into 1 ml nutrient broth and vigorously shaken for 10s. 50 µL of the resulting suspension was used for ARE and 200 µL for VRE screening. Screening for ARE was performed by direct plating of the suspension on Cephalexin Azthreonam Arabinose (CAA) agar (Ford 1994) containing cephalexin 10 mg/L, azthreonam 75 mg/L, amphotericin B 5 mg/L and ampicillin 30 mg/L. Screening for VRE was performed in two steps. The suspension was first inoculated into 5 ml of Enteroccosel® enrichment broth (BBL, Becton Dickinson Microbiology systems) containing vancomycin 8 mg/L and azthreonam 60 mg/L, incubated at 37°C and examined at 72 h. Tubes indicating growth by even the slightest color change to black were subcultured by inoculating 50 µL on CAA agar with vancomycin 8 mg/L. These plates were incubated for 24 h at 37°C. The accuracy of this screening method to detect VRE was tested with the kind assistance of A E van den Bogaard, University Hospital, Maastricht, The Netherlands, who provided seven blinded human fecal samples. We accurately found five of the samples to contain E. faecium vanA and two to be negative for VRE.

Paper III Perineal swabs were directly plated onto bile-esculin agar plates (Becton Dickinson Microbiology Systems) with 10mg/L ampicillin added (BBL-ampi-plate).

Paper V Fecal samples: 0.5 g of feces was put into 4.5 ml of BHI-broth and 0.1ml was cultured on a BBL-ampi-plate. Perineal samples were cultured according to the methods of paper II.

Blood, urinary and wound samples (III, V) Clinical samples were cultured according to routine laboratory procedures, and standardized according to Swedish laboratory standards. Blood cultures

35

were handled manually until 1994, and were thereafter cultured in a Bact-Alert (Organon teknika gen. Diagnostics AB) machine.

Settle plates and VPP pads (III) Shedding of ARE to the near environment was investigated by two methods: Patient bedside tables were sampled in a standardized manner with a sterile 5x5cm viscose-polyester-polyamide (VPP) pad (Hambraeus 1990). VPP pads were “stomached” for 1 min in a plastic bag with 30 ml 2% polysorbate 80 with 0.3% lecithin in PBS according to a previously described method (Hambraeus 1990). 0.5 ml was further spread onto the surface of a blood agar plate and 0.5 ml onto a BBL-ampi-plate. Viable counts were made after incubation of the plates in 37°C for 24 h. Open settle plates were placed on the floor at a distance of 1m from the patient bed and exposed for 24-h (III). Suspected enterococcal colonies were inspected visually and further inoculated onto a BBL-ampi-plate.

Identification of genus and species.

Genus identification Enterococci were identified by colony morphology, Gram-positive coccoid appearance, ability to hydrolyze esculin and absence of catalase production (I, II, II, IV, V). A positive pyroglutamyl-B-napthylamide (PYR) test was used for some strains (II, IV).

Species identification • Enterococcal isolates (or fecal and perineal samples) were cultured on

CAA agar containing ampicillin 30 mg/L and E. faecium colonies were identified by examining colonies surrounded by a zone of color change from red to yellow indicating fermentation of arabinose (II, IV, V).

• Species identification of all blood isolates (III, V) and some other isolates (I) was performed by the API 20 Strep or rapid ID 32 Strep test kits (Bio Mèrieux, Lyon, France). (III, V).

• If in doubt of species, determination of E. faecium, E. faecalis, E.

gallinarium (I, II, IV) and E. casseliflavus (II, IV) was carried out by PCR of the ddl genes as described by Dutka-Malen Genotype (vanA, vanB and vanC-1) was determined by PCR as described by Clarke (I) and (vanA,

36

vanB, vanC-1 and vanC-2) (II, IV) by Dutka-Malen (Clark 1993, Dutka-Malen 1995).

• Ampicillin susceptible E. Faecium were obtained by picking colonies

from around the inhibition zone of a 10 µg ampicillin disc (>20 mm) placed on a CAA plate (II, IV).

Antimicrobial susceptibility testing Antimicrobial susceptibility testing was done by a standardized disk diffusion method (Olsson-Liljequist 1997) (I, III, V). MICs were determined by agar dilution on PDM Antibiotic Sensitivity Medium, (AB Biodisk, Solna, Sweden) (II, IV) and E-test (Biodisk AB, Solna, Sweden) (I, III, V). Resistance to ampicillin was defined as MIC ≥16 mg/L, to vancomycin as MIC ≥8 mg/L and to ciprofloxacin as MIC ≥4 mg/L according to SRGA guidelines (Kahlmeter 2001). High-level resistance to ciprofloxacin was defined by MIC > 16 mg/L (IV).

Identification of gyrA, parC and esp genes.

Sequencing Sequencing was performed for parts of the fluoroquinolone resistance determining regions gyrA and parC (IV). Genomic DNA was extracted by the CTAB method (Ausubel 1990). A 241 bp fragment of the gyrA gene equivalent to nucleotide positions 150-390 of the E. coli gyrA gene, and a 191 bp fragment of the parC gene equivalent to positions 10- 200 of the E. faecalis parC were amplified as previously described (el Amin 1999). After purification of the PCR products (GFX PCR DNA and Gel Band Purification Kit, Amersham Pharmacia Biotech) both strands of each product were sequenced using the ABI PRISM® BigDyeTM Terminator Cycle Sequencing Ready Reaction kit. PCR primers were used as sequencing primers.

Identification of the esp gene To detect the esp gene in the enterococcal DNA (V), two primers were designed to fit the, yet unpublished, partial sequence of the E. faecium variant esp gene. Two VRE control strains, E0060 (esp negative) and E0300 (esp positive) were used. The primer sequences (5’-TTG CTA ATG CAA

37

GTC CAC GTC C-3’ and 5’-TTT GCA TCA ACA CTT GCA TTA CCG A-3’) were used to amplify fragments of 890 bp by PCR. The primer sequences were the same as used by Willems with a few adjustments to better fit the structure of E. Faecium (Willems 2001). PCR for esp: DNA was isolated from bacteria grown on CAA agar by means of DNeasy Tissue Kit (QIAGEN) according to the manufacturer’s instructions. PCR was performed in 50 µL reaction mixtures using 5 µL of DNA from the preparations, 10 pmol of each primer, 5 µL buffer, 5 µL dNTP:s and 0.2 µL Taq polymerase. After an initial treatment of 94°C for 5 min the samples were subjected to 40 cycles of denaturation (94°C for 30s), annealing, (60°C for 30s) and elongation (72°C for 30s) and after that 7 min at 72°C. For this purpose two PCR-devices were used, Eppendorf Mastercycler gradient (Eppendorf), and GeneAmp PCR System 9700 (PE Applied Biosystems).

Typing

Phenotyping method (IV) The PhP-FS, PhPlate Microplate Techniques, Stockholm, Sweden for biochemical fingerprinting of enterococci was previously described (Kuhn 1995, Möllby 1993). In brief, a loopful of a fresh culture of bacteria was suspended in a sterile growth medium containing 0.11% w/v bromothymol blue. Aliquots (150 µl) of the suspensions were inoculated into 24 wells in the ready-made microtiter plates containing 24 different substrates. The plates were incubated at 37 oC and A620 of each reaction was measured after 16, 40, and 64 hours using a microplate reader. After the final reading, the mean value of all readings was calculated resulting in 24 different numerical values for each tested isolate (the biochemical fingerprint). Isolates yielding PhP patterns with correlation coefficients of ≥0.95 following pairwise comparison were defined as belonging to the same PhP type.

Genotyping method (I, II, III, IV, V). PFGE was used in all papers and standard procedures were followed with minor variations in the different papers (Maslow 1993). Briefly, SmaI digested genomic DNA was electrophoresed in 1% agarose gel in 0.5x TBE at 6 V/cm at 14°C using the CHEF DRII or Mapper XA system (Bio-Rad Laboratories, Hercules, CA). Pulse times were generated by the auto-algorithm mode of the apparatus (I, II, III, V) or increased linearly from 3 to 26.5 s for 14.8 h and 0.5–8.5 s for 6.4 h (IV). Lambda ladders and SmaI digests of Staphylococcus aureus NCTC 8325 or another control strain were

38

run in every gel. Ethidium bromide-stained gels were photographed in UV light. DNA band patterns were analyzed visually (I, II, III, V) or with the help of. the Molecular Analyst Fingerprinting software, version 1.6 (Bio-Rad Laboratories, Hercules, CA) (IV). In visual analysis, band patterns were interpreted as indistinguishable (no band differences), closely related (1-3 band differences), possibly related (4-6 band differences) and unrelated (>6 band differences) as suggested by Tenover (Tenover 1995). Banding patterns with <7 band differences were considered as possibly clonally related and comprising a “strain type”(IV, V). The molecular analyst software created Dendrograms, based on bands between 10 and 360 kb, using UPGMA clustering of a similarity matrix based on band-matching Dice coefficients (IV). Isolates showing indistinguishable or closely-possibly related band patterns (<7 band differences, >80% similarity) were regarded as possibly clonally related.

Statistics To identify risk factors, Mantel-Haenzel odds ratios with corresponding 95% confidence intervals [CI] were calculated using the EPI info version 6 (II, III) and 6.04 (V) software (Dean 1994). For logistic regression analysis, the SPSS releases 8.0 (II) and 11.0 (V) for Windows (SPSS Inc.Chicago, Ill.) were used. The Spearman rank correlation test was used for calculation of correlations between sales of antimicrobials and carriage rates of resistant enterococci in the participating hospital wards (II).

Ehtics The Ethics Committee of the Faculty of Medicine, Uppsala University (paper IV, V) as well as local Ethics Committees throughout Sweden (paper II) approved all study sampling of patients and informed consent was obtained from all patients who contributed fecal samples.

39

Results and discussion

Prevalence of VRE and ARE in Sweden (I, II)

The first hospital outbreak of VRE in Sweden (I)

Patients In November 1995, the first highly vancomycin resistant E. faecium isolates were found in urinary (105 cfu/ml) and fecal samples from an 80-year-old woman, receiving hemodialysis in renal unit at ÖMCH. The same month, VRE was isolated in the feces of a 58-year-old woman in the same unit, who 2 months earlier had received a kidney transplant at UUH. In January 1996, VRE were found in the peritoneal fluid and feces of a third patient in the same unit. Finally, in February 1996, a forth patient with VRE in a post-operative wound was recognized in a nearby hospital. Fecal samples were not taken from his patient.

The two first VRE cases were transferred to the Department of Infectious Diseases, ÖMCH. The two others were kept in single rooms at the units where the VRE samples were taken. No treatment for VRE was given to any case. Repeated fecal samples showed that VRE were carried for several months in the fecal flora of the cases. The time to VRE negative culture for each case is presented in Table 4.

Microbiology All VRE isolates were of the VanA phenotype and showed high-level resistance to vancomycin (MIC 256 mg/L) and teicoplanin (MIC 24-256 mg/L). Except for the fecal VRE strain colonizing case 1 all isolates were also resistant to ampicillin (MIC 24-32 mg/L), low-level resistant to gentamicin (MIC for 12-24 mg/L) but susceptible to trimethoprim-sulfamethoxazole (MIC <1 mg/L). All isolates were genotyped as E. faecium. The VRE strains recovered from case 1, 2, and 4 all harbored the

40

vanA gene. The peritoneal fluid isolate of case 3 was PCR negative for vanA, vanB and vanC genes.

PFGE of the first VRE isolate from each of the four cases resolved about 20 distinct fragments from each strain. Isolates from cases 2 and 4 had identical fragment patterns and the urinary isolate of case 1 differed from these by a single band only. The peritoneal fluid isolate from case 3 showed four band differences compared to isolates 2 and 4.

Surveillance cultures In February 1996 (i.e. 2 months after VRE was first isolated), rectal

swabs were taken from all patients and all personnel in the renal unit. The swabs were cultured on selective media but no VRE were found. No further VRE cases were found in the unit or hospitals during the 10-month follow-up period.

Table 4.. Site of isolation, phenotype, genotype, PFGE-pattern, and months from discovery to negative cultures for VRE of 4 cases of VRE in the Örebro VRE outbreak.

Case Site of isolation

Phenotype Genotype PFGE pattern

Months to negative VRE

culture

1 urine VanA vanA A1 2 1 feces VanA ND ND 5 2 feces VanA vanA A 5 3 CAPD1 fluid VanA - A2 0 3 feces VanA ND ND 2 4 wound VanA vanA A ND

N.D. Not determined 1Continuous Ambulatory Peritoneal Dialysis.

Discussion The molecular analysis clearly indicated possible patient to patient spread and we concluded the 4 Örebro VRE cases to represent the first hospital outbreak of VRE in Sweden. The outbreak was probably limited by the infection-control measures taken. One of the isolates lacked the vanA gene for unknown reasons but still expressed high-level resistance to vancomycin. The outbreak illustrated the necessity of early detection by the microbiology laboratory of VRE in cultures from hospitalized patients. In contrast to vanB, the vanA phenotype expresses high MICs to vancomycin, making this type of resistance easier to detect. The screening method used during this outbreak (BBL agar with a 30

41

µg vancomycin disc) would probably not have detected VRE of the vanB resistance phenotype. It was not possible to identify the source of VRE. Acquisition from other patients with non-recognized VRE colonization was one possibility, but the screening of other patients and personnel in the renal unit was negative. However, highly sensitive cultures such as broth cultures were not used and low density colonization may have been undetected. Colonization via intake of meat-products, was another possibility but this was not further investigated.

The first case had one VRE strain in the urine and another, phenotypically different VRE strain in feces, indicating a possible horizontal transfer of vanA. A mobile element (Tn1546) carrying the vanA gene and able to direct its own transfer from the chromosome of one E. faecium strain to another has been described (Handwerger and Skoble 1995). Such horizontal transfer is considered an important mechanism for the spread of vancomycin resistance (Leclercq and Courvalin 1997) (Kim 1999).

Knowledge on the prevalence of colonization with VRE among other patients in hospitals as well as healthy individuals in Sweden was lacking at this time. A “worst case scenario” would mean that highly vancomycin-resistant enterococcal strains already were widespread among Swedish hospitalized patients and that these 4 cases only represented the tip of a so far not recognized VRE iceberg in Sweden.

Fecal carriage rates of VRE and ARE in hospitals (II)

Hospitalized patients, study A. The distribution of hospitals included in the study is presented in Figure 4. 841 patients in 27 hospitals were sampled from feces. The patients were hospitalized in ICUs (22%), surgical/orthopedic (9%), internal medicine (15%), renal (7%), infectious disease (5%), geriatric/rehabilitation (38%) and other units (4%). ARE were found in the fecal flora of 181 (median 21% per ward, range 0-56%) patients and 9 patients (1.1%) carried both VRE and vancomycin susceptible ARE. All ARE and VRE isolates were E. faecium and harbored the vanB gene. Characteristics of patients and non-hospitalized individuals are listed in Table 5.

All VRE carrying patients were hospitalized at Umeå University Hospital, 5 in a renal and 4 in a geriatric unit. Seven of the nine isolates were identical or closely related according to PFGE. Three of the geriatric patients were colonized with indistinguishable VRE isolates (outbreak strain) and one with a closely related strain. Three other VRE positive patients in the renal unit had isolates closely related to the outbreak strain and 2 carried VRE with a common pattern different from the outbreak strain. One of the geriatric VRE

42

patients regularly attended the renal unit for dialysis. This patient also had a history of diarrhea and may have contributed to the spread of VRE between the renal and the geriatric units. None of the VRE carriers had signs of VRE infection.

The sampling of patients from the additional 5 renal units (see methods section) resulted in 113 fecal samples from 133 patients. Only one of these patients carried a VRE strain (E. faecium,vanB).

Non-hospitalized individuals, study B ARE were found in the fecal flora of 40 (6%) of the 670 individuals sampled but only one carried VRE (Table 1). One carrier of VRE, an E. faecium vanA strain, was found. This individual had recently returned from travelling in Africa and was included in the study when attending a hospital in the south of Sweden. He was admitted and treated for acute malaria.

Table 5. Characteristics of hospitalized patients and non-hospitalized individuals in the nation-wide point prevalence studies (II) 1997 and 1998. Number of patients (%) ARE a

No ARE

Study A Patients studied

181

660

Male 98 (54) 301 (45) Age, mean 73 67 Travel outside Scandinavia b

4 (2)

46 (7)

Hospitalization c 71 (39) 159 (24) Study B Patients studied 40 630 Male 18 (45) 298 (47) Age, mean 54 48 Travel outside Scandinaviad 7 (18) 247 (39) Hospitalization e 29 (73) 106 (17)

a Ampicillin resistant enterococci. b Previous 6 months. c Previous hospitalization 3 months back. d Previous 12 months e Previous 12 months. This was associated with ARE carriage (p<0.005).

43

Antimicrobial resistance of ARE and ampicillin susceptible E. faecium Except for glycopeptides, most of the ARE isolates from study A and B were resistant to the antibiotics tested (Table 6). ARE E. faecium isolates from study A had much higher resistance rates for ciprofloxacin (92% vs 14%) and erythromycin (80% vs 45%) compared to the ASE isolates.

The 9 VRE isolates were all resistant to ampicillin, ciprofloxacin and netilmicin but susceptible to teicoplanin. All 5 VRE carriers at the renal unit had received several antibiotics during the previous two weeks (vancomycin 5, fluoroquinolones 3, cephalosporins 2, penicillins 2, imipenem 1, and macrolides 1). One patient was treated with ciprofloxacin for 33 days and with vancomycin both orally and I.V. for 7 days before VRE was isolated. All the geriatric patients were also recently treated with antimicrobials (trimethoprim-sulfamethoxazole, cephadroxil, metronidazole or penicillin G). Due to the low number of VRE carriers found, no further analysis of risk factors for VRE colonization was done.

Table 6. Antimicrobial resistance of ampicillin resistant (ARE) and ampicillin susceptible (ASE) Enterococcus faecium isolates. Number (%) strains StudyA

StudyA Study B

Antimicrobial agent ARE (n=181) ASE (n=177) ARE a (n=40) Imipenem 180 (99) 10 (6) 39 (98) Vancomycin 0 (0) 0 (0) 0 (0) Teicoplanin 0 (0 0 (0) 0 (0) Ciprofloxacin 164 (91) 25 (14) 37 (93) Netilmicin 177 (98) 163 (92) 40 (100) Erythromycin 144 (80) 80 (45) 27 (68)

a ASE isolates from Study B were not analysed.

44

Figure 4. Distribution of the hospitals that contributed fecal cultures to the nation-wide point-prevalence study on carriage rates of ARE and VRE among hospitalized patients in late 1997.

45

Discussion The observed prevalence of VRE colonization among the 954 hospitalized patients in Sweden was low (1.0%), and mainly due to a local outbreak of clonally related isolates in a renal and a geriatric unit in the same teaching hospital. Among 670 individuals attending outpatient clinics, the prevalence was even lower.

The clustering in one hospital of 9 VRE vanB carriers illustrates the propencity of VRE to spread silently in hospitals. At the time of the point-prevalence study, VRE infections were still absent in the Umeå hospital. VRE vanB strains are more difficult to detect by the microbiological laboratory, compared to detection of vanA strains also in surveillance cultures due to lower levels of vancomycin resistance. Our screening method, including use of highly selective broth cultures with a vancomycin concentration of 8 mg/L and subculturing on CAA agar with vancomycin 8 mg/L added, worked well and also detected vanB strains. This method also detects strains with the vanC type resistance and laboratory proceedures must be able to distinguish between strains with intrinsic and acquired resistance. The near absence of VRE vanA in our study was in contrast to the findings in other European countries where VRE vanA has been readily isolated from hospitalized patients, and frequently from non-hospitalized persons and meat-products (see epidemiology section). The fact that ARE were carried by some hospitalized patients in larger hospitals was an expected finding but the high overall carriage rate and the presence of ARE in both small and large hospitals nation-wide (21.5%) was not. The almost uniform resistance to fluoroquinolones in ARE was further surprising (Table 6). Fingerprinting of the ARE isolates was carried out in a sunsequent study (IV) and clonal relationships as well as antibiotic susceptibilities among these isolates are discussed further on. Surprisingly, also 6 % of the non-hospitalized individuals carried ARE. However, almost all of these individuals had received antibiotics and 73 % had been hospitalized during the previous year (Table 5) indicating a probable acquisition of ARE during previous hospital stays.

In summary, the absence of VRE vanA in this large nation-wide investigation of individuals in and outside hospitals was in contrast to findings in countries in continental Europe. As described in the epidemiology section, the prevalence of vanA carrying E. faecium strains among animals and healthy humans in central Europe has been related to the widespread use of a glycopeptide (avoparcin) as growth promoter in these countries. A recent study found VRE vanA strains in Danish but not in Swedish retailed chicken (Quednau 1998) and another pointed out the striking difference in VRE carriage rates between pigs from the Netherlands

46

(39%) and Sweden (0%) (Van den Bogaard 1998). Others demonstrated significant declines of VRE isolated from meat-samples and chickens after the ban of avoparcin in Denmark and Italy, respectively (Pantosti 1999, Bager 1999). Since most of the meat consumed in Sweden is of local origin, our study supported the hypothesis that inadequate use of glycopeptides as growth promoters may well have an impact on carrier rates of VRE in a country. In Swedish hospitals, the selective pressure of glycopeptides was clearly lower compared to reports from U.S hospitals. (Kirst 1998). Sales data from our investigation revealed that only 17 wards in the 27 hospitals studied used more than 0.01 DDD/ bed day (median 0.01, range 0.01- 0.50).

We concluded that ARE colonization was surprisingly common among patients in most hospitals in Sweden but that VRE colonization was still rare. The near absence of VRE and absence of the vanA genotype contrasted with the situation in continental Europe and supported the hypothesis that use of glycopeptides both as growth promoters to livestock and in hospitals affect VRE carrier rates among humans. The VRE vanB outbreak illustrated the propensity for silent spread of such strains in a hospital.

Risk factors for colonization and infection with ARE (II, III, IV)

Antibiotics (II, III, V)

Antibiotics and risk of ARE colonization (II) By univariate analysis, antimicrobial therapy for 5 days or longer during the previous two weeks were strongly associated with ARE colonization among the hospitalized patients of the nation-wide point-prevalence study (110/181 vs. 164/645, OR 4.7; CI 3.3-6.7, Table 7). When comparing the patients treated for more than 5 days with each group of antimicrobials to all the other treated and non-treated patients, cephalosporin, fluoroquinolone and penicillin therapy was all significantly associated with ARE carriage (Table 7). Even after adjustment of antimicrobial therapy for length of hospital stay using logistic regression, antimicrobial therapy for at least five days was independently associated with ARE colonization (adjusted [OR] of 3.8; 95% [CI] 2.6–5.4, p<0.001).

47

Table 7. Association between length of hospital stay, antimicrobial therapy and ARE colonization in hospitalized patients using univariate analysis (Study A). Number of patients (%)

AREa No ARE

Odds ratio

95% CI P value

Patients studied b

181

645

Length of hospital stay

1-2 days c 14 141 1.0 0.4 – 2.3 1.0 3-7 days 35 167 2.1† 1.1 – 4.3 0.02 8-14 39 135 2.9† 1.5 –5.9 < 0.001 > 14 days 93 202 4.6† 2.5 –8.9 < 0.001

> 5 days treatment with d any antibiotic e 110 ( 61 ) 164 ( 25 ) 4.7† 3.3 – 6.7 < 0.001 cephalosporins 28 (15 ) 39 ( 6 ) 2.9† 1.7 – 5.0 < 0.001 fluoroquinolones 24 (13 ) 36 ( 5 ) 2.7† 1.5 – 4.7 < 0.001 penicillins 27 (16 ) 35 ( 5 ) 3.1† 1.8 – 5.5 < 0.001 carbapenems 10 ( 6 ) 19 ( 3 ) 2.0 NS 0.9 – 4.6 0.08 macrolides 9 ( 5 ) 5 (<1 ) metronidazole 9 ( 5 ) 14 ( 2 ) glycopeptides 5 ( 6 ) 9 ( 3 )

a Ampicillin resistant enterococci. b Data missing for 15 patients. c 1-2 days was defined to represent basic risk of colonization. d during preceding two weeks. e Significant also in logistic regression analysis (adjusted [OR] 3.8, p<0.001, see text). CI confidence interval, NS non-significant and † significant odds ratio.

Among the non-hospitalized individuals (II, study B), 37 of 40 ARE carriers (93 %) had been treated with antibiotics during the preceding 12 months on at least at one occasion compared to 305 of 630 (48 %) of ARE non-carriers (p<0.001). ARE carriage among the non-hospitalized was also associated with hospitalization during the previous year (p<0.005).

Correlation of sales of antibiotics to single wards with ARE colonization rates (II)

Data on antimicrobial sales expressed as defined daily doses (DDD) and numbers of patient bed days per year was available from 35 of the 53 wards sampled. The median sales of antimicrobials to these wards was 0.35 (range 0.09-1.40) DDD/ bed day during 1997. The prevalence of ARE carriers did not correlate with total antimicrobial sales (p=0.42) or with sales of cephalosporins (p=0.40) or ampicillin and its deviates (p=0.68). In contrast, ARE carriage rates were correlated with fluoroquinolone sales (p=0.09).

48

After exclusion of wards with small patient samples (<10 or <15 patients) and thus, the risk of large random error in ARE carriage rates, this correlation was even stronger (p= 0.04).

Antibiotics and risk of ARE infection (III, V)

The correlation of antibiotic exposure with ARE infection was investigated in two case-control studies. In the first study of ARE infected patients at UUH (III), exposure to more than two antibiotics was statistically associated with ARE infection in 38 ARE infected cases compared to 38 ASE infected controls (odds ratio [OR], 3.3; 95% confidence interval [CI], 1.2 to 9.5). More than four weeks duration of antibiotic therapy was even more strongly associated (OR 6.9; CI 1.8-28.3). Exposure to cephalosporins, imipenem and fluoroquinolones was more common among cases than controls but only cephalosporin exposure was significantly associated with ARE infection (OR 9.1; CI 2.6-33.7).

In the second study (V), the impact of antibiotic exposure during care in a hematology unit of 34 cases with ARE bacteremia and 68 control patients without ARE bacteremia was investigated. By univariate analysis, ARE colonization was the heaviest risk factor for ARE bacteremia [matched Odds Ratio (OR) 6.38], followed by antibiotic treatment for >6 days (OR 4.21) and treatment with HSCT during the stay (OR 3.60). Fluoroquinolones were used twice as often in cases compared to controls but this difference was not statistically significant, nor was exposure to any other class of antibiotics. By logistic regression analysis, antibiotic treatment did not reach statistical significance when analyzed together with HSCT and ARE colonization (Table 8).

Discussion Our findings illustrate the complexity of the epidemiology of resistant enterococci in hospitals and the consequent difficulties in evaluating an association between antibiotic treatment and presence of ARE and VRE in hospitalized patients.

The conclusion in paper II that ARE carriage rates correlate with the sales of fluoroquinolones to the wards must be interpreted with caution. Data on antibiotic sales to the ward do not reflect the true consumption by each patient. It is well recognized that resistant enterococci can be carried for years in the fecal flora of humans (Bonten 1998a, Montecalvo 1995). Thus, the carriage rates in patients on the wards may reflect previous exposure to antibiotics or other risk factors. On the other hand, the almost uniform resistance to ciprofloxacin among the ARE isolates and the increasing use of fluoroquinolones during recent years is intriguing and certainly theoretically indicates a risk of selection of ARE by wide spread use of fuoroquinolones.

49

As pointed out earlier, others have neglected controlling for length of hospital stay when investigating risk factors for VRE and also had difficulties with the selection of adequate control groups (Carmeli 1999). Adjustment for length of hospital stay is important because both the risk of exposure to multiple risk factors and the risk of ARE transmission between patients increases with the duration of stay. The importance of selecting an adequate control group was demonstrated by some studies included in the meta-analysis that compared VRE infected to VSE infected patients when evaluating vancomycin as a risk factor for VRE infection. But since VSE are eradicated by vancomycin, there is little probability of finding VSE infected controls with a history of recent vancomycin therapy.

In paper III, ASE infected patients were used as controls, and long duration of antibiotic therapy, multiple antibiotics and therapy with cephalosporins were associated with ARE infection. However, the ARE cases in this study had much longer duration of stay in the hospital compared to the controls (mean [median] 29 [22] vs 15 [8] days, p=0,002). Thus, a weakness of the study was that adjustment for this difference was not carried out. On the other hand, since neither ARE, nor ASE are eradicated by cephalosporins, the study was not confounded by the selection of a biased control group.

In paper V, randomly selected patients on the same ward and of the same age and gender were used as controls to ARE bacteremia cases. The analysis showed no differences between cases and controls in previous treatment with HSCTs or previous stays in the unit, indicating that cases and controls were comparable in terms of severity of underlying disease. The length of stay was not found to be a confounding factor. By multivariate analysis, antibiotic treatment still was not significantly associated with ARE bacteremia due to a strong association between such treatment and ARE colonization (Table 8). Timmers et al., recently studied an outbreak of VRE in a hematology unit. A case control study (24 cases and 49 controls) showed antibiotic treatment within 1 month and low albumin levels to be the only independent risk factors in a multivariate model. However, a case was defined by both colonization and infection with VRE and only 5 cases had VRE bacteremia. These findings are in line with our own, of antibiotic treatment to be associated with colonization and other factors to be more important for development of bacteremia in the patients.

Other risk-factors for acquiring ARE (II, III, V)

Duration of stay in the hospital The risk of ARE colonization among hospitalized patients in the nation-wide study (II) was linearly correlated with the length of stay in the hospital

50

(Table 7). As mentioned in previous discussions, the ARE infected patients in paper III had significantly longer durations of median hospital stay compared to ASE infected controls. In the study on hematology patients with ARE bacteremia (V), the risk of ARE bacteremia peaked during day 7-14 of stay in the HU (OR 14.3, CI 5.12-294), to decline during day 14-21 (OR 4.8, CI 1.47-34.8) and was no longer significant for longer stays (OR 2.2, CI 0.74-24.6).

Patient related risk factors In the case-control study (III), indwelling central line catheters, corticosteroid therapy and presence of another infection were all factors significantly associated with ARE infection by univariate analysis. Among hematology patients (V), HSCT during the stay and previously recognized colonization with ARE were independently associated with ARE bacteremia (table 8). Table 8. Multivariate analysis of risk factors for ARE septicemia.

Risk factor Unadjusted

OR Adjusted OR* 95 % C.I.

HSCTa during stay 3.60 6.18 1.03 – 37.0 Colonized with ARE 6.38 6.01 1.45 – 25.0 Antibiotics during >6 days 4.21 1.26 0.30 – 5.27

aHematopoietic Stem Cell Transplantation. *Adjusted OR estimated from a multiple conditional logistic regression analysis. All three factors were included in the model.

Discussion One interpretation of the findings that central line catheters and corticosteroid therapy were associated with ARE infection is that these factors might reflect severe underlying disease in the patients. Such patients are more often treated in hospitals and probably also stay for longer periods. The length of stay in the hospital and prior hospitalization have been found to be associated with ARE colonization and infection also by others (Boyce 1992, McCarthy 1994, Weinstein 1996). The decline in risk for ARE septicemia with time of hospitalization after the first two weeks of stay among the hematology patients might be explained by the fact that HSCT was carried out early during hospitalization and that the numbers of patients with neutropenia peaked during the second week of stay.

51

Clinical impact of ARE infection (V) The second review of the records of the ARE infected cases of paper III, performed four weeks after the initial one, revealed that only 20/38 cases and 22/38 controls had antibiotic therapy initiated after their enterococcal culture findings. For 16 cases and 20 controls the choice of antimicrobials indicated that enterococci were regarded as a possible cause of infection. In two of the 18 cases left without treatment, ARE as well as other bacteria were isolated from wounds, 16 cases had ARE in urinary cultures and half of these had a mixed bacterial flora and/or urinary catheter. One case with ARE in a urinary culture was treated for S. aureus septicemia in the ICU and later died. Autopsy revealed abundant growth in post mortem cultures of ARE from the lung, liver and kidneys.

Among hematology patients (V), death during the stay was more common among cases with ARE bacteremia in the HU compared to controls (11/34 vs. 6/68 crude OR 4.94, matched OR 17.0 95% CI 1.89-153). Of the cases who died, 3 had acute and 2 chronic myeloid leukemia, 4 had acute lymphocytic leukemia or lymphoma and the 2 remaining cases had a diagnosis of aplastic anemia. Four of the cases who died had a recently undergone allogenic HSCT.

Discussion The second visit of the ARE infected cases and controls at UUH in study III showed that the majority of ARE culture findings did not initiate antimicrobial therapy. Moreover, the physicians caring for the cases and controls apparently did not consider the finding of ARE in clinical cultures to be a stronger indication for antimicrobial therapy compared to the finding of ASE. This was encouraging, considering the risk otherwise of unnecessary glycopetide therapy.

In contrast, the increased death rates among hematology patients with ARE septicemia (V) was a highly significant finding. It may have been biased by the severity of illness in cases and controls, but the vast majority of both cases and controls were hospitalized in the unit due to leukemia, and with the same severity of illness.

Our findings are in line with the general opinion of enterococci as relatively low virulent organisms that pose a minor threat to most, but may be the cause of serious infections and death in immunocompromised patients (Dougherty 1984) (Bach 2002, Montecalvo 1994, Linden 1996). Thus, there are strong reasons to improve our knowledge of the significance and epidemiology of both ARE and VRE, especially among physicians that care for immunocompromised patients.

52

Shedding and spread of ARE (I, II, III, V)

Shedding of ARE from colonized patients (III) Skin carriage was present in 26 and shedding from 23 cases. Five of the 23 shedders were heavy and 18 low grade shedders of ARE. PFGE analysis showed that 7 of the 18 low-grade shedders had ARE isolates found on settle plates and bedside tables with PFGE patterns unrelated to the ARE isolate at the site of infection or perineum. Thus there were only 11 true low-grade shedders. Five of the 26 perineal skin carriers were heavy shedders, 9 low-grade shedders and 12 non-shedders (Figure 5). Only two non-perineal carriers shed ARE. Thus, 14 of the 26 perineal carriers but only 2 of the 12 non-perienal carriers shed ARE (p=0.03) indicating perineal skin carriage to be a risk- factor for shedding.

Figure 5. Relation between perineal skin carriage of ARE and shedding of ARE to the near environment in 38 hospitalized ARE infected patients (III). Heavy shedding was defined as growth>5 Cfu of ARE on bedside tables or settle plates left open for 24h in the patient’s room.

All heavy shedders had identical or closely related PFGE patterns of ARE strains cultured from their site of infection, perineum, bedside tables and settle plates, revealing heavy shedding of their infecting strains to the near environment. (Figure.6).

Low-grade shedders(n=2)

Non-shedders(n=10)

Non-skin carriers(n=12)

Heavy shedder(n=5)

Low-grade shedder(n=9)

Non-shedder(n=12)

Skin carriers(n=26)

ARE infected patients(n=38)

53

Figure 6. PFGE patterns of smaI digested chromosomal DNA from ARE isolates from 3 heavy shedders in paper III, illustrating spread to the environment and possible spread between patients. PFGE patterns of ARE isolates from urine, perineum, bedside table and settle plate respectively from shedder 1 (lanes 1-4), shedder 2 (lanes 5-8) and shedder 3 (lanes 9-12). Three lines are Lambda ladders (M).

Spread of VRE and ARE between patients in hospitals (I, II, III, V) That transmission of VRE between patients is a major mechanism for new colonization was supported by the results of the molecular typing with PFGE of the VRE isolates in the outbreaks in Örebro (I) and Umeå (II). At the Umeå hospital, nosocomial spread of VRE between patients in the renal and the geriatric units was highly probable.

In paper III, the fingerprinting of ARE isolates from 38 infected patients on different wards at UUH revealed that 15 of these were infected with isolates that were identical or closely related to the infectious isolates of one or several other cases at the hospital. In paper V, PFGE typing of 29 blood isolates from the years 1994-2001 and patients in the HU revealed that all 8 blood isolates during 1994 and 1995,

M 1 2 3 4 5 6 M 7 8 9 10 11 12 M

54

when ARE first emerged in the HU, were of an identical PFGE strain type (A). ARE of a second distinct PFGE strain type (C) infected 8 patients during the years 1997-1999. A third strain type (E) infected 4 patients during 1998-2000. Two patients were infected in 1999 with ARE of a fourth strain type (G), the remaining 7 patients had ARE bacteremia due to unique strain types (B, D, F, H, I, J, K) (Figure 7)

Figure 7. PFGE strain types of 29 blood-isolates of ampicillin-resistant E. faecium from 29 patients in a hematology unit during the years 1994-2001.

Additional fingerprinting of ARE isolates in surveillance cultures from randomly picked HSCT patients hospitalized in the HU during 1994 showed these to be clonally related, thus supporting the notion of the emergence of an ARE clone in the unit during this period (Figure 8, unpublished data).

0

1

2

3

4

5

6

7

8

1994 1995 1996 1997 1998 1999 2000 2001

Year

No

of c

ases

Type A

Type B

Type C

Type D

Type E

Type F

Type G

Type H

Type I

Type J

Type K

55

Figure 8 PFGE patterns of smaI digested chromosomal DNA from ARE in surveillance cultures (fecal and perineal) from 5 HSCT patients (A-E) treated in the hematology unit at UUH during 1994. The isolates were randomly picked out of a collection of saved isolates from this period. Lambda ladders are indicated with M.

Discussion The results provided strong evidence fthat both ARE and VRE strains were readily transmitted between patients in hospitals. Spread of VRE is extensively documented within hospitals (Handwerger 1993, Timmers 2002, D'Agata 2001). As discussed further on, there is no reason to believe that ARE spreads by a different mechanism than VRE. On the contrary, virulence factors such as the enterococcal surface protein might be more important for the ability to spread than the resistance pattern of the enterococcal strain. At UUH, shedders were identified and the analysis of the molecular epidemiology of ARE infected patients clearly showed intrahospital spread to be present. Opportunities for transmission of ARE between the cases in paper III were identified for some cases but not for others. Two cases met regularly in the hemodialysis unit (one was a heavy shedder), other cases were cared for on the same ward during the same time period. However, it is well recognized that ARE infecting strains only represent the tip of the iceberg of colonizing strains and since the colonization status of fellow patients was unknown, it is hard to draw any conclusions about the mechanisms of transmission based on this investigation in paper III. On the HU at UUH (V), there was clearly clonal spread of ARE during the early 1990s. In 1996, the unit was relocated in a new ward with better

M A A A B C B C D D E E C C M

56

isolation facilities. During a prospective study (part of paper V) on new colonization in this new ward during a 3 months period in 1998, all patients admitted to the ward were sampled from feces and perineal skin at admission and then weekly. Of 61 included patients, 16 (26%) were already colonized at admission, 40 (66%) remained free of ARE colonization and 5 (8%) became colonized with ARE during the study period. Four of these five became colonized with ARE strains with PFGE patterns that were identical to ARE of other colonized patients in the ward, clearly indicating patient-to-patient spread of ARE (data not shown).

We calculated the transmission rate during this period by dividing the number of cases who acquired ARE during the stay to the number of admissions of already colonized patients observed in the study (5/16) and the average duration of stay in the unit of the already colonized patients (26.2 days). This gives an overall estimated transmission rate of 10 newly colonized patients per 1000 patient hospital days. This is a lower transmission rate compared to the VRE transmission rate of 17 per 1000 hospital days found in a 40 bed adult oncology unit in New York during an eight month period. In that investigation, only perianal samples were taken and the true colonization rates might have been higher (Montecalvo 1995).

Skin-colonization was a risk factor for shedding of ARE and heavy shedding from such patients was present at UUH. Others have found skin-colonization associated with diarrhea, bacteremia and transmission of VRE to other patients (Yamaguchi 1994, Beezhold 1997). One of the VRE colonized patients during the Umeå outbreak had diarrhea and her VRE isolate shared PFGE pattern with other VRE colonized in that outbreak.

Is there a clonal origin of E. faecium in Sweden?(IV)

PhP typing of a nation-wide collected sample of E. faecium The 180 ARE isolates from patients in 27 hospitals showed a high degree of homogeneity (0.95) and a low diversity (0.68) and were dominated by one PhP type which was named FMSE1 (73% of isolates, Table 9, Figure 9a). Even among the fecal ARE isolates from outpatients (collection 1) a majority belonged to FMSE1 (54%, Table 9). In contrast, the 169 fecal ASE isolates (collection 1) showed a significantly lower homogeneity (P<0.001) and higher diversity (P<0.001) than the ARE isolates, and only 1% belonged to FMSE1 (Figure 9b, Table 9). Similarly, none of the 29 ASE isolates from healthy individuals in 1990 (collection 5) belonged to this PhP-type (Table 9).

We further analyzed ARE isolates from the beginning of the 1990s, when ARE first occurred in Sweden (collection 2) and 4/5 of these and the single

57

fecal ARE found in 1990 among healthy individuals (JO10) also turned out to belong to FMSE1. Table 9 Homogeneity, diversity and clonality according to PhP typing, and ciprofloxacin resistance of E. faecium. Collection

No of isolates

Homogeneitya Diversitya

Isolates belonging to FMSE1 (%)b

Isolates resistant to ciprofloxacin (%)c

1: ARE hospital isolates, 1998

180 0.95 0.68 73 91

2: ARE outpatient isolates, 1998

39 0.86 0.85 54 92

3: ASE hospital isolates, 1998

169 0.68 0.97 1 35

4: ASE healthy individuals, 1990

29 0.75 0.99 0 24

a For definitions, see Materials and Methods. b Designation of the PhP type that dominated among the ARE isolates. c Ciprofloxacin resistance defined as MIC ≥ 4 mg/L.

PFGE typing By PFGE typing, ARE also had a higher homogeneity than ASE isolates as illustrated by the two dendrograms in Figure 8. When using 80% similarity as the cut-off, 3 ARE from 1990-92 (HA1, JO10 and LI135) and 20 of the 22 ARE hospital isolates belonged to the same PFGE type, designated A. In contrast, 16 PFGE types were identified among the 22 ASE hospital isolates studied.

58

Figure 9. Dendrograms derived from UPGMA clustering of PhP typing data for fecal ARE (n=180, figure 9a, left) and 169 ASE E. faecium isolates (n=169 figure 9b, right) obtained from a recent nation-wide prevalence study in Sweden 1998. The arrows indicate the positions of the PhP type isolate FMSE1 among ARE and ASE.

Mutations in gyrA and parC The sequencing of gyrA and parC fragments in 37 E. faecium isolates revealed two main groups among these.

Figure 9b

59

• The first group consisted of 12 ASE and 11 ARE, all susceptible or with low-level resistance to ciprofloxacin (MICs 1-16 mg/L) and lacking mutations in gyrA and parC. None of these ASE, but 6 of the 11 ARE isolates belonged to PhP type FMSE1.

• The second group consisted of 12 ARE isolates with ciprofloxacin MICs ≥ 32 mg/L which all demonstrated the same mutation in parC but either of two mutations in gyrA. 8 of these 12 isolates belonged to FMSE1. One of the two historical ARE isolates from Linköping 1991 (LI134 and LI135; both of the FMSE1 type) lacked mutations in gyrA or parC (ciprofloxacin MIC 2 mg/L) whereas the other isolate had mutations in both gyrA and parC (ciprofloxacin MIC 64 mg/L).

• Finally, two ARE isolates with ciprofloxacin MICs of 1 mg/L, both belonging to FMSE1, had mutations only in parC (ARE 450 in Figure 9).

Discussion We found evidence for clonal spread and also a possible clonal origin of ciprofloxacin and ampicillin-resistant E. faecium in Sweden. The dominating PhP type FMSE1 was more prevalent inside than outside hospitals (73% vs. 54% of isolates). Therefore, it is highly plausible that the FMSE1 clone emerged and spread mainly in hospitals. The spread in hospitals of ARE has been well recognized by others as well as in this thesis. The clonal spread of ARE in a Norwegian university hospital during a two year period was recently described (Harthug 2000a) and in Finland van genes were incorporated into an epidemic ARE clone present in the Helsinki area (Suppola 1999). VRE have been shown to have the potential of clonal spread within and between acute and long-term care facilities through colonized patients(Handwerger 1993) (Chow 1993, Dunne and Wang 1997, Kim 1999, Trick 1999). Willems et al. recently demonstrated common E. faecium VRE clones, harboring the variant esp gene to spread in hospital in Europe, the USA and Australia. Such clones were almost absent among VRE from community and animal sources in these countries (Willems 2001). Further analysis of isolates from this study using MLST confirmed the presence of such clonal spread (Homan 2002).

60

Figure 9. Dendrograms and PFGE patterns of genomic DNA from E. faecium from patients in 22 hospitals and cities in Sweden. Above: 22 ampicillin-resistant (ARE) and below, 22 ampicillin susceptible (ASE) matched to the 22 ARE for ward and time of isolation. Three ARE isolates from 1990-1992 (JO10, LI135 and HA1, in bold) are added in the upper dendrogram. PFGE type A isolates have band patterns giving Dice coefficients > 80% when compared pairwise to the pattern of ARE0784 (representing the dominating PhP type FMSE1, see MoM). Ciprofloxacin MICs are in mg/L. DNA sequencing results are added when available (mutated codons in bold).

ARE0784 A FMSE1 >32ASE0932 A >32ASE1494 4 ASE0074 B 4 ASE0730 B 4 ASE0241 B 4 ASE0975 2 ASE0671 2 ASE1552 1 ASE0426 8 AGC AGT GAG ASE0162 C 0.25ASE0841 C 0.5ASE0582 C 0.5ASE1156 C 1 ASE1097 2 ASE0620 0.5 ASE0309 4 ASE1207 2 ASE0783 D 2 ASE0377 D 4ASE1335 0.5ASE1430 2 ASE0502 2 NCTC8325lambda

ARE0784 A FMSE1 >32 ATC CGT GAGARE1231 A FMSE1 >32ARE0884 A FMSE1 >32ARE1084 A >32ARE1331 A >32ARE0971 A FMSE1 >32ARE0154 A FMSE1 >32ARE0366 A FMSE1 >32 ATC AGT AAGAREHA1 A FMSE1 >32ARE0187 A FMSE1 >32ARE0941 A FMSE1 2 AGC AGT GAG ARE0493 A FMSE1 >32ARE0716 A FMSE1 >32AREJO10 A FMSE1 1 ARE1153 A FMSE1 >32 ATC CGT GAGARE0060 A FMSE1 >32ARELI135 A FMSE1 >32 ATC CGT GAGARE0271 A FMSE1 >32ARE1388 A FMSE1 >32 ATC AGT AAGARE0622 A >32 ATC AGA GAGARE0302 A FMSE1 >32ARE1489 A 32ARE0450 A FMSE1 1 ATC AGT GAG ARE0746 FMSE1 >32 ATC CGT GAGARE1531 2 AGC AGT GAGNCTC8325lambda

10090807060504030 Isolate PFGE PhP Cipro MIC parC 80 gyrA 83 gyrA 87

61

Interestingly, vanA, VRE isolates of the same FMSE1 type as those found dominating in Swedish hospitals, were recently found in sewage treatment plants in Swedish cities (Iversen 2002). Sewage from local hospitals was treated in these plants. Since VRE vanA were very rarely found in hospitalized patients so far, a reasonable explanation for this finding is that vanA genes were incorporated into the FMSE1 isolates during the passage through the sewage system. Still, this finding demonstrates the propensity for incorporation of additional resistance genes by the FMSE1 clone.

PFGE of genomic DNA is considered the gold standard for epidemiological typing when investigating local outbreaks. However, it is a labor intensive method and very sensitive even for minor alterations in bacterial DNA. It has been demonstrated that one single insertion of a transposon into an E. faecalis strain changed the PFGE pattern of that strain by 8 bands (Thal 1997). Other methods, less sensitive to changes in variable regions of genes may therefore be more suitable when investigating clonal relationships among large number of isolates representing longer periods of time (e.g. >6 months). In contrast to PFGE, the PhP-FS typing system has the advantage of being easy and inexpensive to perform. It has been shown to offer high reproducibility (98%), and according to recent data this method is suitable for typing of enterococci and correlates well with PFGE typing even when investigating local outbreaks (Kuhn 1995, Lund 2002).

High-level ciprofloxacin resistance (MICs>16 mg/L) was found only in ARE and never in ASE isolates. Sequencing of parts of gyrA and parC showed mutations only in ARE (n=12) and never in ASE (n=12) isolates. In our large collection of ASE, we deliberately chose the ASE isolates with the highest MICs of all to be included for sequencing, and still there were no mutations found. ARE isolates with ciprofloxacin MICs of <16 mg/L also generally lacked mutations. Mutations were found both among ARE FMSE1 isolates and other ARE isolates. Interestingly, gyrA and parC mutations were found in only one of two very similar FMSE1 clinical isolates from Linköping 1991, indicating the possible emergence of ciprofloxacin resistance in ARE in the beginning of the 1990s.

Others have found highly ciprofloxacin resistant E. faecium isolates with mutation in parC80 only (Brisse 1999). Topoisomerase IV has been suggested the primary target for fluoroquinolones in enterococci as well as other Gram positive bacteria. Our finding of two ciprofloxacin susceptible (MICs 1 mg/L) isolates with alterations in parC80 indicates that additional mechanisms are necessary for the expression of fluoroquinolone resistance, in E. faecium.

62

Presence and relation to invasiveness of esp in E. faecium (V). In ARE blood isolates from the period 1994-2001, a minority (11/29) were esp positive. Moreover, 8 of these 11 esp positive isolates were of the same PFGE strain type and from patients with bacteremia during 1994-95 (all patients with ARE bacteremia during those years (Table 10). Table 10. Presence of the esp gene in ampicillin-resistant E. faecium isolates from infected and non-infected patients hospitalized in a hematology unit (HU) and other units (OU) at UUHduring the years 1994-2001. The presence of esp in some additional isolates (results not included in paper V) is presented in the lower part of the table.

Source of isolates (collection)1

Year

No of isolates

Isolate from infected site

No of esp positive

isolates (%) Blood HUnew (6) 1994-95 8 Yes 8 (100) Wound OU (5) 1994-95 3 Yes 3 (100) Urine OU (5) 1994-95 12 Yes 11 (92) Surveillance HUold 2 (6) 1994-95 20 No 16 (80) Surveillance HUnew 3 (6) 1998 21 No 13 (62) Surveillance HUnew 4 (6) 2001 7 No 2 (29) Blood HU (6) 1996-2001 21 Yes 3 (14) All isolates in paper V

1994-2001

92 -

56 (60)

Fecal isolates from outpatients (2)

1998

40

No

34 (85)

Fecal isolates from healthy subjects5

1997

11

No

0 (0)

1 See MoM. 2Perianal (n=7), faecal (n=11) and urinary (n=2) isolates obtained during routine surveillance of patients. 3Perianal and faecal isolates obtained from patients during a three months prospective study. 4Perianal isolates obtained during a one-day point-prevalence study. 5Fecal isolates from healthy subjects in Dalarna county with no hospitalization during the previous year. The remaining 3 esp positive blood isolates were from patients with ARE bacteremia during 1999 and two of these were of the same PFGE strain type. There were no esp positive blood-isolates from the other years studied. In colonized patients, esp was frequently found in isolates obtained during 1994-95, when 17/20 of surveillance isolates from 20 patients on the HWold harbored this gene. Six of these patients also had bacteremia with esp

63

positive ARE. In addition, esp was almost uniformly present in the 15 clinical urine and wound isolates from patients on other wards in the hospital during 1994-1995. From 1996 and onward, esp was rarely present in blood isolates, (3/21) but still frequently found in surveillance isolates in 1998 and less so in 2001 (Table 10). We further analyzed the 40 ARE E. faecium isolates from non-hospitalized individuals (collection 1) for the presence of esp and found this gene to be present in 84% of these isolates. However, in fecal E. faecium isolates from healthy individuals in Dalarna with no hospital stays during the previous year, none (0%) harbored esp.

Discussion The variant esp gene was wide-spread among vancomycin susceptible but ampicillin resistant E. faecium isolates in the HU during the 7 year period studied. The gene was found in 71% (45/63) of ARE isolates from colonized sites, urine and wounds and 38% (11/29) of blood isolates (p=0.002).

An esp positive clone probably an emerged in the HU during 1994-1995. During that period, esp was present in almost all ARE isolates from patients in the unit, regardless of cultured body site. In the late 1990s, esp was very rare in blood isolates and not as uniformly found in ARE isolates from colonized patients as earlier. There were limitations of the present study. Before 1995, surveillance cultures were routinely taken from HSCT patients, but not from other patients on the ward. Other surveillance ARE isolates we had available for study (from a prospective 3 month study in 1998 and the one day point-prevalence study in 2001) probably represented only a fraction of all ARE in colonized patients on the ward during the seven years studied. Still, some interesting and maybe worrying conclusions can be drawn. Colonization with esp positive ARE did not increase the risk of bacteremia. The gene was frequently found in surveillance cultures and almost uniformly present in urinary ARE isolates. Our results support the suggestion of Willems et al.that certain lineages of E. faecium are present in hospitals. These may well have certain adherence and biofilm formation properties making them better suited for survival and spread in hospitals. The high rate of esp among ARE from the non-hospitalized individuals (85%) further strengthens the previously mentioned conclusion that these isolates were picked up by the non-hospitalized individuals during previous hospital stays. Further studies of the role of esp among enterococci in hospitalized patients would be of great interest.

64

General conclusions

The purpose of this thesis was to study aspects of enterococci with acquired resistance to antibiotics. When considering all aspects, some of which have been studied in detail, several major conclusions can be drawn:

There was spread and high prevalence in Swedish hospitals of clonally related ARE isolates. This clonal group may well be related to similar vancomycin resistant clones in other countries, although not yet proven. The ARE in this clonal group were almost uniformly E. faecium and resistant also to fluoroquinolones. Mutations encoding high-level fluoroquinolone resistance were present only in ARE islolates and never in matched ASE isolates.

VRE were still rare in the fecal flora of hospitalized patients in Sweden during the period studied, although one silent outbreak of VRE vanB was detected. The reason for the overall absence of VRE vanA strains in Sweden may partly be the result of a wise political decision in 1986 to ban all antibiotic use for growth promoting purposes in livestock. It may also be associated with the absence of MRSA in Sweden. This absence is probably one major reason for the low vancomycin use in Swedish hospitals compared to other countries.

The spread of ARE in hospitals was facilitated by the use of antibiotics, especially fluoroquinolones and cephalosporins. Long duration of stay in hospitals increases the risk of being colonized with ARE. Skin colonization is a risk factor for shedding of ARE to the environment.

Due to relatively low virulence, enterococci are not a major threat to the average hospitalized patient. In contrast, ARE were shown to be able to cause severe septicemia and death in patients with immunosuppression. Risk factors for septicemia were HSCT and colonization with ARE. Colonization with ARE among hematology patients was in turn associated with prior exposure to antibiotics.

65

The virulence gene esp, that was recently reported to be associated with spread of VRE in hospitals in several countries was frequently found among ARE isolates at UUH. The presence of esp did not increase the risk of invasive ARE disease. In parallell with VRE, the presence of this gene in ARE might well be associated with spread and persistance of such isolates in a hospital.

66

Final comment

The strategy in a nation for, if possible, arrest of the emergence of further resistance in enterococci was recently addressed (Ridwan 2002). The measures to be taken depend on the epidemiology of ARE and VRE in each country. The Nordic countries have much in common concerning the prevalence of VRE. The contributors in a recent Nordic meeting on enterococci attended by researchers from both the veterinarian and human medicine agreed on a “search and contain” strategy for future problems with VRE in the Nordic countries. Experiences of others tell us that a successful struggle against VRE include the widespread and consequent use of surveillance culturing, a lot of educational efforts and uniform compliance to hospital-control recommendations (Ostrowsky 2001).

During 2002, two distinct meticillin-resistant Staphylococcus aureus (MRSA) strains, also resistant to high concentrations of vancomycin were isolated in two different states in the U.S. Both strains had acquired resistance due to in vivo transfer of the vanA operon. It is highly suspected that VRE strains provided these genes. This event is by many considered to be the most important in bacterial antibiotic resistance in 20 years. Examples from the past have shown that when a new resistance phenotype is detected it is because it has already been largely spread in nature. (Gonzalez-Zorn and Courvalin 2003).

It seems, we have everything to gain and a lot to loose in the struggle against antimicrobial resistance.

67

Acknowledgements

First, I want to give my sincere thanks to all patients who agreed to be included in the different studies and made all this possible. Second, I wish to express my gratitude to all people in Uppsala and Stockholm who contributed to this thesis. Special thanks to the following:

My tutors, Anna Hambraeus, Otto Cars and Lars G Burman. You taught me the essence of research. I learned to never take anything for granted and to believe in my abilities and thoughts. I have had a very interesting trip together with you three, very different kinds of tutors-You have my warm and sincere gratitude.

Ulla Zettersten, Margareta Edvall and Angela Lagerquist-Widh at the clinical microbiology laboratory in Uppsala. Angela, you showed me this peculiar bacterium for the first time. Ulla Zettersten, without you it would have been nothing, I wish you a lot of rabbits.

All co-writers in the papers. Without your efforts this thesis would have been impossible. A very special thanks to Barbro Olsson-Liljequist, Britt-Marie Hoffman, Sara Haeggman and Inger Kühn, at SMI and KI. I can still miss that special feeling in the old SMI house in Solna. Professor Göran Friman for allocating some research time to me. A warm thanks to Ulrika Ransjö, who let me be the laziest hospital infection control doctor ever during the last months of writing, without this understanding I would not present this thesis today.. My friends Pelle, Erland, Bernice, Hasse, Eva and Örjan who provided stool samples to a crazy researcher during a skiing holiday in Dalarna. Your samples were used to test our screening methods in paper II. If you can’t take shit from your friends, they are no real friends. I hope always to be yours. The following members of the enterococcal study group who willingly cooperated in study II by collecting fecal samples from patients all over Sweden are warmly thanked:

68

Hannover J., Johansson K. (Boden), Johnsson L., Palme´ A., Lundkvist A. (Borås), Sundin C. G., Pettersson B., Montelius S. (Eskilstuna), Hjortzberg-Nordlund C. Lore´ B., Vig I. (Falun), Gnarpe H., Isaksson M., Lekås G. (Gävle), Larsson L, Ek E. (Göteborg), Karell A.C., Andersson G. (Halmstad), Rydberg J., Möller I.B., Johansson B. (Helsingborg), Söre´n L., Lindberg P., Aagesen J. (Jönköping), Eliasson I., Bernhoff L., Lind H. (Kalmar), Blomberg I., Johansson G. (Karlskrona), Kjerstadius T., Fridh G., Romhed L. (Karlstad), Hansson C. (Kristianstad), Isaksson B., Jonasson J., Antonsson Schütz A., Abednazari H. (Linköping), Schalen C., Pettersson A.C. (Lund), Walder M., Svensson K., Svantesson B. (Malmö), Claesson B., Johansson C. (Skövde), Olofsson S., Telander B. (Stockhom Huddinge Hospital), Ransjö U., Rylander M. (Stockhom Karolinska Hospital), Sörberg M., Jörbäck H. (Stockholm Danderyd Hospital), Dornbusch K., Broman E., Hovmöller S. (Sundsvall), Wiksten E., Johansson I., Andersson U.G. (Uddevalla), Lundholm R., Dahlberg A., Wiström J. (Umeå), Hambraeus A., Edvall M. (Uppsala), Ekelöv-Andström E., Hage G. (Visby), Johansson A., Gärd B., Höfer M. (Västerås), Kahlmeter G., Nilsson L. (Växjö), Törnkvist E., Hjerpe-Åhman A., Källman J. (Örebro)

Rob Willems, Bilthoven, Netherlands, who generously sent us the yet unpublished esp sequence and two VRE control strains, E0060 (esp negative) and E0300 (esp positive) is warmly thanked. I hope for future cooperation.

Dennis Kunkel (image copyright Dennis Kunkel Microscopy, Inc. www.denniskunkel.com) who gave permission to use his beautiful photos of enterococci for free.

My struggeling colleagues at the section for infectious diseases in Uppsala.

Bengt Simonsson for valuable criticism of paper V.

My dear mother and father and my sisters Hillevi and Ulrika. Thank you for all support and being the caring relatives you really are.

Marie, Petter, Julia and Meeri. I love you more than ever. At last, “bacillboken” is completed and I will leave the computer, at least for a while. I hope we´ll have a long hot and free summer together.

69

References

Aarestrup, F. M., Ahrens, P., Madsen, M., Pallesen, L. V., Poulsen, R. L. and Westh,

H. (1996) Glycopeptide susceptibility among Danish Enterococcus faecium and Enterococcus faecalis isolates of animal and human origin and PCR identification of genes within the VanA cluster. Antimicrob Agents Chemother, 40, 1938-40.

Aarestrup, F. M., Butaye, P. and Witte, W. (2002) In the Enterococci, Vol. 1 (Ed, Gilmore, M. S.) American Society for Microbiology, Washington, DC, pp. 55-99.

Anonymous (2000) National Nosocomial Infections Surveillance (NNIS) System report, data summary from January 1992-April 2000. Am J Infect Control, 28, 429-48.

Arthur, M., Molinas, C., Depardieu, F. and Courvalin, P. (1993) Characterization of Tn1546, a Tn3-related transposon conferring glycopeptide resistance by synthesis of depsipeptide peptidoglycan precursors in Enterococcus faecium BM4147. J Bacteriol, 175, 117-27.

Austin, D. J. and Anderson, R. M. (1999) Transmission dynamics of epidemic methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococci in England and Wales. J Infect Dis, 179, 883-91.

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Smith, J. A. and Seidman, J. G. (1990) Preparation of genomic DNA from bacteria. In Current protocols in molecular microbiology.(Ed, Stuhl, K.) Greene publisher and Wiley interscience., New York, pp. 2.4.1-2.4.5.

Bach, P. B., Malak, S. F., Jurcic, J., Gelfand, S. E., Eagan, J., Little, C. and Sepkowitz, K. A. (2002) Impact of infection by vancomycin-resistant Enterococcus on survival and resource utilization for patients with leukemia. Infect Control Hosp Epidemiol, 23, 471-4.

Bager, F., Aarestrup, F. M., Madsen, M. and Wegener, H. C. (1999) Glycopeptide resistance in Enterococcus faecium from broilers and pigs following discontinued use of avoparcin. Microb Drug Resist, 5, 53-6.

Bates, J., Jordens, J. Z. and Griffiths, D. T. (1994) Farm animals as a putative reservoir for vancomycin-resistant enterococcal infection in man. J Antimicrob Chemother, 34, 507-14.

Beezhold, D. W., Slaughter, S., Hayden, M. K., Matushek, M., Nathan, C., Trenholme, G. M. and Weinstein, R. A. (1997) Skin colonization with vancomycin-resistant enterococci among hospitalized patients with bacteremia. Clin Infect Dis, 24, 704-6.

70

Bhavnani, S. M., Drake, J. A., Forrest, A., Deinhart, J. A., Jones, R. N., Biedenbach, D. J. and Ballow, C. H. (2000) A nationwide, multicenter, case-control study comparing risk factors, treatment, and outcome for vancomycin-resistant and -susceptible enterococcal bacteremia. Diagn Microbiol Infect Dis, 36, 145-58.

Bonilla, H. F., Zervos, M. A., Lyons, M. J., Bradley, S. F., Hedderwick, S. A., Ramsey, M. A., Paul, L. K. and Kauffman, C. A. (1997) Colonization with vancomycin-resistant Enterococcus faecium: comparison of a long-term-care unit with an acute-care hospital. Infect Control Hosp Epidemiol, 18, 333-9.

Bonten, M. J., Hayden, M. K., Nathan, C., Rice, T. W. and Weinstein, R. A. (1998a) Stability of vancomycin-resistant enterococcal genotypes isolated from long-term-colonized patients. J Infect Dis, 177, 378-82.

Bonten, M. J., Hayden, M. K., Nathan, C., van Voorhis, J., Matushek, M., Slaughter, S., Rice, T. and Weinstein, R. A. (1996) Epidemiology of colonisation of patients and environment with vancomycin-resistant enterococci. Lancet, 348, 1615-9.

Bonten, M. J., Slaughter, S., Ambergen, A. W., Hayden, M. K., van Voorhis, J., Nathan, C. and Weinstein, R. A. (1998b) The role of "colonization pressure" in the spread of vancomycin- resistant enterococci: an important infection control variable. Arch Intern Med, 158, 1127-32.

Bonten, M. J. M., Willems, R. and Weinstein, R., A (2001) Vancomycin-resistant enterococci: why are they here, and where do they come from? The Lancet infectious diseases, 1, 314-25.

Bouza, E., San Juan, R., Munoz, P., Voss, A. and Kluytmans, J. (2001) A European perspective on nosocomial urinary tract infections I. Report on the microbiology workload, etiology and antimicrobial susceptibility (ESGNI-003 study). European Study Group on Nosocomial Infections. Clin Microbiol Infect, 7, 523-31.

Boyce, J. M., Mermel, L. A., Zervos, M. J., Rice, L. B., Potter-Bynoe, G., Giorgio, C. and Medeiros, A. A. (1995) Controlling vancomycin-resistant enterococci. Infect Control Hosp Epidemiol, 16, 634-7.

Boyce, J. M., Opal, S. M., Chow, J. W., Zervos, M. J., Potter-Bynoe, G., Sherman, C. B., Romulo, R. L., Fortna, S. and Medeiros, A. A. (1994) Outbreak of multidrug-resistant Enterococcus faecium with transferable vanB class vancomycin resistance. J Clin Microbiol, 32, 1148-53.

Boyce, J. M., Opal, S. M., Potter-Bynoe, G., LaForge, R. G., Zervos, M. J., Furtado, G., Victor, G. and Medeiros, A. A. (1992) Emergence and nosocomial transmission of ampicillin-resistant enterococci. Antimicrob Agents Chemother, 36, 1032-9.

Bradley, S. J., Wilson, A. L., Allen, M. C., Sher, H. A., Goldstone, A. H. and Scott, G. M. (1999) The control of hyperendemic glycopeptide-resistant Enterococcus spp. on a haematology unit by changing antibiotic usage. J Antimicrob Chemother, 43, 261-6.

Brisse, S., Fluit, A. C., Wagner, U., Heisig, P., Milatovic, D., Verhoef, J., Scheuring, S., Kohrer, K. and Schmitz, F. J. (1999) Association of alterations in ParC and GyrA proteins with resistance of clinical isolates of Enterococcus faecium to nine different fluoroquinolones. Antimicrob Agents Chemother, 43, 2513-6.

71

Brooks, S., Khan, A., Stoica, D., Griffith, J., Friedeman, L., Mukherji, R., Hameed, R. and Schupf, N. (1998) Reduction in vancomycin-resistant Enterococcus and Clostridium difficile infections following change to tympanic thermometers. Infect Control Hosp Epidemiol, 19, 333-6.

Carias, L. L., Rudin, S. D., Donskey, C. J. and Rice, L. B. (1998) Genetic linkage and cotransfer of a novel, vanB-containing transposon (Tn5382) and a low-affinity penicillin-binding protein 5 gene in a clinical vancomycin-resistant Enterococcus faecium isolate. J Bacteriol, 180, 4426-34.

Carmeli, Y., Samore, M. H. and Huskins, C. (1999) The association between antecedent vancomycin treatment and hospital-acquired vancomycin-resistant enterococci: a meta-analysis. Arch Intern Med, 159, 2461-8.

Chirurgi, V. A., Oster, S. E., Goldberg, A. A. and McCabe, R. E. (1992) Nosocomial acquisition of beta-lactamase--negative, ampicillin-resistant enterococcus. Arch Intern Med, 152, 1457-61.

Chow, J. W. (2000) Aminoglycoside resistance in enterococci. Clin Infect Dis, 31, 586-9.

Chow, J. W., Kuritza, A., Shlaes, D. M., Green, M., Sahm, D. F. and Zervos, M. J. (1993) Clonal spread of vancomycin-resistant Enterococcus faecium between patients in three hospitals in two states. J Clin Microbiol, 31, 1609-11.

Clark, N. C., Cooksey, R. C., Hill, B. C., Swenson, J. M. and Tenover, F. C. (1993) Characterization of glycopeptide-resistant enterococci from U.S. hospitals. Antimicrob Agents Chemother, 37, 2311-7.

Collins, M. D., Facklam, R. R., Farrow, J. A. and Williamson, R. (1989) Enterococcus raffinosus sp. nov., Enterococcus solitarius sp. nov. and Enterococcus pseudoavium sp. nov. FEMS Microbiol Lett, 48, 283-8.

Collins, M. D., Farrow, J. A. E. and Jones, D. (1986) Enterococcus mundtii sp. nov. Int. J. Syst. Bacteriol., 36, 8-12.

Collins, M. D., Jones, J., Farrow, E., KIllper-Batz, R. and Schleifer, H. (1984) Enterococcus avium nom. rev., comb.nov.; E casseliflavus nom. rev. comb.nov.; E. durans nom. rev., comb. nov. E. gallinarium comb. nov.; and E. malodoratus sp. nov. J.Syst. Bacteriol., 34, 220-3.

Collins, M. D., Rodrigues, U. M., Pigott, N. E. and Facklam, R. R. (1991) Enterococcus dispar sp. nov. a new Enterococcus species from human sources. Lett Appl Microbiol, 12, 95-8.

Coque, T. M., Tomayko, J. F., Ricke, S. C., Okhyusen, P. C. and Murray, B. E. (1996) Vancomycin-resistant enterococci from nosocomial, community, and animal sources in the United States.

Antimicrob Agents Chemother, 40, 2605-9. D'Agata, E. M., Green, W. K., Schulman, G., Li, H., Tang, Y. W. and Schaffner, W.

(2001) Vancomycin-resistant enterococci among chronic hemodialysis patients: a prospective study of acquisition. Clin Infect Dis, 32, 23-9.

72

de Vaux, A., Laguerre, G., Divies, C. and Prevost, H. (1998) Enterococcus asini sp.

nov. isolated from the caecum of donkeys (Equus asinus). Int J Syst Bacteriol, 48 Pt 2, 383-7.

Dean, A. G., Dean, J. A., Columbier, K. A., Brendel, K. A., Smith, D. C., Burton, A. H., Dicker, R. C., Sullivan, K., Fagan, R. F. and Arner, T. g. (1994) Epi Info version 6, a word processing, database, and statistical program for epidemiology on microcomputers. Centers for Disease Control and Prevention, Atlanta.

Dettenkofer, M., Ebner, W., Hans, F. J., Forster, D., Babikir, R., Zentner, J., Pelz, K. and Daschner, F. D. (1999) Nosocomial infections in a neurosurgery intensive care unit. Acta Neurochir, 141, 1303-8.

Devriese, L. A., Ceyssens, K., Rodrigues, U. M. and Collins, M. D. (1990) Enterococcus columbae, a species from pigeon intestines. FEMS Microbiol Lett, 59, 247-51.

Donskey, C. J., Chowdhry, T. K., Hecker, M. T., Hoyen, C. K., Hanrahan, J. A., Hujer, A. M., Hutton-Thomas, R. A., Whalen, C. C., Bonomo, R. A. and Rice, L. B. (2000a) Effect of antibiotic therapy on the density of vancomycin-resistant enterococci in the stool of colonized patients. N Engl J Med, 343, 1925-32.

Donskey, C. J., Hanrahan, J. A., Hutton, R. A. and Rice, L. B. (1999) Effect of parenteral antibiotic administration on persistence of vancomycin-resistant Enterococcus faecium in the mouse gastrointestinal tract. J Infect Dis, 180, 384-90.

Donskey, C. J., Hanrahan, J. A., Hutton, R. A. and Rice, L. B. (2000b) Effect of parenteral antibiotic administration on the establishment of colonization with vancomycin-resistant Enterococcus faecium in the mouse gastrointestinal tract. J Infect Dis, 181, 1830-3.

Dougherty, S. H. (1984) Role of enterococci in abdominal sepsis. Am J Surg, 148, 308-12.

Drlica, K. and Zhao, X. (1997) DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiol Mol Biol Rev, 61, 377-92.

Dunne, W. M. and Wang, W. (1997) Clonal dissemination and colony morphotype variation of vancomycin-resistant Enterococcus faecium isolates in metropolitan Detroit, Michigan. J Clin Microbiol, 35, 388-92.

Dutka-Malen, S., Evers, S. and Courvalin, P. (1995) Detection of glycopeptide resistance genotypes and identification to the species level of clinically relevant enterococci by PCR. J Clin Microbiol, 33, 24-7.

EARSS (2001) Annual report. www.earss.rivm.nl European Antimicrobial Resistance Surveillance System, Bilthoven, Holland. pp. 61-70.

Edmond, M. B., Ober, J. F., Dawson, J. D., Weinbaum, D. L. and Wenzel, R. P. (1996) Vancomycin-resistant enterococcal bacteremia: natural history and attributable mortality. Clin Infect Dis, 23, 1234-9.

el Amin, N. A., Jalal, S. and Wretlind, B. (1999) Alterations in GyrA and ParC associated with fluoroquinolone resistance in Enterococcus faecium. Antimicrob Agents Chemother, 43, 947-9.

Farrow, J. A. E. and Collins, M. D. (1985) Enterococcus hirae, a new species that includes amino acid assay strain NCDO1258 and strains causing growth depression in young chickens. Int. J. Syst. Bacteriol., 35, 73-5.

73

Ferretti, J. J., Gilmore, K. S. and Courvalin, P. (1986) Nucleotide sequence analysis

of the gene specifying the bifunctional 6'-aminoglycoside acetyltransferase 2&quot;-aminoglycoside phosphotransferase enzyme in Streptococcus faecalis and identification and cloning of gene regions specifying the two activities. J Bacteriol, 167, 631-8.

Fontana, R., Aldegheri, M., Ligozzi, M., Lopez, H., Sucari, A. and Satta, G. (1994) Overproduction of a low-affinity penicillin-binding protein and high-level ampicillin resistance in Enterococcus faecium. Antimicrob Agents Chemother, 38, 1980-3.

Fontana, R., Cerini, R., Longoni, P., Grossato, A. and Canepari, P. (1983) Identification of a streptococcal penicillin-binding protein that reacts very slowly with penicillin. J Bacteriology, 155, 1343-50.

Ford, M., Perry, J. D. and Gould, F. K. (1994) Use of cephalexin-aztreonam-arabinose agar for selective isolation of Enterococcus faecium. J Clin Microbiol, 32, 2999-3001.

Fridkin, S. K., Edwards, J. R., Courval, J. M., Hill, H., Tenover, F. C., Lawton, R., Gaynes, R. P. and McGowan, J. E., Jr. (2001) The effect of vancomycin and third-generation cephalosporins on prevalence of vancomycin-resistant enterococci in 126 U.S. adult intensive care units. Ann Intern Med, 135, 175-83.

Frieden, T. R., Munsiff, S. S., Low, D. E., Willey, B. M., Williams, G., Faur, Y., Eisner, W., Warren, S. and Kreiswirth, B. (1993) Emergence of vancomycin-resistant enterococci in New York City. Lancet, 342, 76-9.

Gonzalez-Zorn, B. and Courvalin, P. (2003) Reflection&Reaction. The Lancet Infectious Diseases, 3, 67-9.

Hallgren, A., Abednazari, H., Ekdahl, C., Hanberger, H., Nilsson, M., Samuelsson, A., Svensson, E. and Nilsson, L. E. (2001) Antimicrobial susceptibility patterns of enterococci in intensive care units in Sweden evaluated by different MIC breakpoint systems. J Antimicrob Chemother, 48, 53-62.

Hambraeus, A., Hoborn, J. and Whyte, W. (1990) Skin sampling-validation of a pad method and comparison with commonly used methods. J Hosp Infect, 16, 19-27.

Handwerger, S., Raucher, B., Altarac, D., Monka, J., Marchione, S., Singh, K. V., Murray, B. E., Wolff, J. and Walters, B. (1993) Nosocomial outbreak due to Enterococcus faecium highly resistant to vancomycin, penicillin, and gentamicin. Clin Infect Dis, 16, 750-5.

Handwerger, S. and Skoble, J. (1995) Identification of chromosomal mobile element conferring high-level vancomycin resistance in Enterococcus faecium. Antimicrob Agents Chemother, 39, 2446-53.

Harthug, S., Digranes, A., Hope, O., Kristiansen, B. E., Allum, A. G. and Langeland, N. (2000a) Vancomycin resistance emerging in a clonal outbreak caused by ampicillin-resistant Enterococcus faecium. Clin Microbiol Infect, 6, 19-28.

Harthug, S., Eide, G. E. and Langeland, N. (2000b) Nosocomial outbreak of ampicillin resistant Enterococcus faecium: risk factors for infection and fatal outcome. J Hosp Infect, 45, 135-44.

74

Harthug, S., Jureen, R., Mohn, S. C., Digranes, A., Simonsen, G. S., Sundsfjord, A.

and Langeland, N. (2002) The prevalence of faecal carriage of ampicillin-resistant and high- level gentamicin-resistant enterococci among inpatients at 10 major Norwegian hospitals. J Hosp Infect, 50, 145-54.

Hayden, M. K. (2000) Insights into the epidemiology and control of infection with vancomycin- resistant enterococci. Clin Infect Dis, 31, 1058-65.

HICPAC (1995) Recommendations for preventing the spread of vancomycin resistance. Hospital Infection Control Practices Advisory Committee. Emerg Infect Dis, 1, 66.

Homan, W. L., Tribe, D., Poznanski, S., Li, M., Hogg, G., Spalburg, E., Van Embden, J. D. and Willems, R. J. (2002) Multilocus sequence typing scheme for Enterococcus faecium. J Clin Microbiol, 40, 1963-71.

Hooper, D. C. (2002) Fluoroquinolone resistance among Gram-positive cocci. Lancet Infect Dis, 2, 530-8.

Huycke, M. M., Sahm, D. F. and Gilmore, M. S. (1998) Multiple-drug resistant enterococci: the nature of the problem and an agenda for the future. Emerg Infect Dis, 4, 239-49.

Hällgren, A., Saeedi, B., Nilsson, M., Monstein, H. J., Isaksson, B., Hahnberger, H. and Nilsson, L. E. (2003) Genetic relatedness among Enterococcus faecalis with transposon-mediated high-level gentamicin-resistance in Swedish intensive care units. JAC, In press.

Iwen, P. C., Kelly, D. M., Linder, J., Hinrichs, S. H., Dominguez, E. A., Rupp, M. E. and Patil, K. D. (1997) Change in prevalence and antibiotic resistance of Enterococcus species isolated from blood cultures over an 8-year period. Antimicrob Agents Chemother, 41, 494-5.

Iversen, A., Kuhn, I., Franklin, A. and Mollby, R. (2002) High prevalence of vancomycin-resistant enterococci in Swedish sewage. Appl Environ Microbiol, 68, 2838-42.

Jensen, L. B., Ahrens, P., Dons, L., Jones, R. N., Hammerum, A. M. and Aarestrup, F. M. (1998) Molecular analysis of Tn1546 in Enterococcus faecium isolated from animals and humans. J Clin Microbiol, 36, 437-42.

Kahlmeter, G. (2001) , Vol. 2003 www.srga.org, . Kalina, A. P. (1970) The taxonomy and nomenclature of enterococci.

Int. J. Syst. Bacteriol., 20, 185-9. Kim, W. J., Weinstein, R. A. and Hayden, M. K. (1999) The changing molecular

epidemiology and establishment of endemicity of vancomycin resistance in enterococci at one hospital over a 6-year period. J Infect Dis, 179, 163-71.

Kirst, H. A., Thompson, D. G. and Nicas, T. I. (1998) Historical yearly usage of vancomycin. Antimicrob Agents Chemother, 42, 1303-4.

Klare, I., Badstubner, D., Konstabel, C., Bohme, G., Claus, H. and Witte, W. (1999) Decreased incidence of VanA-type vancomycin-resistant enterococci isolated from poultry meat and from fecal samples of humans in the community after discontinuation of avoparcin usage in animal husbandry. Microb Drug Resist, 5, 45-52.

Kuhn, I., Burman, L. G., Haeggman, S., Tullus, K. and Murray, B. E. (1995) Biochemical fingerprinting compared with ribotyping and pulsed-field gel electrophoresis of DNA for epidemiological typing of enterococci. J Clin Microbiol, 33, 2812-7.

75

Leclercq, R. and Courvalin, P. (1997) Resistance to glycopeptides in enterococci. Clin Infect Dis, 24, 545-54; quiz 55-6.

Ligozzi, M., Pittaluga, F. and Fontana, R. (1996) Modification of penicillin-binding protein 5 associated with high-level ampicillin resistance in Enterococcus faecium. Antimicrob Agents Chemother, 40, 354-7.

Linden, P. K., Pasculle, A. W., Manez, R., Kramer, D. J., Fung, J. J., Pinna, A. D. and Kusne, S. (1996) Differences in outcomes for patients with bacteremia due to vancomycin-resistant Enterococcus faecium or vancomycin-susceptible E. faecium. Clin Infect Dis, 22, 663-70.

Livornese, L. L., Dias, S., Samel, C., Romanowski, B., Taylor, S., May, P., Pitsakis, P., Woods, G., Kaye, D., Levison, M. E. and et al. (1992) Hospital-acquired infection with vancomycin-resistant Enterococcus faecium transmitted by electronic thermometers. Ann Intern Med, 117, 112-6.

Lucas, G. M., Lechtzin, N., Puryear, D. W., Yau, L. L., Flexner, C. W. and Moore, R. D. (1998) Vancomycin-resistant and vancomycin-susceptible enterococcal bacteremia: comparison of clinical features and outcomes. Clin Infect Dis, 26, 1127-33.

Lund, B., Agvald-Ohman, C., Hultberg, A. and Edlund, C. (2002) Frequent Transmission of Enterococcal Strains between Mechanically Ventilated Patients Treated at an Intensive Care Unit. J Clin Microbiol, 40, 2084-8.

Lynch, C., Courvalin, P. and Nikaido, H. (1997) Active efflux of antimicrobial agents in wild-type strains of enterococci. Antimicrob Agents Chemother, 41, 869-71.

Martinez-Murcia, A. J. and Collins, M. D. (1991) Enterococcus sulfureus, a new yellow-pigmented Enterococcus species. FEMS Microbiol Lett, 64, 69-74.

Maslow, J. N., Slutsky, A. M. and Arbeit, R. D. (1993) In Diagnostic molecular microbiology: principles and applications.(Eds, Persing, D. H., T.F., S., Tenover, F. C. and White, T. S.) American Society for Microbiology, Washington, D.C., pp. 563-72.

McCallum, W. G. and Hastngs, T. W. (1899) A case of acute endocarditis caused by Micrococcus zymogenes (Nov. Spec.) with a description of the microorganism. J. Exp. Med., 4, 521-34.

McCarthy, A. E., Victor, G., Ramotar, K. and Toye, B. (1994) Risk factors for acquiring ampicillin-resistant enterococci and clinical outcomes at a Canadian tertiary-care hospital. J Clin Microbiol, 32, 2671-6.

Melhus, A. and Tjernberg, I. (1996) First documented isolation of vancomycin-resistant Enterococcus faecium in Sweden. Scand J Infect Dis, 28, 191-3.

Moellering, R. C., Jr., Wennersten, C. and Weinstein, A. J. (1973) Penicillin-tobramycin synergism against enterococci: a comparison with penicillin and gentamicin. Antimicrob Agents Chemother, 3, 526-9.

Montecalvo, M. A., de Lencastre, H., Carraher, M., Gedris, C., Chung, M., VanHorn, K. and Wormser, G. P. (1995) Natural history of colonization with vancomycin-resistant Enterococcus faecium. Infect Control Hosp Epidemiol, 16, 680-5.

76

Montecalvo, M. A., Horowitz, H., Gedris, C., Carbonaro, C., Tenover, F. C., Issah,

A., Cook, P. and Wormser, G. P. (1994) Outbreak of vancomycin-, ampicillin-, and aminoglycoside-resistant Enterococcus faecium bacteremia in an adult oncology unit. Antimicrob Agents Chemother, 38, 1363-7.

Montecalvo, M. A., Shay, D. K., Patel, P., Tacsa, L., Maloney, S. A., Jarvis, W. R. and Wormser, G. P. (1996) Bloodstream infections with vancomycin-resistant enterococci. Arch Intern Med, 156, 1458-62.

Morris, J. G., Jr., Shay, D. K., Hebden, J. N., McCarter, R. J., Jr., Perdue, B. E., Jarvis, W., Johnson, J. A., Dowling, T. C., Polish, L. B. and Schwalbe, R. S. (1995) Enterococci resistant to multiple antimicrobial agents, including vancomycin. Establishment of endemicity in a university medical center. Ann Intern Med, 123, 250-9.

Murray, B. E. (1990) The life and times of the Enterococcus. Clin Microbiol Rev, 3, 46-65.

Murray, B. E. (2000) Vancomycin-resistant enterococcal infections. N Engl J Med, 342, 710-21.

Murray, B. E. and Mederski-Samaroj, B. (1983) Transferable beta-lactamase. A new mechanism for in vitro penicillin resistance in Streptococcus faecalis. J Clin Invest, 72, 1168-71.

Murray, B. E., Singh, K. V., Markowitz, S. M., Lopardo, H. A., Patterson, J. E., Zervos, M. J., Rubeglio, E., Eliopoulos, G. M., Rice, L. B., Goldstein, F. W. and et al. (1991) Evidence for clonal spread of a single strain of beta-lactamase-producing Enterococcus (Streptococcus) faecalis to six hospitals in five states. J Infect Dis, 163, 780-5.

Möllby, R., Kuhn, I. and Katouli, M. (1993) Computerised biochemical fingerprinting- a new tool for typing bacteria. Reviews in medical Microbiology., 4, 231-41.

Nachman, S. A., Verma, R. and Egnor, M. (1995) Vancomycin-resistant Enterococcus faecium shunt infection in an infant: an antibiotic cure. Microb Drug Resist, 1, 95-6.

Noskin, G. A., Stosor, V., Cooper, I. and Peterson, L. R. (1995) Recovery of vancomycin-resistant enterococci on fingertips and environmental surfaces. Infect Control Hosp Epidemiol, 16, 577-81.

Olsson-Liljequist, B., Larsson, P., Walder, M. and Miorner, H. (1997) Antimicrobial susceptibility testing in Sweden. III. Methodology for susceptibility testing. Scand J Infect Dis Suppl, 105, 13-23.

Ostrowsky, B. E., Trick, W. E., Sohn, A. H., Quirk, S. B., Holt, S., Carson, L. A., Hill, B. C., Arduino, M. J., Kuehnert, M. J. and Jarvis, W. R. (2001) Control of vancomycin-resistant enterococcus in health care facilities in a region. N Engl J Med, 344, 1427-33.

Pantosti, A., Del Grosso, M., Tagliabue, S., Macri, A. and Caprioli, A. (1999) Decrease of vancomycin-resistant enterococci in poultry meat after avoparcin ban. Lancet, 354, 741-2.

Papanicolaou, G. A., Meyers, B. R., Meyers, J., Mendelson, M. H., Lou, W., Emre, S., Sheiner, P. and Miller, C. (1996) Nosocomial infections with vancomycin-resistant Enterococcus faecium in liver transplant recipients: risk factors for acquisition and mortality. Clin Infect Dis, 23, 760-6.

77

Patino, L. A., Courvalin, P. and Perichon, B. (2002) vanE Gene Cluster of Vancomycin-Resistant Enterococcus faecalis BM4405. J Bacteriol, 184, 6457-64.

Perichon, B., Reynolds, P. and Courvalin, P. (1997) VanD-type glycopeptide-resistant Enterococcus faecium BM4339. Antimicrob Agents Chemother, 41, 2016-8.

Pompei, R., Berlutti, F., Thaller, M. C., Ingianni, A., Cortis, G. and Dainelli, B. (1992) Enterococcus flavescens sp. nov., a new species of enterococci of clinical origin. Int J Syst Bacteriol, 42, 365-9.

Quale, J., Landman, D., Atwood, E., Kreiswirth, B., Willey, B. M., Ditore, V., Zaman, M., Patel, K., Saurina, G., Huang, W., Oydna, E. and Burney, S. (1996) Experience with a hospital-wide outbreak of vancomycin-resistant enterococci. Am J Infect Control, 24, 372-9.

Quednau, M., Ahrne, S., Petersson, A. C. and Molin, G. (1998) Antibiotic-resistant strains of Enterococcus isolated from Swedish and Danish retailed chicken and pork. J Appl Microbiol, 84, 1163-70.

Raze, D., Dardenne, O., Hallut, S., Martinez-Bueno, M., Coyette, J. and Ghuysen, J. M. (1998) The gene encoding the low-affinity penicillin-binding protein 3r in Enterococcus hirae S185R is borne on a plasmid carrying other antibiotic resistance determinants. Antimicrob Agents Chemother, 42, 534-9.

Rhinehart, E., Smith, N. E., Wennersten, C., Gorss, E., Freeman, J., Eliopoulos, G. M., Moellering, R. C., Jr. and Goldmann, D. A. (1990) Rapid dissemination of beta-lactamase-producing, aminoglycoside-resistant Enterococcus faecalis among patients and staff on an infant-toddler surgical ward. N Engl J Med, 323, 1814-8.

Richards, M. J., Edwards, J. R., Culver, D. H. and Gaynes, R. P. (2000) Nosocomial infections in combined medical-surgical intensive care units in the United States. Infect Control Hosp Epidemiol, 21, 510-5.

Ridwan, B., Mascini, E., van Der Reijden, N., Verhoef, J. and Bonten, M. (2002) What action should be taken to prevent spread of vancomycin resistant enterococci in European hospitals? Bmj, 324, 666-8.

Rodrigues, U. and Collins, M. D. (1990) Phylogenetic analysis of Streptococcus saccharolyticus based on 16S rRNA sequencing. FEMS Microbiol Lett, 59, 231-4.

Roghmann, M. C., McCarter, R. J., Jr., Brewrink, J., Cross, A. S. and Morris, J. G., Jr. (1997) Clostridium difficile infection is a risk factor for bacteremia due to vancomycin-resistant enterococci (VRE) in VRE-colonized patients with acute leukemia. Clin Infect Dis, 25, 1056-9.

Schliefer, K. H. and Killper-Baltz, R. (1984) Transfer of Streptococcus faecalis and Streptococcus faecium to the genus Enterococcus nom. rev. as Enterococcus faecalis comb. nov.and Enterococcus faecium comb. nov. Int. J. Syst. Bacteriol., 34, 31-4.

Sexton, D. J., Harrell, L. J., Thorpe, J. J., Hunt, D. L. and Reller, L. B. (1993) A case-control study of nosocomial ampicillin-resistant enterococcal infection and colonization at a university hospital. Infect Control Hosp Epidemiol, 14, 629-35.

Shankar, N., Lockatell, C. V., Baghdayan, A. S., Drachenberg, C., Gilmore, M. S. and Johnson, D. E. (2001) Role of Enterococcus faecalis surface protein Esp in the pathogenesis of ascending urinary tract infection. Infect Immun, 69, 4366-72.

78

Shankar, V., Baghdayan, A. S., Huycke, M. M., Lindahl, G. and Gilmore, M. S. (1999) Infection-derived Enterococcus faecalis strains are enriched in esp, a gene encoding a novel surface protein. Infect Immun, 67, 193-200.

Shay, D. K., Maloney, S. A., Montecalvo, M., Banerjee, S., Wormser, G. P., Arduino, M. J., Bland, L. A. and Jarvis, W. R. (1995) Epidemiology and mortality risk of vancomycin-resistant enterococcal bloodstream infections. J Infect Dis, 172, 993-1000.

Silverman, J., Thal, L. A., Perri, M. B., Bostic, G. and Zervos, M. J. (1998) Epidemiologic evaluation of antimicrobial resistance in community-acquired enterococci. J Clin Microbiol, 36, 830-2.

Simjee, S. and Gill, M. J. (1997) Gene transfer, gentamicin resistance and enterococci. J Hosp Infect, 36, 249-59.

Sitges-Serra, A., Lopez, M. J., Girvent, M., Almirall, S. and Sancho, J. J. (2002) Postoperative enterococcal infection after treatment of complicated intra-abdominal sepsis. Br J Surg, 89, 361-7.

Still, J., Law, E., Friedman, B., Fuhrman, S. and Newton, T. (2001) Vancomycin-resistant organisms on a burn unit. South Med J, 94, 810-2.

Suppola, J. P., Kolho, E., Salmenlinna, S., Tarkka, E., Vuopio-Varkila, J. and Vaara, M. (1999) vanA and vanB incorporate into an endemic ampicillin-resistant vancomycin-sensitive Enterococcus faecium strain: effect on interpretation of clonality. J Clin Microbiol, 37, 3934-9.

Svec, P., Devriese, L. A., Sedlacek, I., Baele, M., Vancanneyt, M., Haesebrouck, F., Swings, J. and Doskar, J. (2001) Enterococcus haemoperoxidus sp. nov. and Enterococcus moraviensis sp. nov., isolated from water. Int J Syst Evol Microbiol, 51, 1567-74.

Teixeira, L. M., Carvalho, M. G., Espinola, M. M., Steigerwalt, A. G., Douglas, M. P., Brenner, D. J. and Facklam, R. R. (2001) Enterococcus porcinus sp. nov. and Enterococcus ratti sp. nov., associated with enteric disorders in animals. Int J Syst Evol Microbiol, 51, 1737-43.

Tenover, F. C., Arbeit, R. D., Goering, R. V., Mickelsen, P. A., Murray, B. E., Persing, D. H. and Swaminathan, B. (1995) Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol, 33, 2233-9.

Thal, L. A., Silverman, J., Donabedian, S. and Zervos, M. J. (1997) The effect of Tn916 insertions on contour-clamped homogeneous electrophoresis patterns of Enterococcus faecalis. J Clin Microbiol, 35, 969-72.

Thiercelin, M. (1899) Sur un diplocoque saprophyte de l`intestine susceptible de devenir pathogene. C.R. Soc. Biol., 5, 269-71.

Timmers, G. J., van der Zwet, W. C., Simoons-Smit, I. M., Savelkoul, P. H., Meester, H. H., Vandenbroucke-Grauls, C. M. and Huijgens, P. C. (2002) Outbreak of vancomycin-resistant Enterococcus faecium in a haematology unit: risk factor assessment and successful control of the epidemic. Br J Haematol, 116, 826-33.

Toledo-Arana, A., Valle, J., Solano, C., Arrizubieta, M. J., Cucarella, C., Lamata, M., Amorena, B., Leiva, J., Penades, J. R. and Lasa, I. (2001) The enterococcal surface protein, Esp, is involved in Enterococcus faecalis biofilm formation. Appl Environ Microbiol, 67, 4538-45.

79

Trick, W. E., Kuehnert, M. J., Quirk, S. B., Arduino, M. J., Aguero, S. M., Carson,

L. A., Hill, B. C., Banerjee, S. N. and Jarvis, W. R. (1999) Regional dissemination of vancomycin-resistant enterococci resulting from interfacility transfer of colonized patients. J Infect Dis, 180, 391-6.

Tyrrell, G. J., Turnbull, L., Teixeira, L. M., Lefebvre, J., Carvalho Mda, G., Facklam, R. R. and Lovgren, M. (2002) Enterococcus gilvus sp. nov. and Enterococcus pallens sp. nov. isolated from human clinical specimens. J Clin Microbiol, 40, 1140-5.

Uttley, A. H., George, R. C., Naidoo, J., Woodford, N., Johnson, A. P., Collins, C. H., Morrison, D., Gilfillan, A. J., Fitch, L. E. and Heptonstall, J. (1989) High-level vancomycin-resistant enterococci causing hospital infections. Epidemiol Infect, 103, 173-81.

Wade, J. J., Desai, N. and Casewell, M. W. (1991) Hygienic hand disinfection for the removal of epidemic vancomycin- resistant Enterococcus faecium and gentamicin-resistant Enterobacter cloacae. J Hosp Infect, 18, 211-8.

van den Bogaard, A. E., Jensen, L. B. and Stobberingh, E. E. (1997) Vancomycin-resistant enterococci in turkeys and farmers. N Engl J Med, 337, 1558-9.

Van den Bogaard, A. E., London, N., Driessen, C. and Stobberingh, E. (1998) The effect of antibiotic growth promoters on the resistance in faecal indicator bacteria in pigs. In In Programme and abstract of the 38th Interscience Conference on Antimicrobial Agents and Chemotheraphy, Vol. C77 American Society for Microbiology, , pp. 91.

van den Braak, N., Kreft, D., van Belkum, A., Verbrugh, H. and Endtz, H. (1997) Vancomycin-resistant enterococci in vegetarians. Lancet, 350, 146-7.

Van der Auwera, P., Pensart, N., Korten, V., Murray, B. E. and Leclercq, R. (1996) Influence of oral glycopeptides on the fecal flora of human volunteers: selection of highly glycopeptide-resistant enterococci. J Infect Dis, 173, 1129-36.

Vancanneyt, M., Snauwaert, C., Cleenwerck, I., Baele, M., Descheemaeker, P., Goossens, H., Pot, B., Vandamme, P., Swings, J., Haesebrouck, F. and Devriese, L. A. (2001) Enterococcus villorum sp. nov., an enteroadherent bacterium associated with diarrhoea in piglets. Int J Syst Evol Microbiol, 51, 393-400.

Wegener, H. C. (1998) Historical yearly usage of glycopeptides for animals and humans: the American-European paradox revisited. Antimicrob Agents Chemother, 42, 3049.

Weinstein, J. W., Roe, M., Towns, M., Sanders, L., Thorpe, J. J., Corey, G. R. and Sexton, D. J. (1996) Resistant enterococci: a prospective study of prevalence, incidence, and factors associated with colonization in a university hospital. Infect Control Hosp Epidemiol, 17, 36-41.

Willems, R. J., Homan, W., Top, J., van Santen-Verheuvel, M., Tribe, D., Manzioros, X., Gaillard, C., Vandenbroucke-Grauls, C. M., Mascini, E. M., van Kregten, E., van Embden, J. D. and Bonten, M. J. (2001) Variant esp gene as a marker of a distinct genetic lineage of vancomycin-resistant Enterococcus faecium spreading in hospitals. Lancet, 357, 853-5.

80

Willems, R. J., Top, J., van Den Braak, N., van Belkum, A., Endtz, H., Mevius, D., Stobberingh, E., van Den Bogaard, A. and van Embden, J. D. (2000) Host specificity of vancomycin-resistant Enterococcus faecium. J Infect Dis, 182, 816-23.

Willems, R. J., Top, J., van den Braak, N., van Belkum, A., Mevius, D. J., Hendriks, G., van Santen-Verheuvel, M. and van Embden, J. D. (1999) Molecular diversity and evolutionary relationships of Tn1546-like elements in enterococci from humans and animals. Antimicrob Agents Chemother, 43, 483-91.

Williams, A. M., Farrow, J. A. E. and Collins, M. D. (1989) Reverse transcriptase sequencing of 16S ribosomal RNA from Streptococcus cecorum. Int. J. Syst. Bacteriol., 39, 495-7.

Wilson, W. R., Karchmer, A. W., Dajani, A. S., Taubert, K. A., Bayer, A., Kaye, D., Bisno, A. L., Ferrieri, P., Shulman, S. T. and Durack, D. T. (1995) Antibiotic treatment of adults with infective endocarditis due to streptococci, enterococci, staphylococci, and HACEK microorganisms. American Heart Association. Jama, 274, 1706-13.

Yamaguchi, E., Valena, F., Smith, S. M., Simmons, A. and Eng, R. H. (1994) Colonization pattern of vancomycin-resistant Enterococcus faecium. Am J Infect Control, 22, 202-6.