Guidelines esbl prod kiemen 2012

Specialist Patient Safety Commission: nosocomial infections and other undesirable events linked to care and practices

REPORT

Recommendations relating to measures to be implemented to prevent the emergence of ESBL enterobacteriaceae and fight against their spread Proposals written up from the point of view of defining a national prevention programme

February 2010

High Council for Public Health

2

TABLE OF CONTENTS ASSIGNMENT LETTER ........................................................................................... 5 TYPE AND CONTEXT OF THE COMPLETE JURISDICTION ..................................... 7 COMPOSITION OF THE WORK GROUP .................................................................. 8 READING GROUP .................................................................................................. 8 WORK METHOD ..................................................................................................... 9 SYNTHESIS OF RECOMMENDATIONS ................................................................. 12 INTRODUCTION...................................................................................................... 14 RECOMMENDATIONS ............................................................................................. 18 CURRENT EPIDEMIOLOGICAL SITUATION - RISK FACTORS ................................. 22 Introduction .................................................................................................... 22 Evolution of ESBL enterobacteriaceae epidemiology...................................... 22 Prevalence and incidence ................................................................................... 22 Types of enzymes involved ................................................................................. 24 Spread of clonal and plasmid groups .................................................................. 24 Circumstances for the appearance of ESBL enterobacteriaceae infections ....... 25 Risk factors for developing an ESBL E.coli infection .......................................... 27 Place of acquiring strains causing ESBL E.coli infections ................................... 27 Cross transmission of strains and ESBL E.coli resistance genes ........................ 28 Differences between ESBL E.coli and MRSA epidemiology ................................ 28 Role of the environment.................................................................................. 28 Animal world ....................................................................................................... 28 Travelling ............................................................................................................ 29 Presence of ESBL enterobacteriaceae in the environment................................. 30 And beyond ESBL enterobacteriaceae... ......................................................... 30 Relationship between ESBL rate and rate of K. pneumoniae resistant to carbapenems ....................................................................................................... 30 Conclusion ....................................................................................................... 30 MICROBIOLOGY ..................................................................................................... 31 Screening patients that are carriers of bacteria producing ESBL ................... 31 Detection of ESBL production by bacteria ...................................................... 32 Conclusion ....................................................................................................... 32 HOW SHOULD WE FIGHT THE SPREAD OF CTX-M ESBL E.coli ESBL? .................... 34 Prevention measures to set up regarding a patient that is infected / colonised by ESBL E.coli .................................................................................................. 34 Detecting carriers of ESBL E.coli ..................................................................... 34 TREATING INFECTIONS - TACKLING COLONISATIONS - GOOD USAGE / MISUSE OF ANTIBIOTICS .................................................................................................... 39 Introduction .................................................................................................... 39 1 - What prevention should be used to limit as much as possible the risk of digestive carrying of ESBL E.coli that could be secondarily responsible for infection? Looking at good usage of antibiotics ............................................. 39 2 - Which probabilistic antibiotherapy should be offered for a PRESUMED E.coli infection and from what level of prevalence should the ESBL risk be integrated? ...................................................................................................... 40 3 - What corrective antibiotherapy should be proposed for a documented ESBL E.coli infection? ...................................................................................... 42

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

3

4 - What treatment strategy should be employed for digestive ESBL E.coli colonisation? ................................................................................................... 45 ANNEXES ............................................................................................................. 46 Table 1: Evolution of the incidence of EBLSE for 1000 HD from 2004 to 2008 - Analysis restricted to establishments that participated in BMR-Raisin surveillance for each of the five years (n=274). Stratification by type of department. 47 Table 2: Evolution of the incidence of EBLSE for 1000 HD from 2004 to 2008 Analysis restricted to establishments that participated in BMR-Raisin surveillance for each of the five years (n=302). Stratification by inter-region. 47 Table 3: Evolution of incidences / 1000 HD of EBLSE per species from 2002 to 2008. 48 Table 4: Evolution of incidences / 1000 HD of E.coli ESBL from 2002 to 2008 (Raisin) Stratification by inter-region ................................................................ 48 Table 5: Department where patient is located for blood cultures where E.coli strains are isolated that are sensitive or resistant to 3rd generation cephalosporins and sensitive staphylococcus aureus (SASM) or resistant to Methicillin (MRSA). EARSS France data 2005-07. .............................................. 49 Table 6: Time between day of hospitalisation and the day that blood cultures are sampled where E.coli strains are isolated that are sensitive or resistant to 3rd generation cephalosporins. EARSS France data 2005-07. ....... 49 Table 7: Percentage of strains that are sensitive (I+R) to 3rd generation cephalosporins among E.coli strains isolated from blood cultures depending on the type of department and the time between day of hospitalisation and the day that blood cultures are sampled. EARSS France data 2002-08. .................. 50 Table 8: Synthesis of studies - share of community origin of EBLSEs ............ 51 Results ................................................................................................................ 51 Table 9: Synthesis of studies centred on community cases of infections or carrying ESBL E.coli ............................................................................................ 52 Community case .................................................................................................. 52 Figure 1: Evolution 2001-2008 of the proportion of strains sensitive to 3rd generation cephalosporins among E.coli strains isolated from bacteraemia, hospital networks participating in EARSS (Réussir, Ile de France and Azayrésistance networks; www.onerba.org/) ........................................................... 53 Figure 2: Proportion (%) of E. coli resistant to 3rd generation cephalosporins in bacteraemia in Europe in 2008 (EARSS data, www.earss.rivm.nl) ............................................................................................ 54 Figure 3: Evolution 2005-2008 of the proportion of strains sensitive to 3rd generation cephalosporins among K . pneum oniae strains isolated from bacteraemia, hospital networks participating in EARSS (Réussir, Ile de France and Azay-résistance networks; www.onerba.org/) ........................................... 55 Figure 4: Epidemiological situation of ESBL E.coli in the world in 2001-2002. 56 Figure 5: Epidemiological situation of ESBL E.coli in the world in 2005. ..... 56 Figure 5: Epidemiological situation of ESBL E.coli in the world in 2005. ..... 57 Figure 6: Epidemiological situation of ESBL E.coli in the world in 2007. ........... 57 Figure 6: Epidemiological situation of ESBL E.coli in the world in 2007. ........... 58 Figure 7: Proportion (%) of resistance to 3rd generation cephalosporins in E.coli according to the % resistance to imipenem in K . pneum onia. Bacteraemia, EARSS 2008. ................................................................................. 59 Bibliographic references ..................................................................................... 62

High Council for Public Health

4

ASSIGNMENT LETTER

High Council for Public Health Specialist Health Safety Commission Technical committee for nosocomial infections and infections linked to care Ref.: 08/373/BGB/BT/SF

Paris, 1st September 2008

File monitored by Béatrice TRAN Tel. 01 40 56 79 53 EmaiL: [email protected]

Subject: Set up a work group on preventing transmission of ESBL enterobacteriaceae.

Dear Sir, Tackling multiresistant bacteria (MRB) is a priority topic for the CTINILS's work programme 2008-2009 validated by our committee in the meeting on 27th February 2008 and by the High Council for Public Health (specialist commission - CS1 "health safety") on 14th March. Among these MRB, enterobacteriaceae producing enlarged spectrum beta-lactamase (ESBL) have now assumed a worrying position. In addition, during its plenary session on 14th May 2008 our committee decided to look into measures to prevent and fight the selection and spread of these ESBL enterobacteriaceae, accompanied by a reflection led by the public powers with a view to defining a national programme to prevent MRB. aims:

I would appreciate it if you could organise and preside over a work group with the following

-

analyse the recent scientific literature on ESBL enterobacteriaceae related infections;

-

report on the national and international epidemiological and microbiological situation regarding ESBL enterobacteriaceae;

-

propose specific recommendations to fight against the selection and spread of these bacteria emerging in our country.

The group will be made up of competent people that you will bring together for this purpose that might include experts from outside the CTINILS/HCSP. The opportunity might arise to have to hold hearings or get external opinions.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

5

The work group report will be presented at the CTINILS plenary session in January 2009, before validation by the High Council for Public Health Specialised "health safety" Commission. I would like to thank you for your commitment to this matter, and I remain yours sincerely,

Dr Bruno GRANDBASTIEN President of the Technical Committee for Nosocomial Infections and infections linked to care

Professor Christian RABAUD CHU Nancy - Hôpitaux de Brabois Service de maladies infectieuses et tropicales Tour Drouet - rue du Morvan 54511 VANDOEUVRE Cedex CC: Dr Jean-Christophe LUCET GH Bichat-Claude Bernard UHLIN 46 rue Henri Huchard 75877 PARIS Cedex 18 Dr Marcelle MOUNIER CHU de Limoges – Hôpital Dupuytren Laboratoire de Bactériologie - UF Hygiène 2 avenue Martin Luther King 87042 LIMOGES Cedex

High Council for Public Health

6

TYPE AND CONTEXT OF THE COMPLETE JURISDICTION 1 It was specified that the group would, throughout their work, aim to search for a priori harmonisation with other reflections in progress or recently finished in close fields, or even in common parts with this complete jurisdiction: -

Prevention of crossed transmission: extra contact precautions (French Hospital Hygiene Company – SFHH - on request from CTINILS); document published in April 2009.

-

SFHH Guide: "Hand Hygiene" (SFHH, O.Keita-Perse); document published in June 2008. Tackling and preventing infections linked to pathogen bacteria that are highly resistant to antibiotics importated into France (juristiction of the HCSP via the DGS, in progress) Surveilling and preventing infections associated with care (update - HCSP, in the process of being published) "MRB Plan" (Strategic national plan 2009-2013 to prevent infections associated with care, being drawn up).

-

-

Furthermore, it was stated, during the plenary session of the HCSP Specialist Patient Safety Commission on 16th April 2009, that it was better to concentrate the group's work on the problem issue of "the emergence of ESBL E.coli, type CTX-M". Fighting against selection and spread of ESBL enterobacteriaceae Recommendations centred on CTX-M ESBL E coli Or "What can we do to delay the progressive replacement of E coli ampiR by CTX-M ESBL E coli ? »

1

Details provided by Bruno Grandbastien, president of the CTINILS and then of the High Council for Public Health's (HCSP) Specialised Patient Safety Commission (CSSP) who sent out the assignment letter received by Jean-Christophe Lucet, Marcelle Mounier and Christian Rabaud: Jurisdiction from GroupiLIN - Subject joining the

preparation of a national "MRB" programme - Subject entering the work programme 2008-2009 for the CTINILS validated by the HCSP.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

7

COMPOSITION OF THE WORK GROUP CTINILS-CSSP

Christian Rabaud: [email protected], coordination

CTINILS

Jean-Christophe Lucet: [email protected] Marcelle Mounier: [email protected]

InVS

Sylvie Maugat: [email protected]

Afssa

Jean-Yves Madec: [email protected]

Onerba

Jérome Robert: [email protected]

BMR-Raisin/AP-HP

Vincent Jarlier: [email protected]

National plan to preserve antibiotic efficacy: Anne Claude Crémieux: [email protected] SPILF

Rémy Gauzit: [email protected] François Caron: [email protected]

SFHH

Hervé Blanchard: [email protected]

CClin

Anne Carbonne: [email protected] Xavier Bertrand: [email protected]

Christophe de Champs Marie-Hélène Nicolas-Chanoine Roland Quentin (microbiologist) Gaëtan Gavazzi (geriatrician) Bertrand Souweine (intensive care) Yannick Aujard (neonatologist)

[email protected] [email protected] [email protected] [email protected] [email protected] [email protected]

READING GROUP Antoine Andremont Serge Alfandari Elisabeth Aslangul Bruno Coignard Didier Gruson Olivia Keita Perse Edouard Bingen Alain Lepape Alexandra Mailles Patrice Nordmann

High Council for Public Health

8

WORK METHOD Four sub-groups were set up in November 2008. Each was aiming to pool knowledge using existing data. On this basis, the sub-group should, on the one hand, make a point of mentioning recommendations but could also, on the other, specify the type of data that appeared to be missing, required to prop up or finalise their recommendations, plus any extra studies that might also be useful. Sub-group no 1: Current epidemiological situation; risk factors Group members Coordination: Xavier Bertrand Rémy Gauzit Gaétan Gavazzi Vincent Jarlier Jean-Yves Madec Sylvie Maugat Marie-Hélène Nicolas-Chanoine Evolution of the frequency of identifying hospital ESBL enterobacteriaceae of the

Enterobacter or Klebsiella type. Analysis of actions implemented to fight the emergence and spread of these ESBL enterobacteriaceae. Results obtained?

ESBL E. coli in hospitals? What type of resistance? What are the consequences? Evolution of the phenomenon? Evolution of the frequency of identifying CTX-M type ESBL E.coli. Hospital's role? Community's role? Uniqueness of the phenomenon? Share of colonisations? Share of infections? Pathogenicity of the strains? Do we know the risk factor for being colonised or infected by an ESBL enterobacteriaceae? In the hospital? Outside the care medium? Particular case of CTX-M type ESBL E. coli circulating outside the hospital? Does this refer to phenomena emerging over the whole country? Importation pathologies spreading secondarily over the country? Is there a link between what is seen in human medicine and what is observed in the animal world? What is the epidemic potential for these different ESBL enterobacteriaceae? Should we run strain epidemics? On "plasmid epidemics"?

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

9

Sub-group no.2: Microbiology Group members Coordination: Christophe De Champs Jean-Christophe Lucet Marcelle Mounier Marie-Hélène Nicolas-Chanoine Jérôme Robert Yves Madec When should we look for the presence of ESBL enterobacteriaceae? Of CTX-M type ESBL E. coli? Where (in which sample) should we look for the presence of bacteria expressing an ESBL? Of a CTX-M type ESBL E. coli? How (using which technique) should we look for the presence of bacteria expressing an ESBL? Of a CTX-M type ESBL E. coli? Sub-group no. 3: How to fight the spread of CTX-M type ESBL E.coli Group members Coordination: Anne Carbonne Hervé Blanchard Gaétan Gavazzi Vincent Jarlier Jean-Christophe Lucet Marcelle Mounier Bertrand Souweine The role of precautions intended to avoid cross transmission? In the hospital? In the community? For home carers? For families? Should we propose a survey for every new case? In the hospital? In the community? According to which modalities? Sub-group no. 4: Treating infections / Tackling colonisations / Good usage – Misuse of antibiotics Group members François Caron Anne-Claude Crémieux Christian Rabaud Should colonisations by CTX-M type ESBL E. coli be treated? If so, in which circumstances? If so, using what means?

High Council for Public Health

10

How can we treat infections due to CTX-M type ESBL E.coli? Who can (or should) treat these infections (primary health care? hospital?)? How can we treat a serious community infection from a urinary starting point? Should we take into account the ESBL risk straightaway? Role of carbapenems? Role of betalactamase inhibitors? Role of furadantin? Role of fosfomycin? Role of aminoglycosides? Role of good use of antibiotics? On which molecules should we focus the recommendations (fluoroquinolone and urine, see Afssaps recommendations, … but also C3G) ? How can we get the message across? To which targets? This chapter was subject to joint work with experts from the work group for anti-infectious medication within the Afssaps Market Authorisation Commission.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

11

SYNTHESIS OF RECOMMENDATIONS •

INFORMATION – TRAINING

1/ Inform the whole medical world about the epidemic spread of ESBL producing E.coli that in time will expose us to the risk of a therapeutic dead end. 2/ Inform microbiologists about the epidemic spread of ESBL E.coli and their resistance genes - Specify that identifying the ESBL resistance mechanism must be suspected and documented when demonstrating resistance to 3rd generation cephalosporins (C3G) and that the result of this research must appear in the report sent by the lab. Provide biologists with the means to participate reliably in surveillance surveys. 3/ Raise awareness among the population about the emergence of a sanitary danger (new faecal danger) that comes from an excessive use of antibiotics and the epidemic spread of ESBL E.coli strains (or from their resistance genes) cause by insufficient respect for basic hygiene rules. •

SURVEILLANCE

4/ Make sure, on a national level, of epidemiological surveillance of ESBL enterobacteriaceae - monitor the ESBL type resistance rate in terms of incidence and prevalence within the E.coli species. •

THERAPEUTIC CONSIDERATIONS - GOOD USAGE AND LEAST USE OF ANTIBIOTICS

5/ Assess selection pressure for antibiotic regimes concerning current infections. Define, assemble and inform on situations when it is recommended not to prescribe an antibiotherapy (respiratory, urinary, etc.). In situations where an antibiotherapy is recommended, specify its optimal spectrum and its duration. Maintain current recommendations concerning the probabilistic treatment of urinary infections - remembering that carbapenems are not part of probabilistic treatment for urinary infections. Immediately review recommendations relating to tackling intra-abdominal infections and review them relating to neonatal infections. In the event of identifying ESBL E.coli, keep back the use of carbapenems for tackling severe infections, by remembering that the use of carbapenems is a "false good solution" - an effective solution therapeutically on an individual scale but a high risk solution to encourage the development of carbapenemases (valid risk on individual and collective scale). Assess new treatment regimes for infections documented as ESBL E.coli: Cephamycin, C3G + betalactamase inhibitor…

High Council for Public Health

12

•

HYGIENE MEASURES

6/ Apply extra "contact" precautions to all infected or colonised patients. Critical points: hand washing hygiene and excreta management. The measures will be applied to health establishments and in EHPAD (medical-social sector). At home and in collectives other than health establishments (schools, etc), emphasis will be put on hand washing hygiene and general hygiene concerning toilets and food. 7/ In health establishments, look for digestive colonisation for ESBL enterobacteriaceae, in the contact subjects for a case - do not try and eradicate digestive carrying of ESBL E.coli by following a decolonisation protocol. •

RESEARCH

8/ Set up complementary studies intended to improve knowledge on risk factors for colonisation of ESBL E.coli. 9/ Commit to complementary work on veterinary and environmental aspects. 10/ Study the role of effluent in the spread of this phenomenon.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

13

INTRODUCTION Escherichia coli is a commensal bacteria in the digestive tract (~108/g of faeces) that is

the most frequent origin of bacterial infections listed in the community, the urinary infection. These infections are more often than not benign but they can be serious in the event of parenchymatous complaints (pyelonephritis, prostatitis) or if they appear in specific fields (maternal-foetal impact of gravid infection). E. coli is also the main bacteria responsible for sometimes serious digestive infections: peritonitis, but also heptobilliary, intestinal (sigmoidite), pancreatic infections, etc. Furthermore, E.coli strains are carriers of specific pathogenicity factors and can be responsible for neonatal meningitis following mother-child transmission during birth (E. coli K1) or diarrhoea (toxinogen E. coli). Since the start of the 2000s, E. coli has progressively become the bacterial species most affected by the emergence of enlarged spectrum betalactamases (ESBL), enzymes that deactivate most betalactamines: CTX-M. Until recently, classic ESBLs (derived from TEM and SHV) were on the whole carried by K. pneumoniae and Enterobacter spp., commensal species usually in low concentration in the digestive tract and above all responsible for nosocomial infections. Because of the abundance and the ubiquitous nature of E. coli, the frequency of nosocomial and community infections in which this species is implicated, the emergence of ESBLs in E. coli widened the stakes for epidemiological surveillance and tackling these infections, singularly complicating any control over their spread. The incidence rates for enterobacteriaceae producing ESBL (EBLSE) in general for E.coli and, to a lesser degree, for K. pneumoniae in particular have largely increased in France since the beginning of the 2000s. The national incidence density for ESBL enterobacteriaceae in 2008 is around 0.4 cases for 1000 hospitalisation days (HD) in the hospitals and for E. coli ESBL it was 0.16/1000 HD. The proportion of strains that are resistant to C3G (= essentially ESBL strains) among the strains responsible for bacteraemia, has increased in France from 1 to 6% for E.coli between 2001 and 2008 and from 5 to 17% for K. pneumoniae between 2005 and 2008. As a consequence, the estimated number of cases of bacteraemia for ESBL E.coli was around 2000 and for ESBL K. pneumoniae bacteria it was around 1000 in France in 2008. If resistance to C3G by ESBL entirely replaced resistance to ampicillin by "classic" betalactamase (TEM-1 or 2, SHV-1) for E.coli, the number of bacteraemia with ESBL E.coli would reach 10 to 15,000 per year in France. The proportion of strains resistant to C3G among strains responsible for E.coli bacteraemia is already higher in certain countries in Europe in its broadest sense (e.g. 10% in Spain, Greece and Portugal, 15% in Italy and Israel, 42% in Turkey) which leads us to fear that France will evolve along the same lines if we do not try and limit the phenomenon. ESBL enterobacteriaceae are also resistant to other antibiotic families, due to the presence of associated genes on the same plasmids or associated chromosome mutations. In France in 2008, overall, ESBL enterobacteriaceae were often resistant to tobramycin (70 %), to gentamicin (50 %), to amikacin (30 %), to ciprofloxacin (75 %) and, as far as E. coli is concerned, to tobramycin (75 %), to gentamicin (35 %) to amikacin (25 %) and to ciprofloxacin (70 %). The emergence of multiresistance within this species therefore puts us up against the risk of a therapeutic dead end.

High Council for Public Health

14

ESBL E.coli infections therefore escape classic therapeutic regimes for confirmed and likely serious enterobacteriaceae infections that are based on C3G, fluoroquinolones and aminoglycosides. From this fact, treatment for documented infections from ESBL E.coli and infections that we fear to be ESBL E.coli tends to intensify the use of carbapenems, considered as the last resort treatment (perhaps along with glycylcyclines). An increase in the use of carbapenems in return exposes us to a risk of resistance emerging to these antibiotics and particularly the emergence of enterobacteriaceae producing carbapenemases, betalactamases that deactivate nearly all betalactamines. For the time being, carbapenemases can be seen above all in the K. pneumoniae species but all enterobacteriaceae are likely to become producers because coding genes are mobile. These enterobacteriaceae are not in general more sensitive than to colistin and the infections that they cause carry a high death rate. In some countries close to France, the proportion of strains resistant to carbapenems in K. pneumoniae is already very high (Greece, Turkey, and Israel). Epidemics of K. pneumoniae producing carbapenemases have appeared in France since 2004 and more singularly since 2007 (three epidemics just in 2007 in Parisian Hospitals (AP-HP)). In Europe, the proportion of strains resistant to C3G in E. coli and the proportion of strains resistant to carbapenems in K. pneumoniae appear correlated. In the industry's current antibiotic development situation, it is not very likely that we will soon have new molecules that are effective on ESBL E.Coli and on other multi-resistant enterobacteriaceae. So, the fight against the emergence of ESBL E. coli is from now on not only a problem but also a public health duty. This fight has a double aim: 1) limiting the number of ESBL E.coli carriers and therefore the number of ESBL E.coli infections. The target to meet seems to no longer involve eradicating a phenomenon that is already installed, but trying to maintain the ESBL enterobacteriaceae rates overall, including ESBL E.coli, as low as possible and for as long as possible to give time to draw up preventive strategies and alternative therapies; 2) limit the emergence of enterobacteriaceae that are resistant to carbapenems resulting from intensive use of carbapenems which would no longer happen if ESBL E.coli incidence rose. The spread of ESBL E.coli is the consequence of two phenomena: cross transmission (spread) and antibiotherapy selection pressure (emergence and above all increase in the density of colonisation). •

Cross transmission of the bacteria itself, or the ESBL genes that the bacteria houses, in the hospital environment, in the medical-social establishments (EHPAD) and in the family or collective environment. Cross transmission is seen by either clonal diffusion of strains (strain epidemics) or by diffusion of resistance genes (gene epidemics) leading to the same result in the end. E.coli acquiring a CTX-M gene that originates from Kluyvera is an extremely rare event because inter-human dissemination of ESBL genes is easy ('mobile' genetic elements: plasmids, transposons …). In the same way, ESBL genes for enterobacteriaceae other than E.coli (e.g. K. pneumoniae) can be incorporated by E. coli. We should therefore consider the risk of spreading ESBL enterobacteriaceae and their genes as a whole.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

15

•

Selection pressure by antibiotics used in human and veterinary medicine Let's remember here that bacterial resistance is the main undesirable effect of antibiotics. After inter-human transmission (or transmission to man from the animal world or the environment), ESBL E.coli strains or ESBL genetic supports are implanted in the "recipients" digestive flora. The resistant strain can then be "selected" and multiplies in the case of active antibiotherapy on the other microorganisms making up the digestive flora. The growth of inoculums for this E.coli ESBL strain within the digestive flora is going to raise the risk of infection due to ESBL E.coli in the carrier, but also raise the risk of spreading ESBL E.coli around them.

Cross transmission of ESBL E.coli, or ESBL genes, is encouraged by the size of the container, much greater than in the cases of Methicillin resistant staphylococcus aureus (MRSA) In their digestive tract, humans host over 108 E. coli per gram of stools, so a total of 1010 to 1011. An ESBL E. coli carrier can thereby eliminate more than 1010 ESBL E. coli to the environment everyday through their excreta. In the case of urinary ESBL E.coli infection, the number of bacteria excreted per day through urine can reach 108 to 109. Interhuman ESBL transmission, or of ESBL genes, for E. coli is therefore easy. With the development of streamlining techniques such as 'everything down the drain' and treating waste water, our societies thought we were safe from "faecal peril" represented by pathogen bacteria for digestive tropism (salmonella, shigella, etc.). ESBL E.coli today represents a new type of faecal peril. ESBL E.coli colonisation or infection is still most often demonstrated in hospitals today. However, it also sometimes appears right at admission or during a consultation in the hospital or, remarkably, in patients without any direct or indirect previous link with a care structure. Even more worrying, CTX-M type ESBL E.coli was identified using stools from carrier children, particularly in developing countries. These epidemiology characteristics underlie the idea that the E.coli strains that produce CTX-M would have emerged outside hospitals. Recent works have clearly shown the spread and digestive carrying of an E.coli strain producing ESBL within a family after this strain was responsible for a urinary infection in one of its members and its presence had been demonstrated in this person's digestive tract. To limit the epidemic diffusion of E-coli and its resistance genes, it is advisable to attack the cross transmission phenomena and over-consumption of antibiotics. In order to reduce the selection pressure exerted by antibiotherapies, it is of course advisable to progress not only making "good usage" of antibiotics, prioritising the use of molecules that exert the weakest selection power over the ESBL E.coli, but also in the "least usage". This is possible when we know that the consumption of antibiotics in France is two to three times higher in the community, per inhabitant and per year, than seen in other European countries. The "antibiotics are not automatic" campaign obtained a significant reduction in the quantities of antibiotics prescribed in the community without the population's state of health having to suffer! Now we have to go one step further. Prescriptions for C3Gs and fluoroquinolone must be the subject of a particular effort in the community and in hospitals. The field is very wide and we should not underestimate the difficulties to implement these actions. For this reason, this is an essential evolution. To fight cross transmission (diffusion of ESBL strains and resistance genes), strategies should associate hospital hygiene measures (prevention of cross transmission; prevention of actual bacteria spreading and their resistance genes by controlling excreta), hygiene measures in the community (prevention of cross transmission in collectives: old High Council for Public Health

16

people's homes, schools, etc.). We should make particularly certain that hygiene measures intended to break this epidemiological chain are followed and if necessary intensify them, not only in hospitals but in all structures that might house people who are colonised or infected, from hospital to home, including EHPADs... All participants, medical or paramedical carers, on a contract or self-employed, people around the patients and patients themselves will play a role in these strategies. The strategies to be implemented have not managed to distinguish the different types of ESBL E.coli (CTX-M, TEM, SHV, etc.) or even distinguish the two main species of "commensal" ESBL enterobacteriaceae (E. coli and K. pneumoniae) due to the intricate epidemic phenomena (diffusion of the same ESBL genes in and among these two species) and the fact that CTX-M type ESBLs largely dominate the other types of ESBL (TEM, SHV, etc.) Strategies centred on one species or a type of ESBL strongly risk being ineffective and will be more difficult to explain and to set up. Finally, it will be necessary to associate these strategies with measures specifically targeting controlling the spread of enterobacteriaceae resistant to carbapenems, as an indirect consequence of the spread of ESBL enterobacteriaceae: limiting selection pressure by these antibiotics by promoting alternative treatments for ESBL enterobacteriaceae infections or those likely to be, particularly urinary infections that are the most frequent and the least serious, by immediately setting up "emergence" type control measures (such as for enterococcus resistant to vancomycin) from the appearance of the first cases. In total, the spread of ESBL enterobacteriaceae, including ESBL E.coli and strategies to limit this spread are typically sustainable development problems in the same way as sparing drinking water resources, and of course implicating the medical world but more widely, the whole population. The reduction in antibiotics consumption must spread to hospitals, in the community, not only for human medicine but also in the animal world. Environmental measures might also be envisaged: food checks, effluents, etc. Working from these statements, the group wished to formulate the following recommendations.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

17

RECOMMENDATIONS 

INFORMATION – TRAINING

1/ Spread throughout the medical world (doctors, dentists, midwives and paramedics, on a contract and freelance) information on the epidemic spread of ESBL E.coli that in time exposes us to the risk of a therapeutic dead-end. This message must be used as a lever to promote good usage and particularly promote less usage of antibiotics. We will insist on the particularly selective character of 3rd cephalosporins (C3G) and fluoroquinolones and on any therapeutic alternatives that might be envisaged. Because, even if it must be encouraged, systematically resorting to an antibiotic reference when demonstrating an ESBL enterobacteriaceae reveals itself to be a lot more complex in EHPAD and in the community than in a health establishment and training actions should be specifically designed to raise awareness among EHPAD coordinating doctors, community doctors working in EHPAD but also more widely, all community doctors. This training, presented in the form of pedagogic kits that can be used all over the country, should be drawn up in partnership with companies that know about geriatrics, hygiene and infectology, articulating with the group in charge of the plan to preserve antibiotic effectiveness. 2/ Raise awareness among all microbiologists concerning the problem of the epidemic spread of ESBL E.coli and its resistance genes and the means that must be implemented to identify this type of resistance. ESBL production in enterobacteriaceae resistant to C3G must be systematically researched and the result of this research must appear in the report sent by the lab. This information is important in terms of therapy (clinic technician) and in epidemiological terms (hygienist). The phenomenon's evolution must be monitored nationally. External quality controls must allow biologists to make sure that they can recognise the production of ESBL and therefore participate reliably in monitoring surveys. We will make sure that microbiologists have at their disposal, as they are developed, fast detection techniques for ESBL enterobacteriaceae and particularly ESBL E. coli in the samples for diagnosis targeting and in the commensal flora. 3/ Raise awareness among the population about the emergence of a sanitary danger that comes from excessive use of antibiotics and the epidemic spread of ESBL E.coli strains and from their resistance genes caused by insufficient respect for basic hygiene rules. The information will mainly bear on the risk of cross transmission for these bacteria and their resistance genes, for which the natural container is mainly digestive and urinary. Awareness will be raised among the population concerning the resurgence of a faecal peril. This information can be drawn up in partnership with Inpes. The information should concern the risk of hand-carrier transmission by including the more general problem issue of preventing cross transmission of infectious agents in the community; air transmission (tuberculosis bacille, flu virus, other respiratory viruses, etc.), hand to hand transmission, contact or exchange of objects (community MRSA, viral gastroenteritis, ESBL E.coli).

High Council for Public Health

18



SURVEILLANCE

4/ Make sure, nationally, that there is epidemiological surveillance on ESBL E.coli and more generally, ESBL enterobacteriaceae: -



The surveillance systems in place (Raisin-BMR, EARSS France, Onerba) must be maintained and consolidated. Specific studies must be envisaged to complete this device: as an example, we aim to specify the frequency of carrying ESBL E.coli in different types of population (old people's home, obstetrics, HAD; patients arriving in A&E), stratifying these measures according to factors such as recent hospitalisation or antibiotherapy history. In the same way, we will evaluate the evolution of the number of infections and serious infections from ESBL E.coli. We will be specifically provided with the means to specify the nosocomial or community character of these infections and evaluate evolution of ESBL E.coli within E.coli infections seen in the community. THERAPEUTIC CONSIDERATIONS – GOOD USAGE AND LESS USE OF ANTIBIOTICS

5/ In a general way, and to reduce the antibiotherapy selection pressure that encourages the emergence, pullulation and spread of ESBL enterobacteriaceae, it is advisable to reduce the volumes of antibiotics used on humans by intensifying actions run within the framework of the antibiotic field. Alongside the concept of "good usage" it is particularly necessary to introduce the concept of "less usage". A strong proposal for this field involves defining, assembling and informing about situations where it is recommended to not prescribe an antibiotherapy (respiratory sphere, urinary, etc.). Furthermore, it will be useful to look again and, if necessary, modify the most frequent first attempt antibiotic treatment recommendations for infections in the aim of reducing the selection pressure on ESBL enterobacteriaceae in commensal flora (particularly digestive flora). It will be necessary to promote resorting to antibiotics other than the C3G and fluoroquinolones. This point might require clinical research. Recommendations concerning first attempt treatment for urinary infections are not questioned right now but should be reconsidered depending on the evolution of the ESBL E.coli rate identified within E.coli infections seen in the community. On the other hand, recommendations relating to tackling intra-abdominal infections seem to be obsolete and should be reviewed as soon as possible. As far as urinary infections are concerned, it is advisable, as far as possible, to prioritise resorting to documented antibiotherapies and limit resorting to probabilistic antibiotherapies; so as far as tackling complicated cystitis is concerned, in the absence of any sign of seriousness, it can be envisaged to delay antibiotherapy to be able to have an antibiogram beforehand. As far as pyelonephritis and prostatitis are concerned, the benchmark probabilistic treatment today is a C3G (parenteral) or a fluoroquinolone (per os or parenteral only in adults). The recommendations envisage adding an aminoglycoside for more severe forms of community urinary infections (serious sepsis, pyelonephritis over obstacle, newborns and babies under 3 months, etc.); this addition partly secures the risk of failure in the event of ESBL E.coli as French strains remain sensitive to aminoglycosides in around 50% of cases.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

19

Faced with a documented or strongly suspected ESBL E.coli infection: (a) it is advisable to prioritise the use of molecules other than carbapenems; (b) as the use of carbapenems is reserved for tackling severe infections: imipenem, meropenem or doripenem (only in the case of complicated urinary infections), ertapenem (AMM in intra-abdominal infections) possibly associated with aminoglycosides. It is best to remember that the use of carbapenems is a false good solution – an effective solution therapeutically on an individual scale but a high risk solution to encourage the development of carbapenemases (valid risk on individual and collective scale). Finally, the evolution of resistances to fosfomycin but also to furans should be particularly monitored as their use exposes us to the appearance of resistance; so resistance to fosfomycin which is currently 2% in France reaches 25% in some Spanish series - the evolution of this resistance appears to be correlated with the growing use of this molecule. 

HYGIENE MEASURES

6/ In terms of fighting cross transmission of ESBL E.coli, several critical points are identified that will be subject to new recommendations: - For fifteen years, control of cross transmission in care environments has above all targeted MRSA, whilst ESBL enterobacteriaceae were also "target" bacteria for MRB control recommendations (1999). It is advisable to put ESBL enterobacteriaceae back at the same level of concern and intervention as MRSA. - As the container for ESBL E.coli is the digestive tract, and urine in the case of infection, the critical points are hand washing hygiene and managing excreta handled by carers. In any case, very specific attention must be given to hand washing hygiene for people who have to work on a patient that is colonised or infected by E.coli ESBL. The use of alcohol hand rub (AHR) will be encouraged within the framework of their recommendations (except for hands that are macroscopically dirty). Hand washing hygiene will also be taught to colonised patients who will be encouraged to disinfect their hands by cleaning with an AHR as often as necessary. - Managing excreta, waste associated with care and dirty laundry, not a key point for controlling MRSA spread, is crucial to control ESBL E.coli. It is advisable to encourage staff handling excreta, waste, dirty laundry to (a) use gloves and disposable aprons, (b) customise them in an ad hoc way and (c) to throw them away as quickly as possible to prevent them becoming recipients for cross transmission. Gloves and aprons are destroyed at the same time. It is best to assure good quality packaging taking into account later transporting or storage. Waste generated in this way must be disposed of using in a subsidiary adapted to the risk. Contrary to regulations, it is recommended to not consider as waste from care activities with a risk of infection (DASRI) any waste that can be assimilated into household rubbish (DAOM) from a patient to whom the extra contact type precautions are applied. So, resorting to the DAOM subsidiary can be envisaged as soon as the excreta and/or the care activity waste have been correctly conditioned. - All these measures will be applied to health establishments and in EHPAD (medical-social sector). At home and in collectives other than health establishments (schools, etc), emphasis will be put on hand washing hygiene and general hygiene concerning toilets and food. - Specific information for health professionals who have to work with patients (hand washing hygiene, disposable aprons, etc.) will be drawn up.

High Council for Public Health

20

7/ Screening for ESBL enterobacteriaceae in care establishments on contact subjects for a case was recommended by the SFHH in intensive care, in MCO and in SSR in the event of an epidemic situation: systematic screening of all patients (see document on "Prevention of cross transmission: extra contact precautions"www.sfhh.net). Outside these situations, systematic contact patient screening for patients presenting an ESBL E.coli infection (or a sample for a positive diagnosis target) should be encouraged (1st level contact: possible room-mates, patients who have to share the same toilets, and patients for whom nursing care is carried out by the same staff and the infected patient). This same attitude should then be applied to the contact patients for level 1 contact patients if they are also proven to be colonised. Extra "contact" precautions will be applied to all infected or colonised patients. Antibiotic treatments aiming to decolonise patients carrying ESBL E.coli are not recommended. 

RESEARCH

8/ Set up complementary studies intended to improve knowledge on the risk factors for colonisation of ESBL E.coli, prevention strategies (recommendations for screening? measures to be associated to it? cost/efficacy ratio of the strategy? therapeutic implications?) and the therapeutic implications, particularly in establishments housing old people and in other collectives at risk from cross transmission (maternity hospitals, nurseries, schools). 9/ Complementary work must be adopted on veterinary and environmental aspects for the ESBL E.coli problem issue. This should particularly analyse the impact of using antibiotics in veterinary medicine and in raising practices on the emergence of ESBL E.coli in animals' digestive tract, in food and also in the environment. Bringing surveillance data from humans together with data from the animal world should be encouraged. In particular, it will be useful to specify the type of enzyme responsible for possible ESBL type resistance in enterobacteriaceae identified in the digestive tract of healthy animals, sick animals and possibly in food resulting from transformation of breeding animals, downstream of the abattoir. 10/ The role of effluent from urban agglomerations, particularly health establishments, in the environmental spread of ESBL E.coli should be studied.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

21

CURRENT FACTORS

EPIDEMIOLOGICAL

SITUATION

-

RISK

Introduction Enlarged spectrum betalactamases (ESBL) are bacterial enzymes that deactivate most betalactam with the exception of Cephamycins and carbapenems. They constitute the main resistance mechanism for beta-lactam in enterobacteriaceae. Often associated with resistance to other families of antibiotics, they considerably restrict therapeutic possibilities. Taking into account the major clinical impact of enterobacteriaceae and particularly the Escherichia coli species, the wide spread of this type of enzyme is a major public health threat. Around fifty years ago the first plasmid beta-lactamase, TEM-1, appeared in E.coli, as an enzyme leading to resistance to amoxicillin which, from then on, was present throughout the world in around 50% of E.coli strains responsible for infections contracted either in hospital or in the community. ESBLs appeared in the 1980s after using C3Gs (particularly cefotaxime, ceftazidime). Different ESBL groups were characterised and most of them are present in Europe. The first ESBL described in Germany in 1983 in Klebsiella pneumoniae and Serratia marcescens, derived from a SHV-1 type penicillinase and was named SHV-2 by analogy (1). In 1984, a strain of K. pneumoniae producing an ESBL derived from TEM type penicillinase, known as TEM-3, by analogy, was characterised in France (2). In both cases, the genetic determining factor was plasmid and transferable by union. Many new ESBL belonging to TEM and SHV groups were subsequently identified (3-5), over 160 and 100 variants in each of these two groups respectively (http://www.lahey.org/studies/webt.htm). TEM and SHV type ESBLs spread throughout the world and many epidemics have been described. The K. pneumoniae and Enterobacter aerogenes species were most frequently involved in these epidemics, the others (E. coli, S. marcescens, Klebsiella oxytoca, Enterobacter cloacae and Citrobacter freundii) being much less frequent (3, 6, 7). During the 90s, new enzymes emerged: CTX-M. They hydrolyse very effectively particularly cefotaxime, which is where they get their name from. These enzymes are coded by genes deriving from chromosome genes naturally present in Kluyvera bacteria (8). The first enzymes of this type have been described since 1989 (9) but did not spread worldwide until 1995 (3). Nowadays, 84 variants of CTX-M have been identified, organised into 5 groups (CTX-M-1, CTX-M-2, CTX-M-8, CTX-M-9 and CTX-M-25) based on the amino acid sequence. The E.coli species is the most often implicated, followed by K. pneumoniae throughout the world but different CTX-M types dominate in different countries (3, 7). Evolution of ESBL enterobacteriaceae epidemiology

Prevalence and incidence National surveillance data on multiresistant bacteria, Raisin (Network for alarm, intervention and surveillance of nosocomial infections; www.invs.sante.fr) shows that, in the stable cohort of 302 care establishments participating in surveillance every year from 2004 to 2008, the incidence of enterobacteriaceae producing isolated ESBL from samples with a diagnosis aim increased from 0.17 to 0.31 for 1000 hospitalisation days (HD) (see table 1 in annexe). The evolution occurs in the same direction in each inter-region (see table 2). Incidence densities in 2008 were much higher in intensive care (1.35) than in short stays on the whole (0.47) and in follow-up care and rehabilitation - long term stays (SSR-SLD) (0.17).

High Council for Public Health

22

In AP-HP hospitals, this incidence remained stable from 1996 to 2001, at around 0.1 for 1000 HD, and then has increased regularly ever since, up to 0.52 in 2008. The stratified incidence densities according to activity are 0.61 in short stay hospitals (0.65 in general medicine, 0.61 in surgery, 1.62 in intensive care) and 0.29 in SSR-SLD hospitals. The incidence of all species of ESBL enterobacteriaceae is therefore from now on similar to that of MRSA that has clearly diminished over the last few years to reach 0.51 for AP-HP for 1000 HD overall (0.51 in short stay hospitals, 1.05 in intensive care and 0.51 in SSR-SLD). Another remarkable evolution occurred in France between 2002 and 2008 in terms of distributing species within ESBL enterobacteriaceae. The role of the E.coli species increased from 18 to 58 %, E. cloacae from 6 to 10 %, whilst E. aerogenes dropped from 36 to 8 % and K. pneumoniae remained stable from 18 to 15 %. This double evolution leads to an increase in the incidence of ESBL E.coli from 0.02 to 0.16 for 1000 HD between 2002 and 2008 nationally (see table 3 in the annex). This incidence increased in all the inter-regions (see table 4) but in 2008, incidence was much higher at APHP (0.28) than in the Eastern (0.13) and Western (0.08) inter-regions. The incidence of ESBLs also rose between 2002 and 2008 for K. pneumoniae (0.02 to 0.04) and E.cloacae (0.01 to 0.03). Surveillance runs by the three federated networks within Onerba (Réussir, Ile de France and Azay-résistance) that included 57 hospitals and contributed to the European surveillance system EARSS (European Antimicrobial Resistance Surveillance System, www.http://www.rivm.nl/rivm.nl/earss) shows, within the E.coli strains isolated from the bacteraemia, a drop in the proportion of strains sensitive to C3G (see figure 1 in the annexe): 98 to 98.5 % in 2002, 2003 and 2004 (out of 2495, 2264 and 5507 strains respectively), 97.5 % in 2005 (5650 strains), 97 % in 2006 (6460 strains), 96 % in 2007 (7645 strains) and 94 % in 2008 (7 990 strains). Resistance to C3G constitutes an indirect marker for the production of ESBL in this species. On this basis, we can estimate that there were, in 2008, in these 57 hospitals, ~500 bacteria for ESBL E. coli (namely ~2 000 in France, by extrapolation). By stratifying the data provided by Onerba, we can see (table 7) that the proportion of strains resistant to C3G within the E.coli strains isolated from the bacteria, that were 6% overall in 2008, were 3% in A&E and in gynaecology-obstetrics, 6% in general medicine, 7% in surgery and 10% in intensive care. This proportion was 4% when blood cultures were sampled during the first two days of the hospital stay, 8% during the rest of the 1st week, 9% during the 2nd week, 9% during the 3rd and 4th week and 14% after 1 month. This data shows that the proportion of resistant strains increases when the context suggests a nosocomial origin (blood cultures sampled in intensive care or later after hospital admission). They do not take into account any possible previous hospitalisations before arrival at the hospital, particularly when blood cultures were sampled in A&E or during the first two days of the hospital stay. The evolution of the E.coli bacteria resistant to C3G, essentially by ESBL, is in the forefront practically throughout Europe, but the proportions affected are still no greater than 5% in a dozen countries (Scandinavia, Switzerland, Holland, Belgium, Poland, etc.) whilst they are already greater than 10% in Italy, Israel, Greece and even over 20% in Rumania, Portugal, Bulgaria and Turkey (see figure 2 in annexe). The proportion of K. pneumoniae bacteria caused by strains resistant to C3G, essentially by ESBL, increased greatly in France between 2005 and 2008, and has reached almost 20% (see figure 3).

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

23

Types of enzym es involved In Europe, up to the end of the 1990s, most enterobacteriaceae ESBL were TEM or SHV type, whilst the main species producing enterobacteriaceae were K. pneumoniae and E. aerogenes (for example K. pneumoniae producing SHV-4 (5, 10)), E. aerogenes producing TEM-24 (4, 6)). This situation has changed radically over the last few years. Currently, E.coli strains producing CTX-M represent the majority of ESBL enterobacteriaceae and are found in all types of hospital services. Apart from North America where TEM and SHV type enzymes seem to still be in the majority, CTX-M are endemic to Latin America, Japan and certain Eastern European countries. In Western Europe, and particularly in France, we are witnessing significant spread of this type of ESBL (11, 12) (see figures 4 to 6 in the annexe). In South America, particularly in Argentina, the CTX-M-2 type enzyme is most often described in enterobacteriaceae and particularly in E.coli. In Asia, CTX-Ms have emerged since 1995 and have become predominant: CTX-M-9 and CTX-M-14 in Japan, CTX-M-14, CTX-M-13, CTX-M-9 and CTX-M-3 in China, CTX-M-14 in Korea and Vietnam and CTX-M-15 and -3 in India (7, 13). The most frequent CTX-M type ESBLs are CTX-M-15, before CTX-M-1 and CTX-M-3 in several European countries (3, 14). In other countries, different types of enzymes are more often found in E.coli such as CTX-M-9 in Spain (15). Spanish, Hungarian, Finnish and also French studies also report significant epidemics linked to hospital closed of K. pneumoniae producing CTXM-15 (16 - 19). French epidemiology is no exception in terms of progressive predominance of CTX-M enzymes. These enzymes emerged in the 2000s particularly in the north of France. Between 2002 and 2003, CTX-M-15, CTX-M-3, CTX-M-10 and CTX-M-14 were predominant in E.coli in the Ile de France region (20). In 2005, CTX-M-15 was in the majority (55 %) among 123 consecutives strains of E. coli isolated in AP-HP hospitals, and this enzyme was also very frequent in K. pneumoniae (43 % of ESBL strains) whilst classic SHV or TEM enzymes remained the majority in E. cloacae and E. aerogenes (AP-HP network data). The presence of CTX-M-15 in K. pneumoniae was also reported in other French hospitals (17, 21). The study carried out in 2004 in Champagne-Ardenne showed that CTX-M type enzymes represented 45% of ESBL in enterobacteriaceae; and in the E.coli species, CTX-M-15 represented 52% of the ESBL (22). In 2006, in the Eastern inter-region, E.coli represented 65% of ESBL enterobacteriaceae isolated from diagnosis targeted sampling within the framework of the BMR-Raisin survey (23); within this species, 91% of enzymes were CTX-M and 51% were CTX-M-15. CTX-M-15 also represented half the ESBLs from E.coli strains responsible for bacteria in an AP-HP hospital between 2001 and 2006 (24). However, CTX-M-15 is not predominant everywhere (25), even if we can see the emergence of CTX-M enzymes everywhere as in 2004 in hospitals in the South of France (26). So, in the Reims CHU and in 2007, among 159 consecutive strains for ESBL E.coli, CTX-M-9 and CTXM-27 were more frequently identified (24% each) than CTX-M-15 (17%) and classic SHV or TEM type enzymes were in the minority ( 80 years old) with demographic data corresponding to these subjects: age and place of residence (EHPAD). A study on prevalence of bacteriuria in hospitalised subjects over 75 years old, organised by the SPILF/SFG inter-group is in progress and will also probably bring answers to some of these questions. o

b/ Patients who have received an antibiotherapy prior to admission: this notion is difficult to take into account operationally. In an epidemic situation, it is required that the conduct to follow will be the same as should be adopted for other ESBL enterobacteriaceae: action against cross transmission associated with good use of antibiotics with consultation of the ES benchmark on prescriptions in the department involved. If a screening strategy is envisaged on admission, it must take into account the risk factors described in the first part: history of treatments with beta-lactam or fluoroquinolone, hospitalisation background, nosocomial context, old age, female, existence of co-morbidities, diabetes, repeated urinary infections, urinary probe, and gynaecological surgery. A screening strategy cannot be systematic: it can only spread in precise situations, in high risk departments hosting patients that also have a high risk of developing an infection in the case of colonisation of an MRB and particularly ESBL enterobacteriaceae. Outside health establishments (institutions; + / - home), several studies should be promoted to effectively tackle the problem issue of ESBL E.coli and work on writing up recommendations in the form of a technical file referring to existing documents, particularly to "recommendations to prevent cross transmission: extra contact precautions" (95). EHPADs and USLDs are a very important potential container for ESBL E.coli. At-risk elderly subjects living in a collective (USLD for health establishments and EHPAD or old people's home for the medical-social sector) number between 700,000 and 1,200,000; 30 to 50% of residents are, have been or will be affected by a bacteriuria that on the whole is due to E.coli: the ESBL rate has still not be given accurately. In the same way, digestive carrying of ESBL E.coli in these residents remains to be explored; it could affect 40% of them according to certain studies (96). The prevalence of ESBL enterobacteriaceae colonisations in EHPAD patients during their general medicine-surgery-obstetrics (MCO) transfer could be evaluated by a simple study involving systematic screening on admission, without systematically setting up constraining measures to deal with it whilst waiting for the results of each screening. On the face of the results from this kind of study, and if colonisation rates seem to be significant (e.g. over 10%), it could be proposed to set up certain measures to receive EHPAD patients in MCO (single room, extra precautions whilst waiting for the screening results) and to complete the study to try and identify other risk factors (e.g. look for a correlation between colonisation and care density - in EHPAD or elsewhere - between colonisation and dependence, etc.)

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

35

which, on a secondary basis, might restrict the population who would be systematically offered screening and adapted care. It therefore seems to be particularly important to analyse the problem issue of ESBL E.coli in EHPADs. It is best to answer certain specific questions during and outside an ESBL E.coli epidemic : - the role of "standard" precautions? - potential need of setting up complementary precautions and if so, when, and which ones...? - need to monitor residents who are known to be carriers (97). These recommendations were shaped into technical files centred on managing stools, urine, sheets, clothing, dirtied environment, etc. technical files intended to complete the document entitled "Prevention of cross transmission" (see CClin Paris North file "Managing excreta in health and medical-social establishments" presented in the annexe). The problem of returning home for a patient who is infected or colonised by ESBL E.coli also requires a specific approach - to answer questions from the people around them, from the family, professionals who will attend the patient in their home. For example, what is the role of AHRs in terms of simply washing hands with soap and water? These are increasingly accessible, in pharmacies, in mass distribution. Recommendations on their use, for the general public, should doubtlessly be envisaged. Technical files are also expected to answer questions on managing laundry, waste, the environment, etc. The last point covered concerns pregnant women and newborns. In maternity hospitals, it has been shown that isolated strains in mother and child were often clonal whilst the strains identified in several mothers were more often diverse. Nevertheless, epidemic phenomena can occur. The population of pregnant women is a young population who has had little previous contact with the care system; for this reason, finding out the prevalence of a possible carrying of CTX-M type ESBL E.coli in this population is very interesting in itself. Nevertheless, this young population, when it has taken antibiotics, has essentially received fluoroquinolone for urinary infections; on that point, this population appears to be particularly interesting in terms of monitoring. For a young population with previous history of hospitalisation (military), the carrying rate was evaluated at around 0.2%. Research on CTX-M ESBL E.coli between the 34th and 37th week of pregnancy has already been done in France in a few centres (E. Bingen) and showed that the rate of CTX-M ESBL E.coli isolated from vaginal samples rose from 0 to 3.5% over the last few years. A survey on hospital labs carried out in Western France showed that CTX-M ESBL E.coli infections remain exceptional in pregnant women, but an increase has been observed in the number of colonisations between 2007 and 2008. In pregnant women, and regarding the national recommendations on tackling bacterial risk during pregnancy (http://www.has-sante.fr/portail/upload/docs /application/pdf/Antenatalprevention.pdf), the circumstances for discovering CTX-M ESBL E.coli are as follows:

In the community or in out-patients:

High Council for Public Health

36

-

-

in urine, when screening by ECBU for asymptotic bacteriuria, or during microbiological documentation of a low lying urinary infection. During isolation of an ESBL enterobacteriaceae (+), the treatment should take into account possible associated resistances and the stage of the pregnancy. in the vaginal flora (vaginal sample), the fortuitous discovery of carrying CTX-M ESBL E.coli can appear during targeted screening of bacterial vaginoses recommended in the 1st term, or to document a vulvovaginal pathology (mycose, vaginose and T. vaginalis infection). In these circumstances and in the absence of risk factors (waters breaking early - RPM), threat of premature labour (MAP), it is currently recommended to treat the initial pathology but not try to "treat" the carrying of bacteria with associated risk for the mother-foetus or newborn involved, regardless of its nature, with a generally administered antibiotherapy in as much as this attitude has been revealed to be barely effective and generates resistance. Should the presence of CTX-M ESBL E.coli in the vaginal flora in these circumstances change this attitude? Taking into account, on the one hand, the low current prevalence of this carrying and on the other hand, the absence of accurate data on the possible specific maternal-foetal or neonatal consequences from these bacterial clones, there is no immediate argument to set this type of carrying apart.

In the maternity hospital, in the hours following hospitalisation : when researching carrying bacteria with high infectious risk in the vaginal flora during RPM. Initial systematic prophylactic antibiotic treatment is only currently recommended in the event of RPM before 34 weeks and possibly between 34 and 37 weeks depending on the obstetrical context (amoxicillin or cephalosporin if there are microbiological or anamnestic arguments) as from 37 weeks, it is recommended to trigger the work. Carrying ESBL E.coli (+) should mean that the antibioprophylaxy should be readapted. in urine (+/- blood cultures) during gravid pyelonephritis. The recommended attack treatment calls on associating C3G + aminoglycosides and is readapted in 48 hours. in the endocervix (+/- blood cultures) during chorioamniotites, obstetrics emergency. The chorioamniotite treatment is more often probabilistic (amoxicillin and/or C3G are recommended depending on the severity of the clinical symptoms and the circumstances it appeared in). In this pathology, the appearance of ESBL E.coli (+) in a vaginal sample carried out beforehand (for example for RPM +/-MAP) means that the treatment can be adapted straightaway. As a consequence, in the population of pregnant women, many questions are raised by the fact that we now isolate CTX-M ESBL E.coli strains in urine and in vaginal flora less exceptionally: What are the risk factors (antibiotherapy before and during pregnancy, trips, geographic origin, diet, etc.)? What are the potential consequences of this type of colonisation? What is the maternal risk when waters break early or for premature labour and what is the impact on the antibioprophylaxy prescribed in these circumstances? What is the risk of chorioamniotitis and the impact on the treatment of this obstetric emergency? What infectious risk is there for the newborn and what difficulties for care given the resistance of the bacteria in the event of neonatal infection?

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

37

-

What is the impact on natural colonisation of the newborn's digestive flora: imbalance of the flora and potentially harmful and lasting effects for the child?

To help to answer these questions and faced with poor data, there is a need for complementary data: The following could be envisaged: documenting the prevalence of carrying (vaginal +/- rectal?) CTX-M ESBL E.coli in women arriving at a maternity hospital to give birth and the obstetrical and paediatric consequences of this carrying. The fact that this is a "captive" population should make it easier to organise these studies; monitoring the prevalence of CTX-M ESBL E.coli demonstrated in urine during pregnancy; setting up a monitoring system that monitors the prevalence of CTX-M ESBL E.coli in the vaginal flora of pregnant women in order to regularly judge the evolution of the phenomenon. This monitoring system could use the vaginal sample recommended for screening group B streptococcus carried out between the 34th and 37th week. The limitations to this proposal come from the fact that vaginal samples are not the best examinations to search for ESBL enterobacteriaceae carrying. Associated rectal swabbing would be more interesting. The comparison of results from the two samples would furthermore allow us to judge the CTX-M ESBL E.coli (often from group ECORB2) capacity to adapt to the vaginal environment in women with rectal carrying. Upstream of these studies, it would be advisable to look at the right attitude to adopt when a pregnant woman accepting to participate is diagnosed as a "carrier", by properly defining the surveillance or care methods for this specific carrying depending on the different obstetric circumstances encountered. In fact, abusive prescription of antibiotics should be avoided whilst assuring patient and newborn safety. Concerning the problem issue of newborns, cases of infections remain exceptional, although the demonstrated number of colonisations is growing. This seems to be a cross transmission problem, post-natal, within maternity hospitals. Today, a great effort is made in these establishments, by means of development care, to encourage the mother-child relationship. This might facilitate cross transmission of some bacteria and increase the risk of faecal-oral transmission. Support for parents in terms of abiding by hygiene measures should therefore be developed. Working from the analysis of vaginal samples, the rate of colonisation for E.coli in pregnant women is 10 to 12% in France. This colonisation would affect 50% of newborns from colonised mothers and so around 5% of the total number of births (840,000 per year). The number of newborns colonised by an ESBL E.coli remains unknown but in some newborns colonised at birth the risk of developing an early infection does not seem any different from what is seen during colonisation by other E.coli. In newborns, the incidence of maternal-foetal E.coli infections, including all antibiotic sensitivity profiles, is 2 to 4 for every 1000 births, higher in premature births. In the event of suspecting an ESBL E.coli infection, the use of carbapenems (outside AMM before 3 months) must be envisaged, particularly depending on the local epidemiology. It is essential to monitor these infections regionally or nationally.

High Council for Public Health

38

TREATING INFECTIONS - TACKLING COLONISATIONS GOOD USAGE / MISUSE OF ANTIBIOTICS Introduction Human E.coli infections, regardless of whether or not the strain produces an ESBL, essentially involves the urinary apparatus, with very different symptoms ranging from ordinary simple cysts that could disappear without any antibiotherapy to uro-sepsis that is life threatening. E.coli is also responsible for other clinical symptoms, particularly intraabdominal infections and neonatal infections. In all cases, the container for these E.coli infections is digestive, with a perfectly documented aggravated risk of resistant strain, particularly by production of ESBL, in the case of prior antibiotherapy having modified this ecosystem. Looking at antibiotics and ESBL E.coli therefore comprises several areas: - What prevention should be used to limit as much as possible the risk of digestive carrying of ESBL E.coli that could secondarily be responsible for infection? - Which probabilistic antibiotherapy should be offered for a PRESUMED E.coli infection and from what level of prevalence should the ESBL risk be integrated? - What corrective antibiotherapy should be proposed for a documented ESBL E.coli infection? - What treatment strategy should be employed for digestive ESBL E.coli colonisation? 1 - What prevention should be used to limit as much as possible the risk of digestive carrying of ESBL E.coli that could be secondarily responsible for infection? Looking at good usage of antibiotics The emergence and diffusion of ESBL E.coli are encouraged by antibiotherapies. Bacterial resistance is the main undesirable effect of antibiotics at an individual and collective level! In addition, in order to reduce the selection pressure due to the antibiotherapy, it is advisable to progress in "good use" of antibiotics, by evolving towards a "reasoned" use or more clearly towards "less use". It is advisable to give details of and raise awareness on situations justifying a non-prescription antibiotic. These situations must be given referring to the urinary sphere, intra-abdominal complaints but it is advisable to go further by also particularly reconsidering all complaints in the high or low respiratory tract (bronchitis, sinusitis, otitis, chronic, obstructive acute broncho-pneumopathy exacerbations, etc. It is advisable to invite knowledgeable agencies and societies to attack this problem issue, and to give details, during a consensus conference or in the form of recommendations for clinical practice, of these situations justifying non-prescription of antibiotics - it is then necessary to assure diffusion of these recommendations and their administration by prescribing clinicians. Each gram of antibiotic prescribed, each gram of antibiotic consumed, induces a selection pressure on the bacteria in the digestive sphere and helps bring about the emergence, the concentration and the spread of resistant bacteria. It has been established that the more we prescribe antibiotics and/or the more we prescribe an antibiotic for a long time, the more we make a damaging impact on the flora. Needless to say that it is difficult, based on data in the literature, to establish a precise hierarchy of antibiotic molecules most at risk regarding the risk of intestinal carrying of ESBL E.coli; prescriptions of C3G and fluoroquinolone appear Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

39

particularly involved, in the community and in hospital; inversely, molecules with a narrow spectrum (penicillin V and M, etc.) are presumed to have less selection risk compared to enlarged spectrum molecules (other penicillin, etc.). Less use of antibiotics is possible and has already been shown within the framework of tacking respiratory infections in the community in France, following the health insurance campaign "antibiotics are not automatic". However the emergence and diffusion of ESBL E.coli demonstrate the limitations of the current plan. It is therefore advisable from now on to go one step further, beyond the single respiratory field, in all situations where antibiotics are used today. At the same time, the reflection must affect the recommendation of an antibiotherapy by clearly stating the situations in which non-prescription is recommended and when an antibiotherapy recommendation is given; it is advisable to give its optimum spectrum and its duration. Dogma might fall, such as systematic antibiotherapy for simple cystitis leading to a medical consultation. A recent study run in England compared different strategies to tackle simple cystitis: antibiotic consumption was lower when the treatment was guided by a urinary strip (80% effective antibiotherapy exposure) or when patients were asked to follow a symptomatic treatment (77% exposure) compared to systematic treatment (97% exposure) or guided by a clinical score on the severity of the symptoms (90% exposure); the duration of the symptoms did not differ between strategies (98), the strategy of the antibiotherapy differed at the subject's initiative was judged positively by 10 patients out of 13 who participated in a specific evaluation questionnaire (99). Since 1991, the French recommendation has been to treat simple cystitis based on a positive result from the urinary strip (100), however, recent studies have shown that this practice is inconsistent in routine; one of the difficulties results, perhaps, from not reimbursing the strip. Developing the use of the urinary strip in compliance with the recommendations could be a first pragmatic step in limiting the antibiotherapy for cystitis symptoms. When an antibiotherapy is recommended, choosing the least "selective" spectrum possible is conceptually interesting. The reflection is very wide-reaching, encompassing all prescriptions as, regardless of the target organ, the antibiotherapy can modify the digestive flora (including for schemas administered parenterally for all digestive secretion molecules). Now, particularly for bio-availability reasons, current French recommendations make room for penicillin A compared to penicillin V and M, including targeting infections that do not implicate negative gram bacterial infections: antibiotherapy by amoxicillin for streptococcus tonsillitis (according to data from a TDR), erysipelas, presumed pneumopathies with pneumococcus… Other countries have used different strategies with, for example, a wider use of penicillin V and M and macrolids in Northern European countries that have lower overall presence of antibioresistance. There is doubtlessly a significant field of research and evaluation there. 2 - Which probabilistic antibiotherapy should be offered for a PRESUMED E.coli infection and from what level of prevalence should the ESBL risk be integrated? The latest French data published (101) is, from now on, in favour of a prevalence of around 2% of community E.coli strains that are resistant to C3G: among 1636 urinary E.coli strains collected from November 2007 to October 2008 by three private labs in Normandy, 34 (2.1%) came back intermediary or resistant to C3G, of which 21 (1.3%) by production of ESBL (20 CTX-M).

High Council for Public Health

40

Furthermore, the EARSS 2009 results relating to hospital identification bacteraemia bring about the appearance of a rate of over 5% of ESBL strains among the E.coli involved in these infections. Community ESBL E.coli are characterised by very frequent multiresistance, affecting cotrimoxazole (70%), fluoroquinolone (70%) and, to a lesser extent, aminoglycosides (amikacin: 20-30%; gentamicin: 40-50%, tobramycin 50-60%); on the other hand, these strains remain very sensitive to fosfomycin and to furans (less than 5% resistance). As far as intra-abdominal infections are concerned that might be due to ESBL E.coli, existing recommendations in France are now old (2000-2002) and must absolutely be reconsidered in the short term. The same goes for recommendations relating to neonatal infections where E. coli represents the second germ in terms of frequency (after urinary infections). A reflection on conduct to follow for urinary infections is given in detail below. In probabilistic antibiotherapy, we should take into account the probability that the infection is either E.coli or that it is an ESBL strain, and finally that the isolate has possible associated resistance. So, the probability that a case of pyelonephritis is finally due to an ESBL E.coli strain resistant to aminoglycosides is currently 0.8%: 0.80 (risk that the infection is due to E.coli) x 0.02 (probability that E.coli is ESBL) x 0.50 (probability that E.coli is also resistant to aminoglycosides). We should also appreciate the loss of opportunity generated by initial antibiotic abstention or a non-adapted probabilistic choice. So the presence of an ESBL E.coli does not bring about any particular difficulty in the majority of clinical situations represented by urinary colonisations and non parenchymatous urinary infections: - urinary colonisations (or asymptotic bacteriuria) come from antibiotic abstention, except in rare situations (pregnancy, programmed surgery, etc.) for which the treatment is straightaway guided by the antibiogram (see infra); - simple cystitis is treated at the 1st attempt by fosfomycin-trometamol, a molecule that currently remains very active on ESBL E.coli; even if this fails, it will have no consequence; - complicated cystitis must be subject, whenever possible, to documented antibiotherapy straightaway (see infra). A delay in setting up an adapted treatment means losing the opportunity for pyelonephritis and prostatitis. The recommended best practices (RBP) given by Afssaps in 2007 to deal with community urinary infections in children kept a C3G (parenteral) in a benchmark probabilistic treatment of pyelonephritis; the RBP in 2008 relating to dealing with community urinary infections in adults kept a C3G (parenteral) or a fluoroquinolone (per os or parenteral) in benchmark probabilistic treatment of pyelonephritis and prostatitis, taking into account a benefit-safety balance for satisfactory use both individually and collectively; at this time, the ESBL rate among community E.coli strains was 1%. These RBPs envisage adding an aminoglycoside for more severe forms: severe sepsis, pyelonephritis over an obstacle, newborns and babies under three months, etc.; this addition partly secures the risk of failure in the case of ESBL E.coli, with the French strains remaining sensitive to aminoglycosides in around 50% of cases, whilst several series of literature prove the efficacy of an aminoglycoside singletherapy in treating urinary infections from sensitive bacteria. So today, with 2% of Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

41

community E.coli producing ESBL, it seems reasonable to hold to these same recommendations: C3G or, in adults or pubescent adolescents, fluoroquinolones in probabilistic treatment of pyelonephritis and prostatitis, with the addition of an aminoglycoside only for severe forms: gentamicin or amikacin intravenously in a daily 30 minute drip, for 3 to 5 days. In time, if the ESBL E.coli endemic gets worse in France, the pyelonephritis/prostatitis reference could essentially evolve in two directions: - systematic addition of an aminoglycoside until receiving the antibiogram; these antibiotics have the advantage of being available in pharmacies; on the other hand, it would then be best to manage their potential nephro and oto-toxicity, which can be complicated in the community (IV drip lasting thirty minutes, need for plasma doses in the event of extended treatments, etc.); - substitution of the C3G or the fluoroquinolone with a carbapenem; this type of evolution towards a wide probabilistic prescription of carbapenems would however expose us to a greater ecological risk: allowing new resistance mechanisms to emerge, even in the community, to one of the last resorts available; furthermore it can raise practical problems in the community relating to the administration of these injectable antibiotics. When should we envisage reviewing the current recommendations? For all therapeutic recommendations tackling infections potentially due to E.coli, it would be interesting to ideally free up individual risk factors for acquiring these strains in order to delay abandoning current strategies as much as possible and the potential evolution towards a wide resort to carbapenems, taking into account the associated ecological risks. For the time being, we can only indicate the need to follow how resistance rates evolve and the kinetics of this evolution in order to appreciate the right time for this review. 3 - What corrective antibiotherapy should be proposed for a documented ESBL E.coli infection? The treatment of severe ESBL E.coli infections raises a problem due to frequent multiresistance. We can see worsened mortality in the case of an inappropriate regime. In practice, this essentially involves IU parenchymatous (PNA, prostatitis), intra-abdominal infections, a fortiori in cases of severe associated sepsis. A recent review of the literature available in "The Lancet Infectious Diseases" (63) reminds us of how poor the clinical data in this field is (absence of randomised tests, small series reported, with sometimes diverging results). The in vitro resistance to C3G is seen clinically as an over-incidence of failure in treatment by cefotaxime, ceftriaxome, or ceftazidime (102-104). For the time being, C3Gs must not be used 2, carbapenems represent the benchmark treatment; alternatives are few and far between and have been poorly validated. Concerning carbapenems, it is advisable to specify that they are not interchangeable and that there are ertapenem-resistant and imipenem-sensitive strains From the perspective of a parenchymatous infection to deal with in the community, different problems are raised: o

2

Even if some experts (105) question the opportunity of considering isolates expressing an ESBL but presenting relatively low CMIs for C3Gs (< = 1) as sensitive.

High Council for Public Health

42

- the use of carbapenems, far from being ideal, must be seen as a "false good solution": this is an effective solution therapeutically on an individual scale but a high risk solution to encourage the development of carbapenemases (worthy risk on individual and collective scale). This is therefore a treatment that should only be envisaged with precaution, by evaluating the safe use/ expected benefit ratio case by case; - this only refers to products for parenteral administration, for hospital prescription; - to date, a single product can be administered every 24 hours (ertapenem) but this does not have AMM in Europe for urinary infection (as opposed to the United States). Reflection points for possible molecules to be used when tackling ESBL E.coli infections Today, it is advisable to take a good look at the molecules that could be used when tackling ESBL E.coli infections whilst already recognising that none of the proposals will be ideal. o

This is not a matter of recommendations or therapeutic strategies validated today, but strands of research that might constitute study paths (94). a/ Fosfomycin-trometamol and nitrofurantoïn (molecules reserved for less severe infections). Current French isolates remain most often sensitive to fosfomycin and nitrofurantoïn. With these two molecules, eradication rates expected today are as good regarding ESBL enterobacteriaceae as E.coli that does not produce ESBL. So, among 28 patients suffering from EBLSE cystitis treated by a single 3g dose of fosfomycin-trometamol, 26 (93%) were cured (106). However, this does not exclude a few failures as already reported. We must remember that if furans are recommended according to the RBP for both simple (in 2nd attempt treatment) and complicated (1st attempt) cystitis, fosfomycintrometamol is only recommended for simple cystitis (1st attempt and from then on without age restrictions in women over 65 without co-morbidity). Long regimes of fosfomycin-trometamol in treating complicated cystitis with ESBL enterobacteriaceae could be studied in order to find an alternative to furans and spare the parenchymatous urinary infection treatments using carbapenems and aminoglycosides (93). A series of 52 cases colligated in Turkey demonstrate the efficacy of administrating fosfomycin-trometamol over several days (3 g/day for 3 days) in treating complicated cystitis with ESBL enterobacteriaceae, with clinical and biological eradication rates of 94% (49/52) and 79% (41/52) respectively (107). So, longer administration regimes for fosfomycin-trometamol would be likely to bring better eradication rates, and the safe usage profile of the molecule would also envisage this. However, an increase in resistance to fosfomycin in parallel to growth in consumption was described in certain countries (108), particularly in the United States with, among 46 community isolates for ESBL enterobacteriaceae, 9% resistance to fosfomycin (despite very limited use of the product in this country) and 25% resistance to furans (109); showing the importance of close surveillance on cross resistances in France. b/ Cephamycins (cefoxitine) A risk of resistance by "impermeability" was described very early in some ESBL enterobacteriaceae, particularly in K. pneumoniae producing TEM or SHV (110) but it is rarer in E.coli; nevertheless, very few uses in the clinic have been reported for ESBL enterobacteriaceae. Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

43

c/ Temocillin Studies carried out in vitro show over 90% of ESBL enterobacteriaceae sensitive to Temocillin (111-113). d/ Cephalosporin associations (see monobacatames or mecillinam) + betalactamase inhibitor These associations have potential interest. The most effective beta-lactamase inhibitor is clavulanic acid but it only has an AMM in France in association with amoxicillin or ticarcillin. The inhibitor that has benefited from an AMM in France, without association with an antibiotic, is sulbactam. Although this particular feature can offer an advantage, the poor efficacy of sulbactam presents a greater disadvantage. The ceftriaxome + sulbactam association has already been the subject of clinical studies. The paths that it seems should be explored in France involve associating a C3G particularly with clavulanic acid (remove using the amoxicillin + clavulanic acid association to get clavulanic acid). The cefoperazone-sulbactam association (Sulperazon®) is currently being researched in the United States and in Asia Sanford 2010). In the same way, ceftriaxome + sulbactam and cefepime + tazobactam associations are available or being researched in India and it will be interesting to find out about their efficiency in treating ESBL E.coli infection. The interest of Cephamycin + beta-lactamase inhibitor associations must be evaluated, as cephamycins have greater intrinsic activity than C3G. Monobactame + beta-lactamase inhibitors should be studied because these associations (a) present an alternative for patients who are allergic to carbapenems; (b) exert lower selection pressure than seen with carbapenems. e/ Piperacillin + tazobactam association The interest of the piperacillin + tazobactam association still has to be specified. If studies have shown that use of this association in non complicated urinary infections leads to therapeutic success even in the presence of ESBL E.coli, cases of failure have been reported secondarily (108). f/ Tigecycline Tigecycline is active in vitro regarding strains (115); although it obtained the AMM in treating complicated intra-abdominal infections, the data remains limited for more severe patients (with bacteraemia, co-morbidity or severe prognostic score); no AMM has been awarded in treatment of urinary infection; if renal excretion is in the minority, some clinical successes have been reported in treating uro-sepsis with ESBL (116). g/ Colimycin The use of colimycin must preferentially be reserved for treating multiresistant negative gram bacterial infections (particularly bacteria secreting carbapenemases) and being uniquely sensitive to colimycin. It is also very complicated to match: hospital prescription, questioning the optimal posologies, potential toxicity, etc.

High Council for Public Health

44

4 - What treatment strategy should be employed for digestive ESBL E.coli colonisation? No element today speaks out in favour of setting up systematic decolonisation of people carrying ESBL E.coli. The interest in setting up a policy of decolonising carrier patients in at-risk departments and/or in the case of an epidemic remains to be discussed. Setting up clinical research studies in the field must be encouraged.

Preventing the emergence of ESBL enterobacteriaceae and fighting their spread/February 2010

45

ANNEXES

High Council for Public Health

46

Table 1: Evolution of the incidence of EBLSE for 1000 HD from 2004 to 2008 Analysis restricted to establishments that participated in BMR-Raisin surveillance for each of the five years (n=274). Stratification by type of department.

EBLSE Incidence for 1000 HD

Establishment MCO (including psy) of which intensive care SSR-SLD All

Evolution

2004 0.19

2005 0.23

2006 0.22

2007 0.33

2008 0.40

p*