The Burden and Genetic Variability of Extended-Spectrum ß-Lactamase (ESBL) - Producing Pathogens in Swiss Children

Background \& Rationale:

Increasing resistance to antibiotics is a growing problem in public health and is associated with treatment failures, increased health care costs, and prolonged hospital stays. Over the past years, one mechanism of resistance has been of particular concern: ESBL, which is produced mainly by Escherichia coli and Klebsiella species but also found in other gram negative bacteria. ESBLs are plasmid-encoded ß-lactamases conferring resistance to penicillins, cephalosporins and aztreonam.

ESBL-producing enterobacteria are usually susceptible to carbapenems, such as ertapenem, imipenem or meropenem. Thus, treatment with carbapenems remains safe if the presence of ESBL is recognized. However, the confirmation of ESBL is not always a routine procedure in all laboratories. Furthermore, ESBL strains with high resistance to ertapenem and decreased activity to imipenem and meropenem have already been described. The mechanism among these strains however, appears to be a combination of the expression of ESBL and impermeability or increased efflux for carbapenems. Infection with ESBL-producing E. coli or K. pneumoniae is potentially hazardous for patients. Mortality is significantly higher following a bloodstream infection caused by ESBL producing E. coli compared to non-ESBL producing strains.

ESBLs are encoded on plasmids: classically, these plasmid mediated enzymes have been of the TEM and SHV type. However, in recent years CTX-M-type ESBLs have been increasingly identified worldwide throughout the world. CTX-M enzymes predominantly hydrolyze cefotaxime, but are weakly active against ceftazidime. However, some ESBLs of the CTX-M family also display increased hydrolytic activities against ceftazidime, as is the case for CTXM-15 and CTX-M-32.

The prevalence of ESBL producing enterobacteria and particularly E. coli has significantly increased in recent years. This increase is primarily due to the spread of ESBL of the CTX-M-type. Genes encoding CTX-M enzymes are associated with mobile genetic elements allowing these enzymes to spread throughout the community. Well established primer sets are used in polymerase chain reaction (PCR) tests for the detection of the blaTEM, blaSHV, and blaCTX-M genes, and are used to differentiate ESBL strains.

ESBL-producing organisms were first detected in Europe in the 1980s. Although the initial reports came from Germany, the vast majority of descriptions in that first decade came from France. Consequently, it was not surprising that the first large outbreak occurred in France in 1986. Other outbreaks with ESBL-producing organisms have now been reported from almost every European country. In southern Europe, the prevalence of ESBL in hospitals is estimated around 25% for Klebsiella species, while in the community it is 1.7% for E. coli, and 4% for K. pneumoniae. In the USA, ESBL-producing organisms were first reported in 1988. Similarly to Europe, these strains expanded over the following years, and in 2002, 6.1% of the K. pneumoniae isolates were found to be resistant to third-generation cephalosporins. Furthermore, in at least one tenth of intensive care units, ESBL producing K. pneumoniae quickly exceeded 25% of all isolated strains. In contrast, in regular inpatient areas prevalence rates remained around 5.7% of all isolated K. pneumoniae. ESBL strains have also been documented in South Africa, Israel, Saudi Arabia, and a variety of North African countries. In South America, the SENTRY antimicrobial surveillance program recovered ESBL in up to 10% of E. coli and 55% of K. pneumoniae. National surveys in Asia have revealed the presence of ESBL-producers in 5 to 8% of E. coli in Korea, Japan, Malaysia, and Singapore and in 12 to 24% in Thailand, Taiwan, the Philippines, and Indonesia. The prevalence in China ranged between 27% and 34% for E. coli and up to 38.3% for K. pneumoniae already in 2002.

This worldwide distribution is reflected by the fact that ESBL producing enterobacteria are present in the community, and in adult as well as in children's hospitals. However, in this younger population, ESBL-producing strains are isolated not exclusively but essentially in neonatology, where various outbreaks have been described.

Studies in neonates in India revealed ESBL-strains in half of the infants with early-onset sepsis, and in 82% with late-onset sepsis. A study of blood samples from suspected cases of neonatal sepsis detected ESBL in 87% of Klebsiella spp., 73% of Enterobacter spp. and 64% of E. coli strains.

If considering resistance to 3rd generation cephalosporins as an indicator for ESBL, the prevalence of ESBL producing E.coli and Klebsiella pneumoniae among children in Switzerland was estimated 2.2% and 4.3% in 2006, respectively (www.search.ifik.unibe.ch). The frequency of ESBL producing enterobacteria in Geneva in the same year was 1.8% for E. coli and 4.9% for Klebsiella spp. (personal communication Dr. P. Rohner). In the Department of Paediatrics in Geneva, the majority of ESBL producers were found among children from "Terre des Hommes", a non-profit humanitarian organization, which brings African children to Switzerland for cardiac surgery. These children all had a history of multiple hospital stays in their country of origin, and possibly repeated antimicrobial therapies. However, most ESBL strains can be expected from urinary tract infections on a national basis. The new recommendations of the entire oral treatment with 3rd generation cephalosporins without initial intravenous application for uncomplicated febrile urinary tract infections in children will be effective by 2008. Therefore, most children will be treated fully on an outpatient basis by their paediatricians or general practitioners. Treatment over all might increase by the simplified therapeutic procedure and some general practitioners might switch from trimethroprim-sulfamethoxazole to 3rd generation cephalosporins.

Risk factors for ESBL colonization or infection are well defined and similar to other resistant microorganisms such as methicillin-resistant S. aureus (MRSA): serious illness with prolonged hospital stays, invasive medical devices (urinary catheters, endotracheal tubes, central venous lines), and repeated or prolonged antibiotic use, especially 3rd generation cephalosporins. Selective antibiotic pressure is clearly associated with the emergence and dissemination of ESBL-producing enterobacteria. Children with bloodstream infections due to ESBL producing K. pneumoniae were almost 6 times more likely of prior exposure to an extended-spectrum cephalosporin in the 30 days before infection.

Enterobacteria can be transferred by direct contact or through fomites. Various outbreaks have been described incriminating sources such as other patients, artificial fingernails or ultrasonography gel. The identification of ESBL-producing bacteria should prompt the use of isolation and contact precaution measures. The application of such measures together with antibiotic stewardship was successful to overcome most outbreaks with ESBL-producing enterobacteria. However, the success of infection control measures depends largely on the recognition of risk and of carrier states. Therefore, not only infected children should be tracked but colonized patients as well, as they are capable of transmitting ESBL-producing bacteria to vectors such as health-care workers, or directly to other patients.

Finally, additional costs and use of resources are important and ESBL infections are expensive. Average infection-related costs per patient are significantly greater for ESBL patients than for control patients without ESBL. Additional costs attributed to the infection of an ESBL producer were estimated to 16'450 $ per patient and hospital stay. This cost increase was mainly due to increased length of stay.

The prospective surveillance of ESBL-producing enterobacteria in Switzerland is essential because ESBL-producing enterobacteria represent already a threat for the patients. This includes delayed treatment and therefore cure, and increased health costs. ESBL will probably be even more of a concern if no action is taken now in order to decrease its prevalence.

Methods:

All children aged 0 to 16 years admitted to a paediatric ward and either colonized or infected with an ESBL-producing strain will be included in the study. Identification of ESBL-producing strains will be performed in local microbiology laboratories using the local available detection methods. All strains will then be collected at the investigation centre for microbiological confirmation of ESBL-production and analysis of the genes associated with ESBL phenotype by sequencing. Because microbiological methods differ between laboratories, the investigation centre will therefore directly contact the local microbiology laboratories in order to obtain information about specific methods used to isolate the ESBL strains and feedback will be given after examination of the strains at the study centre. Hospitals reporting cases to SPSU will receive a case report form (CRF) for each patient in order to get patient data, medical history and implemented infection control measures. Strains isolated as potential ESBL should be stored and some colonies be sent to the study centre for microbiological analysis. The shipment will be organised and paid by the study centre. The shipment of stains is on a voluntary basis but recommended for quality control and to analyse the molecular resistance mechanisms. In order to ease logistic obstacles, hospitals are asked to only provide contact information of their respective laboratories; these however will be addressed directly by the study centre in order to obtain microbiological information and to organize shipment of the strains. Data management and entry will be performed by the study centre. This is an observational study. The study design with SPSU precludes any intervention.

Analysis:

Demographics will be reported by using standard descriptive statistics (mean, median, percentage) for the following variables: age (gestational age), sex, country of origin, last stay outside Switzerland, past hospitalisations, direct transfer from a hospital abroad, prior antibiotic use, and history of urinary tract infections. Risk factor analysis for ESBL infection vs. ESBL colonisation will be provided using standard statistical methods such as logistic regression analysis for the following variables: age (gestational age), gender, main diagnosis, comorbidity, immune-deficiency, length of hospital stay, length of neonatal stay, length of intensive care unit stay, medical devices (central venous lines, endotracheal tubes, urinary catheters), and antibiotic use (especially third generation cephalosporins). Furthermore, information about isolation precaution measures and outbreak investigation will be collected.

Case Definition:

All children (age range: birth until 16 years of age) colonized or infected with any ESBL-producing enterobacteria according to antibiotic testing of local microbiology laboratory, will be included. All isolated strains will be sent to the University Hospitals of Geneva's Microbiology Laboratory (Head: Prof. J. Schrenzel) for confirmation by specific ESBL culture techniques (double disc diffusion technique) and for identification of resistance genes by in-house sequencing. There are no exclusion criteria.

Ethical issues:

This is a descriptive study without having any direct effect on patient care. All information obtained in this study is already documented in the patients' chart and therefore no signed consent form will be asked. All data will be analysed anonymously and stored in a safe, password-protected computer. We will obtain an approval from the Ethical Committee of the University Hospitals of Geneva before the beginning of the study.