Infections Caused by ESbL-Producing Enterobacteriaceae in Italy

RESEARCH PLAN AND METHODOLOGY

Population All patients with infections (i.e. bacteremia, pneumonia, abdomen infections, skin infections, and urinary tract infections) caused by ESBL-producing Enterobacteriaceae.

Exclusion criteria Children < 16 years

Trial design Multicenter prospective cohort study. Patients corresponding to the study definition will be detected by daily inspection of microbiological databases by one dedicated physician. Epidemiological variables will be detected for all patients at study admission. To identify the risk factors for infections (i.e. bacteremia, pneumonia, abdomen infections, skin infections, and urinary tract infections) caused by ESBL-producing Enterobacteriaceae a case-control study will be performed. A case patient will be defined as adult patient ( 16 years of age) who has an infection caused by ESBL-producing Enterobacteriaceae according to the definitions of the Center for Diseases Control and Prevention in Atlanta (CDC). In the Clinical Microbiology Laboratories, definition of ESBL production by the Clinical Microbiology Laboratories will be according to the CLSI guidelines (3), or to the BSAC guidelines (http://www.bsac.org.uk) for species other than Klebsiella spp., E. coli and Proteus mirabilis. One control will be selected for each case. Control will be chosen among all adult patients admitted to the same hospital during the same period (within 30 days) and in whom ESBL-producing Enterobacteriaceae were not isolated during their hospital stay. If more than one control will be available per case, patients with the date and time of admission closest to the case will be chosen.

To identify the risk factors for mortality a second case-control study will be performed comparing infected patients who died to infected patients who survived (evaluable in patients with bacteremia, wound infections, pneumonia, and meningitis, only). This analysis will be performed adjusting the results for inadequate empiric antimicrobial therapy and site of infection (see following definitions).

Data collection Medical records of in-patient admissions and microbiology and pharmacy databases will be reviewed. Data collected at the study enrollment will include: patient demographics, transfer from another hospital, resident of a long term facility or nursing home, previous hospitalization within one year, ambulatory status, requirement for chronic hemodialysis, presence of a central venous catheter (CVC), and ICU stay and surgical procedures in the 30 days prior to study inclusion. A composite score of co-morbid illnesses will be derived using the Charlson score. The severity of illness at presentation will be quantified using the criteria of McCabe and Jackson. Antibiotics administered during a 30-day period prior to study enrollment and for at least 48 hours will be recorded. For the risk factors analysis, oral and intravenous antibiotic exposure will be analysed by individual antibiotics and by classes, and will include penicillins, vancomycin, cephalosporins, antibiotics with predominantly anaerobic activity (metronidazole and clindamycin), aminoglycosides, quinolones, and carbapenems. Length of stay (LOS) after infection diagnosis, management of infections, timing and type of antimicrobial therapy, results of follow-up clinical cultures (if any), hematogenous complications and death will be also extracted from medical records. Microbiology records will be also reviewed for recovery of vancomycin resistant enterococci (VRE) or methicillin-resistant S. aureus (MRSA) in the 12 months prior to the study enrollment. All patients with infections caused by ESBL-producing Enterobacteriaceae will be followed up to hospital discharge or death.

Mortality will be defined as death occurring during the study hospitalization. Appropriate antimicrobial therapy will be defined as the initiation of therapy with activity against the ESBL-producing Enterobacteriaceae (according to the results of the antimicrobial susceptibility pattern of the isolate) from the day before to 2 days after the initial positive clinical culture result.

Outcomes Primary outcomes

1. Risk factors for the development of infections caused by ESBL producers bacteria;
2. Risk factors for inadequate initial antimicrobial therapy;
3. Overall mortality and 30-day mortality among patients receiving inadequate initial antimicrobial therapy (evaluable in patients with bacteremia, abdominal infections and pneumonia, only);
4. Variability by site of infection of overall mortality and 30-day mortality among patients receiving inadequate initial antimicrobial therapy (evaluable in patients with bacteremia, abdominal infections and pneumonia, only).

Secondary outcomes

1. Lenght of hospitalization;
2. Costs of inadequate initial therapy;
3. Days of defervescence.

Statistical analysis At the end of the study two different analysis will be performed: the first one will consider patients with bacteremia, abdominal infections and pneumonia, and it will be focused on mortality; the second analysis will consider all infected patients and it will be focused on risk factor analysis and secondary outcomes.

Sample size justification Assumptions

Patients with bacteremia, pneumonia, abdomen infections, skin infections and urinary tract infections (first analysis):

Mortality among patients with inadequate initial therapy: 25% Mortality among patients with appropriate initial therapy: 13%

All included patients:

Percentage of bacteremia: 20% Percentage of pneumonia: 10% Percentage of abdomen infections: 15% Percentage of skin infections: 3% Percentage of urinary tract infections: 52%

Primary outcomes Risk factors for mortality Difference of mortality between controls (inappropriate therapy) and cases (appropriate therapy): we anticipated a reduction from 25% to 13%.

Secondary outcomes Difference of length of hospitalization between cases (appropriate therapy) and controls (inappropriate therapy): we anticipated a decrease from 80% (length of hospitalization < 15 days) to 40%.

Difference of defervescence between cases (appropriate therapy) and controls (inappropriate therapy): we anticipated a decrease from 85% (defervescence < 3 days) to 30%.

Assuming that p1=p2, where p1 is the proportion in population 1 and p2 is the proportion in population 2 and that alpha=0.05 (two-sided), power=0.80.

Required sample size for analysis of risk factors for mortality: 366 patients with bacteremia, wound infections, pneumonia, and meningitis. To include 366 patients with the above reported infections the total sample size is likely to be 813 patients infected with ESBL-producing Enterobacteriaceae.

Microbiological analysis of bacterial strains

All ESBL-producing isolates obtained from patients included in the study at the Clinical Microbiology Laboratories of each Clinical Center will be collected and subjected to further investigation including:

• confirmation of the identification at the species level;
• determination of the antimicrobial resistance profile against a standard set of drugs;
• analysis of the ESBL determinants;
• analysis of clonal relationship between bacterial isolates of the same species. Results of microbiological analysis will be used for correlation with clinical data to assess the epidemiological and clinical impact of the ESBL-producing strains.

Experimental methodology

Confirmation of identification of the bacterial isolates at the species level will be carried out by standard phenotypic methods (10). In case of ambiguity of results, identification will be confirmed by 16S rDNA sequencing (16).

Antimicrobial susceptibility testing will be carried out by broth microdilution as recommended by the CLSI (3), and categorical assignment will be carried out using the CLSI breakpoints (3). The following agents will be included: ertapenem, imipenem, meropenem, cefepime, ceftazidime, cefotaxime, piperacillin/tazobactam, amoxicillin/clavulanate, ciprofloxacin, levofloxacin, gentamicin, amikacin.

The nature of the ESBL determinants will be investigated by molecular analysis using PCR and sequencing, as described previously (9, 15). The presence of ESBL determinants of the TEM-, SHV, CTX-M- and PER-type (i. e. those known to be circulating in Italy (9, 15)) will be investigated. In case of negative results with the available probes, the ESBL phenotype will be reconfirmed and, if positive, the nature of -lactamase deerminants will be studied by means of: (i) analytical isoelectric focusing (IEF) on crude extracts using previously described methods (14); (ii) DNA microarray technology, using a microarray for β-lactamase genes that has been developed in our laboratory, which includes probes for uncommon ESBL genes such as VEB, GES, BES, SFO, and TLA enzymes. In case of new ESBL genes undetected by the above methods (an unlikely occurrence), In the case of production of ESBL but unaccounted for by the presence of known resistance determinants, the resistance determinants will be studied by means of a shotgun-cloning approach followed by mapping and sequencing.

Clonal relatedness among isolates of the same species will be investigated by high-throughput genotyping techniques such as RAPD (Random Amplified Polymorphic DNA) and/or by AFLP (Amplified Fragment Length Polymorphism) analysis and/or REP PCR (19-21). In selected isolates, the clonal relationships will be performed by macrorestriction analysis of genomic DNA by pulsed-field gel electrophoresis (PFGE).

Plasmid analysis and investigation of resistance determinants to non β-lactam agents (by molecular techniques) will be considered in selected cases, should these information be relevant for correlation with clinical and epidemiological data.