Risk factors for bacteremia due to extended-spectrum beta-lactamase-producing Escherichia coli in a Turkish hospital

Introduction: Bloodstream infection caused by extended-spectrum beta-lactamase (ESBL)-producing pathogens has become a serious concern worldwide. The purpose of this study was to identify risk factors for bacteremia due to ESBL-producing Escherichia coli in a Turkish hospital. Methodology: We retrospectively reviewed the medical records of patients with E. coli bacteremia in a tertiary care centre from January 2007 to October 2011. Data from patients such as demographic features, underlying conditions, and antibiotic exposure were analysed. Results: A total of 113 patients with bacteremia due to E. coli were included and data from patients with ESBL-producing E. coli (case patients) were compared to those with non-ESBL-producing E. coli (control patients). The frequency of ESBL producers was 38.9% (44/113). Exposure to fluoroquinolones and cephalosporins, history of intra-abdominal surgery intervention, and presence of central venous catheteter and urinary catheteter were more frequently detected among case patients (P < 0.05). Independent risk factors for bacteremia due to ESBL-producing E. coli were exposure to fluoroquinolones (OR 13.39, 95% CI 1.28-140.03) and cephalosporins (OR 3.48, 95% CI 1.0311.74). Conclusions: Previous use of fluoroquinolone and cephalosporin in patients with bacteremia caused by E. coli increased the risk for ESBLproducing strains.


Introduction
Extended-spectrum beta-lactamases (ESBLs) are a heterogeneus group of enzymes responsible for the resistance of enterobacteria to broad-spectrum betalactam antibiotics.These enzymes were first identified in Germany in 1983 [1].During the past two decades, antimicrobial-resistant strains producing ESBLs have emerged among the Enterobacteriaceae, especially among Escherichia coli and Klebsiella pneumoniae [2].ESBLs inactivate cephalosporins, monobactams and penicillins, while ESBL-producing microorganisms are also often resistant to fluoroquinolones and aminoglycosides [2,3].Infections caused by ESBL-producing microorganisms have limited treatment options, and when treated with inappropriate empirical therapy they are associated with increased mortality [4].Therefore, identification of risk factors that are predictive of resistance to empirically prescribed antimicrobials should facilitate attempts to define more effective management strategies for serious infections [5].We aimed to evaluate the risk factors for ESBL-producing bacteremia and the impact of ESBL production on outcome in patients with E. coli bacteremia.

Methodology
This retrospective study was conducted at Suleyman Demirel University Research Hospital, Isparta, which is a 550-bed medical centre facility located in Southern Turkey.Hospital records were searched to identify all adult inpatients ( ≥ 18 years old) diagnosed with bloodstream infection caused by ESBL-producing and non-ESBL-producing E. coli strains between January 2007 and October 2011.Cases of bloodstream infections caused by ESBLpositive/negative E. coli in at least one blood culture specimen from a patient with systemic inflammatory response syndrome (e.g., fever, tachycardia, tachypnea, and leukocytosis) were considered.Each patient was included as a case patient only once.The date of the first positive blood culture was regarded as the date of bacteremia onset.If E. coli strains were isolated on multiple occasions, only the first episode of infection was considered.Controls were identified among hospitalized patients with bloodstream infection who were infected with ESBL-negative E. coli during the same period.Antimicrobial susceptibility testing and ESBL confirmatory testing were performed using the disk diffusion method following the recommendations of the Clinical and Laboratory Standards Institute [6].Hospital-onset bacteremia was considered when the infection occurred more than 48 hours after admission to the hospital, while community-onset bacteremia was defined as infection that occurred less than 48 hours after admission to the hospital.Data obtained included age, sex, underlying conditions, primary source of infection, immunosuppressive treatments, presence of central venous catheteter, presence of urinary catheter, intra-abdominal operation and/or urinary interventions performed within 90 days prior to the episode of bacteremia, antibiotic use, and another hospitalisation within 90 days prior to the episode of bacteremia, intensive care unit stay, and survival.
Statistically analyses were performed using SPSS 15.0 (SPSS, Chicago, USA).All P-values were twotailed and a P-value of < 0.05 was considered statistically significant.The chi-square test was used to compare dichotomous data unless the cell size was < 5, in which case Fisher's exact test was used.Univariate and multivariate logistic regression (stepwise regression) analyses were used to compare matched cases and controls.Variables found to be significant on univariate analysis were included in the multivariate model.Odds ratios and their 95% confidence intervals were calculated.
The study protocol was approved by the ethics committee of the Suleyman Demirel University, Faculty of Medicine.

Results
During the study period, 113 adult inpatients had bloodstream infection due to E. coli.The study included 58 male and 55 female subjects.The most common underlying diseases were solid tumours (n = 35; 30.9%), followed by diabetes mellitus (n = 27; 23.89%) and chronic renal failure (n = 17, 15.04%).The frequency of ESBL producers was 38.9% (44/113).Seventeen patients with bacteremia due to ESBL-producing E. coli had community-onset bacteremia (17/44; 38.6%).No significant differences between community-and hospital-onset bacteremia due to ESBL-producing E. coli were found in all parameters such as gender, age, underlying diseases and conditions, mortality rate, site of infection, urinary catheter, and prior to antibiotic use in 90 days (p > 0.05) (Table 1).Mortality rates of both groups (community-onset and hospital-onset ESBL-producing E. coli bacteremia) were similar.The source of most infections of ESBL-producing E. coli bacteremia was the intra-abdominal region (n = 18; 40.9%), followed by the urinary tract (n = 13; 29.5%).
Univariate analysis revealed that significant factors associated with bacteremia due to ESBL-producing E. coli were presence of urinary catheter, presence of central venous catheter, intra-abdominal surgery within 90 days, prior cephalosporin use in 90 days, and prior fluoroquinolone use in 90 days (p < 0.05).Although the mortality rate of ESBL-producing E. coli bacteremia group was higher than that of the ESBLnegative E. coli bacteremia group, no significant difference was found (Table 2).
Multivariate analysis using a logistic regression model showed that fluoroquinolone and cephalosporin exposure within 90 days were the significant independent factors associated with ESBL-producing E. coli bacteremia (p < 0.05) (Table 3).

Discussion
Bloodstream infections due to Gram-negative pathogens may be associated with high mortality [7]; therefore, treatment of these infections is important.On the other hand, treatment of infections due to resistant microorganism strains may be more difficult.Clinicians must decide the appropriate antibiotic for the patient; thus determination of these risk factors contributes to empiric antimicrobial treatments for life-threatening infections.In this study, we aimed to identify risk factors for ESBL-producing microorganisms in hospitalized patients with bacteremia.
The source of hospital-onset Gram-negative bacteremia is usually urinary tract infections.Although our results demonstrate that most common infections in hospital-onset ESBL-producing E. coli bacteremia is intra-abdominal infection, this was not statistically significant compared to community-onset bacteremia (p > 0.05).This situation may be explained by the fact that the majority of cases in both of groups had undergone an intra-abdominal surgery.The source of bacteremia in nearly half of the cases in the community-onset bacteremia group was the urinary tract.Furthermore, the most detected primary infection site was the urinary tract in the ESBL-negative E. coli group, whereas the ESBL-producing E. coli group involved the intra-abdominal site.
Fluoroquinolone exposure was found to be independently associated with ESBL-producing E. coli bacteremia in the present study (p = 0.032, OR 12.85 95 % CI 1.24-132.75).In a prospective study, the use of beta-lactams or fluoroquinolones was detected as an independent risk factor among hospitalized patients for ESBL-producing strains [8].Previous retrospective case-control studies have also shown similar results [3,4].
Although elongated intensive care unit stay is commonly a risk factor for nosocomial infections due to multidrug resistant microorganisms such as Acinetobacter baumannii and ESBL-producing Escherichia coli, our results were not consistent with this [9,10] and may be explained by short duration of stay in intensive care unit by our patients.
In the presented study, hospitalization within 90 days before onset of bacteremia infection was not found to be associated with ESBL production.The emergence of this unusual outcome may be connected to some factors such as antimicrobial chemotherapy, invasive procedures and hospitalization duration in our patients.
Prior use of cephalosporins have been identified as a risk factor for ESBL-producing bacteria as well as fluoroquinolones [3,4,11,12].Use of cephalosporin was associated with an increased risk of ESBLproducing E. coli bacteremia by multivariate analysis in the present study (p = 0.044).
Mortality of bloodstream infections caused by ESBL-producing organisms was higher than that of infections caused by ESBL-negative organisms [13,14].Mortality rates due to ESBL-producing microorganism were found as about 12.1% to 50.9% in several published reports [1,12,[15][16][17].In the present study, however, the rate of mortality in the ESBL-producing E. coli bacteremia group was higher than that of the ESBL-negative E. coli bacteremia group (p > 0.05); similar underlying conditions in both groups may be the reason for this observation.However, treatments in both groups were not analyzed.
This study has several limitations, one of which is that prior studies have demonstrated that colonization of ESBL-producing microorganisms is a significant factor associated with bacteremia with ESBL-  producing isolates.In this retrospective study, we were unable to observe the colonization of ESBL-producing bacteria in the patients' data.The second limitation is that the patients were followed in different clinics; some data was lost for some patients.Therefore, we did not determine these factors to contribute to mortality.

Conclusion
This study showed that community-onset ESBLproducing E. coli bacteremia is an important problem for our region.Furthermore, previous uses of fluoroquinolone and cephalosporin are positive risks of ESBL-producing E. coli bacteremia.

Table 1 .
Comparison of patients' characteristics with community-onset and hospital-onset bacteremia due to ESBLproducing E. coli ESBL: extended-spectrum beta-lactamase, DM: diabetes mellitus, COPD: chronic obstructive pulmonary disease, CRF: chronic renal failure, CLD

Table 2 .
Characteristic of patients with bacteremia due to ESBL-producing E. coli versus non-ESBL-producing E. coli

Table 3 .
Risk factors on multivariate analysis for bacteremia due to ESBL-producing E. coli ESBL: extended-spectrum beta-lactamase; OR: odds Ratio; CI: Confidence interval; CVC: central venous catheter *statistically significant at p < 0.05