Extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae (ESBLE) has emerged in hospitals at the end of the 1980s, causing outbreaks and hyperendemic situations in many centers. In Israel, several recent data have suggested that ESBLE infections are currently emerging in the community.1-3 Because a large proportion of ESBLE is also resistant to other classes of antibiotics, including fluoroquinolones, trimethoprim-sulfamethoxazole, and aminoglycosides,1,3,4 the choice of treatment of these infections is limited. In a previous study, we described the risk factors for ciprofloxacin and gentamicin resistance among patients who had Enterobacteriaceae bloodstream infection on admission to the emergency department.5 The aims of the present study were to describe the risk factors for bloodstream infection due to ESBLE and the factors related to the outcome of these patients, emphasizing the role of adequate versus inadequate empirical treatment.
PATIENTS AND METHODS
The study evaluated prospectively all patients who were admitted in the emergency department and were found to have a positive blood culture drawn on admission, during a 1-year period. For each patient, a 25-variable questionnaire was completed, containing pertinent demographic, clinical, bacteriological, and laboratory features and outcome (30-day in-hospital mortality). We attempted to apply the Acute Physiology and Chronic Health Evaluation (APACHE) III score, originally developed for intensive care unit patients, for our patient population.6 However, not all data were available, in particular, the acid-base status, Po2, urine output, and Glasgow Coma Scale. Consequently, the range of our modified APACHE-III score is 0 to 209 instead of the 0-to-299 range of the original score system. Sufficient clinical data for our modified APACHE-III score were available for 100 of 131 patients.
Blood cultures were managed using the BACTEC 9240 system (Becton & Dickinson, Sparks, MD). Bacterial identification and antibiotic susceptibilities were determined using disk-diffusion methods and applying the National Committee for Clinical Laboratory Standards guidelines for the determination of breakpoints.7 Susceptibility testing included ceftazidime, ceftriaxone, and imipenem but not ertapenem. The presence of ESBLs was determined with the double-disk method.7 Extended-spectrum β-lactamase-producing Enterobacteriaceae was defined as multidrug-resistant if it was resistant to all of the following 3 antibiotic classes: trimethoprim-sulfamethoxazole, aminoglycosides, and fluoroquinolones.8
Statistical analysis: Continuous variables were assessed using the t test, and categorical variables with the χ2 or Fisher exact test. Multivariate analysis was done by forward-stepwise (likelihood ratio) binary logistic regression. For all analyses, α = 0.05. All analyses were performed using the SPSS for Windows (version 10; SPSS Inc, Chicago, Ill).
During the study period, approximately 7500 blood culture sets were collected from the adult and pediatric emergency departments, from approximately 4700 patients. Of these cultures, 245 patient-specific cultures were positive, of which 131 (53%) had Enterobacteriaceae bacteremia. Among these, 28 isolates (21%) were ESBLE. The baseline characteristics and the risk factors for non-ESBL versus ESBL among the Enterobacteriaceae are presented in Table 1. One hundred patients (76%) had at least 1 risk factor for health care-associated infection, including nursing home residence, recent hospitalization or medical procedure, immunosuppression, and permanent catheterization. The risk for ESBLE was higher among these patients (24% vs 13%, respectively), but the difference was not statistically significant.
Resistance to other antimicrobial agents, including trimethoprim-sulfamethoxazole, ciprofloxacin, and gentamicin, was more common in ESBLE than in non-ESBLE (P < 0.01). Among the ESBLE, multidrug resistance was more common in K. pneumoniae isolates (10 of 12 vs 4 of 16 in other bacteria; P = 0.006), but imipenem resistance was not found.
The most common agents that were administered as an empirical treatment (data were available for 125 patients) were gentamicin (n = 38; 30%) and ciprofloxacin (n = 36; 29%) but never carbapenem. Empirical treatment was adequate in 79 (63%) of the patients. The variables that were associated with inadequate empirical treatment are presented in Table 2. In multivariate analysis, ESBL and the presence of urinary or other nonvascular catheters were identified as risk factors for an inadequate treatment.
The in-hospital 30-day mortality rate was 16% (n = 21; data were available for 130 patients). The variables that were associated with increased mortality are presented in Table 3. Neither ESBLE isolation nor inadequate antibiotic treatments were found to be associated with a high mortality rate. Only a highly modified APACHE-III score was associated with increased mortality in the multivariate analysis.
In the present study, we found a 21% rate of bloodstream infection due to ESBLE in patients on admission to the emergency department. This rate is higher than those previously found in bloodstream infections in southern and central Israel (13.7% and 5%, respectively)1,2 and is similar to the rate found in urine cultures (25%) in a long-term care facility in central Israel.9
The risk factors for ESBLE in our study are similar to those found in previous outpatient-based studies, including older age,3,4 nursing home residence,1 recent use of antibiotics (in particular, second-generation cephalosporins and ciprofloxacin),3,4,10 and presence of a permanent urinary catheter.2 Recent hospitalization, which was found to be a risk factor in some of these studies,3,4 was more common in patients with ESBLE, although it did not reach statistical significance. Surprisingly, in our study, diabetes mellitus was found to be an indicator for antibiotic susceptibility rather than resistance. We cannot offer a plausible explanation for that. Most of these risk factors indicate that acquisition of the infection, in most cases, was health care-related and not truly community acquired. As pointed out by Siegman-Igra et al,11 we found that the majority (76%) of our patients had at least 1 risk factor for health care-related infection, and these patients had a higher risk for ESBLE. Nursing homes and long-term care facilities are known reservoirs for ESBLE9,12; although initially ESBLE may be imported from health care facilities,13 there may be subsequent spreading of resistance genes carrying plasmids12 inside the nursing home. Thereafter, whenever these patients require hospitalization, they may introduce these strains back into the hospital,1 and a vicious cycle accelerating the spread of ESBLE is formed. Although the use of antibiotics should be limited as much as possible, all other risk factors are nonmodifiable. Thus, the main tool for limiting the spread of these organisms is the increased adherence to infection-control practice inside the hospitals and in long-term care facilities.
Extended-spectrum β-lactamase-producing Enterobacteriaceae was the main risk factor for inadequate treatment in our study as has been previously reported.14,15 As shown in our study, ESBLE strains usually have more coresistance to other classes of antibiotics than non-ESBLE strains. This feature is explained by the transfer of plasmids that carry a complex of several resistance-acquiring genes.12 Because carbapenems were not administered in the emergency department, it is not surprising that most of these patients received inadequate empirical antibiotic treatment. This raises the dilemma of whether carbapenems should be used more often as empirical treatment of patients with suspected sepsis with the previously mentioned risk factors for ESBLE. This approach is undesirable, because an empirical treatment of suspected ESBLE infections with imipenem has been associated with significant increases in carbapenem resistance in other organisms (eg, Pseudomonas aeruginosa and Acinetobacter baumannii8) and may also contribute to the recent emergence of carbapenem-resistant Enterobacteriaceae in Israel.1 Therefore, it was important to investigate the effect of inadequate treatment on the outcome of these patients. We found that neither ESBLE isolation nor inadequate treatment had a significant effect on the 30-day in-hospital mortality. The only significant predictor for a higher mortality rate was the modified APACHE-III score, which reflects the severity of the patients' clinical condition on admission. The importance of the severity of the patients' state at presentation, as opposed to the lack of significant effect of inadequate empiric treatment, on the mortality rate in ESBLE sepsis was also found in other studies.14,16 As reported by Hyle et al,14 we found a lower mortality rate in patients with urinary tract infection (UTI) (10% vs 28% in UTI and non-UTI infections, respectively; Table 3), but this variable was not significant in the multivariate analysis. Although Anderson et al17 found increased in-hospital mortality in ESBLE sepsis when administration of adequate treatment was delayed for more than 72 hours, they did not find a significant effect of inadequate treatment during the acute phase or a significant advantage of empirical treatment with carbapenems at extended-spectrum cephalosporins. Schwaber et al15 reported a higher mortality rate, higher hospital costs, and longer length of stay of patients with ESBLE sepsis compared with non-ESBLE patients. A meta-analysis done by the same group also found that delay in adequate treatment was associated with increased mortality.18 Our study is limited in that we were only able to use a modified and, therefore, limited clinical severity score and that the only outcome measure that we evaluated was the 30-day mortality rate. Hence, it is possible that other measures were adversely affected by ESBL or by the inadequate empirical treatment.
In conclusion, although the proportion of bloodstream infections due to ESBLE in patients admitted to the emergency department with sepsis is high, especially in patients with health care-related risk factors, it is possible that inadequate empirical antibiotic treatment does not have an adverse effect on the 30-day mortality rate of these patients, although other outcome measures may be affected. Because of their role as a "last resort" in many severe gram-negative infections and to decrease the risk for increased resistance to these agents, we believe that routine use of carbapenems as an empirical treatment in the emergency department may not be an optimal approach. These agents should be considered for suspected sepsis in critically ill patients with risk factors for ESBLE. Other agents, such as aminoglycosides, may be preferred as an empirical treatment of suspected gram-negative sepsis.
1. Ben-Ami R, Schwaber MJ, Navon-Venezia S, et al. Influx of extended-spectrum beta-lactamase-producing Enterobacteriaceae into the hospital. Clin Infect Dis
2. Borer A, Gilad J, Menashe G, et al. Extended-spectrum beta-lactamase-producing Enterobacteriaceae strains in community-acquired bacteremia in Southern Israel. Med Sci Monit
3. Colodner R, Rock W, Chazan B, et al. Risk factors for the development of extended-spectrum beta-lactamase-producing bacteria in nonhospitalized patients. Eur J Clin Microbiol Infect Dis
4. Rodriguez-Bano J, Navarro MD, Romero L, et al. Epidemiology and clinical features of infections caused by extended-spectrum beta-lactamase-producing Escherichia coli
in nonhospitalized patients. J Clin Microbiol
5. Benenson S, Yinnon AM, Schlesinger Y, et al. Optimization of empirical antibiotic selection for suspected Gram-negative bacteraemia in the emergency department. Int J Antimicrob Agents
6. Knaus WA, Wagner DP, Draper EA, et al. The APACHE-III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest
7. Clinical and Laboratory Standard Institute (CLSI). Performance Standards for Antimicrobial Susceptibility Testing. 6th Informational Supplement
. NCCLS document M100-S16. Wayne, PA: CLSI; 2006.
8. Hyle EP, Lipworth AD, Zaoutis TE, et al. Risk factors for increasing multidrug resistance among extended-spectrumβ-lactamase-producing Escherichia coli
species. Clin Infect Dis
9. Mendelson G, Hait V, Ben-Israel J, et al. Prevalence and risk factors of extended-spectrum beta-lactamase-producing Escherichia coli
and Klebsiella pneumoniae
in an Israeli long-term care facility. Eur J Clin Microbiol Infect Dis
10. Calbo E, Romani V, Xercavins M, et al. Risk factors for community-onset urinary tract infections due to Escherichia coli
harbouring extended-spectrum beta-lactamases. J Antimicrob Chemother
11. Siegman-Igra Y, Fourer B, Orni-Wasserlauf R, et al. Reappraisal of community-acquired bacteremia: a proposal of a new classification for the spectrum of acquisition of bacteremia. Clin Infect Dis
12. Wiener J, Quinn JP, Bradford PA, et al. Multiple antibiotic-resistant Klebsiella
and Escherichia coli
in nursing homes. JAMA
13. Arpin C, Dubois V, Maugein J, et al. Clinical and molecular analysis of extended-spectrum b-lactamase-producing enterobacteria in the community setting. J Clin Microbiol
14. Hyle EP, Lipworth AD, Zaoutis TE, et al. Impact of inadequate initial antimicrobial therapy on mortality in infections due to extended-spectrum beta-lactamase-producing Enterobacteriaceae: variability by site of infection. Arch Intern Med
15. Schwaber MJ, Navon-Venezia S, Kaye KS, et al. Clinical and economic impact of bacteremia with extended-spectrum b-lactamase-producing Enterobacteriaceae. Antimicrob Agents Chemother
16. Kang CI, Kim SH, Park WB, et al. Bloodstream infections due to extended-spectrum beta-lactamase-producing Escherichia coli
and Klebsiella pneumoniae
: risk factors for mortality and treatment outcome, with special emphasis on antimicrobial therapy. Antimicrob Agents Chemother
17. Anderson DJ, Engemann JJ, Harrell LJ, et al. Predictors of mortality in patients with bloodstream infection due to ceftazidime-resistant Klebsiella pneumoniae
. Antimicrob Agents Chemother
18. Schwaber MJ, Carmeli Y. Mortality and delay in effective therapy associated with extended-spectrum b-lactamase production in Enterobacteriaceae bacteraemia: a systematic review and meta-analysis. J Antimicrob Chemother