Since their development in the 1960s, fluoroquinolones have been frequently prescribed for infectious disease. Together with trimethoprim-sulfamethoxazole (TMP-SMZ), fluoroquinolones are included among the antibiotics recommended by the Infectious Disease Society of America (IDSA) for urinary tract infection [1]. Despite this, resistance to TMP-SMZ and fluoroquinolones has been reported in urinary tract infection worldwide [2-5]. Ampicillin has been frequently used for urinary tract infection as an alternative drug owing to its activity against gram-negative microorganisms. Nonetheless, as acquired resistance to beta-lactam class antibiotics was reported [6], cephalosporins stable to beta-lactamase were developed, and thus the choice of antibiotics for urinary tract infection was widened [7,8]. However, after the report of extended spectrum beta-lactamase (ESBL)-producing organisms in 1983 in Germany, anxiety over the emergence of bacteria resistant to cephalosporins has been growing more and more [9].

According to the beta-lactamase classification by Bush et al, ESBL belongs to group 2be and is an enzyme that inactivates extended spectrum beta-lactam antibiotics such as cefotaxime, ceftazidime, and aztreonam [10]. The prevalence of ESBL-producing bacteria is gradually on the rise [11].

Therefore, we investigated the current situation of resistance of Escherichia coli to antibiotics and the prevalence of ESBL-producing bacteria in patients treated for urinary tract infection at our hospital as outpatients and as inpatients. In addition, by examining the past history of antibiotic administration, the past history of catheterization, and the past history of urological treatments and assessing their effects on the emergence of ESBL-producing bacteria, we made efforts to establish measures to prevent the emergence of resistant bacteria.

The development of antibiotics has contributed greatly to reducing mortality caused by infection; nonetheless, as the use of antibiotics becomes generalized, the vicious circle of the development of new antibiotics and the emergence of resistant bacteria cannot be severed. In fact, according to an epidemiological study conducted by the National Nosocomial Infections Surveillance (NNIS), the resistance rate of most bacteria to antibiotics has increased [13].

Comparison of two studies conducted on ambulatory patients well explains the increasing resistance of E. coli to TMP-SMZ and ciprofloxacin in America [4,5]. In the report of the European Antimicrobial Resistance Surveillance System (EARSS), similarly, in most European countries, the prevalence of E. coli resistant to fluoroquinolones, the third-generation cephalosporins, and aminoglycosides is on the increase every year [3].

In Korea, according to the study reported by Ryu et al, a trend for a reduction in sensitivity of E. coli to the thirdgeneration cephalosporins and fluoroquinolones was observed, and E. coli resistance to antibiotics was greater in I-UTI [14]. An increase in resistance to antibiotics in O-UTI has been shown already in other Korean studies [15-17].

We investigated the antimicrobial susceptibilities of E. coli in I-UTI compared with O-UTI. Because of the possibility of contamination from a nosocomial component, we used the term 'outpatient' instead of 'community-acquired.' The sensitivity of E. coli to cephalosporins and fluoroquinolones was lower in I-UTI. In contrast, the emergence of ESBL-producing bacteria was observed to be higher in I-UTI (Table 1). In O-UTI, sensitivity of E. coli to the third-generation cephalosporins did not show a great difference from the studies reported in 2008 by Lee [17]; nonetheless, in comparison with other previous studies, slightly reduced patterns were observed [15,16]. In comparison with Europe, concerning the second-generation cephalosporin drug cefuroxime, sensitivity was comparable to that in a multinational study conducted by Naber et al [18]. Nonetheless, in comparison with the 2008 EARSS report, the sensitivity of E. coli to the entire class of third-generation cephalosporins was observed to be lower, excluding Bulgaria (29%) and Turkey (42%) [3].

The current situation of resistance to antibiotics has reached a serious point In urinary tract infection, and presently, multidrug-resistant bacteria including ESBL-producing bacteria can be readily encountered in clinics. Antibiotics that can be used for the treatment of multidrug-resistant bacteria including ESBL-producing bacteria In urinary tract infection are limited [19]. In our study, similarly, excluding imipenem and cefotetan, antibiotics with sensitivity higher than 50% to ESBL-producing E. coli were absent (Table 2). In addition, infection with ESBL-producing bacteria raises mortality, and it not only prolongs hospital stay but also increases relative treatment costs [20]. Therefore, it is very important to assess the risk factors for the emergence of multidrug-resistant bacteria, including ESBL-producing bacteria, and to provide means to prevent such resistance.

Recent studies showed that the emergence of ESBL-producing bacteria is more frequent in patients with a past history of hospitalization, a past history of exposure to antibiotics, a past history of catheterization, and a past history of urogenital surgery [21-23]. In the present univariate analysis, female gender, I-UTI, recent history of hospitalization, recent history of Foley catheterization, and recent exposure to specific antibiotics were observed to be associated with the emergence of ESBL-producing bacteria (Table 3). In the multivariate analysis, recent hospitalization was shown to be the greatest risk factor, followed by (in order) Foley catheterization, I-UTI, exposure to cefaclor, exposure to cefminox, and female gender (Table 4). Rodriguez-Bano et al reported in a case-control study conducted on 147 patients with O-UTI that the strongest risk factor for the emergence of ESBL-producing E. coli was a past history of hospitalization within the recent 1 year (odds ratio, 18.2; 95% confidence interval, 1.9 to 30.1) [24]. In our study, similarly, the probability of infection with ESBL-producing E. coli was higher in patients with a recent history of hospitalization than in patients without a history by approximately 2.5 times (Table 4). This suggests that blocking of infection within hospitals could reduce the prevalence of ESBL-producing bacteria and could lower the spread to communities. Thus, efforts to reduce the opportunities for inpatient infection with ESBL-producing bacteria should be considered very important. Siegel et al recommended that to block the spread of multidrug-resistant bacteria within hospitals, hospitals should restrict contact with patients infected with multidrug-resistant bacteria internally and provide detergents containing alcohol in the area and require their use [25]. In addition, they mentioned that hand hygiene education for medical staff and guardians should be required. Furthermore, Siegel et al emphasized the need to monitor the trend of multidrug-resistant bacteria by routine bacterial culture and also to establish organized report systems.

A past history of exposure to a specific antibiotic may be another risk factor for the emergence of ESBL-producing bacteria, and this topic is currently an active research area. Calbo et al reported the risk factors of exposure to the second-generation cephalosporins in oral form and the second-generation cephalosporin drug cefuroxime in injection form [21]. In our study, similarly, the second-generation cephalosporin drug cefaclor in oral form and the second-generation cephalosporin drug cefminox in injection form were shown to be risk factors. Among ESBL patients, the percentages of patients already exposed to these drugs were 12.2% and 14.4%, respectively. The percentage of patients already exposed to ciprofloxacin, one of the fluoroquinolones, was 17.2%; nonetheless, it was not associated with the emergence of ESBL-producing bacteria. Goossens et al explained the association of the level of use of antibiotics with bacteria acquiring resistance and emphasized the adequate use of antibiotics [26]. In fact, it has been reported that the appropriate use of antibiotics against gram-negative bacteria decreases the frequency of multidrug-resistant bacteria [25]. On the other hand, in clinics, prescription of antibiotics empirically for pyuria in asymptomatic bacteriuria patients or catheterized patients is seen occasionally, and the IDSA suggests the standard of prescription of antibiotics for such patients [27]. Therefore, to prevent the emergence of multidrug-resistant bacteria in urinary tract infection, efforts should be made to strictly observe the indications for the administration of antibiotics and to use antibiotics appropriately for a minimal period through culture for cases requiring antibiotics.

Foley catheterization is well known to be a risk factor for urinary tract infection. About 20% of urinary tract infection associated with Foley catheterization occurs due to unskilled sterilization techniques during the insertion of the catheter [28]. Infection may occur due to contamination of the collecting system through the lumen of the catheter [29]. Other cases have been reported to occur through the biofilm formed between catheter and the urethral mucosa [30]. Some reports emphasized that Foley catheterization may play an important role as a risk factor for the appearance of ESBL-producing bacteria [22]. In our study, similarly, the risk of emergence of ESBL-producing bacteria appeared to be approximately 2.4 times higher in patients with Foley catheterization (Table 4). Therefore, unnecessary Foley catheterization should be avoided in patients with urinary tract infection to prevent the emergence of ESBL-producing bacteria. If unavoidable, Foley catheterization should be performed by use of sterile techniques, and a closed drainage system should be applied during the minimal indwelling period.

Female gender appeared to be a risk factor for the emergence of ESBL-producing bacteria, and the proportion of women was about 8% higher in the ESBL group than in the non-ESBL group, although some studies have shown no statistical difference in gender or a slightly higher incidence of ESBL-producing bacterial infection in men [21-24]. Because many studies had a relatively small scale and were conducted retrospectively in limited age groups, large, prospective clinical studies with all age groups will be required to minimize the selection bias.

A past history of urogenital surgery, regardless of the type of the surgery, had no association with ESBL-producing bacterial infection in this study. Therefore, further clinical studies are needed because the fraction of the ESBL group with a urogenital operative history was too low to avoid the error of false generalization.

The total prevalence of ESBL-producing microorganisms in urinary tract infection was approximately 13%. In O-UTI and I-UTI, the prevalence of ESBL-producing E. coli was 12.1% and 23.1%, respectively. The risk appeared to be increased in cases with I-UTI, previous hospitalization, a history of Foley catheterization, previous exposure to specific antibiotics, and female gender. Among the risk factors, a past history of hospitalization appeared to be the greatest risk factor.

Therefore, to prevent the spread of multidrug-resistant microorganisms, especially ESBL-producing species in urinary tract infection, medical institutions should make efforts to develop administrative and educational programs and to provide appropriate guidelines for the prescription of antibiotics as well as urinary catheterization and to follow these guidelines strictly.