Clin Microbiol Infect 2010; 16: 147–151
The aim of this prospective cohort study was to determine the risk factors for community-acquired urinary tract infections (UTIs) caused by extended-spectrum β-lactamase (ESBL)-positive Escherichia coli and the distribution of the ESBL enzyme types. Structured forms were filled in for patients diagnosed with community-acquired UTI in four different geographical locations in Turkey. The forms and the isolates were sent to the central laboratory at Baskent University Hospital, Ankara. Antimicrobial susceptibility was determined according to the CLSI criteria. PCR and DNA sequencing were used to characterize the blaTEM, blaCTX-M and blaSHV genes. Multivariate analysis was performed using logistic regression. A total of 510 patients with UTI caused by Gram-negative bacteria were included in this study. ESBLs were detected in 17 of 269 (6.3%) uropathogenic E. coli isolates from uncomplicated UTIs and 34 of 195 (17.4%) E. coli isolates from complicated UTIs (p <0.001). According to multivariate analysis, more than three urinary tract infection episodes in the preceding year (OR 3.8, 95% CI 1.8–8.1, p <0.001), use of a β-lactam antibiotic in the preceding 3 months (OR 4.6, 95% CI 2.0–0.7, p <0.001) and prostatic disease (OR 9.6, 95% CI 2.1–44.8, p 0.004) were found to be associated with ESBL positivity. The percentages of isolates with simultaneous resistance to trimethoprim–sulphamethoxazole, ciprofloxacin and gentamicin were found to be 4.6% in the ESBL-negative group and 39.2% in the ESBL-positive group (p <0.001). Forty-six of 51 ESBL-positive isolates (90.2%) were found to harbour CTX-M-15. Therapeutic alternatives for UTI, particularly in outpatients, are limited. Further clinical studies are needed to guide the clinicians in the management of community-acquired UTIs.
The incidence of community-acquired urinary tract infections (UTIs) due to extended-spectrum β-lactamase (ESBL)-producing Escherichia coli has increased worldwide [1,2]. Recently, the emergence of an intercontinental clone of an E. coli strain producing CTX-M-15 has gained much interest, and Turkey was reported to be concerned by this emergence of a community-acquired urinary isolate [3–5]. Besides their capacity for rapid and widespread dissemination, ESBL-producing strains are increasingly associated with resistance to non-β-lactam antimicrobials, and cause therapeutic difficulties for outpatients .
The aim of this prospective cohort study was to determine the risk factors for ESBL positivity in community-acquired uropathogenic E. coli strains and the distribution of the enzyme types.
Materials and Methods
Design of the study
The present study is a prospective cohort study, and was carried out between 1 January 2007 and 31 December 2007. Patients with the diagnosis of UTI were included consecutively. UTI was defined as a growth of 105 CFU/mL in urine culture, pyuria, and at least one of the following symptoms: dysuria, frequency or urgency. Each patient was included once in the study, and only one isolate was obtained from each patient. The four centres participating in the study were located in large cities from three geographically distinct regions of Turkey. All four centres were tertiary-care hospitals of Başkent University. In Turkey, patients may be admitted to any hospital (primary-care or tertiary-care hospital) with any complaint (including dysuria), because there is not yet an admission chain scheduled in the healthcare system. In this way, consecutive outpatients admitted to the participating centres were included in this study.
A structured form was used to collect data about demographic characteristics, urinalysis, signs of complication, antibiotic usage, and underlying diseases. Informed consent from participants was obtained on a written form. All forms and the identified Gram-negative uropathogens were sent to the central laboratory (Baskent University Ankara Hospital Microbiology Laboratory). This project was approved by the Başkent University Research Committee and was funded by Başkent University Research Foundation.
Patients between 18 and 65 years of age with the diagnosis of community-acquired uncomplicated or complicated UTI were included in the study. Uncomplicated UTI is defined as an infection in a structurally and neurologically normal urinary tract . Complicated UTI is defined as an infection in a urinary tract with functional or structural abnormalities, and UTI in men and pregnant women . Male patients with urolithiasis, patients who had more than three episodes of UTI in the preceding year, pregnant patients, patients with a urinary catheter and patients who had undergone an operation because of urolithiasis or bladder malignancy in the preceding year were considered to have complicated UTI. Hospital stay within the preceding month was a criterion of exclusion.
Standard biochemical reactions were performed as the first step for the identification of bacteria, and BBL Crystal Enteric/NF 4.0 identification kits (Becton Dickinson) were used when needed. All isolates were stored at −20°C. Antibiotics were grouped into two sets according to the requirements of clinical practice; one was tested against isolates from uncomplicated UTIs, and the other against isolates from complicated UTIs. The antibiotics tested against both types of isolate were ampicillin, amoxycillin–clavulanate, cefazolin, cefuroxime, ceftriaxone, cefixime, gentamicin, ciprofloxacin, trimethoprim–sulphamethoxazole, imipenem, meropenem, and ertapenem. The antibiotics tested only against isolates from uncomplicated UTIs were nitrofurantoin and fosfomycin. The antibiotics tested only against isolates from complicated UTIs were piperacillin–tazobactam, cefepime, cefoperazone, ceftazidime, and amikacin.
Antibacterial susceptibility testing was performed according to the CLSI criteria . Quality control was ensured by testing E. coliATCC25922, Pseudomonas aeruginosaATCC27853 and ESBL-producing Klebsiella pneumoniaeATCC 700603 in every batch. All zone sizes for these strains were within the ranges given by the CLSI for the antibiotics tested in this study. ESBL determination was performed phenotypically with ceftazidime/ceftazidime clavulanate and cefotaxime/cefotaxime clavulanate disks, as recommended by the CLSI .
Characterization of enzyme types
PCR and DNA sequencing were used to characterize the blaTEM, blaCTX-M and blaSHV genes. DNA sequencing was performed with the DYEnamic ET Terminator Cycle Sequencing Kit (Amersham Biosciences Corp., NJ, USA) and an ABI PRISM 310 Genetic Analyzer at Iontek Ltd, Turkey. The primers used have been described previously .
Data were analysed using the SPSS (version 11.0) software package. Categorical variables were compared using chi-square tests, although Fisher’s exact test was used when data were sparse. Significance was set at p <0.05, using two-sided comparisons. A multivariate model was used. Multivariate analysis included age, gender, and all significant factors found in univariate analysis. Significance was set at p <0.05 for inclusion in the multivariate analysis. A backward stepwise method was used for the elimination of the risk factors. ESBL positivity was the dependent variable in logistic regression.
A total of 16 130 urine samples were submitted to the microbiology laboratories of four centres during the 1-year study period. A total of 510 patients were included in this study. E. coli was isolated from 464 (90%) patients, and Klebsiella spp. were isolated from 27 (5%) patients. The mean age was 43 years, with a range of 18–65 years. E. coli was the causative agent in 269 of 289 (93%) uncomplicated UTIs and in 195 of 221 (88%) complicated UTIs (p 0.05).
ESBL positivity was detected in 17 of 269 (6.3%) uropathogenic E. coli strains isolated from uncomplicated UTIs and in 34 of 195 (17.4%) E. coli strains isolated from complicated UTIs (p <0.001). The rates of resistance in E. coli strains against the tested antibiotics are shown in Table 1.
|Uncomplicated UTI, n = 269 (%)||Complicated UTI, n = 195 (%)||p|
|Ampicillin||143 (53)||129 (66)||0.005|
|Amoxycillin–clavulanic acid||39 (14)||52 (27)||0.001|
|Cefazolin||39 (14)||49 (25)||0.004|
|Cefuroxime||38 (14)||48 (25)||0.004|
|Ceftriaxone||25 (9)||36 (18)||0.004|
|Cefixime||25 (9)||36 (18)||0.004|
|Gentamicin||28 (10)||38 (19)||0.006|
|Ciprofloxacin||59 (22)||80 (41)||<0.001|
|Trimethoprim–sulphamethoxazole||93 (35)||98 (50)||0.001|
The results of the univariate analysis are shown in Table 2. Multivariate analysis included all significant factors found in univariate analysis (more than three UTI episodes in the preceding year, prostatic disease, UTI in the preceding year, use of a β-lactam antibiotic in the preceding 3 months), and revealed that more than three UTI episodes in the preceding year (OR 3.8, 95% CI 1.8–8.1, p <0.001), use of β-lactam antibiotics in the preceding 3 months (OR 4.6, 95% CI 2.0–10.7, p <0.001) and prostatic disease (OR 9.6, 95% CI 2.1–44.8, p 0.004) were associated with ESBL positivity. The Hosmer and Lemeshow test (p 0.807) was used as a goodness-of-fit measure.
|Total, n = 464 (%)||ESBL positivity, n = 51 (%)||p|
|Age over 50 years||161 (35)||18 (11)||0.925|
|Complicated UTI||195 (42)||34 (17)||<0.001|
|Male gender||59 (13)||9 (15)||0.262|
|Urolithiasis||45 (10)||9 (20)||0.073|
|Pregnancy||20 (4)||1 (5)||0.712|
|Urological operation||28 (6)||6 (21)||0.108|
|Prostatic disease||7 (2)||3 (42)||0.032|
|More than three UTI episodes in the preceding year||47 (10)||14 (30)||<0.001|
|Urinary catheter||16 (3)||2 (13)||0.692|
|UTI in the preceding year||125 (27)||22 (18)||0.006|
|Use of antibiotic in the preceding 3 months||91 (20)||19 (21)||0.001|
|Use of β-lactam antibiotic in the preceding 3 months||31 (7)||11 (35)||<0.001|
|Use of quinolones in the preceding 3 months||29 (6)||3 (10)||1.000|
The rates of ciprofloxacin resistance among ESBL-positive and ESBL-negative strains were 84% and 23%, respectively (p <0.001); the corresponding values for trimethoprim–sulphamethoxazole resistance were 69% and 39% (p <0.001), for gentamicin resistance they were 57% and 9% (p <0.001), and for amoxycillin–clavulanate resistance they were 76% and 13% (p <0.001).
Multidrug resistance, defined as resistance to at least two of trimethoprim–sulphamethoxazole, ciprofloxacin or gentamicin, was seen in 21.1% of E. coli isolates in the ESBL-negative group and in 70.6% in the ESBL-positive group (p <0.001). Isolates resistant to all three antibiotics were found at a rate of 4.6% in the ESBL-negative group and 39.2% in the ESBL-positive group (p <0.001).
Forty-six of 51 ESBL-positive isolates (90.2%) were found to harbour CTX-M-15, four isolates harboured CTX-M-3, and the remaining isolate harboured SHV-12. Six of 46 CTX-M-15-positive isolates also harboured TEM-116.
ESBL-producing E. coli has become the most worrisome causative agent of community-acquired UTIs. The present study is a prospective cohort study evaluating the prevalence of and risk factors for both uncomplicated and complicated UTIs. ESBL positivity was detected in 6.3% of E. coli isolates from uncomplicated UTIs and in 17.4% of E. coli isolates from complicated UTIs. More than three UTIs in the preceding year, use of a β-lactam antibiotic in the preceding 3 months and prostatic disease were found to be associated with ESBL positivity. One of the most striking findings of this study is the high rate of resistance to other drug classes; that is, 39.2% of ESBL-producing isolates were found to be resistant to trimethoprim–sulphamethoxazole, ciprofloxacin and gentamicin. CTX-M-15 was the most common (90.2%) ESBL type harboured by uropathogenic E. coli strains.
In our previous study carried out in 2004, ESBL positivity was found to be 5% in isolates from uncomplicated UTIs and 12% in isolates from complicated UTIs . The same increasing trend was observed in Spain . Recurrent UTI, a major risk factor determined in our study, was also found to be an independent risk factor for ESBL positivity in two other studies from Spain [1,7]. Previous use of aminopenicillins , cephalosporins , cefuroxime , second-generation and third-generation cephalosporins , penicillins  or cephalosporins  was associated with ESBL positivity in other studies. Quinolone use was found to be an independent risk factor in many studies, including one from Turkey [1,7,12,13], but was not found to be associated with ESBL positivity in a recent study, just as in the present study . This is somewhat surprising, because misuse of quinolones along with other antibiotics was documented in a study from Turkey . Long-term-care facilities were found to be risk factors in other studies but, as nursing home care is not common in Turkey, no long-term-care facility was included in this study.
One of the most worrisome aspects of ESBL-positive bacteria concerns the high rates of resistance to non-β-lactam antibiotics, particularly quinolones, trimethoprim–sulphamethoxazole, and aminoglycosides. Ciprofloxacin resistance, in a study from Spain in 2006, was reported to be 31.5% in ESBL-positive and 9.1% in ESBL-negative E. coli isolates . Ciprofloxacin resistance rates, published in 2004 from Israel, were reported to be even higher: 39% in ESBL-positive and 16% in ESBL-negative E. coli isolates . The results of the present study show an unfortunately high rate, 84%, of ciprofloxacin resistance among ESBL-positive E. coli isolates. Multidrug-resistant isolates, which cause difficult-to-treat infections, are also increasing in parallel. In a study from Spain, the percentage of multidrug-resistant isolates was reported to be higher than 70%, similar to the rate of 70.6% found in our study .
The most important consequence of such high resistance rates is the ensuing difficulty in the management of UTI patients. The number of drugs available for use in the outpatient setting is limited. Nitrofurantoin and fosfomycin seem to be the only choices in appropriate cases. Fortunately, in the present study, only three of all 464 uropathogenic E. coli isolates were resistant to nitrofurantoin, and only one isolate was resistant to fosfomycin, which is similar to what has been observed in many other studies [1,10,15,16]. Amoxycillin–clavulanate was used with success for patients with cystitis due to ESBL-producing E. coli in a recent study  where the resistance rate was 29%, and for lower urinary tract infection according to one case report , but as 76% of our ESBL-positive isolates were found to be resistant, this does not seem to be a good choice in our region.
The first CTX-M-15-producing E. coli isolate from Turkey was reported in 2005 and was isolated from the urine of a hospitalized patient . In a recent study from Turkey, CTX-M-15 was found in 53% of uropathogenic E. coli isolates from community-acquired UTIs . This increase in the frequency of CTX-M-15 in such a short period should be taken note of by clinicians.
The ESBL positivity and ciprofloxacin resistance rates are higher than those reported in other studies. One of the reasons for these high rates is that data were obtained from tertiary-care facilities. Another reason is probably related to the definition of ‘community-acquired’ infections, because, as stated by Friedman et al. , there seems to be a need to change the term ‘community-acquired’ to ‘healthcare-associated’, particularly in some instances. Therefore, our estimates regarding the resistance rates might be exaggerated, because we considered both tertiary-care admissions and healthcare-associated infections.
In conclusion, the widespread and rapid dissemination of ESBL-producing uropathogenic E. coli seems to be an emerging issue worldwide. Further clinical studies are needed to guide clinicians in the management of community-acquired UTI cases.
This project was funded by Başkent University Research Foundation. The authors declare that they have no conflicts of interest.