Abstract
Background. Women suffering from recurrent urinary tract infections (rUTIs) are routinely treated for asymptomatic bacteriuria (AB), but the consequences of this procedure on antibiotic resistance are not fully known. The aim of this study was to evaluate the impact of AB treatment on antibiotic resistance among women with rUTIs.
Methods. The study population consisted of 2 groups of women who had previously been enrolled in a randomized clinical trial: group A was not treated, and group B was treated. All women were scheduled for follow-up visits every 6 months, or more frequently if symptoms arose. Microbiological evaluation was performed only in symptomatic women. All women were followed up for a mean of 38.8 months to analyze data from urine cultures and antibiograms.
Results. The previous study population consisted of 673 women, but 123 did not attend the entire follow-up period. For the final analysis, 257 of the remaining 550 patients were assigned to group A, and 293 to group B. At the end of follow-up, the difference in recurrence rates was statistically significant (P < .001): 97 (37.7%) in group A versus 204 (69.6%) in group B. Isolated Escherichia coli from group B showed higher resistance to amoxicillin–clavulanic acid (P = .03), trimethoprim-sulfamethoxazole (P = .01), and ciprofloxacin (P = .03) than that from group A.
Conclusions. This study shows that AB treatment is associated with a higher occurrence of antibiotic-resistant bacteria, indicating that AB treatment in women with rUTIs is potentially dangerous.
Asymptomatic bacteriuria (AB) is defined as the presence of bacteria in the urine of an individual without signs or symptoms of urinary tract infections (UTI) [1]. AB treatment is generally recommended only during pregnancy and at preoperative evaluation of men before urological procedures [2, 3]. Despite these recommendations, overuse of antibiotics in AB treatment is common. This is particularly true in women with recurrent UTIs (rUTIs), in whom antibiotics are routinely administered, with a high risk of selecting multidrug-resistant bacteria [4, 5]. For many years, the inappropriate use of antibiotics has been recognized to be a major problem, leading to higher healthcare costs as well as increased antimicrobial resistance [4, 6] through the selection and spread of drug-resistant microorganisms with severe consequences for patient health [7, 8].
Optimization of antibiotic usage would not only prevent increased resistance, it would also limit costs. In particular, AB is treated in women with rUTIs without any clear indication for treatment and with a high risk of selecting multidrug-resistant bacteria [4, 5]. In a randomized controlled clinical study, Cai et al [4] demonstrated that during a 1-year period treatment of AB in patients affected by rUTIs was associated with a higher rate of symptomatic UTIs, suggesting that AB may play a protective role in preventing symptomatic recurrence. Although AB treatment in patients with rUTIs is associated with a higher rate of symptomatic UTIs [4], we have no data on the relationship between AB treatment and the risk of higher antibiotic resistance. Considering the available evidence, we investigated whether AB treatment in women affected by rUTIs increases the rate of antibiotic-resistant bacteria.
MATERIALS AND METHODS
Study Population, Design and Ethical Considerations
The study population consisted of 550 women previously enrolled in a randomized clinical trial to establish the impact of AB treatment on the recurrence rate in young women affected by rUTIs [4]. In our former study, 673 women were randomized into 2 groups: group A (untreated) and group B (treated) [4]. In the present study, we analyzed results in all the women included in the first study who were not lost to follow-up. Our report compares 2 parallel groups differing in exposure to the study factor in an observational, analytical, and longitudinal study to assess any possible cause-effect relationships [9]. The 2 groups of patients were identified on the basis of the 2 randomized groups in the previous study: group A included untreated women, and group B, treated women [4].
Thirty-six women in group A and 45 in group B were lost to follow-up after the previous study, and 19 in group A and 23 in group B were lost after the present study. This left a total of 550 women for analysis, 257 in group A and 293 in group B. Figure 1 shows the study flow chart, in accordance with the STROBE Statement (http://www.strobe-statement.org/). At the end of the previous study [4], clinical evaluation and microbiological analysis were performed in all the women to evaluate symptomatic recurrence and analyze data from urine cultures and antibiograms. The main result was the measure of resistance to antibiotics at the end of the study period. Written informed consent was obtained from all patients before treatment. The study was conducted in line with Good Clinical Practice guidelines, with the ethical principles laid down in the latest version of the Declaration of Helsinki. The local ethics committee was asked to evaluate the study but deemed it exempt from their approval.
Kaplan–Meier curve analysis performed to calculate the probability of being recurrence-free in the 2 groups. The arrow indicates the start of the present study, 12 months after the start of the previous study. Abbreviations: HR, hazard risk; SD, standard deviation.
Study Schedule and Data Collection
All women were asked to participate in this second trial at the conclusion of the previous study. The characteristics of the study population have been described in detail elsewhere [4]. After enrollment, all women gave their written informed consent and were scheduled for follow-up visits, every 6 months, or more often in the case of recurring symptoms. No microbiological evaluation was scheduled for asymptomatic patients to avoid any inappropriate antibiotic treatment. At each follow-up visit, the women completed a baseline questionnaire and underwent urological examination. As in the previous study schedule [4], all episodes of symptomatic UTIs were recorded and treated with antibiotics, depending on the organism, according to susceptibility testing in line with the European Association of Urology guideline recommendations [10]. The treated women were not excluded from analysis but monitored during follow-up. If >1 episode occurred, all data were collected and analyzed. All microbiological data from urine cultures and antibiograms for the 2 groups were collected and compared to evaluate any prevalence of antibiotic-resistant bacterial strains. All isolated strains from women who experienced multiple episodes were considered in the final analysis.
Microbiological Considerations
In cases of symptomatic recurrence, clean-catch midstream urine samples were collected at room temperature at urological examination and immediately taken to the laboratory under refrigerated conditions. All urine samples were analyzed for common bacteria and yeasts, aliquoted for DNA extraction and polymerase chain reaction testing for Chlamydia trachomatis, Neisseria gonorrhoeae, and urogenital Mycoplasma. Leukocyte esterase and nitrite levels were also measured from all the urine samples using dipstick assay (Bayer Multistik Pro Reagent Strips), in accordance with Hawn et al [11]. Microbiological culture was performed according to the procedure described by Hooton et al [12]. Urine DNA extraction and purification were performed using the DNeasy1 Tissue Kit (Qiagen), as described by Mazzoli et al [13]. The C. trachomatis chromosomal DNA polymerase chain reaction procedure amplifying the omp1 gene sequence was performed on 10 mL of the sample extraction mixture, in accordance with the procedure described by Pannekoek et al [14]. Our sexually transmitted diseases laboratory is registered under the United Kingdom National External Quality Assessment for molecular detection of C. trachomatis (Quality Assurance Laboratory, Health Protection Agency Centre for Infection).
According to Hawn and Nicolle, the presence or degree of pyuria had no prognostic significance, and it was therefore not considered [11, 15]. All susceptibility testing used the Vitek II semiautomated System for Microbiology (BioMerieux). The Kirby–Bauer disk diffusion method was used according to the recommendations by the Clinical Laboratory Standard Institute [16]. For microbiological diagnosis, ≥105 colony-forming units/mL was considered the cutoff. Resistance patterns for amoxicillin–clavulanic acid, trimethoprim-sulfamethoxazole, ciprofloxacin, levofloxacin, aminoglycosides (amikacin, gentamicin), piperacillin-tazobactam, fosfomycin, and cefotaxime against all isolated pathogens were analyzed. Antimicrobial resistance rates were recorded throughout the study period.
Questionnaires
Patient quality of life (QoL) was measured by using an Italian version of the Quality of Well-Being, a validated, multiattribute health scale [17]. This scale was selected because it has been successfully applied to acute illnesses, whereas other QoL scales, such as the Short Form (36) Health Survey (SF-36), are more suitable in chronic cases [18]. Higher scores on the Quality of Well-Being reflect better QoL.
Statistical Analysis
As the null hypothesis, we assumed there would be no difference in resistant bacterial strains between the 2 groups. To determine the number of persons needed for enrollment, sample size was calculated with an assumed relative risk of 3, a confidence level of 95%, an α error level of .05 (2 sided), and power of 0.80. The calculation yielded 138 individuals per group. Fisher exact and χ2 tests were used to assess the significance of all parameters, with differences considered significant at P < .05. Categorical variables were presented as percentages and compared using χ2 analysis. Continuous variables were given as means (standard deviations) and compared using Student t or Mann–Whitney U tests. Relative risks and 95% confidence intervals were estimated by applying log-binomial regression and Cox regression with a constant in the time variable [19]. All reported P values were 2 sided. Statistical analyses were performed using SPSS 11.0 software for Apple-Macintosh (SPSS).
RESULTS
At the end of follow-up, data from 257 women in group A and 293 in group B were analyzed. Nineteen women in group A and 23 in group B were lost during follow-up and consequently excluded from analysis. Apart from marital status (P < .001) and the number of previous UTIs per year (P < .001), the 2 groups were comparable in demographic, clinical, and laboratory parameters (Table 1); number of sexual partners (P = .71); and sexual encounters per week (P = .24).
Patient Clinical and Laboratory Characteristics at Enrollment Time
| Characteristic | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Age, median (SD), y | 38.5 (6.4) | 38.9 (7.2) | .49 (0.46/548) |
| Marital status | <.001 (10.9/1) | ||
| Married | 146 (56.8) | 124 (42.3) | … |
| Single | 111 (43.2) | 169 (57.7) | … |
| No. of partners | .71 (0.13/1) | ||
| 1 | 234 (91.1) | 263 (89.7) | … |
| ≥2 | 23 (8.9) | 30 (10.3) | … |
| Sexual encounters per week, mean (SD), No. | 1.9 (1.1) | 2.0 (0.9) | .24 (1.1/548) |
| Contraceptive use | 121 (47.1) | 139 (47.4) | .93 (0.007/1) |
| Oral hormonal | 63/121 (52.1) | 73/139 (52.5) | … |
| Condom | 37/121 (30.6) | 35/139 (25.2) | … |
| Coitus interruptus | 21/121 (17.3) | 31/139 (22.3) | … |
| Current smoker | 81 (31.5) | 96 (32.7) | .82 (0.049/1) |
| Parity | .30 (1.04/1) | ||
| Nulliparity | 52 (20.3) | 71 (24.2) | … |
| Multiparity | 205 (79.7) | 222 (75.8) | … |
| No. of UTIs per year | <.001 (0.02/1) | ||
| ≤3 | 89 (34.6) | 35 (11.9) | … |
| >3 | 168 (65.4) | 258 (88.1) | … |
| Start of recurrent UTI history, mean (SD), mo from the first episode | 19 (4.3) | 19 (5.0) | 1 (0.0/1) |
| Bacterial strains | .85 (0.03/1) | ||
| Escherichia coli | 102 (39.6) | 113 (38.6) | … |
| Enterococcus faecalis | 85 (33.0) | 97 (33.2) | … |
| Enterococcus faecium | 31 (12.3) | 38 (12.9) | … |
| Klebsiella spp. | 14 (5.5) | 17 (5.9) | … |
| Streptococcus agalactiae | 11 (4.2) | 14 (4.7) | … |
| Serratia spp. | 14 (5.5) | 14 (4.7) | … |
| Characteristic | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Age, median (SD), y | 38.5 (6.4) | 38.9 (7.2) | .49 (0.46/548) |
| Marital status | <.001 (10.9/1) | ||
| Married | 146 (56.8) | 124 (42.3) | … |
| Single | 111 (43.2) | 169 (57.7) | … |
| No. of partners | .71 (0.13/1) | ||
| 1 | 234 (91.1) | 263 (89.7) | … |
| ≥2 | 23 (8.9) | 30 (10.3) | … |
| Sexual encounters per week, mean (SD), No. | 1.9 (1.1) | 2.0 (0.9) | .24 (1.1/548) |
| Contraceptive use | 121 (47.1) | 139 (47.4) | .93 (0.007/1) |
| Oral hormonal | 63/121 (52.1) | 73/139 (52.5) | … |
| Condom | 37/121 (30.6) | 35/139 (25.2) | … |
| Coitus interruptus | 21/121 (17.3) | 31/139 (22.3) | … |
| Current smoker | 81 (31.5) | 96 (32.7) | .82 (0.049/1) |
| Parity | .30 (1.04/1) | ||
| Nulliparity | 52 (20.3) | 71 (24.2) | … |
| Multiparity | 205 (79.7) | 222 (75.8) | … |
| No. of UTIs per year | <.001 (0.02/1) | ||
| ≤3 | 89 (34.6) | 35 (11.9) | … |
| >3 | 168 (65.4) | 258 (88.1) | … |
| Start of recurrent UTI history, mean (SD), mo from the first episode | 19 (4.3) | 19 (5.0) | 1 (0.0/1) |
| Bacterial strains | .85 (0.03/1) | ||
| Escherichia coli | 102 (39.6) | 113 (38.6) | … |
| Enterococcus faecalis | 85 (33.0) | 97 (33.2) | … |
| Enterococcus faecium | 31 (12.3) | 38 (12.9) | … |
| Klebsiella spp. | 14 (5.5) | 17 (5.9) | … |
| Streptococcus agalactiae | 11 (4.2) | 14 (4.7) | … |
| Serratia spp. | 14 (5.5) | 14 (4.7) | … |
Abbreviations: SD, standard deviation; UTI, urinary tract infection.
a Data represent No. (%) of patients unless otherwise specified. Group A was untreated; group B, treated.
Patient Clinical and Laboratory Characteristics at Enrollment Time
| Characteristic | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Age, median (SD), y | 38.5 (6.4) | 38.9 (7.2) | .49 (0.46/548) |
| Marital status | <.001 (10.9/1) | ||
| Married | 146 (56.8) | 124 (42.3) | … |
| Single | 111 (43.2) | 169 (57.7) | … |
| No. of partners | .71 (0.13/1) | ||
| 1 | 234 (91.1) | 263 (89.7) | … |
| ≥2 | 23 (8.9) | 30 (10.3) | … |
| Sexual encounters per week, mean (SD), No. | 1.9 (1.1) | 2.0 (0.9) | .24 (1.1/548) |
| Contraceptive use | 121 (47.1) | 139 (47.4) | .93 (0.007/1) |
| Oral hormonal | 63/121 (52.1) | 73/139 (52.5) | … |
| Condom | 37/121 (30.6) | 35/139 (25.2) | … |
| Coitus interruptus | 21/121 (17.3) | 31/139 (22.3) | … |
| Current smoker | 81 (31.5) | 96 (32.7) | .82 (0.049/1) |
| Parity | .30 (1.04/1) | ||
| Nulliparity | 52 (20.3) | 71 (24.2) | … |
| Multiparity | 205 (79.7) | 222 (75.8) | … |
| No. of UTIs per year | <.001 (0.02/1) | ||
| ≤3 | 89 (34.6) | 35 (11.9) | … |
| >3 | 168 (65.4) | 258 (88.1) | … |
| Start of recurrent UTI history, mean (SD), mo from the first episode | 19 (4.3) | 19 (5.0) | 1 (0.0/1) |
| Bacterial strains | .85 (0.03/1) | ||
| Escherichia coli | 102 (39.6) | 113 (38.6) | … |
| Enterococcus faecalis | 85 (33.0) | 97 (33.2) | … |
| Enterococcus faecium | 31 (12.3) | 38 (12.9) | … |
| Klebsiella spp. | 14 (5.5) | 17 (5.9) | … |
| Streptococcus agalactiae | 11 (4.2) | 14 (4.7) | … |
| Serratia spp. | 14 (5.5) | 14 (4.7) | … |
| Characteristic | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Age, median (SD), y | 38.5 (6.4) | 38.9 (7.2) | .49 (0.46/548) |
| Marital status | <.001 (10.9/1) | ||
| Married | 146 (56.8) | 124 (42.3) | … |
| Single | 111 (43.2) | 169 (57.7) | … |
| No. of partners | .71 (0.13/1) | ||
| 1 | 234 (91.1) | 263 (89.7) | … |
| ≥2 | 23 (8.9) | 30 (10.3) | … |
| Sexual encounters per week, mean (SD), No. | 1.9 (1.1) | 2.0 (0.9) | .24 (1.1/548) |
| Contraceptive use | 121 (47.1) | 139 (47.4) | .93 (0.007/1) |
| Oral hormonal | 63/121 (52.1) | 73/139 (52.5) | … |
| Condom | 37/121 (30.6) | 35/139 (25.2) | … |
| Coitus interruptus | 21/121 (17.3) | 31/139 (22.3) | … |
| Current smoker | 81 (31.5) | 96 (32.7) | .82 (0.049/1) |
| Parity | .30 (1.04/1) | ||
| Nulliparity | 52 (20.3) | 71 (24.2) | … |
| Multiparity | 205 (79.7) | 222 (75.8) | … |
| No. of UTIs per year | <.001 (0.02/1) | ||
| ≤3 | 89 (34.6) | 35 (11.9) | … |
| >3 | 168 (65.4) | 258 (88.1) | … |
| Start of recurrent UTI history, mean (SD), mo from the first episode | 19 (4.3) | 19 (5.0) | 1 (0.0/1) |
| Bacterial strains | .85 (0.03/1) | ||
| Escherichia coli | 102 (39.6) | 113 (38.6) | … |
| Enterococcus faecalis | 85 (33.0) | 97 (33.2) | … |
| Enterococcus faecium | 31 (12.3) | 38 (12.9) | … |
| Klebsiella spp. | 14 (5.5) | 17 (5.9) | … |
| Streptococcus agalactiae | 11 (4.2) | 14 (4.7) | … |
| Serratia spp. | 14 (5.5) | 14 (4.7) | … |
Abbreviations: SD, standard deviation; UTI, urinary tract infection.
a Data represent No. (%) of patients unless otherwise specified. Group A was untreated; group B, treated.
Clinical Findings
At the end of the follow-up period (mean, 38.8 months), recurrence occurred in 97 (37.7%) of patients in group A and 204 (69.6%) in group B, with a statistically significant difference (df = 1; χ2 = 54.88; P < .001). Kaplan–Meier curve analysis showed that the probability of recurrence was higher in group B than in group A (hazard risk, 4.36; standard deviation, 2.1; P = .003) (Figure 2). At microbiological analysis, the most common isolated pathogen proved to be Enterococcus faecalis in group A (58.8%) and Escherichia coli in group B (45.5%) (Table 2). The 2 groups differed in terms of isolated strains (P < .001). Six patients in group B exhibited clinical and laboratory parameters indicative for pyelonephritis (2.05%), compared with 2 in group A (0.8%). There was no significant difference between the groups in the upper UTI rate (P = .37).
Clinical Characteristics and Microbiological Data at the End of Follow-up
| Clinical and Microbiological Data | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Symptomatic UTIs | 97 (37.7) | 204 (69.6) | <.001 (54.8/1) |
| QoL score, mean (SD) | 0.83 (0.05) | 0.54 (0.03) | <.01 (83.6/548) |
| Pyelonephritis | 2 (0.77) | 6 (2.0) | .37 (0.78/1) |
| Bacterial strains isolated from symptomatic patients | <.001 (35.6/1) | ||
| No bacterial growth | … | … | … |
| Escherichia coli | 26 (26.8) | 93 (45.5) | … |
| Enterococcus faecalis | 57 (58.8) | 47 (23.0) | … |
| Enterococcus faecium | 7 (7.2) | 2 (0.9) | … |
| Klebsiella spp. | 7 (7.2) | 16 (7.9) | … |
| Streptococcus agalactiae | … | 16 (7.9) | … |
| Serratia spp. | … | 8 (3.9) | … |
| Cytrobacter spp. | … | 5 (2.5) | … |
| Proteus miralibilis | … | 9 (4.5) | … |
| Pseudomonas spp. | … | 8 (3.9) | … |
| Antibiotics tested on E. colib | |||
| Amoxicillin–clavulanic acid | 1/26 (3.8) | 23/93 (24.7) | .03 (4.2/1) |
| Trimethoprim-sulfamethoxazole | 3/26 (11.5) | 32/93 (34.4) | .01 (5.8/1) |
| Amikacin | 3/26 (11.5) | 10/93 (10.7) | .90 (0.01/1) |
| Ciprofloxacin | 5/26 (19.2) | 41/93 (44.0) | .03 (4.3/1) |
| Levofloxacin | 3/26 (11.5) | 14/93 (15.1) | .89 (0.01/1) |
| Gentamicin | 1/26 (3.8) | 15/93 (16.1) | .19 (1.6/1) |
| Piperacillin-tazobactam | 2/26 (7.6) | 9/93 (9.6) | .75 (0.09/1) |
| Imipenem | 0 | 0 | … |
| Cefotaxime | 4/26 (15.3) | 15/93 (16.1) | .92 (0.008/1) |
| Clinical and Microbiological Data | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Symptomatic UTIs | 97 (37.7) | 204 (69.6) | <.001 (54.8/1) |
| QoL score, mean (SD) | 0.83 (0.05) | 0.54 (0.03) | <.01 (83.6/548) |
| Pyelonephritis | 2 (0.77) | 6 (2.0) | .37 (0.78/1) |
| Bacterial strains isolated from symptomatic patients | <.001 (35.6/1) | ||
| No bacterial growth | … | … | … |
| Escherichia coli | 26 (26.8) | 93 (45.5) | … |
| Enterococcus faecalis | 57 (58.8) | 47 (23.0) | … |
| Enterococcus faecium | 7 (7.2) | 2 (0.9) | … |
| Klebsiella spp. | 7 (7.2) | 16 (7.9) | … |
| Streptococcus agalactiae | … | 16 (7.9) | … |
| Serratia spp. | … | 8 (3.9) | … |
| Cytrobacter spp. | … | 5 (2.5) | … |
| Proteus miralibilis | … | 9 (4.5) | … |
| Pseudomonas spp. | … | 8 (3.9) | … |
| Antibiotics tested on E. colib | |||
| Amoxicillin–clavulanic acid | 1/26 (3.8) | 23/93 (24.7) | .03 (4.2/1) |
| Trimethoprim-sulfamethoxazole | 3/26 (11.5) | 32/93 (34.4) | .01 (5.8/1) |
| Amikacin | 3/26 (11.5) | 10/93 (10.7) | .90 (0.01/1) |
| Ciprofloxacin | 5/26 (19.2) | 41/93 (44.0) | .03 (4.3/1) |
| Levofloxacin | 3/26 (11.5) | 14/93 (15.1) | .89 (0.01/1) |
| Gentamicin | 1/26 (3.8) | 15/93 (16.1) | .19 (1.6/1) |
| Piperacillin-tazobactam | 2/26 (7.6) | 9/93 (9.6) | .75 (0.09/1) |
| Imipenem | 0 | 0 | … |
| Cefotaxime | 4/26 (15.3) | 15/93 (16.1) | .92 (0.008/1) |
Abbreviations: QoL, quality of life; SD, standard deviation; UTI, urinary tract infection.
a Data represent No. (%) of patients unless otherwise specified. Group A was untreated; group B, treated.
bE. coli was the most commonly isolated bacterial strain.
Clinical Characteristics and Microbiological Data at the End of Follow-up
| Clinical and Microbiological Data | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Symptomatic UTIs | 97 (37.7) | 204 (69.6) | <.001 (54.8/1) |
| QoL score, mean (SD) | 0.83 (0.05) | 0.54 (0.03) | <.01 (83.6/548) |
| Pyelonephritis | 2 (0.77) | 6 (2.0) | .37 (0.78/1) |
| Bacterial strains isolated from symptomatic patients | <.001 (35.6/1) | ||
| No bacterial growth | … | … | … |
| Escherichia coli | 26 (26.8) | 93 (45.5) | … |
| Enterococcus faecalis | 57 (58.8) | 47 (23.0) | … |
| Enterococcus faecium | 7 (7.2) | 2 (0.9) | … |
| Klebsiella spp. | 7 (7.2) | 16 (7.9) | … |
| Streptococcus agalactiae | … | 16 (7.9) | … |
| Serratia spp. | … | 8 (3.9) | … |
| Cytrobacter spp. | … | 5 (2.5) | … |
| Proteus miralibilis | … | 9 (4.5) | … |
| Pseudomonas spp. | … | 8 (3.9) | … |
| Antibiotics tested on E. colib | |||
| Amoxicillin–clavulanic acid | 1/26 (3.8) | 23/93 (24.7) | .03 (4.2/1) |
| Trimethoprim-sulfamethoxazole | 3/26 (11.5) | 32/93 (34.4) | .01 (5.8/1) |
| Amikacin | 3/26 (11.5) | 10/93 (10.7) | .90 (0.01/1) |
| Ciprofloxacin | 5/26 (19.2) | 41/93 (44.0) | .03 (4.3/1) |
| Levofloxacin | 3/26 (11.5) | 14/93 (15.1) | .89 (0.01/1) |
| Gentamicin | 1/26 (3.8) | 15/93 (16.1) | .19 (1.6/1) |
| Piperacillin-tazobactam | 2/26 (7.6) | 9/93 (9.6) | .75 (0.09/1) |
| Imipenem | 0 | 0 | … |
| Cefotaxime | 4/26 (15.3) | 15/93 (16.1) | .92 (0.008/1) |
| Clinical and Microbiological Data | Patients, No. (%)a | P Value (t or χ2/df) | |
|---|---|---|---|
| Group A (n = 257) | Group B (n = 293) | ||
| Symptomatic UTIs | 97 (37.7) | 204 (69.6) | <.001 (54.8/1) |
| QoL score, mean (SD) | 0.83 (0.05) | 0.54 (0.03) | <.01 (83.6/548) |
| Pyelonephritis | 2 (0.77) | 6 (2.0) | .37 (0.78/1) |
| Bacterial strains isolated from symptomatic patients | <.001 (35.6/1) | ||
| No bacterial growth | … | … | … |
| Escherichia coli | 26 (26.8) | 93 (45.5) | … |
| Enterococcus faecalis | 57 (58.8) | 47 (23.0) | … |
| Enterococcus faecium | 7 (7.2) | 2 (0.9) | … |
| Klebsiella spp. | 7 (7.2) | 16 (7.9) | … |
| Streptococcus agalactiae | … | 16 (7.9) | … |
| Serratia spp. | … | 8 (3.9) | … |
| Cytrobacter spp. | … | 5 (2.5) | … |
| Proteus miralibilis | … | 9 (4.5) | … |
| Pseudomonas spp. | … | 8 (3.9) | … |
| Antibiotics tested on E. colib | |||
| Amoxicillin–clavulanic acid | 1/26 (3.8) | 23/93 (24.7) | .03 (4.2/1) |
| Trimethoprim-sulfamethoxazole | 3/26 (11.5) | 32/93 (34.4) | .01 (5.8/1) |
| Amikacin | 3/26 (11.5) | 10/93 (10.7) | .90 (0.01/1) |
| Ciprofloxacin | 5/26 (19.2) | 41/93 (44.0) | .03 (4.3/1) |
| Levofloxacin | 3/26 (11.5) | 14/93 (15.1) | .89 (0.01/1) |
| Gentamicin | 1/26 (3.8) | 15/93 (16.1) | .19 (1.6/1) |
| Piperacillin-tazobactam | 2/26 (7.6) | 9/93 (9.6) | .75 (0.09/1) |
| Imipenem | 0 | 0 | … |
| Cefotaxime | 4/26 (15.3) | 15/93 (16.1) | .92 (0.008/1) |
Abbreviations: QoL, quality of life; SD, standard deviation; UTI, urinary tract infection.
a Data represent No. (%) of patients unless otherwise specified. Group A was untreated; group B, treated.
bE. coli was the most commonly isolated bacterial strain.
Study schedule according to the Consolidated Standards of Reporting Trials (CONSORT) group recommendation (http://www.consort-statement.org/Media/Default/Downloads/CONSORT%202010%20Statement/CONSORT%202010%20Statement%20(BMJ).pdf). Abbreviation: RCT, randomized clinical trial.
A total of 301 symptomatic patients were treated; the most commonly used antibiotics were fosfomycin trometamol in 85 patients (28.2%), nitrofurantoin in 59 (19.6%), amoxicillin–clavulanic acid in 37 (12.2), cotrimoxazole in 35 (11.6%), and ciprofloxacin in 30 (9.9%). None of the patients reported any adverse effects during antimicrobial therapy; these findings were recorded but not taken into account. There was no difference in the use of antibiotics compared with the previous study (P = .69). However, there was a significant difference in the QoL questionnaire results between the 2 groups at the end of the follow-up period (t = 83.6; df = 548; P < .01).
Antibiotic Resistance Patterns
A total of 301 isolates were analyzed from all patients with symptomatic UTIs during the entire study period. As seen in Table 2, the most common isolated bacteria were Enterobacteriaceae. Table 2 also includes microbiological data, the bacterial spectrum, and resistance against all tested antibiotics for E. coli strains in patients who reported symptomatic rUTIs. At the end of the follow-up period, the proportions of E. coli isolates resistant to amoxicillin–clavulanic acid (P = .03), trimethoprim-sulfamethoxazole (P = .01), and ciprofloxacin (P = .03) were higher in group B than in group A. The difference in antibiotic susceptibility between the 2 groups was significant at 2-year follow-up. Before this cutoff period, no significant differences in antibiotic susceptibility were found.
DISCUSSION
Main Findings
We showed that AB treatment is associated with a higher prevalence of antibiotic-resistant bacteria. To our knowledge, ours is the first study to address this issue in women affected by rUTIs. We demonstrated that the higher rates of symptomatic UTIs in the treated group depended on AB treatment during the previous study, indicating that the incorrect use of antibiotics in AB patients is potentially dangerous, with long-lasting negative effects.
Comparison With Previous Reports
Asscher et al [20] found that in nonpregnant premenopausal women without diabetes antimicrobial treatment of AB can improve short-term microbiological outcome but does not lower the risk of symptomatic UTIs. Cai et al reported in 2012 [4] that AB treatment in patients with rUTIs is associated with a higher rate of symptomatic UTIs after 12 months as well as modification of the isolated bacteria. In particular, growth of multidrug-resistant E. coli was noted after treatment of AB from E. faecalis, regardless of the antibiotic administered [4]. In nonpregnant premenopausal women without diabetes, treatment of AB may be more detrimental than beneficial. Although treatment of AB has been clearly demonstrated to offer no advantage concerning recurrence, no data on antibiotic-resistant bacteria have been collected so far.
Sundvall et al [21] determined that antibiotic treatment during the previous month and hospitalization during the previous 6 months predicted higher resistance rates among uropathogens in elderly nursing home residents. In a systematic review in 2014, Dull et al [22] showed that the risk of symptomatic UTI was slightly higher patients with AB than in nonbacteriuric controls, but treatment of the asymptomatic colonization did not reduce the risk of subsequent symptomatic infection. They also stated that the use of antibiotics in such cases only increases the risk of progressively resistant infections, with no clinical benefit [22].
Beerepoot et al [23] found increased antibiotic resistance rates in these E. coli isolates after 1 month in premenopausal women affected by rUTIs and treated with trimethoprim-sulfamethoxazole, showing that oral administration of antibiotics can cause ecological disturbances in the normal intestinal microflora and promote antimicrobial-resistant strains. Evaluation of the fecal flora may thus be useful in predicting the risk of resistant strains in patients with UTIs. Beerepoot et al [23] did not assess the fecal flora, which could be considered a study limitation. Modification in the normal intestinal flora and the emergence of antimicrobial resistant bacteria may explain why AB treatment can have negative effects after 12 months, with important consequences for the correct use of antibiotics.
In 2005, the Infectious Diseases Society of America published guidelines discouraging screening and treatment in premenopausal nonpregnant women, women with diabetes mellitus, elderly institutionalized residents in long-term care facilities, elderly ambulatory adults, patients with spinal cord injury, and individuals with indwelling urethral catheters [24]. This evidence was based on randomized clinical trials to evaluate the clinical impact of AB treatment and not the impact on microbiological changes. Our results enrich data discouraging the use of antibiotics in women with AB and affected by rUTIs [25]. International guidelines on AB should be implemented to improve the efficacy of antibiotic treatment and reduce bacterial resistance [26, 27].
Study Strengths and Limitations
The present study had some strengths. First, patients who had participated in a previous study were motivated to enroll in this second investigation and could be accurately followed up. Second, because the study was carried out at a single institution, all tests were performed in the same microbiological laboratory. The long follow-up period allowed the differences in bacterial resistance rates between the 2 groups to be highlighted. Indeed, the time lag between antibiotic therapy and possible changes in antibiotic resistance should be borne in mind.
Nevertheless, our study also had a few limitations. There was no control group of healthy patients, and we analyzed a specific population of patients who had already undergone multiple cycles of antibiotic therapy for rUTIs. However, all the enrolled patients resembled the patients with rUTIs commonly found in general clinical practice. We investigated exclusively findings in outpatient women attending our urological institution, and we do not know the effect of AB nontreatment in the general physician setting. Finally, at the time of enrollment, the 2 groups differed in marital status. This parameter is not clinically relevant, however, because the 2 groups proved similar in the number of sexual partners and sexual encounters per week, and marital status is not a prognostic parameter for rUTIs.
In conclusion, our findings show that AB treatment is associated with a higher prevalence of antibiotic-resistant bacteria and is therefore detrimental in women with rUTIs.
Notes
Acknowledgments. We are grateful to all members of the Santa Maria Annunziata Hospital sexually transmitted diseases laboratory for their technical laboratory assistance.
Potential conflicts of interest. All authors: No potential conflicts of interest.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
