Amoxicillin-Clavulanate vs Ciprofloxacin for the Treatment of Uncomplicated Cystitis in Women:  A Randomized Trial FREE

Acute uncomplicated urinary tract infections (UTIs) are among the most commonly encountered bacterial infections in women. Management of these infections has been made more complicated in recent years by increasing antimicrobial resistance, especially to β-lactams and trimethoprim-sulfamethoxazole. The current standard for the empirical treatment of acute uncomplicated cystitis is trimethoprim-sulfamethoxazole for 3 days. Because the prevalence of resistance to trimethoprim-sulfamethoxazole among uropathogens is increasing, however, fluoroquinolones are increasingly being used as empirical therapy for the management of cystitis.^1- 2

Increasing fluoroquinolone resistance in many pathogens causing health care–associated and community-acquired infections has raised concerns about the future usefulness of this potent and frequently prescribed class of antibiotics.³ It is important, therefore, that fluoroquinolone-sparing antibiotics be used when appropriate. In the case of acute cystitis, nitrofurantoin and fosfomycin tromethamine have been advocated as fluoroquinolone-sparing alternatives to trimethoprim-sulfamethoxazole.⁴ Amoxicillin is not recommended for the empirical treatment of UTIs because of the high prevalence of resistance associated with this drug,⁴ but use of amoxicillin-clavulanate is increasing.² However, there are few data on the efficacy of short-course amoxicillin-clavulanate regimens for treatment of cystitis, and there are no published studies on the efficacy of amoxicillin-clavulanate in the treatment of acute cystitis in the current era of high-level amoxicillin resistance. We therefore compared, via a single-blind randomized trial, 3-day regimens of amoxicillin-clavulanate and ciprofloxacin for treatment of women with acute uncomplicated cystitis and used long-term follow-up to evaluate recurrent infection. We also evaluated the effects of these antibiotics on vaginal colonization with Escherichia coli and the relationship of these effects to recurrent infection.

The study was conducted at the University of Washington Student Health Center and at Group Health Cooperative from July 1998 to May 2002. The University of Washington Student Health Center provides primary care for approximately 85% of the 34 000 students enrolled in the university. Group Health Cooperative is a mixed-model health maintenance organization with approximately 475 000 enrollees located in western Washington State.

This treatment trial fulfilled 1 of the 5 primary aims of a large randomized longitudinal study to examine the effects of antimicrobials on UTI recurrence and vaginal flora. For investigating the question of difference in UTI recurrence between the antibiotic treatment groups, the sample size was estimated to provide greater than 90% power to detect a difference of 20% or more between the 2 groups.

Women were eligible for enrollment if they were healthy, between 18 and 45 years of age, and had dysuria, frequency, and/or urgency. Women were ineligible for enrollment if they were pregnant or if they had evidence of pyelonephritis (such as temperature greater than 38°C, severe back pain, or costovertebral angle tenderness), history of significant allergy to a fluoroquinolone or penicillin, a chronic illness requiring medical supervision, a known anatomic or functional abnormality of the urinary tract, or had received systemic or vaginal topical antimicrobials within the previous 14 days.

The Human Subjects Review Committees of the University of Washington and Group Health Cooperative approved the study, and all patients gave written informed consent.

At the initial visit, participants underwent a directed history and physical examination, an interview using a standardized study questionnaire, a midstream urine specimen collection to evaluate bacteriuria and pyuria, and a vaginal swab specimen collection to evaluate bacterial colonization. Race and ethnicity were self-reported on the questionnaire. Participants were randomized to treatment assignments by the statistician, using a blocked randomization scheme with varying block sizes not revealed to clinic personnel. Assignments were placed in sealed, sequentially numbered envelopes, which were opened at the time of enrollment. Women received either amoxicillin-clavulanate (500 mg/125 mg twice daily) or ciprofloxacin (250 mg twice daily) each for 3 days. They were asked to return to the clinic every 2 weeks for 4 months or until they were retreated for symptomatic persistent or recurrent UTI. At each visit, the tests performed at the initial visit were repeated. At the first follow-up visit (2 weeks), participants were asked how many study pills they had taken.

Methods for collecting urine and vaginal specimens and isolating, identifying, and quantifying urine and vaginal uropathogens have been previously described.^5- 6 Vaginal cultures were considered positive for E coli if there was ≥1+ growth on a semiquantitative scale. The Kirby-Bauer disc method was used to determine antimicrobial susceptibility of causative strains. The leukocyte esterase test was performed on each midstream urine specimen and categorized as positive if at least trace.

Genomic DNA isolation, restriction endonuclease digestion, gel electrophoresis, and Southern blot hybridization were performed as previously described.⁷ Total E coli ribosomal RNA (Sigma, St Louis, Mo) was radioactively labeled by extension of random primers using reverse transcriptase.⁸ Ribotyping was performed using PvuII and Eco RI. Only isolates showing 100% match in the restriction fragment length polymorphism patterns were classified as identical to each other.

The primary study outcome was study drug efficacy based on clinical cure. Women were considered to have clinical cure if they did not have symptomatic persistent or recurrent UTI. A woman was considered to have a persistent UTI if, following study drug treatment, she had persistent symptoms warranting retreatment. A woman was considered to have a recurrent UTI if she had resolution of her initial UTI symptoms after treatment but then developed recurrent symptoms warranting retreatment. A persistent or recurrent UTI was considered to be culture-confirmed if there were at least 10² colony-forming units (CFU) of uropathogens per milliliter of urine.

Secondary study outcomes were (1) study drug efficacy based on microbiological cure at the first posttreatment visit (2 weeks); (2) vaginal colonization with E coli at all posttreatment visits; and (3) the association between vaginal colonization with E coli and symptomatic persistent or recurrent UTI. Women were considered microbiologically cured if, at an asymptomatic 2-week follow-up visit, they had less than 10⁵ CFU/mL of all uropathogens and at least a 1-log drop in colony count of the causative uropathogen compared with the urine culture at enrollment. Women who had persistent or recurrent UTI symptoms warranting retreatment before the 2-week visit were also classified as microbiologically cured if their urine cultures had less than 10² CFU/mL and they were not taking antibiotics at the time of the urine culture.

The associations between antimicrobial susceptibility of the initially infecting strains and clinical and microbiological outcomes were evaluated. For culture-confirmed UTIs, uropathogens with the highest colony count were considered the causative pathogen, but if 2 or more organisms had the same colony count (within <1 log) they were considered copathogens. Coagulase-negative staphylococci, α-hemolytic streptococci, lactobacilli, diphtheroids, and mixed gram-positive flora were categorized as nonuropathogens at enrollment and follow-up.

All women who were enrolled and randomized to receive treatment were included in analyses of demographic and behavioral characteristics and UTI history, using means and frequency distributions. Women were included in the analysis if they met enrollment criteria, had at least 10² CFU/mL of uropathogens in their enrollment urine, and returned to the clinic for at least 1 postrandomization clinic visit. Infecting uropathogens at the enrollment UTI were also described, as well as their antibiotic susceptibility profiles.

Kaplan-Meier curves were constructed to examine the cure rates over the follow-up period, and the likelihood ratio test was used to test for differences between treatment groups. Comparisons of proportions between the 2 treatment groups were tested using the χ² statistic. Women were followed up until they were retreated for persistent or recurrent UTI symptoms or the end of the study. SAS version 8.2 (SAS Institute, Cary, NC) was used for all analyses. Statistical significance was considered to be P<.05.

Three hundred seventy women (335 at the student health center and 35 at Group Health Cooperative) were enrolled and randomized to receive treatment (Figure 1). Women randomized to the 2 treatment groups and included in the efficacy analyses were similar with respect to baseline characteristics (Table 1). Forty-eight women were excluded from the analyses: 39 had less than 10² CFU/mL at enrollment and 9 had no follow-up. The 322 women eligible for the analyses (160 randomized to receive amoxicillin-clavulanate and 162 randomized to receive ciprofloxacin) were followed for a median of 103 days (range, 2-125 days). Compliance with at least 5 of the 6 doses of study drug was 99% in both groups. Women taking fewer than 6 doses of study drug were included in the analyses.

Among the 370 women enrolled, 331 women had 1 or more uropathogens present in quantities of at least 10² CFU/mL. Escherichia coli was the only pathogen or copathogen in 270 (82%), group B streptococci in 28 (8%), Staphylococcus saprophyticus in 26 (8%), enterococci in 8 (2%), and Klebsiella species and Proteus mirabilis in 7 (2%) each. Overall susceptibility data for the 323 isolates tested are shown in Table 2. Susceptibility data for the E coli strains were almost identical. Among the 99 isolates that were resistant to ampicillin, 36% were susceptible, 33% were intermediate, and 30% were resistant to amoxicillin-clavulanate. Among the 31 isolates that were intermediately resistant to ampicillin, 65% were susceptible to amoxicillin-clavulanate.

Clinical Outcomes. The rate of clinical cure was significantly lower in the amoxicillin-clavulanate group (93 [58%] of 160) than in the ciprofloxacin group (124 [77%] of 162) (P<.001) (Figure 2). The women in the amoxicillin-clavulanate group had a significantly lower clinical cure rate compared with women in the ciprofloxacin group regardless of whether the patients were infected with strains that were susceptible (65 [60%]of 109 women in the amoxicillin-clavulanate group vs 114 [77%] of 149 women in the ciprofloxacin group; P  = .004) or nonsusceptible (34 women) (P = .001 vs ciprofloxacin) to amoxicillin-clavulanate (Figure 2). Although there was a trend toward a better clinical outcome in women treated with amoxicillin-clavulanate if they were infected with strains susceptible to amoxicillin-clavulanate, this difference was not statistically significant (P = .17) (Figure 2). In addition, among the 109 women in the amoxicillin-clavulanate group infected with amoxicillin-clavulanate–susceptible uropathogens, cure rates were not associated with amoxicillin susceptibility of the infecting strains.

Persistent UTIs occurred in 8 women treated with amoxicillin-clavulanate (3 had susceptible strains at enrollment) and in 1 woman treated with ciprofloxacin (the strain was susceptible) (Table 3). Recurrent UTIs occurred in 59 women in the amoxicillin-clavulanate group and 37 women in the ciprofloxacin group. Among the 67 persistent or recurrent UTIs in the amoxicillin-clavulanate group, 33 (50%) occurred within the first 2 weeks (Table 3) and 34 occurred during weeks 3 through 10 following treatment compared with 3 (8%) and 35 (92%), respectively, of the 38 persistent or recurrent UTIs in the ciprofloxacin group. Persistent and recurrent UTIs were culture-confirmed in 51 (76%) women treated with amoxicillin-clavulanate compared with 30 (79%) women treated with ciprofloxacin. Pyuria was present in 92% of the women who had persistent or recurrent UTI in each of the 2 treatment groups.

In the amoxicillin-clavulanate group, all 3 pairs of initial and persistent UTI strains tested and 16 (73%) of 22 pairs of initial and recurrent UTI strains tested had the same ribotype.

Five women infected with amoxicillin-clavulanate–susceptible strains of E coli at enrollment who were treated with amoxicillin-clavulanate had recurrent UTIs 7 to 44 days after treatment caused by E coli strains that were nonsusceptible to amoxicillin-clavulanate. Isolates from only 1 such patient had the same ribotype. Two women treated with ciprofloxacin whose original uropathogens were susceptible to ciprofloxacin developed recurrent UTI with ciprofloxacin-nonsusceptible strains 20 and 29 days after treatment—both were different species from the originally infecting strains.

Two women in the amoxicillin-clavulanate group who at 2 weeks after treatment had persistent bacteriuria with at least 10⁵ CFU/mL of E coli and pyuria but no urinary symptoms were subsequently diagnosed with and treated for pyelonephritis. Because they were asymptomatic, these 2 women did not receive treatment at the 2-week visit. One woman presented 3 days after her 2-week follow-up visit to the emergency department with subjective fever and chills and acute onset of back pain and was found to have a temperature of 38.5°C, bilateral mild costovertebral angle tenderness, pyuria, and greater than 10⁵ CFU/mL of E coli, which was susceptible to amoxicillin-clavulanate. Her initial, 2-week follow-up, and pyelonephritis strains were identical by ribotype. The other woman presented to an outside hospital 5 days after her 2-week visit with subjective fever and shaking chills, urgency, and low back pain and reported that she had a temperature of 39.8°C. We were not able to obtain her outside medical records or laboratory isolate, but her initial and 2-week isolates were identical by ribotype.

Microbiological Outcomes. Microbiological cure was also inferior in the amoxicillin-clavulanate group. At the 2-week posttreatment visit, microbiological cure was observed in 118 (76%) of 156 women in the amoxicillin-clavulanate group compared with 153 (95%) of 161 women in the ciprofloxacin group (P<.001) (Table 3). In the amoxicillin-clavulanate group, microbiological cure occurred in 77 (73%) of 106 women infected with susceptible strains compared with 27 (82%) of 33 women infected with nonsusceptible strains (P = .30). In the amoxicillin-clavulanate group, 66% of the women with microbiological failure had symptomatic persistent or recurrent UTI compared with 13% in the ciprofloxacin group (Table 3).

Vaginal Colonization withE coli. Vaginal colonization with E coli was present at enrollment in 82% (130/159) of the women in the amoxicillin-clavulanate group compared with 86% (139/162) in the ciprofloxacin group (Figure 3). By the 2-week posttreatment follow-up visit, however, 45% (68/151) of the women in the amoxicillin-clavulanate group compared with 10% (16/153) in the ciprofloxacin group had vaginal colonization (P<.001). This difference did not narrow until week 10 after treatment. Antimicrobial susceptibility testing was not performed on vaginal isolates at enrollment and, thus, we were not able to determine the relationship between antimicrobial susceptibility of the vaginal strains at enrollment and the prevalence of posttreatment colonization.

There was a trend toward presence of vaginal colonization with E coli at the 2-week posttreatment visit and subsequent recurrent UTI over the next month. Thus, recurrent UTIs occurred in 44 (21%) of 213 women who did not have vaginal colonization with E coli at the 2-week visit compared with 22 (31%) of 70 who had colonization (P = .06).

Adverse Events. When asked an open question about whether the participant had adverse effects related to study medications, women in the amoxicillin-clavulanate group were more likely to report symptoms (27%) compared with women in the ciprofloxacin group (19%) (P = .06). Differences were due mostly to increased loose stools and vaginal symptoms in the women treated with amoxicillin-clavulanate. Loose stools were reported in 8% and 0.6%, vaginal symptoms in 9% and 3%, nausea in 5% and 8%, any central nervous symptoms in 3% and 6%, and fever or rash in 1% and 0% of amoxicillin-clavulanate–treated and ciprofloxacin-treated women, respectively. Twenty-two amoxicillin-clavulanate–treated women required treatment for adverse effects compared with 10 ciprofloxacin-treated women. Only 2 women in the amoxicillin-clavulanate group and 1 woman in the ciprofloxacin group discontinued study medication due to adverse effects.

This is the first comparison of a 3-day regimen of amoxicillin-clavulanate vs a 3-day regimen of a fluoroquinolone for the treatment of acute uncomplicated cystitis in women. Amoxicillin-clavulanate was markedly inferior to ciprofloxacin in clinical and microbiological cure rates as well as in eradication of vaginal E coli. Of concern, 2 women developed same-strain pyelonephritis soon after treatment with amoxicillin-clavulanate (same-strain is presumed in 1 woman based on a urine culture 5 days before the episode of pyelonephritis). The difference in clinical cure occurred primarily within the first 2 weeks of treatment and was most likely due to the inferior activity of amoxicillin-clavulanate in eradicating E coli from the vagina. At the 2-week follow-up visit, 45% of amoxicillin-clavulanate–treated women had vaginal colonization compared with 10% of ciprofloxacin-treated women (Figure 3). In this regard, amoxicillin-clavulanate is similar to ampicillin and amoxicillin, drugs that also do not effectively eradicate E coli from the vaginal reservoir and are associated with high recurrence rates.⁹

Although not evaluated in this study, another factor that may have contributed to the poorer results seen with amoxicillin-clavulanate is the much shorter serum half-life of amoxicillin compared with ciprofloxacin (1.3 hours for amoxicillin, 1 hour for clavulanate, and 4 hours for ciprofloxacin) and, thus, the shorter period of time that amoxicillin is present in high concentrations in urine. Although concerns have been raised about the effectiveness of amoxicillin-clavulanate in the treatment of UTIs caused by amoxicillin-resistant E coli,¹⁰ we were unable to demonstrate that amoxicillin-resistant strains were associated with higher failure rates.

Finally, it has been demonstrated in an experimental mouse model that E coli, following inoculation of large quantities into the bladder, can penetrate into the bladder epithelium and produce intracellular colonies of bacteria and eventually biofilms.^11- 12 If epithelial invasion and biofilm formation occurs in human cystitis, it is possible that amoxicillin-clavulanate is less effective than ciprofloxacin in eradicating uropathogens from these sites, resulting in higher clinical failure rates.

The worldwide increase in trimethoprim-sulfamethoxazole resistance in E coli and concerns about fluoroquinolone resistance in UTIs and other more serious infections³ highlight the need for alternative antibiotics for the treatment of cystitis. Nitrofurantoin and fosfomycin tromethamine have both been advocated as fluoroquinolone-sparing alternatives to trimethoprim-sulfamethoxazole, but neither appears to be as effective as trimethoprim-sulfamethoxazole or fluoroquinolones.^4,13 Although β-lactam antibiotics have long been used for the treatment of UTIs, lower UTI cure rates have generally been observed compared with trimethoprim-sulfamethoxazole and fluoroquinolones regardless of the duration of treatment.⁴ Amoxicillin-clavulanate has been used for the treatment of UTIs and other infections since the early 1980s, but most published trials of its efficacy have been very small, in children or in mixed populations of patients, including those with complicated UTIs.^14- 23 In studies of uncomplicated cystitis, a 3-day regimen of amoxicillin-clavulanate and a 7-day regimen of trimethoprim-sulfamethoxazole have been shown to be superior to a single-dose regimen of amoxicillin-clavulanate.^24- 25 Exposure to amoxicillin-clavulanate has been shown previously to select for subsequent amoxicillin-clavulanate–resistant E coli UTIs.²⁶ In our study, 5 women originally infected with amoxicillin-clavulanate–susceptible strains who were treated with amoxicillin-clavulanate were found to have recurrent UTIs with strains nonsusceptible to amoxicillin-clavulanate, but only one was genetically identical to the initial strain.

We chose to compare 3-day regimens of amoxicillin-clavulanate and ciprofloxacin for the treatment of acute uncomplicated cystitis because of the need for safe and effective fluoroquinolone-sparing antibiotics, the absence of a large trial comparing a 3-day regimen of amoxicillin-clavulanate to current standard therapy, and increasing amoxicillin-clavulanate use for the treatment of UTIs over the past decade.² The prevalence of amoxicillin-clavulanate resistance among causative strains at baseline (9%) was in the same range as that described in recent European countries (3.4% [range, 0%-9.3%]).²⁷ The rate of amoxicillin-clavulanate resistance appeared to have limited impact on both clinical and microbiological outcomes. In fact, amoxicillin-clavulanate was found to be significantly inferior to ciprofloxacin even in women infected with uropathogens susceptible in vitro to amoxicillin-clavulanate.

Strengths of this study include its large sample size, well-defined study population, long follow-up period, low dropout rate, high rate of medication adherence, and genetic analysis of strains causing persistent and early recurrent UTIs. The single-blind study design is a potential weakness of the study if there was a clinical bias against amoxicillin-clavulanate and a subsequent lower threshold to treat women in this group who had mild symptoms. This seems unlikely since the percentage of UTIs that were culture-confirmed and the percentage that were associated with pyuria were almost identical in the 2 treatment groups.

Trimethoprim-sulfamethoxazole should continue to be the first-line treatment for acute cystitis if the woman has no history of allergy to the drug and if the likelihood of trimethoprim-sulfamethoxazole resistance is low.^4,28 In areas where the likelihood of trimethoprim-sulfamethoxazole resistance is high (>20%) or in women who have risk factors for trimethoprim-sulfamethoxazole resistance, nitrofurantoin or a fluoroquinolone is an appropriate choice. Amoxicillin-clavulanate in a 3-day regimen is not as effective as a fluoroquinolone for the treatment of acute uncomplicated cystitis, even in those women with UTIs caused by susceptible uropathogens, and should be considered only when use of other first- and second-line antibiotics is not feasible.