1580 for Preventing Recurrent Urinary Tract Infection: A Meta-Analysis X Julia Steinrücken1, Andrew Atkinson1, Annette Kuhn2, Jonas Marschall1

1 Department of Infectious Diseases, Bern University Hospital, Bern, Switzerland 2 Department of Gynecology, Bern University Hospital, Bern, Switzerland Contact Information: J. Steinrücken, MD. E-mail: [email protected]

Background Figure 1: Forest plots for Placebo-controlled (left) and Head-to-head (right) studies • Urinary tract infections (UTI) are a common healthcare problem • Recurrent UTI (RUTI) in healthy non-pregnant women is defined as ≥ 3 UTI episodes during one year The scientific literature on randomized trials for prophylaxis of RUTI has not been screened systematically in >10 years • Providers may choose from a wide range of options, including and non- antibiotic preparations • Long-term antibiotics have been proposed as optimal prevention strategy for RUTI • Preferable antibiotic choices are poorly characterized

Methods • Search terms: “recurrent”, “UTI”, “prophylaxis”, “antibiotic”, and “RCT” • Screening of: MEDLINE, EMBASE, the Cochrane Library, clinicaltrials.gov and reference lists of retrieved articles • We considered any published RCT in adults where antibiotics were used as RUTI Summarizing the 14 placebo-controlled (PC) studies, the risk ratio for developing In the 9 head-to-head trials of different prophylactic antibiotics, was the single UTI was 0.18 (95% CI 0.11-0.29); the corresponding overall risk reduction was 52% (NNT 1.99, most common comparator [to (3 studies), cefaclor (2) or (TMP) / prophylaxis 95% CI 1.83-2.29). After 2004, a single RCT, comparing fosfomycin to placebo, was conducted and trimethoprim combination (SMZ / TMP) (4)]; there was no significant • Data on RUTI episodes in both comparators (antibiotic vs. antibiotic or placebo) were and revealed an absolute UTI risk reduction of 68% (NNT 1.5). treatment difference [RR 1.10 (0.84-1.44)] * not listed: Seppänen et al. (Cinoxacin vs trimethoprim )1988 extracted from the selected articles • Statistical analyses: random effects model, with the results expressed as risk ratio Heterogeneity was examined using the Q and I2 statistics, and considered significant for p-values less than 0.1. Funnel plots and the Egger test were employed to investigate (RR) with 95% confidence intervals (CI) potential publication bias, with a sensitivity analysis based on the “Trim and Fill” method, and the Copas selection model. All analyses were performed in R.

Flow Chart Table 1: Selection of studies for the meta-analysis Conclusions

Number of Risk ratio 95% CI p-value • Over the last decade, Total hits in Pubmed, Type of comparison only very few RCTs have MEDLINE, EMBASE, the studies (RR) Cochrane Library, been added to the scientific clinicaltrials.gov and reference A. Placebo-controlled 14 0.18 (0.11, 0.29) < 0.001 literature lists of retrieved articles (101) • Antibiotic prophylaxis for - no antibiotic prophylaxis (56) A1. Placebo-controlled excluding cinoxacin° 9 0.15 (0.10, 0.23) < 0.001 the prevention of recurrent excluded - complicated UTI (6) B. Head-to-head 10 UTI confers a 50% risk - study design not meeting inclusion reduction (NNT=2) criteria (12) B1. Nitrofurantoin vs others 9 1.10 (0.83, 1.44) 0.49 • Head-to-head trials were - other (2) mainly published for controlled trials (21) B2. Trimethoprim (± Sulfamethoxazole) vs others 5** 0.77 (0.52, 1.15) 0.21 nitrofurantoin vs. B3. Norfloxacin vs others 3 1.17 (0.59, 2.33) 0.66 comparators and show no difference in RUTI rates C. Post-coital vs continuous 1 1.39 (0.24, 8.07) 0.71 • Nitrofurantoin, norfloxacin and TMP/SMZ D. Post-coital vs nothing 1 0.004 (0.0006, 0.03) < 0.001 Placebo - controlled Head - to - head Postcoital prophylaxis appear to be (14 *) (10 *) (2) Total number of comparisons 26* from 24 distinct RCTs. interchangeable options * The three-arm study by Stamm et al. was included twice in the placebo controlled comparison (vs Nitro and TMP-SMZ), and once in the head-to- head comparison (TMP vs Nitro). (in terms of efficacy) ** Two studies compare TMP with Nitro, two TMP-SMZ with Nitro, one TMP with CNO

* Stamm et al. Urinary prophylaxis with trimethoprim and trimethorpim-sulfamethoxazole: efficacy, Disclosure Statement: The authors have nothing to disclose. The study was conducted without dedicated funding influence on the natural history of recurrent bacteriuria, and cost control. 1982 Literature: Albert et al. Antibiotics for preventing recurrent UTI in non-pregnant women. Cochrane Reviews, 2004.