1283. Making a Clinical Ranking of Treatments for Uncomplicated Urinary Tract Infections in Primary Care: Network Meta-analysis of Randomized Trials
Session: Poster Abstract Session: Urinary Tract Infections
Saturday, October 22, 2011
Room: Poster Hall B1
  • bartknottnerus_uti_network_111018.pdf (34.6 kB)
  • Background: For clinical treatment decisions, ideally, a physician needs to know the effectiveness of each possible treatment in comparison with all relevant alternatives. However, since classic meta-analyses are based on direct, head-to-head comparisons of treatments, they do not provide data that rank the different clinical competitors. Using a novel application of logistic regression, we performed a network meta-analysis to compare different antibiotic regimens for uncomplicated urinary tract infections (UTIs) in primary care simultaneously and derive a clinically applicable ranking.

    Methods: We included all relevant randomized controlled trials (RCTs) that compared any oral antibiotic or placebo regimens for UTI in primary care, and connected them in a network structure. Using the studies’ contingency tables, an individual-patient-based logistic regression analysis was performed for five binary outcomes: early clinical, early bacteriological, late clinical, and late bacteriological cure; and adverse effects. To preserve randomization within each trial, a dummy variable for each study was included. Fore each of the five outcomes, a ranking of the regimens within the network was made.

    Results: Not all treatments could be connected, which resulted in two separate sub-networks consisting of 13 different treatments from nine RCTs in total. The following significant results were found: low clinical effectiveness of a 3-day amoxicillin-clavulanate regimen, low bacteriological effectiveness of a 3-day nitrofurantoin regimen, and a high rate of adverse events of a 7-day trimethoprim-sulfamethoxazole (TMP/SMX) regimen.

    Conclusion: Network meta-analysis using logistic regression is a useful and relatively simple tool to derive a clinically applicable ranking of treatments. The method may enhance the practical usefulness of clinical guidelines, reveals future research needs, and can be used for meta-analyses of RCTs in other research areas.

    Subject Category: J. Clinical practice issues

    Bart Knottnerus, MD1, Larissa Grigoryan, MD PhD2, Suzanne Geerlings, MD, PhD1, Eric Moll van Charante, MD PhD1, Theo Verheij, MD PhD2, Alphons Kessels, MD MSc3 and Gerben ter Riet, MD PhD4, (1)Academic Medical Center - University of Amsterdam, Amsterdam, Netherlands, (2)Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, Netherlands, (3)Maastricht University Medical Center, Maastricht, Netherlands, (4)Department of General Practice, Academic Medical Center - University of Amsterdam, Amsterdam, Netherlands


    B. Knottnerus, None

    L. Grigoryan, None

    S. Geerlings, None

    E. Moll van Charante, None

    T. Verheij, None

    A. Kessels, None

    G. ter Riet, None

    << Previous Abstract | Next Abstract

    Findings in the abstracts are embargoed until 12:01 a.m. EST Thursday, Oct. 20 with the exception of research findings presented at IDSA press conferences.