1070. High Incidence of Community-Onset Urinary Tract Infections Following Hospitalization
Session: Poster Abstract Session: Surveillance of HAIs: Evaluating National Strategy
Friday, October 4, 2013
Room: The Moscone Center: Poster Hall C
  • McGregor_CO UTI_ IDWeek 2013_Final.pdf (156.2 kB)
  • Background: Urinary tract infections (UTI) represent a large portion of healthcare-associated infections.  Surveillance for healthcare-associated UTI (hUTI) is typically limited to those infections that are hospital onset (HO).  The purpose of this study was to assess the potentially unidentified fraction of hUTI that is community onset (CO). 

    Methods:   We assembled a retrospective cohort of adults hospitalized at Oregon Health & Science University between May 2009 and Dec 2011 who receive primary care in the Department of Family Medicine.  Pregnant females and patients with a history of chronic cystitis, chronic pyelonephritis, or UTI diagnosis in the past 30 days were excluded.  UTIs were identified by ICD-9 diagnosis codes.  Patients not admitted or diagnosed with a UTI during the index hospitalization were considered at risk for CO-hUTI following discharge.  HO-hUTI were identified during the index admission using culture or urinalysis dates to establish time of infection onset as greater than 48 hours post admission.  Cumulative incidence of hUTI was calculated as the proportion of HO-hUTI and CO-hUTI among the total cohort at risk.  The proportion of CO-hUTI among all hUTI was calculated to represent the potentially unidentified fraction of hUTI. Uropathogen susceptibilities were also compared (Fisher’s exact test).

    Results: Among the 4,126 patient admissions in the cohort, 46 (1.1%) had HO-hUTI during the index admission and 73 (1.8%) experienced a CO-hUTI.  Of cases of CO-hUTI, 54.8% were diagnosed within 14 days post-discharge.  The cumulative incidence of HA-UTI in this cohort was 3.1% (95%CI: 2.6-3.7%).  Thus, 73/120 (60.8%) of hUTI were CO and represent the potentially unidentified fraction.  Escherichia coliwas the most frequently isolated pathogen at 39.1% of HO-hUTI and 34.9% of CO-hUTI.  Susceptibility to trimethoprim/sulfamethoxazole (88.9 v. 80.0%; p=0.64) and ciprofloxacin (83.3% v. 73.3%; p=0.67) did not differ between HO- and CO-hUTI.

    Conclusion: Current surveillance methodologies for healthcare-associated infections are typically limited to the period of hospitalization. This approach may underestimate true hUTI incidence.  Further work is needed to confirm the probable CO-hUTI infections as true cases of hUTI.

    Jessina C. Mcgregor, PhD1,2, John M. Townes, MD3, David T. Bearden, PharmD1, Miriam R. Elman, MPH1, John D. Heintzman, MD4, Jodi A. Lapidus, PhD2 and Craig D. Williams, PharmD1,4, (1)Dept. Pharmacy Practice, Oregon State University College of Pharmacy, Portland, OR, (2)Dept. Public Health & Preventive Medicine, Oregon Health & Science University School of Medicine, Portland, OR, (3)Div. Infectious Diseases, Dept. Medicine, Oregon Health & Science University School of Medicine, Portland, OR, (4)Dept. Family Medicine, Oregon Health & Science University School of Medicine, Portland, OR


    J. C. Mcgregor, None

    J. M. Townes, None

    D. T. Bearden, None

    M. R. Elman, None

    J. D. Heintzman, None

    J. A. Lapidus, None

    C. D. Williams, None

    Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.