Blue Ribbon Abstract Award: Are CDC Criteria Useful for Diagnosis or Surveillance of Catheter Associated Urinary Tract Infections?

BACKGROUND: Urinary tract infections (UTI) are reported to be the most common nosocomial (N) infection. Catheter-associated UTI (CA-UTI) is the most common N UTI, yet there are no good criteria for its diagnosis. Current definitions often identify those with colonized Foley catheters (FC) and may overstate the incidence of N UTI. The goals of this study were to test different case definitions of CA-UTI and to develop a more relevant definition of CA-UTI. METHODS: The microbiology database was queried for unique (+) urine cultures (UrCx) from FC placed in-hospital over two 3-month periods (April–June 2002 and 2003). Records were queried for: urinalysis, symptoms, other infections, treatment, and UTI diagnosis. UTI <48 hours of admission were excluded. Data were analyzed using CDC criterion 1 and criterion 2* for N UTI and “clinical criteria.” Clinical criteria were defined as: ≥1 of the following without another cause: fever (>38°C) or symptoms with a (+) UrCx for a pathogen, and pyuria (≥10 WBC/mm3) and treatment with an appropriate antimicrobial. For this study, the reference standard for CA-UTI was: symptoms, bacteriuria with a pathogen (≥105 CFU/mL), pyuria (≥10 WBC/mm3), appropriate antimicrobial therapy and absence of other infections. RESULTS: 93 potential cases were found; 26 met the CA-UTI case definition. CDC criterion 1 was sensitive (26/26, 100%) but there were 43 false positives for a positive predictive value (PPV) = 38%. CDC criterion 2 was not sensitive (1/26, 4%) and PPV = 20% (1/4). NPV = 64% (63/98). Clinical criteria were sensitive (26/26, 100%) and specific (63/67, 94%). PPV = 87% (26/30) and NPV = 100% (63/63). The incidence of CA-UTI using CDC criteria was 2.7 times greater than that using clinical criteria. CONCLUSIONS: For CA-UTI, CDC criteria had limitations. Compared to case definition, CDC criterion 1 was sensitive but not specific; criterion 2 was specific but not sensitive. Both had low PPV. Clinical criteria were sensitive and more specific. Compared to our case definition, current CDC criteria overestimate the incidence of CA-UTI. Development of a more useful definition of CA-UTI for clinical and infection control use is needed. *RP Gaynes, TC Horan. Surveillance of nosocomial infections. In Mayhall CG, ed., Hospital Epidemiology and Infection Control. Baltimore: Williams & Wilkins, 1996, 1017-31, App-A-1-14.