Control of an Outbreak of Infection with the Hypervirulent Clostridium difficile BI Strain in a University Hospital Using a Comprehensive "Bundle" Approach

Abstract

BACKGROUND: In June 2000, the hospital-acquired Clostridium difficile (CD) infection rate in our hospital (University of Pittsburgh Medical Center-Presbyterian, Pittsburgh, PA) increased to 10.4 infections per 1000 hospital discharges (HDs); the annual rate increased from 2.7 infections per 1000 HDs to 7.2 infections per 1000 HDs and was accompanied by an increase in the frequency of severe outcomes. Forty-seven (51%) of 92 HA CD isolates in 2001 were identified as the "epidemic BI strain." A comprehensive CD infection control "bundle" was implemented to control the outbreak of CD infection. METHODS: The CD infection control bundle consisted of education, increased and early case finding, expanded infection-control measures, development of a CD infection management team, and antimicrobial management. Process measures, antimicrobial usage, and hospital-acquired CD infection rates were analyzed, and CD isolates were typed. RESULTS: The rates of compliance with hand hygiene and isolation were 75% and 68%, respectively. The CD management team evaluated a mean of 31 patients per month (11% were evaluated for moderate or severe disease). Use of antimicrobial therapy associated with increased CD infection risk decreased by 41% during the period 2003-2005 (P<.001). The aggregate rate of CD infection during the period 2001-2006 decreased to 4.8 infections per 1000 HDs (odds ratio, 2.2; 95% confidence interval, 1.4-3.1; P<.001) and by 2006, was 3.0 infections per 1000 HDs, a rate reduction of 71% (odds ratio, 3.5; 95% confidence interval, 2.3-5.4; P<.001). During the period 2000-2001, the proportion of severe CD cases peaked at 9.4% (37 of 393 CD infections were severe); the rate decreased to 3.1% in 2002 and further decreased to 1.0% in 2006--a 78% overall reduction (odds ratio, 20.3; 95% confidence interval, 2.8-148.2; P<.001). In 2005, 13% of CD isolates were type BI (20% were hospital acquired), which represented a significant reduction from 2001 (P<.001). CONCLUSIONS: The outbreak of CD infection with the BI strain in our hospital was controlled after implementing a CD infection control "bundle." Early identification, coupled with appropriate control measures, reduces the rate of CD infection and the frequency of adverse events.