495. Predictors of C. difficile Infection and Impact of Primary Prophylaxis Among Asymptomatic C. difficile Colonized Patients: A Cross-Sectional Study

BACKGROUND: Patients who are colonized with C. difficile are at risk of developing C. difficile infections (CDI), but factors associated with disease onset are poorly understood. The objectives of this study were to identify predictors of hospital-onset CDI (HO-CDI) among asymptomatic C. difficile colonized patients and explore the potential benefit of primary prophylaxis to prevent CDI. METHODS: We performed a retrospective cross-sectional study of C. difficile colonized patients admitted to a tertiary academic institution in Quebec City between November 2013 and January 2017. Colonization status was determined upon hospital admission through a systematic screening program by detecting the TcdB gene by PCR on a rectal swab. Primary prophylaxis was defined as the preventive use of ≥1 dose of oral vancomycin or metronidazole in a patient without diarrhea. The choice and dosing of prophylaxis were left to the discretion of the treating physician. Univariate and multivariate logistic regression analyses were used to determine independent predictors of HO-CDI. RESULTS: Of 513 C. difficile colonized patients, 39 (7.6%) developed a HO-CDI, with a 30-day attributable mortality of 18%. We found that an increased length of hospital stay (adjusted odds ratio [aOR] per day, 1.03; P = 0.006), exposure to multiple classes of systemic antibiotics (aOR per class of antibiotic, 1.45; P = 0.03), the use of opioid analgesics (aOR, 2.70; P = 0.01) and cirrhosis (aOR, 5.57; P = 0.007), were independently associated with an increased risk of HO-CDI in multivariate analysis, whereas the use of laxatives was associated with a lower risk of CDI (aOR, 0.36; P = 0.01). Among the antimicrobials, B-lactam with B-lactamase inhibitors (OR, 3.65; P < 0.001), first-generation cephalosporins (OR, 2.38; P = 0.03), and carbapenems (OR, 2.44; P = 0.03) correlated with the greatest risk of HO-CDI. In contrast, patient age, exposure to proton pump inhibitors, and the use of prophylaxis were not significantly associated with occurrence of HO-CDI in this specific population. CONCLUSION: This study identifies several variables that are specifically associated with the development of CDI among C. difficile colonized patients. Whether modifying these risk factors could help prevent CDI should be further investigated. DISCLOSURES: S. Trottier, CIHR: Grant Investigator, Research grant. V. Loo, Merck: Consultant and Scientific Advisor, Consulting fee. Y. Longtin, Merck: Grant Investigator, Research grant. Becton Dickinson: Grant Investigator, Grant recipient.