Clostridium difficile infection (CDI) is one of the most common nosocomial infections worldwide. While exposure to antimicrobials is the most important risk factor for CDI, the magnitude of the risk from different antimicrobials has not been well quantified through big data analysis of large healthcare systems.
We conducted a retrospective cohort study of all inpatients with no recent history of CDI in the US Department of Veterans Affairs Health Care System admitted between January 1, 2008 and December 31, 2013. For patients with multiple hospitalizations, only the first hospitalization during the study period was considered. Patients were followed until the development of hospital-acquired CDI (HA-CDI), death, or discharge, whichever came first. HA-CDI was defined as a laboratory test indicating the presence of toxin or toxin genes from a stool sample collected on hospital day 4 or later. Antimicrobial exposures were assessed daily for current (“on”) or recent (“post”) exposures by class. The impact of time-varying antimicrobial exposure on the risk of HA-CDI was assessed using multivariable Cox proportional hazards regression models with robust covariance estimation. Patient factors, such as age and comorbidity, were included as adjustors. Only patient-days at risk for HA-CDI (i.e., day 4 or later) were included.
Approximately 2.8 million patient-days from 476,679 patients were included in the analysis. Table 1 shows the impact of on- and post-antibiotic exposures by class on the risk of HA-CDI after accounting for patient factors, including concomitant antimicrobial exposures.
We observed risks of HA-CDI associated with cephalosporins and fluoroquinolones lower than previously reported. Tetracycline exposure appears to be protective. This big data analysis from the nationwide VA healthcare system helps to better quantify the risk of CDI during and after receiving different categories of antimicrobials. This work could better guide antimicrobial selection and antimicrobial stewardship efforts, potentially reducing the risk of CDI among patients
M. Leecaster, None
B. Sauer, None
M. Rubin, None