The epidemiology of Clostridium difficile infection in Australia

Background: Clostridium difficile was traditionally considered a nosocomial pathogen and research on C. difficile infection (CDI) has largely focused on symptomatic hospital-associated (HA)-CDI. Recent studies have pointed out the importance of asymptomatic C. difficile colonisation and community-associated (CA)-CDI in C. difficile epidemiology, yet our current understanding of these components is limited. Therefore, the objectives of my research were: 1) to identify risk factors associated with asymptomatic C. difficile colonisation; 2) to compare the predominant C. difficile ribotypes between symptomatic and asymptomatic patients; 3) to determine the spatio-temporal distribution of CA-CDI in Queensland and to examine its association with medication exposure at a population level; and 4) to investigate novel therapeutical options and risk factors for CDI. Methods: I analysed datasets from different sources: 1) a prospective three-year repeated cross-sectional study in hospitalised patients; 2) a three-year longitudinal surveillance study of symptomatic CDI in the hospitals and the communities; 3) CDI data from Sullivan Nicolaides Pathology and quantities of medications prescribed in Queensland from the Pharmaceutical Benefits Scheme; and 4) published data. Depending on the nature of the data and the objectives, I analysed the data using multivariate regression models, regression models built in a Bayesian framework to incorporate spatially unstructured random effects, and meta-analytical models. Results: Seven percent of admitted patients to two Australian tertiary hospitals were asymptomatically colonised by C. difficile. Toxigenic C. difficile (TCD)-colonisation was associated with gastro-oesophageal reflux disease, higher number of hospital admissions, and antimicrobial exposure; whereas, non-toxigenic C. difficile (NTCD)-colonisation was associated with chronic obstructive pulmonary disease and chronic kidney failure. Asymptomatic C. difficile colonisation was seasonal with a higher prevalence in summer than winter. The predominant C. difficile ribotypes isolated in the hospital setting corresponded with those isolated in the community. Similarly, ribotypes isolated from symptomatic patients matched those isolated from asymptomatic patients in the hospitals. The proportion of positive CDI stool specimens increased 3-fold during the last decade in Queensland. The distribution of CDI had no evidence of spatial clustering at the postcode level and the observed increase of CA-CDI was not associated with variation in medication exposure at a population level. Faecal microbiota transplantation (FMT) was more effective for CDI recurrence/relapse when administered via colonoscopy/enema than gastroscopy/nasogastric tube. Lower levels of vitamin D were associated with CDI as well as severe forms of CDI. Conclusions: I provided the first prevalence estimates of asymptomatic C. difficile colonisation in Australian hospitals. I also provided evidence that patient characteristics differed between asymptomatic NTCD- and TCD-colonisation. I found that the predominant ribotypes circulating in the communities concorded with those circulating in the hospitals. The findings suggested that asymptomatic colonised patients can act as a means of transmission between the hospital and community settings. I identified that over the past decade CDI has significantly increased in Queensland and antibiotic restriction policy in the community might have little effect on CA-CDI. I provided evidence of low levels of vitamin D is a risk factor for CDI and FMT for CDI recurrence/relapse should be preferably delivered via colonoscopy/enema