Spatial clustering and livestock exposure as risk factor for community-acquired Clostridium difficile infection

Abstract Objectives Clostridium difficile infections (CDI) account for 1.5% of diarrhoeic episodes in patients attending a general practitioner in the Netherlands, but its sources are unknown. We searched for community clusters to recognize localized point sources of CDI. Methods Between October 2010 and February 2012, a community-based prospective nested case–control study was performed in three laboratories in the Netherlands with a study population of 2 810 830 people. Bernoulli spatial scan and space–time permutation models were used to detect spatial and/or temporal clusters of CDI. In addition, a multivariate conditional logistic regression model was constructed to test livestock exposure as a supposed risk factor in CDI patients without hospital admission within the previous 12 weeks (community-acquired (CA) CDI). Results In laboratories A, B and C, 1.3%, 1.8% and 2.1% of patients with diarrhoea tested positive for CDI, respectively. The mean age of CA-CDI patients ( n  = 124) was 49 years (standard deviation, 22.6); 64.5% were female. No spatial or temporal clusters of CDI cases were detected compared to C. difficile –negative diarrhoeic controls. Except for one false-positive signal, no spatiotemporal interaction amongst CDI cases was found. Livestock exposure was not related to CA-CDI (odds ratio, 0.99; 95% confidence interval, 0.44–2.24). Ten percent of CA-CDIs was caused by PCR ribotype 078, spatially dispersed throughout the study area. Conclusions The absence of clusters of CDI cases in a community cohort of diarrhoeic patients suggests a lack of localized point sources of CDI in the living environment of these patients.