Introduction: Bile acids (BA) influence germination and growth of Clostridium difficile. Ursodeoxycholic acid (UDCA), a BA minor in human, used for cholestatic liver diseases, inhibits germination and growth of C. difficile in vitro, but was never tested in vivo with an infectious challenge versus control. We hypothesized that UDCA could prevent CDI. We evaluated the effects of UDCA on C. difficile in vitro and in hamsters, with pharmacokinetics study and with an infectious challenge. Then, we studied CDI incidence in UDCA–treated patients.Methods: We evaluated germination and growth of C. difficile, with 0.01, 0.05, and 0.1% UDCA. We analyzed fecal BA of hamsters receiving antibiotics and UDCA (50 mg/kg/day), antibiotics, or UDCA alone. Then, we challenged with spores of C. difficile at D6 hamsters treated with UDCA (50 mg/kg/day) from D1 to D13, versus control. In human, we analyzed the database of a cohort on CDI in acute flares of inflammatory bowel disease (IBD). As PSC-IBD patients were under UDCA treatment, we compared PSC-IBD patients to IBD patients without PSC.Results: In vitro, UDCA inhibited germination and growth of C. difficile at 0.05 and 0.1%, competing with 0.1% TCA (with 0.1%: 0.05% ± 0.05% colony forming unit versus 100% ± 0%, P < 0.0001). In hamsters, UDCA reached high levels only when administered with antibiotics (43.5% UDCA at D5). Without antibiotics, UDCA was in small amount in feces (max. 4.28%), probably because of UDCA transformation into LCA by gut microbiota. During infectious challenge, mortality was similar in animals treated or not with UDCA (62.5%, n = 5/8, P = 0.78). UDCA percentage was high, similar and with the same kinetics in dead and surviving hamsters. However, dead hamsters had a higher ratio of primary over secondary BA compared to surviving hamsters. 9% (n = 41/404) of IBD patients without PSC had a CDI, versus 25% (n = 4/12) of PSC-IBD patients treated with UDCA.Conclusion: We confirmed the inhibitory effect of UDCA on growth and germination of C. difficile in vitro, with 0.05 or 0.1% UDCA. However, in our hamster model, UDCA was inefficient to prevent CDI, despite high levels of UDCA in feces. Patients with PSC-IBD treated with UDCA did not have less CDI than IBD patients.