We investigated the feasibility of increasing ursodeoxycholic acid (UDCA) in the enterohepatic circulation of pigs by administering living bacteria capable of epimerising endogenous amidated chenodeoxycholic acid (CDCA) to UDCA. We first demonstrated that combining Bifidobacterium animalis DN-173 010, as a bile salt-hydrolysing bacterium, and Clostridium absonum ATCC 27555, as a CDCA to UDCA epimerising bacterium, led to the efficient epimerisation of glyco- and tauro-CDCA in vitro, with respective UDCA yields of 55·8 (se 2·8) and 36·6 (se 1·5)%. This strain combination was then administered to hypercholesterolaemic pigs over a 3-week period, as two daily preprandial doses of either viable (six experimental pigs) or heat-inactivated bacteria (six controls). The main effects of treatment were on unconjugated bile acids (P=0·035) and UDCA (P<0·0001) absorbed into the portal vein, which increased 1·6–1·7- and 3·5–7·5-fold, respectively, under administration of living compared with inactivated bacteria. In bile, UDCA did not increase significantly, but the increase in biliary lithocholic acid with time in the controls was not observed in the experimental pigs (P=0·007), and the same trend was observed in faeces. All other variables (biliary lipid equilibrium, plasma lipid levels and partition of cholesterol between the different lipoprotein classes) remained unaffected by treatment throughout the duration of the experiment. In conclusion, it is feasible to increase the bioavailability of UDCA to the intestine and the liver by administering active bacteria. This may represent an interesting new probiotic activity, provided that in future it could be expressed by a safe food micro-organism.