Ursodeoxycholic acid (UDCA) is the only approved treatment for primary biliary cirrhosis (PBC). However, the benefit from UDCA therapy on the progression of PBC from its early stage towards extensive fibrosis and cirrhosis has not been clearly shown. The aim of this study was to assess the effect of UDCA therapy on liver fibrosis progression in PBC. A Markov model was used to analyze the progression rates between early and late histologic stages in 103 patients with PBC enrolled in a randomized, double-blind, placebo-controlled trial of UDCA. Early stage was defined by the presence of portal and periportal lesions without extensive fibrosis, whereas late stage was defined by the presence of numerous septa, bridging fibrosis, or cirrhosis. A total of 162 pairs of liver biopsy specimens were studied. The model accurately described the observed data. UDCA therapy was associated with a 5-fold lower progression rate from early stage disease to extensive fibrosis or cirrhosis (7% per year under UDCA vs. 34% per year under placebo, P <.002), but was not associated with a significant difference in regression rates (3% per year under both UDCA and placebo). At 4 years, the probability of UDCA-treated patients to remain in early stage disease is 76% (95% confidence interval: 58%-88%), as compared with 29% (15%-52%) in placebo-treated patients. In conclusion, UDCA therapy significantly delays the progression of liver fibrosis in PBC. Markov modeling should prove useful in assessing the efficacy of future medical treatments in clinical trials involving histologic endpoints.