BACKGROUND & AIMS: Ursodeoxycholic acid (UDCA) improves laboratory liver test results in patients with primary biliary cirrhosis (PBC). Few studies have assessed the prognostic significance of biochemical data collected following UDCA treatment. We performed a prospective multicenter study of patients with PBC treated with UDCA to compare prognosis with biochemical response. METHODS: PBC was classified as early (pretreatment bilirubin and albumin levels normal), moderately advanced (one level abnormal), or advanced (both levels abnormal). Biochemical response was defined as proposed by Pares (decrease in alkaline phosphatase [ALP] level>40% of baseline level or normal level), Corpechot (ALP level<3-fold the upper limit of normal [ULN], aspartate aminotransferase level<2-fold the ULN, bilirubin level<1-fold the ULN), and our group (Rotterdam; normalization of abnormal bilirubin and/or albumin levels). RESULTS: The study included 375 patients, and median follow-up time was 9.7 (range, 1.0-17.3) years. The prognosis for early PBC was comparable with that of the Dutch population and better than predicted by the Mayo risk score. Survival of responders was better than that of nonresponders, according to Corpechot and Rotterdam criteria (P<.001). Prognosis of early PBC was comparable for responders and nonresponders; prognosis of responders was significantly better in those with (moderately) advanced disease. CONCLUSIONS: Prognosis for UDCA-treated patients with early PBC is comparable to that of the general population. Survival of those with advanced PBC with biochemical response to UDCA is significantly better than for nonresponders. Thus, UDCA may be of benefit irrespective of the stage of disease. Prognostic information, based on bilirubin and albumin levels, is superior to that provided by ALP levels.