BACKGROUND & AIMS: Noninvasive surrogate end points of long-term outcomes of patients with primary biliary cirrhosis (PBC) are needed to monitor disease progression and evaluate potential treatments. We performed a meta-analysis of individual patient data from cohort studies to evaluate whether patients' levels of alkaline phosphatase and bilirubin correlate with their outcomes and can be used as surrogate end points. METHODS: We performed a meta-analysis of data from 4845 patients included in 15 North American and European long-term follow-up cohort studies. Levels of alkaline phosphatase and bilirubin were analyzed in different settings and sub-populations at different time points relative to the clinical end point (liver transplantation or death). RESULTS: Of the 4845 patients, 1118 reached a clinical end point. The median follow-up period was 7.3 years; 77% survived for 10 years after study enrollment. Levels of alkaline phosphatase and bilirubin measured at study enrollment (baseline) and each year for 5 years were strongly associated with clinical outcomes (lower levels were associated with longer transplant-free survival). At 1 year after study enrollment, levels of alkaline phosphatase that were 2.0 times the upper limit of normal (ULN) best predicted patient outcome (C statistic, 0.71) but not significantly better than other thresholds. Of patients with alkaline phosphatase levels <= 2.0 times the ULN, 84% survived for 10 years compared with 62% of those with levels >2.0 times the ULN (P < .0001). Absolute levels of alkaline phosphatase 1 year after study enrollment predicted patient outcomes better than percentage change in level. One year after study enrollment, a bilirubin level 1.0 times the ULN best predicted patient transplant-free survival (C statistic, 0.79). Of patients with bilirubin levels <= 1.0 times the ULN, 86% survived for 10 years after study enrollment compared with 41% of those with levels >1.0 times the ULN (P < .0001). Combining levels of alkaline phosphatase and bilirubin increased the ability to predict patient survival times. We confirmed the predictive value of alkaline phosphatase and bilirubin levels in multiple subgroups, such as patients who had not received treatment with ursodeoxycholic acid, and at different time points after study enrollment. CONCLUSIONS: Levels of alkaline phosphatase and bilirubin can predict outcomes (liver transplantation or death) of patients with PBC and might be used as surrogate end points in therapy trials.