OBJECTIVE Portal hypertension can complicate primary biliary cirrhosis, but studies evaluating the direct measurement of the portohepatic gradient (PHG) are rare. The aim of the study was to determine the prevalence and prognostic value of portal hypertension in patients treated with ursodeoxycholic acid. METHODS A total of 132 patients from a local "PBC clinic" were enrolled in this cohort study. The PHG and biochemical values were measured at inclusion and every 2 years. Factors associated with survival were analyzed. RESULTS Mean PHG at inclusion was 7.2 +/- 5.8 mm Hg. It was higher than normal (6 mm Hg) in 46 patients (34.9%) and higher than 12 mm Hg (variceal bleeding risk limit) in 26 patients (19.7%). There was a difference between the 3 subgroups in the probability of survival free of liver transplantation (P < .0003). After 2 years of treatment, a decreased or stable PHG (hazard ratio, 4.64; 95% confidence interval, 2.01-10.72) and normalization of aspartate aminotransferase (AST) level (hazard ratio, 2.89; 95% confidence interval, 1.03-8.05) were predictive of better survival on multivariate analysis. "Responders" (stable or improved PHG and normalized AST level at 2 years) have a 15-year survival similar to that of a control Quebec female population. CONCLUSIONS Significant portal hypertension is a common complication of primary biliary cirrhosis. Changes in the PHG and normalized AST level after 2 years of ursodeoxycholic acid treatment can be used to identify a subgroup of responders with survival comparable to that of a control population.