The natural history of pruritus in primary biliary cirrhosis (PBC) remains poorly defined. The aim of this investigation was to evaluate outcomes of pruritus in clinical trials for ursodeoxycholic acid (UDCA). In a UDCA-placebo trial begun in 1988 (n = 180), a 55% prevalence rate for pruritus was observed. Serum alkaline phosphatase level and Mayo risk score were independent risk factors for pruritus (P < 0.0001). Among placebo-treated patients (n = 91), the annual risks for development or improvement/resolution of pruritus were 27% and 23%, respectively. For UDCA-treated patients (n = 89), a trend toward improvement in pruritus was observed after 1 year compared to placebo (30% vs. 23%, P = 0.08) but not at 2 or 3 years. In a 3-dose UDCA trial begun in 1995 (n = 155), the overall prevalence of pruritus was significantly lower at 37% when compared to UDCA-placebo participants (P < 0.001). Baseline serum alkaline phosphatase level and Mayo risk score were again independent risk factors for pruritus (P < 0.0001). Among 13 (3.9%) patients with refractory pruritus, symptom resolution (n = 5) or improvement (n = 8) was associated with the use of oral rifampin (300 or 600 mg daily). Two patients treated with rifampin developed biochemical evidence for hepatotoxicity necessitating drug withdrawal. Although less prevalent among recently diagnosed individuals, more than one third of PBC patients develop pruritus. No significant risk reduction in developing pruritus with UDCA therapy was observed compared to placebo-treated patients. The long-term administration of rifampin for refractory pruritus is associated with occasional hepatotoxicity.