We entered 60 patients with primary biliary cirrhosis in a double-blind randomized controlled trial to determine whether colchicine is therapeutically effective. Thirty patients had early disease (Stages 1 and 2), and 30 had advanced disease (Stages 3 and 4). Fifteen patients with early disease and 15 with advanced disease received colchicine (0.6 mg twice daily), and the remainder received placebo. Patients were studied about every two months; those remaining in the blind phase at two years underwent repeat liver biopsy and were then placed on open-label colchicine (0.6 mg twice daily). With a few exceptions, the results in patients with early disease were similar to those in patients with advanced disease; hence, data on patients in all stages were combined in the main analysis. During the two-year study period the colchicine-treated patients, as compared with the placebo-treated patients, had improvement in levels of serum albumin, serum bilirubin, alkaline phosphatase, cholesterol, and aminotransferases. However, there was no such improvement in the severity of symptoms or physical findings; moreover, there was no significant difference in the histologic changes noted at liver biopsy in the two treatment groups. At four years after entry, the cumulative mortality from liver disease was 21 percent in patients given colchicine and 47 percent in those given placebo (P = 0.05). The only side effect of colchicine was diarrhea, noted in three patients. The consistent and significant improvement in a number of markers of liver disease and the apparent decreased mortality from liver disease suggest that colchicine may provide some long-term clinical benefit in patients with primary biliary cirrhosis. However, the failure of colchicine to reduce hepatic inflammation and fibrosis leaves uncertain the effect of the drug on the longterm outcome of this disease.