Testing for antimitochondrial antibodies is the most useful laboratory procedure in the diagnosis of primary biliary cirrhosis; nevertheless, 5-10% of patients with typical features of primary biliary cirrhosis do not have detectable antimitochondrial antibodies, their condition being referred to as antimitochondrial antibody-negative primary biliary cirrhosis or "autoimmune cholangitis". Uncertainty exists whether antimitochondrial antibody-positive and -negative primary biliary cirrhosis represent distinct entities. We reviewed studies that compared: (i) the clinical, laboratory and histological characteristics of antimitochondrial antibody-positive and -negative primary biliary cirrhosis; (ii) the response to treatment of both conditions; and (iii) the response of autoimmune cholangitis to ursodeoxycholic acid and immunosuppressive therapy. Antimitochondrial antibody-positive and -negative primary biliary cirrhosis were characterized by similar clinical, laboratory and histological abnormalities, clinical course and survival. Antimitochondrial antibody status did not seem to affect the response to ursodeoxycholic acid. At present, the efficacy of therapies for autoimmune cholangitis has not been established in controlled trials. Of 52 patients with autoimmune cholangitis treated with ursodeoxycholic acid in 13 uncontrolled studies, 83% had serum biochemical improvement. Also, a favourable effect of immunosuppressive drugs occurred in 57% of 54 patients with autoimmune cholangitis in 17 uncontrolled studies. Each of these trials included very few patients and most evaluated the effects of treatment on surrogate markers of disease only. No marker that consistently distinguished patients who would respond favourably to ursodeoxycholic acid or immunosuppression was apparent. Consequently, treatment is, at present, empirical. However, ursodeoxycholic acid may be given when histology reveals bile duct lesions, whereas immunosuppressive therapy should probably be reserved for patients exhibiting interface hepatitis.