To evaluate the efficacy of ursodeoxycholic acid as adjunctive therapy in type 1 autoimmune hepatitis, 37 patients who had experienced treatment failure, repeated relapse, or incomplete response were randomized to ursodeoxycholic acid (13-15 mg/kg daily) or placebo for 6 months in addition to their usual corticosteroid schedule. Serum aspartate transaminase (70% vs. 31%, P =.04) and alkaline phosphatase (47% vs. 7%, P =.02) levels improved more commonly in the 21 patients randomized to ursodeoxycholic acid. Mean serum levels, however, were similar before and after the treatment period. The frequency of dose reduction or corticosteroid withdrawal was comparable in both groups (29% versus 31%, P >.9), and clinical improvement (48% vs. 44%, P >.9) or its absence (52% vs. 56%, P >.9) occurred as commonly in patients receiving ursodeoxycholic acid or placebo. The modified histological activity score (3.5 +/- 0.8 vs. 3. 5 +/- 0.9) and the modified fibrosis score (2.4 +/- 0.4 vs. 2.4 +/- 0.4) were similar before and after treatment with ursodeoxycholic acid and no different than after placebo therapy. We conclude that ursodeoxycholic acid can improve certain laboratory tests in problematic patients with type 1 autoimmune hepatitis when administered adjunctively for 6 months. Short-term therapy, however, does not facilitate reduction in the dose of corticosteroids or its withdrawal, affect clinical outcome, or reduce histological activity.