The influence of intravenous infusion of branched-chain amino acids (BCAAs) on brain function in patients with liver cirrhosis and acute hepatic encephalopathy was examined using a double-blind, randomized study design. Five medical centers in France and Sweden participated, and 50 patients were studied. The patients received either BCAAs (40 gm per day) in 5% glucose or 5% glucose alone (placebo) for 5 days or until “wake up”. Nutritional support was provided with equal proportions of carbohydrate and fat. During BCAA adminstration, plasma concentrations of aromatic amino acids and methionine fell (20 to 40%, p < 0.05 to 0.01), and the ratio of BCAAs to aromatic amino acid concentrations increased significantly. Clinical improvement was seen in 14 of 25 BCAA-treated patients and in 12 of 25 patients receiving placebo (N.S.). EEG responses were similar in the two groups during treatment. In the BCAA group, 10 of 25 patients died in the course of the study, compared to 5 of 25 in the placebo group (N.S.); six patients died from encephalopathy in the BCAA group as compared to three among placebo-treated patients. It is concluded that BCAA adminstration, in the dose and composition employed in the present study, reduces the concentrations of aromatic amino acids but neither improves cerebral function nor decreases mortality in patients with hepatic encephalopathy.