Patients with moderately severe or severe alcoholic hepatitis, described in a companion paper in this issue, had serial studies of energy and protein metabolism and elemental balances before and during treatment for 21 days with one of four randomly assigned regimens: 1) standard therapy, consisting of abstinence, a balanced, nutritionally adequate diet, and multivitamins; 2) oxandrolone (20 mg orally four times a day) plus standard therapy; 3) nutritional supplementation, consisting of 2 liters of 3.5% crystalline amino acids in 5% dextrose given by peripheral vein (PPN) plus standard therapy; and 4) a combination of the other three treatments. Dietary and intravenous intakes and weights were recorded daily, and weekly averages were calculated. Anthropometric measurements and blood studies were done weekly; blood studies included white blood cell counts and differentials, serum prealbumin, transferrin, and total protein and plasma aminograms. Four-days complete balance studies and measures of 15N,1-13C-leucine metabolism also were performed at baseline and after the treatment period. Major findings were as follows: a) Intakes of total calories and protein were significantly higher in PPN-treated than in other groups. b) All patients had positive elemental balances, both at baseline and at the end of the treatment period. However, those treated with PPN (with or without oxandrolone) had higher positive balances of nitrogen, potassium, phosphorus, and magnesium, indicating improvement in lean body mass. c) Anthropometric measurements showed no significant changes, but measures of the visceral protein compartment (serum prealbumin, transferrin, total protein, total lymphocyte count) improved significantly with time. For most of these variables, increases were significantly greater in those treated with PPN with or without oxandrolone than in the other groups. However, for prealbumin, the increase was greatest in the oxandrolone-treated group d) PPN treatment produced dramatic increases in levels of branched-chain amino acids and improvement in the ratio of plasma branched chain to aromatic amino acids. Other treatments had no effect on plasma aminograms. e) Metabolism of 15N,1-13C-leucine was normal and was not affected significantly by treatment. Therapy with PPN and/or oxandrolone was tolerated well. We conclude that PPN has favorable effects on energy and protein metabolism in florid alcoholic hepatitis; oxandrolone has lesser effects, although it may exert some additional action and particularly increases serum prealbumin levels. The results support the use of nutritional supplementation in therapy of moderately severe or severe alcoholic hepatitis.