Fourteen patients with liver cirrhosis received oral prednisone or prednisolone (0.3 mg per kg) randomised on two consecutive days. Serum prednisone and prednisolone were measured over the following four hours. Mean serum prednisolone concentration after oral prednisone decreased with impaired liver function estimated by galactose elimination capacity (r = 0.64, P less than 0.03). Mean serum prednisolone concentration after oral prednisone in the seven patients with severely impaired liver function was only 53% (P less than 0.05) of that observed in the seven patients with slightly impaired liver function. Conversely, mean serum prednisone concentration after oral prednisone in the patients with severely impaired liver function was 74% higher (P = 0.05) than in patients with slightly impaired liver function. Mean serum prednisolone after oral prednisolone was independent of liver function. As only prednisolone exerts glucocorticoid activity, our results indicate that prednisolone should be preferred to prednisone in the treatment of patients with impaired liver function.