Saiboku-To, a mixture of ten different herbal extracts, has been used in Japan and Czechoslovakia for corticosteroid-dependent severe asthma to reduce the maintenance doses of corticosteroid. Magnolol has been considered to be an active component of Saiboku-To as an inhibitor of 11 beta-hydroxysteroid dehydrogenase and T-lymphocyte proliferation resulting in corticosteroid-sparing. To investigate the relationship between magnolol and the clinical effects of Saiboku-To, urinary magnolol excretion was compared in responders and non-responders under long-term Saiboku-To treatment. The clinical outcome of the Saiboku-To treatment was evaluated in nine asthmatic patients at 52 weeks after the onset of the treatment, using individual fluctuation of asthmatic points obtained from the patients' diary cards. Three patients whose clinical conditions were improved by the treatment were termed responders and six others were termed non-responders. The difference in the amounts of the total magnolol excreted were not significant; however, free (or non-conjugated) amounts of magnolol excreted in the responders were 7 times those in the non-responders (P < 0.05). These results suggest that the magnolol might be responsible for the therapeutic effect of Saiboku-To, indicating practical bioavailability in the responders.