Postmenopausal women with metastatic breast cancer were treated with trilostane, initially 240 mg daily increasing after 3 days to 480 mg daily and after a further three days to 960 mg daily. After 3 days at this dose dexamethasone 1 mg daily was added and this combination was continued until disease progression occurred. Partial remission was seen in 26% and stabilization of previously progressive disease in a further 13% of the first twenty-three patients studied. During therapy with trilostane alone significant increases in DHEAS, androstenedione, 17-hydroxypregnenolone, progesterone, testosterone and oestradiol were seen. A significant fall in oestrone concentration occurred at the same time. After dexamethasone was added the elevated steroid concentrations fell back to the baseline while oestrone remained depressed below this and testosterone was also significantly lowered. No change was seen in cortisol or ACTH concentration while patients were on trilostane alone but cortisol levels were undetectable after dexamethasone was added though, in most patients, ACTH remained detectable. There was no change in the ratio of delta 5:delta 4 steroids at any stage of therapy but a highly significant increase in the androstenedione: oestrone ratio was seen. We conclude that in long-term use in vivo it is difficult to demonstrate that trilostane inhibits 3 beta-hydroxysteroid dehydrogenase but it may produce inhibition of aromatase.