CGS 16949A (fadrozole hydrochloride), a potent cytochrome P450-mediated steroidogenesis inhibitor, blocks aromatase at low doses, but other biosynthetic steps at higher concentrations. Recent studies demonstrated inhibition of C11-hydroxylase, corticosterone methyloxidase-II, and deoxycorticosterone to corticosterone conversion with this agent at some-what higher concentrations than those required for blockade of aromatase. Based upon phase I studies, we postulated that relatively selective inhibition of aromatase might be possible if sufficiently low doses of CGS 16949A were used. A phase II study in 54 postmenopausal women with metastatic breast cancer examined the effects of low dose CGS 16949A on estrogen, mineralocorticoid, and glucocorticoid secretion. Two dose schedules and two dose levels were chosen based upon our prior dose escalation protocol study. Plasma estrone, estradiol, and estrone sulfate as well as urinary estrone and estradiol fell equally with 1.8-4 mg CGS 16949A given either on a twice daily or three times daily dose schedule. Isotopic kinetic studies demonstrated an 84% decrease in the rate of conversion of androstenedione to estrone to 0.40 +/- 0.07% (patients receiving 1.8-4 mg CGS 16949A daily). With these three regimens, basal levels of aldosterone and cortisol did not change significantly over a 12-week period of observation. Clinical examination, plasma electrolytes, and urinary sodium/potassium ratios suggested no biological evidence of mineralo-corticoid deficiency. ACTH-stimulated cortisol concentrations, however, were blunted at each dose level compared to pretreatment values. Nonetheless, peak responses exceeded 550 nmol/L, or a basal to peak difference of 190 nmol/L or greater, in 97% of instances. This probably reflected inhibition of C11-hydroxylase, since basal and ACTH-stimulated levels of 11-deoxycortisol were increased in response to CGS 16949A. Androstenedione and 17 alpha-hydroxyprogesterone also exhibited an upward trend in response to drug treatment. ACTH-stimulated aldosterone levels were blunted to a greater extent than those of cortisol, probably as a reflection of corticosterone methyloxidase type II blockade. Overall, the results suggest that CGS 16949A, at doses of 1.8-2 mg daily, blocks aromatase effectively and does not produce clinically important inhibition of cortisol or aldosterone biosynthesis. Thus, this agent can probably be used safely without glucocorticoid or mineralocorticoid supplementation.