A major obstacle to the understanding of the mechanisms of action of aromatase inhibitors in breast cancer is the observation that plasma estrogens are sustained at about 30-50% of their control levels despite 85-95% inhibition of the conversion of tracer androstenedione (A) to estrone (E1). The discrepancy could be due to lack of sensitivity of current RIAs. Due to low levels of plasma estradiol (E2) (mean about 20 pM) and E1 (mean about 75 pM) in postmenopausal women, it is difficult to develop RIA methods with the sensitivity required to detect > 90% suppression from baseline. In contrast, the plasma level of the estrogen conjugate estrone sulphate (E1S) is substantially higher (mean level about 400 pM). This paper describes a new assay to measure plasma E1S in the low range aiming to detect > 95% suppression of E1S from baseline values in patients treated with aromatase inhibitors. E1S was separated from unconjugated estrogens, hydrolysed and purified as unconjugated E1. E1 was subsequently reduced to E2, purified, and measured by a highly sensitive RIA using oestradiol-6-(O-carboxymethyl) oximino-(2(-)[125I]iodohistamine as ligand. The sensitivity limit of the method was 2.7 pM. Patients on treatment with the aromatase inhibitors formestane or aminoglutethimide or both drugs in concert were found to have plasma levels of E1S ranging from 3 to 274 pM with a mean suppression of 78, 86 and 95%, respectively, compared to baseline, a lower suppression than that reported in previous trials with these drugs.