Summary Twelve premenopausal patients with advanced breast cancer were randomised to receive 3.75 or 7.5 mg of a slow release formulation of the luteinising hormone releasing hormone agonist leuprorelin once every 4 weeks. All patients were oestrogen receptor positive or unknown. Serum (Jacobi et al., 1987;Smith et al., 1985; Donnelly & Milstead, 1987) and breast cancer (Harvey et al., 1985;Klijn, 1984;Williams et al., 1986). This is due to the 'down-regulation' of LHRH receptors by the drugs, which results in a fall in gonadotrophin levels and a consequent reduction in the level of gonadal steroid production (Sandow, 1983). The response rate is higher in premenopausal patients in breast cancer as might be expected for a drug with this action. The small number of responses in postmenopausal patients may be due to reduced ovarian androgen secretion with a consequent fall in circulating oestrogen levels (Dowsett et al., 1988;Crighton et al., 1989).Leuprorelin is a synthetic nonapeptide LHRHa (D-Leu6 DES Gly NH2iO GnRH ethylamide) which lacks the aminoacid glycine at position 10 and has leucine substituted for glycine at position 6 of natural LHRH. Earlier clinical studies required daily administration in aqueous solution (Harvey et al., 1985), but a more convenient sustained release formulation has now been developed, in which the agonist is microencapsulated in polylactic and polyglycolic acid. It has been found that the compound is equally effective endocrinologically in prostatic cancer as once monthly doses of either 7.5 or 3.75 mg (Isurugi et al., 1988). This is the first report of the use of the slow-release formulation in premenopausal women. A detailed endocrine study was conducted to compare the doses of 3.75 and 7.5 mg in 12 patients, to determine whether there was any contraindication to further study of the lower dose in a larger group of patients. Blood samples were drawn from patients prior to treatment and 1, 2, 3, 4, 6, 8, 10 and 12 weeks after starting treatment. The samples were allowed to clot and the resultant serum was stored at -20°C until analysis. The following analyses were performed by previously described immunoassays: luteinising hormone (LH) and follicle stimulating hormone (FSH) (Ferguson et al., 1982); oestradiol (Dowsett et al., 1987), oestrone (Harris et al., 1983). Serum levels of androstenedione were measured using the Biogenesis kit. This is a direct assay employing an iodinated tracer in which the only cross-reactions of greater than 0.1 % were to 11-deoxycortisol (1.2%) and isoandrosterone (0.3%). Withinand between-assay coefficients of variation were 7.5% and 8.6%, respectively. Serum levels of testosterone were measured using the St Thomas Hospital Testosterone kit (Wheeler et al., 1983). In this assay, the serum is first extracted with ether and this extract is subject to immunoassay using an iodinated tracer. The antiserum cross-reacts 20% with 5adihydrotestosterone. All other cross-reactions were less than 0.1%. Within-and between-assay coefficients of variation were ...