The efficacy and side-effects of megestrol acetate and medroxyprogesterone acetate in postmenopausal patients with advanced breast cancer were compared in a prospectively randomized study. The dosage of MA was 2 X 80 mg p.o. or MPA 2 X 500 mg p.o. daily, given as a secondary hormonal treatment, mostly after previous treatment with tamoxifen. Ninety-eight patients entered the study and 92 were evaluable for effect, 48 patients on MA and 44 on MPA. Age, main tumor site and prior treatment were not different, but there was a preponderance of ER-negative tumors in the MA group. Responses appeared to be more frequent in the MPA-treated group (25% vs. 43%), predominantly in bone lesions, 12% for MA and 45% for MPA. Median progression-free survival was comparable, 15 vs. 10 months, and overall survival was not different (20 vs. 16 months). Toxicity was frequent, occurring in 83% vs. 74% of patients: increased appetite, nausea and dizziness in more than 20%, and a preponderance of pyrosis and breathlessness on MA and hot flashes, sweating and tremors on MPA. Cushingoid symptoms were present in about a quarter of the patients treated for more than 3 months. The occurrence of thrombo-embolic episodes and cardiovascular events was evenly distributed. Patients on MPA had more often increase in body weight, systolic blood pressure and serum creatinine than those treated with MA. It is concluded that MPA may be more effective for treatment of bone metastases, at the expense of more progestational side-effects. The occurrence of Cushingoid effects is frequent but similar in both arms, while the incidence of cardiovascular or thrombo-embolic events cannot be related to the use of either compound.