The efficacy of tamoxifen (TAM) was compared to that of progestins (medroxyprogesterone acetate, MPA, and megestrol acetate, MA) in the treatment of metastatic breast cancer in postmenopausal women by a quantitative analysis of the results of published randomized clinical trials. Seven studies involving a total of 801 subjects compared TAM with MPA. Overall, the frequency of complete and partial response was 9 and 18%, respectively, in the women treated with TAM, versus 9 and 28% in those given MPA. Considering complete and partial responses together, the frequency of response was 29% in the TAM group and 39% in the MPA group, the corresponding pooled odds ratio (OR) of response being 1.5 (95% confidence interval, CI, 1.1-2.0). The median duration of response was greater in the TAM-treated patients; however, the difference was small (14 vs. 11 months). The probability of response to MPA treatment was about 3-fold higher compared with the response to TAM treatment in the subgroup with bone metastases (OR 3.4), and 2-fold higher in the subgroup with visceral metastases (OR 2.2), but the difference in the OR estimates was not statistically significant. The response to the two drugs was similar in the subgroup with metastases in soft tissues. Four studies compared TAM with MA, taking in 463 subjects. The overall frequency of complete and partial response was 35% in the patients who received TAM compared with 29% in those treated with MA. The corresponding pooled OR was 0.8 (95% CI 0.5-1.1). Analysis of the results according to site of metastases revealed no significant difference in the frequency of complete or partial response in the two treatment groups.