This epidemiologic case-control study examined the relationship between replacement estrogen use and breast cancer risk in 2 population groups in Hawaii. No significant associations were observed when 161 Caucasian cases were compared with either their neighborhood controls (RR = 0.9; 95% Cl = 0.5-1.3) or their hospital controls (RR = 0.7; 95% Cl = 0.4 to 1.1) and when 183 Japanese cases were compared with either their neighborhood controls (RR = 1.1; 95% Cl = 0.7-1.6) or their hospital controls (RR = 1.0; 95% Cl = 0.6-1.4). The results indicate that the use of replacement estrogens cannot account for the large difference in breast cancer incidence between the 2 Hawaiian ethnic groups. However, further data analysis involving neighborhood controls was suggestive of a possible increase in breast cancer risk with estrogen use for certain sub-groups of women who are at high risk for the disease. These included estrogen users with a family history of breast cancer or a history of benign breast disease. These findings are in agreement with other studies which have used non-hospitalized controls. Because the numbers of cases in this study are not substantial, it is recommended that a large population-based case-control study be undertaken to clarify the relationship between breast cancer risk and replacement estrogen use, especially in sub-groups of women at high risk for the disease.