We followed 49 women who underwent a minimum of 2 years estrogen replacement therapy (ERT) after diagnosis and treatment for localized breast cancer. Forty-three women were treated with oral ERT. In this group, the median age at the time of cancer diagnosis was 46 years (range 26 to 66 years), and ERT was begun a median of 84 months after diagnosis (range 0 to 286 months). The patients were followed for a median of 144 months after cancer diagnosis (range 46 to 324 months), and the median duration of ERT was 31 months (range 24 to 142 months). For six women, ERT was administered as a vaginal cream application. In this group, the median age at time of cancer diagnosis was 46 years (range 38 to 57 years), and ERT was begun a median of 49 months after diagnosis (range 24 to 61 months). The patients were followed for a median of 95 months after cancer diagnosis (range 72 to 154 months), and the median duration of ERT was 47 months (range 27 to 80 months). One patient experienced disease recurrence; she had received surgery for a stage I, estrogen receptor (ER)-positive lesion. The patient began ERT 30 months after cancer diagnosis and developed a recurrent ER-negative tumor 56 months after initiation of ERT. She remained alive without evidence of disease for 10 years since initial diagnosis of breast cancer. Despite the inherent limitations of retrospective experiential data and the need for prospective, randomized trials to assess the safety of ERT, the present observations suggest that ERT does not appear to have a pronounced adverse effect on cancer outcome. Nevertheless, until appropriate clinical trials determine that ERT is safe, caution is needed.