Estrogen Replacement Therapy in Breast Cancer Survivors-Reply

Abstract

<h3>In Reply.</h3> —We thank the many authors of letters written in response to our work. Meta-analyses, which overcome type II error, have not demonstrated an association between ERT and risk of breast cancer. In general, prospective studies are preferable to case-control studies. However, there are problems with all studies, including prospective ones. For example, the Nurses' Health Study found an increase in breast cancer risk only in those who were drinking alcohol.¹The Swedish study wasn't a true prospective cohort trial.²Files from a prescription database were merged with files from a cancer registry. It is not known whether "hormone users" were actually taking hormones. We need to weigh the entire body of evidence. We should not discount the results of 25 case-control studies because we think two prospective ones are better. Most studies of adjuvant tamoxifen in postmenopausal women evaluated tamoxifen vs no treatment, whereas those in