Clinical and pathologic findings were compared in 43 postmenopausal endometrial carcinoma patients who had received exogenous estrogens prior to diagnosis and 79 similar patients unexposed to estrogens. Estrogen non-users were more likely to manifest lower parity, later menopause, obesity, hypertension, and diabetes, all of which have been considered to be constitutional risk factors for the development of endometrial carcinoma. Although estrogen users and non-users had similar extent of disease as judged by clinical stage, there was a tendency to more myometrial invasion in hysterectomy specimens from non-users, as well as greater frequency of unfavorable histologic types and grades of tumor. At short-term follow-up, more recurrences occurred in non-users, and this tendency appeared to be independent of clinical stage, histologic type, histologic grade, or modality of treatment. The significance of these and other observation to the determination of the risk-benefit ratio for estrogen administration is discussed.