EXOGENOUS ESTROGENS AND ENDOMETRIAL CANCER: A CASE-CONTROL STUDY AND ASSESSMENT OF POTENTIAL BIASES

Abstract

Eighty-eight cases with newly diagnosed carcinoma of the endometrium and 177 age-matched neighborhood controls were interviewed to test the hypothesis that exogenous estrogens lead to an increased risk of endometrial cancer. Forty-five per cent of the cases and 22% of the controls reported a history of estrogen use which yielded an odds ratio of 2.9 (confidence interval (Cl) 1.7–5.1). Women with five or more years of estrogen use had an odds ratio of 8.6 (Cl 3.2–23.0). Approximately 80% of the estrogen users hqd used conjugated equine estrogens. For these women the odds ratio was 4.0 (Cl 1.9–8.4) for daily dosages of more than 1 mg of estrogen. Several sources of bias which might affect the estrogen association were investigated. These included comparability of cases and controls, selection procedures, difference between estrogen users and nonusers, exclusion of controls who had hysterectomy, source of estrogen information, and differential recall. The concept of medical surveillance was evaluated by access to medical care and prior history of dilatation and curettage. The strong association between exogenous estrogen use and endometrial cancer remained after consideration for the effects of these biases.