Review of epidemiologic studies of endometrial cancer and exogenous estrogen

Abstract

Epidemiologic literature on the association of exogenous estrogen and endometrial cancer is reviewed. Descriptive studies have documented fluctuations in the incidence of endometrial cancer, mainly of localized disease, associated with estrogen use. Etiologic studies have established an association between estrogen use during menopause and the occurrence of endometrial cancer. Although the association appears to be a valid one, several biases may have falsely increased the magnitude of this association. The association also appears to be strongest for local disease and weakest for the most invasive disease, which implies that the etiology for the more invasive endometrial cancers is largely unaccounted for by estrogen use. A need for a prospective study to define other potential risks and benefits of estrogen therapy is clear. However, appreciation of factors known to modify the risk of endometrial cancer from exogenous estrogen can help the clinician to use these preparations judiciously.

PIP: Controversy over the relationship between the use of exogenous estrogens and endometrial cancer persists and will be dissipated only by well designed prospective studies. Most epidemiological investigations, both descriptive and etiological, have reported a relationship between estrogen use and endometrial cancer, but the association may have been artifically inflated by a number of factors. The evidence for an association between less invasive forms of endometrial cancer and estrogen use is more convincing. Descriptive studies have reported an increase in the incidence of endometrial cancer in a number of regions in the U.S. from 1960 to the mid-1970s, and this observed increase has been attributed to the corresponding increase in estrogen usage during the same period. The high endometrial cancer rates may have been inflated by the inclusion of nonmalignant hyperplasia cases stemming from the recent adoption of more sensitive diagnostic criteria. The association between endometrial cancer and estrogen use revealed in a number of case control studies may have been exaggerated because: 1) cancer is more likely to be detected among women taking estrogens since they are under closer medical scrutiny; 2) women with a typical hyperplasia may have been assigned disease status; and 3) retrospectively ascertained medication records on women with cancer may be more complete than those obtained for controls. Prospective studies are needed to clarify the degree of association despite the high cost involved in this type of undertaking. In the meantime, it is suggested that clinicians in treating menopausal women 1) refrain from estrogen therapy if pretreatment endometrial biopsy findings indicate the presence of hyperplasia; 2) use low estrogen doses; 3) consider the addition of progesterone in the treatment regimen; and 4) perform yearly biopsies on women receiving treatment. Line graphs depict 1) duration of estrogen use by risk ratio of endometrial cancer and 2) comparison of cumulative risk of endometrial cancer by years of follow-up for a hypothetical 50 year old patient treated with estrogens and one not treated.