The Risks of Oral Contraceptives and Estrogen Replacement Therapy

THE RISKS OF ORAL CONTRACEPTIVES AND ESTROGEN REPLACEMENT THERAPY FREDRIC L. COE andJOAN H. PARKS* So many millions ofwomen use oral contraceptive pills (OCP) or estrogen replacement (ERT) after menopause that all side effects matter, especially any that promote major illnesses, such as cancer, heart attack, or stroke. The risks of OCP and ERT have been studied rather well, but not everyone understands what is known, partly because published reports are scattered and statistically elaborate, and partly because some have contradicted each another and some are clearly wrong. We offer our summary of the main studies linking cancer, heart attack, stroke, and venous thromboembolism that we believe clearly define the serious risks of OCP and ERT. How Risk Is Measured If women who use OCP or ERT are matched, somehow, with women who do not, and the two cohorts are observed over years, one can count the occurrences of diseases among the users and the controls and estimate risk. Because each woman is observed only so long, and women are changeable—they stop OCP, or start it, or become pregnant—each contributes "woman-years" to categories such as "never user," "past user," "current user." The tally of disease occurrences per woman-year is a mosaic pieced together of years from many women, each one contributing time as her habits have dictated. The cohorts must be matched in age, by behaviors that affect risk such as smoking, the timing and number of pregnancies, and by matters of fate, like family history. We will tell you about eight cohort studies, one of them of the most specialized type, a random-allocation, double-blind drug trial. Work supported by NIH POl 33949. *Nephrology Program, University of Chicago, Pritzker School of Medicine, Chicago, Illinois 60637.© 1989 by The University of Chicago. AU rights reserved. 003 1 -5982/90/330 1-0650$0 1 .00 86 I Fredric L. Coe andJoan H. Parks · Risks ofEstrogen Therapy The common alternative to cohorts uses "case controls." Investigators cull out every case of a disease as it is diagnosed in some region of the country or in a cluster of hospitals. By some device, they select two matching women for each case, from the same locale or group of hospitals , or clinics, one of whom does not have the disease but is like the woman who does in age, risk factors, and all other ways possible. Among the "cases" and the "controls," how many use OCP, or ERT, or have used one or both? If cases use OCP or ERT more than controls, or have used OCP or ERT, OCP or ERT may be the guilty party; why else be more frequent in cases if the controls were chosen fairly? How you express risk depends on study design. Using cohorts, you measure new incidents of disease per woman per year, the incidence rate. Risk is the ratio of incidence rates of—let us say—OCP users or past users to rates of nonusers or never users. For case control, risk is inferred from enrichment; the greater the frequency of, for example, OCP use in cases versus controls, the greater the risk. The actual result is a risk ratio that is based on the frequency rates and complicated computation . Guilt is by association. What Are OCP and ERT? All OCPs used today contain an estrogen and a progestogen. Normally , the ovaries produce estradiol, which is the most potent natural hormone that can cause the "estrogenic" response: uterine growth, thickening of vaginal mucosa, thinning of cervical mucus, development of the ductal system of the breasts, and female characteristics ofhair and body fat distribution and skin texture. The other estrogen in blood, estrone, is made mainly by other tissues, from androstenedione, the immediate precursor of testosterone. Twenty-nine of 38 current preparations [1] use ethinyl estradiol, 20-50 µg daily; this compound is estradiol bearing an added ethinyl group at carbon-17. The rest use mestranol , identical except that a methoxy group replaces a hydroxyl group at carbon-3. The physiologic progestogen is progesterone, produced mainly by the ovarian corpus luteum. Progestational effects of the hormone prepare the uterine lining for implantation of the fertilized egg, inhibit uterine contractions, and stimulate...