LONG-TERM CONSEQUENCES OF ESTROGEN AND ESTROGEN-PROGESTIN REPLACEMENT

Except for the control of smoking, hormone replacement therapy (HRT) is potentially one of the most useful means of preventing major causes of morbidity and mortality in older women, who are likely to spend more than one-third of their lives in a postmenopausal (estrogen-deficient) stage. A full assessment of the overall long-term effects of noncontraceptive estrogens-with and without cyclically added progestins-therefore is needed urgently. This paper briefly reviews the available epidemiologic findings to date, with emphasis on a prospective study in Uppsala, Sweden. An increase in the risk of endometrial cancer which is substantially dependent on dose and duration of the exposure has been established, whereas a biologically plausible protective effect of progestins has been documented in only a few studies. A moderate, perhaps twofold, increase in the incidence of breast cancer has been found in several recent studies, notably among long-term and/or current users of potent estrogens; there is no evidence that progestins counteract this adverse effect. The risk of osteoporotic hip fractures can be reduced substantially, at least during ongoing treatment starting early after menopause. A 25 to 50 percent reduction in cardiovascular morbidity and mortality has been shown consistently and might outweigh any conceivable adverse effect of HRT. Notwithstanding these largely encouraging epidemiologic data-including evidence of lower all-cause mortality in users than in nonusers of HRT-it is argued here that randomized control trials are needed to provide the necessary basis for widespread preventive use of HRT. Critical assessment of observational studies is hampered by our deficient knowledge of the differences in risk at baseline between women with and without climacteric symptoms, and by the fact that healthy women, e.g. without cardiovascular disease, are more likely to be prescribed HRT. Systematic errors therefore might both conceal and exaggerate risks and benefits. Several contexts in which randomized control trials might be informative, ethical, and feasible are suggested.