HRT gained wide acceptance over 30 years ago and gradually became a staple in the treatment of estrogen deficiency.  Patients were pleased by the relief of menopausal symptoms offered by HRT, and physicians were generally eager to embrace this fairly convenient therapy that delayed or prevented the onset of osteoporosis through positive effects on bone turnover. In the decades since HRT was introduced, its efficacy in preventing or ameliorating vasomotor instability and genitourinary atrophy has been well established, and its efficacy in preventing osteoporosis has been confirmed by a sizable body of studies.

Early population-based retrospective analyses suggested that HRT reduced overall cardiovascular mortality through beneficial effects on the heart and vasculature. Initial enthusiasm for the use of HRT to address the increased risk of coronary heart disease (CHD) associated with estrogen deficiency was reflected in a position statement by the American Collegeof Physicians (Guidelines for counseling…,  1992)  proposing that all postmenopausal women be offered HRT. 

However,  conflicting reports periodically emerged that raised doubts about the cardiovascular benefits of HRT and suggested that HRT might increase the risk of not only thromboembolism but also breast cancer. Over the past several years, a series of prospective trials designed to answer these questions has led to a paradigm shift in the management of estrogen deficiency and its sequelae.

Observational studies have differed in their assessments regarding the impact of HRT on CHD risk. For example, the Framingham Heart Study (Wilson et al., 1985) found no decrease in CHD risk in estrogen users, but the first part of the Nurse’s Health Study (Stampfer et al., 1985) found a significant decrease in CHD events with estrogen use. In the Heart and Estrogen/Progestin Replacement Study (Hulley et al., 1998; Grady et al., 2002),  a large randomized trial specifically designed to study the effect of HRT on secondary prevention of CHD, postmenopausal patients with established heart disease were followed for 6.8 years. HRT use was associated with an increased risk of CHD events in the first year and a decreased risk of CHD events during years 3-5, but the cumulative risk of CHD was no different from that of the placebo group.

Another large randomized trial,  the landmark Women’s Health Initiative (WHI)  trial (Rossouw et al., 2002; Anderson et al., 2004), evaluated the use of HRT for primary prevention of CHD.  The study was made up of two parallel trials that compared estrogen or estrogen plus progesterone with placebo in healthy postmenopausal women. In patients treated with estrogen plus progesterone, a significant increase of 29% was observed in CHD events, most of which were nonfatal myocardial infarctions; no significant difference was observed in CHD deaths or revascularization procedures. The estrogenonly arm showed a 9% decrease in CHD events, but this finding was not statistically significant.

The question of thromboembolic risk associated with HRT was addressed by the Women’s Estrogen for Stroke Trial (Viscoli et al., 2001), a prospective randomized study of secondary prevention of ischemic stroke in postmenopausal women. In this trial, estrogen therapy did not alter the risk of nonfatal stroke. However, compared to women in the placebo group, women in the estrogen group were more likely to have a fatal stroke and had worse neurological deficits after nonfatal stroke.  The effect of HRT on primary prevention of ischemic stroke was explored in the WHI trial, which found a 41% increase in the risk of ischemic stroke with estrogen and progesterone and a 39% increase in risk with estrogen alone; increases in risk were mostly due to nonfatal events. The risk of Deep venous thrombosis was significantly increased in both the estrogen-plus-progesterone and estrogen-only arms of the WHI trial.

Apprehension that HRT may increase the risk of developing breast cancer is the most frequent reason women cite for avoiding HRT.  The current consensus is that very prolonged estrogen use may contribute to breast cancer risk and that this risk is augmented by the use of progesterone. The estrogen-plus-progesterone arm of the WHI trial was terminated early because of a significant 26% increase in the risk of invasive breast cancer but showed no change in the risk of in situ breast cancers.

The estrogen-only arm of the WHI trial showed a 23% reduction in invasive breast cancer risk that narrowly missed statistical significance (P =.06). The Million Women Study (Beral, 2003), an observational cohort study of over one million women aged 50-64 years, tracked breast cancer incidence and the breast cancer mortality rate in HRT users in the United Kingdom. Results of the study showed that the relative risk of breast cancer was greater in current users than never users of HRT and increased with total duration of use. The relative risks of breast cancer for less than 5 years of use and for 5 or more years of use were 1.70 (95% confidence interval [CI], 1.56 to 1.85) and 2.21 (95% CI, 2.06 to 2.37), respectively, in women receiving estrogen and progesterone and 1.21 (95%  CI,  1.07 to 1.37)  and 1.34 (95%  CI,  1.23 to 1.40),  respectively,  in women receiving estrogen only. Past users of HRT were at no increased risk of breast cancer, regardless of the duration of prior use.

The implication of these recent studies is that the evidence no longer favors HRT in the routine treatment of menopause.  Many alternatives to HRT are available to address the cardiovascular and skeletal sequelae of estrogen deficiency.  However,  HRT remains the most effective agent for the treatment of vasomotor and other climacteric symptoms, and if a woman is unable to find relief from disabling symptoms with alternative therapies, a joint decision between the physician and the informed patient to prescribe HRT remains a reasonable course of action.

Gilbert G. Fareau and Rena Vassilopoulou-Sellin
Estrogen replacement therapy in breast cancer patients:  a time for change? Proc Am Soc Clin Oncol 1996;15:121.

References