Estrogen Replacement in Coronary Artery Disease: More Bad News

Abstract & Commentary

Synopsis: Women with heart disease should not use conjugated estrogen, alone or in combination with medroxyprogesterone acetate, with an expectation of cardiovascular benefit.

Source: Herrington DM, et al. N Engl J Med 2000;343: 522-529.

The heart and estrogen/progestin replacement Study (HERS) was a major disappointment for health care providers.1 Examination of other databases as well as an early report from the Womens’ Health Initiative appear to support the HERS data, suggesting a potential adverse outcome in the first year or two following hormone replacement therapy (HRT) initiation. The ERA study represents another important look at this issue. Herrington and colleagues at six centers, randomized 309 postmenopausal women with documented coronary disease to estrogen (E), progesterone (P), E plus P, or placebo. The primary end point was angiographic progression of atherosclerosis (mean minimal coronary artery diameter). Secondary end points were related to angiographic factors. In addition, clinical events were tracked. Eligible women had to have at least one greater than 30% narrowing of a major epicardial artery and be postmenopausal for five years. Follicle-stimulating hormone (FSH) levels were used to confirm absence of estrogen activity; most subjects were older than 55 years of age. Obvious contraindications to estrogen therapy were exclusion criteria, as well as uncontrolled hypertension or diabetes. A second angiogram was available for review in 248 or 80% of the initial cohort. The mean follow-up between the first and second angiogram was 3.2 ± 0.6 years. Compliance to therapy was highest in the E ± P and placebo groups (85% compared to E alone of 75%). Subjects were followed with annual mammography and gynecological examination including measurement of uterine wall thickness. The baseline population had a mean age of 66 (range, 42-80). Hormone replacement reduced LDL cholesterol by 9.4% and 16.5 % in E ± P and E alone, respectively. HDL levels increased by 14-19% and triglcyerides rose 6-10% in both active treatment groups. Insignificant changes were noted with placebo. Primary analysis of 2317 proximal coronary segments (mean of 9.3 per patient) showed no significant differences in minimal diameter change from baseline in the active-treatment and placebo groups, nor in the increase in percent stenosis. Adjustments were made for a variety of variables, and the analysis remained the same. New lesions developed in 30% of E and 20% of E and P, vs. 33% of placebo women. Baseline segments with minimal to no angiographic disease showed no differences over time among the three groups. There was considerable uterine bleeding in the unexposed E women, who had a greater incidence of uterine hyperplasia. Cancer or fractures were not different among the groups, although there was a trend for more fractures in placebo patients. Clinical events were no different among the groups; those who underwent coronary revascularization were equally distributed. Herrington et al conclude that estrogen replacement had no effect of slowing progression of coronary atherosclerosis in these women, consistent with the HERS trial. They point out that E alone vs. E ± P made no difference in the results. This raises the possibility that estrogen may be "more effective in preventing the development of atherosclerosis than in slowing the progression of disease." In the absence of new clinical trial data, they conclude that "women with heart disease not use conjugated estrogen, alone or in combination with medroxyprogesterone acetate, with an expectation of cardiovascular benefit."

Comment by Jonathan Abrams, MD

This important study first reported at the American College of Cardiology meetings in March 2000, provides additional evidence that HRT is not beneficial for slowing coronary heart disease, in spite of the large epidemiologic and observational database, as well as a considerable basic and animal science. While the ERA trial does not suggest harm, it had a relatively short follow-up with a small cohort. For instance, only 204 patients took active hormone treatment compared to 105 placebo women. In contrast to HERS, no adverse early clinical events were noted. The side effects of vaginal bleeding and uterine hyperplasia are well known in women who take unprotected estrogen without a progestin. HERS and ERA, two major estrogen trials reported within the past 18-24 months, represent the only available randomized clinical trial database and are overwhelmingly neutral to negative with respect to benefits of HRT in women with established coronary artery disease (CAD). In both of these studies, women were well into their sixties with established vascular disease at baseline. It may be that initiation of HRT much earlier in the lifespan of coronary atherosclerosis, with dominant early lesions or no angiographic coronary atherosclerosis, might have different results, particularly over a long follow-up period of 5-10 years. Perhaps the Women’s Health Initiative will be fruitful, although the results will not be available for many years. A number of ongoing trials remain, some of which are angiographic in nature, examining the benefits of HRT in women. As of now, one must agree with the conclusion of Herrington et al, as well as many other experts, that physicians should not specifically prescribe hormone replacement treatment to women with vascular disease with the hopes of slowing progression of CAD.

The angiographic data in ERA contrasts with the many lipid lowering trials using similar methodologies. These individually, and in the aggregate, have consistently shown slowing of disease progression with statins. In addition, the overall experience in these regression studies have predicted event reductions that were subsequently shown in the large, randomized, clinical trials using statins, such as 4S, CARE, LIPID, representing thousands of patients with established CAD. The mechanisms for failure of HRT to support published experimental and observational data are unknown. Recent animal and human data indicate a pro-inflammatory effect of estrogen; several studies have documented elevated c-reative protein (CRP) levels in women using estrogen that may reflect more activation of CAD. Herrington et al suggest that estrogen benefits on the vessel wall may be seen primarily in women who have a relatively healthy endothelium. Thus, initiating HRT in women with established vascular disease, long-standing hypertension, or in the elderly, may lead to disappointment. Data in younger women with mild to no disease is unavailable. In conclusion, HRT should not be administered solely in hope of preventing or slowing coronary disease. However, the benefits of hormone therapy for women who have postmenopausal symptoms, genital/urinary problems relating to estrogen deficiency or osteoporosis, are well established. The risk of breast cancer with HRT remains controversial and at best is only a modest one. Furthermore, some evidence suggests that women who develop breast cancer who are taking estrogen may have a more benign form of the disease. Experimental data regarding dementia and HRT is of interest, but it is too preliminary to know if prevention of cognitive dysfunction will ultimately turn out to be an indication for estrogen or estrogen surrogates.


  1. Hulley S, et al. JAMA 1998;280:605-613.