As the system for financing and providing healthcare in the United States undergoes reform, increased emphasis is being placed on evidence-based medicine, that is, medical management based on the results of carefully designed, conducted, and analyzed, large-scale, multicenter, randomized, placebo-controlled clinical trials. This approach has the dual advantages of avoiding risk by not recommending courses of clinical therapy that are as yet unproven and avoiding cost by not paying for unproven therapy.
The practice of evidence-based medicine can be very comfortable. For example, because of the results of three multicenter trials of coronary bypass surgery, cardiologists can confidently recommend revascularization for symptomatic patients with hemodynamically significant lesions of the left main coronary artery. Unfortunately, one of the most difficult aspects of being a practicing physician is the frequent need to make decisions on the basis of incomplete data. This is the dilemma that faces physicians who follow the recommendation of the American College of Physicians that all postmenopausal women be considered for estrogen replacement therapy1 and the recommendation of the National Cholesterol Education Program Adult Treatment Panel II that postmenopausal women with hypercholesterolemia receive estrogen replacement.2 The results of large, randomized, controlled trials are not available.
What evidence supports this recommendation of the College of Physicians? Several recent reviews have addressed this issue.3 4 5 6 Thus, only the main points will be addressed here. Little controversy exists concerning the conclusion that bilateral oophorectomy results in an increased incidence and prevalence of coronary atherosclerosis. Over four decades ago, Wuerst and his coworkers7 reported that at autopsy, women who had had both ovaries removed had significant coronary stenosis as frequently as men of the same age, whereas women with intact ovaries had far fewer lesions until 10 to 15 years past menopause. The Framingham Study observed that the incidence of cardiovascular disease increased about fourfold in women who underwent menopause before the age of 40 years.8
Controversy does exist about the effect of natural menopause on cardiovascular risk. While it is clear that the risk of cardiovascular disease increases with age, the magnitude of the contribution of the loss of ovarian function to this result remains a matter of discussion. The Framingham Study observed that the risk of cardiovascular disease increased approximately twofold in women who underwent menopause in their fifties; however, other epidemiological studies have not shown an increased risk.9
More than 30 observational studies have examined the effect of estrogen replacement therapy on cardiovascular events and on all-cause mortality.3 5 There have been 13 case-controlled studies using either hospital- or population-based controls. Ten of these studies found that women receiving estrogen had fewer cardiovascular events than those who did not. However, only one of these studies was statistically significant. There have also been 17 prospective cohort studies, using either internal or external controls. Sixteen of these studies found a protective effect of estrogens. The exception was the Framingham Study, which found an increased risk of cardiovascular disease in estrogen users, which might be explained by different ages and end points.10
Laboratories in four different institutions have carried out cross-sectional studies to examine the effect of estrogen replacement therapy on the degree and extent of coronary atherosclerosis.11 12 13 14 Significantly less coronary disease was found in women who used estrogen replacement in all four of the studies.
In one of these studies, the effect of estrogen replacement therapy on all-cause mortality in women whose coronary status had been determined by coronary angiography was examined over a period up to 10 years.15 Women who did not have coronary artery disease at baseline had a good survival rate regardless of whether estrogen replacement therapy was used. Five-year survival was 98% in women who had never used estrogen replacement therapy and those who had used estrogen at some time since angiography. By 10 years, survival was 91% in the never-users and 98% in the ever-users, a difference that did not reach statistical significance. In contrast, the women who had coronary disease and used estrogens had significantly better survival rates. Among women who had mild to moderate coronary lesions, survival was 85%, whereas the group that used estrogen had a survival of 96% (P=.027). The difference was greatest in the women with severe coronary artery disease, with 5-year survival of 81% in the never-users and 97% in the users, a difference increasing to 60% survival in the never-users and 97% in the users at 10 years (P=.007). In this study, Cox's multiproportional hazards model was used to estimate survival as a function of multiple covariables, and estrogen use was found to have a significant effect on survival that was independent of other risk factors.
Data from two other large prospective cohort studies also support the use of estrogen replacement therapy for the secondary prevention of cardiovascular disease. In the Lipid Research Clinics Study,16 which included a cohort of 2270 women followed an average of 8.5 years, the mortality rate was 12.8 per 10 000 in estrogen users who did not have clinical evidence of cardiovascular disease at baseline, whereas the mortality rate was 30.2 per 10 000 in estrogen nonusers, a reduction of 58%. In women with prevalent cardiovascular disease, a 79% reduction was observed, with a death rate of 13.8 per 10 000 in the users and 66.3 per 10 000 in the women who did not use estrogen.
Similarly, the Leisure World Study,17 which included 8881 postmenopausal women followed over an average of 7.5 years, found an all-cause mortality rate of 21.8 per 1000 in women without a history of angina or myocardial infarction who used estrogen and 26.7 per 1000 if they had not used estrogen, a reduction of approximately 18%. In those women who had a history of myocardial infarction or angina pectoris, all-cause mortality was 41.7 per 1000 in the estrogen nonusers and 27.5 per 1000 in the estrogen users, a reduction of 34%.
A meta-analysis by Stampfer and Colditz3 of all the studies cited above found an overall risk reduction of approximately 44% when estrogen users were compared with those who had never used estrogen replacement. A better relative risk of 0.50 (95% CI, 0.45 to 0.59) was found in current users.5 Thus, the data are strong and consistent that estrogen replacement therapy is associated with a lower cardiovascular and total mortality rate. However, most of these studies have used estrogen alone. In contemporary practice, to avoid the risk of abnormal endometrial proliferation and endometrial cancer, estrogen is given with a progestin, administered either continuously or cyclically. Because the addition of a progestin blunts the increase in HDL cholesterol associated with estrogen administration, an effect that is believed to play an important role in the improved cardiovascular outcome, concern has been expressed that the benefits observed in earlier studies are not relevant to current practice. The clinical data addressing this issue are limited. Five studies have either used hormone replacement therapy or have compared groups who used hormone replacement therapy with those who have used estrogen replacement therapy alone. In the only published randomized trial of estrogen therapy, that of Nachtigall et al,18 medroxyprogesterone acetate was also used to treat 84 pairs of hospitalized women. In this study, the risk of myocardial infarction was lowered to 0.33. However, this was not statistically significant. Hunt et al19 conducted a prospective cohort study in the United Kingdom that involved 4544 women. In this study, the relative risk of death due to cardiovascular disease was reduced to 0.37 (95% CI, 0.15 to 0.58). In a population-based case-controlled study, Thompson et al20 found no significant difference in the rate of myocardial infarction or stroke. In contrast, a recent Swedish study observed almost equal reductions in the risk of first myocardial infarction when women who received estrogen replacement therapy were compared with those receiving estrogens plus progestin.21 In the Puget Sound area, Psaty et al22 found that estrogen replacement therapy lowered the relative risk of myocardial infarction to 0.69 (95% CI, 0.02 to 0.47). Women who were receiving estrogen plus progestin had an almost identical reduction of risk to 0.68 (95% CI, 0.38 to 1.22). Recently, 16-year follow-up data from the 59 337 participants of the Nurses' Health Study showed that the relative risk for major coronary heart disease was 0.39 (95% CI, 0.19 to 0.78) when women who used estrogen and progestin were compared with women who did not receive hormone replacement therapy. The relative risk in those who used estrogens alone was 0.60 (95% CI, 0.43 to 0.83).23 Recently, 16-year follow-up data from the 59 337 participants of the Nurses' Health Study showed that the relative risk for major coronary heart disease was 0.39 (95% CI, 0.19 to 0.78) when women who used estrogen and progestin were compared with women who did not receive hormone replacement therapy. The relative risk in those who used estrogens alone was 0.60 (95% CI, 0.43 to 0.83).22a22A Therefore, although very little evidence is available, that available thus far does not suggest that the addition of progestins eliminates the effect of estrogen replacement therapy on cardiovascular risk.
An impressive body of evidence has grown to support the biological plausibility of a cardioprotective effect due to estrogen.23 Initially, this effect was thought to be due solely to an effect on HDL cholesterol, which rises during estrogen replacement. In addition, LDL cholesterol levels and lipoprotein(a) are lowered by estrogen replacement therapy, whereas triglyceride levels rise. The addition of progestin has little effect on the LDL cholesterol–lowering properties of estrogen. However, the rise in HDL is blunted, less so when micronized progesterone is used. Current estimates put the lipid-altering properties of estrogen as accounting for approximately 25% to 33% of the observed cardioprotective effect. Thus, an extensive search for other mechanisms has been carried out and continues. Estrogen replacement therapy has been found to restore endothelial function after menopause by upregulating the transcription of nitric oxide synthase with increased release of nitric oxide, which makes coronary vessels less likely to develop spasm. Data obtained from experimental animals suggest that progestins might interfere with this effect. Estrogen replacement therapy also results in enhanced production of prostaglandin I2 and decreased production of thromboxane A2. Estrogen lowers plasma endothelin levels and blocks endothelin receptors. Part of the vasodilatory effect of estrogens probably is due to calcium blockade and to the opening of potassium channels. Estrogen replacement therapy increases insulin sensitivity, thus lowering plasma insulin and glucose levels, which have been associated with premature atherosclerosis. After menopause, plasma fibrinogen and plasminogen activator inhibitor levels rise, both of which are independent cardiovascular risk factors. Estrogen replacement therapy reverses this rise. However, estrogen replacement therapy also lowers antithrombin III and raises factors VII and X; thus, the overall effect on the clotting factors might be nil. Estrogen is an antioxidant and decreases oxygen free radicals in arterial walls, which may account for the effect on vasomotor tone observed within 30 minutes of estrogen administration, an effect that cannot be mediated by genomic mechanisms in such a short time. Estrogen influences the proportions of collagen, elastin, and procollagen. Estrogen enlarges collateral channels by enhancing migration and proliferation of endothelial cells and potentiates the angiogenic effect of fibroblast growth factor.
It has been proposed that selection bias might account for the better outcome observed in women who use estrogen because they have more education, affluence, access to healthcare, and fewer risk factors for cardiovascular diseases.24 The magnitude of the difference is unlikely to explain a 50% reduction of cardiovascular end points. However, one large study concluded that “the antecedent data, obtained 15 years previously, also suggest that current (1985) estrogen users were not significantly healthier than current nonusers with regard to age at menopause or risk factors for heart disease before estrogen use.” A follow-up study of the same population showed that a higher proportion of the women who continued to use estrogen had also increased their level of exercise and underwent more screening tests.25
Despite the observed benefits on blood vessels and bones, estrogen replacement therapy is not without risk. Estrogen increases the risk of endometrial carcinoma approximately sixfold, an effect that is eliminated by the addition of progestins. Controversy continues over whether estrogen replacement increases the risk of breast cancer, a disorder that afflicts 1 woman in 8. Various meta-analyses have found either no effect or an increase in risk of approximately 30%, depending on which studies are included or excluded from the meta-analysis. Two recent large-scale studies similarly revealed conflicting results. The large Nurses' Health Study reported an increase in the risk of breast cancer of approximately 40%, which is particularly high in older women who had been using hormones for at least 5 years.26 The inclusion of progestin had no effect on this risk. In contrast, a study involving the general female population of King County, Washington, found that the use of estrogen with progestin did not appear to be associated with an increased risk of breast cancer in middle-aged women. The relative odds were 0.9 (95% CI, 0.7 to 1.3). For those who had used estrogen/progestin for more than 8 years, the risk was reduced to 0.4 (95% CI, 0.2 to 1.0).27 Thus, although a link with breast cancer has been sought, conclusive evidence is lacking. On this issue as well, there are no randomized trials.
Another risk is that of deep vein thrombosis and pulmonary embolism. In a British study involving 103 cases of idiopathic venous thrombolism and 178 female control subjects, current users were found to have an odds ratio of 3.5 (95% CI, 1.8 to 7.0) compared with those who did not use hormone replacement.27a27A In a similar case-control study the Group Health Cooperative of Puget Sound found a relative risk of 3.2 (95%, CI, 1.5 to 6.8) comparing current estrogen users with nonusers.27b27B The Nurses' Health Study observed a twofold increase in the risk of thromboembolism among current postmenopausal estrogen users.27c27C Although the three studies found that estrogen use was associated with an increase in the relative risk of venous thromboembolism, the absolute risk was low, because venous thrombolism occurred infrequently. When weighed against a 44% reduction in cardiovascular disease, a highly prevalent disorder, the increased risk of venous thromboembolism does not contraindicate estrogen replacement but does point out the need for attention to a prior history of idiopathic thrombosis.
Several studies are under way that one hopes will establish whether estrogen replacement therapy definitely reduces the risk of cardiovascular disease in women with and without coronary artery lesions and whether it increases the risk of breast cancer. The Heart and Estrogen/Progestin Replacement Study will be completed in about 1½ years. In this study, women with documented coronary artery disease have been randomized to receive either placebo or a combined, continuous conjugated equine estrogen/progestin regimen. Two studies of estrogen replacement therapy are under way in women with coronary artery disease to determine whether estrogen replacement therapy delays the progression or induces the regression of atherosclerotic lesions. Another study will examine whether estrogen replacement therapy alters the accuracy of cardiac diagnostic tests in postmenopausal women.
A fifth very important study is the Women's Health Initiative. This nationwide study is attempting to recruit 27 500 women to assess the effect of hormone replacement therapy. The women with uterus intact will be randomized to receive either placebo or combined estrogen/progestin therapy, using conjugated estrogens in a daily dose of 6.25 mg and medroxyprogesterone acetate in a dose of 2.5 mg. The women who have had a hysterectomy will be treated with estrogen alone. It is anticipated that treatment will be observed over a period of 9 years unless interim analyses of the emerging results indicate that the trial should be stopped either because of benefits or adverse outcomes. Unfortunately, this trial faces several challenges. First of all, it may prove difficult to recruit a large number of women who are willing to be randomized into estrogen replacement therapy because many have already decided whether they will accept estrogen replacement. In a survey conducted by Ravnikar,31 20% to 30% of women never filled their prescriptions for hormone therapy, and the overall compliance rate was only 30%. In addition, among those who agree to participate, industry data suggest that the percentage of women who continue estrogen replacement therapy for more than 1 year is only approximately 40%. Many women discontinue therapy because they do not wish to have regular menstrual periods or do not tolerate combined continuous therapy because of unpredictable bleeding. Furthermore, if the trial is stopped after a few years because of a definite favorable effect on cardiovascular events, the issue of breast cancer, which appears to require a much longer period of therapy to evolve, will remain unanswered. The Women's Health Initiative will test only one therapeutic regimen, a conjugated equine estrogen that contains nine steroid components in addition to 17-β estradiol. Thus, it will not be clear whether all forms of estrogen replacement therapy will have the same effect. In an article in this issue of Circulation, Dr Rossouw,3232 of the Women's Health Initiative, calls for caution and “putting the brakes on the bandwagon” of estrogens for the prevention of coronary heart disease. Even with the unargued beneficial effect on osteoporosis and the known benefits on vascular lability and vaginal dryness, data collected by industry do not suggest that it is much of a bandwagon. At present in the United States, only approximately 20% of women take estrogen replacement therapy. This varies from a low of 4% in New England to a high of 30% in California. It is my opinion that there will not be widespread use until the results of randomized trials are in. Certainly there was limited use of antihypertensive agents until the Veterans Administration Trial and the Hypertension Detection and Follow-up Trial provided conclusive evidence that the treatment of mild to moderately elevated levels of blood pressure did more good than harm. Similarly, despite the widespread appreciation of the dangers of elevated cholesterol levels, physicians did not recommend, and patients did not seek, cholesterol-lowering therapy to any great degree until the results of the Lipid Research Clinic Study and the Helsinki Heart Study became available; even now, with the impressive results of atherosclerosis regression trials and improved outcome studies published, relatively few individuals with known coronary artery disease are receiving the aggressive treatment recommended by the National Cholesterol Education Program. If the results of the Heart and Estrogen/Progestin Replacement Study and the Women's Health Initiative conclusively prove that estrogen replacement therapy reduces the risk of coronary artery disease, the challenge to the medical profession will be to persuade women to comply with the appropriate recommendation as well as with recommendations about cigarette smoking, diet, exercise, and control of blood pressure and cholesterol.
The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association.
- Copyright © 1996 by American Heart Association
American College of Physicians. Guidelines for counseling post-menopausal women about preventive hormone therapy. Ann Intern Med. 1992;117:1038-1041.
Grady D, Rubin SM, Petitti DB, Fox CS, Black D, Ettinger B, Ernster VL, Cummings SR. Hormone therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med. 1992;117:1016-1037.
Wuerst JH Jr, Dry TJ, Edwards JE. The degree of coronary atherosclerosis in bilaterally oophorectomized women. Circulation. 1953;7:801-809.
Lerner DJ, Kannel WB. Patterns of coronary heart disease morbidity and mortality in the sexes: a 26-year follow-up of the Framingham population. Am Heart J. 1986;113:383-390.
Sullivan JM, Vander Zwaag R, Lemp GF, Hughes JP, Maddock V, Kroetz FW, Ramanathan KB, Mirvis DM. Postmenopausal estrogen use and coronary atherosclerosis. Ann Intern Med. 1988;108:358-363.
Bush TL, Barrett-Connor E, Cowan LD, Criqui MH, Wallace RB, Suchindran CM, Tyroler HA, Rifkind BM. Cardiovascular mortality and noncontraceptive use of estrogen in women: results from the Lipid Research Clinics Program Follow-up Study. Circulation. 1987;75:1102-1109.
Thompson SG, Meade TW, Greenberg G. The use of hormonal replacement therapy and the risk of stroke and myocardial infarction in women. J Epidemiol Community Health. 1989;43:173-178.
Grodstein F, Stampfer MJ, Manson JE, Colditz GA, Willett WC, Rosner B, Speizer FE, Hennekens CH. Postmenopausal estrogen and progestin use and the risk of cardiovascular disease. N Engl J Med. In press.
Gerhard M, Ganz P. How do we explain the clinical benefits of estrogen? From bedside to bench. Circulation. 1995;92:5-8.
Barrett-Connor E. Postmenopausal estrogen and prevention bias. Ann Intern Med. 1991;115:455-456.
Daly E, Vessey MP, Hawkins MM, Carson JL, Gough P, Marsh S. Increased risk of venous thromboembolism in hormone replacement therapy users. Lancet. In press.
Jick H, Jick SS, Gurewich V, Myers MW, Vasilakis C. Risk of idiopathic venous thromboembolism among users of postmenopausal estrogens. Lancet. In press.
Rossouw JE. Estrogens for prevention of coronary heart disease: putting the brakes on the bandwagon. Circulation.. 1996;94:2982-2985.