The Journal of Clinical Endocrinology & Metabolism Vol. 84, No. 6 1806-1810
Copyright © 1999 by The Endocrine Society
Does Estrogen Have a Role in the Prevention of Cardiovascular Disease?
Jacques E. Rossouw
Women’s Health Initiative
National Heart, Lung, and Blood Institute Bethesda,
Maryland 20892
 |
Introduction
|
|---|
PREVENTIVE cardiology can point to a proud
tradition of evidence-based practice. In particular, large randomized
clinical trials were required before accepting the benefits of
antihypertension or cholesterol lowering therapies. Demonstration of
the association of high blood pressure or high blood cholesterol with
coronary heart disease (CHD), and that drugs lowered these risk
factors, were deemed insufficient; we also wanted to know that
treatment lowered CHD events and did so without increasing the risk of
noncoronary events (1, 2, 3, 4). This tradition is not being upheld in the
case of estrogen. In actuality, we do not currently know whether
estrogen will prevent cardiovascular disease, and we will not know
until the large clinical trials now underway are completed. Despite
this lack of certainty, most physicians in the United States are
convinced that estrogen works, are being advised to prescribe it for
prevention of CHD, and are doing so in increasing numbers (5). Numerous
articles on this topic in the scientific literature encourage this
trend, and the supposed benefits of estrogen are widely touted by
written and electronic news media. Respected bodies responsible for
putting out practice guidelines have also succumbed to the lure of
estrogen for the prevention of CHD (6, 7). While these guidelines pay
lip service to the need for clinical trials, they nonetheless take an
encouraging stance towards estrogen treatment especially for women at
high risk of CHD. The purpose of this contribution is to offer reasons
why a more cautious attitude toward the widespread use of long-term
estrogen might be in order.
|
Evidence that estrogen may prevent CHD
|
|---|
Physicians have some justification for believing in estrogen
therapy. Taken at face value, the flood of observational studies,
together with studies of intermediate mechanisms, may appear
convincing. Indeed, it is quite possible that these studies are correct
in their prediction of benefit. But it is at least equally possible
that they are wrong. The studies may be wrong even to the extent that
there may be no benefit whatsoever for CHD prevention, or that any
benefit is insufficient to offset the adverse effects.
Observational studies. A large number of observational studies
have suggested that women who have ever taken estrogen appear to have a
30–50% lower risk of CHD than women who have never used estrogens (8, 9). Benefit appears to be strongest in current users (8, 10). Fewer
studies have data on estrogen use in combination with a progestin, but
apparent risk reductions are similar to those for estrogen alone
(10, 11, 12). The data for stroke are less consistent (13). For all types
of strokes combined there appears to be no net effect, although in the
Nurses’ Health Study there was a significant trend towards increased
strokes with increasing dose, as well as an excess of thrombotic
strokes overall in hormone users (10). Recent observational studies
have indicated that hormone users may have a 3-fold excess risk for
venous thromboembolism (14, 15, 16). Finally, while the observational
studies suggest that there may be only a modest excess incidence of
breast cancer in ever-users, current users appear to be at higher risk,
and prolonged current use appears to confer a substantial excess risk
of death from breast cancer (17).
Because CHD is more common than any of the other hormone-related
outcomes discussed here, it follows that a strong effect on the
prevention of CHD will translate into overall benefit even if there are
adverse trends for less common outcomes (Fig. 1
).Estimates of overall mortality or years of life saved that accept the
observational data at face value will tend to show overall benefit (18, 19). However, these estimates are only as good as the data on which
they are based, and all the observational data share a potentially
serious flaw: women who take estrogen are healthier than women who do
not.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 1. Risk of death (odds ratio and 95% confidence interval) among current postmenopausal hormone users compared to never-users in the Nurses’ Health Study (19 ). Note the apparently reduced risk for all cancer as well as that for coronary heart disease. Values are adjusted for age, age at menopause, type of menopause, body mass index, diabetes, high blood pressure, high cholesterol, smoking, past use of oral contraceptives, family history of myocardial infarction, family history of breast cancer, parity, age at menarche, and time period.
|
|
The effect of the numerous biases that affect these studies will be to
overestimate benefits and underestimate risks. Thus, estrogen users
will appear to have a lower risk of death from almost all causes,
including cancers and other diseases with no plausible biological
relationship to estrogen (20, 21). Relative risk for all cancers in
estrogen users can be used as an index of bias (20). The studies with
the lowest relative risks for CHD death also have the lowest risk for
cancer death (Fig. 2).On average at least 20 percentage points of the CHD risk reduction can
be explained by bias (20). The recent mortality findings from the
Nurses’ Health Study share the same bias, and fit well on the
regression line of previous studies (19, 20). Unintended selection of
healthy women may help explain these findings and may contribute to an
exaggerated estimate of a cardioprotective effect. Importantly, a
better health profile has been shown to be present even before women
are prescribed estrogens (22).
View larger version (16K):
[in this window]
[in a new window]
|
Figure 2. Relationship between relative risk for total cancer and cardiovascular disease in observational studies, illustrating that the unintended selection of healthy women for estrogen therapy may have influenced the reported beneficial effect on cardiovascular disease [adapted from Postuma et al. (20 )].%*Note: the Nurses’ Health Study mortality findings fell on the same regression line as the other studies (19 ).
|
|
Compliance bias is a well established phenomenon. Patients who are
compliant, (i.e. good at pill-taking—irrespective of
whether the pills are active or placebo) have a lower mortality (23).
By definition, women who continue to take estrogen over a period of
many years are compliant. Furthermore, hormone users are likely to be
followed more carefully, and disease discovered and treated at an early
stage, lowering mortality (surveillance bias). Finally, women who have
recently stopped estrogen have higher relative risks for all causes of
death, indicating cessation of treatment when risk factors or early
disease become manifest (21). Thus, the selection of who is placed on
treatment, who stays on treatment and under surveillance, and who is
taken off treatment, rather than the treatment itself may account for
much of the lower subsequent risk of disease, including heart disease,
in women who remain currently on estrogen. These biases may be very
powerful, e.g. compliance bias alone may account for a 60%
reduction in mortality, which may be erroneously attributed to the
effects of estrogen (23).
Though the best studies attempt to do so, it is impossible to fully
correct for all the potential biases in analyzing observational data
(24). For example, it is not possible to fully correct for compliance
bias in observational studies comparing current users to nonusers or
ever-users, as current users (especially long-term users) are defined
by the very fact that they are compliant. Furthermore, combining the
results of the observational studies in meta-analyses is not helpful
and may even be misleading. If there is a systematic bias in study
data, then combining the data of all studies ascribes a significance
level to the bias, but does not illuminate the basic question (25).
Currently, we simply cannot tell from the observational studies whether
there is a real benefit of estrogen on CHD, or what the size of any
effect might be. The only way to obtain a reliable estimate of the real
effects of estrogen is through randomized controlled clinical trials,
which if large enough, will eliminate the possibility that differences
between study groups account for the results. Even though insufficient
for public health recommendations, the observational data are certainly
strong enough to justify the need for trials of clinical outcomes.
Trials of intermediate outcomes. While these trials constitute
stronger evidence than observational studies, they too are not
definitive. In the past, they have frequently provided misleading
information about clinical questions (26). While we now have good data
from such trials that estrogen with and without progestin have
generally favorable effects on blood lipids (27), it is by no means
clear that these lipids are the mediators of any estrogen effects on
vascular health. For example, the animal data suggest that changes in
high density lipoprotein cholesterol may not mediate the effects on the
vessel wall (28), and in humans the observational studies suggest that
the benefit from estrogen is not attenuated by adding a progestin, even
though progestins greatly decrease the impact of estrogen on high
density lipoprotein cholesterol. Similarly, while the relevance of
estrogen’s effects on vessel wall physiology to clinical events
remains to be established, progestins do counter those effects (29, 30).
Favorable changes in blood lipids or vessel wall physiology will not
necessarily translate into less vascular pathology, as they could be
neutralized by one or more of the myriad other effects of estrogen. The
data on coagulation factors are not reassuring, as potentially
favorable effects on fibrinogen and plasminogen may well be
counterbalanced by increases in Factor X, Factor VII, and decreases in
antithrombin III (26, 31). Changes in coagulation factors might explain
the apparently increased risks for thrombotic strokes and venous
thromboembolism in current estrogen users (10, 16), and the excess of
venous thromboembolism and CHD found in the early estrogen trials
conducted in men (32, 33). While less well established than the
standard risk factors, there is a growing body of evidence that
coagulation factors play a role in the etiology of CHD (34).
|
Why do we need clinical trials?
|
|---|
Importantly, until recently, no study has directly and reliably
measured the effects of these hormones on the clinical manifestations
of CHD in women, and any projection of potential benefit based on
observational epidemiology or surrogate biological variables is highly
speculative. Treatments suggested by favorable epidemiologic
associations or effects on intermediate biological outcomes may not
confer benefit, or they may even harm patients in terms of actual
disease outcomes when put to the test in a clinical trial (25). The
disappointing and even alarming results of recent trials of
beta-carotene (35, 36) and meta-analyses of trials of short-acting
calcium channel blockers are good examples (37).
Because of their inherent weaknesses, further observational studies or
trials of intermediate biological outcomes will not answer the
important public health questions. Will postmenopausal hormone
replacement therapy reduce CHD incidence, and/or will it increase the
risk of breast cancer, and what are the overall benefits and risks of
long-term use? To what extent do the benefits and risks apply to older
women? Should most postmenopausal women be prescribed hormone
replacement therapy?
Only randomized trials of sufficient size and duration and with
clinical outcomes can answer these questions. Trials done to date have
been mainly in men and have used higher doses of estrogen than those in
current use. Nonetheless, the results were not encouraging (Fig. 3)(32, 33, 38, 39, 40, 41). In the large Coronary Drug Project, both estrogen
arms were stopped early because of adverse trends in CHD, venous
thromboembolism, cancer, and total mortality (32, 33). A summary
analysis of the small number of clinical outcomes in 22 short-term
intermediate outcome trials in women also failed to support the idea
that estrogen was cardioprotective (42).
View larger version (22K):
[in this window]
[in a new window]
|
Figure 3. Meta-analysis of previous trials of estrogen to prevent coronary heart disease (32 33 38 39 40 41 ). In aggregate the trials do not suggest a cardioprotective effect of estrogen. Two large trials in men were stopped early because of excess mortality trends, and excesses of coronary heart disease, venous thromboembolism, and cancer (32 33 ).
|
|
The large clinical outcome trials now being done will provide unbiased
estimates of the benefits and risks of hormone replacement therapy
(HRT). The first of these, the Heart and Estrogen/Progestin Study
(HERS), was a secondary prevention trial that published its results in
1998 (43). HERS randomized 2,763 postmenopausal women with established
CHD and an intact uterus to either a placebo or to conjugated
equine estrogens (Premarin), 0.625 mg plus
medroxyprogesterone, 2.5 mg daily and followed the
participants for an average of 4.1 yr. The overall result was that
there was no effect on the primary outcome of CHD, and a significant
3-fold increase in the incidence of venous thromboembolism. There were
no significant effects on any other outcome, including breast cancer;
however, the trial was too small and too short in duration to have
adequate statistical power to examine effects on these less common
outcomes. For CHD, there was an interesting trend in that the results
were in an adverse direction early on, then gradually reversed to a
beneficial direction in the final years. This trend was statistically
significant, but as it was not hypothesized before starting the trial
the finding should be viewed with some caution. Nevertheless, it is of
interest because it may suggest something about the mechanisms at play.
The early adverse trend may be explained by procoagulant effects
predominating early on, and causing an increased rate of CHD women with
diseased arteries. Over the longer term, the favorable lipid effects of
HRT may have predominated. It can be speculated that, if the trial had
continued for longer, the overall effect on CHD might have been
favorable. On the other hand, it is equally possible that unfavorable
effects on breast cancer might have emerged and countered what would
have been at best a modest overall benefit for CHD. In any event, the
result of no overall benefit for CHD was unexpected, and well
illustrates the potential pitfalls of taking observational data at face
value. The HERS findings do not lend support to the use of HRT for the
secondary prevention of CHD. At a minimum, the HERS findings should
caution physicians to lower their expectations as to what HRT might
achieve.
Clearly, more information than could be obtained from a single trial is
needed before evidence-based treatment recommendations can be made. The
HERS findings applied only to women with existing heart disease, HERS
did not test estrogen alone, and the trial was of relatively short
duration. Therefore, HERS could not address the issue of overall
benefit and risk. Two primary prevention trials are currently ongoing:
in the USA the Women’s Health Initiative (WHI) completed recruitment
in 1998 and plans to have results after 2005 (44), and in the United
Kingdom, a trial called Women’s Intervention Study of Long Duration
Oestrogen after the Menopause (WISDOM) commenced recruitment in 1999
and plans to have results in 2005 (Madge Vickers, Ph.D., oral
communication). Both of these trials are large, with over 27,000 women
in WHI and 34,000 in WISDOM, and an average follow-up period of 9 yr.
Each of these trials will randomize participants to estrogen alone,
estrogen with a progestin, and placebo. The combined results will
provide answers to the burning question: should most postmenopausal
women consider HRT?
|
Standard treatments for the prevention of CHD
|
|---|
For the prevention of CHD and stroke, there are better proven and
safer alternatives to estrogen. These include diet, weight control,
smoking cessation, and exercise. Cholesterol lowering with statins have
been shown to be very effective in preventing a second heart attack in
women in two trials, even in women who do not have very high levels of
cholesterol (Fig. 4)(44, 45). Blood pressure control prevents strokes and heart disease in
women as well as in men. Low dose aspirin and beta blockers are
effective in secondary prevention in women. Thus, there is no
imperative to turn to an unproven and potentially harmful treatment
such as estrogen. Available standard treatments should be used, while
continuing to research the potential value of estrogen in place of or
in addition to existing treatments.
View larger version (17K):
[in this window]
[in a new window]
|
Figure 4. In secondary prevention trials cholesterol lowering with statins significantly lowered coronary event rates in both men and women (44 45 ).
|
|
|
References
|
|---|
-
Collins R, Peto R, MacMahon S, et al. 1990 Blood pressure, stroke, and coronary disease, II. Short-term reductions
in blood pressure: overview of randomized drug trials in their
epidemiologic context. Lancet. 335:827–838.[CrossRef][Medline]
-
Law MR, Wald NJ, Thompson SG. 1994 By how much and
how quickly does reduction in serum cholesterol lower risk of ischaemic
heart disease? BMJ. 308:367–373.[Abstract/Free Full Text]
-
Law MR, Thompson SG, Wald NJ. 1994 Assessing
possible hazards of reducing serum cholesterol. BMJ. 308:373–379.[Abstract/Free Full Text]
-
Gould AL, Rossouw JE, Santanello NC, Heyse JF, Furberg
CF. 1995 Cholesterol reduction yields clinical benefit: a new look
at old data. Circulation. 91:2274–2282.[Abstract/Free Full Text]
-
Wysowski DK, Golden L, Burke L. 1995 Use of
menopausal estrogens and medroxyprogesterone acetate in the United
States, 1982–1995. Obstet Gynecol. 85:6–10.[Abstract]
-
American College of Physicians. 1992 Guidelines
for counseling post-menopausal women about preventive hormone therapy. Ann Intern Med. 117:1038–1041.
-
Expert Panel. 1993 Summary of the Second Report of
the National Cholesterol Education Program (NCEP) Expert Panel on
Detection, Evaluation, and Treatment of High Blood Cholesterol in
Adults (ATP-II). JAMA. 269:3015–3023.[CrossRef][Medline]
-
Stampfer MJ, Colditz GA. 1991 Estrogen replacement
therapy and coronary heart disease: a quantitative assessment of the
epidemiologic evidence. Prev Med. 20:47–63.[CrossRef][Medline]
-
Grady D, Rubin SM, Petitti DB, et al. 1992 Hormone
therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med. 117:1016–1037.
-
Grodstein F, Stampfer MJ, Manson JE, et al. 1996 Postmenopausal estrogen and progestin use and the risk of
cardiovascular disease. N Engl J Med. 335:453–461.[Abstract/Free Full Text]
-
Falkeborn M, Persson I, Adami HO, et al. 1992 The
risk of acute myocardial infarction after estrogen and
estrogen-progestogen replacement. Brit J Obstet Gynecol. 99:821–828.[Medline]
-
Psaty BM, Heckbert SR, Atkins D, et al. 1994 The
risk of myocardial infarction associated with the combined use of
estrogens and progestins in postmenopausal women. Arch Intern Med. 154:1333–1339.[Abstract]
-
Paganini-Hill A. 1995 Estrogen replacement therapy
and stroke. Progr Cardiovasc Dis. 38:223–242.[CrossRef][Medline]
-
Daly E, Vessey MP, Hawkins MM, Carson JL, Gough P, Marsh
S. 1997 Risk of venous thromboembolism in users of hormone
replacement therapy. Lancet. 348:977–980.
-
Jick H, Derby LE, Myers MW, Vasilakis C, Newton KM. 1997 Risk of hospital admission for idiopathic venous thromboembolism
among users of postmenopausal estrogens. Lancet. 348:981–983.
-
Grodstein F, Stampfer MJ, Goldhaber SZ, et al. 1996 Prospective study of exogenous hormones and risk of pulmonary embolism
in women. Lancet. 348:983–987.[CrossRef][Medline]
-
Colditz GA, Hankinson SE, Hunter DJ, et al. 1995 The use of estrogens and progestins and the risk of breast cancer in
postmenopausal women. N Engl J Med. 332:1589–1593.[Abstract/Free Full Text]
-
Col NF, Eckman MH, Karas RH, et al. 1997 Patient-specific decisions about hormone replacement therapy in
postmenopausal women. JAMA. 277:1140–1147.[Abstract]
-
Grodstein F, Stampfer MJ, Colditz GA, et al. 1997 Postmenopausal hormone therapy and mortality. N Engl J Med. 336:1769–1775.[Abstract/Free Full Text]
-
Postuma WFM, Westendorp RG, Vanderbrouke JP. 1994 Cardioprotective effect of hormone replacement therapy in
postmenopausal women: is the evidence biased? BMJ. 308:1268–1269.[Abstract/Free Full Text]
-
Sturgeon SR, Schairer C, Brinton LA, Pearson T, Hoover
RN. 1995 Evidence of a healthy estrogen user survivor effect. Epidemiology. 6:227–231.[Medline]
-
Matthews KA, Kuller LH, Wing RR, Meilahn EN, Plantinga
P. 1996 Prior to use of estrogen replacement therapy, are users
healthier than nonusers? Am J Epidemiol. 143:971–978.[Abstract/Free Full Text]
-
Horwitz RI, Viscoli CM, Berkman L, et al. 1990 Treatment adherence and risk of death after myocardial infarction. Lancet. 336:442–455.[Medline]
-
Sotelo M, Johnson S. 1997 The effects of hormone
replacement therapy on coronary heart disease. Endocr Metab Clin N Am. 26:313–328.
-
Shapiro S. 1994 Meta-analysis/Shmeta-analysis. Am J Epidemiol. 140:771–777.[Abstract/Free Full Text]
-
Fleming TR, DeMets DL. 1996 Surrogate end points in
clinical trials: are we being misled? Ann Intern Med. 125:605–613.[Abstract/Free Full Text]
-
Writing Group. 1995 Effects of estrogen or
estrogen/progestin regimens on heart disease risk factors in
postmenopausal women. The Postmenopausal Estrogen/Progestin
Interventions PEPI) Trial. JAMA. 273;199–208.
-
Adams MR, Kaplan JR, Manuck SB, et al. 1990 Inhibition of coronary artery atherosclerosis by 17-beta estradiol in
ovariectomized monkeys. Arteriosclerosis. 10:1051–1057.[Abstract/Free Full Text]
-
Leiberman EH, Gerhard MD, Uehata A, et al. 1994 Estrogen improves endothelium-dependent, flow-mediated vasodilatation
in post-menopausal women. Ann Intern Med. 121:936–941.[Abstract/Free Full Text]
-
Rosano GMC, Sarrel PM, Poole-Wilson PA, Collins P. 1993 Beneficial effect of estrogen on exercise-induced myocardial
ischaemia in women with coronary artery disease. Lancet. 342:133–136.[CrossRef][Medline]
-
Medical Research Council’s General Practice Research
Framework. 1996 Randomized comparison of estrogen versus estrogen
plus progestogen hormone replacement therapy in women with a
hysterectomy. BMJ. 312:473–478.[Abstract/Free Full Text]
-
Coronary Drug Project Research Group. 1970 The
Coronary Drug Project: initial findings leading to modifications of its
research protocol. JAMA. 214:1303–1313.[CrossRef][Medline]
-
Coronary Drug Project Research Group. 1973 The
Coronary Drug Project: findings leading to discontinuation of the 2.5
mg/day estrogen group. JAMA. 226:652–657.[CrossRef][Medline]
-
Ridker PM. 1996 The pathogenesis of atherosclerosis
and acute thrombosis: relevance to strategies of cardiovascular disease
prevention. In: Prevention of Myocardial Infarction. New York: Oxford
University Press; Manson JE, Ridker PM, Gaziano JM, Hennekens CH, eds.
32–54.
-
Hennekens CH, Buring JE, Manson JE, et al. 1996 Lack of effect of long-term supplementation with beta carotene on the
incidence of malignant neoplasms and cardiovascular disease. N
Engl J Med. 334:1145–1149.[Abstract/Free Full Text]
-
Omenn GS, Goodman GE, Thornquist MD, et al. 1996 Effects of a combination of beta carotene and vitamin A on lung cancer
and cardiovascular disease. N Engl J Med. 334:1150–1155.[Abstract/Free Full Text]
-
Furberg CD, Psaty BM, Meyer JV. 1995 Nifedipine. Dose-related increase in mortality in patients with
coronary heart disease. Circulation. 92:1326–1331.[Abstract/Free Full Text]
-
Marmorston J, Moore FJ, Kuzma OT, Magidson O, Weiner
J. 1960 Effect of Premarin on survival in men with myocardial
infarction. Proc Soc Exp Biol Med. 105:618–620.
-
Oliver MF, Boyd GS. 1961 Influence of reduction in
serum lipids on prognosis of coronary heart disease. Lancet.
ii:499–905.
-
Schoch HK. 1969 The U.S. Veterans Administration
Cardiology Drug-Lipid Study: an interim report. In: Holmes WL, Carlson
LA, Paoletti R, eds. Drugs Affecting Lipid Metabolism. New York: Plenum
Press; 405–420.
-
Nachtigall LE, Nachtigall RH, Nachtigall RD, Beckman
EM. 1979 Estrogen replacement therapy II: a prospective study in
the relationship to carcinoma and cardiovascular and metabolic
problems. Obstet Gynecol. 54:74–79.[Medline]
-
Hemminki E, McPherson K. 1997 Impact of
postmenopausal hormone therapy on cardiovascular events and cancer:
pooled data from clinical trials. BMJ. 915:149–153.
-
The Women’s Health Initiative Study Group. 1998 Design of the Women’s Health Initiative clinical trial and
observational study. Control Clin Trials. 19:61–109.[CrossRef][Medline]
-
Scandinavian Simvastatin Survival Study Group. 1994 Randomized trial of cholesterol lowering in 4444 patients with coronary
heart disease: the Scandinavian Simvastatin Survival Study. Lancet. 344:1383–1389.[CrossRef][Medline]
-
Sacks FM, Pfeffer MD, Moye LA, et al. 1996 The effect of pravastatin on coronary events after myocardial
infarction in patients with average cholesterol levels. Cholesterol and
Recurrent Events Trial investigators. N Engl J Med. 335:1001–1009.[Abstract/Free Full Text]
Top
Introduction
Evidence that estrogen may...
