Evidence for primary and secondary prevention of coronary artery disease in women taking oestrogen replacement therapy

European Heart Journal (1996) 17 {Supplement D), 9-14 Evidence for primary and secondary prevention of coronary artery disease in women taking oestrogen replacement therapy T. L. Bush The increasing use of oestrogen replacement therapy in women has focussed attention on the cardioprotective properties it has demonstrated. Historically, it has been shown that women enjoy a certain protection from heart disease, a phemen, however, which has t been studied extensively. Women at every age have less coronary artery disease (CAD) than men, even when various risk factors are accounted for, although the presence of diabetes carries equal mortality for both sexes. However, women who do develop CAD have a greater risk of mortality than men with CAD. Other gender differences include a later age of onset of CAD for women, and a difference in the type of atherosclerotic lesions developed. Most striking is the fact that, in women, high-density lipoprotein (HDL) seems to be a more potent predictor of major cardiovascular events than low-density lipoprotein (LDL), or total cholesterol. The Postmepausal Oestrogen and Progesterone Interventions (PEPI) Trial looked at changes in HDL, fibrigen, blood pressure and serum insulin resulting from oestrogen Introduction In the 1960s, the Coronary Drug Project and studies in England clearly showed an increase in thromboembolic disease and myocardial infarction (MI) in men with a previous history of MI given high doses of oestrogen' 1 " 31. Although studies in women had yet to be conducted, it was inferred from these results in men and from studies of oral contraceptive (OC) use in women that oestrogen replacement therapy in women would naturally increase the risk of heart disease. The many studies completed in the years since have proven us wrong; it appears that oestrogen therapy does t increase the incidence of heart disease in women, but, conversely, all evidence points to it being significantly cardioprotective. Correspondence: Trudy L. Bush MD, The JHU Womens Research Core, Lutherville, MD, The JHU Womens Research Core, Lutherville, MD, use. Four regimens were compared against placebo in 875 women. The results showed that HDL was increased significantly, LDL decreased significantly, fibrigen levels decreased significantly, and blood pressure and serum insulin levels were essentially unaffected by oestrogen and oestrogen/progestin interactions. The Heart and Oestrogen/ Progestin Replacement (HERS) Study, currently underway, is a secondary prevention trial testing the protective effect of hormone therapy in women with documented CAD. This trial may definitively answer the question of whether hormones protect against CAD. After HERS, it may be unethical to continue conducting placebo-controlled trials in a therapy that has such documented cardioprotective benefit. (Eur Heart J 1996; 17 (Suppl D): 9-14) Key Words: Coronary artery disease (CAD), oestrogen replacement therapy, oestrogen, cardioprotection. Differences in heart disease between men and women It has long been recognised that the incidence of heart disease is significantly lower in women than in men; particularly women in the pre-mepausal. However, women at every age have less coronary artery disease (CAD) than men, even after mepause. More importantly, even when considering various risk factors, women continue to have lower rates of CAD than men (Fig. 1). This seems to be true for nearly all risk factors including elevated cholesterol levels, left ventricular hypertrophy (LVH), and hypertension. The only time women lose their natural female protection against CAD is in the presence of diabetes, which equalises the risk of CAD mortality between men and women. That is, women with diabetes die like men with diabetes. A second major difference between men and women in their development of heart disease is the age of onset. On average, in the U.S., women develop heart disease 7 to 10 years later than men. 0195-668X/96/0D00O9+O6 $18.00/0 %) 1996 The European Society of Cardiology

10 T. L. Bush o O 30-35- 40-45- 50-5560Age group 65-70- 75-80- 85+ Figure 1 Comparison of death rates in the United States due to cardiovascular disease for men ( ) and women ( ), by age group191. The third distinction between CAD in men and 100.0 women is the type of atherosclerotic lesions detected. Data from a study by Roberts et al. showed that women tend to develop plaque that contains significantly more cellular fibrous tissue compared to men, whose plaque contains more dense fibrous tissue (Fig. 2)[41. The clinical significance of this finding is t yet kwn; however, it has clear implications for possible differentiation in the treatment of atherosclerosis in men compared to women. A fourth and most interesting difference between the two genders is that in women, high-density lipoprotein (HDL) cholesterol seems to be a more potent predictor of major cardiovascular events than lowdensity lipoprotein (LDL), or total cholesterol. Data from the Lipid Research Clinics Program (LRC) showed that, of 2000+ women studied, those who had total cholesterol levels over 240 mg. dl ~', but who had HDL levels over 5 0 m g. d l ~ ', had increased risk of CAD151. Likewise, women with total cholesterol levels Women Men less than 200 mg. dl ~ ', but whose HDL levels were less Native coronary arteries than 50 mg. dl ~', had an increased risk of disease. We of atherochose an HDL level of 50 as the cutoff point in this Figure 2 Differences in type and composition sclerotic plaques in men and women'41. = d e n s e fibrous analysis, because earlier analyses in the Framingham tissue; = cellular fibrous tissue; =calcific deposits; Heart Study and the Lipid Research Clinics Program 3 = pultaceous debris; O=foam cells, lymphocytes. found that an HDL of <50 mg. dl ~~' was as predictive of heart disease for women as an HDL of <35 mg. dl ~ ' was for men16'71. triglyceride levels also supports the finding that HDL is 51 When the LRC participants were stratified by the primary predictor of heart disease in women'. A last, and possibly most important difference LDL level, going from <130, 130-159, and 160 or greater, it was apparent that, irrespective of LDL levels, between men and women is hormonal. The impact of the higher HDL levels continued to exert a cardioprotective presence of endogeus or exogeus oestrogen on the effect on women. Finally, analysis of the data for risks of heart disease has recently given way to what I

Prevention of CAD in women 11 Pneumonia/influenza ~~\ Diabetes V\ 0 50 100 150 200 250 300 350 Death rate per 100 000 per year Figure 3 Leading causes of death in women in the United States (1979)1'1. has been called the oestrogen/coronary heart disease hypothesis. Many observational studies have already shown that post-mepausal women who take oestrogen replacement therapy have a markedly lower risk of coronary heart disease than women who do t[8]. Multivariate analyses indicate that this protection cant be explained by selection bias, but that oestrogen itself prevents heart disease indirectly by modifying certain risk factors, and probably directly, as well, through its effects on vessel walls. Major causes of death in women The leading cause of death in women in Western countries including the United States is heart disease (Fig. 3). Data collected by the U.S. government National Center for Health Statistics clearly show that heart disease and cerebrovascular disease each outrank all other major causes of death, and that heart disease carries a significantly higher mortality among women than all cancers combined191. In real terms, this means that of 2000 healthy women over the age of 50, each year 20 will develop clinical coronary disease, and 12 of those will die from this condition. Of those same 2000 women, 11 will develop osteoporotic fracture, six will develop breast cancer, and three will develop endometrial or uterine cancer. The number of deaths per year from these three diseases are: one from fracture, two from breast cancer, and one from endometrial cancer, respectively. These data demonstrate a clear impact on a general medical practice; because heart disease in women is likely to cause three times as many deaths as the other causes mentioned, great attention needs to be given to the prevention of CAD. While women do show a lower rate of heart disease than men of the same age, the progsis for women with diagsed CAD is worse than for men, with significantly greater mortality. The reasons for this are t yet understood, although it is hypothesized that the later age at which women are diagsed with CAD may partially represent a delay in treatment; further, women have smaller hearts than men, and small size may be a risk factor. However, even taking these facts into account, there still appears to be a higher mortality for women than men. Prevalence of CAD An unappreciated fact about heart disease is that a large number of women around the world are already living with clinical heart disease. Among younger women between the ages of 45 and 64, the estimated prevalence in the U.S. is one in seven, while among women 65 and over, the prevalence increases to one in three. Only 50% of heart disease is predictable by kwn risk factors; therefore, approximately half of all women who go on to develop CAD have identifiable risk factors at the time of diagsis. Such statistics indicate the necessity for changes in the way we treat heart disease in women. A preventive approach based solely on the presence of risk factors is inappropriate; instead, a more global use of prophylactic lifestyle changes, and perhaps pharmacological therapies to prevent the development of heart disease in all women may be needed. Pharmacological treatment of high blood pressure and hyperlipidaemia is already indicated. Oestrogen replacement therapy, which has significant protective effects against heart disease, is a candidate for more universal use in post-mepausal women. The protective effects of oestrogen The argument against the cardioprotective effects of oestrogen continues despite a large and consistent body of evidence supporting such a link'5'7'101. One of the most convincing pieces of evidence lies in the fact that the lower incidence of heart disease in women compared with men is seen at virtually every age and across different cultures. Furthermore, this phemen is t Accidents

12 T. L. Bush Table 1 Observational cohort studies of oestrogen and cardiovascular disease First author, year Potocki, 1971 Hammond, 1979 Lafferty, 1985 Persantine trial, 1986 Wilson, 1986 Bush, 1987 Eaker, 1987 Petitti, 1987 Criqui, 1988 Boysen, 1988 van der Giezen, 1990 Henderson, 1991 Stampfer, 1991 Wolf, 1991 Falkeborn, 1992 Sturgeon, 1995 Ettinger, 1996 Country Poland Holland Holland Sweden Risk estimate 0-50 0-35 0-20 0-20 1-40 0-40 0-90 0-70 0-60 0-50 0-70 0-60 0-40 Significant? limited to humans. In a classic study by Pick and Stammler in the 1940s, roosters who were fed oestrogentreated, cholesterol-laden chicken feed developed virtually atherosclerotic plaque, while those fed a chicken feed diet with cholesterol but without oestrogen did show substantial plaque development 1 " 1. Williams et al. later showed a similar protective effect of oestrogen in monkeys 1 ' 21. Similarly, in autopsy studies of men who had been treated with oestrogen for prostatic cancer, surprisingly little atherosclerosis was found when compared to the arteries of men who had t taken oestrogen. Metabolic and mechanistic studies have also shown that oestrogen has an effect on CAD response, including vasodilation. Additionally, there is evidence that oestrogen may act as a calcium channel blocker, and as an antioxidant, mechanisms which could protect against CAD. Oral oestrogens have been found to affect insulin and glucose levels, to raise HDL by approximately 15%, and to decrease LDL by a similar 15%. The final and most impressive block of evidence in support of a cardioprotective effect of oestrogen comes from the large number of recently completed observational epidemiological studies of women. To date, more than 20 cohort studies have been completed, and with the exception of the Framingham Study, all have found that oestrogen use prevented CAD (Table 1). When the Framingham data were reanalysed by Eaker and Castelli using a different definition of risk (i.e., HDL level was t adjusted for), even the Framingham Heart Study showed a cardioprotective effect of oestrogen' 131. In addition to data on CAD, recent data from Ettinger et al. show a significant reduction in all-cause mortality in women taking oestrogen replacement; this reduction in all-cause mortality is somewhat smaller than that seen for CAD mortality 1141. The PEPI Trial Despite this large body of observational data, there remains a belief that women who take oestrogen are in some way different from those who do t, and that it is this difference and t the oestrogen causing the cardioprotective effect. The PEPI Trial, initiated in 1987, was designed to test in a randomized, double-blinded, placebo-controlled study, the impact of a variety of hormonal regimens on coronary artery disease risk factors. My colleagues and I wished to test a variety of hormonal regimens and conduct a long-term follow-up, since most of the studies before 1987 were t longer than one year. To maintain a reasonable cost and workable amount of data for the study, it was decided that PEPI would focus on four CAD risk factors only: HDL cholesterol, fibrigen levels, blood pressure, and serum insulin. Specifically, PEPI asked the question, What are the long-term effects of oestrogen and oestrogen/progestin combinations on these specific risk factors for heart disease? Study methods Eight hundred and seventy-five women were enrolled into either the placebo arm or one of four separate treatment arms: upposed oestrogen, continuous oestrogen with cyclical progestin, continuous oestrogen with continuous progestin, or oestrogen plus micronized progesterone. The treatment cycle was 28 days. The first group was given placebo; the second, conjugated equine oestrogen (CEE) 0-625 mg.d" 1 ; the third group took CEE 0-625mg. d ~' plus medroxyprogesterone acetate (MPA) 10 mg over days 1-12; the fourth group took CEE 0-625 mg.d" 1 plus MPA 2-5mg.d~'; and the fifth group was given CEE 0-625 mg.d~' plus micronized progesterone (MP) 200 mg over days 1-12. The data were collected and analysed in an intention-to-treat manner, which is a conservative methodology. In such an analysis, the subjects were analysed as if they remained on the assigned regimen throughout the trial period. Results One of the most obvious findings of the PEPI Trial was that all drug regimens significantly increased HDL cholesterol levels, with upposed oestrogen and the oestrogen/micronized progestin combination showing the greatest increases, at 5-6% and 41%, respectively. These increases were followed by CEEZMPA(cyc) at 1-6% and CEE/MPA(con) at 1-2%, while patients taking placebo actually saw a cumulative decrease in HDL of 1-2% (Table 2). All active regimens showed a reduction in LDL cholesterol. These findings all represent a significant cumulative difference from placebo (/ ) <0001), which also showed a reduction in LDL of 4-1%. In

Prevention of CAD in women 13 Table 2 Impact of various oestrogen regimens on receptors of heart disease in 2000 women in the PEPI Trial CAD Parameter Placebo CEE CEE + MPA(cyc) CEE + MPA(con) CEE+MP HDL(mg.dr') LDL(mg.dr') Fibrigen (gl) Systolic BP Diastolic BP Fasting insulin 2 h insulin Fasting glucose 2 h glucose -1-2 -41 010 + 1-2 00 3-8 -13-7 -0-5 -01 + 5-6* - 14-5* -0-2* + 0-5 -0-7 -1-7 -80-2-8* + 2-0 + 1-6* - 17-7* +006* +0-7 -10 + 1-3 + 13-4 -2-7 + 7-5* + 1-2* - 16-5* + 01* + 1-8 +0-2 -3-8 + 1-2 -21 + 6-9* +4-1* - 14-8* + 001* +0-1 -0-6 -3-5 -25-1 -2-5 + 3-0 'Significantly different from placebo; PEPI = Postmepausal Oestrogen and Progesterone Interventions; HDL=high-density lipoprotein; LDL = low-density hpoprotein; BP=blood pressure; h=hour. Regimens: CEE=conjugated equine oestrogens 0.625mg.d~'; MPA(cyc) = medroxyprogesterone acetate 10mg.l2d~'; MPA(con)=medroxyprogesterone acetate 2.5 mg. d" 1 ; MP=micronized progesterone 200mg. 12 d" 1. general, oestrogen vs placebo increased HDL by 5-6 mg. dl ~' and decreased LDL by 15 mg. dl ~'. One surprising result was the slight but significant impact seen on fibrigen levels. Fibrigen levels increase with increasing age, and, as expected, they increased in the placebo group by 010%. What was t expected was the fact that all active groups had increases in fibrigen levels. Ather finding of the PEPI Trial was that ne of the oestrogen regimens tested demonstrated any significant impact on blood pressure (either systolic or diastolic), or on fasting or 2-h insulin levels. Conclusions of PEPI The findings of the PEPI Trial have provided a substantial amount of information, t only in support of the oestrogen/coronary heart disease hypothesis, but also in terms of the quality of the impact of oestrogen on CAD risk factors in women. The PEPI Trial showed a clinically significant effect on standard risk factors for heart disease, such as LDL and HDL levels, a moderate effect on fibrigen levels, and impact on blood pressure or blood insulin levels. We interpret these last two findings as being suggestive that hypertension status or diabetic status should t be considered as contraindications for hormone therapy. In addition, women in the treatment arms of the trial showed a lower weight gain over the 3-year trial period than the women who took placebo. Women on upposed oestrogen gained the least weight, while women taking combination therapy gained intermediate amounts of weight, compared with placebo. This finding suggests that age, and t the introduction of hormone therapy, is the cause of weight gain in post-mepausal women. Current areas of study Now underway in the United States is the Heart and Oestrogen/Progestin Replacement Study (HERS), an industry-sponsored secondary prevention trial of combined oestrogen and progestin therapy in women with documented coronary artery disease. Currently, nearly 3000 women are enrolled at 18 clinical centres. The trial began in 1993, and the results are expected to become available in 1999. Discussion If the results from HERS are positive, one question that may finally be raised is whether it is feasible or even ethical to continue conducting placebo-controlled trials with a therapy that has already demonstrated benefit. In my opinion, if HERS shows a significant benefit, there would certainly be a place for major dosing studies; however, it would become ethically difficult to deny women hormone therapy in the face of the evidence of its cardioprotective effect. Although oestrogen replacement therapy has been, until recently, prescribed largely for treatment of vasomotor symptoms and other discomforts associated with mepause as well as for the prevention of osteoporosis, the public health benefit in terms of prevention of heart disease is perhaps the most significant. Controversies still exist concerning the use of oestrogen replacement therapy, the most significant of which has been the hypothesized association with breast cancer. Numerous studies over the last 50 years have looked at the association between oestrogen therapy and breast cancer risk and definitive conclusion has been reached. Recently, the Nurses' Health Study, an ongoing surveillance study, reported that oestrogen users were 1-4 times more likely than n-users to develop breast cancer 1 ' 51. In contrast, a large case-control study from Seattle which found increase in risk was published 3 weeks later 1 ' 6 '. This controversy will t be resolved soon. In interpreting the available data, it seems likely that a potential increased risk may exist, but it is likely that it is minimal or it would have been seen more

14 T. L. Bush conclusively by w. As a clinical epidemiologist, I must conclude that we have reached the limits of epidemiology when we assess the associations on the issue of oestrogen replacement and breast cancer. Further, we probably will t have any answers on this issue until we have made some major breakthrough in our understanding of the mechanisms of hormonal actions in the development of breast cancer. The Women's Health Initiative (WHI) is a U.S. government-sponsored study currently underway, which many hope will answer the hormone therapy/breast cancer association question. It appears unrealistic to expect, however, that this trial, designed to measure the effects of oestrogen on more common end-points such as heart disease and fracture, will continue for the necessary length of time (12-15 years) to assess the occurrence of breast cancer in this population. The nature of such trials is that they are terminated when the significant benefits are seen. Finally, an uncontested risk of upposed oestrogen replacement therapy is the increased incidence in endometrial cancer. However, the addition of sufficient progestin therapy to an oestrogen regimen reduces this risk to at or below that in the population of n-users. Overall, population studies conducted over 20 years strongly suggest that oestrogen replacement therapy is associated with a 50% reduction in the major causes of morbidity and mortality in women over 50. As public health practitioners, we have the opportunity to use this powerful intervention in our patients to significantly reduce the burden of CAD disease in women. References [1] The Coronary Drug Project. Initial findings leading to modifications of its research protocol. JAMA 1970; 214: 1303-13. [2] Oliver MF, Boyd GS. Influence of reduction of serum lipids on progsis of coronary heart disease: a five-year study using oestrogens. Lancet 1961; 2: 499-505. [3] McDowell F, Louis S, McDevitt E. A chemical trial of Premarin in cerebrovascular disease. J Chronic Disease 1967; 20: 679-84. [4] Mautner SL, Lin F, Mautner GC, Roberts WC. Comparison in women versus men of composition of atherosclerotic plaques in native coronary arteries and in sapheus veins used as aortocoronary conduits. J Am Coll Cardiol 1993; 21: 1312-8. [5] The Writing Group for the PEPI Trial. Effects of oestrogen or oestrogen/progestin regimens on heart disease risk factors in postmepausal women. JAMA 1995; 273: 199-208. [6] Bass KM, Newschaffer CJ, Klag MJ, Bush TL. Plasma lipoprotein levels as predictors of cardiovascular death in women. Arch Intern Med 1993; 153: 2209-16. [7] Bush TL, Barrett-Conner E, Cowan LD et at. Cardiovascular mortality and ncontraceptive use of oestrogen in women: results from the Lipid Research Clinics Program Follow-up Study. Circulation 1987; 75: 1102-9. [8] Bush TL. Noncontraceptive oestrogen use and risk of cardiovascular disease: an overview and critique of the literature. In: ICorenman SG, ed. The Mepause Sero Symposium; 1990; Nowell (MA). [9] Vital statistics of the United States, II: Mortality, Part B. Washington, DC: U.S. Department of Health, Education and Welfare; Publication: NIH-79-1102. [10] Wenger NK, SperoffL, Packard B. Cardiovascular health and disease in women. N Engl J Med 1993; 329: 247-56. [11] Pick R, Stamler J, Rodbard S, Katz LN. Oestrogen-induced regression of coronary atherosclerosis in cholesterol-fed chicks. Circulation 1952; 6: 276-80. [12] Williams JK, Adams MR, Herrington DM, Clarkson TB. Short-term administration of oestrogen and vascular responses of atherosclerotic coronaries. J Am Coll Cardiol 1992; 20: 452-7. [13] Eaker ED, Castelli WP. Differential risk for coronary heart disease among women in the Framingham Study. In: Eaker E, Packard B, Wenger N et ah, eds. Coronary heart disease in women. New York; Haymarket Doyma, 1987: 122-30. [14] Ettinger B. Overview of the efficacy of hormonal replacement therapy. Am J Obstet Gynecol 1987; 156: 1298-303. [15] Colditz GA, Hankinson SE, Hunter DJ et al. The use of oestrogens and progestins and the risk of breast cancer in postmepausal women. N Engl J Med 1995; 332: 1589-93. [16] Stanford JL, Weiss NS, Voigt LF, Dating JR, Habel LA, Rossing MA. Combined oestrogen and progestin hormone replacement therapy in relation to risk of breast cancer in middle-aged women. JAMA 1995; 274: 137^t2.