Risks and benefits of hormone replacement therapy B de Lignie`res & EA MacGregor1 1

Hoˆpital Necher, Paris, France and 1The City of London Migraine Clinic, London, UK

de Lignie`res B & MacGregor EA. Risks and benefits of hormone replacement therapy. Cephalalgia 2000; 20:164–169. London. ISSN 0333-1024 Menopause, the permanent cessation of menstruation, is due to ovarian failure, which may lead to oestrogen deficiency diseases, particularly osteoporosis, cardiovascular disease and cerebrovascular disease. Mortality and morbidity caused by these conditions can be modified by using hormone replacement therapy, but the benefits of this therapy must be weighed against the increased risk of breast cancer and the symptomatic sideeffects the treatment may cause. The combination of transdermal oestrogen and natural progesterone offers the most favourable risk-to-benefit profile. u Hormone replacement therapy, oestradiol, progestins, risks, benefits B de Lignie`res, Hoˆpital Necher, Pavillon Apert, 149 Rue de Sevres, Paris 75015, France. Received 10 February 2000, accepted 25 May 2000

Introduction The ovaries produce oestradiol and progesterone on a cyclical schedule for an average of 40 years, from puberty to menopause. These hormones have striking influences on sexual target organs, such as the uterus, breast, vagina and skin, and also on the brain, urinary tract, kidneys, liver, bones, myocardium and arterial walls. Human ovaries stop secreting hormones when women are approximately 50 years of age. As a result of prolonged life expectancy, women today live for 30 years or more with objective plasma hormonal deficiency. However, more than 50% of these postmenopausal women are able to synthesize oestradiol outside the ovaries and do not experience the full consequences of oestradiol deprivation (1–3). Extraovarian oestradiol synthesis is demonstrated in human adipose tissue and may occur in the brain, bones and arterial walls. Early symptoms of oestrogen deficiency include hot flushes, night sweats, headaches and panic attacks. These are perceived either as minor unpleasant disturbances or as unacceptable severe events, according to individual characteristics, such as culture, health care levels, and genetically determined extra-ovarian oestradiol synthesis. Oestradiol deprivation increases the prevalence of the most common bone (osteoporosis) and cardiovascular (myocardial ischaemia) diseases in the older population (4), which is a matter of great concern. For a large percentage of the post-menopausal population, oestrogen deficiency should be considered 164

an endocrine disease that has serious potential consequences and requires replacement. Since the 1960s, efforts have been made to identify and offer exogenous steroid replacement of ovarian oestradiol to women at risk of oestrogen deficiency. In recent years, an increasing number of different oestrogen formulations and doses have become available. Oestrogen-only therapy should only be given to women who have had a hysterectomy, since unopposed oestradiol stimulation (i.e. no progesterone) has harmful consequences for the uterine endometrium, i.e. the potential development of endometrial hyperplasia or carcinoma. To ensure adequate endometrial protection, different schedules associating oestrogens and progesterone or related drugs (progestins) have been developed. These different combinations of hormone replacement therapy (HRT) given at different doses, schedules and routes of administration may elicit different efficacy/ safety ratios (4–6).

What is lacking and should be replaced? Oestrogen In the reproductive years, the ovaries produce a specific oestrogenic steroid, 17b-oestradiol, on a cyclic 4-week schedule. Plasma levels of oestradiol range from 40 to 80 pg/ml during the first week of the ovarian cycle (early follicular phase) and from 80 to 300 pg/ml during the second week (mid-and late follicular phases, includ# Blackwell Science Ltd Cephalalgia, 2000, 20, 164–169

Risks and benefits of HRT ing the periovulatory peak). During the third and fourth weeks, the oestradiol level fluctuates between 100 and 150 pg/ml (early and mid-luteal phases) to 40 pg/ml a few days before menstruation (late luteal phase). The mean oestradiol level across a normal 28-day cycle is about 80 pg/ml. The levels of the main oestradiol metabolite, oestrone, are within the same range and the oestradiol:oestrone ratio remains close to unity. Oestradiol concentrations are identical in the vessels of the forearm, uterus and liver, but are five to 10 times higher in the efferent ovarian veins. Many women experience uncomfortable symptoms when oestradiol reaches extreme values during menstruation or at the ovulatory peak, but very few have discomfort when the plasma values remain at mean follicular phase levels (between 60 and 150 pg/ml). The periovulatory oestradiol peak is necessary to trigger luteinizing hormone and follicle stimulating hormone surges and induce ovulation in pre-menopausal women, but it does not have to be reproduced for full bone and artery wall protection after menopause. Even when oestradiol replacement is required for oocyte donation in cases of premature menopause, an oestradiol peak does not have to be stimulated to reach full efficacy on target organs. After the menopause, oestradiol plasma values remain consistently lower than 40 pg/ml. However, some tissues, such as adipose tissue, gain the ability to locally synthesize a significant amount of oestrone and oestradiol, with the tissue concentration 10–20 times higher than the plasma concentration. The consequences of ovarian insufficiency on some oestrogen targets, such as bone and artery wall, may be substantially attenuated in some individuals, even those who have very low oestradiol plasma levels. Full oestradiol replacement is expected to restore 17b-oestradiol plasma levels up to around 80 pg/ml during a 28-day course of treatment.

Progesterone Pre-menopausally, physiological production of progesterone increases rapidly after ovulation, and plasma values reach 3 ng/ml in the early luteal phase and 10 ng/ml or more in the mid-luteal phase. The high plasma values that are reached in the mid-luteal phase are necessary to induce secretory transformation of the endometrium and allow initiation and maintenance of pregnancy. However, the low plasma values of the early luteal phase are sufficient to suppress the oestradiol receptors within the endometrial glands, to block mitosis, and to arrest endometrial proliferation, which is the aim of progesterone substitution in post-menopausal women. Moreover, low progesterone doses are less likely to re-induce the angiogenic cycle in endome# Blackwell Science Ltd Cephalalgia, 2000, 20, 164–169

165

trium and thus less likely to re-induce bleeding. Therefore, an optimal progesterone replacement provides effective control of endometrial growth without secretory transformations and cyclic bleeding (7).

Benefits Benefits on menopausal symptoms According to numerous large clinical trials, a positive effect of oestrogen treatment on menopausal symptoms, such as hot flushes, sleep disturbances, asthenia, mood instability, vaginal dryness and urinary tract infection, can be expected. However, side-effects such as bleeding, mastodynia, bloating, mood instability and leucorrhea occur in many women who use oestrogen or oestrogen plus progestin replacement therapy. Progestins, particularly when given cyclically, may be associated with pre-menstrual-like symptoms (8). These side-effects, which severely reduce acceptability, are probably related to the large inter-individual variation in the bioavailability of both oestrogens and progesterone. HRT must be customized, not standardized, to achieve optimal quality of life benefits. On the other hand, HRT monitoring should not be based solely on control of menopausal symptoms, since oestrogen levels may not be sufficiently high to obtain the expected benefits on bone or cardiovascular disease (9–11). Post-menopausal women with severe oestrogen-dependent bone loss may not experience hot flushes or vaginal dryness, probably because of the non-uniform localized, instead of systemic, mode of oestrogen synthesis after menopause.

Benefits on bone Adequate oestrogen substitution in post-menopausal women should be an important part of preventive medicine in reducing bone loss in the short term and risk of fracture in the long term (12). Benefits are only partly related to an increase in bone density and are mostly linked to improvement in the bone structure that is not fully appreciated by current bone density measurements (13). In untreated post-menopausal women around 70 years of age, small differences in serum oestradiol levels correlate with large differences in the risk of fracture, suggesting that low dose exogenous oestrogen may be protective for a minority of women with otherwise undetectable serum oestradiol concentrations (14). However, a significant reduction of the hip-, spine- and wrist-fracture incidence of 30–70% has been observed until now only in current users of relatively high doses of HRT. A recent case-control study, based on 1327 women

166

B. de Lignieres and E.A. MacGregor

(including 120 HRT ever users) with hip fracture and 3262 randomly selected controls (including 456 HRT ever users) showed a dose-ranging effect of oestrogens in 70-year-old women. The results showed that patches delivering less than 50 mg oestradiol per day, or tablets including less than 2 mg of oestradiol or 0.625 mg of conjugated oestrogens, were unlikely to provide protection against hip fractures (15). The addition of a progestin to the oestrogen further reduced the risk of hip fracture and seemed to be more effective if it was structurally related to progesterone rather than androgen. The benefits appear to be restricted only to current users, since never and past users had similar fracture risks 5–10 years after HRT was stopped. This implies that relatively high doses of oestrogens must be used indefinitely to actually reduce the risk of fracture.

Benefits on cardiovascular risks Greater use of HRT has recently been advocated for preventing cardiovascular disease in post-menopausal women. A 50% reduction in the relative risk of almost all vascular events has been promised to all HRT users, especially to those at high risk. However, several recent papers conclude that no data are strong enough to support this claim (16–20). These predicted benefits were extrapolated from consistent experimental data, mostly using natural oestradiol and progesterone, from epidemiological human data evaluating the influence of age at menopause as an independent risk factor, and from observational studies comparing current users to past and never users of conjugated equine oestrogens (CEE) with or without synthetic progestins. Current CEE users involved in the largest observational studies were probably healthier than never users before starting the treatment, according to the prescribing practices of physicians. They are also likely to have had a lower incidence of side-effects and adverse events during treatment compared with past users. Similarly, the compliance with placebo has been shown to select the subjects with 50% less vascular morbidity and mortality than non-compliers (4, 18, 19). Some recent surveys, comparing the relatively small group of current users of CEE or other oral oestrogens to the large group of never and past users or to oestriol users, did not find any significant decrease in the risk of coronary events (16–18), but did find a significant increase in deep vein thrombosis (21), pulmonary embolism (22) and transient and ischaemic strokes (23, 24). It is now suspected that popular formulations that contain the synthetic progestins medroxyprogesterone acetate (MPA) and norethisterone (NTA) inhibit the expected benefits of oestrogens within the artery wall

(25, 26). The Heart Oestrogen-progestin Research Study (HERS), the first prospective randomized study evaluating the potential cardiovascular benefits for women with pre-established coronary heart disease, has shown that they exhibit no improvement from a treatment combining CEE and MPA, but have significantly more thromboembolic events (27). Thus the expected cardiovascular benefits have not been confirmed and it is difficult to continue recommending them as first line treatments. Do other steroids, doses, schedules and routes of administration produce better results? No long-term comparative studies are underway. The Women’s Health Intervention Study, expected to deliver results around 2008, has been designed to study only CEE and MPA. The only means presently available to predict the efficacy:safety ratio of one specific HRT for each individual is to assess the likely effects on lipids and haemostasis.

Plasma lipids The protective effect of oestradiol on cardiovascular diseases can be explained by its influence on plasma lipids. Oral oestrogens induce a parallel increase in HDL cholesterol and triglycerides, mostly through a decrease in hepatic lipase activity (5). This mechanism tends to induce an accumulation of free cholesterol, phospholipids, and triglycerides within the HDL particles by decreasing their hepatic clearance. The marked increase in triglycerides is not favourable in women over 50 years of age (4, 5). Similarly, oral oestrogens induce a fast decrease in LDL cholesterol, which is expected to be beneficial. However, because the ratio of apoprotein B to LDL cholesterol increases, creating smaller and denser particles that are theoretically more atherogenic, the actual benefit is unclear (5). Based on plasma lipid criteria, using non-oral oestradiol instead of oral CEE will prevent the triglyceride increase and the LDL particle size decrease and replace endogenous oestradiol secretion more closely. Also, switching from MPA to progesterone has prospectively been shown to significantly improve HDL-C metabolism (28).

Haemostasis After menopause, there is a spontaneous increase in some procoagulant factors, such as fibrinogen and factor VII, and in antifibrinolytic factors, such as PAI1. This reflects a tendency to an increased thrombotic risk, which HRT is expected to prevent or correct. Oral # Blackwell Science Ltd Cephalalgia, 2000, 20, 164–169

Risks and benefits of HRT oestrogen substitution decreases fibrinogen and PAI1, but, due to first pass effect, increases factor VII, fibrinopeptides, and fragment 1+2 of prothrombin and decreases anticoagulant factors, such as antithrombin III and protein S and C (29, 30). The mean tendency in oral oestrogen users is towards coagulation activation, fully or partly balanced by fibrinolysis activation, with a resultant variable increased thrombotic risk (30). Non-oral administration of oestradiol induces only favourable changes in coagulation factors, including factor VII, and in fibrinolytic factors (30, 31). However, these changes are related to oestradiol serum levels, and optimal efficacy on haemostatic balance is obtained only when oestradiol reaches a concentration similar to that of a mid-follicular phase (31). Many transdermal system users do not reach these values.

Artery intima-media thickness Carotid artery intima-media thickness changes can be measured by high-resolution ultrasound and are correlated to carotid and coronary by atherosclerotic progression. In women less than 60 years of age, oestrogen substitution does not have a significant influence on this measurement, but the intima-media thickness may become a good marker of the influence of HRT on atherosclerosis in women over 65 (32, 33). However, no study has validated this hypothesis.

Vasomotor effects Vascular benefits of HRT include re-establishing the physiological vasoactive responses to ischaemia, exercise or cold. The methodology used to measure vasodilation in the forearm after a standardized ischaemic or cold test is precise enough to differentiate optimal and insufficient circulating levels of oestradiol and optimal and suboptimal progestin effects (26, 34, 35) The vasodilating effect of oestrogen is likely to be at least partly related to nitric oxide (NO) activity. Both the vasodilation and plasma levels of NO correlate with the circulating levels of oestrogens. As during the menstrual cycle, plasma oestradiol levels less than 80 pg/ml during replacement therapy are unlikely to provide optimal efficacy on both NO concentration and vasomotor effects (34, 36, 37) Also some progestins such as MPA and NTA inhibit the vasodilating effects of oestrogens and should not be selected as first choice drugs (25, 26, 35)

Other benefits Recent studies also suggest that HRT reduces the risk # Blackwell Science Ltd Cephalalgia, 2000, 20, 164–169

167

of colon cancer (38) and Alzheimer’s dementia (39). Oestrogen use has been associated with longer survival (40), although this finding could relate to selection biases of users.

Risks Risks for endometrium In the mid 1970s, it was established that unopposed oestrogen use results in a high risk of endometrial cancer (7). Adding a progestin at least 10 days a month was shown to reduce this endometrial risk, a fact that has been consistently confirmed by subsequent studies. However, in two recent case-control studies conducted in non-hysterectomized women, the use of oestrogen, either unopposed or opposed by fewer than 10 days of progestin treatment each month, has been identified in 42% to 59% of all cases of endometrial cancer and in 71% to 74% of the cases discovered in HRT users. This unsafe use of oestrogens was also found in 47% to 65% of HRT users among controls (41, 42). Similarly, in the Nurses Health Study no more than 25% of current oestrogen users were also progestin users, whatever the dose and schedule, while 45% of them were not hysterectomized (23). It is clear that large discrepancies between theory and practice remain, despite almost 30 years of consistent warning. Even among users of opposed oestrogens, the overall relative risk of endometrial cancer is around 1.6–2.0 due to the 25% of women who use less than 10 days of progestins (40, 41). This situation is probably related to the low acceptability of the current progestins and emphasizes the need for new compounds with safer profiles (such as progesterone) (43).

Risks for breast A small (R.R.=1.35) but significant increase in the risk of breast cancer has been linked to post-menopausal oestrogen use, according to the re-analysis of 51 studies (44). How much the biases that tend to increase or decrease the apparent relative risk have influenced the results is not known (4). The excess in risk becomes measurable during the fifth year of treatment and then increases with the age of users but not with the duration of treatment. Surprisingly it disappears 5 years after stopping oestrogen. The addition of a progestin does not seem to modify the risk. However, users of opposed oestrogens were a minority in the studies re-analysed and, contrary to endometrial data, there is almost no information about the monthly duration of the progestin treatment. Whether using a short or long sequence of progestin has any influence on risk for breast cancer is

168

B. de Lignieres and E.A. MacGregor

not know. Some recent in vivo human studies have shown that progesterone counteracts the stimulating effect of oestradiol on the breast epithelial cell cycle, with a kinetic only slightly different from endometrial glands (45, 46). These data suggest that more than 10 days per month of progesterone may be protective in women treated with oestradiol but may not be in women treated predominantly with oestrone. It is not known if the risk can be predicted and the treatment individually monitored according to changes in breast density as measured by simple mammograms (47). Women who have breast cancer diagnosed while on HRT tend to have better prognosis criteria, especially when they are using oestrogens opposed by progesterone or some progestins (48–51). These observations seem to suggest some degree of protection and exclude a harmful stimulating effect of progesterone or some progestins on malignant cells. Clinical trials of different HRTs in women who have a history of breast cancer are in progress in several countries.

Conclusion In 2000, it no longer seems reasonable to estimate the benefits and risks of HRT as a single treatment. The results of the formulations that have been used for the last 30 years are disappointing. Almost 50% of users of CEE and MPA drop out after only a few months or years of treatment, and when long-term treatment is required to provide full bone protection, only a few women over 65 remain effectively treated. Probably because of the pharmacological impact of oral oestrogen on the liver, creating a pregnancy-like situation, unnecessary metabolic side-effects are induced such as increase in triglycerides, decrease in LDL particle size, and coagulation activation. Also the doses of currently used progestins inhibit the vasodilating effect of oestrogens. As a result, the expected cardiovascular benefit has not been obtained and thromboembolic risks may have been raised. Because of the poor acceptability of the progestins in current use, too many women do not use them or use them incorrectly. Consequently, the relative risk of endometrial cancer has remained high despite recommendations and warnings. The relative risk of breast cancer has also probably been raised in women who have used HRT for 5 years or more. Thus, the overall results are not good enough to justify long-term treatment for many women. However, relatively new HRT formulations that have actions closer to the physiological model of ovarian secretion have been available for several years. It is possible to increase serum oestradiol levels through nonoral administration without pharmacological distur-

bances in lipids and haemostasis. It is also possible to administer sufficient natural progesterone to protect the endometrium while inducing fewer side-effects than with the usual synthetic progestins. More favourable progestins have also been developed. New methods of prescribing HRT, customized instead of standardized, and individually monitored, must be considered. Unfortunately, the cost of such sophisticated HRT is expected to increase.

References 1 Rink JD, Simpson ER, Barnard JJ, Bulun SE. Cellular characterization of adipose tissue from body sites of women. J Clin Endocrinol Metab 1996; 81:2443–7. 2 Bayard F, Clamens S, Meggetto F, Blaes N, Delsol G, Faye JC. Estrogen synthesis, estrogen metabolism and functional estrogen receptors in rat arterial smooth muscle cells in culture. Endocrinology 1995; 136:1523–9. 3 Jakob F, Siggelkow H, Homann D, Kohrle J, Adamski J, Schutze N. Local estradiol metabolism in osteoblast- and osteoclast-like cells. J Steroid Biochem Molec Biol 1997; 61:167– 74. 4 Barrett-Connor E. Hormone replacement therapy. Br Med J 1998; 317:457–61. 5 De Lignie`res B. The case for a non plasma lipoproteins etiology of vascular risk in estrogen replacement therapy. Curr Opin Obstet Gynecol 1993;5:389–95. 6 La Rosa JC. Has HRT come of age? Lancet 1995; 345:76–7. 7 De Lignie`res B, Moyer D. Influence of sex hormones on hyperplasia/carcinoma risks. In: Lobo R, ed. Treatment of the postmenopausal woman. New York: Raven Press, 1994: 373– 83. 8 Magos AL, Brewster E, Singh R et al. The effects of norethisterone in postmenopausal women on oestrogen replacement therapy: a model for the premenstrual syndrome. Br J Obstet Gynaecol 1986; 93:1290–6. 9 De Lignie`res B. Estrogen replacement therapy must be customized. Eur Menop J 1996; 3:21–5. 10 Polo-Kantola P, Erkkola R, Helenius H, Irjala K, Polo O. When does estrogen replacement therapy improve sleep quality? Am J Obstet Gynecol 1998; 178:1002–9. 11 Vihtamaki T, Tuimala R. Can climacteric women self-adjust therapeutic estrogen doses using symptoms as markers? Maturitas 1998; 28:199–203. 12 Marcus R. New perspectives on the skeletal role of estrogen. J Clin Endocrinol Metab 1998; 83:2236–8. 13 Cauley JA, Lucas FL, Kuller LH et al. Bone mineral density and risk of breast cancer in older women. JAMA 1996; 276:1404–8. 14 Cummings SR, Browner WS, Bauer D et al. Endogenous hormones and the risk of hip and vertebral fractures among older women. New Engl J Med 1998; 339:733–8. 15 Michaelsson K, Baron JA, Farahmand BY et al. Hormone replacement therapy and risk of hip fracture: population based case-control study. Br Med J 1998; 316:1858–63. 16 Schairer C, Adami HO, Hoover R, Persson I. Cause-specific mortality in women receiving hormone replacement therapy. Epidemiology 1997; 8:59–65. 17 Hemminki E, McPherson K. Impact of postmenopausal

# Blackwell Science Ltd Cephalalgia, 2000, 20, 164–169

Risks and benefits of HRT

18

19 20 21

22

23

24

25

26

27

28

29

30

31

32

33

34

hormone therapy on cardiovascular events and cancer: pooled data from clinical trials. Br Med J 1997; 315:149–53. Sidney S, Petitti DB, Quesenberry CP. Myocardial infarction and the use of estrogen and estrogen-progestogen in postmenopausal women. Ann Intern Med 1997; 127:501–8. Petitti DB. Hormone replacement therapy and heart disease prevention. JAMA 1998; 280:650–2. Khaw KT. Hormone replacement therapy again. Br Med J 1998; 316:1842–4. Daly E, Vessey MP, Hawkins MM et al. Risk of venous thromboembolism in users of hormone replacement therapy. Lancet 1996; 348:977–80. Grodstein F, Stampfer MJ, Goldhaber SZ et al. Prospective study of exogenous hormones and risk of pulmonary embolism in women. Lancet 1996; 348:983–7. Grodstein F, Stampfer MJ, Manson JE et al. Postmenopausal estrogen and progestin use and the risk of cardiovascular disease. New Engl J Med 1996; 335:453–61. Pedersen AT, Lidegaard O, Kreiner S, Ottesen B. Hormone replacement therapy and risk of non-fatal stroke. Lancet 1997; 350:1277–83. Adams MR, Register TC, Golden DL, Wagner JD, Williams JK. Medroxyprogesterone acetate antagonizes inhibitory effects of conjugated equine estrogens on coronary artery atherosclerosis. Arterioscler Thromb Vasc Biol 1997; 17:217–21. Sorensen KE, Dorup I, Hermann AP, Mosekilde L. Combined hormone replacement therapy does not protect women against the age-related decline in endothelium-dependent vasomotor function. Circulation 1998; 97:1234–8. Hulley S, Grady D, Bush T et al. Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. JAMA 1998; 280:605–13. Writing Group for the PEPI Trial. Effects of estrogen or estrogen/progestin regimens on heart disease risk factors in postmenopausal women. JAMA 1995; 273:199–208. Caine YG, Bauer KA, Barzegar S et al. Coagulation activation following estrogen administration to postmenopausal women. Thromb Haemostas 1992; 68:392–5. Scarabin PY, Alhenc-Gelas M, Plu-Bureau G et al. Effects of oral and transdermal estrogen/progesterone regimens on blood coagulation and fibrinolysis in postmenopausal women. Arterioscler Thromb Vasc Biol 1997; 17:3071–8. Lindoff C, Peterson F, Lecander I, Martinsson G, Astedt B. Transdermal estrogen replacement therapy: beneficial effects on hemostatic risk factors for cardiovascular disease. Maturitas 1996; 24:43–50. Nabulsi AA, Folsom AR, Szklo M et al. No association of menopause and hormone replacement therapy with carotid artery intima-media thickness. Circulation 1996; 94:1857–63. Espeland MA, Applegate W, Furberg CD et al. Estrogen replacement therapy and progression of intimal-medial thickness in the carotid arteries of postmenopausal women. Am J Epidemiol 1995; 142:1011–9. Hashimoto M, Akishita M, Eto M et al. Modulation of endothelium-dependent flow-mediated dilatation of the

# Blackwell Science Ltd Cephalalgia, 2000, 20, 164–169

35

36

37

38 39

40

41

42

43

44

45

46

47

48 49

50

51

169

brachial artery by sex and menstrual cycle. Circulation 1995; 92:3431–5. Sullivan JM, Shala BA, Miller LA, Lerner JL, McBrayer JD. Progestin enhances vasoconstrictor responses in postmenopausal women receiving estrogen replacement therapy. J N Am Menopause Soc 1995; 2:193–9. Cicinelli E, Ignarro LJ, Lograno M et al. Acute effects of transdermal estradiol administration on plasma levels of nitric oxide in postmenopausal women. Fertil Steril 1997; 67:63–6. Clarkson TB, Hughes CL, Klein KP. The nonhuman primate model of the relationship between gonadal steroids and coronary heart disease. Prog Cardiovasc Dis 1995; 38:189–98. Newcomb PA, Storer BE. Postmenopausal hormone use and risk of large bowel cancer. J Natl Cancer Inst 1995; 87:1067–71. Yaffe K, Sawaya G, Lieberburg I, Grady D. Estrogen therapy in postmenopausal women: effects on cognitive function and dementia. JAMA 1998; 279:688–95. Cauley JA, Seeley DG, Browner WS. Estrogen replacement therapy and mortality among older women. Arch Intern Med 1997; 157:2181–7. Pike MC, Peters RK, Cozen W et al. Estrogen-progestin replacement therapy and endometrial cancer. J Nat Cancer Inst 1997; 89:1110–6. Shapiro JA, Weiss NS, Beresford SAA, Voigt LF. Menopausal hormone use and endometrial cancer, by tumor grade and invasion. Epidemiology 1998; 9:99–101. Grady D, Ernster VL. Hormone replacement therapy and endometrial cancer: are current regimens safe? J Nat Cancer Inst 1997; 89:1088–9. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer. Lancet 1997; 350:1047–59. Chang KJ, Lee TTY, Linares-Cruz G, De Fournier S, de Lignie`res B. Influences of percutaneous administration of estradiol and progesterone on human breast epithelial cell cycle in vivo. Fertil Steril 1995; 63:785–91. Foidart JM, Colin C, Denoo X et al. Estradiol and progesterone regulate the proliferation of human breast epithelial cells. Fertil Steril 1998; 69:963–9. Harvey JA. Do different hormone replacement therapy regimens have different mammographic effects. Radiology 1998; 207:277–8. Salmon RJ, Remvikos Y, Ansquer Y, Asselain B. HRT and breast cancer. Lancet 1995; 346:1702–3. Bonnier P, Romain S, Giacalone PL et al. Clinical and biologic prognostic factors in breast cancer diagnosed during postmenopausal hormone replacement therapy. Obstet Gynecol 1995; 85:11–7. Magnusson C, Holmberg L, Norden T, Lindgren A, Persson I. Prognostic characteristics in breast cancers after hormone replacement therapy. Breast Cancer Res Treat 1996; 38:325–34. Holli K, Isola J, Cuzick J. Hormone replacement therapy and biological aggressiveness of breast cancer. Lancet 1997; 350:1704–5.