10.1.2 Endocrinology of the menopause and hormone replacement therapy

A major endocrine function of the human ovary is the production of oestradiol, a hormone essential for the development of the secondary sex characteristics, for normal reproduction, and for the integrity of the cardiovascular, skeletal, and central nervous systems in particular. Oestradiol is a product of the granulosa cells, and hence its secretion is dependent largely on the presence of ovarian follicles. The number of those follicles falls steeply in the last 10 years or so of reproductive life (1), to approach zero at around the time of final menses (Fig. 10.1.2.1). This results in a profound decline in oestradiol production, to levels less than 10% of those observed during reproductive life. The question of whether the consequences of this decline are to be regarded as ‘natural,’ or as giving rise to a pathological state of oestrogen deficiency, is a controversial one. This chapter describes the endocrine changes which take place from the mid-reproductive years through to the postmenopausal years, and addresses the consequences of these changes and their possible prevention.

The perimenopause is defined by the World Health Organization as the phase extending from the onset of symptoms of the ensuing menopause to one year after the final menstrual period. It may be divided into two phases—the early menopause transition, characterized by menstrual cycle irregularity, cycle lengths being 7 or more days different from the regular cycles of reproductive age, and the late transition, marked by the occurrence of at least one episode of > 60 days without a menstrual bleed. The median age of onset of the transition is 45.5–47.5 years, with an average duration of four years (2). The perimenopause represents the years of transition from fertile, ovulatory cycles of the mid-reproductive years to a stable postmenopausal low oestrogen state, and is characterized by dynamic and complex endocrine physiology. The menopause is defined as the permanent cessation of menstruation resulting from the loss of ovarian follicular activity. It is designated retrospectively after 12 months of amenorrhoea, and occurs at an average age of 51 years, ranging from 35 to 58 years. The stable endocrine physiology of the postmenopause is well established, with high gonadotropins, low sex steroids, and undetectable levels of inhibin-B and Anti-mullerian hormone (AMH). These alterations in oestrogen physiology in particular induce clinical symptoms, and have long-term health implications.

There are striking differences between the pituitary and ovarian hormone levels of women of reproductive age and postmenopausal women, with a complex series of changes occurring in the intervening transition. Using the early follicular phase of the menstrual cycle as a reference period, follicle-stimulating hormone (FSH) levels postmenopausally are 10–15 times higher, luteinizing hormone levels 3–5 times higher, oestradiol levels 90% lower, and AMH, inhibin-A and inhibin-B levels more than 90% lower (mostly undetectable) (Fig. 10.1.2.2). These three phases, mid- reproductive, transitional, and postmenopausal, are discussed below.

The hormonal dynamics of the hypothalamic–pituitary–ovarian axis control reproductive physiology. An understanding of the physiology of the mid-reproductive years underpins that of the events of the perimenopause and menopause. Hypothalamic pulsatile secretion of gonadotropin-releasing hormone controls pituitary luteinizing hormone and FSH release, which regulate ovarian function. Gonadotropins are subject to predominantly negative feedback control by the sex steroids oestrogen and progesterone, with FSH also subject to negative feedback by the inhibins. Oestrogen and the peptide hormones (the inhibins and activins) are produced by ovarian follicles. Progesterone is produced by the corpus luteum following the maturation of the dominant ovarian follicle, and androgens (primarily testosterone and androstenedione) are secreted by the ovarian theca cells.

Two distinct inhibin subtypes, A and B, composed of a common α subunit and one of two β subunits, display functional, structural, and molecular differences, and are involved in paracrine regulation of the gonads and negative feedback on pituitary FSH. Inhibin levels have been inversely correlated with FSH, with FSH administration stimulating inhibin production. The physiology of the inhibins in the menstrual cycle has been documented (3). In brief, inhibin-A levels remain constant for most of the follicular phase, rise to a mid-cycle peak in parallel with oestradiol and go on to reach maximal levels during the luteal phase. Inhibin-B is maximal in the early follicular phase, in close relationship to FSH, falls in the late follicular phase, has a small mid-cycle peak, and declines to low levels in the luteal phase (Fig. 10.1.2.3).

The hormonal milieu during the perimenopause has been clarified substantially in recent years (2). Traditional concepts focused on gradually declining oestrogen, stimulating rising FSH; however, current understanding is more complex, with evidence that oestrogen as well as FSH levels rise in the perimenopause (4). Cross-sectional studies, in regularly cycling ovulating women 20–50 years of age, have noted FSH increasing and inhibin-B decreasing with age (especially in the early follicular phase). These changes are accompanied by minimal rises in luteinizing hormone, and no appreciable change or an overall slightly higher oestradiol level. Perimenopausal women also have highly variable hormonal profiles. Longitudinal data based on urine steroid profiles have highlighted two important observations: firstly, that oestrogen levels fluctuate dramatically both within and between individuals, and secondly, that hyperoestrogenism is a frequent occurrence, often coinciding with elevated FSH. These observations may explain the fluctuating symptoms observed in these women, as well as symptoms of hyperoestrogenism and high rates of dysfunctional uterine bleeding.

With the paradox of elevated oestrogen and FSH levels, a role for declining inhibin, stimulating rising FSH, was hypothesized (2). Subsequently, evidence for a reduced inhibin reserve with age was supported by studies on women undergoing in vitro fertilization (IVF), and in those undergoing gonadotropin hyperstimulation. The fall in inhibin levels correlates with physiological changes occurring in the ovary as follicle numbers (the source of inhibin production) decline exponentially. Combining autopsy and oophorectomy studies and applying mathematical analysis, a steady decline in follicle numbers from early years up to age 40 has been demonstrated; this is followed by an accelerated decline (1). This decline is reflected in changes in the circulating concentrations of AMH, a marker of follicular number and hence ovarian reserve, which is inversely related to FSH concentrations. However, AMH it is not involved in FSH feedback regulation, in contrast to inhibin-B. It has been shown recently that a fall in AMH to levels undetectable by current assays is a significant marker of the approach of menopause, on average 5 years later (5). It is hypothesized that the accelerated follicular depletion rate is due to an increased rate of atresia of primordial follicles, and not an increase in follicles reaching ovulation. This is consistent with the observed increase in anovulatory cycles as women approach the menopause (2). At this time, the primary event that stimulates this process of accelerated follicular depletion is not understood, and remains a topic of ongoing research.

Progesterone is produced primarily from the corpus luteum post ovulation, and is thus a marker of ovulatory cycles. Anovulatory cycles with low progesterone levels in the presumptive luteal phase occur with increasing frequency as menopause approaches. Three to seven per cent of cycles were noted to be anovulatory in women aged 26–40 years, but 12–15% were anovulatory between the ages of 41 and 50, possibly because of increased rates of follicle atresia. Ultimately, progesterone levels decline as the menopause approaches (2). This reduced frequency of ovulatory cycles, observed in the setting of complex endocrine changes, also results in reduced fertility. Anovulatory cycles are characterized by markedly elevated levels of FSH and luteinizing hormone (into the typical post-menopausal range) and low levels of oestradiol. The increasing frequency of anovulatory cycles close to menopause results in the decline in mean oestradiol levels, and the increase in mean FSH, which are seen in the 2 years prior to final menses and the first 1–2 years afterwards. The late perimenopausal drop in oestradiol levels is thus not a continuous decline, but rather an irregular fall, where low levels in one follicular phase of an anovulatory cycle may be followed by normal levels in an ensuing ovulatory cycle. This accounts at least in part for the unpredictability of hormone concentrations in an individual woman or between women, and hence the unreliability of hormone measurements in characterizing menopausal status during the transition (6). An additional source of hormonal variability is the occurrence, in a proportion of ovulatory cycles in the transition, of what has been termed a luteal out of phase (LOOP) event. In a LOOP event, a second wave of follicular recruitment is seen, with ovulation occurring again in the late luteal phase or at the time of next menses. This may be associated with markedly raised oestradiol levels in the perimenstrual phase (7). In summary, current understanding encompasses the exponential decline of oocyte numbers, falling ovarian inhibin and AMH production, rising FSH, falling progesterone, and fluctuating but maintained oestrogen levels until late into the perimenopause. The significant role of declining inhibin levels (primarily inhibin-B), with reduced negative feedback on the pituitary, stimulating increased FSH production and thus maintaining oestrogen production, has been appreciated only in recent years. The maintenance of oestrogen levels during the perimenopausal years would be appropriate in teleological terms to reduce undesirable long-term health outcomes including osteoporosis and atherosclerosis.

The menopause is characterized by ovarian failure with exhaustion of follicles, low levels of sex steroids and inhibins and elevated gonadotropins. The primary source of sex steroids after the menopause is the adrenal gland, with aromatization of androgens primarily adrenal androstenedione in adipose tissue mainly responsible for oestrogen production. Oestrogen levels are higher in women with increased body mass index (BMI) with resulting increases in oestrogen-dependent tumours, including ovarian and breast cancer, and lower rates of oestrogen deficiency-associated osteoporosis.

The important androgenic steroids in menopause physiology are the pre-androgen DHEAS and the androgen, testosterone. The circulating concentrations of DHEAS, which acts as a precursor of testosterone in peripheral tissues, fall progressively with increasing age from the late teens into late old age, and are not influenced by the menopause transition. Testosterone also falls by about 50% between ages 20 and 45 years (8), but the levels do not fall in relation to the perimenopause or menopause per se (9).

Assessment of testosterone status in women requires a sensitive and specific total testosterone assay, together with a measurement of sex hormone binding globulin (SHBG), a high affinity testosterone binding protein, from which the free testosterone concentration may be calculated. Testosterone levels increase with obesity, while SHBG falls. SHBG is increased by oral oestrogen administration, hence causing a fall in free testosterone, which may be a factor in the effects of oral oestrogens on sexual function.

The definitions of both the menopause and the perimenopause are based on clinical features related to symptoms and menstrual bleeding patterns. Predicting the stage of an individual woman in her reproductive life cycle is also ideally based on clinical features (7). Longitudinal, population-based hormone profiles reflect clinical changes in menstrual flow and frequency (2), yet in individuals, isolated levels of FSH, inhibin, and oestrogen are unhelpful as they fluctuate and therefore are unreliable (2, 6). Hence, isolated hormone profiles are not recommended in clinical assessment. If still desired, FSH assays from the early follicular phase have been shown to be the most reliable. A more rational approach in the perimenopausal individual is the acquisition of longitudinal symptom data on women who present with symptoms or alterations in menstrual flow patterns. Daily symptom diaries have proven very helpful, and can be a useful exercise in self-education for women as well as an assessment tool for clinicians.

Changes in the circulating concentrations of oestradiol and progesterone, particularly the former, may give rise to symptoms during the menopausal transition and after the menopause. Most emphasis has been given to symptoms of oestrogen deficiency, which may be relieved by oestrogen administration. For women with an intact uterus, a progestin is added for uterine protection from hyperplasia and malignancy, often termed EPT (oestrogen and progestin therapy), and for hysterectomized women ET (oestrogen only therapy). Both forms of treatment are included in the generic term hormone replacement therapy (HRT).

Conventionally, benefits and risks are considered for short-term administration for symptom relief—generally less than 5 years—and long-term administration of 5 years or more, either for longer term symptom relief, when required by symptom persistence, or when disease prevention is the aim. The benefits of short-term therapy include symptom relief, with consequent improvement in quality of life, improvement in bone mineral density, and probably a decrease in cardiovascular risk (Box 10.1.2.1). The risks associated with short-term therapy are minimal, the only significant increased risk being of venous thromboembolism, particularly with EPT, bearing in mind that baseline risk in women of perimenopausal age is low. The benefits of long-term therapy may include prolonged symptom relief, prolonged protection against osteoporotic fracture, and possibly prolonged cardioprotection. Potential risks include an increased risk of being diagnosed with breast cancer, an increased risk of stroke and an increased risk of cardiovascular disease (CVD) if EPT is initiated many years after menopause.

Substantial concerns regarding HRT were raised in 2002 following the publication of the Women’s Health Initiative, a randomized controlled trial of EPT (10) undertaken to determine whether EPT was truly cardioprotective in older postmenopausal women, and whether long-term EPT increased breast cancer risk. This was specifically not a trial of HRT for symptom relief, the average age of the participants being 63 years. The relevant results from that trial are considered further in the following sections.

The clinical features and management of the perimenopausal woman offer some unique challenges. Women in the perimenopausal years are more likely to seek medical consultation than their pre- or postmenopausal counterparts (11), with a marked increase in most menopausal symptoms. Symptoms may reflect either oestrogen excess (breast tenderness, menorrhagia, migraine, nausea, shorter cycle length. and a shorter follicular phase) or oestrogen deficiency. Interestingly, symptoms such as vasomotor disturbances and migraines demonstrate the most instability during these years, probably reflecting fluctuating hormone profiles. Dysfunctional uterine bleeding (DUB), with persistent elevation of unopposed oestrogens, occurs most frequently in perimenopausal women, who have the greatest maximal thickness of endometrium and have the highest incidence of hysterectomy. A similar hormone profile and endometrial pattern are seen in anovulatory dysfunctional bleeding in adolescence. Menorrhagia occurs in 20% of perimenopausal women compared to 9% of women in other phases of reproductive life, with differences in uterine vessel structure documented in perimenopausal women. Genitourinary symptoms are less prominent in perimenopausal women whose mean oestrogen levels are maintained.

The menopause involves both a psychosocial and physical transition which is very variable both between different cultures and also between individuals. Most of the symptoms reported by women around the time of the menopause fall into three groups: vasomotor phenomena, such as hot flushes, night sweats, and palpitations; symptoms of genitourinary atrophy such as vaginal dryness, dyspareunia, dysuria, and urinary frequency; and psychological symptoms, such as anxiety, impaired concentration and memory, loss of confidence, and depressed mood. Other symptoms include insomnia, increased sleep apnoea, breast discomfort, sensory disturbances such as formication (the sensation of ants crawling under the skin), joint pain and stiffness, and changes in libido. Most of these symptoms are attributable to oestrogen deficiency, although the mechanism of some, such as vasomotor symptoms, is poorly understood. The combination of symptoms can impact significantly on the general sense of wellbeing experienced by many women at this time of life. About 20–25% of women are severely affected by symptoms, while in 20–25% the menopause has no symptomatic impact.

Genitourinary symptoms resulting from low oestrogen levels contribute to problems of vaginal dryness, dyspareunia, urinary frequency, and an increased susceptibility to recurrent urinary tract infections, which become increasingly common with ageing. Oestrogen via either systemic or local administration can provide relief to women faced with these problems. Genitourinary atrophy can also contribute to low libido, which is no doubt multifactorial, ranging from physical changes in vaginal lubrication to changes in neurotransmitters, motivational-affective, and relationship factors.

CVD is uncommon in premenopausal women, but its incidence increases exponentially with age, and it is the leading cause of death among women worldwide. Whether the menopause is a significant aetiological factor in the occurrence of heart disease is controversial. The effect of HRT on the cardiovascular system in humans is informed by positive observational studies and extensive interventional animal data. There are now several large randomized trials studying the effect of HRT on CVD. These suggest overall neutral or adverse effects on CVD. However, all are based on the use of HRT many years after the menopause; there is growing evidence that early use of HRT from the time of menopause onwards may be preventative for CVD, generating the ‘critical window hypothesis’, but more research is needed (12, 13).

Animal studies have highlighted the complexity of HRT effects on the cardiovascular system. Factors contributing to this complexity include sex steroid receptor subtypes, variable tissue distributions, genomic and nongenomic actions, and variable organ effects regulated by coactivators and corepressors (14). The interpretation of oestrogen’s effects is further complicated by the diverse variety of oestrogen preparations available, the routes of administration, doses, combinations with different progestins, age of first use, duration of use, and many other factors, making interpretation of the literature on CVD and HT challenging (12).

Despite the complexities, the effects of HRT on atherosclerosis were best appreciated following work on oophorectomized monkeys (15). In several studies, these monkeys were randomized to placebo, oral, or transdermal oestrogen alone or combined HRT, with continuous or cyclic progestin. They were fed an atherogenic diet for two years, and at necropsy a comprehensive assessment of coronary atherosclerosis was undertaken. Oestrogen alone reduced atherosclerotic plaque by 50% (transdermal therapy) or 70% (oral therapy) compared to those on placebo. The effects of continuous medroxyprogesterone acetate (MPA) were to negate the beneficial changes seen with oestrogen. Natural progestin or cyclic MPA did not have these effects. Invasive vascular reactivity studies have also demonstrated that atherosclerotic arteries exhibit an abnormal constriction response to acetylcholine administration, which is reversed by oestrogen addition in the monkey model. Animal studies provided considerable insight into the mechanisms of the effects of sex steroids on the cardiovascular system (Box 10.1.2.2). More recent animal studies show that oestrogen inhibited progression of atherosclerosis in rabbits, but only in the early stages and not when commenced after established atherosclerotic disease (16). The antiatherosclerotic effects appear to be mediated by the endothelium, with oestrogen apparently most effective with an intact healthy or early postmenopausal endothelium, supporting the ‘critical window’ hypothesis.

A large number of cohort and case–control studies have compared the risk of myocardial infarction, related events such as bypass grafting or angioplasty, and death from CVD in users and non-users of HRT. The studies vary considerably in their endpoints and design, and in the methods used to eliminate the effects of confounding variables. Nevertheless, most studies have found a reduced risk of CVD in HRT users compared to non-users, and several studies, including the Nurses’ Health Study, report a reduction in overall mortality in women who take HRT (17). In two meta-analyses (18, 19), the relative risk (RR) of coronary artery disease in women who had ever used HRT compared with those who had not was estimated at 0.65 (95% CI 0.59 to 0.71) and 0.64 (95% CI 0.59 to 0.68). In women currently taking HRT, the RR was estimated at 0.5 (95% CI 0.45 to 0.59).

These studies suggest that the use of HRT may reduce a woman’s risk of coronary artery disease by as much as 50%. There are, however, many caveats. Almost all these studies compared women who had elected to take HRT with women who had elected not to. There is considerable evidence that, even in socioeconomically homogeneous populations, these two groups of women differ in education, exercise, blood pressure, cholesterol, and participation in preventive health measures (20). Thus, HRT users may be at lower risk of CVD than nonusers independently of HRT (the ‘healthy user effect’). Adjustment for known confounding variables has relatively little effect on the estimated relative risk of coronary artery disease, however, and it seems unlikely that the healthy user effect would account for all of the 35–50% reduction in coronary risk.

The initial randomized controlled trial on the effects of HRT on vascular disease, the heart and oestrogen/progestin replacement study (HERS), was reported in 1998 (21). This was a well designed, double-blind, placebo-controlled, randomized study of combined oral HRT use in 2763 postmenopausal women for the secondary prevention of CVD. Mean age was 66.7 years and all participants had pre-existing CVD. HERS failed to demonstrate any differences in CVD outcomes, including myocardial infarction, coronary revascularization, unstable angina, congestive cardiac failure, stroke or transient ischaemic attack (TIA), or peripheral arterial disease, between the placebo and active treatment groups, despite an improvement in lipids with HRT (21). In the context of previously published literature, these results were unexpected. Interestingly, clinically significant effects of HRT on the haemostatic system were confirmed, with an increase in venous thrombosis similar to that seen with the oral contraceptive pill. The HRT group had increased CVD in the first year (RR of 1.52), falling over time to a relative risk of 0.67 by year four. HRT is possibly a double-edged sword, therefore, with prothrombotic effects that negate any potential atherosclerotic benefits in women with pre-existing plaques. These are prone to rupture, leading to arterial thrombosis and CVD events. The limitations of the HERS study included the older age for HRT commencement and the progestin used. The recommendations following the HERS study were that women with pre-existing CVD should not commence HRT; however if they are already on HRT they should not cease.

This was followed by one of the largest randomized controlled trials in women’s health, the Women’s Health Initiative (WHI) (10). The WHI was established in 1991 by the National Institutes of Health to address the most common causes of death, disability, and impaired quality of life in postmenopausal women. The clinical trial arms in the WHI were designed to test the effects of postmenopausal HRT, dietary modification, and calcium or vitamin D supplements on heart disease, fractures, and breast or colorectal cancer. It was not designed to look at HRT effects in symptomatic women around menopause. Data from WHI, on the effects of combined oral HT (n=16 608), showed that women receiving 0.625 mg of conjugated equine oestrogens (CEE) daily with the addition of 2.5 mg MPA had higher rates of CVD, cerebrovascular disease, and venous thrombosis, compared to the placebo group (10). Absolute excess risks per 10 000 person-years attributable to oestrogen plus progestin were 7 more coronary heart disease (CHD) events, 8 more strokes, and 8 more pulmonary embolisms (PEs). The risk of CHD was especially elevated, with a hazard ratio (HR) of 1.29 (95% CI 1.02 to 1.63). In the ‘CEE only’ arm of the WHI trial in hysterectomized women, (n=10 739), the HR for CVD in CEE treated women (adjusted for age and prior disease) was 0.91 (95% CI 0.75 to 1.12), suggesting that MPA combined with oestrogen increases the CVD risk in HRT treated women (22). More recently, WHI study data analysis has suggested that women under the age of 60 years did not have an increased risk of CVD, further supporting the ‘critical window’ hypothesis (23) (Fig. 10.1.2.4). These data suggest that younger, healthy women can be started on HRT with the reassurance that there is no good evidence of increased CVD risk (19, 24). This is also consistent with observational data suggesting that women who took HRT from menopause were protected from CVD, yet those who commenced HRT years after menopause in the WHI had increased CVD (23).

With respect to cerebrovascular diseases, the HR for ischaemic stroke was 1.44 (95% CI 1.09 to 1.90) in the WHI study with combined HRT (10). With oestrogen alone, there were 168 strokes in the CEE group and 127 in the placebo group. For all strokes, the intention-to-treat HR for CEE versus placebo was 1.37 (95% CI 1.09 to 1.73); the HR was 1.55 (95% CI 1.19 to 2.01) for ischaemic stroke alone (22). Data on both HRT preparations used in the WHI study suggest an excess risk of ischaemic stroke. This was apparent across all categories of baseline stroke risk, including younger and more recently menopausal women, and in women with prior or current use of statins or aspirin. However, a more recent analysis (23) of the stroke data indicated that there was no significant increase in risk in women aged 50–59 (HR 1.13, 95% CI 0.73 to 1.76).

Venous thromboembolism (VTE) (25, Box 10.1.2.3) is an important factor in assessing the benefit-to-risk profile of HRT (Box 10.1.2.3). In WHI, subjects taking combined oral HRT showed a twofold higher rate of VTE, compared to those on placebo (HR 2.06, 95% CI 1.57 to 2.70) (10). In the CEE-only arm of WHI, the risk of VTE was increased by a factor of 1.3 (HR 1.32, 95% CI 0.99 to 1.75) (22). Comparison of the VTE risk in the CEE-only arm with that of the CEE+MPA arm of the WHI trial showed that the risk was significantly higher for the latter (P=0.03), even after adjusting for other risk factors.

Importantly, the risk of VTE with HRT is multiplicative with existing VTE risk factors. The attributable incidence of VTE with oral combined HRT is greatest in women who are obese, aged over 60, or who have thrombophilias, including the factor V Leiden variant. The risk of VTE in women taking oral oestrogen-only preparations is minimal (8 additional cases per 10 000 person-years compared to placebo), and does not appear to be exacerbated by increased BMI, age, or other VTE risk factors. A case-control study of VTE in women taking oestrogen provided the first evidence for a differential association of oral and transdermal oestrogen with VTE risk. The study evaluated 271 consecutive cases with a first documented episode of idiopathic VTE, and 610 matched controls. After adjustment for potential confounders, odds ratios (ORs) for VTE in current users of oral and transdermal oestrogen compared with nonusers were 4.2 (95% CI 1.5 to 11.6) and 0.9 (95% CI 0.4 to 2.1), respectively. Tibolone, another HRT preparation, may also have benefits over CEE; however, for both transdermal therapy and tibolone, further randomized trials are needed.

Osteoporosis is a significant cause of morbidity and mortality in postmenopausal women. At age 50, a woman has a 60% lifetime risk of sustaining an osteoporotic fracture, and a 16% risk of hip fracture (26). These risks are partly attributable to the accelerated bone loss that occurs after the menopause as a result of oestrogen deficiency. The use of HRT to modulate bone metabolism stemmed from the principal role of prevention of menopausal symptoms.

Numerous observational studies suggest that users of HRT have a reduced risk of osteoporotic fracture. The RR of hip fracture in women who had taken HRT at some point was estimated at 0.75 (95% CI 0.68 to 0.84) compared with those who had never used it (17). The protective effect in current users of HRT was greater, with a reduction in the risk of hip fracture of about 50% (27). HRT users also have a reduced risk of forearm fractures (26). More recently, the WHI randomized controlled trials have confirmed HRT’s benefits on bone, showing a reduction in hip fractures with combined HRT with an HR of 0.66 (95% CI 0.45 to 0.98) (10), and a reduction with oestrogen alone (HR 0.65, 95% CI 0.45 to 0.94) (22).

Several studies have tried to establish the minimum dose of oestrogen required to prevent bone loss in postmenopausal women, and CEE 0.625 mg/day or equivalent is usually recommended (Box 10.1.2.4). There is, however, good evidence that smaller oestrogen doses (CEE 0.3 mg/day or equivalent) still have a bone-sparing effect, at least on the spine, when combined with generous calcium supplementation (28). Lower oestrogen doses are therefore appropriate in women who cannot tolerate conventional bone-sparing doses. In addition, there is evidence that progestogens, particularly C-19 derivatives such as norethisterone, augment the effects of lower oestrogen doses on bone density. However, a recent extensive review of controlled trial data noted no differential effects between norethisterone and other progestins (29).

It is important to note that if HRT is stopped, bone loss resumes. Bone density in older women who took HRT in their early postmenopausal years and then stopped is only slightly higher than in women who never took HRT (30). The median age of hip fracture is 79 years, and women who start HRT at the time of menopause and continue for only 10 years or so may gain little or no protection against fracture in their late seventies and eighties. For the most effective protection against fracture, therefore, HRT may be most appropriate soon after menopause, in high risk women, transitioning to other agents after 5 years or so as women age. Recently available fracture risk calculators, including the WHO FRAX calculator (http://www.shef.ac.uk/FRAX), provide guidance on overall fracture risk and on when to intervene with therapy. For more details on HRT and osteoporotic fracture prevention see ‘medical management of the postmenopausal woman’, below.

Perhaps the most controversial risk of postmenopausal hormone therapy is a possible increased incidence of breast cancer in long-term users. The most comprehensive of the analyses of this risk from observational studies was that published by the Collaborative Group on Hormonal Factors in Breast Cancer (31). The Group reanalysed data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer, from 21 countries. The major analysis was of 53 865 postmenopausal women with a known age of menopause, of whom one-third had used hormone therapy at some time. Among current users of hormone therapy or those who had ceased within one to four years, the RR of having breast cancer diagnosed increased by a factor of 1.023 for each year of use, with the risk being increased 35% for women who had used oestrogen for five years or longer. The increase was in fact comparable with the effect on breast cancer incidence of a delay in menopause; in non-prior users of HRT, RR of breast cancer increased by a factor of 1.028 for each later year of menopause occurrence. The excess risk of breast cancer disappeared within five years after cessation of hormone use. The relatively increased risk was associated with long durations of use and was greater for women of lower body weight. Cancers diagnosed in hormone users were less advanced clinically than those in non-prior users. In terms of quantitative effects, the authors estimated that approximately 45 women per 1000 in North America and Europe would develop breast cancer between the ages of 50 and 70. If hormone usage was for 5, 10, or 15 years, the cumulative excess numbers of breast cancers diagnosed were calculated to be 2, 6, and 12. It must be emphasized that this large database consisted of epidemiological studies either with a case control or cohort design. One additional large cohort study has been published since that analysis, the Million Women Study in the UK (32). Although it was reported that an increase in breast cancer risk was seen within a year of enrolment, it must be noted that the women on EPT in that study had been on their therapy for an average of about 6 years prior to study entry.

More rigorous evidence has come from the EPT and ET randomized controlled trials conducted within the US WHI (10). For the 74% of participants in the EPT trial who had not previously used EPT, there was no increase in breast cancer risk after an average of 5.6 years of follow-up (adjusted HR 1.02; 95% CI 0.77 to 1.36) (Fig. 10.1.2.5). For prior EPT users, the true increase in risk is difficult to assess, as prior users assigned to placebo had a decreased annualized breast cancer incidence rate of 0.25% per year, compared to the rate in non-prior users of 0.36% per year (33) (Fig. 10.1.2.5). Prior users had variable periods of EPT use prior to entry into the trial, so that a true estimate of risk in relation to exposure was not possible. From a practical standpoint, it can be concluded that for a population of older postmenopausal women similar to those enrolled in WHI, no increase in breast cancer risk is seen in the initial 5–6 years of therapy. Many publications since the WHI report have quoted a figure of an increase in breast cancer risk of 8 per 10 000 women per year, a misleading figure based on the whole population of prior and non-prior users. Because the effect on breast cancer risk of EPT may be modified based on the individual subject’s BMI, the true increase in risk over the initial 5 years of therapy is not known with certainty. It has also been suggested that breast cancer risk is increased the sooner after menopause that HRT is initiated, whereas there may be no short-term increase in risk when HRT is initiated some years after final menses (34). It is also important to note that the WHI trial examined only one regimen, CEE 0.625 mg with MPA 2.5 mg daily. Observational data from France suggests that the nature of the progestin given in EPT may be important from the standpoint of breast cancer risk. No significant increase in risk over a period of 8 years was seen in women taking oestradiol plus progesterone or dydrogesterone, whereas increases were seen with several other progestins (35). Furthermore, randomized controlled trial data for the drug tibolone indicates a reduction in breast cancer risk in subjects over 60 years of age treated with 1.25 mg daily (36).

ET was not associated with an increased breast cancer risk in the WHI trial (37). In fact the breast cancer rate in the ET arm was decreased in comparison with the placebo arm, not quite reaching statistical significance, though the decrease was significant in non-prior hormone users and in those who were adherent to treatment. Thus for a hysterectomized woman starting ET, she can be reassured about the breast safety of the treatment in the first 5 or more years.

Particularly vexing is the question of treating oestrogen deficiency symptoms in women living with a diagnosis of breast cancer. This again has been the subject of intense controversy, with a diagnosis of breast cancer generally being regarded as a contraindication to hormone therapy (38). Nevertheless, a number of small studies have suggested that concern about such risks may be exaggerated. Present practice would suggest that oestrogens should not generally be given to women with a diagnosis of breast cancer, but that they could be considered if other therapies were unhelpful and her quality of life as a result of oestrogen deficiency symptoms is so poor that she is willing to accept the possible increase in the risk of breast cancer spread or recurrence. It must be emphasized that the evidence for that risk remains to be proven (39).

The long-term use of unopposed oestrogen in postmenopausal women with an intact uterus results in at least a fourfold increase in the risk of endometrial cancer (17). Several studies have shown that the use of progestogen for 10 days or more per month negates this increased risk. A recent large case–control study reported, however, that long-term use of oestrogen and sequential progestogen treatment still carries an increased risk of endometrial cancer (40). The RR in this study was 2.5 (95% CI 1.1 to 5.5), for women taking oestrogen for five or more years with a progestogen for 10 days or more per month. Thus, it is possible that the protective effect of cyclical progestogen on the endometrium of women taking long-term oestrogen is incomplete. In contrast, no increased risk has been observed in women taking long-term combined continuous treatment.

Several observational studies have suggested that HRT users have a reduced risk of death from colon cancer. Such an effect is biologically plausible, since oestrogens inhibit the growth of colonic cancer cells in vitro, and oestrogens and progestogens alter bile acid production. A significant reduction in risk was seen in the EPT arm of WHI (10), while there was no effect for women on ET, thus providing randomized trial evidence for the benefit of EPT on colon cancer risk.

Recent years have seen a dramatic increase in public interest in dementia of the Alzheimer’s type (DAT). Late-onset DAT, with dementia symptoms appearing after age 65, is much more common than early-onset illness, and its prevalence doubles about every 4.5 years. About 1.5 to 3 times as many women have DAT as do men. There is currently a great deal of interest in the possibility that exogenous oestrogens may be protective against the risk of developing DAT, and may also be of therapeutic benefit in patients with early disease. The biological basis for a possible beneficial effect of oestrogens is beyond the scope of this description. Four epidemiological studies have been reported in which information on postmenopausal oestrogen use was collected prospectively before the presumed onset of symptoms of dementia. In three of these, a reduction in risk of developing DAT was reported, varying between 30% and 60%. In the fourth study no significant benefit of oestrogen was reported overall; however, when oral oestrogen users were analysed as a subset, a 30% reduction was again reported. As with the other risks and benefits of long-term hormone therapy, current data are based entirely on epidemiological studies and not on randomized, prospective, controlled trials. The WHI did include a sub-study, entitled WHIMS, in which a group of women aged over 65years were shown not to benefit in terms of cognitive function from either EPT or ET, with an increased risk of dementia being observed in association with treatment (41). This study provides no evidence regarding the effects of EPT or ET started around the time of menopause and continued long term.

Many women present to medical practitioners around the time of the menopause, a convenient time for health intervention measures. Perimenopausal women should be offered advice on general behaviours, which may improve health and prolong life, such as the benefits of exercise and healthy diet. Routine preventive practices, such as blood pressure measurement, cervical smear, mammography, assessment of vitamin D status (serum 25-hydroxy vitamin D), and plasma cholesterol measurement should also be offered. Recommendations for calcium intake after the menopause range from 1000 to 1500 mg/day. As a minimum, calcium supplements should be advised when the daily intake is less than 1000 mg/day. The likely benefits and risks of HRT (Box 10.1.2.1) should be discussed, as well as the uncertainty surrounding them.

General advice centred on education, support and advice on lifestyle and dietary issues is essential in the approach to management of the peri- as well as the postmenopausal woman. There are specific pharmacological considerations in the perimenopausal woman (24). When needed, pharmacological alternatives in this group include conventional hormone replacement therapy, the low dose oral contraceptive pill, or cyclic progestin or progesterone therapy alone. In light of generally maintained oestrogens until late in the perimenopausal period, replacement of cyclic progesterone alone is useful in the management of irregular menses and menorrhagia. It also reduces hot flushes, and is often well-tolerated in this age group.

The use of conventional HRT to induce cycle regularity, reduce hormone fluctuations and reduce symptoms is more difficult than in the predictable oestrogen deficiency state of the postmenopausal woman. Whilst HRT may be appropriate in the later stages of the transition, with hormone profiles more akin to those of postmenopausal women, women in the earlier stages of the transition with more erratic hormone profiles are likely to encounter difficulties. Low dose oral contraceptive pills provide a viable alternative, as they offer the advantage of suppressing hypothalamic–pituitary function, thereby reducing turbulent endogenous hormonal activity. Low dose preparations (20 mcg of ethinyl oestradiol or vaginal/transdermal preparations) appear safe in non-smoking, non- hypertensive women of healthy weight up to the age of 50 years. The notable increased risk of VTE’s and cerebrovascular accidents is primarily expressed in smokers, hypertensive women, and women who are obese. The issue of when to change from oral contraception to HRT is difficult. A practical rule of thumb is that oral contraception can be employed in suitable women up to the age of 50–51; this can then be withdrawn temporarily, with intervening barrier contraception. If amenorrhoea occurs for longer than three months, particularly in the presence of oestrogen deficiency symptoms it is probably reasonable to change to HRT with continuous oestrogen and cyclic progestogen, though it must be explained that HRT does not have contraceptive efficacy.

Currently, many women take HRT for a few years after the menopause. Short term systemic oestrogen therapy for two to three years is a well-established and effective treatment for oestrogen-deficiency symptoms, giving good relief in most cases. Although psychological symptoms also improve, significant depressive symptoms may require specific antidepressant treatment. Women with primarily genito-urinary symptoms may be treated effectively with vaginal oestrogens alone.

Short term HRT alleviates menopausal symptoms, however it probably has little impact on the rates of CVD or osteoporotic fracture until some 15–30 years later. The beneficial effects of oestrogen on fracture rates and CVD would theoretically be maximized if all women started HRT at the menopause and continued lifelong. This cannot be recommended, however, until the effects of long-term oestrogen and combined oestrogen–progestogen on CVD and breast cancer risk have been clarified, and, in any case, would be unacceptable to many women.

Advice to women considering HRT therefore needs to be individualized (24). The major factors to take into account are the presence or absence, and the severity, of oestrogen deficiency symptoms, risk factors for venous thrombosis (overweight, smoking, and age), presence of pre-existing vascular disease, risk factors for CVD (especially family history of premature CVD, dyslipidaemia, hypertension, smoking, and diabetes mellitus), risk factors for osteoporosis (including, where appropriate, bone density measurement), and risk factors for breast cancer (particularly family history, alcohol intake, and obesity).

Women with a history of osteoporotic fracture or with proven osteoporosis by bone mineral density (BMD) measurement (BMD 2.5 SD or more below the young adult mean), and an elevated overall fracture risk (see osteoporosis section above), would have their risk of further fracture reduced by an initial period of HRT, before going on to other therapies after 5 years or so. Women with low BMD (T scores between −1 and −2.5) would also benefit from starting HRT at the time of the menopause, and continuing treatment for at least 5 years before switching to another therapy, e.g. a selective oestrogen receptor modulating agent (SERM), and this should be discussed. For the large majority of women with normal BMD or relatively low fracture risk at menopause, osteoporotic fracture may effectively be prevented by simple measures including adequate calcium, Vitamin D, and if appropriate, periodic measurement of bone density. Medical therapy, of which HRT is an option (with risks and benefits), may be initiated when fracture risk is sufficiently elevated.

Women presenting with low BMD in their sixties and early seventies are a group in whom low dose parenteral oestradiol treatment may be considered. This approach might maximize the cardiac and skeletal benefits of HRT while reducing the risk of adverse effects on breast and endometrium; however this warrants further study before it can be recommended (42).

Women with established CVD should not be commenced on oral HRT pending further data (10). Those with CVD risks should be advised that the definitive data is still pending and should be available within the next five years from controlled primary prevention studies. In those at risk of CVD, transdermal oestradiol, which lacks prothrombotic effects, may be the best option if HRT is considered.

A family history of breast cancer affecting first degree relatives is not a contraindication to HRT, but should be weighed up carefully in the assessment of the risks and benefits for the individual woman. A personal history of breast cancer is, however, generally considered a contraindication (38).

A history of deep vein thrombosis or pulmonary embolism in the absence of predisposing factors such as age of over 60 years, surgery, immobilization, obesity, or a family history of thromboembolism, suggests an underlying predisposition to thromboembolic disease. In women with such a history, thromobophilic states such as activated protein C resistance (factor V Leiden defect) and prothrombin mutations should be excluded. A history of venous thromboembolism attributable to oral contraceptives is a relative contraindication to HRT. Age, smoking and increased BMI are also potent risk factors. If oestrogen is prescribed for women with risk factors for venous thromboembolism, either tibolone or transdermal estrogen preparations are preferable, as they do not alter circulating concentrations of haemostatic factors, and have been shown not to increase VTE risk in studies (43) to date, although a definitive randomized controlled trial in this area is lacking.

Women who have had a hysterectomy require oestrogen replacement only. Although progestogens may give some additional benefits on bone density, they are not widely prescribed for hysterectomized women other than in those with a history of endometriosis. Women with an intact uterus require progestogen therapy in addition to oestrogen in order to prevent endometrial hyperplasia and carcinoma. The usual doses of commonly used oestrogen and progestogen preparations are shown in Box 10.1.2.5; several combined preparations are also available.

In late peri- and early postmenopausal women, a regimen of continuous oestrogen and cyclical progestogen for at least 10 days per month is widely used. A typical US regimen is CEE 0.3–0.625 mg daily, and MPA 10 mg on days 1–12 per calendar month. In the UK and Europe, oestradiol with either progesterone, dydrogesterone, or norethisterone are more commonly prescribed. In women who are several years postmenopausal, and who do not desire the return of menses, combined continuous regimens of oestrogen and progestogen or tibolone can be offered. The dose of progestogen is typically half that used in cyclical regimens. With this regimen, most women have amenorrhoea during long-term use, but irregular vaginal bleeding is common in the first few months of treatment (44). In older women, lower doses of oestrogen and progesterone (for example, CEE 0.3 mg, or oestradiol 1 mg daily or alternate daily, and dydrogesterone 5 mg, norethisterone 0.35–0.7 mg, or MPA 2.5–5 mg daily) can be used. The dosage of oestrogen can be increased gradually over 4–8 weeks, if necessary. A recent addition to the therapeutic possibilities is a newer progestin, drospirenone 2 mg, which has antimineralocorticoid and antiandrogenic properties, particularly suitable for women with hypertension or mild acne and hirsutism.

Common side effects of oestrogens include breast tenderness; this may improve after dose reduction, a change to tibolone or transdermal therapy, or with the passage of time. Some women experience nausea, which may be minimized if HRT is taken at night, or transdermally, rather than in the morning. Symptoms of fluid retention, bloating, and mood changes (which may be similar to premenstrual symptoms) are often caused by the progestogen component, and may improve with dose reduction or a different progestogen. Transdermal preparations and oestradiol implants may be helpful in women who have difficulty tolerating oral HRT, or whose symptoms fail to respond to oral treatment. If subcutaneous implants are used, serum oestradiol levels should be measured before the insertion of repeat implants, to avoid a progressive increase in oestradiol levels and the poorly understood phenomenon of oestrogen tachyphylaxis. As transdermal and subcutaneous oestradiol preparations are not subject to hepatic first pass metabolism, they are preferred in women with a history of VTE, liver dysfunction or cholelithiasis. Tibolone also does not appear to increase VTE risk, based primarily on observational and mechanistic data (36).

There is some evidence that androgen deficiency contributes to loss of libido in postmenopausal women, particularly after bilateral oophorectomy (45). If libido does not return to normal after adequate oestrogen replacement, then a trial of testosterone therapy is worthwhile. With the regimens shown in Box 10.1.2.5, virilizing effects rarely occur. It should be noted, however, that the long-term effects of testosterone treatment in women (particularly with respect to breast cancer risk) are not known.

There are other options that need to be considered when treating menopausal women. Long-term cardio-protection can be achieved with lipid-lowering agents, aspirin, and other well proven medications. Osteoporosis is increasingly treatable using a wide variety of medications including SERMs, bisphosphonates, strontium ranelate, parathyroid hormone, and vitamin D derivatives, which are beyond the scope of this discussion. Finally the option of ‘alternative therapies,’ primarily for menopausal symptoms is becoming increasingly available. These range from phyto-oestrogens (natural dietary plant based compounds with weak affinity for the oestrogen receptor) to ‘natural progesterone’ creams and a variety of herbal preparations. These compounds are described as alternative as they have generally been inadequately tested for both safety and efficacy. Limited studies so far have demonstrated weak oestrogenic effects of phyto-oestrogens, with no overall benefit on menopausal symptoms in randomized controlled trials. Current research suggests that other health benefits are related to vegetable proteins in whole food sources like soy, rather than to phyto-oestrogens, and there is little evidence to support the use of isolated phyto-oestrogen supplements. Remifemin may offer some benefits with improved side effect profiles, however much more research is needed to resolve these issues and demonstrate that such treatments are efficacious and safe.

The issues surrounding short-term hormone therapy for symptomatic women around the time of menopause are generally noncontroversial. In contrast, much debate surrounds the question of long-term hormone therapy, given primarily with the aims of cardiovascular and/or bone protection. Advances are currently being made in the development of new agents, such as compounds with varying target-site oestrogen receptor specificity, the SERMs. One promising development is a combination of low dose oestrogen and a new SERM, bazedoxifene. These have relatively protective vascular and bone effects, but appear to be inhibitory to the development of breast and endometrial cancer. Other approaches are available for cardiovascular protection (e.g. the statins) and for reduction of osteoporotic fracture risk (the bisphosphonates, strontium ranelate). The next few years should see clarification of many of these current controversies.