Abstract
Introduction
At menopause, the permanent cessation of menstruation, sex steroid hormone levels are low and gonadotropin levels are elevated. The average age of menopause is between 51 and 52 years, with a range of 40–60 years. The timing of menopause is both genetically and environmentally influenced. It occurs about 2 years earlier in women who smoke cigarettes than in those who do not smoke (1). The prevailing view has been that normal menopause results from the depletion of ovarian follicles that can be stimulated to ovulate. An alternative view is that age-related changes in the central nervous system initiate menopause and the exhaustion of ovarian follicles is a consequence of the altered temporal organization of neural signals (2).
Some believe that an enormous excess of follicles is required for normal ovarian function (3). When the number of follicles falls below a critical number, fine-tuning is lost, more follicles enter the growing pool, the number of ova ovulated during the cycle and the length of time between cycles becomes more variable, and fertility declines (2). Follicular loss is not linear throughout the reproductive life-span. There is a doubling of the rate of follicle loss in women at around the age of 35, when about 25 000 primordial follicles remain and perimenopausal symptoms begin to appear.
Ovarian function decreases gradually; the time between the onset of menstrual irregularity and menopause is called the perimenopause. In the perimenopausal period, plasma concentrations of FSH are elevated and are associated with increased serum oestradiol levels and urine levels of oestrogen conjugates. As the follicles become depleted and inhibin levels remain low, they no longer respond to elevated FSH levels and oestradiol levels fall. This leads to the failure of endometrial development and absence of uterine bleeding.
In a longitudinal study of 160 women over a 12-year period, the menstrual pattern changed from regular cycles to irregular cycles to cessation of menses (4). Increased FSH and, to a lesser extent, LH levels occurred about 5 years before the menopause. Oestrogen levels fell 6–12 months prior to menopause. This was associated with an increase in the oestrone to oestradiol ratio, reflecting the continued formation of oestrone from androstenedione in peripheral fat. Replacement with exogenous oestrogen does not reduce FSH to pre-menopausal levels, since FSH release is mainly suppressed by inhibin and inhibin levels remain low post-menopausally. Therefore, measurement of FSH cannot be used to determine whether sufficient exogenous oestrogen is being given to produce physiological replacement (1). Menopause, according to an alternative perspective, may be due to age-related changes in the central nervous system that drive the exhaustion of ovarian follicles.
An alternate view is that the hypothalamus initiates the events leading to menopause. Hypothalamic changes such as hot flushes, sleep disturbances, and changes in the GnRH secretion pattern, occur in women between 35 and 40 years of age, at the same time that the rate of follicular loss increases. Increased follicular loss may be due to a change in neuroendocrine control that governs the dynamics of follicular reawakening, recruitment, growth and differentiation. Elevated FSH levels are not always temporally correlated with decreased oestradiol or inhibin levels in individual women.
The pattern of pulsatile GnRH secretion determines the secretory patterns of LH and FSH. The frequency of LH pulses mirrors the GnRH pulse frequency. In middle-aged, regularly menstruating women, LH pulse frequency decreases and the width of the peak increases during the follicular phase, before changes in plasma oestradiol levels occur. FSH concentrations are also elevated during the early follicular phase (2). Imprecise hypothalamic GnRH pulse generation results from the uncoupling of the neurochemical factors that coordinate GnRH secretion. Elevated FSH levels during the follicular phase have been assumed to indicate changes in ovarian oestradiol and inhibin feedback. However, decreased GnRH pulse frequency or lowered GnRH concentrations can lead to a preferential FSH release.
Deterioration of the neural pacemaker in the biological clock located in the suprachiasmatic nuclei of the hypothalamus, or its coupling to outputs, may cause these desynchronizations (2). This may initiate the gradual disintegration of the temporal organization of neurotransmitter rhythms that are critical for stable, precise and regular cyclic LH secretion (2). Hot flushes occur in normally cycling women during the fourth decade, when more than 25 000 ovarian follicles should be present. In rodent models, changes in neurotransmitter activity, neurotransmitter receptor densities, and levels of mRNAs that encode for GnRH-regulating neuropeptides occur during middle age, when animals still exhibit regular oestrous cyclicity (2). This suggests that the hypothalamus is at least as important as the ovary in triggering menopause.
Hormonal replacement
The menopause is associated with both early and late symptoms (Table 1) (5, 6). Hot flushes, a vasomotor change, correlate with bursts of activity in hypothalamic pacemaker neurones leading to pulses of GnRH and thus, LH (7, 8). Hormonal replacement with oestrogens (oestrogen replacement therapy, ERT), alone or in combination with progestins (HRT), is often used to treat symptoms and prevent osteoporosis (9, 10). ERT may delay the onset and decrease the risk of Alzheimer's disease (11) and improve cognition, but this is controversial (12). Oestrogen therapy may decrease the risk of coronary artery disease (CAD) and hip fracture, but long-term, unopposed oestrogen therapy increases the risk of endometrial carcinoma (13). The increased endometrial cancer risk can be avoided by adding a progestin to the oestrogen regimen for women who have a uterus, probably without reducing the benefit for CAD (13–16). The effect of ERT on the risk of breast cancer is uncertain, with studies still yielding inconsistent results (13, 14). There appears to be no increased risk with short-term oestrogen use, but the risk of breast cancer may increase slightly with long-term use.
Menopausal symptoms
Six meta-analyses (1) concluded that there is no increased risk of developing breast cancer in women who had ever used oestrogen, while four meta-analyses (1) suggested that there was a 30% increased risk of developing breast cancer with long-term oestrogen use. Current users of oestrogen replacement (mainly conjugated equine oestrogen for more than 5 years) had a significantly increased risk of developing breast cancer in the Nurses' Health Study (1995) (17). Long-term prior oestrogen users had no increased risk. However, a case-control study from Washington state (18) found that long-term oestrogen use (20 years or more), or oestrogen-progestin use (8 years or more), did not increase the risk of breast cancer (relative risk being 1.0 and 0.4, respectively). No increased risk of breast cancer with oestrogen alone or with oestrogen plus progestin was found in a case control study of 3000 women (19). In a subset analysis, restricted to recent users of HRT, no increased breast cancer risk was observed with short- or long-term oestrogen or oestrogen plus progestin therapy. These recent studies suggest that post-menopausal oestrogen use is not a risk factor for breast cancer. Three meta-analyses examined the risk of oestrogen replacement and breast cancer in women with a family history of breast cancer. One (20) found a significantly increased risk of breast cancer in women oestrogen users with a family history of breast cancer compared with non-users. A more recent analysis (21) found no increased risk.
The American Heart Association recently reviewed cardiovascular disease in women (22). Prospective studies show a consistent reduced risk (35–50%) of coronary heart disease (13, 23, 24) and a less consistent reduced risk of stroke in women using ERT. Some of this reduction could be due to selection artifact. Petitti et al. (25), in a case control study of women with stroke, found that HRT (oestrogen alone or oestrogen/progesterone) was not associated with either an increase or decrease in ischaemic stroke risk.
In a recent study (26), 232 women who had used ERT for an average of 17 years had nearly 50% lower mortality from all causes than age-matched non-users (95% CI 0.38–0.76); reduced mortality was largely due to less fatal cardiovascular disease. In a case control study of women hospitalized at Kaiser Permanente for acute myocardial infarction from 1991 to 1994, Sidney et al. (27) estimated the relative risk for incident acute myocardial infarction in relation to the current use of oestrogen and oestrogen-progestogen. The odds ratio for myocardial infarction in current users of oestrogen or oestrogen-progestogen compared with women who had never used these agents was 0.96 (95% CI 0.66–1.40) after adjustment for confounders. The odds ratio for myocardial infarction in past users of oestrogen or oestrogen-progestogen was 1.07 (CI 0.72–1.58). The duration of hormone use was unrelated to the odds ratio for myocardial infarction. HRT was not associated with a statistically significant decrease in the odds ratio for myocardial infarction associated with current use of oestrogen or oestrogen (27).
The Nurses' Health Study (1997, 28), showed a reduced risk of nearly every disease except breast cancer in women using ERT. A recent Markov analysis (29) postulated a 3-year increase in life expectancy associated with the use of post-menopausal oestrogen. Women who are better educated and more affluent (well-established cardioprotective factors) are more likely to be prescribed ERT and to take it. This ‘healthy woman selection bias’ was seen in the Healthy Women's Study (30). Women who later elected to take HRT were, when pre-menopausal, significantly more educated and had significantly more favourable levels of HDL cholesterol, blood pressure, fasting insulin, body weight, alcohol intake and physical activity. Thus, the amount of protection attributed to oestrogen may be exaggerated (22).
In previous studies (31, 31–34, 34–36), estimates of the relative risk for myocardial infarction in users of oestrogen-progestogen varied considerably. Sidney et al. (27) summarized directly comparable studies of the effect of oestrogen on myocardial infarction — those that examined current users of oestrogen with both fatal and non-fatal coronary heart disease and adjusted for cardiovascular risk factors (Table 2). Relative risk estimates range from 0.60 in the most recent analysis of data from the Nurses' Health Study (36) to 0.97 in the 1976 hospital-based case-control study by Rosenberg and coworkers (37). Sidney et al.'s (27) estimate of the odds ratio for myocardial infarction in current users of oestrogen is closer to 1.0 than are many of the estimates from similar studies, but it is within the range of estimates found in previous studies.
Breast cancer risk in women using HRT
Results of case-control and cohort studies; current use of oestrogen replacement therapy that adjusted for coronary risk factors – age, hypertension, stroke and diabetes.
Table adapted from Sidney et al. (27).
Most studies describing cardiovascular protection included women using unopposed conjugated equine oestrogen. Two recent US studies found similarly reduced cardiovascular risks for women taking oestrogen plus a progestin, usually medroxyprogesterone acetate, suggesting that this combination is equally protective (33, 38). Several new studies suggest that medroxyprogesterone acetate may mask or at least partially reverse oestrogen-induced benefits. In the post-menopausal oestrogen/progestin 3-year placebo-controlled randomized double-blind trial intervention study (39), HDL cholesterol levels rose significantly more in the women who were assigned to unopposed conjugated equine oestrogen than in those who were assigned to conjugated equine oestrogen plus continuous or cyclic medroxyprogesterone acetate.
An increased risk of venous thromboembolism in women taking ERT was demonstrated in two case-control studies (40, 41) and one prospective study of pulmonary emboli (41). All showed a similar three- to four-fold increased risk associated with ERT. Although the relative risk was high, the absolute risk was low, owing to the infrequency of events with or without hormone therapy.
Two ongoing trials in the United States of America (USA) have been designed to quantify the cardioprotective effect of ERT unconfounded by healthy woman selection bias. HERS is a 5-year randomized placebo-controlled secondary prevention trial of conjugated equine oestrogen plus medroxyprogesterone acetate in 2763 post-menopausal women who already have coronary heart disease. Unopposed oestrogen is not being evaluated (42). The primary outcome was the occurrence of non-fatal myocardial infarction or coronary heart disease death. Secondary cardiovascular outcomes included coronary revascularization, unstable angina, congestive heart failure, resuscitated cardiac arrest, stroke or transient ischaemic attack, and peripheral arterial disease. All-cause mortality was also considered.
During an average follow-up of 4.1 years, treatment with oral conjugated equine oestrogen plus medroxyprogesterone acetate did not reduce the overall rate of coronary heart disease events in post-menopausal women with established coronary disease. The treatment did increase the rate of thromboembolic events and gall bladder disease. Based on the finding of no overall cardiovascular benefit and a pattern of early increase in risk of coronary heart disease events, the authors did not recommend starting this treatment for the purpose of secondary prevention of coronary heart disease. However, given the favourable pattern of coronary heart disease events after several years of therapy, continued treatment was recommended for women already receiving this treatment.
There is a great deal of observational data on ERT and HRT used for primary prevention. Women with coronary disease are likely to have unmodifiable risk factors, such as diabetes and obesity, that influence the tendency to thrombosis. If this is so, any procoagulant effect of hormones would be greatest in women who have coronary disease. The heart and oestrogen/progestin replacement study results should not be immediately extrapolated to ERT and HRT used for primary prevention (43).
The Women's Health Initiative, a placebo-controlled primary prevention trial in 27 500 post-menopausal women, was scheduled to complete randomization in 1998 (44). The three major outcomes are coronary vessel disease, osteoporosis and breast cancer. Active treatments are conjugated equine oestrogen for women without a uterus or conjugated equine oestrogen plus continuous medroxyprogesterone acetate for women with an intact uterus, as compared with placebo. The Women's Health Initiative is planned to end in 2006.
Yaffe et al. (12) reviewed the literature on the effects of ERT on cognitive function and dementia in post-menopausal women. Five observational studies and eight trials addressed the ERT effect on cognitive function in non-demented post-menopausal women. Cognition improved in perimenopausal women, possibly because menopausal symptoms improved, but there was no clear benefit in asymptomatic women. Ten observational studies measured the effects of ERT use on the risk of developing dementia. Meta-analysis of these studies suggests a 29% decreased risk of developing dementia among oestrogen users, but the findings of the studies are heterogeneous. Four positive trials of ERT in women with Alzheimer's disease were small, of short duration, non-randomized, or uncontrolled. All of the studies have methodological problems and have produced conflicting results, reflecting, in part, the healthy women bias.
The menopause presents a particular set of problems in women in whom oestrogen replacement is indicated but its use leads to a worsening of migraine symptoms. Women should use the lowest dose of oestrogen that will relieve vasomotor symptoms, prevent vaginal and urethral epithelial atrophy, maintain the collagen content of the skin, reduce the rate of bone resorption, and prevent acceleration of atherosclerosis. The most commonly used oestrogen, Premarin™, is a mixture of oestrogens, including the equine oestrogen equilin (Table 3) (45). Pure oestrones, oestradiols, and synthetic ethinyl oestradiol are also available (46). Oestrogen replacement should be physiological, not pharmaco-logical (1). The optimal long-term oestrogen dose given to asymptomatic women to reduce the risk of osteoporosis and cardiovascular disease (physiological replacement dose) is 0.625 mg of conjugated equine oestrogen or oestrone sulphate or 1 mg of micronized oestradiol, and probably 0.05 mg transdermally. Higher oestrogen doses may be needed for 1 or 2 years to relieve hot flushes. Oestrogens can be taken sequentially for 25 days a month with the addition of a progestational agent on days 16–25 to induce bleeding; alternatively, oestrogens and progesterone can be taken continuously (47). Oestrogens can also be given for 5 days each week (skipping the weekend) to control breast tenderness.
Hormonal preparations
Oestrogens are available orally or parenterally in the form of injection, vaginal cream, oestrogen ring or transdermal patch (45). Parenteral administration of oestrogens produces fewer hepatic effects and a higher, more physiological serum oestradiol-to-oestrone ratio than oral administration (46, 48). Experimental implants and transdermal patches provide stable blood oestrogen levels (8, 9); however, oestradiol pellets for implantation are available only on an investigational basis (49). Adjunct hormones include progesterone (50) for prevention of endometrial cancer (not needed after hysterectomy) and androgens (51) to combat decreased libido and sexual responsiveness and, perhaps, fatigue, depression and headache.
Progesterone itself is associated with side-effects, including increased headache and other CNS side-effects and, in women using progesterone for contraception, is one of the major reasons for discontinuation. The benefits of oestrogen replacement therapy exceed any possible risk, but many post-menopausal women have not taken it, and, if they did, often discontinued it within 12 months. Contraindications to oestrogen therapy include breast or endometrial cancer, active thrombo-phlebitis, and undiagnosed abnormal uterine bleeding. A prior history of thromboembolic events associated with exogenous oestrogen is also a contraindication (1).
Women without a uterus are not at risk for endometrial cancer and can use oestrogens unopposed, either cyclically or continuously. Post-menopausal women with a uterus should take progestins either sequentially or continuously. In the sequential regimen, the progestin is given daily for the first 10–14 days of the month. In the combined regimen, both the oestrogen and the progestin are given daily. The cyclic regimens produce monthly withdrawal bleeding in about 80% of women. The continuous regimen, wherein oestrogens and progestins are given daily, usually results in irregular bleeding and spotting during the first 3 months, but with longer use nearly all women become amenorrheic (52). Oestrogen and progestogens are available in combination oral tablets and in a transdermal patch containing oestradiol and norethindrone acetate.
Selective oestrogen receptor modulators (SERMs) are a new class of drugs that differ from each other and from classic oestrogen. Raloxifene, the newest SERM, has oestrogen-like effects on bone and on lipid metabolism, but no effect on the breast or uterus (53). It, like oestrogen, is associated with venous thrombosis. Tamoxifen is also a SERM with oestrogen-like effects on the uterus, but not on the breast. In clinical trials, there was no increase in reports of migraine or headache in women receiving raloxifene compared with those receiving placebo or HRT. In fact, the incidence of headache was statistically lower in women receiving Raloxifene (60 mg) than in those receiving placebo. Raloxifene may be a choice for women who cannot tolerate standard oestrogens (53–55).
Although migraine prevalence decreases with advancing age (56), migraine can either regress or worsen at menopause (57). Neri et al. (58) investigated 556 consecutive post-menopausal women who attended an outpatient clinic and found that headache was present in 13.7%. Most (82%) had had headaches prior to the onset of menopause. Many (62%) had migraine without aura; the remainder had tension-type headaches. None of the women had migraine with aura or cluster headache. Women with prior migraine generally (2/3) improved with physiological menopause. In contrast, surgical menopause usually (2/3) resulted in a worsening of migraine. Other studies have shown that hysterectomy or oophorectomy is not an effective treatment for migraine at any age (51, 59) despite recent suggestions to the contrary (60, 61). Oestrogen (and progestin) replacement therapy can exacerbate migraine (62–64) or, alone (65) or with testosterone (66), relieve it. This has been confirmed in one (67), but not another (68), double-blind study. The use of drugs for the treatment of migraine in menopausal women who do not need replacement oestrogens should be guided by their cardiac and renal status (47). Refractory cases may be treated with hormonal replacement (69).
Headache management can be difficult in women who require hormonal replacement therapy for menopausal symptoms but develop headaches as a result of the therapy. Several empirical strategies may be utilized (Table 4). Reducing the dose of oestrogen or changing the type of oestrogen from a conjugated oestrogen to pure oestradiol, to synthetic ethinyl oestradiol, or to a pure oestrone may significantly reduce headache. Aylward (63), in a controlled double-blind cross-over trial of menopausal women, found that oral oestropipate decreased the frequency and intensity of headache, whereas ethinyl oestradiol increased the headache. Changing from interrupted to continuous administration may be very effective if the headaches are associated with oestrogen withdrawal. Techniques may be combined. Kudrow (62) reported a 58% improvement in headache control with a reduced continuous dose of oestrogen. Parenteral oestrogens, with or without adjunct hormones, can be effective. Greenblatt (66) studied post-menopausal women with oral oestrogen-induced headaches and found that their headaches could be improved by switching from oral to parenteral oestrogens (oestradiol) and adding androgens (testosterone). The oestradiol cutaneous patch, which provides a physiological ratio of oestradiol to oestrone and a steady-state concentration of oestrogen, has been associated anecdotally with fewer headache side-effects; however, this has not been proven in any controlled study (45, 70, 71). The new SERM, raloxifene, can also be used if a women requires, but cannot tolerate, non-selective oestrogen.
Treatment of hormonal replacement headache
Progestins, used to prevent endometrial hyperplasia, can cause headache in addition to other symptoms of PMS, particularly if used cyclically. Giving a lower dose of a progestin (medroxyprogesterone, 2.5 mg vs. 7.5 mg) continuously can often control this. Another strategy is to change the type of progestin. Women who received norethindrone had less depression than those who received medroxyprogesterone acetate (72). Women with a uterus who have intolerable mental symptoms with progestins may be placed on an oestrogen-only regimen in conjunction with an annual endometrial biopsy or vaginal ultrasonography to measure the endometrial thickness. If the endometrial echo complex is less than 4 mm, it may not be necessary to perform a biopsy since the endometrial cancer risk is very low (1).
Another strategy is to use targeted drug delivery. The new adhesive vaginal gel containing micronized progesterone in an emulsion system was designed to maximize progesterone's therapeutic effect on the uterus while minimizing the potential for systemic side-effects. The vaginal route is traditionally used for local therapy or as an alternative route for systemic drug administration. However, the uterine effects of vaginal progesterone gels exceed the response due to circulating levels of progesterone due to a direct first pass effect. It has been shown experimentally that radio-labelled progesterone (73), applied to the cuff of vaginal tissue remaining attached to the cervix after hysterectomy, diffused to the entire uterus in the absence of systemic circulation. Endometrial progesterone extraction was higher in the luteal than in the proliferative phase. Progesterone vaginal gel absorption has a long absorption half-life of 25–50 h, an elimination half-life of 5–20 min, and a dose-dependent bioavailability of 20–28%. The direct uterine progesterone delivery allows lower systemic progestin levels that provide endometrial protection with fewer progesterone side-effects. It is available as a 4% (45 mg) and 8% (90 mg) formulation (Crinone™) that can be given every other day (54).