Hormone replacement therapy—estrogens

General information

For a complete account of adverse reactions to estrogens, readers should consult the following monographs as well as this one:

Diethylstilbestrol
Estrogens
Hormonal contraceptives—emergency contraception
Hormonal contraceptives—oral
Hormone replacement therapy—estrogens + androgens
Hormone replacement therapy—estrogens + progestogens.

Estrogen replacement therapy was until recently widely recommended for the prevention of osteoporosis in middle-aged and older women [ ]. Long-term estrogen therapy also reduces the incidence of ischemic heart disease in such women [ ]. However, it has always been difficult to know in which women such prophylactic use is likely to be needed, and this dilemma is compounded by the prospect of adverse reactions, which can include any of the acute reactions listed in the estrogen monograph, but also in some cases long-term reactions such as tumorigenicity.

The multiplicity of hormone replacement therapy regimens in use (involving one or two drugs, continuous or intermittent treatment, and various forms of administration) has made it difficult to express any general conclusion about the benefit-to-harm balance [ ], and there has been a thoughtful review of the obstacles to assessing these matters objectively and scientifically, including questions of both ethics and trial design [ ]. Even the ultimate effect of hormone replacement therapy on the incidence of ischemic heart disease remains subject to dispute [ ], and incidental reports of cardiac complications, sometimes in women with entirely healthy coronary vessels, continue to cause concern [ ]. For such reasons, much work has been devoted to determining the lowest effective dose of estrogen needed to achieve particular results.

An extraordinarily thorough review of all significant controlled studies of estrogen replacement therapy has been published by the Cochrane Collaboration [ ]. The participants totalled 2511 and the trials lasted 0.25–3 years. The primary purpose was to confirm efficacy, but data on adverse events were also collected. Withdrawal because of adverse events, commonly breast tenderness, edema, joint pains, and psychological symptoms, was not significantly higher with HRT than placebo (OR = 1.38; 95% CI = 0.87, 2.21). Breast tenderness and withdrawal bleeding were the only significant problems in terms of frequency. The studies did not justify the conclusion that there are serious adverse reactions, such as thrombosis and malignancies.

The desirability of limiting treatment to the minority of women who really need it is now regularly echoed [ ]. As has been stressed in a brief review [ ], the Women’s Health Initiative study was stopped because of safety concerns regarding combined treatment with estrogens + progestogens, and it was an entirely fair conclusion that the risks of the treatment outweighed the benefits. The outcome of the Women’s Health Initiative has had drastic consequences for the acceptance of hormone replacement therapy among women who have been informed of the findings [ ]. All the same, the risks were difficult to express in exact terms, and this has led to criticism of the work and its conclusions [ ]. There was undoubtedly an increase in breast cancer among users, but no greater than the increase reported in observational studies; there was an increase in vascular events, but many of the women in the study had risk factors for cardiovascular disease. Since no further long-term trials of combined hormone replacement therapy appear to be in progress, the ultimate truth regarding the benefit to harm balance may remain in doubt for very long time. “Hormone replacement therapy can still be prescribed for menopausal symptoms and for osteoporosis prevention, but the need for continued use should be reviewed annually” [ ], and “a consensus is growing that postmenopausal women may be treated with HRT only … for disturbing symptoms of the ovarian hormone insufficiency syndrome, rather than … for menopause per se” [ ].

A number of groups continue to stress the possibility, raised in the Nurses' Health Study and elsewhere, that by reducing the dosages used in hormonal replacement one might eliminate the risks while retaining efficacy [ ], but so long as this has not been shown to be true it will not provide a way ahead.

The Million Women Study in the UK [ ] has been criticized on various grounds by some proponents of hormone replacement therapy, notably because in their view the number of deaths from breast cancer was too small and the follow-up too short to justify the belief that HRT increases the risk of death from breast cancer; discrepancies between this and other studies have also been stressed [ ]. The interpretation of the study in the editorial that accompanied it has also been criticized as being unduly pessimistic. However, the fact remains that the study is not the only source of serious doubts about the benefit to harm balance of hormone replacement therapy.

With such disagreements and doubts it can be difficult to provide a woman with the properly balanced information which she needs if “informed consent” to hormonal replacement treatment is to be meaningful [ ].

The evidence of adverse reactions emerging from randomized trials in and around 2002 has resulted in a dramatic fall in the use of hormone replacement therapy, the use of some formulations falling by two-thirds [ ]. Continuing debate on the benefit to harm balance of hormonal therapy at the time of the menopause or subsequently has in the recent past hardly yielded new conclusions [ ].

Assessment of the efficacy and safety of hormonal replacement therapy has always been bedevilled by the question of adherence to therapy, which is often poorly recorded. Particularly after the acute symptoms of the menopause have passed and been forgotten, some women lose their motivation to continue hormonal treatment. Added to this is the now incontrovertible evidence that HRT can in some cases be harmful, and presentation of the risks in the mass media has undoubtedly further reduced adherence. An Italian group examined this problem prospectively in 138 women who agreed to be enrolled in a longitudinal study intended to last for 24 months [ ]. Only 72 were still taking the treatment after 1 year and only 56 at the end of the study, although only three reported that they had experienced no benefit. Type of work, surgical menopause, and previous use of oral contraceptives were significantly associated with better adherence. The occurrence of real or supposed adverse reactions and a fear of breast cancer were the commonest reasons cited for early discontinuation.

Drug studies

Observational studies

A Japanese study of the use of estriol 2 mg/day for 12 months in 68 postmenopausal women with climacteric symptoms showed a significant effect in relieving hot flushes, night sweats, and insomnia [ , ]. There were significant falls in serum follicle stimulating hormone (FSH) and luteinizing hormone (LH) concentrations, but no effect on lipids, bone demineralization, or blood pressure. There was slight vaginal bleeding in 14% of women treated during a natural menopause, but histological and ultrasound evaluation showed no changes in the endometrium or breasts. It is evident, however, that higher doses might be needed when treating women of other races with a higher body weight. Other workers have found that when given with a progestogen over long periods, estriol 2.0 mg/day seems much less likely to cause undesirable lipid changes than are equine conjugated estrogens, which can cause increased HDL cholesterol and triglyceride concentrations [ ].

Comparative studies

In a randomized, multicenter study in Denmark 376 perimenopausal women with climacteric symptoms were randomly allocated to oral sequential combined treatment with regimens based on estrogen plus either desogestrel or medroxyprogesterone acetate [ ]. Both treatments effectively alleviated menopausal complaints within 6 months and gave good cycle control. Bleeding pattern and mood disturbances were more favorably affected by desogestrel, but overall the differences in adverse reactions (irregular bleeding and a slight tendency to hypotension) were not large. It should be noted, however, that with cyclic combined hormone replacement therapy treatment, the bleeding pattern alone does not seem to be a reliable means of distinguishing cases in which the endometrium is atrophic or inactive from those in which it is proliferative or hyperplastic [ ].

Organs and systems

Cardiovascular

The overall benefit to harm balance for medium-term hormone (estrogen + progestogen) replacement therapy (2 years) has been assessed in a novel although theoretical study using a Markov model. Data from the Women’s Health Initiative were used to simulate the effect of short-term treatment on life expectancy and quality-adjusted life expectancy (QALE) in 50-year-old menopausal women with intact uteri [ ]. Quality-of-life (QOL) utility scores were derived from the literature. The investigators assumed that HRT affected quality of life only during the perimenopause, when it reduced symptoms by 80%. Among asymptomatic women, short-term therapy was associated with net reductions in life expectancy and QALE of 1–3 months, depending on the risk of cardiovascular disease. Women with mild or severe menopausal symptoms gained 3–4 months or 7–8 months of QALE respectively. Among women at low risk of cardiovascular complications, HRT extended QALE, even if menopausal symptoms lowered quality of life by as little as 4%. Among women at increased risk of cardiovascular disease, HRT extended QALE only if symptoms lowered quality of life by at least 12%. The authors’ conclusion was that hormone therapy is associated with reduced survival but gains in QALE for women with menopausal symptoms, and that women who are expected to benefit from short-term HRT can be identified by the severity of their menopausal symptoms and their risk of cardiovascular disease.

Coronary heart disease

The debate about the relation, if any, between HRT and coronary heart disease has been reviewed [ , ]. A paradoxical finding in the Women’s Health Initiative (WHI) study was that while HRT resulted in an improvement in blood lipid concentrations there was no reduction in the incidence of coronary heart disease. A re-examination of the findings in 2005 generated the hypothesis that the key to the paradox could lie in effects on specific lipid subgroups rather than on lipids as a whole [ ]. This hypothesis was tested by an evaluation of differences in coronary calcification, lipids, and lipoprotein subclasses among menopausal HRT users and non-users in a longitudinal study. Lipoprotein subclasses and coronary artery calcification (the latter measured using electron beam computed tomography) were studied in HRT users (49%) and non-users in a total of 243 women from the Healthy Women Study, who were about 8 years postmenopausal. The distribution of calcification scores was not significantly different between users and non-users and neither were there differences between the groups as regards any LDL subclass. However, regardless of HRT use, women with detectable calcification of the coronary arteries had higher concentrations of VLDL and small LDL particles, higher LDL particle concentration, and smaller mean LDL size compared with women with no detectable calcification. The fact that HRT users had higher concentrations of VLDL particles (triglycerides) and did not have a better LDL subclass distribution could explain the fact that HRT was not associated with a difference in coronary calcification in this study or with a reduction in coronary heart disease risk in randomized clinical trials.

Thromboembolic disease

One of the most serious aspects of the thromboembolic complications now widely acknowledged as being associated with HRT is that their emergence coincides with the development of the conclusion that the role of HRT in reducing the risk of coronary heart disease is at best unproven. A form of treatment that was originally viewed as potentially beneficial to the cardiovascular system is at present on balance perhaps harmful [ ].

A US group examined potential risk factors for venous thromboembolic events in women assigned to HRT in the Postmenopausal Estrogen/Progestin Interventions (PEPI) study, a 3-year double-blind study in 875 postmenopausal women designed to assess the effects of HRT on heart disease risk factors (HDL cholesterol, fibrinogen, blood pressure, and insulin) [ ]. Women with a history of estrogen-associated venous thromboembolic events were excluded. Ten women, all assigned to HRT, had a venous thromboembolic event during the study. Only baseline fibrinogen varied significantly between those who had a venous thromboembolic event while assigned to HRT event (mean 2.49 g/l) and those who did not have an event (mean 2.81 g/l). Adjusting for covariates did not affect this finding. As the authors remarked, the lower fibrinogen concentrations among women who subsequently reported venous thromboembolic events may be a marker for a specific, but as yet undefined, coagulopathy that is magnified in the presence of exogenous hormones. However, larger studies are needed to confirm this hypothesis.

Since much of the evidence of thromboembolic complications with HRT relates to the use of conjugated equine estrogens, the degree of risk when natural 17-beta-estradiol was used instead has been examined in Norway in a population-based case-control study involving consecutive women, aged 44–70 years, discharged from a University Hospital between 1990 and 1996 with a diagnosis of deep venous thrombosis or pulmonary embolism [ ] Women with cancer-associated thrombosis were excluded. Random controls were used. The material comprised 176 cases and 352 controls, that is two controls for each case. All the women who received HRT had been given estradiol. The frequency of HRT use was 28% (50/176) in cases and 26% (93/352) in controls. The estimated matched crude odds ratio was 1.13 (CI = 0.71, 1.78), which shows no significant association of overall use of estradiol-based HRT and thromboembolism. However, when the duration of exposure to HRT was taken into account by stratification, there was an increased risk of thromboembolism during the first year of use, with a crude odds ratio of 3.54 (CI = 1.54, 5.2). This effect was reduced by extended use to a crude odds ratio of 0.66 (CI = 0.39, 1.10) after the first year of use. The authors concluded that the use of estradiol for HRT was associated with a three-fold increase in the risk of venous thromboembolism, but that this increased risk was restricted to the first year of use. One is bound to wonder whether this shift in risk was genuine or reflects only the limitations of the study.

Among the less common forms of thromboembolism that have been reported is occlusion of the retinal vein, familiar with the oral contraceptives but unusual with HRT [ ].

QT interval prolongation

Hormone replacement therapy can cause prolongation of the electrocardiographic QT interval. In a prospective study of the incidence and extent of this effect, 3103 women were followed at intervals over 9 years and data were collected on their use of hormonal replacement and the QT interval, prolongation being defined as an increase of not less than 10% [ ]. The QT _c interval was moderately but significantly longer in users of hormone replacement therapy, the risk of QT prolongation being nearly twice that in never-users. The consequences of these changes for the user’s health could not be assessed.

The effects of HRT on cardiovascular function have been prospectively studied over 1 year in 46 healthy postmenopausal women, mean age 55 years, who took either estrogen replacement therapy alone (n = 23) or progestogen + estrogen replacement therapy (n = 23) [ ] The doses used were 0.625 mg/day of conjugated equine estrogen with or without medroxyprogesterone acetate 2.5 mg/day. The controls were 25 health premenopausal women, mean age 35 years. Long-term estrogen-only replacement increased the QT interval, QT dispersal, and the index of parasympathetic activity; there was also a small non-significant increase in the incidence of dysrhythmias. Long term use of progestogen + estrogen did not affect the QT interval, QT dispersion, or the frequencies of ventricular dysrhythmias or parasympathetic activity, but it did increase the incidence of supraventricular tachycardias. These findings support the idea that estrogen may directly modulate ventricular repolarization and that progestogens do not.

Respiratory

Since female reproductive hormones appear to influence respiratory disease in some as yet poorly defined way, a prospective cohort study over 8 years sought to determine whether postmenopausal hormone use was associated with an increased rate of newly diagnosed asthma or chronic obstructive pulmonary disease [ ] During 546 259 person-years of follow-up, current use of estrogen alone was associated with an increased rate of asthma (multivariate rate ratio = 2.29; 95% CI = 1.59, 3.29) compared with those who never used hormones. Current users of estrogen + progestogen had a similarly increased rate of newly diagnosed asthma. Rate ratios increased with certainty of diagnosis of asthma. In contrast, rates of newly diagnosed obstructive pulmonary disease were the same among hormone users and non-users.

Ear, nose, throat

Visual hallucinations have been associated with estrogen in a patient with Charles Bonnet syndrome [ ].

An 84-year-old woman with poor visual acuity secondary to bilateral, non-exudative, age-related macular degeneration had non-threatening visual hallucinations 2 weeks after starting oral estrogen for osteoporosis. The estrogen was withdrawn and the hallucinations subsided. She was given estrogen twice more and each time the hallucinations recurred.

In this patient estrogen may have promoted release phenomena and triggered the hallucinatory episodes.

Nervous system

Over the years there have been reports that headache or classical migraine is either alleviated or exacerbated by HRT. This was studied in 50 menopausal women with headaches who were randomized to either transdermal estradiol 50 micrograms for 7 days a month or oral conjugated estrogens 0.625 mg/day for 28 days, both regimens being supplemented with medroxyprogesterone acetate 10 mg/day during the latter half of each month [ ]. In patients with episodic tension headache there was no significant change in headache pattern. However, in women with migraine without aura the frequency and duration of the attacks increased significantly during HRT in the subgroup using the oral formulation; the transdermal formulation had no effect on the migraine pattern.

Sensory systems

The possibility that estrogen replacement therapy might cause dryness of the eyes has been reviewed in the light of data from the large Women’s Health Study in the USA [ ], and this work has been further reviewed [ ]. Questionnaires were sent to 25 665 participants. For every 3-year increase in the duration of replacement therapy there was a 15% increase in the incidence of dry eye syndrome. The risk was greater in women who used estrogen alone than in those who used combined estrogen + progestogen regimens. The evidence was not statistically strong, but it suggests that there is some correlation and that users and prescribers should be aware of it.

Psychological

The mechanisms underlying the mood changes that are often associated with menstruation, the menopause, and hormonal therapy are not understood, but there is some evidence that they are associated with the response to the neuroactive steroid pregnanolone and that progestogens and estrogens might alter this response. In a randomized, double-blind, crossover study 26 postmenopausal women with climacteric symptoms took oral estradiol 2 mg/day continuously for two cycles and either vaginal progesterone 800 mg or a placebo during the last 14 days of each cycle [ ]. Before treatment and again at the end of each treatment cycle pregnanolone was administered intravenously, and its effects on saccadic eye velocity, saccade acceleration, saccade latency, and self-rated sedation were examined. During treatment with either estradiol alone or with added progesterone the effect of pregnanolone on saccadic eye movements and self-rated sedation was increased. Saccadic eye velocity, saccade acceleration, and sedation responses to pregnanolone were also increased in women who usually experienced cyclicity of mood during HRT treatment, but not in those with no history of mood cyclicity.

Metabolism

Lipids

Changes in lipids have been observed with both HRT and oral contraceptives and have sometimes been promoted as potentially advantageous, but it is not clear how significant such changes are, at least in biochemical terms. A review and pooled analysis of 248 prospective studies available up to the year 2000 has provided data on this issue in postmenopausal women [ ]. All estrogen-only regimens raised HDL cholesterol and lowered LDL and total cholesterol. Oral estrogens raised triglycerides. Transdermal 17-beta-estradiol lowered triglycerides.

The unfavorable effects of estrogens on blood lipids are particularly serious in women with hypertriglyceridemia. The changes in women with previously raised lipids have been quantified [ ]. In 56 women, median age 52 years, with serum triglyceride concentrations over 4.5 mmol/l (400 mg/dl) despite diet and drug treatment, or with a history of hypertriglyceridemic acute pancreatitis. Of the 56 women, 23 had taken some form of hormone replacement therapy and one had been treated with selective estrogen receptor modulators; in these women at entry the median fasting triglyceride concentration averaged 1270 mg/dl and in the previously untreated women 1087 mg/dl. After baseline testing, the hormonal treatment was withdrawn and they received a very low fat diet, gemfibrozil 1.2–1.5 mg/day, and omega-3-fatty acids 4–12 g/day. After 2–4 weeks of this treatment, the median triglyceride concentration in those who had taken hormones fell from 1270 to 284 mg/dl and in the non-hormone group from 1087 to 326 mg/dl. The authors stressed the seriousness of the lipid complications that can occur with estrogens when women already have a tendency, familial or otherwise, to hypertriglyceridemia. Triglycerides must be measured before beginning hormone therapy; hormone therapy is contraindicated in women with a triglyceride concentration of 500 mg/dl or more. Triglyceride-lowering diets and drugs are likely to fail as long as hormones are being given.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here