Postmenopausal hormone therapy: the pros and cons
Funding: Unrestricted research grant support was provided by Proctor & Gamble, Servier, and Solvay and Wyeth.
Conflicts of interest: Susan Davis is a consultant and on the advisory board for Servier, Cellergy, Acrux and Proctor & Gamble. She is also an investigator for Proctor & Gamble, Servier, Acrux and Organon.
Susan Davis, The Jean Hailes Foundation, 179 Carnish Road, Clayton, Vic. 3168, Australia. Email: firstname.lastname@example.org
The use of postmenopausal hormone therapy (HT) has recently generated much debate following the results of various large randomised controlled trials. This has challenged physicians to reassess the use of HT in each individual woman within the realm of her symptomatology and risk profile for long-term illnesses. There are several areas of definite benefit or risk and many areas of uncertainty. The results of recent randomised controlled trials pertain to the specific HT regimens used and the characteristics of the study groups, such that the conclusions cannot be extrapolated to the use of HT in general. (Intern Med J 2004; 34: 109−114)
Postmenopausal oestrogen with or without progestin hormone therapy (HT) has been used for decades to treat menopausal symptoms successfully, in particular hot flushes, sweats, insomnia and urogenital atrophy, and there are no equally effective alternatives.1 Much concern and controversy has been generated following the results of several large randomised controlled trials (RCT). However, inferences from these results can only be applied to a given study group and need to be interpreted in the context of absolute risk. We have reviewed the risks and benefits of conventional HT as oestrogen alone or oestrogen plus progestin. Other approaches, such as androgens, selective oestrogen receptor modulators (raloxifene) and tissue specific-steroids (tibolone), are beyond the scope of this article.
HT is predominantly used by Australian women for the relief of menopausal symptoms.2 These include hot flushes, sweats, mood changes and those symptoms related to urogenital atrophy. In many women, relief of postmenopausal symptoms is imperative to an improved quality of life (QOL) and well-being.
At least 50% of postmenopausal women suffer symptoms of urogenital atrophy, including vaginal dryness, dyspareunia, recurrent urinary tract infections and stress incontinence.3 Oestrogen receptors have been demonstrated in the lower genito-urinary tract and pelvic floor. Systemic oestrogen therapy alleviates symptoms of urogenital atrophy. Vaginal application of oestrogen provides an alternative approach that is safe and acceptable and avoids systemic exposure to exogenous oestrogen.4 This is important in a treatment that is often required long term as urogenital atrophy is a lifelong disorder.
Osteoporosis is a significant cause of morbidity and mortality in postmenopausal women. At age 50 years, a woman has a 60% lifetime risk of sustaining an osteoporotic fracture and a 16% risk of hip fracture.5 These risks are partly attributable to the accelerated bone loss, that occurs after the menopause as a result of oestrogen deficiency.
HT reduces bone turnover, increases bone mineral density (BMD) and decreases vertebral and hip fracture rates by up to 35%.6 Importantly, the beneficial skeletal effects of oestrogen therapy are not attenuated by increased age.7 Progestin does not appear to enhance the effects of oestrogen on BMD.
In the Women's Health Initiative (WHI) study comparing the effects of 0.625 mg continuous conjugated oestrogens (CEE) plus 2.5 mg medroxyprogesterone acetate (MPA) versus placebo, the absolute risk of all fractures was reduced by 44/10 000 HT users per year and that of hip fractures by 5/10 000 HT users per year.8 Despite these findings this study group was not selected on the basis of osteoporosis risk. Currently, HT is the only agent indicated for Pharmaceutical Benefits Scheme support for the prevention of fractures in women with osteoporosis and no prior fracture.
If HT is stopped, bone loss resumes.9 As the median age for hip fracture is 79 years,5 women who use HT short-term for symptom relief may gain little or no protection against fracture beyond the age of 70 years.10 For the most effective protection against fracture, HT should be either started at the menopause and continued long term or started later in life when the risk of fracture is greater.11 Alternatively, HT can be initiated to manage menopausal symptoms with progression to other agents as women age.
Accumulating evidence suggests that postmenopausal hormone use might decrease the risk of colorectal cancer. The Nurses Health Study showed apparent protection against colorectal cancer with HT, predominately in current users with a substantial reduction in protection associated with cessation of HT.12 Similarly, in the Cancer Prevention Study II a reduction in risk of fatal colon cancer was observed among oestrogen users.13 These findings have been confirmed in a RCT with an absolute risk reduction of six cases per 10 000 HT users per year.8 With such a small absolute benefit, the use of HT to prevent colorectal cancer cannot be advocated.
The data regarding oestrogen use in preventing cognitive decline are promising but by no means conclusive. In a meta-analysis of the role of HT in cognitive function, women with menopausal symptoms had improvements in verbal memory, vigilance, reasoning and motor speed.14 No effects were observed in asymptomatic women. The meta-analysis was limited by inadequate numbers of RCT, the predominant use of oral CEE and small study sizes. The results of the Womens’ Health Initiative Memory Study (WHIMS), a RCT, found that CEE (0.625 mg) plus MPA (2.5 mg) did not significantly worsen nor improve cognitive function compared with placebo.15 This study had some methodological limitations and the findings are not surprising given that the women in whom HT is not routinely commenced were aged over 65 years.
The role of HT in dementia needs to be evaluated independently for primary and secondary prevention.
A recent meta-analysis of observational studies suggested that the use of HT in healthy postmenopausal women decreased the risk of dementia. Furthermore, oestrogen users tended to develop Alzheimer's disease later than non-users and HT use beyond 1 year was associated with a greater reduction in risk. However, most studies in this review had important methodological limitations.14
In the Cache County Study, a prospective observational study of 1889 elderly women (mean age 74.5 years), prior HT use and current HT use exceeding 10 years were associated with reduced risk of Alzheimer's disease.16
The WHIMS also assessed the role of CEE (0.625 mg) plus MPA (2.5 mg) in the prevention of dementia.15 In this study of 4532 women, average age 71 years, the absolute risk of dementia was 45 cases per 10 000 HT users per year versus 22 cases per 10 000 women per year in the placebo group. The causative role of HT in this group is questionable given the generation time required for the development of dementia. In summary, HT should not be initiated in older women for the prevention of dementia but this does not apply to initiating HT in younger women.
Further research assessing the role of HT in primary prevention of dementia is currently ongoing in a large RCT: preventing postmenopausal memory loss and Alzheimer's with replacement oestrogens (PREPARE).
The role of HT in treating established dementia is less optimistic. To date, two RCT of oestrogen therapy in women with Alzheimer's disease have been reported and both showed no benefit.17 Further research is required in this area before firm conclusions can be derived.
Quality of life
The WHI study findings of the effects of CEE plus MPA on health-related QOL have recently been reported. There was a statistically significant but not clinically relevant improvement in sleep disturbance, physical functioning and bodily pain at 1-year follow up.18 The findings were no longer significant at the 3-year follow up. There was no significant improvement in general health, limitations (either physical or emotional) on usual role-related activity, vitality, social functioning, mental health, depressive symptoms or sexual satisfaction. However, it is important to note that the WHI study was not designed to test the effect of HT on menopausal symptoms, which plays a large role in QOL for many peri/postmenopausal women. The majority of women in the WHI study were 15 years postmenopause, only 12.0% had moderate or severe vasomotor symptoms and the effect on urogenital symptoms was not assessed. Women with vasomotor symptoms were actively discouraged from entering the study. If symptomatic, these were unlikely to be disabling as the women were willing to be randomly assigned to placebo. Among the women with vasomotor symptoms at 1-year follow up, 76.7% of HT users had improvement compared with 51.7% in the placebo group (P < 0.001). At 3-year follow up, 71% of HT users had improvement compared with 52.8% in the placebo group (P < 0.0001), thereby improving QOL. However, this was not included as a QOL measure in the WHI study. Furthermore, treatment for vasomotor symptoms is usually required only short-term in women younger than those in the WHI study. Therefore, the risks associated with 5 years of HT use in older women seen in the WHI study should not deter the use of HT for short-term relief of debilitating symptoms.
The increased risk of endometrial cancer with unopposed oestrogens in women who have not had a hysterectomy is well recognised and this practice is not recommended.19
The lifetime risk of developing breast cancer in Australia is approximately 1 in 12 and increases with age, with the greatest incidence being in women over the age of 60 years. Hence, breast cancer is a disease that mostly affects women in their postmenopausal years when they are in a state of relative oestrogen deficiency.
In a large meta-analysis evaluating over 52 000 women with breast cancer and 108 000 controls, no significant increase in breast cancer risk was observed for women using HT for less than 5 years.20 For women who used oestrogen with or without a progestin beyond 5 years, a relative risk (RR) of 1.35 was reported.
The WHI trial comparing the effects of 0.625 mg CEE plus 2.5 mg MPA with placebo was prematurely terminated after 5.2 years average follow up because the statistical boundary for a possible increase in breast cancer risk had been reached.8 Overall, women treated with CEE/MPA had a 1.26-fold greater risk of invasive breast cancer than women treated with placebo. This translates to an absolute risk of eight extra cases per 10 000 women per year. However, 12 305 women in the study had not used HT prior to commencing the study and for these women the risk of breast cancer was not increased (RR 1.05). Thus, the findings from the WHI trial are consistent with the epidemiological findings cited above.
It remains unclear whether the small but apparent increase in risk after 5 years of HT is a result of oestrogen alone or a result of co-prescription of a progestin and the extent to which this risk reflects the mode of oestrogen administration. Several studies now indicate that the concurrent use of a progestin may confer a significantly greater risk of breast cancer than the use of oestrogen alone.21−23 In the recent Million Women Study, a cohort study of over a million women aged 50−64 years, current users of HT at recruitment were more likely than those who never used HT to develop breast cancer.24 The relative risk was 2.00 for combined HT users and 1.30 for users of oestrogen only preparations. The risk of breast cancer increased with increasing duration of use of HT, however, past users of HT were not at increased risk. Results from the continued oestrogen-only arm of the WHI study in hysterectomised women may clarify this issue.
Counterintuitively, a family history of breast cancer does not appear to increase the risk of developing breast cancer with HT use.8 Furthermore, mortality as a result of breast cancer is not increased with HT use25 and observational studies have found no evidence that HT use after breast cancer increases recurrence or breast cancer mortality.26
The risk of acute coronary events for women increases exponentially with age and following the menopause. Women who have undergone a premature surgical menopause have double the risk of coronary artery disease compared with those who have undergone natural menopause.27 However, the extent to which oestrogen insufficiency contributes to this increased risk is uncertain. The role of HT for the primary prevention of cardiovascular disease (CVD) is controversial; however, most would agree that HT has no role in secondary CVD prevention.
There is strong epidemiological evidence that unopposed oestrogen may reduce CVD risk in menopausal women without established disease.27 This appears to be directly attributable to a beneficial change in plasma lipid profile with a reduction in low-density lipoprotein and lipoprotein(a) and an increase in high-density lipoprotein28 along with favourable effects of oestrogen on endothelial function.29 When progestin is combined with oestrogen, the beneficial effect on lipid profile is diminished.19 However, in epidemiological research, concurrent progestin use does not appear to attenuate the beneficial effects of oestrogen on the cardiovascular system.30
In the WHI study, the absolute risk of an acute myocardial event was increased by seven cases per 10 000 women per year in the HT users in the earlier years of use.8 Sixty-six per cent of women in this study were over 60 years of age at commencement, almost 70% were overweight, 36% had treated hypertension, 13% had dyslipidaemia and 4.4% had diabetes mellitus. Thus, although most of these women had no history of CVD, many had pre-existing known cardiovascular risk factors and were not truly indicative of ‘healthy post menopausal women.’
Once adverse cardiovascular conditions have developed beyond menopause, the benefits of oral oestrogen therapy are greatly diminished such that the potential procoagulant effects appear to outweigh favourable lipid effects. In a study of hypercholesterolaemic postmenopausal women with poor endothelial function, presumably secondary to their hyperlipidaemia, treatment with oral oestrogen did not result in normalisation of endothelial function.31 This contrasts normalisation of impaired endothelial function with oestrogen therapy in otherwise healthy postmenopausal women.32
In the heart and oestrogen/progestin replacement study (HERS), a RCT, CEE 0.625 mg/day plus MPA 2.5 mg/day did not show any benefit over placebo in women with established coronary artery disease.33 As with the WHI trial, there were more events in the first year of HT use, which was attenuated over time. The subjects were at high risk, with a mean age of 67 years; 20% were diabetic, more than 55% were significantly overweight, 13% were smokers and 10% had congestive cardiac failure.
In contrast, the oestrogen in the prevention of re-infarction trial has shown neither increased risk nor protection against cardiovascular events for 2 years after a primary event with oestradiol valerate (2 mg) compared with placebo.34 In this study, treatment was initiated within weeks of a first myocardial infarct. The results of the ongoing WHI oestrogen-only arm may help to further clarify the role of progestins.
The Papworth study was the first RCT to evaluate the effects of transdermal oestrogen therapy in women with established coronary heart disease (CHD).35 The primary outcome measures were unstable angina, myocardial infarction or cardiac death. During the first 2 years of follow up, the HT group had a higher, but not statistically significant, event rate than the placebo group. Thus, despite the different mode of administration there appears to be no beneficial effect of transdermal oestrogen in established CHD.
In summary, the use of HT for primary and secondary prevention of atherosclerotic vascular disease should be strongly discouraged. The issue of risk of adverse cardiovascular events in typical HT users who are younger peri/postmenopausal woman being treated for menopausal symptoms is not known. However, CVD events are very uncommon in otherwise healthy women before 60 years of age and any small increase in risk in this otherwise low risk group is unlikely to translate into any clinically significant absolute increase in events.
Published epidemiological studies have indicated that users of oestrogen therapy are at reduced risk of stroke. However, in the WHI study continuous combined CEE plus MPA was associated with a 1.41-fold increase in risk of stroke in women with a mean age of 63 years, equivalent to an absolute increase of eight cases per 10 000 women per year.8
In the women's oestrogen and stroke trial of 664 postmenopausal women commenced on 1 mg 17-β oestradiol or placebo within weeks of a stroke, there was an increase in strokes during the first 6 months and strokes tended to be more severe in the oestradiol group.36 In HERS, combined continuous HT had no effect on stroke risk in postmenopausal women.33 Clearly, the risk of stroke is increased in older women and HT should not be commenced for primary or secondary prevention.
For both HERS33 and the WHI study,8 the most concerning outcome was a two to threefold increase in venous thromboembolic events (VTE), which was attenuated over time in HERS II.37 Of note in both protocols, HT was not ceased if immobilisation occurred. In HERS, nearly all VTE occurred following major surgery or hospitalisation. The data for the WHI study have not been released. Clearly women should cease oral HT before major surgery or during an immobilising illness. Of significance is the decreased risk of VTE noted with aspirin or statin use in HERS, suggestive of a protective effect of both therapies against VTE for HT users.33
Potentially, the increased risk of VTE may be focused in those with a predisposing genetic susceptibility for thrombosis; however, further research is needed to assess whether these factors are identifiable and which women are most at risk.
The routine screening of women for haemostatic abnormalities prior to commencing HT is not recommended, but a thorough history of family or personal experience of VTE or recurrent abortion is essential to screen specific women for known mutations.
The American Cancer Society's Cancer Prevention Study II, a prospective US cohort study with mortality follow up in 211 581 postmenopausal women, noted a small increase in ovarian cancer mortality with 10 or more years of oestrogen therapy.38 Ovarian cancer death rates were 6.4 and 3.8 per 10 000 women for baseline and former users of oestrogen therapy for 10 or more years, respectively, and 2.6 per 10 000 women for those who have never used oestrogen therapy. This study relied on recall of prior HT use; data were not available on type of HT use and only a small proportion of women had used HT for more than 10 years.
In a recent cohort study of 44 241 postmenopausal women, there was a slight increase in absolute risk of ovarian cancer in women using oestrogen therapy for 10 or more years compared to non-users, of two more cases per 10 000 woman years of HT use.39 This increased risk was predominantly confined to women with a prior hysterectomy as these women were more likely to be prescribed long-term unopposed oestrogen therapy. However, the role of hysterectomy in the development of ovarian cancer is unclear. This study had important methodological limitations; results were presented as a cumulative of total person years of oestrogen use and were not stratified into the number of women in each group of long-term, short-term and non-users of oestrogen therapy. It also relied on recall of prior HT use.
Overall, the absolute risk of ovarian cancer in HT users remains very small and may be isolated to long-term unopposed oestrogen therapy. In addition, many of the women in these observational cohort studies were likely to be on higher oestrogen doses than what is currently recommended. Additional data are required to clarify the risk of ovarian cancer with short- and long-term combined HT use.
Gall bladder disease
In HERS, oral HT users had an almost twofold increase risk of gall bladder disease compared with non-users.33
The issues surrounding HT are complex because of conflicting data pertaining to the proposed benefits and risks, the diversity of doses, formulations and modes of administration of oestrogens and progestins available and the uncertainty of the role of progestins.
The results of the WHI study and HERS have generated much concern and controversy with regards to managing the menopause, among both women and their doctors. However, both of these studies investigated the long-term risk of HT in a particular subset of women, predominantly asymptomatic, elderly and obese, that is not truly representative of the women for which we prescribe HT. In addition, they studied only one of the many HT doses and regimens available for women. The results need to be considered in terms of absolute risk and within the clinical setting of an individual's menopausal symptoms, QOL and well-being, as well as her risk profile for long-term illnesses.
Based on available information of the risks/benefits, HT should be predominantly prescribed short-term for the relief of menopausal symptoms and in selected women for the prevention of osteoporotic fractures.
After 2−5 years of treatment, each woman should, in discussion with her treating physician, consider the pros and cons of continuing therapy.