Abstract
Much new information on menopausal hormone therapy (MHT) has become available since the publication of the first report of the Women's Health Initiative in 2002 and a consensus is now emerging. The risk of breast cancer depends on type of MHT, duration of use, body mass, breast density and interval between menopause and starting MHT. The risk of breast cancer is generally increased by MHT, particularly in lean women with no previous MHT who start estrogen–progestin therapy near the menopause and continue for several years, but there is no increased risk 5 years after stopping MHT. The risks of venous thrombo-embolic disease (VTE), stroke and coronary heart disease (CHD) depend on age on starting MHT, dose, nature and route of administration of MHT. The risk of VTE is increased in women over 60 years of age and in women who are obese or have had a VTE but may not be increased by transdermal estrogens. The risk of stroke is very small in women under 60 years of age and may not be increased by low dose oral and low dose transdermal estrogen, MHT is of benefit in preventing atherosclerosis and CHD in healthy younger postmenopausal women (under the age of 60) but is not of benefit, and may be harmful, in older women with clinical or subclinical atherosclerosis. MHT prevents bone loss and osteoporosis but is not generally recommended in women over 60 because of the risks of VTE, stroke and CHD, and if stopped at 60 years does not prevent fractures in later life. MHT reduces the overall mortality in women under 60. MHT is by far the most effective treatment, and greatly improves the quality of life in women with menopausal symptoms. A simplified approach to MHT is suggested as a framework for the care of women at and after the menopause.
In the last decade the various outcomes of the Women's Health Initiative (WHI) randomized controlled trial of estrogen (conjugated equine estrogens [CEE]) plus a progestin (medroxyprogesterone acetate [MPA]) and of CEE-only in postmenopausal women in the USA have been extensively analyzed and have led to a better understanding of the risks and benefits of hormone therapy at and after the menopause.
The CEE + MPA arm of the WHI was terminated after a mean 5.2 years because of a significant increase in the risk of breast cancer, coronary heart disease (CHD), stroke and deep vein thrombosis, although the risks of hip fracture and colorectal cancer were reduced [1]. The CEE-only arm of the WHI in hysterectomized women was terminated after a mean 6.8 years because of a significant increase in the risk of stroke, although the risks of breast cancer and CHD were not increased and the risk of hip fracture was reduced [2]. The women recruited in the WHI had a significant number of medical risk factors and women in the CEE-only arm had a higher percentage of risk factors than those in the CEE + MPA arm, namely BMI >30 (45 vs 34%), treatment for hypertension (48 vs 35%) and treatment for diabetes (7.7 vs 4.4%). In the CEE-only arm 70% of the women were age 60 years or more and in the CEE + MPA arm, 66% women were age 60 years or more at the time of recruitment. The women in the WHI were on average older and had more risk factors than those in the general population who usually start hormone therapy at the time of the menopause and are under 60 years of age. The publication of the principal results of the CEE + MPA arm in July 2002 led to a substantial fall in the use of hormone therapy in menopausal women worldwide. The principal results of the CEE-only arm were published in April 2004 and were largely ignored by the public. The further analyses of the WHI include investigations on the effects of duration of therapy, prior hormone therapy and the time gap between onset of the menopause and initiation of therapy. The health outcomes after stopping therapy in both the CEE + MPA and CEE-only arms have also been analyzed. A consensus is now emerging on the benefits and risks of postmenopausal hormone therapy (please refer to the Endocrine Society scientific statement 2010, 2010 Position Statement of the North American Menopause Society, International Menopause Recommendation 2011 and The Public Assessment Report of the UK Medicines and Healthcare Products Regulatory Agency 2011 [3–5,101]).
Much valuable information has become-available in the past decade from studies in addition to those emanating from the WHI, including the Million Women Study (MWS) of breast cancer in the UK and the California Teachers Study (CTS) in the USA [6,7]. This review endeavors to summarize current understanding in light of the evidence currently available and to offer a possible explanation of the findings. A simplified approach to the care of women at and after the menopause is suggested as a framework for the care of women at and after the menopause (
Throughout this paper, 95% CIs are included within brackets.
Symptoms of the menopause
The majority of women who seek advice and treatment at or around the menopause do so because of vasomotor symptoms. Vasomotor symptoms – hot flushes, night sweats and sleep disturbances – can be very distressing, sometimes incapacitating and may substantially affect quality of life. With appropriate adjustment of dose and route of administration, the great majority of women can be completely relieved of symptoms by menopausal hormone therapy (MHT). Vasomotor symptoms are subjective and difficult to quantify, and it is difficult to balance the immediate benefit of relief of often distressing symptoms against the long-term risks of breast cancer, cardiovascular and thromboembolic diseases.
The WHI questionnaires on well-being were completed at baseline and after 1 year [8]. Women aged 50–54 years who had moderate-to-severe vasomotor symptoms showed improvement and overall there was a small but significant benefit of MHT on health-related quality of life. In the randomized controlled WISDOM trial, 3721 women aged 50—69 years were recruited from general practices in the UK, Australia and New Zealand and were treated with CEE 0.625 mg plus either MPA 2.5 or 5 mg daily or a placebo [9]. The quality of life was measured by a nine-component questionnaire at recruitment and after 1 year. Therapy was discontinued in 271 women because of vaginal bleeding or breast tenderness, leaving 1043 and 1087 women, respectively, in the treated and placebo arms. After 1 year there was a small but highly significant improvement in vasomotor symptoms, insomnia, joint and muscle pains, vaginal dryness and sexual functioning.
The menopause is associated with atrophic changes in the urogenital tract, particularly the vulva and vagina, giving rise to vaginal dryness, dyspareunia and problems with intercourse, sexuality and personal relationships. These changes often respond to treatment with estrogens locally. Changes in the lower urinary tract may be associated with urge incontinence, stress incontinence and recurrent urinary tract infections in postmenopausal women. The extent to which urological changes are due to the menopause or aging and the use of estrogens in the treatment of urinary symptoms, however, remains controversial [10,11].
Vasomotor symptoms usually start at or around the time of the menopause and may last for several years. The severity of symptoms decreases with time and the majority of women requiring MHT for relief of menopausal symptoms are under the age of 60 when the risks of MHT are small. A recent study found that the most common age of onset of moderate-to-severe hot flushes was 45–49 years of age and that the mean duration was 10.2 years, exceeding the time frame generally accepted in clinical practice [12]. The occurrence of hot flushes may affect the clinical outcomes of the menopause including breast cancer, cardiovascular disease (CVD) and osteoporosis and these relationships need further investigation [13–15]. MHT is by far the most effective treatment for menopausal symptoms, but in older postmenopausal women MHT may carry risks as well as benefits. For many women, however, irrespective of age, the benefit of the relief of menopausal symptoms often outweighs the possible risks of MHT even when these have been carefully explained to them.
Breast cancer
Breast cancer is the most common invasive cancer in women, with a lifetime risk of approximately one in eight and a risk of dying from the disease of one in 200 in the USA [16,102]. Many studies have shown that the risk of breast cancer is increased by MHT and also by many other factors including a late menopause, occurrence of benign breast disease (BBD) and family history of breast cancer [17].
In the 1997 Collaborative Group reanalysis of 51 studies of 52,705 women with breast cancer and 108,441 women without, the relative risk (RR) in women who had used MHT for 5 years or more was RR: 1.35 (1.21–1.59). After stopping MHT, the increased risk was decreased and had disappeared after 5 years [18]. In the Million Women Study the risk was greater in estrogen–progestin users than in all other types of MHT but there was no increased risk in past MHT users [19].
In the CEE + MPA arm of the WHI, the risk of breast cancer was increased with RR: 1.26 (1.00–1.59), but in the CEE arm, RR was decreased to 0.88 (0.62–1.04) [1,2].
In an autopsy study of 110 young or middle-aged women who had died accidentally, small occult neoplastic breast tumors were found in 39% of the women [20]. It is hypothesized that a small but increasing proportion of these occult cancers become invasive over time. It is suggested that the administration of estrogen and progestin in postmenopausal women accelerates the growth of pre-existing occult tumors but does not initiate malignant change. This would account for the fact that the incidence of breast cancer is increased in current users of estrogen and estrogen–progestin but is not increased in past users. The increased incidence of breast cancer in overweight and obese women has been attributed to the increased endogenous production of estrogens in fat and muscle [21]. In overweight and obese women, the incidence of breast cancer approaches a maximum and the incidence of breast cancer is, in general, not increased further by the administration of exogenous estrogens. Kuhl and Stevenson have suggested that the increased production of estrogens in overweight and obese women is insufficient to account for the increased incidence of breast cancer [22]. They suggest that the increased incidence is due to the elevated insulin and IGF-levels in overweight and obese women associated with the increased incidence of the metabolic syndrome and hyperinsulinemia. Elevated levels of insulin and IGFs may stimulate tumor growth and the administration of estrogen decreases the levels of insulin and the risk of cancer [23,24]. This may explain the decreased incidence of breast cancer in the CEE-only arm of the WHI. The addition of MPA may reverse the beneficial effect of estrogen on insulin levels because of its glucocorticoid activity and increase the risk of breast cancer as with the CEE + MPA arm of the WHI.
The risk of breast cancer with MHT depends on a number of factors, including:
Type of MHT
Duration of use
Incidence of mammography and breast density
Body weight and BMI
Interval between menopause and starting of MHT
Previous MHT use
The effect of stopping MHT
Type & route of administration of MHT
In the Collaborative Reanalysis of 51 Epidemiological Studies of breast cancer there was no marked variation in the results according to hormone type or dosage of MHT, but 80% of the women had used oral estrogen-only formulations [18].
In the MWS, the risk of breast cancer was RR: 2.00 (1.88–2.13) for estrogen-progestin users and RR: 1.30 (1.21–1.45) for estrogen-only users compared with never users [19]. There was little difference in results between different estrogens and different progestins and between sequential and continuous combined regimes. There was also no significant difference between different routes of administration. The risks of breast cancer with oral, transdermal and implant routes of administration of estrogen-only were, respectively, RR: 1.32 (1.21–1.45), RR: 1.24 (1.11–1.39) and RR: 1.65 (1.26–2.16), compared with never users. The risk of breast cancer in estrogen–progestin users was substantially greater than in all other types of MHT. The number of additional cases of breast cancer in estrogen–progestin users was calculated as five to six cases per 1000 women for 5 years use and 15–19 cases per 1000 women for 10 years use.
In the WHI there a was marked difference between the CEE + MPA arm where the risk of breast cancer was significantly increased with RR: 1.26 (1.00–1.59) and the CEE-only arm where the risk of breast cancer was decreased, RR: 0.77 (0.59–1.01) [1,2].
Duration of use
In virtually all the studies of MHT, with the exception of the CEE-only arm of the WHI, the cumulative incidence of breast cancer increased progressively with increasing duration of use.
In the Collaborative Analysis of 51 studies, the risk of breast cancer increased by RR: 1.023 (1.011–1.036) per year of use. In women who had used MHT for 5 years or more, the risk was RR: 1.35 (1.21–1.49) and the risk was approximately the same as a delay in the menopause for 5 years [18].
In the MWS, the risk of breast cancer in current MHT users compared with never users increased progressively with increasing duration of use (
Relative risk of breast cancer by duration of use.
Adjusted for age, time since menopause, parity, age at first birth, BMI and family history of breast cancer.
RR: Relative risk.
Data taken from [19].
In the intervention phase (mean 5.2 years) of the CEE + MPA arm of the WHI, the incidence of invasive breast cancers increased progressively and was 0.43% cases per year in the treated group and 0.33% cases per year in the placebo group. Postintervention (mean 3 years), there was a downward trend in the cumulative risk in the treated group but it was not statistically significant [25].
In the intervention phase (mean 7.1 years) of the CEE-only arm, the incidence of breast cancer was 0.28% cases per year in the treated group and 0.35% cases per year in the placebo group [26]. Postintervention (mean 4 years), the incidence of invasive breast cancer was 0.26% cases per year in the treated group and 0.34% cases per year in the placebo group. The risk of invasive breast cancer was reduced during and after treatment in women in the CEE-only arm with RR: 0.79 (0.61–1.02) and RR: 0.75 (0.51–1.09), respectively, and the reduction in risk persisted over 7 years after stopping therapy.
Mammography & effect of MHT
Mammographic screening reduces the mortality due to breast cancer by approximately 15% compared with nonscreened populations, according to the US Preventive Services Task Force Report of 2009, and by up to 35% in a recent community study in Nijmegen in the Netherlands [27,28]. Clinical breast examination and mammography are both important in the early detection of breast cancer. Clinical breast examination has a sensitivity of 40–69% and a specificity of 88–99%, and mammography has a sensitivity of 77–95% and a specificity of 94–97% [28]. MHT increases breast density and reduces both the sensitivity and specificity of mammography. In an ancillary study of the WHI, 413 women assigned active treatment and 211 who received placebo had mammograms that were digitalized to calculate the degree of density. After 1 year the mean density in women was decreased by 0.9% in the placebo group and increased by 1.6% in users of CEE-only and by 6% in users of CEE + MPA [29,30]. The percentage of women with abnormal mammograms in the CEE + MPA arm of the WHI trial was higher than in the placebo group (35.0 vs 23.0%) and resulted in more than one in ten having otherwise avoidable mammograms and one in 25 having avoidable breast biopsies [31]. The breast cancers in the CEE + MPA group were larger and more advanced than in the placebo group [31]. In the CEE-only arm, the number of cases of breast cancer was too small to assess the clinical significance of mammography. The breast density in both arms of the WHI decreased on stopping therapy. The decrease on short-term suspension of active therapy (1–2 months), however, was small and did not reduce recall rates [32].
In a study of 587,369 women who had a total of 1,349,027 mammograms in California, 14,090 women were diagnosed with breast cancer [33]. Breast cancer risk was low in women aged 55–59 years with low breast density (Breast Imaging Reporting and Data System-1). The 5-year breast cancer risks were 0.8% (0.6–0.9%) in non-MHT users and 0.9% (0.7–1.1%) in MHT users. In women aged 55–59 years with high breast density (Breast Imaging Reporting and Data System-4), the 5-year risk of breast cancer was 2.4% (2.0–2.8%) for non-MHT users, 3% (2.6–2.8%) for estrogen-only users and 4.2% (3.7–4.6%) for estrogen plus progestin users.
Breast density on mammography is increased by MHT, particularly in women treated with estrogen and progestin, and is a risk factor for breast cancer as it delays diagnosis and the cancers are diagnosed at a more advanced stage.
Body weight & BMI
Many studies have found that the incidence of breast cancer increases with increasing body weight and increasing BMI.
In the WHI, 80% of the women were overweight or obese and no significant associations were found in the risk of breast cancer with increasing BMI. In the CEE + MPA arm, the risks for breast cancer for BMI <25, 25–29 and >30 were, respectively, RR: 1.29 (0.94–1.77), RR: 1.34 (1.04–1.75) and RR: 1.14 (0.90–1.44) [34]. In the CEE-only arm, the risks of breast cancer for BMI <25, 25–29 and >30 were RR: 0.94 (0.49–1.76), RR: 0.61 (0.37–1.03) and RR: 0.89 (0.63–1.25), respectively [34,35].
In the collaborative reanalysis of 51 studies, the incidence of breast cancer in postmenopausal women was increased by 3.1% per kg/m2 of BMI [18]. In women with a duration of MHT of >5 years and last use <5 years, the risks of breast cancer in women with a weight of <65 kg and >65 kg were, respectively, RR: 1.65 (standard error [SE]: 0.065) and RR: 1.06 (SE: 0.103). The risks of breast cancer in the same women with a BMI of <25.0 kg/m2 and >25.0 kg/m2 were RR: 1.52 (SE: 0.083) and RR: 1.02 (SE: 0.103).
In the MWS, the risks of breast cancer in current estrogen plus progestin users in women with a BMI of <25 kg/m2 or >25 kg/m2 were, respectively, RR: 2.51 (2.12–2.53) and RR: 1.78 (1.64–1.96). In current estrogen-only users in women with a BMI of <25 kg/m2 or >25 kg/m2, respectively, the RR was 1.53 (1.05–1.2) and 1.17 (1.05–1.20) [19]. The increased risk of breast cancer associated with MHT was significantly greater in thin women with a BMI <25 kg/m2, than in overweight and obese women with a BMI >25 kg/m2 [102]. In both the collaborative reanalysis of 51 studies and the MWS, the risk of breast cancer was increased in overweight and obese women but was not increased further by MHT.
Effect of time from menopause to initiation of MHT: ‘gap time’
In the CEE + MPA arm of the WHI, the risk of breast cancer in women who started therapy within 5 years of the menopause was greater than in women who started therapy 5 years or more after the menopause, with RR: 1.41 versus 1.15 (p = 0.08)
Relative risk of breast cancer by years before initiation of menopausal hormone therapy.
MHT: Menopausal hormone therapy; MWS: Million Women Study; RR: Relative risk; WHI: Women's Health Initiative.
In the CEE-only arm, the risk of breast cancer in women who started therapy within 5 years of the menopause was greater (RR: 1.06; 0.74–1.51) than in those who started therapy 5 years or more after the menopause in whom the risk was reduced (RR: 0.63; 0.42–0.93)
In the MWS, the risk of breast cancer in current users of estrogen plus progestin who started therapy within 5 years of the menopause was RR: 2.04 (1.95–2.14) and was significantly greater than in those who started therapy 5 years or more after the menopause (RR: 1.53; 1.38–1.70). The risk of breast cancer in users of estrogen only who started therapy within 5 years of the menopause was RR: 1.43 (1.35–1.51) and was greater than in those who started therapy 5 years or more after the menopause (RR: 1.05; 0.89–1.24)
There is evidence that a period of estrogen deprivation after the menopause sensitizes breast cancer cells to the proapoptotic effect of estrogens and reduces the subsequent incidence of breast cancer. This may account for the lower risk of breast cancer in women who start MHT after a gap of ≥5 years after the menopause [38].
Time from the menopause to starting MHT may be an important factor affecting breast cancer risk. Most women, however, start MHT either before or at the menopause for the relief of vasomotor symptoms.
Effect of prior MHT
In the CEE + MPA arm of the WHI, 26% of women had a history of prior MHT and in the CEE-only arm, 50.6% had a history of prior MHT. All women recruited to the WHI who were currently taking MHT had a 3-month treatment-free ‘washout’ period before starting the trial. Prior MHT had significant effects on both the absolute incidence and the RRs of invasive breast cancer in the CEE + MPA
Women's Health Initiative: conjugated equine estrogens + medroxyprogesterone acetate invasive breast cancer annualized incidence (%) and relative risk by prior menopausal hormone therapy.
p for interaction 0.027.
Adjusted for age, race/ethnicity, BMI, physical activity, smoking, alcohol use, parity, oral contraceptive use, family history of breast cancer, family history of fracture, mammography and presence of moderate-to-severe symptoms.
CEE: Conjugated equine estrogens; MHT: Menopausal hormone therapy; MPA: Medroxyprogesterone acetate
RR: Relative risk.
Data taken from [39].
*CEE + MPA
In the CEE + MPA arm, the RR of breast cancer was increased in women with prior MHT (RR: 1.96; 1.17–3.27) but not in those with no prior MHT (RR: 1.02; 0.77–1.36). In the women in the CEE + MPA arm who received placebo, the annualized incidence of breast cancer was lower in women with prior MHT than the incidence in women with no prior MHT (0.25 vs 0.36%). With prior MHT, the RR of breast cancer was significantly increased due to the low incidence of breast cancer in the placebo group and not to an increased incidence of breast cancer in the group treated with CEE + MPA.
The Kaplan–Meier estimates showed that in women with prior MHT, the cumulative incidence in women treated with CEE + MPA increased progressively but that the cumulative incidence of breast cancer in women who received placebo remained relatively low throughout the years of follow-up
In women with no prior MHT in the CEE + MPA arm, the annualized incidence in women treated with CEE + MPA was similar to the annualized incidence of breast cancer who received placebo (0.41 vs 0.36%) and the RR of breast cancer with CEE + MPA was not increased (RR: 1.02; 0.77–1.36).
In the Kaplan–Meier estimates, the cumulative incidence of breast cancer in women with prior MHT treated with CEE + MPA increased at the same rate as in women with prior MHT who received placebo during the first 4 years of the study
CEE-only
In the CEE-only arm, the RR of breast cancer was reduced in women with no prior MHT (RR: 0.65; 0.46–0.92) but not in women with prior MHT (RR: 1.02; 0.70–1.50) [34].
In women with prior MHT in the CEE-only arm, the annualized incidence of breast cancer was 0.29% in those treated with CEE-only and was 0.28% in women who received placebo, and the RR of breast cancer of women was not increased or decreased (RR: 1.02; 0.70–1.50).
In women with no prior MHT in the CEE-only arm, the annualized incidence of breast cancer with CEE was 0.27% and with placebo was 0.40%, and the RR of breast cancer in women was significantly reduced with RR: 0.65 (0.45–0.92).
In the Kaplan-Meier estimates in the CEE-only arm of the WHI, the cumulative incidence of breast cancer was reduced in all women with prior MHT (both treated and placebo) and in women with no prior MHT treated with CEE-only. The only group in which the cumulative incidence of breast cancer was not reduced was women with no prior MHT who received placebo [34]. All the women who had been previously treated with MHT and also those who were treated with CEE-only during the WHI, had a reduced incidence of breast cancer.
In the WHI, all women had a mammography before recruitment and all women with a history of breast cancer were excluded. The relatively low incidence of breast cancer in women with prior MHT in both arms of the WHI may be explained by the exclusion of the women who had already developed breast cancer. It is suggested that in women with prior MHT, an increased proportion of those with occult breast tumors sensitive to estrogens and progestins would have developed invasive cancers before recruitment to the WHI and these women would have been excluded. This would leave a deficit in the number of women with occult breast tumors sensitive to estrogens and progestins, and a consequent reduction in the subsequent incidence of breast cancer in these women. In women with no prior MHT, the proportion of women with occult tumors sensitive to estrogens and progestins would not be reduced and the incidence of breast cancer would then be relatively increased by MHT. This may explain the reduced incidence of breast cancer with prior MHT in the WHI.
The interaction between ‘gap time’ and prior MHT on the risk of breast cancer in the WHI has been examined in two studies
Prior menopausal hormone therapy and ‘gap time’ for women treated with conjugated equine estrogens + medroxyprogesterone acetate.
RR for interaction, no vs yes: p =0.53.
RR for interaction <5 years vs >5 years: p = 0.02.
MHT: Menopausal hormone therapy; RR: Relative risk.
Data taken from [40].
Prior menopausal hormone therapy and ‘gap time’ for women treated with conjugated equine estrogens only.
RR for interaction <5 years vs >5 years: p = 0.20.
MHT: Menopausal hormone therapy; RR: Relative risk.
Data taken from [41].
In women treated with CEE + MPA, the risk of breast cancer was significantly increased if commenced within 5 years of the menopause, but was not increased after a ‘gap time’ of 5 years before commencement [40].
In women who received CEE only with no prior MHT, the RR of breast cancer was significantly decreased in women who started therapy more than 5 years from the menopause, but was not decreased in women who started therapy within 5 years of the menopause [41].
‘Gap time’ and prior MHT appear to have separate independent effects on the risk of breast cancer in the CEE + MPA and CEE-only arms of the WHI.
Effect of cessation of MHT
In both the WHI and MWS, the risk of breast cancer declined towards that of never users within 2–3 years of stopping MHT [25,26]. In all the studies of breast cancer, there is no persisting increase in risk of breast cancer in past users ≥5 years after stopping MHT This finding suggests that MHT increases the proportion of occult tumors that become invasive but does not initiate malignant change. In the CEE-only arm in the WHI, the reduction in the overall risk of breast cancer persisted over a mean 10.7 years of follow-up and was significantly lower than in the placebo group (RR: 0.77; 0.62–0.95) [42]. The persisting reduction of the incidence of breast cancer after stopping MHT in the CEE-only arm of the WHI accords with the hypothesis that estrogens given after a period of estrogen deprivation are proapoptotic and reduce the number of occult breast tumors that subsequently become invasive.
In summary, the risk of breast cancer is significantly increased with MHT particularly in lean women without prior MHT who start estrogen–progestin therapy near the menopause and continue therapy for several years. The increased risk of breast cancer decreases on stopping MHT and after 5 years, the risk is the same as in never users. In overweight and obese women, the risk of breast cancer is increased but in general is not increased further by MHT. Regular mammography reduces the mortality from breast cancer by approximately 15%.
The risk of breast cancer depends upon many factors in addition to the effect of MHT, including current age, age at menarche, age at first pregnancy, age at menopause, parity, lactation, ethnicity, BMI, smoking, alcohol use, oral contraceptive use, BBD and family history of breast cancer. In the Nurses Health Study (NHS), analysis of 121,700 nurses aged 30–55 years at recruitment, the main predictors of breast cancer risk up to the age of 70 were BBD, family history of breast cancer and MHT
Endometrial cancer
Administration of unopposed estrogen (without added progestin) in peri- and postmenopausal women with an intact uterus results in an approximate 50% increase in the incidence of endometrial hyperplasia or carcinoma within 3 years [43]. The majority of the endometrial carcinomas are endometriod, well differentiated and hormonally responsive. The incidence of papillary, serous and clear cell endometrial carcinomas, which are usually poorly differentiated and not hormonally sensitive, is the same in women receiving MHT and in untreated women. Many trials, including the WHI, have reported that both continuous combined, as well as sequential estrogen–progestin therapy, protect against the increased risk of endometrial cancer associated with the use of unopposed estrogen in women with an intact uterus [1,2]. In a Finnish study of 224,015 women aged more than 50 years who had used estradiol–progestin therapy between 1994 and 2006, 1364 type I and 38 type II endometrial cancers were identified through the Finnish Cancer Registry [44]. The use of continuous combined estrogen–progestin therapy for more than 3 years was associated with a significant 76% reduction in type I cancers. By contrast, the use of sequential estrogen–progestin therapy for at least 5 years was associated with a significant 69% increase in incidence if progestin was added monthly and 276% if added trimonthly. This suggests that sequential MHT should be reserved for short-term use only and continuous combined MHT should be given for long-term use in all postmenopausal women. This potentially important finding urgently merits further investigation.
Ovarian cancer
All the major observational studies of MHT, including the NHS, the MWS and a recent Danish study, have found a small but significant increase in the incidence of ovarian cancer associated with estrogen and estrogen–progestin therapy in peri- and postmenopausal women [45]. Some studies have found differences between estrogen–progestin therapy and estrogen-only therapy and others between sequential and continuous combined therapy. Some studies have found an increase in incidence with increasing duration of MHT. Virtually all have found no increase in past users. In the randomized, controlled WHI trial of CEE + MPA, there was a small but nonsignificant increase in the RR of ovarian cancer (RR: 1.58; 0.77–3.24). In the large Danish study of 960,887 women with 3068 incident ovarian cancers, the absolute increase in incidence was 0.12 (0.1–0.17) per 1000 women years, approximating to one extra case for 8300 women receiving MHT per year and accounted for 5% of all cases of ovarian cancer in Denmark [45]. No differences were found between different hormone therapies or different routes of administration. The risk of ovarian cancer was not altered by duration of use and no increase was found 6 years after stopping MHT. The increased risk of ovarian cancer with MHT is very small and does not materially affect the balance of risk and benefits of MHT.
Colorectal cancer
In the WHI there was a significant decrease in the risk of colorectal cancer in the CEE + MPA arm (RR: 0.63; 0.46–0.92), but not in the CEE-only arm [1,2]. The cancers in the CEE + MPA arm, however, were more advanced with a greater number of positive nodes. In the 3 years follow-up after stopping CEE + MPA, the decreased risk of colorectal cancer was no longer significant. In the CEE-only arm, the risk of colorectal cancer was not decreased but was significantly increased postintervention with RR: 1.83 (1.08 −3.12) in women aged 70–79 years at entry [25]. In an 8-year follow-up of 118,404 nurses in the NHS, past use of estrogens postmenopausally was associated with a reduced risk of colorectal cancer (RR: 0.05; 0.3–1.0) [46]. In a review of epidemiological data on incidence and mortality of colon cancer, it was concluded that estrogen use reduced the incidence of colon adenoma and carcinoma by approximately 30% and that mortality of cancer-related mortality was significantly reduced [47]. The mechanisms by which estrogen and/or progestin may reduce the incidence of colon cancer are not clear, and in light of the findings of the WHI, the possible protective role of postmenopausal estrogen and progestin therapy needs further investigation.
Lung cancer
In the postintervention study of the CEE + MPA arm of the WHI, the mortality due to lung cancer was significantly increased, although the incidence of lung cancer was not increased. Lung cancer was a major factor in the increased mortality in the postintervention phase of women treated with CEE + MPA [25,48]. The increase in mortality was primarily in women with nonsmall cell lung and was confined to women over the age of 60. The absolute death rate was greater in smokers than in nonsmokers. There was no increase in lung cancer in the CEE-only arm of the WHI and the basis of the increased risk with CEE + MPA remains largely unexplained.
Cardiovascular disease
The main cardiovascular outcomes of the WHI were venous thromboembolic disease (VTE), stroke and CHD. In the CEE + MPA arm, the risks of VTE, stroke and CHD were all significantly increased; the risks were, respectively, RR: 2.11 (1.26–3.55), RR: 1.41 (0.86–2.31) and RR: 1.29 (0.85–1.97) [1]. In the CEE-only arm, the risks of VTE and stroke were increased, with RR: 1.33 (0.86–2.08) and RR: 1.39 (0.9–1.99), respectively, but the risk of CHD was not increased (RR: 0.91; 0.72–1.15) [2].
The analyses of the effects of MHT on CVD have focused on:
Age at initiation of MHT
Body weight
Route of hormone administration: oral versus transdermal
Nature and dose of hormone administered
The effect of the addition of progestin to estrogen
The effect of stopping MHT
VTE
In the further analysis of the CEE + MPA arm of the WHI, the risk of venous thrombosis increased progressively with age: age 50–59 years, RR: 1.27 (1.19–4.33); age 60–69 years, RR: 4.28 (2.38–7.72); and age 70–79 years, RR: 7.46 (4.32–14.38) [49]. The risk of venous thrombosis also increased progressively with increasing BMI: BMI <25, RR: 1.78 (0.91–3.51); BMI 25–30, RR: 3.80 (2.08–6.94); and BMI >30, RR: 5.6 (3.12–10.11). The presence of factor V Leiden further increased the risk of VTE (RR: 6.39; 3.09–14.49), but the risk of VTE was not increased by other genetic variants. After stopping CEE + MPA, the risk of VTE was the same as in never users [25].
In the CEE-only arm, there was a nonsignificant increase in the risk of thromboembolic disease (RR: 1.33; 0.91–2.12) and the risk was reduced postintervention [26].
Three meta-analyses have confirmed the increased risk of VTE associated with MHT but have reached differing conclusions on the effects of different types of MHT and the different routes of administration and on the risk of VTE in the first year of use.
In a meta-analysis of 31 randomized, controlled trials of estrogen plus progestin and estrogen-only, the overall risk of VTE was increased (RR: 2.05; 1.44–1.92). The risk of VTE was not significantly related to age but the risk of VTE with estrogen plus progestin was double that with estrogen-only [50].
In a meta-analysis of eight observational studies, the risk of VTE was increased in current oral estrogen users but not in current transdermal estrogen users with RR: 2.5; 1.9–3.4 vs RR: 1.2 (0.9–1.7). The risk of VTE was higher in the first year of use (RR: 4.0; 2.9–5.7) than after the first year of use (RR: 1.2; 1.3–3.8), but no difference was found between users of estrogen plus progestin and of estrogen only [51].
In a meta-analysis of nine randomized trials, the risk of VTE was increased in oral estrogen users (RR: 2.1; 1.4–3.10) and was further increased in obese women and in women with thrombogenic genetic mutations but was not increased in past users [51].
In the ESTHER multicenter case–control study of women aged 45–70 years, the risk of VTE was significantly greater in current oral estrogen users than in current transdermal estrogen users with RR: 4.2 (1.5–11.6) versus RR: 0.90 (0.4–2.1). There was also a fourfold increase in VTE with the addition of nonpregnane derivatives (RR: 3.9; 1.5–10.0) but no significant increase in risk of VTE with the addition of micronized progesterone or pregnane derivatives [52]. Kuhl and Stevens have put forward evidence that MPA (due to its glucocorticoid activity), but not other progestins, up regulates thrombin receptors and thrombin-induced production of tissue factor and increases procoagulant activity in blood vessel walls [22].
There is general agreement that oral estrogens increase the risk of VTE. It is suggested that administration of estrogens orally results in a ‘first pass effect’ on hepatic protein synthesis coinciding with the peak blood levels of estrogen 3–4 h after administration. This ‘first pass effect’ increases circulating procoagulant factors and impairs antithrombotic mechanisms, including a rise in activated protein C, resulting in an increased incidence of VTE. With transdermal administration, estrogen is released continuously and the peaks in estrogens levels are avoided and the consequent ‘first pass effects’ on hepatic metabolism are avoided or reduced.
In a recent ‘Green-top Guideline’ of the Royal College of Obstetricians and Gynecologists on ‘Venous Thromboembolism and Hormone Replacement Therapy’ based on 87 publications, it was concluded that “the available evidence suggests transdermal preparations are associated with a substantially lower risk of VTE than oral preparations” [103].
Stroke
Stroke is a leading cause of morbidity and mortality and the incidence increases significantly after the menopause and with increasing age.
In both the CEE + MPA and CEE-only arms of the WHI, the risks of ischemic stroke were increased with RR: 1.31 (1.02–1.68) and RR: 1.37 (1.09–1.73), respectively [1,2]. The risk of hemorrhagic stroke, however, was not increased. The increased risk of ischemic stroke was independent of age and of all basal risk factors and was not altered by current use of aspirin or statins [53]. In the postintervention phases of both the CEE + MPA arm (mean 3 year follow-up) and the CEE-only arm (mean 7.1 year follow-up), the risk of stroke was not increased [25,26].
In a meta-analysis of 28 trials, the risks in MHT users versus never users were:
Total stroke: 1.29 (1.13–1.47)
Nonfatal stroke: 1.23 (1.06–1.46)
Stroke leading to disability or death: 1.56 (1.11–2.26)
Fatal stroke: 1.28 (0.87–1.88)
Ischemic stroke: 1.29 (1.06–1.56)
MHT was not associated with an increased risk of hemorrhagic stroke (RR: 1.07; 0.65–1.75) or in transient ischemic attacks (RR: 1.02; 0.78–1.34) [54].
In another meta-analysis of 31 studies, the risks for all types of cerebrovascular disease was increased in MHT users compared with never users (RR: 1.24; 1.09–1.34), but there were no differences with different types of MHT [50].
In a third meta-analysis of seven randomized controls, the use of MHT was associated with an increased overall risk of stroke (RR: 1.29; 1.13–1.58) and there was no significant heterogeneity with any subgroups [55].
In the NHS, the risk of stroke was increased in current users of estrogen plus progestin (RR: 1.27; 1.04–1.56) and of estrogen-only (RR: 1.39; 1.18–1.63). The increase in risk was confined to ischemic stoke and the risk of hemorrhagic stroke was not increased. The increase in risk of stroke was not significantly influenced by age or by time since the menopause. There was, however, a strong association with estrogen dosage (p < 0.001) and there was no increased risk of stroke in women receiving a low dose of CEE (0.3 mg daily) compared with never users [56].
In a recent case–control study of women aged 50–79 years from the UK General Practice Research Base, the rate of stroke was 2.85 per 1000 women/year. The risk of stroke was increased with current use of both low and high estrogen dose oral MHT (RR: 1.28; 1.15–1.42). The risk was also increased with high estrogen dose transdermal patches (>50 μg/day; RR: 1.89 [1.15–3.11]) but was not significantly increased with low-dose estrogen transdermal patches (50 μg/day or less; RR: 0.81 [0.62–1.05]) [57].
MHT is associated with an increased risk of ischemic stroke but not of hemorrhagic stroke. There is also no persisting increase in risk after stopping MHT. The pattern of the increased risk of stroke with MHT is similar to that of VTE. It is hypothesized that the increased risk of stroke with MHT is related to the effect of oral estrogens in increasing procoagulant factors and the risk of thrombosis. The absolute incidence of stroke is very small in women under the age of 60, and appears not to be increased with low-dose oral estrogens (CEE 0.3 mg/day) and with low-dose estrogen patches (estradiol 50 μg/day or less).
CHD
Most but not all cohort, retrospective and prospective observational studies have demonstrated a significant 40–60% reduction in CHD and a reduction in mortality due to CHD in postmenopausal women taking MHT [58–60].
Observational studies, however, are liable to inherent bias, particularly the ‘healthy user effect’, and randomized controlled trials are necessary to exclude possible bias. The randomized controlled Heart and Estrogen/Progestin Replacement Study (HERS) was initiated to determine the effect of CEE + MPA in women with known CVD [61]. The study showed no significant decrease, or increase, in CHD but did show a significant increase in VTE. In the WHI randomized controlled trial, the risk of CHD was significantly increased in women treated with CEE + MPA (RR: 1.29; (1.02–1.63) but was decreased in those treated with CEE-only (RR: 0.91; 0.75–1.12) [1,2].
Several hypotheses have been advanced to account for the difference between the observational studies and the HERS and WHI trials including the effects of:
Age and time interval between the menopause and start of MHT
Health status at initiation of MHT
The effect of addition of MPA to estrogen in MHT
Age at menopause and the effect of bilateral oophorectomy
Age & time interval between the menopause & start of MHT
In both the CEE + MPA and CEE-only arms of the WHI, both age and the time interval between the menopause and the start of MHT had significant effects on the risk of CHD [61,63]. Although the number of women in the 50–59 age group was small, women under age 60 years appeared to have a lower risk of CHD than older women. Women who started treatment with CEE + MPA or CEE-only within 10 years of the menopause had a lower risk of CHD than women who started treatment later. In the postintervention phase of CEE + MPA the risk of CHD decreased to RR: 1.04 (0.89–1.21) over a mean of 3 years of follow-up [25]. In the CEE-only arm the risk of CHD was unchanged during therapy and over a mean 4 years follow-up with RR: 0.91 (0.75–1.12) and RR: 0.97 (0.75–10.9), respectively [26].
In the NHS of 70,533 postmenopausal women, including 1258 women with major coronary events, the risk of CHD was significantly lower in current users of oral CEE than in never users (RR: 0.61; 0.52–0.71) and the reduced risk was maintained through 10 years of use (Table 6). The risk for CHD was similar with CEE 0.3 mg/day (RR: 0.58; 0.37–0.92) and CEE 0.625 mg/day (RR: 0.54; 0.44–0.67) [64].
Relative risk of coronary heart disease by years of hormone use in the Nurses Health Study.
Adjusted for age, BMI, diabetes, hypertension, high cholesterol, age at menopause, cigarette smoking and parental history of premature coronary heart disease.
Data taken from [60].
In the NHS, the risks of CHD in women who started CEE + MPA or CEE-only near the menopause, were reduced relative to those who started therapy 10 years or more after the menopause [65].
In a pooled analysis of 23 randomized clinical trials, MHT reduced the risk of CHD in younger women (within 10 years of the menopause or under age 60 years; RR: 0.68 [0.48–0.9]), but not in older women (more than 10 years since the menopause or over age 60; RR: 1.03 [0.91–1.16]) [66].
The risk of CHD in the CTS of 71,327 women (mean age 63 years, range 36–94 years), the age of the current HRT users at the time of recruitment to the study was more important than the age at which they started therapy. In the CTS, the mortality from ischemic heart disease was markedly reduced by current use of MHT in younger women (under age 65 years; RR: 0.38 [0.22–0.67]), but had no effect in older women (age 75–84 years; RR: 0.93 [0.77–1.12]; p = 0.008). The formulation, duration and past use of MHT and the interval between the menopause and start of MHT had no effect on the risk of CHD [7].
Cardiovascular status on initiation of MHT
Many factors, in particular hypercholesterolemia, affect the occurrence and degree of atherosclerosis. In a nested control study from both arms of the WHI, blood lipids were measured in 271 patients with CHD and in 707 controls. Women with a low-density lipoprotein: high-density lipoprotein ratio of >2.5 had an increased risk of CHD (RR: 1.73; 1.18–2.53), but those with a low-density lipoprotein: high-density lipoprotein ratio of <2.5 had no increase in risk of CHD (RR: 0.60; 0.43–1.06) [67]. The baseline state of the cardiovascular system is an important factor in determining the effect of MHT on CHD in postmenopausal women.
Effect of addition of MPA to estrogen in MHT
In the WHI, the risk of CHD was increased in women treated with CEE + MPA but was not increased, and was possibly decreased, in women with CEE-only even though the women in the CEE-only arm had a greater number of cardiovascular risk factors. Addition of MPA does not produce any unfavorable effects on CEE-induced changes in various hemostatic factors. MPA however, has been shown to upregulate thrombin receptors in blood vessel walls and to activate extrinsic coagulation. Thrombin receptors are found in thrombocytes, vascular epithelium and smooth muscle cells and activation of the receptors by thrombin plays an important role in the development of thrombosis and atherosclerosis. Thrombin receptors are upregulated by low concentrations of glucocorticoids and by progestins with gluococorticoid activity such as MPA [68]. In cynomolgous monkeys, CEE prevented the vasoconstrictor effect of acetylcholine on healthy and atherosclerotic coronary arteries but the beneficial effect of CEE was negated by the addition of MPA [69]. The addition of MPA to estrogens appears to significantly decrease the benefit and to increase the risk of CHD in postmenopausal women.
Effect of bilateral oophorectomy & premature menopause on CHD
The effect of bilateral oophorectomy on CHD is of major importance, particularly in premenopausal women. Over 50 years ago, Oliver and Boyd published a study of 36 women who had had bilateral oophorectomy and 35 women who had had unilateral oophorectomy under the age of 35 years, whom they followed up 20 years later [70]. The women who had both ovaries removed, but not those who had one ovary conserved, had a much higher incidence of clinical coronary artery disease and a significant elevation of serum lipids. In 1959, Oliver and Boyd concluded that: “Estrogen-replacement therapy should be given to all women under the age of 50 who have had both ovaries removed or irradiated.”
In the NHS of 29,380 women with benign disease, 13,035 (44.4%) had hysterectomy with ovarian conservation and 16,345 (55-6%) had hysterectomy with bilateral salpingo-oophorectomy (BSO) [71]. Over 24 years of follow-up, the risks of CHD, stroke and total mortality were all increased in women who had hysterectomy and BSO compared with women who had hysterectomy and ovarian conservation: CHD (RR: 1.17; 1.02–1.35), stroke (RR: 1.14; 0.98–1.33) and total mortality (RR: 1.12; 1.03–1.35). In women who had BSO before the age of 50 without MHT, the risks were further increased: CHD (RR: 1.98; 1.05–2.32), stroke (RR: 2.19; 1.16–4.14) and total mortality (RR: 1.40; 1.01–1.96). With a lifespan of 35 years after surgery, one additional death would be expected for every nine women who had a hysterectomy with BSO.
In a nation wide study of Swedish women below the age of 50, the risk of CHD in later life following hysterectomy with conservation of ovaries was RR: 1.18 (1.13–1.23) but with oophorectomy was RR: 2.22 (1.01–4.83) [72].
The results of the Multi-Ethnic Study of Atherosclerosis of 2509 women aged 45–84 years followed for an average of 7 years were recently presented to the Endocrine Society [73]. Of the women, 28% reported an early menopause either natural or following bilateral oophorectomy. In women who had a premature menopause (<46 years), 5.9% had CVD events and in the women who had not had a premature menopause (>46 years), 2.6% had CVD events. Women in the premature menopause group had double the risk of CHD compared with those who did not have a premature menopause.
Perimenopausal women aged 45–55 years have little or no detectable coronary artery calcium (CAC) but the amount increases after the menopause, particularly in women not taking MHT. In a substudy of the CEE-only arm of the WHI, CAC was measured by computed tomography 1.3 years after the trial was stopped. In a multivariate analysis of women who reported no previous MHT, the RR for any CAC in women with a history of BSO was approximately doubled (RR: 2.0; 1.2–3.4) compared with those with no history of BSO [74]. MHT use within 5 years of BSO was associated with a lower prevalence of CAC (p = 0.04).
BSO before the menopause results in an abrupt reduction of endogenous estrogen and androgen production [75]. Low levels of endogenous estrogens and testosterone are associated with the development of atherosclerosis and CHD [76].
A strong case can be made for MHT in all women who have a premature menopause, particularly those who have had a BSO.
Estrogens appear to have two opposing effects on the incidence and risk of CHD. Administration of estrogens in postmenopausal women may have beneficial effects including favorable effects on lipids, coronary calcification, endothelial function, blood pressure and arterial compliance [60]. However, estrogens, particularly when given orally, may have a harmful effect by increasing the production of procoagulant proteins in the liver and inducing hypercoagulation and by changes in blood vessel walls that favor arterial and venous thrombosis [77]. Estrogens may also have proinflammatory effects including increase in C-reactive protein and matrix metalloproteinases that may result in rupture of atherosclerotic plaques [78,79]. Atherosclerosis progresses from fatty streaks to fatty plaques, then to atherosclerotic plaques and to unstable plaques. Clinical CHD results from rupture and thrombosis of vulnerable plaques [80]. Estrogens may have beneficial effects in preventing atherosclerosis in younger menopausal women but have harmful effects in older postmenopausal women with existing atherosclerosis resulting in CHD [81].
The interval between the menopause and the initiation of MHT may be critical in determining whether estrogens in postmenopausal women are beneficial or harmful [82]. After the menopause, estrogen deficiency, in addition to accelerating atherosclerosis, may result in the loss of estrogen receptors in blood vessel walls [83]. The loss of receptors increases with increasing duration of estrogen deficiency and is further increased in areas of atherosclerosis [84]. Loss of estrogen receptors reduces the beneficial effects of estrogens on the cardiovascular system. In cynomolgous monkeys on an atherogenic diet, both CEE + MPA and CEE-only started immediately after bilateral oophorectomy reduced coronary atherosclerosis by 50–70.5% but no beneficial effect was seen if HRT was delayed for 2 years [85,86].
Compared with oral administration, transdermal administration of estrogens may have less tendency to cause hypercoagulation and thrombosis of the coronary arteries. Transdermal estrogens may also have a beneficial effect on plasma lipids and in preventing atherosclerosis [87]. It has been suggested, however, that the apparent reduced effect on coagulant factors with transdermal administration of estrogens is due to the lower blood levels of estrogens rather than to the route of administration. Two ongoing randomized, controlled trials should provide valuable information on the effects of estrogens on CHD. In the Kronos Early Estrogen Prevention (KEEPS) study trial transdermal and oral administration of estrogens are being investigated and in the Early versus Late Intervention with Estradiol (ELITE) trial, early and late administration of estrogens is being investigated.
In summary, administration of estrogen appears to be of major benefit in preventing atherosclerosis and CHD in healthy younger peri- and postmenopausal women (under age 60 years) but appears to be of no benefit and may be harmful, in older women (over the age of 60 years) who may have pre-existing atherosclerosis, clinical or subclinical, with vulnerable atherosclerotic plaques. The addition of a progestin in combined therapy may reduce the benefit of estrogen on CHD. Women with a premature menopause (under age 45 years) are at increased risk of atherosclerosis and MHT is essential to prevent CHD.
Cognitive function
Many women at and after the menopause complain of forgetfulness, memory loss and mood changes. The incidence of Alzheimer's disease (AD) is greater in women than in men and it has been suggested that this gender difference may be related to the menopause and estrogen deficiency [88]. Estrogens have several possible neuroprotective effects on the aging brain including inhibition of β-amyloid formation, stimulation of cholinergic activity and reduction of oxidative cell damage [89].
In the Women's Health Initiative Memory Study (WHIMS), 4532 women aged 65 years and older and free of dementia (a subset of women in the CEE + MPA arm of the WHI) were given the Modified Mini-Mental State Examination [90]. In women treated with CEE + MPA, 40 women were diagnosed with probable dementia compared with 21 women in the placebo group (RR: 2.05; 1.21–3.48). In a separate study in the WHIMS trial, 1416 women 65 years and older and free from probable dementia were assessed by a battery of tests of specific cognitive functions [91]. CEE + MPA had a negative effect on verbal memory compared with placebo (p > 0.01), but there was a trend towards a positive effect on figural memory (p = 0.012). CEE + MPA did not significantly affect other cognitive domains or depressive symptoms.
Similar studies were carried out on 3200 women aged 65–70 years and free from probable dementia in the CEE-only arm of the WHI [92]. After a mean 5.4 years of follow-up, Modified Mini-Mental State Examination scores were lower in women who received CEE than in the placebo group (p = 0.04). CEE-only had an adverse effect on cognition in women aged 65 years and older and the effect was greater in women with lower cognitive function at the start of the trial. MRI was performed 1–4 years after conclusion of the trial on 1400 of the women who were free of dementia and cognitive impairment at the start of the investigation [93]. Fifty three of the women who developed dementia or cognitive impairment were found to have relatively smaller hippocampi and total brain volumes compared with women without dementia or cognitive impairment. It was concluded that both CEE + MPA and CEE-only increased the risk of probable dementia and did not prevent mild cognitive impairment in the groups of older postmenopausal women in the WHI trials.
A possible beneficial effect of MHT on cerebral function in younger postmenopausal women, however, cannot be entirely ruled out. In a systematic review of 29 studies in women symptomatic from the menopause, MHT improved verbal memory, reasoning and motor speed but did not improve cognition. A meta-analysis of observational studies also suggested that MHT was associated with a decrease in the risk of dementia (RR: 0.66; 0.53–0.82) [94].
In the Multi-Institutional Research in Alzheimer's Genetic Epidemiology (MIRAGE) case–control study of 532 female patients with AD and 819 controls, there was a significant association between MHT and age on the risk of AD. In a stratified analysis there was a significant benefit of MHT on AD in women aged 50–63 years (RR: 0.35; 0.19–0.66), but not in older women over the age of 63 [95].
In the Cache County prospective study of incident dementia in 1889 women (mean age 74.5 years) and 1357 men (mean age 73.2 years) followed up for 3 years, 88 women (4.7%) and 35 men (2.6%) developed AD [96]. Women who had used MHT had a reduced risk of AD compared with nonusers (RR: 0.59; 0.36–0.96). Almost all the benefit of MHT was in former users (RR: 0.33; 0.15–0.65) and the benefit increased with increasing duration of use up to 10 years. No benefit was seen in current users of MHT unless it had exceeded 10 years. The authors suggest that the benefit of MHT may take years to appear and that there may be a considerable latency between treatment and effect. MHT may possibly have a beneficial effect in prevention of AD in younger postmenopausal women but may have a deleterious effect in older women.
In a recent retrospective French study of 3130 naturally postmenopausal women aged 65 years and older, MHT did not influence the incidence of dementia but current MHT users performed significantly better than never users in verbal fluency, working memory and psychomotor speed [97]. Global cognitive function, verbal memory and executive function, however, were not altered. The majority of women in this study used transdermal estrogen–progestin preparations. A longer duration of MHT appeared more beneficial but initiation of MHT close to the menopause was not associated with better cognition. The KEEPS trial is examining the effect of oral and transdermal MHT on cognitive function in recently menopausal women. MHT cannot currently be recommended for the prevention or treatment of cognitive impairment or AD in postmenopausal women.
Osteoporosis & fracture
One of the undisputed benefits of both the CEE + MPA and CEE-only arms in the WHI was the reduction in the risk of hip fracture. In the CEE + MPA arm the risk was RR: 0.66 (0.45–0.98) and there were five fewer hip fractures per 10,000 person-years [1]. Postintervention (mean 2.4 years) in the CEE + MPA arm, the risk of hip fracture was RR: 1.11 (0.90–1.27) and there was no further benefit after 1 year [25]. In the CEE arm the risk of hip fracture was RR: 0.61 (0.41–0.91) and there were six fewer fractures per 10,000 person-years [2]. Postintervention (mean 4 years) the risk of hip fracture was RR: 1.11 (0.88–1.82) and all benefit was lost after 3 years [26].
The UK Medicines and Healthcare Products Regulatory Agency Medicine (MHRA) stated that: “The incidence of hip fractures in women younger than 60 years is very low. It does not increase significantly until age 70 years and peaks in those older than 85 years. If women stop using HRT by the time they reach 60 years, its effect on bone density will wear off within a few years and it is unlikely to provide benefit when most needed. If HRT is continued beyond age 60 years for the sole purpose of preventing osteoporosis, many women will likely be exposed to an unacceptable risk due to the marked increase in baseline risk of cardiovascular events and the increasing risk of breast cancer and ovarian cancer with long-term HRT use” [101].
Current MHT use prevents bone loss and helps to prevent osteoporosis and fractures and is beneficial in the prevention and treatment of osteoporosis in women under age 60 years. It cannot be recommended for long-term prevention in women over age 60 years as other effective nonhormonal therapies, albeit with their own risks and benefits, are available. MHT is the preferred therapy in the prevention of bone loss in all women with a premature menopause.
Balance of risk & benefits & mortality
The WHI
Several authorities have endeavored to draw up a balance sheet of the risks and benefits of MHT The WHI Investigators calculated a ‘global index’ of the combined risks of CHD, stroke, pulmonary embolism (PE), breast cancer, colorectal cancer, endometrial cancer, hip fracture and death due to other causes [1,2].
In the CEE + MPA arm of the WHI the global index of risk was increased (RR: 1.12; 1.02–1.24). The absolute excess incidence per 10,000 person-years was seven more CHD events, eight more strokes, eight more PE's, eight more invasive breast cancers, five fewer colorectal cancers but five fewer hip fractures [1]. In the analysis of the health outcomes postintervention after 3 more years of follow-up, the global index was unchanged (RR: 1.14; 0.97–1.27) [25]. The risks of CHD, stroke, deep vein thrombosis and PE were no longer increased with RR: 0.95 (0.73–1.26), RR: 1.16 (0.83–1.61), RR: 1.07 (0.66–1.75), RR: 1.07 (0.62–1.86), respectively. The overall risk for malignancies however, was increased from RR: 1.03 (0.92–1.15) during therapy to RR: 1.24 (1.04–1.48) postintervention and the overall mortality was increased from RR: 0.97 (0.81–1.24) during therapy to RR:1.15 (0.95–1.39) postintervention. Most of the deaths postintervention were due to cancer: 101 in the CEE + MPA group and 69 in the placebo group. Only 27 of the deaths in the CEE + MPA group and 16 of the deaths in the placebo group were due to breast, colorectal, endometrial or ovarian cancer. The number of deaths from ‘other cancers’, mainly lung cancers, was increased postintervention: 33 deaths in the CEE + MPA versus 15 in the placebo group. This remains to be explained.
In the CEE-only arm the global index was not significantly increased or decreased (RR: 1.03; 0.93–1.14). The absolute incidences per 10,000 person-years were 12 more strokes and six fewer hip fractures and the incidence of adverse events in the global index was a nonsignificant two additional events per 10,000 person years [2]. Postintervention, the global index of risk was unchanged (RR: 1.02; 0.89–1.1) [26]. The risks of breast cancer remained reduced (RR: 0.75; 0.51–1.09), the risk of CHD was unchanged (RR: 0.97; 0.75–1.25), the risk of stroke was no longer elevated (RR: 0.89; 0.64–1.24) and the risk of deep vein thrombosis was reduced (RR: 0.63; 0.41–0.98). The risk of hip fracture, however, was no longer reduced (RR: 1.27; 0.88–1.82). There were significant interactions between increasing age and the global index (p = 0.009) and the risk of CHD (p = 0.05). The overall health outcomes were more favorable for younger women (age 50–59 years) compared with older women (age 70–79 years).
The WHI Investigators concluded that CEE + MPA and CEE-only should not be used for the prevention of chronic disease in postmenopausal women. The women in the WHI were on average older (over 60 years of age) and had many more risk factors than women who are usually treated with MHT who are, in general, younger (under 60 years of age) and are healthy with very few risk factors.
Mortality
In the WHI, the total mortality was reduced in younger women age 50–59 years in both the CEE + MPA and the CEE-only arms of the trial with RR: 0.69 (0.44–1.07) and RR: 0.71 (0.46–1.01), respectively, and was significantly reduced when the data for both arms was combined (RR: 0.70; 0.51–0.96) [62]. The total mortality in both groups was not significantly increased in older postmenopausal women over the age of 60. The mortality, however, may be increased in subgroups, for example in women who developed breast cancer and women treated with CEE + MPA who had an increased mortality from lung cancer postintervention in the WHI. Breast cancer and lung cancer are the two leading causes of cancer death in women.
In an analysis of the pooled data of 26,708 postmenopausal women from 30 trials, the total mortality in postmenopausal women who had received MHT under age 60 was significantly reduced (RR: 0.61; 0.39–0.95) and was not increased in older women over the age of 60 years (RR: 1.03; 0.90–1.18) [66]. The overall mortality associated with MHT was RR: 0.98 (0.87–1.12).
In the CTS of 71,327 postmenopausal women (mean age 63 years), the overall all-cause mortality was reduced in current MHT users (RR: 0.83; 0.79–0.87), but was not reduced in former MHT users compared with never users [7]. The mortality was decreased in current MHT users in women age 36–59 years (RR: 0.54; 0.46–0.62), but was not decreased in older women age 75–84 years (RR: 1.00; 0.91–1.09).
The reduction in total mortality with MHT in women age 50–59 years provides substantial evidence of the overall positive balance of benefits and risks, and reassurance of the safety of MHT in younger postmenopausal women.
UKMHRA
The Public Assessment Report of the UKMHRA is the most recent independent synthesis and meta-analysis of evidence from randomized trials and observational studies from Europe and from North America. The report summarizes the evidence reviewed by the Agency on the risks and benefits of MHT together with the Agency's recommendations on MHT [101].
The MHRA calculated a ‘crude estimate’ of balance of risks by adding the extra cases of breast cancer, endometrial cancer, ovarian cancer, stroke, CHD and VTE caused by MHT, and the number of fractures and colorectal cancers prevented by MHT in 1000 women aged between 50–59 years and 60–69 years over 5 and 10 years of use. The risks in women with a uterus treated with estrogens only were also included (Tables 7 & 8).
Comparison of overall balance of benefits and risks associated with estrogen only and combined HRT for 5 years of use.
Obtained by subtracting the baseline risk in non-HRT users from the absolute risk in HRT users.
Data taken from [101].
Comparison of overall balance of benefits and risks associated with estrogen only and combined HRT for 10 years of use.
Obtained by subtracting the baseline risk in non-HRT users from the absolute risk in HRT users.
Data taken from the Medicines and Healthcare Products Regulatory Agency [101].
The MHRA concluded that “though crude, these estimates demonstrate that for most women without a uterus the balance of risks and benefits in users of estrogen-only HRT is likely to be favorable. For women with a uterus the balance of risk and benefits is likely to be less favorable particularly for those who are older than 60 years and those who have used MHT for a long time”. A major deficiency of the MHRA estimate is that it does not include the important benefit of the relief of vasomotor symptoms and does not take into account the significant reduction in total mortality with MHT in women under the age of 60.
The MHRA noted that in women with a premature menopause “evidence for the risk of HRT in women who have a premature menopause is limited. However, the baseline risk of adverse events in these younger women is very low and the balance of benefits and risks may therefore be more favorable than in older women.” In the management of women with a premature menopause, it is also necessary to take into account the important benefit of MHT in prevention of atherosclerosis and osteoporosis as well as the relief of menopausal symptoms.
The final conclusions of the MHRA report were: “Generally the much lower baseline risk of CHD and other adverse events in healthy younger women who use HRT to relieve menopausal symptoms means that the overall risk of HRT is very low. With increasing age, however, their baseline risk for all cardiovascular events increases substantially, so older HRT users have a much greater overall risk of these events. Furthermore risk of breast cancer, ovarian cancer and endometrial cancer due to HRT increases with longer duration of use. The balance of risks and benefits of HRT therefore differs for every woman according to her need for treatment, age at starting HRT, duration of use and type of HRT, estrogen only or estrogen plus a progestogen.”
The MHRA report, like many others, though necessarily and correctly cautious, underestimates the immeasurable benefit of the improvement in quality of life conveyed by MHT in women with menopausal symptoms.
Future perspective
The main development in the next 5–10 years may well be the publication of the KEEPS randomized, controlled trial of oral versus transdermal therapy. If transdermal administration of estrogen carries less risk of increasing blood coagulation factors, VTE and stroke and prevents atherosclerosis and reduces CHD, the balance of benefits and risks of MHT will be significantly altered and require a reappraisal of oral MHT.
The role of androgen therapy in menopausal women, particularly those who have had a bilateral oophorectomy, may become of increasing importance in women with loss of libido and other sexual problems.
The Kronos Study includes a comprehensive assessment of mood, sleep and sexual function. If MHT is shown to improve cerebral function if given for 10 years or so after the menopause, and is shown to reduce the incidence of AD in later life, the approach of many women and health professionals to MHT will be radically altered.
Executive summary
Outcomes of the Women's Health Initiative (WHI) trial with conjugated equine estrogens (CEE) and medroxyprogesterrone acetate (MPA) and CEE-only extensively analysed in last 10 years and a consensus on menopausal hormone therapy (MHT) is now emerging.
Relief of menopausal symptoms is a major benefit and improves quality of life; duration of vasomotor symptoms may exceed generally accepted time frame.
Breast cancer risk influenced by many factors:
– Type of MHT: risk increased by CEE + MPA but decreased by CEE in WHI
– Duration of use: risk Increases with increasing duration of MHT
– BMI: risk increases with increasing BMI but not increased further by MHT in WHI as majority of women overweight or obese
– Cessation of MHT: no increased risk in past users 5 years after stopping MHT
– Benign breast disease and family history of breast cancer as important as MHT
Endometrial cancer may be increased with sequential but decreased with continuous combined MHT; further investigation required.
Colorectal cancer incidence decreased with CEE + MPA but not by CEE-only in WHI; not a major benefit of MHT.
Lung Cancer mortality increased post intervention of CEE + MPA: needs investigation.
Venous thromboebolism (VTE): Risk increases with increasing age but very small under age 60, may be less with low dose oral and transdermal MHT.
Stroke risk increased by MHT; increases with increasing age but risk very small under 60; may be even less with low dose oral and transdermal MHT.
Coronary heart disease (CHD): estrogens decreases atherosclerosis and CHD in healthy younger menopausal women but increases risk CHD in older women.
Cognitive function impaired by MHT in WHI: incidence of Alzheimer's may possibly be decreased by MHT starting at menopause in younger women: needs investigation.
Osteoporosis MHT prevents bone loss and osteoporosis but is not recommended after age 60 because of increased risk VTE, stroke and CHD.
Balance of Benefits and Risks UK Public Assessment Report “the much lower baseline risks in healthy younger women who use HRT to relieve menopausal symptoms means overall risk from HRT very low; balance of benefits and risk is specific to every woman's circumstances.” Immeasurable benefit of relief of menopausal symptoms and improved quality of life must be taken into account.
Simplified Approach to MHT suggested as a framework for care.
Footnotes
The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.