Valerie A Flores ,
Endocrine Reviews, Volume 42, Issue 6, December 2021, Pages 720–752, https://doi.org/10.1210/endrev/bnab011
Published:
15 April 2021
Abstract
Hormone therapy (HT) is an effective treatment for menopausal symptoms, including vasomotor symptoms and genitourinary syndrome of menopause. Randomized trials also demonstrate positive effects on bone health, and age-stratified analyses indicate more favorable effects on coronary heart disease and all-cause mortality in younger women (close proximity to menopause) than in women more than a decade past menopause. In the absence of contraindications or other major comorbidities, recently menopausal women with moderate or severe symptoms are appropriate candidates for HT. The Women’s Health Initiative (WHI) hormone therapy trials—estrogen and progestin trial and the estrogen-alone trial—clarified the benefits and risks of HT, including how the results differed by age. A key lesson from the WHI trials, which was unfortunately lost in the posttrial cacophony, was that the risk:benefit ratio and safety profile of HT differed markedly by clinical characteristics of the participants, especially age, time since menopause, and comorbidity status. In the present review of the WHI and other recent HT trials, we aim to provide readers with an improved understanding of the importance of the timing of HT initiation, type and route of administration, and of patient-specific considerations that should be weighed when prescribing HT.
The use of hormone therapy (HT) in menopausal women has, in recent decades, been one of the most contentious topics in women’s health. A plethora of observational data had suggested that HT was not only effective against common menopausal symptoms such as hot flushes and night sweats, but also offered benefit against chronic disorders such as osteoporosis, coronary artery disease, dementia, and even all-cause mortality (1-4). However, as randomized trials of HT were conducted, some of the previously purported long-term health benefits of HT were called into question. The Women’s Health Initiative (WHI) primary prevention hormone trials, by demonstrating HT-related risks in older postmenopausal women, led to a seismic shift in menopause management, and a prescribing pattern that shifted considerably in the years following the WHI HT trials (5, 6). Prior to these randomized clinical trials, observational studies consistently demonstrated lower rates of coronary heart disease (CHD) and all-cause mortality among women using HT, compared to nonusers. As such, HT was initially heralded for use in the prevention of CHD and other chronic diseases, in addition to treatment of menopausal symptoms. However, the large-scale WHI trials, conducted in postmenopausal women across a broad age range (50-79 years, mean age 63) did not confirm the cardiovascular and all-cause mortality benefits that had previously been suggested by observational studies. On the contrary, although benefits for fracture reduction were confirmed, HT was found to increase the risk of stroke and venous thromboembolic disease. For CHD and breast cancer, the results varied by formulation. Subsequent analyses of the WHI indicated that the trial findings for CHD and all-cause mortality were influenced by age or time since menopause, with more favorable results in younger than older women (especially for the estrogen-alone trial [E-alone]). Given that most postmenopausal women in observational studies began HT early in menopause, the age-stratified results from the WHI trials helped to reconcile results from these different sources of evidence.
As we discuss in the present review, recent data from the WHI and other randomized trials have provided insights into the role of age, timing of HT initiation, formulation and route of administration, and assessment of comorbidities when considering prescribing HT for menopausal women.
Menopausal Hormone Therapy: Indications for Treatment
Management of Menopausal Symptoms
HT remains the most effective treatment option available for the management of menopausal vasomotor symptoms (VMS) and the genitourinary syndrome of menopause (GSM) (7-9). Both conditions are highly prevalent in postmenopausal women, affecting 80% and 50%, respectively, and adversely affecting health and quality of life. HT is approved by the US Food and Drug Administration (FDA) for both of these indications, as well as for prevention of bone loss, and treatment of premature hypoestrogenism. In a metanalysis of RCTs in aging women (on average age 50 years), it was found that both oral conjugated equine estrogens (CEEs) or transdermal estradiol (E2) (with or without the addition of a progestin) were effective in ameliorating hot flashes, reducing symptoms by 70% to 95% (10). For the treatment of GSM, a systematic review found that vaginal estrogen products were the most effective form of treatment for genitourinary symptoms, with superiority over vaginal lubricants and moisturizers (11). Clinical management guidelines, available HT formulations, and alternative treatment options are discussed in the following sections. Ongoing debate about the benefit:risk profile of HT when used to prevent osteoporosis, another FDA-approved indication for HT, are also addressed in detail.
Menopausal Hormone Therapy: Observational Studies and Clinical Trials
Observational Studies of Hormone Therapy and Cardiovascular Risk
Cardiovascular disease (CVD) risk increases for women following menopause, and the loss of ovarian estrogen has been postulated to contribute to this risk. In observational studies, decreased CHD risk was nearly consistently demonstrated in postmenopausal women using HT compared to nonusers of HT (3, 12-48). This seemed plausible because of the low risk of CHD in premenopausal compared to postmenopausal women and the findings from small-scale RCTs indicating that estrogens increase high-density lipoprotein cholesterol (HDL-C) and decrease low-density lipoprotein cholesterol (LDL-C), thus potentially slowing the risk of atherosclerosis (49). In the large-scale Nurses’ Health Study, estrogen therapy (ET) at doses of 0.3 mg or 0.625 mg of CEEs was associated with reduced risk for CHD, even after adjusting for physical activity, diet, and other lifestyle and medical factors; the same held true for combined estrogen and progestin therapy (EPT) (50). Meta-analyses of observational studies demonstrated 40% to 50% reductions in CHD comparing HT users to nonusers (51) (see subsequent sections). Observational studies showed inconsistent results for HT and stroke, with an increased risk found in the Nurses’ Health Study (52) and in the General Practice Research Database (53), but a reduced risk in several other studies (56, 60-64). The overall supportive evidence for cardioprotection from HT in observational studies led to increasing prescription rates for these hormones (54-57) and generated hypotheses for testing in randomized trials of HT and CVD outcomes.
The Women’s Health Initiative Hormone Trials
The 2 WHI randomized controlled trials (RCTs) of HT (funded by the US National Institutes of Health) were designed to examine the effects of HT on the prevention of CHD and other chronic diseases in postmenopausal women across a broad range of midlife and older age groups (5, 6). The WHI-E + P trial compared the effects of continuous combined estrogen-progestin regimen (0.625 mg of CEE and 2.5 mg of medroxyprogesterone acetate [MPA] daily) vs placebo in 16 608 postmenopausal women (aged 50-79 years; mean age 63.3 years) with an intact uterus. The WHI-E-alone trial assessed the effects of CEE (0.625 mg daily) therapy (ET) alone or placebo in postmenopausal women lacking a uterus (had previously undergone hysterectomy for noncancerous reasons) (58, 59). The WHI-E + P trial was discontinued after 5.6 years of intervention (3 years earlier than planned) because of evidence of net harm from HT exceeding a predetermined threshold, and in the absence of evidence of benefit for CHD, the primary end point). Increased risks of breast cancer, stroke, deep vein thrombosis (DVT) and pulmonary embolism (PE)—in the absence of coronary benefit—led to the premature termination of the WHI-E + P trial despite significant reductions in incident fractures and colon cancer. Subsequent analyses showed a complex matrix of benefits and risks (Table 1).
Table 1.
Women’s Health Initiative estrogen-progestin and estrogen-alone trials, intervention phasea (62, 63)
| Estrogen + progestin | Estrogen alone | |
|---|---|---|
| Outcome | HR (95% CI) | HR (95% CI) |
| CVD | ||
| Coronary heart diseaseb | 1.18 (0.95-1.45) | 0.94 (0.78-1.14) |
| Myocardial infarction | 1.24 (0.98-1.56) | 0.97 (0.79-1.21) |
| Coronary revascularizationc | 0.95 (0.78-1.16) | 1.00 (0.83-1.19) |
| Stroke | 1.37 (1.07-1.76) | 1.35 (1.07-1.70) |
| Pulmonary embolism | 1.98 (1.36-2.87) | 1.35 (0.89-2.05) |
| Deep vein thrombosis | 1.87 (1.37-2.54) | 1.48 (1.06-2.07) |
| Cardiovascular mortality | 1.08 (0.78-1.48) | 1.01 (0.78-1.31) |
| All cardiovascular eventsd | 1.13 (1.02-1.25) | 1.11 (1.01-1.22) |
| Cancer | ||
| Breast cancer | 1.24 (1.01-1.53) | 0.79 (0.61-1.02) |
| Colorectal cancer | 0.62 (0.43-0.89) | 1.15 (0.81-1.64) |
| Endometrial cancer | 0.83 (0.49-1.40) | NA |
| Cancer mortality | 1.10 (0.86-1.42) | 0.96 (0.75-1.22) |
| All cancer typese | 1.02 (0.91-1.15) | 0.93 (0.81-1.07) |
| Other outcomes | ||
| Hip fracture | 0.67 (0.47-0.95) | 0.67 (0.46-0.96) |
| All fracture | 0.76 (0.69-0.83) | 0.72 (0.64-0.80) |
| Diabetes | 0.81 (0.70-0.94) | 0.86 (0.76-0.98) |
| Gallbladder disease | 1.57 (1.36-1.80) | 1.55 (1.34-1.79) |
| Probable dementiaf | 2.01 (1.19-3.42) | 1.47 (0.85-2.52) |
| Other (non-CVD, noncancer) mortalityg | 0.59 (0.39-0.90) | 1.34 (0.93-1.94) |
| All-cause mortality | 0.97 (0.81-1.16) | 1.03 (0.88-1.21) |
| Global indexh | 1.12 (1.02-1.24) | 1.03 (0.93-1.13) |
Abbreviations: CVD, cardiovascular disease; HR, hazard ratio; NA, not applicable (because of hysterectomy).
aMedian length of randomized treatment was 5.6 years for estrogen-progestin and 7.2 years for estrogen alone.
bCoronary heart disease is defined as nonfatal myocardial infarction or coronary death.
cCoronary revascularization is defined as coronary artery bypass grafting or percutaneous coronary intervention.
d“All cardiovascular events” is a composite outcome of myocardial infarction, stroke, coronary revascularization, angina, heart failure, carotid artery disease, peripheral vascular disease, venous thromboembolism (pulmonary embolism, deep vein thrombosis), and cardiovascular mortality.
eAll cancer types except nonmelanoma skin cancer.
fProbable dementia was assessed in women aged 65 years and older.
gOther (non-CVD, noncancer) mortality is a composite outcome of dementia mortality, chronic obstructive pulmonary disease (COPD) mortality, accident and injury mortality, and other mortality of known cause not due to CVD, cancer, dementia, COPD, or accident or injury.
hGlobal index is a composite outcome of coronary heart disease, stroke, pulmonary embolism, breast cancer, colorectal cancer, endometrial cancer (in the estrogen-progestin trial), hip fracture, and all-cause mortality.
Modified from Manson JE, et al. JAMA. 2013;310(13):1353-1368. Copyright© (2013) American Medical Association. All rights reserved.
The WHI E-alone trial included 10 739 women (aged 50-79 years; mean age 63.6) and was stopped 1 year early (after 6.8 years) because of an increased risk of stroke and the absence of benefit for CHD. As in the E + P trial, increased risks of stroke, DVT, and PE were noted. However, unlike the E + P trial, there were no observed increases in risk of CHD or breast cancer with the use of E-alone (for E + P, hazard ratio [HR] was 1.18 [95% CI, 0.95-1.45] and HR 1.24 [95% CI, 1.01-1.53], respectively; for E-alone, HR was 0.94 [95% CI, 0.78-1.14] and 0.79 [95% CI, 0.61-1.02], respectively), and results for all-cause mortality were neutral in both trials. Similar to the E + P trial, significant decreases in fracture risk were seen with E-alone, compared to placebo (see Table 1). Improvements in self-reported VMS and sleep were similar to what was seen in the E + P trial (6).
Analyses following discontinuation of the WHI HT trials identified HT-related increases in the risk of dementia in the E + P and the pooled E + P and E-alone trials (not with E-alone), nonsignificant increases in the risk of ovarian and lung cancer in the E + P trial (60, 61), increased urinary incontinence and gallbladder disease both in E + P and E-alone trials, and reductions in risk of type 2 diabetes in hormone users in both trials (62, 63).
Women’s Health Initiative Hormone Trials—Postintervention and Cumulative Follow-up
Although the WHI hormone trials were stopped early, participants have had continued follow-up following the end of hormone interventions (62). The risk for CHD was determined to be neutral both in the postintervention and 13-year cumulative follow-up periods of the E + P trial and E-alone trials (62). Invasive breast cancer risk remained significantly elevated in the postintervention and cumulative follow-up period of the E + P trial, with a decline in the risk over time since EPT cessation (64). Interestingly, in the E-alone trial, the reduced risk of invasive breast cancer in hormone users achieved statistical significance during cumulative follow-up (62). Results for stroke, PE, DVT, and colorectal cancer remained neutral after 13 years of cumulative follow-up in both hormone trials, and for all-cause mortality, results were neutral at all follow-up time points in both trials (62, 63). The risk reduction in fractures was attenuated during the postintervention period for both trials, although the positive effects of HT persisted over the 13-year cumulative follow-up for the E + P trial. Dementia risk was neutral in postmenopausal women aged 50 to 55 years at random assignment during the postintervention follow-up. In addition, the decreased rate of diabetes in HT users was no longer evident following discontinuation of HT, while urinary incontinence symptoms were still elevated in women assigned to HT in both trials. Risks of gallbladder disease showed persistent elevations in the E + P trial but became neutral in the E-alone trial (62).
Several of the adverse findings from the WHI HT trials in the overall cohort (intervention and cumulative follow-up) were unexpected and inconsistent with earlier observational data. Before investigating factors that may explain the inconsistencies between WHI findings and prior observational data, it is important to examine the results of other major RCTs that assessed HT and health outcomes.
Hormone Therapy: Secondary Prevention Trials
The Heart and Estrogen-progestin Replacement Study (HERS) was a randomized, placebo-controlled trial designed to determine if EPT (CEEs + MPA) treatment reduced CHD in women with existing coronary disease (62, 65). The 4-year secondary prevention trial included 2763 women, who were on average age 66.7 years. While there was no difference in the risk of CHD (myocardial infarctions or coronary deaths) in the intervention compared to placebo groups, a 50% increased risk of CHD during the first year of the intervention was found. An increased risk of venous thrombotic events was also observed. In HERS II, HERS participants underwent an open-label, 2.7-year mean follow-up after stopping HT; rates of coronary events in the HT group were not significantly different from those in the placebo group (66).
The Estrogen Replacement and Atherosclerosis trial examined cardiovascular effects of CEEs alone or CEE + MPA compared to placebo in postmenopausal women (309 participants, average age 65.8 years) with known coronary artery disease. While this 3-year trial found improvements in lipid biomarkers with use of HT compared to placebo (decrease in levels of LDL-C and increase in HDL-C levels), there was no difference in atherosclerosis progression across treatment groups (67).
The Women’s Estrogen-progestin Lipid-Lowering Hormone Atherosclerosis Regression Trial (WELL-HART) examined the ability of daily oral E2 alone or oral E2 plus sequential MPA to slow atherosclerosis progression compared to placebo in 226 postmenopausal women with an average age of 63.5 years. There was no difference in atherosclerosis progression among the 3 groups (68).
Three additional secondary prevention randomized trials failed to identify any decrease in CHD with the use of menopausal HT (69-71).
Elucidating the Divergent Findings on Hormone Therapy and Cardiovascular Disease in Randomized Controlled Trials vs Observational Studies
Results of RCTs differed from observational data in terms of CHD effects of HT. While the bulk of observational data suggested that HT offered cardioprotection and was associated with a reduced risk of CHD, to the surprise of many, these assumptions were not supported by the primary and secondary prevention trials discussed earlier. However, it is important to recognize that in observational studies women initiating HT were younger and closer to menopause onset, as they began HT in early menopause primarily for managing hot flashes and other vasomotor symptoms. Indeed, in the Postmenopausal Estrogen/Progestin Interventions (PEPI) randomized clinical trial, in which enrolled women were on average age 56 years, the effects of HT on CVD biomarkers tended to be favorable. The PEPI trials assessed the effects of ET, EPT (with cyclic vs continuous MPA or cyclic micronized progesterone) or placebo on several CVD risk factors (HDL, fibrinogen, insulin, and systolic blood pressure) (72). The 3-year trial included 875 healthy postmenopausal women aged 45 to 64 years. The study found that E-alone resulted in the greatest increases in HDL-C and decreases in LDL-C, although all HT regimens favorably affected these biomarkers when compared to placebo; triglyceride levels increased in all hormone intervention groups compared to placebo. Fibrinogen levels were not elevated in the HT groups but were mildly elevated in the placebo group. Insulin levels did not differ across treatment or placebo groups, and there was no difference in blood pressure levels. It is important to recognize, however, that because different progestogens were used, this alone could have affected CVD outcomes. Thromboembolic disease, and breast cancer, though not primary outcomes, were assessed as adverse events; there was no difference in these adverse events across groups, but statistical power was limited (72).
In the decades following the WHI hormone trials, we have learned much about the relevance of age, and time since onset of menopause and of aging-related comorbidities as determinants of cardiovascular effects of HT; population differences in these critical modulators may contribute to the discrepant cardiovascular effects of HT reported in RCTs and observational studies, as discussed further in the subsequent section.
Meta-Analyses of Randomized Controlled Trials of Hormone Therapy in Relation to Cardiovascular Disease, Venous Thromboembolism, Breast Cancer, Fracture, and All-Cause Mortality
Coronary heart disease/cardiovascular disease/venous thromboembolism
Following the publication of the WHI results, there was a sharp decline in prescription rates and use of HT. However, in a meta-analysis of 23 trials (including 39 049 women), the effects of HT on CHD varied by the woman’s age and/or time since onset of menopause. In the meta-analysis, HT initiated less than 10 years from the onset of menopause was associated with a 32% reduction of CHD, whereas HT initiated 10 or more years since menopause onset did not reduce risk (73). Additionally, a separate meta-analysis found no association between HT and cardiac death or stroke (74). Venous thromboembolism (VTE) risk was increased in postmenopausal women using oral HT (ET or EPT); however, there was no significant excess risk in women using nonoral HT (75).
Breast cancer
In a meta-analysis of 58 studies conducted by the Collaborative Group on Hormonal Factors in Breast Cancer (CGHFBC), there was an increased risk of breast cancer in HT users, particularly EPT users, which equated to a 6% to 10% increase per 5 years of HT use (76). The risk was also increased for ever-users of EPT. In a separate meta-analysis, which included 2 RCTs, EPT was again found to confer a greater increased risk of breast cancer when compared to ET (77). In a meta-analysis including 12 RCTs, there was an increase in breast cancer mortality in EPT users; however, there was no increase in breast cancer risk with ET use alone (78). Importantly, in the WHI E-alone trial, breast cancer risk was decreased (79) with long-term follow-up. Though not assessed by the CGHFBC, nor the prior-mentioned meta-analysis, the E3N French Cohort Study also found different relative risks (RRs) of breast cancer based on HT type and years of use. In EPT users, the RR of breast cancer for 2 to 5 years of hormone use was 1.59 (95% CI, 1.39-1.95) (80). In ET users, the RR of breast cancer for 2 to 5 years of hormone use was 1.13 (95% CI, 0.70-1.8) (80). However, a major limitation of the CGHFBC (and other prior studies) was the lack of assessment of the effect of underlying breast cancer risk on attributable risk (81). As such, the CGHFBC data was recently reassessed, taking into consideration the effect of underlying risk of breast cancer on attributable risk (81). To do so, women were divided into low (1.5%), intermediate (3%), and high (6%) underlying risk of breast cancer over 5 years. Using this approach, the attributable risk in EPT users for 5 to 9 years was 12, 42, and 85 per 1000 in each respective group (81). The attributable risk in ET users was lower—4.8, 9.9, and 19 per 1000 in each respective group (81). These results highlight the importance of examining the innate risk of breast cancer for each woman, as for those with low underlying risk, the overall risk of breast cancer with use of hormone therapy is lower than that predicted from other studies (including the CGHFBC). Last, it is equally important to recognize that many of the analyzed studies were observational studies with the potential for differential surveillance (ie, mammographic screening) for breast cancer in HT users and nonusers, and as such the potential for residual confounding by other factors.
Bone health
In the WHI E + P and E-alone trials, the interventions reduced the risk of hip fracture by approximately one-third (62). In a metanalysis of RCTs assessing fracture risk in women using oral CEEs, transdermal, or oral E2 (with or without the addition of a progestin) there was a 20% to 37% reduced risk of hip, vertebral, and total fracture (82). While both CEE and E2 were effective at reducing total fracture risk, E2 resulted in a slightly greater decrease in risk (82). There was some attenuation of protection following cessation of HT, as well as a more pronounced reduced risk of fracture in those using HT before age 60 years; importantly, there was no increased risk of rebound fractures (82, 83).
All-Cause mortality
Use of HT and effect on mortality has also been assessed in a meta-analysis and systematic review of 43 RCTs that demonstrated that HT does not affect risk of death from all causes (74). A Bayesian meta-analysis of 19 RCTs and 8 observational studies of HT in younger postmenopausal women (mean age < 60 years) demonstrated a 25% reduction in mortality in women taking HT compared to placebo (84). As will be discussed in subsequent sections of this review, age at HT initiation is an important factor to consider when balancing risks and benefits of HT use in postmenopausal women.
Evidence for or Against the “Timing Hypothesis”
In Animal Models
The timing hypothesis was first proposed by Thomas Clarkson in the 1990s (85). Using a cynomolgus monkey model, he found that initiating CEE immediately after bilateral ovariectomy resulted in a 70% reduction in coronary atherosclerosis, when compared to ovariectomized monkeys receiving placebo (86, 87). Conversely, monkeys that did not initiate treatment until 2 years following bilateral oophorectomy (the equivalent of > 6 years in humans) did not have a reduction in coronary atherosclerosis; estrogen did not cause plaque regression (86, 87). Nonhuman primates represent an ideal animal model for studying menopause because they have a more than 90% average genetic coding sequence identical to humans (88, 89). In addition, they experience surgical menopause after ovariectomy and respond to ET in a manner similar to that of women (88, 89). With respect to cognitive function, several monkey models have been used to study the potential neuroprotective effects of HT based on timing of initiation. When ovariectomized monkeys are given HT within 6 months of oophorectomy, cognition improves; however, HT given 2 or more years following oophorectomy does not improve cognition (90). Human studies (as discussed later) have demonstrated similar findings. The mechanism by which early initiation of estrogen has favorable cardiovascular and neurological effects is felt to relate to its ability to play an anti-inflammatory/protective role only prior to an inflammatory insult, and prior to a prolonged hypoestrogenic state. For example, arteries from younger women/women without significant plaque buildup are able to respond favorably to exogenous estrogen production—with increased vasodilatory capacity and decreased inflammatory activity. Similarly, exogenous estrogen administration prior to the onset of cognitive dysfunction (ie, closer to the onset of menopause) has an anti-inflammatory/neuroprotective effect. However, with age and following long periods of estrogen deprivation, cardiac and brain function is adversely affected by estrogen administration. In blood vessels, there is increased vasoconstriction and potential for plaque disruption in vessels following delayed exposure to E2, and in the brain there is inability to mediate an anti-inflammatory response, with some data supporting a paradoxical effect of estrogen-mediated inflammation (91-93).
In clinical trials—by age group and time since menopause
Subsequent analyses of the WHI trials—stratifying HT use by age and time since menopause—as well as a review of newer studies have provided additional clarity on HT use related risks in postmenopausal women. The “timing hypothesis” postulates that the timing of HT initiation in relation to time since the final menstrual period (FMP) differentially affects clinical outcomes. Thus, the more proximate the timing of initiation of HT to the FMP (within 5 years of menopause onset), the more likely that HT will confer organ-specific protection (eg, cardioprotection or neuroprotection). Conversely, HT-related risks are amplified when HT is introduced in older menopausal women who are remote from menopause onset (more than 10 years since FMP) (9, 94). A secondary analysis of the WHI trial by age group (50-59, 60-69, 70-79 years at baseline) found that women in the 50 to 59 age group did not have an increased risk of CHD with HT (95-97) (Table 2, Figure 1). On the contrary, women in this age group had a protective effect with CEEs alone, and no significant effect with EPT, and a statistically significantly reduced risk for total mortality in the two HT trials pooled (95, 96). Furthermore, in a 10-year follow-up of these younger postmenopausal women in the ET arm of the WHI, there was a significantly reduced risk of myocardial infarction and CHD, and total mortality was also in the direction of risk reduction (see Table 2, Figure 1) (97, 98). Conversely, in women who were more than 20 years from the time of menopause at the time of random assignment to ET or EPT, use of HT was associated with a significantly increased risk of CHD (see Table 2) (95). As previously mentioned, subsequent metanalyses have also demonstrated reduced risk of CHD in women initiating HT within 10 years of menopause onset.
Table 2.
Health outcomes in the Women’s Health Initiative estrogen-progestin and estrogen-alone trials, according to age at study entry, intervention phasea (62)
| Estrogen + progestin | Estrogen alone | |||
|---|---|---|---|---|
| Outcome | HR (95% CI) | P, trend by age | HR (95% CI) | P, trend by age |
| CVD | ||||
| Coronary heart diseaseb | ||||
| 50-59 y | 1.34 (0.82-2.19) | .81 | 0.60 (0.35-1.04) | .08 |
| 60-69 y | 1.01 (0.73-1.39) | 0.95 (0.72-1.24) | ||
| 70-79 y | 1.31 (0.93-1.84) | 1.09 (0.80-1.49) | ||
| Myocardial infarction | ||||
| 50-59 y | 1.32 (0.77-2.25) | .55 | 0.55 (0.31-1.00) | .02 |
| 60-69 y | 1.05 (0.74-1.47) | 0.95 (0.69-1.30) | ||
| 70-79 y | 1.46 (1.00-2.15) | 1.24 (0.88-1.75) | ||
| Coronary revascularizationc | ||||
| 50-59 y | 1.03 (0.63-1.68) | .67 | 0.56 (0.35-0.88) | .06 |
| 60-69 y | 0.85 (0.64-1.13) | 1.13 (0.88-1.46) | ||
| 70-79 y | 1.08 (0.77-1.51) | 1.07 (0.79-1.43) | ||
| Stroke | ||||
| 50-59 y | 1.51 (0.81-–2.82) | .50 | 0.99 (0.53-1.85) | .77 |
| 60-69 y | 1.45 (1.00-2.11) | 1.55 (1.10-2.16) | ||
| 70-79 y | 1.22 (0.84-1.79) | 1.29 (0.90-1.86) | ||
| Pulmonary embolism | ||||
| 50-59 y | 2.05 (0.89-4.71) | .61 | 1.53 (0.63-3.75) | .28 |
| 60-69 y | 1.69 (1.01-2.85) | 1.72 (0.94-3.14) | ||
| 70-79 y | 2.54 (1.27-5.09) | 0.85 (0.39-1.84) | ||
| Cancer | ||||
| Breast cancer | ||||
| 50-59 y | 1.21 (0.81-1.80) | .68 | 0.82 (0.50-1.34) | .89 |
| 60-69 y | 1.20 (0.89-1.62) | 0.73 (0.51-1.07) | ||
| 70-79 y | 1.37 (0.90-2.07) | 0.86 (0.52-1.43) | ||
| Colorectal cancer | ||||
| 50-59 y | 0.79 (0.29-2.18) | .66 | 0.71 (0.30-1.67) | .02 |
| 60-69 y | 0.61 (0.37-0.99) | 0.88 (0.53-1.47) | ||
| 70-79 y | 0.58 (0.31-1.08) | 2.24 (1.16-4.30) | ||
| Cancer mortality | ||||
| 50-59 y | 0.71 (0.38-1.33) | .37 | 0.78 (0.43-1.40) | .06 |
| 60-69 y | 1.26 (0.88-1.80) | 0.77 (0.53-1.12) | ||
| 70-79 y | 1.13 (0.73-1.75) | 1.34 (0.90-1.97) | ||
| All cancerd | ||||
| 50-59 y | 0.97 (0.76-1.23) | .77 | 0.89 (0.66-1.19) | .39 |
| 60-69 y | 1.11 (0.93-1.31) | 0.89 (0.73-1.08) | ||
| 70-79 y | 0.94 (0.75-1.17) | 1.04 (0.81-1.33) | ||
| Other outcomes | ||||
| All fracture | ||||
| 50-59 y | 0.82 (0.68-1.00) | .83 | 0.90 (0.72-1.11) | .33 |
| 60-69 y | 0.70 (0.61-0.81) | 0.63 (0.53-0.75) | ||
| 70-79 y | 0.79 (0.66-0.95) | 0.71 (0.58-0.87) | ||
| Diabetes | ||||
| 50-59 y | 0.85 (0.66-1.09) | .10 | 0.83 (0.67-1.04) | .99 |
| 60-69 y | 0.61 (0.49-0.77) | 0.91 (0.76-1.09) | ||
| 70-79 y | 1.35 (0.98-1.88) | 0.82 (0.62-1.07) | ||
| Other (non-CVD, noncancer) mortalitye | ||||
| 50-59 y | 0.53 (0.22-1.27) | .65 | 0.51 (0.20-1.26) | .002 |
| 60-69 y | 0.55 (0.27-1.13) | 1.15 (0.65-2.03) | ||
| 70-79 y | 0.67 (0.35-1.26) | 2.59 (1.36-4.92) | ||
| All-cause mortality | ||||
| 50-59 y | 0.67 (0.43-1.04) | .20 | 0.70 (0.46-1.09) | .04 |
| 60-69 y | 1.07 (0.81-1.41) | 1.01 (0.79-1.29) | ||
| 70-79 y | 1.03 (0.78-1.36) | 1.21 (0.95-1.56) | ||
| Global indexf | ||||
| 50-59 y | 1.12 (0.89-1.40) | > .99 | 0.84 (0.66-1.07) | .02 |
| 60-69 y | 1.13 (0.97-1.31) | 0.99 (0.85-1.15) | ||
| 70-79 y | 1.12 (0.95-1.32) | 1.17 (0.99-1.39) |
Abbreviations: CVD, cardiovascular disease; HR, hazard ratio.
aMedian length of randomized treatment was 5.6 years for estrogen-progestin and 7.2 years for estrogen alone.
bCoronary heart disease is defined as nonfatal myocardial infarction or coronary death.
cCoronary revascularization is defined as coronary artery bypass grafting or percutaneous coronary intervention.
dAll cancer types except for nonmelanoma skin cancer.
eOther (non-CVD, noncancer) mortality is a composite outcome of dementia mortality, chronic obstructive pulmonary disease (COPD) mortality, accident and injury mortality, and other mortality not due to CVD, cancer, dementia, COPD, or accident or injury. Age-specific effect estimates for these individual outcomes could not be computed because the numbers of events within age strata were small.
fGlobal index is a composite outcome of coronary heart disease, stroke, pulmonary embolism, breast cancer, colorectal cancer, endometrial cancer (in the estrogen-progestin trial), hip fracture, and mortality.
Modified from Manson JE, et al. JAMA. 2013;310(13):1353-1368. Copyright© (2013) American Medical Association. All rights reserved.