Women’s Health Initiative

The two main WHI trials were designed as randomized, double-blind trials comparing estrogen plus progestin or estrogen only against placebo.⁷⁴ The primary outcome for both trials was CHD, and invasive breast cancer was the primary adverse event. Secondary outcomes included hip and other types of fractures; additional cardiovascular disease outcomes such as stroke and thromboembolic disease; endometrial, colorectal, and other types of cancer; and mortality. A global index was developed to estimate the effect of hormone therapy on overall health and included the earliest occurrence of CHD, invasive breast cancer, stroke, pulmonary embolism (PE), endometrial cancer, colorectal cancer, hip fracture, or death due to other causes.

Participants were recruited by population-based direct mailing campaigns and media awareness programs to one of 40 clinical centers in the United States between 1993 and 1998. Eligibility criteria included postmenopausal status, aged 50 to 79 years at initial assessment, having an intact uterus for the estrogen plus progestin trial or a previous hysterectomy for the estrogen only trial, plans to reside in the area for 3 years, and ability to provide written informed consent. Women were excluded if they had a medical condition associated with a predicted survival of less than 3 years, breast cancer or other cancer within the last 10 years, alcoholism, dementia, or transportation limitations. Participants using hormone therapy at the baseline assessment were required to go through a 3-month washout period prior to randomization.

Participants enrolled in the two trials had different cardiovascular and breast cancer risk factors at baseline (Table 4). Those in the estrogen only trial had more risk factors for cardiovascular disease. These included high body mass index (BMI); history of previous myocardial infarction (MI), stroke, and thromboembolic events; high systolic and diastolic blood pressure; treatment for hypertension; high cholesterol levels requiring medication; and treatment for diabetes. Participants in the estrogen only trial also had characteristics that reduce risk for breast cancer, including more women with previous hysterectomies and bilateral oophorectomies, less with nulliparity, and fewer first pregnancies at age 30 years and older. However, more participants in the estrogen only trial had relatives with breast cancer and high BMI—factors that increase risk for breast cancer. Interpretation of results requires consideration of these differences, and acknowledgement that WHI was not a head-to-head trial of estrogen only versus estrogen plus progestin.

Table 4

Baseline Characteristics of Participants in Randomized, Controlled Trials.

Participants were followed by telephone 6 weeks after randomization and every 6 months thereafter, with in-clinic visits each year to assess outcomes and adverse events. At each 6-month visit, reports of outcomes were obtained using self-administered questionnaires. Participants had mammography and clinical breast examinations annually, and electrocardiography at baseline and at 3 and 6 years. Adherence to study medication was assessed by weighing returned pill bottles, and study medications were withheld from participants who did not follow the protocol. Permanent discontinuation of study medications was required for participants who developed breast cancer, endometrial pathology, deep vein thrombosis (DVT), PE, malignant melanoma, meningioma, or triglyceride level >1000 mg/dL or received prescriptions of estrogen, testosterone, or selective estrogen receptor modulator medications from their personal physicians.

WHI estrogen plus progestin trial. The estrogen plus progestin trial randomized 16,608 participants; 8,506 received 0.625 mg per day of conjugated equine estrogen (CEE) plus 2.5 mg per day of medroxyprogesterone acetate (MPA) in a single tablet and 8,102 received matching placebo.¹² There were no significant differences in baseline characteristics between groups. The mean age of participants was 63.3 years, the majority were white (84 percent in both groups), and most had never used hormone therapy in the past (74 percent in both groups).

Study participants, physicians, and outcome assessors were blinded to the assignment of medication throughout the trial. However, clinic gynecologists were unblinded when managing adverse effects, such as persistent vaginal bleeding for 3,444 (40.5 percent) participants in the estrogen plus progestin group and 548 (6.8 percent) in the placebo group. Participants who had a hysterectomy after randomization for reasons other than cancer were switched to estrogen only or placebo without unblinding (248 [2.9 percent] using estrogen plus progestin and 183 [2.3 percent] using placebo). A small imbalance in the number of participants in each group was a consequence of an early protocol change eliminating an estrogen only intervention for participants with a uterus.

The trial was planned for 8.5 years; however, it was stopped early in 2002 by the Data and Safety Monitoring Board after an average of 5.2 years. The Board concluded that increases in breast cancer, CHD, stroke, and PE outweighed reductions in fractures and colon cancer. At the time the trial was stopped, vital status was known for 16,025 participants (96.5 percent), 449 (2.7 percent) had died, and 583 (3.5 percent) were lost to followup or stopped providing outcomes data for more than 18 months. Forty-two percent of participants in the treatment group and 38 percent in the placebo group had stopped taking study medications at some time during the trial. Data from these participants were included in the intention-to-treat analysis.

After discontinuation of the trial, assessment of outcomes continued until 2005, the end of the predefined trial period.³³ Postintervention data were available for 8,052 participants from the original 8,506 participants randomized to receive estrogen plus progestin (95 percent), and for 7,678 of 8,102 participants in the original placebo group (95 percent). Participants lacking postintervention data did not differ by treatment group or from participants providing data. Participants were followed for an average of 2.4 years during the postintervention trial.

After the postintervention trial ended, participants provided consent for the extension trial for continued assessment of breast cancer incidence until 2009.²⁵ Data were available for 6,545 participants from the estrogen plus progestin group and 6,243 from the placebo group who consented (83 percent of surviving participants from the original cohort). Baseline characteristics of participants consenting to the extension phase were similar to characteristics of the original WHI cohort. Participants who consented were slightly younger and more likely to be white compared with those who did not consent.

Only data from participants in the postintervention and extension phases of the WHI estrogen plus progestin trial that were analyzed based on originally assigned medication groups using an intention-to-treat analysis were used for this report.

WHI estrogen only trial. The estrogen only trial randomized 10,739 participants; 5,310 received 0.625 mg per day of CEE and 5,429 received matching placebo.⁴² There were no significant differences in baseline characteristics between groups. The mean age of participants was 63.6 years, most were white (75 percent in both groups), and approximately half had never used hormone therapy in the past (52 percent in both groups). As with the estrogen plus progestin trial, participants, physicians, and outcome assessors were blinded throughout the trial.

The trial was planned for 9 years; however, it was stopped early in 2004 by the Data and Safety Monitoring Board after an average of 6.8 years because of increased incidence of stroke in the estrogen group. At the time the trial was stopped, vital status was known for 95 percent of participants, 5.4 percent had died, and 54 percent had stopped taking their study medication.

After discontinuation of the trial, assessments of outcomes continued until 2005, the end of the predefined trial period.⁴⁹ Postintervention data were available for 4,794 participants from the original 5,310 participants randomized to receive estrogen (90 percent), and for 4,877 of 5,429 participants in the original placebo group (90 percent). Participants lacking postintervention data did not differ by treatment group or from participants providing data. Participants were followed for an average of 3.9 years during the postintervention trial.

After the postintervention trial ended, participants provided consent for the extension trial for continued assessment of breast cancer incidence until 2009.⁴⁹ Data were available for 3,778 participants from the estrogen group and 3,867 from the placebo group who consented (75 percent of surviving participants from the original cohort). Baseline characteristics of participants consenting to the extension phase were similar to the characteristics of the original WHI cohort. However, there were slight differences between groups, with more participants in the estrogen only group having no previous pregnancies (9.3 vs. 8.0 percent; p=0.04) and fewer with bilateral oophorectomies (39.0 vs. 41.8 percent; p=0.01) than in the placebo group.

Only data from participants in the postintervention and extension phases of the WHI estrogen only trial that were analyzed based on originally assigned medication groups using an intention-to-treat analysis were used for this report.

Women’s Health Initiative Memory Study

WHIMS was a concurrent trial of participants enrolled in either the WHI estrogen plus progestin trial or the WHI estrogen only trial.⁵⁵ The primary outcome was all-cause dementia. Secondary outcomes included mild cognitive impairment and global cognitive function. Enrollment was limited to women age 65 years or older and free of probable dementia.

In addition to the measures and laboratory tests required for the main WHI trials, participants in WHIMS completed the Modified Mini-Mental State Examination (3MSE) at baseline and annually thereafter. The 3MSE has 15 parts with 46 items and scores ranging from 0 to 100, with a higher score reflecting better cognitive function. WHIMS used the 3MSE to screen for global cognitive impairment and track changes. If participants scored at or below education-adjusted cut-off points, they completed an expanded neuropsychological battery. This included the modified Consortium to Establish a Registry for Alzheimer’s Disease and a clinical examination. These participants were also examined by a physician with experience diagnosing dementia, who reviewed all data and classified the participants as having no dementia, mild cognitive impairment, or probable dementia based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria. Participants diagnosed with probable dementia had an unenhanced computed tomography scan of the brain and laboratory blood tests to rule out possible reversible causes. If dementia was present after these tests, the physician diagnosed the most probable etiology based on reviewing all data. The diagnosis of vascular dementia, Alzheimer’s disease, and other dementia-related classifications was based on DSM-IV criteria. A panel of adjudicators independently reviewed all cases of probable dementia, as well as 50 percent of cases of mild cognitive impairment and a random sample of 10 percent of cases without dementia. All cases were discussed until a consensus was reached.

Of the 4,894 eligible participants from the WHI estrogen plus progestin trial, 4,532 (93 percent) agreed to participate in WHIMS; 2,229 in the estrogen plus progestin group and 2,303 in the placebo group.⁵⁵ Nearly half of the participants were aged 65 to 69 years (47 percent in both groups) and most had never used hormone therapy in the past (78 percent in both groups). After 6 years of followup, only 32.3 percent of participants in the estrogen plus progestin group and 61.4 percent in the placebo group still adhered to their study medication. Of the 3,200 eligible participants from the WHI estrogen only trial, 2,947 (92 percent) agreed to participate in WHIMS; 1,464 in the estrogen group and 1,483 in the placebo group.⁵⁴ Most participants were white (83 percent in estrogen group and 84 percent in placebo group), nearly half were aged 65 to 69 years (44 percent in estrogen group and 45 percent in placebo group), and half had never used hormone therapy in the past (54 percent in estrogen group and 55 percent in placebo group). After 7 years of followup, only 36.8 percent of participants in the estrogen group and 45.1 percent in the placebo group still adhered to their study medication.

Data were analyzed according to assigned groups at WHI enrollment and until the trial stopped in 2002 for the estrogen plus progestin trial and 2004 for the estrogen only trial. WHIMS did not include any postintervention or extension trial data.

Women’s Health Initiative Study of Cognitive Aging

WHISCA was a concurrent trial of women enrolled in either the WHIMS estrogen plus progestin trial or the estrogen only trial.⁵⁸ The primary outcome was age-related changes in specific cognitive function. Enrollment was limited to English-speaking women without probable dementia at 14 of the 39 WHIMS centers. WHISCA was initiated 3 years after WHI randomization. In addition to the WHI and WHIMS measures, participants in WHISCA completed a battery of cognitive tests consisting of the Primary Mental Abilities scale, the Benton Visual Retention Test, the California Verbal Learning Test (CVLT), the Positive and Negative Affect Schedule, the Geriatric Depression Scale, verbal fluency (letter and semantic), attention and working memory ability (digits forward and backward), spatial rotation ability (card rotations), and fine motor speed (finger tapping).

Of the 2,089 eligible participants from the WHIMS estrogen plus progestin trial, 1,416 (67.8 percent) agreed to participate in WHISCA; 690 in the estrogen plus progestin group and 726 in the placebo group.⁵⁸ The mean age of participants was 74 years, the majority were white (92 percent in estrogen plus progestin group and 93 percent in placebo group), and most had never used hormone therapy in the past (79 percent in estrogen plus progestin group and 77 percent in placebo group). After 3 years of followup, 42.2 percent in the estrogen plus progestin group and 44.1 percent in the placebo group completed final assessments. Of the 1,361 eligible participants from the WHIMS estrogen only trial, 866 (64 percent) agreed to participate in WHISCA; 434 in the estrogen group and 452 in the placebo group.⁵⁷ The mean age of participants was 74 years, the majority were white (86 percent in estrogen group and 87 percent in placebo group), and half had never used hormone therapy in the past (50 percent in estrogen group and 54 percent in placebo group). After 4 years of followup, 75.1 percent in the estrogen group and 71.9 percent in the placebo group completed final assessments.

Data were analyzed according to assigned groups at WHI enrollment and until the trial stopped in 2002 for the estrogen plus progestin trial and 2004 for the estrogen only trial. WHISCA did not include any postintervention or extension trial data.

Heart and Estrogen/Progestin Replacement Study

HERS was a randomized, double-blinded, secondary prevention trial of estrogen plus progestin compared with placebo in women with established coronary artery disease.^59,62 The primary outcome was the occurrence of nonfatal MI or CHD death. Secondary outcomes included other CHD outcomes, vascular disease, cancer, thromboembolism, gallbladder disease, fractures, mortality, uterine bleeding, and side effects of hormone therapy.

Participants were recruited from outpatient and community settings at 20 clinical centers in the United States between 1993 and 1994. Postmenopausal women younger than age 80 years with an intact uterus were eligible. Women were excluded if they reported a CHD event within 6 months of randomization, hormone therapy was contraindicated or used within 3 months of recruitment, they were participating in another trial, or they were thought to be unlikely to adhere to the protocol.

Participants were followed every 4 months to assess compliance, obtain data, and refill medications. Evaluations included annual general and cardiac examinations with electrocardiography; blood tests at first, third, and final visits; annual breast examinations, screening mammography, and pelvic examinations with Papanicolaou tests; and endometrial evaluations at second and final annual visits. Permanent discontinuation of study medications was required for participants who developed endometrial pathology; endometrial, cervical, breast, or ovarian cancer; DVT, PE, or prolonged immobilization; or active gallbladder disease.

There were 2,763 participants randomized; 1,380 received 0.625 mg per day of CEE and 2.5 mg per day of MPA in a single tablet and 1,383 received matching placebo. Baseline characteristics did not differ significantly between groups. The mean age of participants was 66.7 years, the majority were white (88 percent in estrogen plus progestin group and 90 percent in placebo group), and most had never used estrogen in the past (76 percent in estrogen plus progestin group and 77 percent in placebo group).⁶² Participants, physicians, and outcome assessors were blinded throughout the trial. Symptoms were addressed by separate medical staff.

During the middle years of the trial, the incidence of venous thromboembolic events in the estrogen plus progestin group exceeded safety thresholds. The Data Safety and Monitoring Board recommended against extending the trial beyond its scheduled closeout date. HERS ended in 1998 after 4.1 years of followup. Participants were informed of their treatment assignment and the main trial results. Blinded medication was stopped and participants were advised to make individual decisions about using open-label hormone therapy after discussing it with their personal physicians.^60,61

At the time the trial ended, vital status was available for all 2,763 participants. Self-reported adherence to study medications was 75 percent in the hormone group and 81 percent in the placebo group after 3 years, and pill counts indicated that 70 percent of participants in the hormone group were 80 percent adherent to their medications at the end of year 3. There were 36 (3 percent) participants in the estrogen plus progestin group and 110 (8 percent) in the placebo group who had discontinued their study medication and reported taking estrogen outside of the study.⁶² Closeout assessments were completed by 1,222 participants in the estrogen plus progestin group and 1,228 in the placebo group. Unadjusted intention-to-treat analysis was performed for the primary analysis. Secondary analyses using multivariate hazard models were adjusted to examine possible confounders by controlling for baseline covariates.⁶²

Of the surviving HERS participants, 2,321 (93 percent) agreed to enroll in HERS II followup and provided consent; 1,156 participants in the estrogen plus progestin group and 1,165 participants in the placebo group. Baseline characteristics of participants in HERS II were not significantly different between groups. The mean age of participants was 67 years, the majority were white (89 percent in estrogen plus progestin group and 91 percent in placebo group), and a small proportion had used estrogen in the past (25 percent in estrogen plus progestin group and 23 percent in placebo group). ⁶⁰

Open-label followup continued in HERS II to collect additional data about CHD events. Documentation of clinical events in HERS II was the same as for HERS. Participants were contacted by telephone every 4 months and asked about health outcomes and their use of medications and hormones. Final telephone contacts were completed with 99 percent of survivors. Only data analyzed according to the original HERS assigned medication groups were included in this review.^61,65

HERS II was planned to continue for 4 years. Data were reviewed annually by a review committee who decided to discontinue HERS II after the second review because they determined that no useful information was likely to result from continuing until the end of the fourth year.⁶⁰ The mean duration of followup in HERS II was 2.7 years. Among participants in the estrogen plus progestin group, 81 percent adhered to medications during the first year, declining to 45 percent during the sixth year, compared with none in the first year and 8 percent in the sixth year for participants in the placebo group taking open-label hormone therapy.^60,61

Women’s International Study of Long Duration Oestrogen After Menopause

WISDOM was a randomized, placebo-controlled and head-to-head, double-blind trial comparing estrogen plus progestin with placebo and estrogen only.⁶⁶ This review uses results comparing estrogen plus progestin against placebo, because WISDOM did not report results comparing estrogen only against placebo, and comparisons of estrogen plus progestin against estrogen only are limited by baseline differences between participants.⁶⁶ Primary outcomes were major cardiovascular disease, osteoporotic fractures, and breast cancer. Secondary outcomes included other types of cancer, death from all causes, venous thromboembolism, cerebrovascular disease, dementia, and quality of life. Outcomes were subsequently changed to exclude stroke and to include unstable angina.

Postmenopausal women aged 50 to 69 years were recruited from general medical practices in the United Kingdom (384 sites), Australia (91 sites), and New Zealand (24 sites) between 1999 and 2001. Women were excluded for hormone therapy use within 6 months; MI or cardiovascular event within 6 months; DVT or PE; history of cancer; renal impairment; liver or gallbladder disease; history of hepatitis B, hepatitis C, or HIV; current pregnancy; contraceptive use within 12 months; fasting triglyceride level >5.5 mmol/L; using selective estrogen receptor modulator medications; and any other conditions or circumstances limiting informed consent or trial procedures.

Participants were followed at 4, 14, 27, 40, and 52 weeks and every 6 months thereafter to assess outcomes and adverse events using a validated questionnaire administered by study nurses. At the annual visit, risk factors were also assessed and participants’ continued eligibility was confirmed.

There were 4,385 participants randomized; 2,196 received 0.625 mg per day of CEE and 2.5 to 5.0 mg per day of MPA in a single tablet and 2,189 received matching placebo. Baseline characteristics did not differ significantly between groups. The mean age of participants was 63.3 years, almost all were white (99 percent in both groups), and half had never used hormone therapy previously (53 percent in estrogen plus progestin group and 54 percent in placebo group).

Participants, physicians, and outcome assessors were blinded throughout the trial. Unblinding occurred to manage adverse effects such as persistent vaginal bleeding for 712 (38 percent) participants in the estrogen plus progestin group and 66 (4 percent) participants in the placebo group.

While 10 years of treatment was planned, the trial was prematurely closed after publication of early results from WHI in 2002 showed increased risks with estrogen plus progestin. The median followup time was 11.9 months (interquartile range, 7.1 to 19.6). Because the trial closed early, the study was not adequately powered for most outcomes and dementia was never assessed.

Oestrogen in the Prevention of Reinfarction Trial

ESPRIT was a randomized, double-blind trial of estradiol valerate compared with placebo in women with recent MI.⁶⁷ The primary outcomes were first nonfatal reinfarction, cardiac death, or death from another cause within 2 years of study entry. Secondary outcomes included uterine bleeding, endometrial cancer, breast cancer, stroke, other thromboembolic events, fractures, and compliance with treatment.

Participants were recruited by trained research nurses from coronary care units or general medical wards in participating hospitals in England and Wales between 1996 and 2000. Postmenopausal women aged 50 to 69 years who were recently discharged from the hospital after a MI, within 31 days of admission, and had had no previous MI were eligible. Exclusion criteria included use of hormone therapy; vaginal bleeding 12 months before admission; history of breast, ovarian, or endometrial carcinoma; active thrombophlebitis; history of DVT or PE; and acute or chronic liver disease, Rotor syndrome, Dubin-Johnson syndrome, or severe renal disease.

Participants were followed at 3, 6, 12, and 18 months to assess outcomes and adverse events using a questionnaire administered by their physicians. When a potential reinfarction was noted, trained research nurses obtained medical records to ascertain whether or not a reinfarction had occurred.

There were 1,017 participants randomized; 513 received 2 mg per day of estradiol valerate, while 504 received matching placebo. Baseline characteristics were similar between groups. The mean age of participants was 62.6 years, almost all were white (97 percent), and most had never used hormone therapy in the past (88 percent in estrogen group and 90 percent in placebo group). Participants, physicians, and outcome assessors were blinded throughout the trial. A separate team of clinicians investigated uterine bleeding and other adverse effects. Participants were followed for 2 years after randomization and assessed using an intention-to-treat analysis.

Estrogen Memory Study

EMS was a randomized, double-blind trial of estrogen plus progestin given in a cyclical regimen with estrogen alone compared with placebo.⁶⁸ The primary outcome was short-delay verbal recall (scored 0–16) on the CVLT. Secondary outcomes included immediate recall (sum of five immediate recall trials), new list recall (scored 0–16), cued recall (sum of four cued recall trials), recognition memory of the CVLT (true positives plus true negatives divided by total number of words), as well as the 3MSE.

Postmenopausal women aged 60 years or older who had their last menstrual cycle at least 12 months earlier were recruited from a single center in Toronto between 2000 and 2006. The trial included women with normal to below normal scores on the short-delay recall trial of the Rey Auditory Verbal Learning Test to increase the probability of detecting cognitive decline. Exclusion criteria included women with dementia or a clinical history of a neurological, systemic, or psychiatric condition that affects cognition, as well as any conditions that could be exacerbated by estrogen. Participants were followed annually to assess outcomes and adverse events, measure serum estradiol levels, and assess adherence to study medications. Participants who took less than 80 percent of their study medication were considered to have discontinued it.

There were 142 participants randomized; 70 received 1 mg of 17-β estradiol micronized per day for 4 days followed by 1 mg of 17-β estradiol plus 0.35 mg norethindrone per day in a single tablet for 3 days, repeated in this cycle every week, and 72 received matching placebo. Baseline characteristics did not differ significantly between groups. The mean age of participants was 75 years, the majority were white (96 percent in treatment group and 90 percent in placebo group), and most had never used hormone therapy in the past (69 percent in treatment group and 76 percent in placebo group).

Participants, physicians, and outcome assessors were blinded throughout the 2-year trial. A total of 62 (88.6 percent) participants in the treatment group and 66 (91.6 percent) in the placebo group completed 2-year assessments, and outcomes were assessed using an intention-to-treat analysis.

Ultra-Low-Dose Transdermal Estrogen Assessment

ULTRA was a randomized, double-blind trial of transdermal estradiol compared with placebo.⁶⁹ Primary outcomes were bone mineral density and endometrial hyperplasia, both intermediate outcomes not meeting eligibility criteria for this review. However, clinical fractures were secondary outcomes. Cognitive function was measured, but not identified as an a priori outcome, and assessed by the 3MSE, Modified Boston Naming Test, Brief Visuospatial Memory Test, Logical Memory Immediate and Delayed, Word List Memory, Trails B, and verbal fluency. Participant recruitment was not reported. Postmenopausal women aged 60 to 80 years with an intact uterus, at least 5 years postmenopausal, and with osteoporosis but normal bone mineral density for their age were included. Exclusion criteria included unexplained uterine bleeding; abnormal mammography suggestive of breast cancer; history of metabolic bone disease; cancer; CHD; venous thromboembolism; uncontrolled hypertension; thyroid disease; liver disease; fasting triglyceride of >300 mg/dL; fasting glucose of >180 mg/dL; prior use of fluoride, calcitonin, or bisphosphonates; or estrogen or progestin use within 3 months. Participants were followed every 4 months for 2 years and cognitive function was assessed at annual visits.

There were 417 participants randomized; 208 received 0.014 mg per day of estradiol transdermally and 209 received matching placebo. All participants received oral supplements of 400 mg calcium twice daily and 400 IU vitamin D once daily. Baseline characteristics did not differ between groups, except that participants in the treatment group had baseline bone mineral density at the lumbar spine that was 2 percent lower than the placebo group. Mean age of participants was 67 years and the majority were white (93 percent in treatment group and 92 percent in placebo group). Participants, physicians, and outcome assessors were blinded throughout the trial.

At the end of the trial, 191 (91.8 percent) participants in the treatment group and 185 (88.5 percent) in the placebo group completed the trial; of these, 18 (9.4 percent) in the treatment group and 24 (13 percent) in the placebo group had stopped taking their study medications.

Summary

Results of trials indicated benefits for hormone therapy that vary by regimen. For participants using estrogen and progestin in WHI, hip, vertebral, and total fractures and diabetes were significantly reduced compared with placebo. For participants using estrogen only in WHI, invasive breast cancer incidence and death and hip, vertebral, and total fractures were reduced. Participants in HERS using estrogen and progestin had reduced diabetes, but not fractures.

Evidence

Benefits of hormone therapy are summarized in Table 5 according to regimen (estrogen with progestin [E+P] or estrogen only [E]), trial (WHI or other trials), and outcome (cancer, diabetes, and fractures). Results are expressed as hazard ratios (HRs) or rate ratios (RRs), with 95 percent confidence intervals (CIs). Results are further described in evidence tables in Appendixes C2–C12.

Table 5

Results of Hormone Therapy Trials.

Participants using estrogen only in WHI had reduced invasive breast cancer incidence (HR, 0.77 [95% CI, 0.62 to 0.95])⁴⁹ and mortality (HR, 0.37 [95% CI, 0.13 to 0.91]).⁷³ Colorectal cancer was reduced for participants using estrogen plus progestin (HR, 0.75 [95% CI, 0.57 to 1.00]),³³ although results were of borderline statistical significance. Colorectal cancer was not reduced for participants using estrogen only in WHI⁴⁹ or in the HERS trial of estrogen plus progestin.⁶¹

The incidence of diabetes was reduced for participants using estrogen plus progestin in WHI (HR, 0.79 [95% CI, 0.67 to 0.93])³⁸ and HERS (HR, 0.65 [95% CI, 0.48 to 0.89]),⁶³ but not for estrogen only in WHI.⁴³ Diabetes diagnosis was based on self-report in WHI and by fasting glucose levels in HERS (≥6.9 mmol/L).

Both estrogen plus progestin and estrogen only reduced hip, vertebral, and total fractures in WHI^12,24,42,75 but not HERS.⁶¹ For estrogen plus progestin, estimates included a hazard ratio of 0.67 (95% CI, 0.47 to 0.95) for hip, 0.68 (95% CI, 0.48 to 0.96) for vertebral, and 0.76 (95% CI, 0.69 to 0.83) for total fractures. Results for the estrogen only trial were similar (Table 5). For most fracture outcomes in WHI, the confidence intervals included 1.00 when estimates were adjusted for multiple outcomes.

Summary

Results of trials indicated several important harms for hormone therapy that vary by regimen. For participants using estrogen and progestin in WHI, CHD, stroke, thromboembolic events (DVT and PE), invasive breast cancer, death from breast cancer, death from lung cancer, gallbladder disease, probable dementia, and urinary incontinence were significantly increased compared with placebo. For participants using estrogen only in WHI, stroke, thromboembolic events (DVT), gallbladder disease, and urinary incontinence were significantly increased compared with placebo. Participants in HERS using estrogen and progestin had increased urinary incontinence. Results are further described in evidence tables in Appendixes C2–C12.

Evidence

Harms of hormone therapy are summarized in Table 5 according to regimen (estrogen with progestin [E+P] or estrogen only [E]), trial (WHI or other trials), and outcome (cardiovascular events, thromboembolic events, cancer, death, gallbladder disease, cognitive function, and urinary incontinence). Results are expressed as hazard or rate ratios with 95 percent confidence intervals.

Contrary to the cardioprotective effects initially hypothesized, participants randomized to estrogen plus progestin in WHI had increased CHD, including nonfatal MI and CHD death, although results were of borderline statistical significance (HR, 1.22 [95% CI, 0.99 to 1.51]).³³ CHD was not increased for participants randomized to estrogen only.⁴⁹

Stroke was significantly increased for both estrogen plus progestin (HR, 1.34 [95% CI, 1.05 to 1.71])³³ and estrogen only (HR, 1.36 [95% CI, 1.08 to 1.71])⁴⁹ in WHI. Thromboembolic events, DVT and PE, were also increased in both WHI trials, but to higher levels for estrogen plus progestin³³ than estrogen only.⁴⁹

Although breast cancer was reduced among participants in WHI using estrogen only, incidence was increased among participants using estrogen plus progestin (HR, 1.25 [95% CI, 1.07 to 1.46]).²⁵ Other types of cancer, including lung, endometrial, ovarian, and cervical, were not significantly increased in the estrogen plus progestin trial.^23,27,33 The estrogen only trial reported results for lung cancer that were not significantly increased.⁴⁵ Invasive breast, lung, and endometrial cancer were not increased in HERS II.⁶¹

All-cause mortality was not significantly increased in the WHI estrogen plus progestin, WHI estrogen only, HERS II, or ESPRIT trials. Death from breast cancer (HR, 1.96 [95% CI, 1.00 to 4.04])⁴⁵ and lung cancer (HR, 1.71 [95% CI, 1.16 to 2.52])²⁷ were increased for participants using estrogen plus progestin in WHI, although results for breast cancer mortality were of borderline statistical significance.

Gallbladder disease, cholecystectomy, and cholecystitis were all significantly increased in WHI for participants using estrogen plus progestin and estrogen only.²⁹ Incidence was higher for estrogen only users (gallbladder disease: HR, 1.79 [95% CI, 1.44 to 2.22]).²⁹

Measures of impaired cognitive function were significantly increased for participants in WHI using estrogen plus progestin for probable dementia (HR, 2.05 [95% CI, 1.21 to 3.48]), but not mild cognitive impairment.⁵⁵ These measures were not significantly increased for estrogen only.⁵⁴

Studies evaluating urinary incontinence used self-reported measures. The incidence of overall urinary incontinence was increased for participants using estrogen plus progestin (RR, 1.39 [95% CI, 1.27 to 1.52]) and estrogen only (RR, 1.53 [95% CI, 1.37 to 1.71]) after 1 year of treatment in WHI.³⁵ Further analysis indicated increased risk for different types of urinary incontinence, including stress, urgency, and mixed for both the estrogen plus progestin and the estrogen only trials. In a subsample of estrogen plus progestin users continent at baseline, incontinence persisted during 3 years of followup.³⁵ Weekly, stress, and urge incontinence were increased among estrogen plus progestin users in HERS (weekly odds ratio [OR], 1.6 [95% CI, 1.3 to 1.9]),⁶⁴ but urinary incontinence was not significantly increased in ULTRA.⁷¹

Summary

Subgroups defined by the key question include women with premature menopause; women with surgical menopause; age of use; types, doses, and modes of delivery of hormone; and presence of comorbidities. Trials did not report results for most of these subgroups, and post hoc subgroup analyses of trial results based on these characteristics were restricted to age and a limited number of comorbidities (Appendix C13). These included increased risk for breast cancer for participants randomized to estrogen plus progestin with prior oral contraceptive use, prior estrogen plus progestin use, or smoking; increased CHD risk for participants randomized to estrogen plus progestin with high low-density lipoprotein (LDL) cholesterol levels or randomized to estrogen only with high C-reactive protein levels; increased thromboembolic disease for participants randomized to estrogen plus progestin who were older, obese, or possessed Factor V Leiden; and increased urinary incontinence for older participants using either regimen.

Evidence

Breast cancer. In the WHI estrogen plus progestin trial, invasive breast cancer incidence was reported by baseline characteristics for age (50–59, 60–69, and 70–79 years), BMI (normal, overweight, and obese categories), Gail risk score, prior estrogen plus progestin use, and time since menopause. Among these analyses, there were no significant differences based on age, BMI, or Gail risk score.²⁵ However, breast cancer incidence was increased for participants who entered the study with prior estrogen plus progestin use (HR, 1.85 [95% CI, 1.25 to 2.80]) compared with participants with no prior use (HR, 1.16 [95% CI, 0.98 to 1.37]).

A separate analysis of the estrogen plus progestin trial reported subgroup analyses by prior estrogen plus progestin use, age, Gail risk score, prior oral contraceptive use, recency of hormone use, BMI, smoking status, and use of nonsteroidal anti-inflammatory drugs (NSAIDs).²⁶ Among the subgroup analyses, participants who received estrogen plus progestin had higher rates of invasive breast cancer if they had prior use of oral contraceptives for less than 5 years, prior use of menopausal estrogen plus progestin for 5 or more years, or were current smokers. Another analysis of estrogen plus progestin subgroups based on first-degree family history of breast cancer found no significant interactions.³¹

An evaluation of the effect of age on breast cancer incidence in the estrogen only trial indicated no significant differences by age.⁴²

Colorectal cancer. A subgroup analysis of the WHI estrogen plus progestin trial showed no differences in invasive colorectal cancer incidence by age, race or ethnic group, family history of colorectal cancer, prior use of menopausal estrogen plus progestin, BMI, waist circumference, smoking status, current alcohol use, dietary selenium, diabetes, use of NSAIDs, or history of polyp removal.²⁸

An analysis of the estrogen only trial indicated that subgroup analyses based on history of polyp removal, height, and waist circumference yielded statistically significant interactions.⁵⁰ There were no statistically significant results based on age, race or ethnic group, family history of colorectal cancer, BMI, smoking status, alcohol intake, dietary selenium, treated diabetes, use of NSAIDs, previous use of menopausal hormones, previous use of oral contraceptives, and bilateral oophorectomy.

Cardiovascular disease. Subgroup analyses in the WHI estrogen plus progestin trial indicated no significant interactions for CHD, except for participants with elevated LDL cholesterol at baseline.³⁷ Many risk factors related to CHD and other baseline characteristics were evaluated and found to not be significantly related. CHD events were significantly increased during the first year of the trial compared with later years (year 1: HR, 1.81 [95% CI, 1.09 to 3.01] vs. year 6 or later: HR, 0.70 [95% CI, 0.42 to 1.14]).³⁷ Similar analyses for the estrogen only trial indicated that participants with elevated levels of C-reactive protein at baseline had a greater risk for CHD with estrogen only, but all other analyses were not significant.⁴⁸

An additional subgroup analysis of the WHI estrogen plus progestin and estrogen only trials indicated that participants initiating hormone therapy closer to menopause had reduced CHD risk compared with participants initiating later.³⁹ Although the reduction in risk for women within 10 years of menopause was not statistically significant, the trend of increasing risk with increasing duration of time from menopause was.

Risk for stroke was similar in all subgroups evaluated for the WHI estrogen plus progestin and estrogen only trials.^41,47 For thromboembolic disease, use of estrogen plus progestin increased the risks associated with older age, being overweight or obese, or possessing Factor V Leiden.³⁰ Analysis of subgroups in the WHI estrogen only trial indicated no associations with venous thrombosis.⁴⁶

Fractures. The protective effect of estrogen plus progestin in WHI did not differ by age, BMI, smoking status, history of falls, personal and family history of fracture, calcium intake, prior hormone therapy, bone mineral density, or fracture risk score.²⁴

Cognitive function and dementia. In WHIMS, analyses were conducted on 17 subgroups using adjusted models that indicated no significant subgroup differences between estrogen plus progestin and placebo.⁵³ For estrogen only, 17 subgroups were also analyzed.⁵² Results indicated that participants at or below the screening cut-off point for 3MSE scores at baseline in the estrogen only group had significantly lower changes in 3MSE scores compared with placebo (difference, −1.52 [95% CI, −2.29 to −0.75]; p=0.006).

Urinary incontinence. In WHI, participants differed significantly at baseline for parity in the estrogen only trial, and emphysema in the estrogen plus progestin trial. Adjusting for these baseline differences did not change results for urinary incontinence outcomes. Further subgroup analysis did not show a significant effect for race or ethnicity. In the estrogen plus progestin trial, older age, increasing duration since menopause, prior menopausal hormone therapy use, and beta blocker use were significantly associated with stress urinary incontinence for estrogen plus progestin users.³⁵ The absence of diabetes was significantly associated with urge urinary incontinence. In the estrogen only trial, older age and increasing duration since menopause were significantly associated with stress urinary incontinence, and increasing age with urge and mixed urinary incontinence for estrogen users. A significant interaction was also found between smoking, estrogen only use, and mixed urinary incontinence.

In HERS, participants assigned to placebo were older and further from menopause than participants taking hormone therapy at baseline, but were otherwise similar. Adjusting for these differences did not change results. Participants taking estrogen plus progestin were more likely to report weekly, urge, and stress urinary incontinence compared with those taking placebo. The adverse effect on urge urinary incontinence increased with time. However, for participants younger than age 60 years, the effect of estrogen plus progestin on urinary incontinence was minimal and not significantly elevated (OR, 1.31 [95% CI, 0.85 to 2.04]).⁶⁴