Menopause and Hormone Replacement Therapy
General information on menopause
Clinically diagnose menopause based upon amenorrhea for 6 months, symptoms such as vasomotor flushes, vaginal dryness, and insomnia.
Regarding symptoms associated with menopause, a panel convened by the NIH concluded that evidence establishes causality only for vasomotor symptoms and vaginal dryness, with some positive evidence for causality with regards to sleep disturbances. Evidence does NOT establish causality for mood symptoms (depression, anxiety, and irritability), cognitive disturbances, somatic symptoms (back pain, tiredness, and stiff joints), urinary incontinence, or sexual dysfunction (Ann Intern Med. 2005. 142. 1003-1013).
A meta-analysis of 2 longitudinal and 8 cross sectional studies with >80% follow up showed that vasomotor symptoms on average start to increase 2 years before the final menstrual period, peak at 1 year after the final menstrual period, and return to baseline 8 years after the final menstrual period. 50% of women continue to report vasomotor symptoms 4 years after the final menstrual period and 10% report symptoms up to 12 years after the final menstrual period (J Gen Intern Med. 2008. 23. 1507-1513).
Confirm a diagnosis of menopause with FSH greater than 40 pg/nl, and serum pregnancy test negative. An alternative to a FSH level is measurement of vaginal pH – a value greater than 4.5 in women without vaginitis is diagnostic (Am J Obstet Gynecol. 2004. 190. 1272-1277).
Confirm that symptoms are indeed due to estrogen deficiency by checking estradiol level to see if it is low. BE AWARE that many perimenopausal symptoms are associated with a low progesterone level, and not a low estrogen level.
Hormone replacement therapy (HRT)
Historically, three indications for hormone replacement therapy (HRT).
Relief of symptoms of menopause (hot flashes, urogenital atrophy, moodiness, insomnia). Therapy is effective in controlling symptoms in more than 90% of women (Acta ObstetGynecol Scand. 1997. 76. 442-448). As of 7/02, this is considered the only indication for HRT.
Prevention of osteoporosis/hip fractures. NO LONGER CONSIDERED A GOOD INDICATION AS OF JULY, 2002 BECAUSE THERE ARE OTHER PHARMACEUTICAL AGENTS WITH FEWER RISKS. SEE BELOW FOR DETAILS.
Prevention of coronary artery disease (CAD). AS OF JULY, 2002 BASED ON NEW DATA, NO LONGER CONSIDERED AND INDICATION. SEE BELOW FOR DETAILS.
Contraindications to estrogen use include pregnancy, breast cancer, estrogen-dependent cancer of any kind, unexplained vaginal bleeding, history of deep venous thrombosis, and active hepatocellular (liver) disease.
A17 year observational cohort study which evaluated data from 2755 breast cancer survivors, of whom 174 had used HRT after diagnosis, the mortality rate was actually lower in the group of HRT users (16 per 1000 women years versus 30 per 1000 women years), suggesting that HRT use in self-selected symptomatic breast cancer survivors may actually be safe (J Natl Cancer Inst. 2001. 93. 754-762).
HOWEVER, an open randomized trial which had enrolled 434 women was terminated early on 12/17/03 after 6 years when investigators determined that 26 women in the treatment group experienced recurrence, compared with only 7 in the control group. At this point in the study, 345 of the 434 women had at least one follow up visit and the recurrence occurred at a median of 2.1 years after initiation of HRT (Lancet. 2004. 363. 453-455).
Risk versus benefit - at 18 years of follow up in women in both arms (estrogen alone and estrogen plus progesterone) of the WHI trials (details in the sections just below), hormone therapy for 5.6 to 7.2 years did not affect all-cause, cancer, or cardiovascular mortality (JAMA. 2017. 318. 927-938, and ACP Journal Club. 2018. 168. JC4).
The risks of hormone replacement therapy include breast cancer, heart disease, stroke, thromboembolic events, Alzheimer’s disease, biliary tract surgery, ovarian cancer, urinary incontinence, and endometrial cancer if unopposed estrogen is used in women with an intact uterus.
Breast cancer:
Analysis of data from 46,355 postmenopausal women who participated from 1980-1995 in the Breast Cancer Detection Demonstration Project, a prospective observational study, shows a 1% increase in the relative risk of breast cancer with each year of estrogen use alone, but an 8% increase in the relative risk of breast cancer with each year of estrogen-progesterone use (JAMA. 2000. 283. 485-491). The increased risk of breast cancer is largely limited to current or recent users of hormone replacement therapy.
In the Women's Health Initiative estrogen plus progestin trial, a prospective, randomized trial of 16,608 postmenopausal women aged 50-79 with an intact uterus, the relative risk of breast cancer in women on combined HRT arm (Premarin 0.625 mg per day plus Provera 2.5 mg per day) was 1.26 (1.00-1.59) at 5.2 years of follow up, which translates into an absolute increase in risk of 8 cases of breast cancer per year in 10,000 women taking combined HRT. There were 38 cases in the treatment group versus 30 cases in the control group per 10,000 woman years. The mean age of the women in the study was 63 (JAMA. 2002. 288. 321-333).
Further analysis of the WHI data showed that the cancers in the combined HRT group were diagnosed at a more advanced stage, despite nearly 90% compliance with annual mammography screening in both the treatment and placebo groups. There were 199 invasive cancers in the treatment group compared with 150 in the placebo group (p=0.003). In the treatment group, average tumor diameter of the invasive cancers was 1.7 cm compared to 1.5 cm in the placebo group (p=0.04). In the treatment group, 25.9% of the cancers showed regional spread compared to 15.8% of the cancers in the placebo group (p=0.04). The histology and grade of the invasive tumors were similar in the treatment versus placebo group. The incidence of breast cancer in the treatment group was actually lower in the first two years of the study, but this may be due to the decreased sensitivity of mammography in this group (see below). There were not any easily identified subgroups at higher risk (JAMA. 2003. 289. 3243-3253).
Adjustment of the relative risk of breast cancer for baseline risk factors for breast cancer, including age, race/ethnicity, BMI, physical activity level, smoking, alcohol use, parity, OC use, family history of breast cancer, frequency of screening mammography, and vasomotor symptoms changed the relative risk slightly from 1.26 to 1.20 (0.94-1.53). This adjusted 95% confidence interval indicates that the increased risk of breast cancer in the Women's Health Initiative estrogen plus progestin trial is not statistically significant (Maturitas. 2006. 55. 103-115).
After a total mean follow-up of 11.0 years shows that cumulative breast cancer incidence was higher in the treatment arm, the cancers associated with HRT were more often invasive (p=0.004), more commonly node positive (p=0.03), and mortality from breast cancer in the treatment arm of the WHI appeared to be increased, as compared with the placebo arm (p=0.049). All cause mortality also appeared increased in the treatment arm (p=0.045) [JAMA. 2010. 304. 1684-1692 and editorial 1719-1720).
After a median of 20 years of follow up, the breast cancer incidence was higher in the treatment group (HR 1.28; 95% CI, 0.37 - 0.97) but there was no statistically significant difference in breast cancer mortality (HR 1.35; 95% CI, 0.94 - 1.95) [JAMA. 2020. 324. 369-380 and editorial 347-349].
In the HERS II trial, the increased relative risk of breast cancer in the combined HRT group at 6.8 years of follow-up was comparable in magnitude to the increased risk seen in the WHI (JAMA. 2002. 288. 58-66).
In the U.K. Million Women Study, a fair quality study, the relative risk of breast cancer in current users of combined HRT was 2.00 (1.91-2.09) compared with those who never used HRT (Lancet. 2003. 362. 419-427).
In the Black Women’s Health Study, a prospective epidemiologic study, data on 32,559 women over age 40, collected through biennial questionnaires, showed that use of estrogen alone and also estrogen plus progesterone was associated with an increased risk of breast cancer. The magnitude of risk increased with duration of use, and the association was stronger among leaner women (Arch Intern Med. 2006. 166. 760-765).
A population-based case-control study in three counties in western Washington state indicate that the increased risk of breast cancer associated with combined hormone replacement therapy is due primarily to an increase in invasive lobular cancers, and is independent of whether the progesterone component was taken in a continuous or a sequential manner(JAMA. 2003. 289. 3254-3263).
Hormone replacement therapy increases breast density in some women, and decreases both the sensitivity and the specificity of mammography screening for breast cancer (J Natl Cancer Inst. 1996. 88. 643-649; Ann Intern Med. 2003. 138. 168-175).
In the PEPI trial, the 16% to 26% of women taking estrogen plus progesterone who experienced an increase in breast density did so within the first year (Ann Intern Med. 1999. 130. 262-269).
Combinations of estrogen and progesterone increase breast density more than estrogen alone, and continuous combined regimens increase the density more than sequential combined regimens (J Clin Oncol. 1997. 15. 3201-3207).
A study of 103,770 women from Australia showed that the sensitivity of two year mammographic screening in women aged 50-69 given hormone replacement therapy was 64.3% whereas in those not given HRT it was 79.8% (Lancet. 2000. 355. 270-274).
In the WHI combined treatment arm, after 1 year 9.4% of women in the treatment group had an abnormal mammogram compared to 5.4% in the placebo group (p<0.001) [JAMA. 2003. 289. 3243-3253]. At 5.6 years of follow up, 35% of women in the treatment group had an abnormal mammogram compared to 23% in the placebo group (p<0.001). Furthermore, the breast biopsy rate (for abnormal mammograms) was 10% in the treatment group compared to 6.1% in the placebo group (p<0.001) [Arch Intern Med. 2008. 168. 370-377].
Discontinuing HRT 1 month or 2 months prior to screening mammography does NOT increase detection rates of breast cancer, based on data from a RCT, even though it is associated with decreased mammographic breast density. In this study, hot flashes recurred in 85% of those randomized to discontinue HRT in advance of screening mammography
In the WHI estrogen-alone arm, a prospective, randomized trial in 10,739 postmenopausal women aged 50-79, with prior hysterectomy, Premarin 0.625 mg per day was not associated with an increased risk of breast cancer at 6.8 years of follow up. There was actually a strong trend in this trial toward a decreased risk of breast cancer in the treatment group, with a relative risk of 0.77 (0.59-1.01). There were 7 fewer cases of breast cancer per 10,000 woman years in the treatment group, but this did not reach statistical significance (JAMA. 2004. 291. 1701-1712).
Further analysis of the data came to the same conclusion, but did show a significant increase in abnormal mammograms requiring ‘short-interval’ mammographic follow-up in the treatment group, compared with the control group (36.2% vs. 28.1%, p<0.001) [JAMA. 2006. 295. 1647-1657].
At the cessation of this arm of the trial, 7645 surviving participants (78%) who completed the intervention phase gave consent for continued observational follow up through a mean of 10.7 years. Median Premarin use in this subgroup was 5.9 years (but as per an accompanying editorial, median adherent time of women in this arm of the study was only 3.5 years). The decreased risk of breast cancer did persist in users of Premarin, with a relative risk at 10.7 years of follow up of 0.77 (0.62-0.95) [JAMA. 2011. 305. 1305-1314 and editorial 1354-1355].
After a median of 20 years of follow up, the breast cancer incidence was lower in the treatment group (HR 0.78; 95% CI, 0.65 - 0.93) and breast cancer mortality was lower in the treatment group (HR 0.60; 95% CI, 0.37 - 0.97) [JAMA. 2020. 324. 369-380 and editorial 347-349].
In the E3N-EPIC study, a prospective population study of 54,548 women in France born between 1925 and 1950, at an average follow up of 5.8 years, the relative risk of breast cancer in women using estrogens and synthetic progestins was 1.4 whereas the risk for those using estrogen plus micronized progesterone was 0.9. The risk for those using estrogen alone was 1.1 (not statistically significant). This data would suggest that micronized progesterone is safer from a breast health standpoint (Int J Cancer. 2005. 114. 448-454).
The USPSTF III update (2005) concluded that there is good quality evidence that combined HRT increases incidence of breast cancer, but its effects on breast cancer mortality remain uncertain, with the results of 2 good-quality cohort studies conflicting. Results are also conflicting in the two studies of estrogen alone and effects on breast cancer incidence (Ann Intern Med. 2005. 142. 855-860).
Data in Finland obtained by cross-referencing cancer diagnosis in the comprehensive registry with pharmacy data (to identify women older than age 50 who filled at least one prescription for estrogen-only hormone therapy and then eliminating women who took CEE) showed that estradiol use for 5 or more years was associated with an increased relative risk of breast cancer of 1.44, which translates into 20-30 additional cases of breast cancer per 10,000 women during 10 years of follow up. This data encompassed 648,022 patient-years of observation (Obstet Gynecol. 2006. 108. 1354-1360).
Heart disease:
Historically, prospective, observational data suggested that estrogen reduces the risk of heart disease by 35-50%. The most convincing of the observational studies was the Nurses Health Study which is methodologically sound and has followed 70,533 postmenopausal women over 20 years (Ann Intern Med. 2000. 133. 933-941). Premarin 0.625 mg daily increases HDL cholesterol 10-15% and decreases LDL cholesterol by 10-15%. Thus, there was a seemingly strong theoretic rationale to support the conclusion of a cardioprotective effect of HRT.
However, the combination HRT arm of the Women's Health Initiative estrogen plus progestin trial, a prospective, randomized trial of 16,608 postmenopausal women aged 50-79 with an intact uterus but without heart disease, was terminated two years early, in 2002, because an independent review board actually found an increase in the risk of heart disease! At 5.2 years the relative risk in the active treatment group was 1.29 (1.02-1.63) for heart attacks, which translates into an absolute risk of 7 additional heart attacks per 10,000 women per year. There were 37 events in the treatment group versus 30 events in the control group per 10,000 woman years. The mean age of the women in the study was 63 (JAMA. 2002. 288. 321-333). NOTE the subgroup analysis by years from menopause described just below (JAMA. 2007. 297. 1465-1477).
Furthermore, a randomized, prospective secondary prevention trial of postmenopausal women with established heart disease, the HERS Trial, which followed 2763 postmenopausal women for 4.1 years, found no overall difference in cardiovascular endpoints between the women treated with a combination of daily Premarin 0.625 mg and Provera 2.5 mg versus women treated with placebo (JAMA. 1998. 280. 605-613). In the HERS trial, an increase in cardiovascular risk was noted in the first year, postulated to be due to the thrombotic effects ofhormone replacement therapy, with a trend noted of decreased risk of CHD events with increased duration of use of HRT. However, the HERS II, which was unblinded and followed the women in the HERS trial for an additional 2.7 years, found that those who chose to take HRT did not sustain the lower rates of CHD noted in the final years of the HERS trial, such that after 6.8 total years of combined hormone replacement therapy, there was no reduction in CHD events in women with pre-existing CHD (JAMA. 2002. 288. 49-57).
Finally, another secondary prevention trial, the Estrogen Replacement in Atherosclerosis Trial, in which 309 women with angiographically verified coronary artery disease, followed for a mean of 3.2 years, had baseline and follow-up quantitative angiograms, showed no benefit of estrogen alone or estrogen plus progesterone on the progression of coronary artery disease (N Engl J Med. 2000. 343. 522-529).
However, in the Danish Osteoporosis Prevention Study (DOPS), a RCT of 1006 recently postmenopausal women, mean age of 50, 10 years of therapy with estradiol (and, if the woman had a uterus, norethisterone), “hormone replacement therapy reduced a composite of death, myocardial infarction, or heart failure after 10 years of therapy and another 6 years of follow up. The rates of stroke, venous thromboembolism, and cancer were the same after 10 years of therapy and another 6 years of follow up (BMJ. 2012. 345. e6409).
The USPSTF III (2002) concluded that HRT does not decrease and may in fact increase the risk of CHD, but its effects on CHD mortality remain uncertain (Ann Intern Med. 2002. 137. 834-839).
In the WHI estrogen-alone arm, a prospective, randomized trial in 10,739 postmenopausal women aged 50-79, with prior hysterectomy, the risk of heart disease was identical at 6.8 years of follow up in the treatment and control groups, with a relative risk in the treatment group of 0.91 (0.75-1.12) [JAMA. 2004. 291. 1701-1712].
A more detailed analysis of the same data concluded that there was no protection against either MI or coronary death, (Arch Intern Med. 2006. 166. 357-365), but subgroup analysis (described just below) showed a strong trend toward protection in the subgroup of women age 50-59 (Arch Intern Med. 2006. 166. 399-404).
Data at a mean of 10.7 years in the 7645 surviving participants (78%) who completed the intervention phase gave consent for continued observational follow up continued to show a lack of protective benefit of Premarin, with a relative risk in the intervention group of 0.97 (0.75-1.25). However, as reported in the 2006 paper looking at subgroup analysis of women age 50-59 at trial onset, data suggested a net reduction in heart disease in this younger subgroup of women (JAMA. 2011. 305. 1305-1314 and editorial 1354-1355).
Esterified estrogen (EE) might be safer than conjugated equine estrogen (CEE), based on a case-control study conducted in GHC, a staff-model HMO in which there was a formulary change – there was a trend in the first 6 months on estrogen toward a higher risk of MI with CEE (odds ratio 2.33, 95% confidence interval 0.93-5.82) [Arch Intern Med. 2006. 166. 399-404].
Estrogen and heart disease – there is data to support the concept that estrogen has different effects on blood vessels in younger postmenopausal women (50-59 years old), as compared with older postmenopausal women (NOTE the average age of women in the WHI study was 63 years old) [Editorial. HRT and the young heart. N Engl J Med. 2007. 356. 2639-3641].
Basic research and animal research on the vascular biology of estrogen supports this concept of a differential effect of estrogen in younger versus older postmenopausal women (Curr Opin Cardiol. 1994. 9. 619-626; N Engl J Med. 1999. 340. 1801-1811).
This concept of a differential effect of estrogen in younger versus older postmenopausal women can help to explain the discrepancy between observational data showing a protective effect of estrogen and the WHI showing no protective effect.
Subgroup analysis of the WHI estrogen-alone arm showed that there was a trend toward a lower risk of heart disease in women age 50-59 at baseline, with a relative risk of 0.63, but a confidence interval of 0.36-1.08. The number of women 50-59 in this study who received Premarin was 1637 and the number who received placebo was 1673. The relative risk in women age 60-69 was 0.94 and the relative risk in women age 70-79 was 1.11 (Arch Intern Med. 2006. 166. 399-404).
Subgroup analysis in which data from the WHI estrogen-alone arm and the WHI estrogen plus progesterone arm was combined and analyzed by decade after menopause found that (1) in women less than 10 years since menopause, the hazard ratio for CHD was 0.78, (2) in women 10-19 years since menopause, the hazard ratio for CHD was 1.10, and (3) in women greater than 20 years since menopause, the hazard ratio for CHD was 1.35 (JAMA. 2007. 297. 1465-1477).
However, re-analysis of the WHI data looking at the subgroup of women who initiated either Premarin alone or Premarin + Provera within 5 years of menopause found that the risk of CHD was slightly increased regardless of whether hormone therapy was instituted within 5 years of menopause or more than 5 years after menopause (Am J Epidemiol. 2009. 170. 12). This study differed from the above studies published in JAMA and Archives of Internal Medicine in that this study analyzed the time between menopause and first use of hormone therapy, whereas the above studies published in JAMA and Archives of Internal Medicine analyzed the time between menopause and study entry into the WHI (Commentary by Tori Hudson. Townsend Letter. 4/10. Pg 114).
In the WHI-CASS, an ancillary substudy of the WHI in which 28 of the 40 centers participated, coronary artery calcium scores were determined in 1064 women age 50-59, after a mean of 7.4 years of treatment and 1.3 years after the trial was completed. Surgically induced menopause occurred a mean of 11 years prior to the entry of this cohort of women in the WHI. Mean scores were lower in women receiving Premarin (83.1) than in women receiving placebo (123.1), and this 42% difference was statistically significant (p=0.02). In the subset of women adherent to estrogen treatment, calcium scores were 61% lower (N Engl J Med. 2007. 356. 2591-2602).
2007 Consensus statements of the North American Menopause Society and the IMS endorse the timing hypothesis and recognize the potential beneficial cardiovascular effects of estrogen in younger menopausal women (Menopause. 2007. 2. 168-182; Climacteric. 2007. 10. 181-194).
Stroke:
In the Women's Health Initiative estrogen plus progestin trial, the relative risk of stroke in women on combined HRT was 1.41 (1.07-1.85) at 5.2 years, which translates into an absolute increase in risk of 8 strokes per year in 10,000 women taking combined HRT. There were 29 strokes in the treatment group and 21 strokes in the control group per 10,000 woman years (JAMA. 2002. 288. 321-333).
Further analysis of the WHI data showed that the increased risk was for ischemic strokes, with a nonstatistically significant decrease in the risk of hemorrhagic stroke. Subgroup analysis indicated that the increased risk of stroke was present in all age groups, in all categories of baseline stroke risk, and in women with and without hypertension, prior history of cardiovascular disease, use of hormones, statins, or aspirin. Other risk factors for stroke, including smoking, high blood pressure, diabetes, lower use of vitamin C, and blood-based biomarkers of inflammation did not modify the effect of estrogen plus progesterone on stroke risk (JAMA. 2003. 289. 2673-2684).
In the Nurses Health Study involving 70,533 postmenopausal women, the risk for ischemic stroke was increased among hormone users compared with never-users (Ann Intern Med. 2000. 133. 933-941).
The USPSTF III (2002) concluded there is fair evidence that HRT increases the risk for stroke (Ann Intern Med. 2002. 137. 834-839).
In the WHI estrogen-alone arm, a prospective, randomized trial in 10,739 postmenopausal women aged 50-79, with prior hysterectomy, the study was terminated approximately one year early at 6.8 years of follow up because of an increased relative risk of stroke, 1.39 (1.10-1.77). This translates into 12 additional strokes per 10,000 woman years (JAMA. 2004. 291. 1701-1712). Recalculation of the hazard ratio in a subsequent publication yielded a result of 1.33, with a nominal 95% confidence interval of 1.05-1.66, but an adjusted 95% confidence interval of 0.97-1.99 (JAMA. 2007. 297. 1465-1477).
Esterified estrogen (EE) might be safer than conjugated equine estrogen (CEE), based on a case-control study conducted in GHC, a staff-model HMO in which there was a formulary change – there was a trend toward a higher risk of ischemic stroke with CEE (odds ratio 1.57, 95% confidence interval 0.98-2.53) [Arch Intern Med. 2006. 166. 399-404].
Transdermal estrogen might be safer than oral estrogen - in a nested case control study in postmenopausal women, "current use of hormone therapy with oral, but not transdermal, estrogen was associated with increased risk for stroke" (Stroke. 2016. 47. 1734-1741, as cited in ACP Journal Club. 2016. 165. JC46).
Thromboembolic events:
In the Women's Health Initiative estrogen plus progestin trial, the relative risk of thromboembolic events in women on combined HRT was 2.13 (1.39-3.25) at 5.2 years of follow-up, which translates into an absolute increase in risk of 18 thromboembolic events per year in 10,000 women taking combined HRT. There were 34 events in the treatment group versus 16 events in the control group per 10,000 woman years. The absolute increase in the risk of pulmonary embolus, a subcategory of thromboembolic events, was 8 per year (JAMA. 2002. 288. 321-333).
In the HERS trial of 2763 postmenopausal women on combined HRT, the relative risk of thromboembolic events was 2.66 at 4.1 years of follow-up (JAMA. 1998. 280. 605-613). In HERS II it dropped to 1.40, so the overall risk at 6.8 years of follow up was 2.08 (JAMA. 2002. 288. 49-57).
A meta-analysis of 12 studies demonstrate that current use of postmenopausal estrogen is associated with a relative risk of venous thromboembolism of 2.14, based on an additional 1.5 events per 10,000 women each year, from a baseline risk of 1.3 events per 10,000 woman-years. Six studies that reported risk according to duration of use found the highest risk in the first year of use, with a relative risk of 3.49 (Ann Intern Med. 2002. 136. 680-690). As an aside, tamoxifen (Nolvadex) is associated with a threefold increased risk of pulmonary embolism (J Natl Cancer Inst. 1998. 90. 1371-1388) and raloxifene (Evista) is associated with a threefold increased risk of venous thromboembolism (JAMA. 1999. 281. 2189-2197).
The USPSTF III (2002) concluded there is good evidence that HRT increases the risk for venous thromboembolism (Ann Intern Med. 2002. 137. 834-839).
Transdermal estradiol does NOT appear to increase the risk of thromboembolic events, based on data from a case-control study. This may be due to the reduced plasma estrone-to-estradiol ratios achieved by transdermal preparations, or it may be due to less alteration of clotting factors (Lancet. 2003. 362. 428-432).
Esterified estrogen does NOT appear to increase risk of thromboembolic events, based on data from a case control study (JAMA. 2004. 292. 1581-1587).
In the WHI estrogen-alone arm, there was a nonsignificant increase in the risk of pulmonary embolus, with a relative risk of 1.34 (0.87-2.06) [JAMA. 2004. 291. 1701-1712]. Further analysis of the data showed that there was a significant increased risk of DVT of 1.47 (1.06-2.06), and that the risk venous thromboembolism (DVT + pulmonary embolus) was highest in the first two years on conjugated estrogens (Arch Intern Med. 2006. 166. 772-780).
Alzheimer's disease:
Historically, epidemiologic data subjected to a systematic review and meta-analysis (JAMA. 2001. 285. 1489-1499) and prospective, observational data (Lancet. 1996. 348. 429-432; Neurology. 48. 1517-1521; JAMA. 2002. 288. 2123-2129) suggested a benefit. The magnitude of the benefit appeared to be about a 30% reduced risk. In the most recent prospective study (JAMA) only former users and long time current users (>10 years) appeared to benefit, suggesting that there may be a critical period for use prior to the onset of preclinical disease.
Additionally, strong biological evidence supports the beneficial effects of estrogen on the brain, including neurotrophic effects, reductions in beta-amyloid accumulation, enhanced neurotransmitter release and action, and protection against oxidative damage (J Clin Endocrinol Metab. 1999. 84. 1790-1797; Nat Med. 1998. 4. 447-451). Estrogen receptors are located throughout the brain, especially in areas involved in learning and memory (J Comp Neurol. 1997. 388. 507-525).
The USPSTF III (2002) concluded there is insufficient evidence to determine whether HRT reduces the risk for dementia or cognitive dysfunction in otherwise healthy women (Ann Intern Med. 2002. 137. 834-839).
However, in the Women's Health Initiative Memory Study (WHIMS), a RCT of women over age 65 and free of probable dementia, recruited from 39 of the 40 WHI clinical centers, hormone replacement therapy was associated with an increased risk of dementia and a more rapid deterioration in cognitive function. The results are as follows:
In the Premarin plus Provera group, in which 4532 community-dwelling women aged 65-79 were enrolled, at 4.05 years of follow up the relative risk for probable dementia was 2.05 (1.21-3.48, p=0.01). This translates into an absolute risk of 23 additional cases of dementia per 10,000 women per year (45 cases vs. 22 cases per 10,000 women per year). Alzheimer disease was the most common classification of dementia in both the treatment group and the placebo group. Subgroup analysis found similar results among multiple subgroups evaluated. Treatment effects on mild cognitive impairment did not differ between the treatment group and the placebo group, with 63 cases per 10,000 women per year in the treatment group compared with 59 cases in the placebo group, for a relative risk 1.07 (0.74-1.55) [JAMA. 2003. 289. 2651-2662].
In the Premarin group, in which 2947 community-dwelling women aged 65-79 were enrolled, at 5.4 years of follow up the relative risk for probable dementia was 1.49 (0.83-2.66). This translates into an absolute risk of 9 additional cases of dementia per 10,000 women per year (28 cases vs. 19 cases per 10,000 women per year). Mild cognitive impairment was more common in the treatment group, with 76 cases vs. 58 cases, for a relative risk of 1.34 (0.95-1.89) [JAMA. 2004. 291. 2947-2958].
When data from the Premarin plus Provera trial were pooled with data from the Premarin trial, as per the original WHIMS protocol, the relative risk for probable dementia was 1.76 (1.19-2.60, p=0.005). Re-analysis of the data after excluding participants with baseline low Modified MMSE scores did not change the statistical outcomes. The relative risk for mild cognitive impairment when the data was pooled was 1.25 (0.97-1.60) [JAMA. 2004. 291. 2947-2958].
In the Premarin plus Provera group, at 4.2 years of follow up, there was no difference in global cognitive function between the treatment group and the placebo group, as measured by the Modified MMSE exam. The scores tended to improve in the treatment and placebo group, actually with slightly greater improvement in the placebo group which was statistically significant (p=0.03) but not considered to be clinically significant (JAMA. 2003. 289. 2663-2672).
In the Premarin group, at 5.4 years of follow up, global cognitive function deteriorated more in the treatment group than the placebo group (p=0.04) [JAMA. 2004. 291. 2959-2968].
When the data from the Premarin plus Provera trial were pooled with data from the Premarin trial, global cognitive function deteriorated more in the treatment group than the placebo group (p=0.006) [JAMA. 2004. 291. 2959-2968].
Randomized trials of Premarin 0.625 mg a day or higher for 4-12 months in women diagnosed with mild to moderate Alzheimer’s showed no benefit and possible harm, as per an editorial (JAMA. 2004. 291. 3005-3007). These results were all reported after the initiation of WHIMS (Neurology. 2000. 54. 295-301; JAMA. 2000. 283. 1007-1015; Neurology. 2000. 54. 2061-2066; Cochrane Database Syst Rev. 2002. CD0031220.
Gallbladder disease:
Risk is increased by a factor of 1.8 in the short term and 2.5 in the long term (Obstet Gynecol. 1994. 83. 5-11).
In HERS II the relative risk of biliary tract surgery was 1.48 at 6.8 years of follow up (JAMA. 2002. 288. 49-57).
The Women's Health Initiative estrogen plus progestin trial at 5.6 years of follow-up found that the relative risk of a gallbladder event (hospitalization for gallbladder disease or gallbladder-related procedure) was 1.59 (1.28-1.97) which translates to an excess of 20 events per 10,000 person years (55 vs. 35). In the WHI estrogen-alone arm, the relative risk of a gallbladder event was 1.67 (1.35-206) which translates to an excess of 31 events per 10,000 person years (78 vs. 47) [JAMA. 2005. 293. 330-339].
The USPSTF III (2002) concluded there is fair evidence that HRT increases the risk for cholecystitis (Ann Intern Med. 2002. 137. 834-839).
Ovarian cancer:
One study found no effect of HRT on ovarian cancer mortality (Int J Cancer. 1996. 67. 327-332).
However, a good-quality cohort study reported an increased risk of ovarian cancer mortality in women who had taken hormone replacement therapy for 10 years or more (JAMA. 2001. 285. 1460-1465).
Analysis of data from 46,355 postmenopausal women who participated from 1979-1998 in the Breast Cancer Detection Demonstration Project, a prospective observational study, shows an increased risk of ovarian cancer with estrogen-only replacement therapy, but no increased risk in women taking combined estrogen-progesterone hormone replacement therapy. The relative risk of ovarian cancer in women who took estrogen-onlyreplacement therapy for 10-19 years was 1.8, and the relative risk of ovarian cancer in women who took estrogen-onlyreplacement therapy for 20 or more years was 3.2 (JAMA. 2002. 288. 334-341).
The Women's Health Initiative estrogen plus progestin trial at 5.6 years of follow-up found a trend toward an increased incidence of invasive ovarian cancer in women on estrogen/progesterone, with a hazard ration of 1.58 [0.77-3.24] (JAMA. 2003. 290. 1739-1748).
In a prospective observational study of 910,000 women over age 50 in Denmark followed for an average of 8 years, current users of HRT had a RR of ovarian cancer of 1.38 and previous users had a RR of 1.15 (JAMA. 2009. 302. 298).
The USPSTF III (2002) concluded evidence was insufficient to determine the effect of HRT on ovarian cancer (Ann Intern Med. 2002. 137. 834-839).
Urine incontinence (UI):
In women continent at baseline in the WHI, the relative risk of stress UI at one year was1.87 (1.61-2.18) in women on estrogen plus progesterone and 2.15 (1.77-2.62) in women on estrogen alone (JAMA. 2005. 293. 935-948)..
In women continent at baseline in the WHI, the relative risk of mixed UI at one year was1.49 (1.10-2.01) in women on estrogen plus progesterone and 1.79 (1.26-2.53) in women on estrogen alone (JAMA. 2005. 293. 935-948)..
In women continent at baseline in the WHI, the relative risk of urge UI at one year was neutral at 1.15 (0.99-1.34) in women on estrogen plus progesterone but 1.32 (1.10-1.58) in women on estrogen alone (JAMA. 2005. 293. 935-948)..
Among women experiencing urine incontinence at baseline in the WHI, the frequency worsened at one year in both the estrogen plus progesterone group and the estrogen alone group (JAMA. 2005. 293. 935-948)..
Women receiving either estrogen plus progesterone or estrogen alone were more likely than women receiving placebo to report at one year that urine incontinence limited their daily activities and bothered them (JAMA. 2005. 293. 935-948).
Endometrial cancer:
Results of a meta-analysis of 29 good-quality observational studies reported a relative risk of 2.3 for users of unopposed estrogen (Obstet Gynecol. 1995. 85. 304-313). However, the risk of death is probably not increased because of early detection. Hysterectomy is usually curative - women on unopposed estrogen probably face a lifetime probability of 20% of requiring hysterectomy; addition of progesterone eliminates the risk. In women on unopposed estrogen, yearly endometrial biopsy is quite effective for early detection. Transvaginal ultrasound may eliminate the need for many endometrial biopsies.
The Women's Health Initiative estrogen plus progestin trial (JAMA. 2002. 288. 321-333) and the HERS II trial (JAMA. 2002. 288. 58-66) showed that combined estrogen/progestin treatment does not increase the risk of endometrial cancer.
The USPSTF III (2002) concluded that unopposed estrogen, but not estrogen-progestin therapy, increases the risk for endometrial cancer (Ann Intern Med. 2002. 137. 834-839).
Lung cancer: Post hoc analysis of the Women's Health Initiative estrogen plus progestin trial showed that at a mean of 7.9 years of follow up, estrogen plus progesterone increased lung cancer mortality but not incidence in postmenopausal women. Incidence was increased, but this increase did not reach statistical significance. The increased deaths were specifically detected for non-small cell lung cancer and women over age 60 (Lancet. 2009. 374. 1243-1251).
Cataract surgery: In a 98 month study in 30,861 postmenopausal women, those who had ever used HRT had a 14% higher risk for cataract extraction, and current users of HRT had a18% higher risk, compared with women who never used HRT (Lindblad E. Ophthalmology. Epub 1/4/10).
Nephrolithiasis – in the Women’s Health Initiative, Premarin was associated with a RR of nephrolithiasis of 1.21, relative to placebo (39 events per 10,000 person years as compared with 34 events per 10,000 person years). When the data was analyzed for the subgroup of women who adhered to therapy, RR was 1.39. The increased risk was independent of coadministration of Provera (Arch Intern Med. 2010. 170. 1678-1685).
The benefits of hormone replacement therapy include relief of symptoms of menopause, prevention of hip fracture, a decreased risk of colon cancer, and a decreased risk of diabetes.
Mortality in younger women – Bayesian meta-analysis of 19 RCT’s (n=16,283; mean follow up 5 years) and 8 observational studies (n=212,171; mean follow up 14 years) concluded that consistent evidence from randomized and observational studies shows that hormone therapy reduces mortality in younger (age < 60) postmenopausal women. The absolute risk reduction was 0.75%, NNT 134 (Am J Med. 2009. 122. 1016-1022).
Hip fracture:
In the Women's Health Initiative estrogen plus progestin trial, the relative risk of hip fracture in women on combined HRT was 0.66 (0.45-0.98) at 5.2 years of follow up, which translates into an absolute risk reduction of 5 hip fractures per year in 10,000 women taking combined HRT. There were 10 cases in the treatment group versus 15 cases in the control group per 10,000 woman years (JAMA. 2002. 288. 321-333).
Further analysis of data in the Women's Health Initiative estrogen plus progestin trial determined at 5.6 years of follow-up that estrogen plus progesterone increased bone mineral density and decreased the risk of fracture (hip + vertebrae + wrist) independent of baseline risk. The relative risk for any fracture was 0.76 in the treatment group (0.69-0.83). There were 47 fewer total fractures per 10,000 women per year in the treatment group (JAMA. 2003. 290. 1729-1738).
The HERS II trial surprisingly found no reduction in fractures in HRT users (JAMA. 2002. 288. 58-66).
Historical note - in terms of effect on bone density, some studies suggest that the response to estrogens may be greatest in those women furthest from menopause (ObstetGynecol. 1990. 76. 290-295; J Clin Endocrinol Metab. 1995. 80. 776-782).
In a RCT of 40 women with a mean age of 62 and with established osteoporosis, BMD at the lumbar spine increased by 10.6% (p<0.01) and BMD at the hip increased by 5.5% (p<0.1) after two years of conjugated estrogen 0.625 mg daily (Obstet Gynecol. 1990. 76. 290-295).
In a RCT of 75 women with a mean age of 65 and with established osteoporosis, BMD at the lumbar spine increased by 5.3% (p<0.001) after one year of transdermal estradiol 0.1mg/week (Ann Intern Med. 1992. 117. 1-9).
Ultra-low dose micronized 17 beta-estradiol, 0.25 mg/day shown in a 3 year RCT in 167 healthy, community-dwelling women over age 65 to increase density in the hip, spine, and total body, and to reduce bone turnover, as measured by metabolic markers. This lower dose is presumably safer, although this is unknown. The effect of this low dose on osteoporotic fractures is also unknown (JAMA. 2003. 290. 1042-1048).
In the WHI estrogen-alone arm, the relative risk of hip fracture was decreased with treatment, relative risk 0.61 (0.41-0.91). This translated into 6 fewer hip fractures per 10,000 women treated per year. The relative risk of any fracture was 0.70 (0.63-0.79), translating into 56 fewer fractures per 10,000 women treated per year (JAMA. 2004. 291. 1701-1712).
The USPSTF III update (2005) concluded that there is good quality evidence that combined HRT increases bone density (Ann Intern Med. 2005. 142. 855-860).
Colorectal cancer:
In the Women's Health Initiative estrogen plus progestin trial, the relative risk of colon cancer in women on combined HRT was 0.63 (0.43-0.92) at 5.2 years of follow up, which translates into an absolute risk reduction of 6 cases of colon cancer per year in 10,000 women taking combined HRT. There were 10 cases in the treatment group versus 16 cases in the control group per 10,000 woman years (JAMA. 2002. 288. 321-333).
In the HERS II trial, the relative risk of colon cancer was reduced but not statistically significant at 0.81 [0.46-1.45] (JAMA. 2002. 288. 58-66).
The USPSTF III (2002) concluded there is fair evidence that HRT reduces colorectal cancer incidence (Ann Intern Med. 2002. 137. 834-839).
In the WHI estrogen-alone arm, the risk of colon cancer was unchanged with treatment, relative risk 1.08 (0.75-1.55) [JAMA. 2004. 291. 1701-1712].
Diabetes:
In the HERS Trial, a secondary prevention trial in 2763 postmenopausal women, HRT was protective in the subgroup of 2029 women who did not have diabetes at baseline, reducing the risk of onset of diabetes by 35% from 9.5% in the placebo group to 6.2% in the HRT group. The number needed to treat to prevent one case of diabetes was 30 (Ann Intern Med. 2003. 138. 1-9).
A subanalysis of the Women's Health Initiative estrogen plus progestin trial found that women on HRT had a relative risk of 0.79 (0.67-0.93) for development diabetes after 5 years (277 cases vs. 324 cases). This translates into 15 fewer cases of diabetes per 10,000 women per year in the treatment group. In the subgroup of women compliant with HRT throughout the follow up period, the decreased risk was 33% (Diabetologia. 2004. 47. 1175-1187).
A subanalysis of the WHI estrogen-alone arm found that women on estrogen had a relative risk of 0.88 (0.77-1.01) for development of diabetes after 6 years, a difference which was not significant. In the subgroup of women compliant with estrogen throughout the follow up period, the decreased risk was 27%, a difference which was significant (Diabetologia. 2006. 49. 459-468).
Historical note - estrogen was historically considered advisable in women after a hysterectomy because of observational data showing that estrogen therapy post-hysterectomy adds an average of 1.1 years of life. However, in the WHI estrogen-alone arm, the estimated excess risk for all monitored events in the global index was 2 events per 10,000 woman years, indicating neither harm nor benefit from estrogen monotherapy (JAMA. 2003. 289. 2663-2672).
Side effects of hormone replacement therapy:
Side effects of estrogen include bloating, headaches, and breast tenderness in 5-10% women. Symptoms are often self limited after a few months. If intolerance, try half dose for 2 months, then increase to full dose.
Side effects of progesterone include breast tenderness, bloating, weight gain, irritability, and depression. If side effects occur with one formulation, then try a different formulation.
Modes of administration of hormone replacement therapy:
Estrogen can be given by mouth, by patch, or by injection.
Injections are expensive, inconvenient, and produce unnecessarily high blood levels.
The advantage of a patch is that it avoids “first pass” metabolism of estrogen in the liver.
Progesterone historically was always given by mouth, is now also available in patch form.
Formulations of estrogen (and usual replacement doses) include:
Conjugated estrogens (Premarin) 0.625 mg
Estropipate (Ogen, Orto-Est) 0.625 mg
Micronized estradiol (Estrace) 0.5 mg
Transdermal estradiol (Climara, Estraderm, Vivelle) 0.5 mg
Ethinyl estradiol 0.05 mg
Esterified estrogens 0.3 mg
Formulations of progesterone include:
Provera (medroxyprogesterone) 2.5-10 mg/day
Norethindrone 0.35mg/day
Norgestrol 0.075 mg/day
Micronized progesterone (Prometrium) 200 mg/day
Dosage options if a decision is made to treat (trend toward decreased doses and increased duration of Provera treatment).
Premarin 0.625 mg days 1-25, Provera 5-10 mg days 14-25.
Premarin 0.625 mg days 1-31, Provera 5-10 mg days 1-12.
Withdrawal bleeding is acceptable from days 5-15.
If bleeding occurs before day 10, increase progesterone dose.
Premarin 0.625 mg + Provera 2.5 mg every day.
There is no long term withdrawal bleeding, but 1/3 of patients have unpredictable bleeding for the first year.
Data in diabetics indicates that this regimen causes more disruption of insulin levels and blood sugar control.
Premarin 0.3 mg + Provera 1.5 mg every day.
The Women's HOPE Study, which enrolled over 2600 healthy but symptomatic postmenopausal women with an intact uterus showed both an improvement in vasomotor symptoms and vaginal atrophy at 3 weeks (Fertil Steril. 2001. 75. 1065-1079) and an increase in bone mineral density of the spine, hip, and total body at 24 months, compared to baseline (JAMA. 2002. 287. 2668-2676). The lower dose provided endometrial protection similar to standard doses despite higher rates of amenorrhea (Fertil Steril. 2001. 75. 1080-1087).
Older, less rigorous studies had shown no benefit from doses of conjugated equine estrogen this low.
HRT and quality of life:
In the Postmenopausal Estrogen/Progesterone Intervention Trial, HRT was not associated with an improvement in quality of life (Obstet Gynecol. 1998. 92. 982-988).
In the HERS trial, a RCT with 2763 patients with known coronary artery disease, women with the symptom of flushing or hot flashes showed a measurable improvement in emotional quality of life on HRT, but women without flushing at baseline showed a measurable deterioration in quality of life on HRT.
In the Women's Health Initiative estrogen plus progestin trial, in which only 12% of women reported symptoms of menopause at baseline, HRT did not have a clinically meaningful effect on health-related quality of life, except in the subgroup with symptoms (New Engl J Med. 2003. 348. 1839-1854).
In the Women’s Health Initiative estrogen only arm in postmenopausal women with hysterectomy, oral conjugated equine estrogens (Premarin) did not improve any of the HRQOL variables at one year (Arch Intern Med. 2005. 165. 1976-1986).
HRT discontinuation and quality of life: (JAMA. 2005. 294. 183-193 and 245-246)
Data below based on information collected in surveys mailed 8-12 months after the stop date of the estrogen + progesterone arm of the WHI to each of the 8405 women who were still taking study pills (either HRT or placebo) when the study was terminated 7/8/02. 9351 women were enrolled in this arm of the study; 89.9% were still taking pills on the study termination date. Mean age of these women at trial stop date was 69.1 years, and women had been taking pills on average for 5.7 years.
In women who discontinued estrogen + progesterone, 21.2% experienced moderate to severe vasomotor symptoms. In this group, 55.5% of the women with these symptoms at baseline reported them upon withdrawal of estrogen + progesterone.
In women who discontinued placebo, 4.8% experienced moderate to severe vasomotor symptoms. In this group, 21.3% of the women with these symptoms at baseline reported them upon withdrawal of placebo.
In women who discontinued placebo, 38.3% reported pain or stiffness – note these symptoms may not be caused by an estrogen-deficient state.
Among women who did not have vasomotor symptoms at baseline, only 6.4% reported symptoms upon discontinuation of estrogen + progesterone.
Among the 63.3% of WHI participants who reported at least one moderate or severe symptom after discontinuation of estrogen + progesterone, lifestyle changes such as increasing exercise, practicing yoga, meditation, or breathing exercises, and using fans were perceived as helpful.
One can speculate that part of the reason for symptoms upon withdrawal of estrogen + progesterone is due to abrupt discontinuation and that tapering the dose may decrease the likelihood of withdrawal symptoms.
Bio-identical hormone replacement therapy
By definition, this refers to administration of synthetic estrogen or progesterone which is identical in biologic structure to what the body produces.
Bio-identical hormone replacement therapy is a more accurate term than natural hormone replacement therapy.
Bio-identical hormones are synthesized from beta-sitosterol, which is derived from soybean, or from diosgenin, which is extracted from wild yam
Some individuals will experience fewer side effects with bio-identical HRT, compared with Premarin and Provera
Effectiveness of bio-identical HRT:
A systematic review and meta-analysis of 32 treatment trials found that the use of Premarin (conjugated equine estrogens) and oral and transdermal 17-beta-estradiol have consistent and comparable effects on the treatment of hot flashes in menopausal women with symptoms (JAMA. 2004. 291. 1610-1620).
Bio-identical progesterone should be as effective as synthetic progestins at prevention of endometrial cancer as long as an adequate serum level is achieved.
Micronized oral progesterone shown to be effective in the PEPI trial.
Three studies have shown that if progesterone cream is used instead of oral progesterone along with estrogen replacement therapy, serum levels of progesterone are insufficient to prevent estrogenic stimulation of the uterus (Lancet. 1998. 351. 1255-1256; Lancet. 1999. 354. 1447-1448; Maturitas. 2002. 41.1-6).
Wild yam products which contain only diosgenin, a precursor to progesterone, are not effective because diosgenin cannot be converted to hormonally active progesterone in the human body.
Safety of bio-identical HRT:
Hypothesized as safer than Premarin and Provera, and there are some in-vitro studies to indicate greater safety (J Complem Med. 2004. 3[5]. 44-45), but no clinical trial data comparing bio-identical HRT to Premarin and Provera.
Micronized oral progesterone has been shown in the PEPI trial to have a less adverse effect on lipid profiles than Provera, but this is a surrogate endpoint.
There was no increase in risk of breast cancer in the estrogen arm of the French MISSION study, at an average follow up of 8 years. This is a prospective observational study (Gynecologic Endocrinology. 2007. 23. 391-397).
Progesterone cream
Dr. John Lee published data on transdermal progesterone cream and osteoporosis reversal (Lancet. 1991. 336. 1327).
A one year RCT in 102 postmenopausal women looking at the efficacy of 1/4 teaspoon or 20 grams of progesterone cream per day, rubbed on the skin, found that 83% of treatment subjects noted improvement or resolution of vasomotor symptoms compared with 19% of placebo control subjects(p<0.001). Eight women in this study had induction of postmenopausal bleeding by progesterone cream. There was no benefit in terms of bone density. Both placebo and control groups took a multivitamin and calcium 1200 mg per day (Obstet Gynecol. 1999. 94. 225-228).
Proponents of progesterone creams recommend using only those creams which contain at least 400 mg of progesterone per ounce of cream. Over the counter creams contain 50 mg/ounce or less.
Until there is further research into the risks of progesterone cream in terms of induction of postmenopausal bleeding, some advocates suggest 3 weeks on and one week off as a safer regimen.
NOTE that wild yam preparations contain diosgenin, which can be converted to progesterone in a laboratory, but not in the human body.
Page Updated August 2, 2020