From Dynamed
Hormonal replacement therapy (HRT)
Updated 2011 Apr 13 12:00:00 AM: no significant differences in most outcomes after 10 years comparing estrogen vs. placebo in postmenopausal women (JAMA 2011 Apr 6) view update
DHEA appears to have no effect on bone mineral density, lower extremity strength or function in older frail women (J Am Geriatr Soc 2010 Sep) view update
HRT associated with increased risk of breast cancer at 11 years (JAMA 2010 Oct 20) view update
Related Summaries:
for issues regarding addition of progestin to estrogen, cyclic vs. continuous therapy, preparations and doses, see HRT doses and preparations
HRT and cardiovascular disease
HRT and osteoporosis
HRT and venous thromboembolism
HRT and breast cancer
Menopause
Overview:
also called estrogen replacement therapy (ERT), sometimes HRT used to imply addition of progestins
estrogen proven to relieve menopausal symptoms
progestin should be added if patient has uterus to reduce risk for endometrial cancer
long-term use of combined estrogen/progestin (HRT) has more risks than benefits in healthy postmenopausal women
risks include breast cancer, myocardial infarction, stroke, and venous thromboembolism
benefits include prevention of osteoporotic fractures and colorectal cancer
no differences in overall mortality or quality of life
long-term use of estrogen alone has benefits that do not outweigh risks overall
risks include stroke and possibly deep vein thrombosis
benefits include prevention of osteoporotic fractures and possibly lower risk of invasive breast cancer
no differences in overall mortality, cardiovascular disease, colorectal cancer or quality of life
discontinuation of HRT associated with recurrence or onset of menopausal symptoms
Evidence Summaries
Cochrane review:
hormone replacement therapy (HRT) not indicated for routine management of chronic disease (grade A recommendation [consistent high-quality evidence])
based on Cochrane review
systematic review of 19 randomized placebo-controlled trials lasting ≥ 1 year in 41,904 perimenopausal or postmenopausal women
all statistically significant results based on 2 largest trials
combined continuous HRT use in relatively healthy women
associated with increased risk of
venous thromboembolism (up to 5.6 years)
coronary events (up to 4 years)
stroke (after 3 years)
breast cancer (after 5 years)
gallbladder disease (after 5.6 years)
dementia among women > 65 years old (up to 4.2 years)
associated with decreased risk of
fractures (at 5.6 years)
decreased risk of colon cancer (after 5 years)
subgroup analysis of 2,839 women aged 50–59 years compared with similar size placebo group found increased risk for venous thromboembolism, though absolute risk was low
in women with cardiovascular disease, long-term combined continuous HRT increased risk of venous thromboembolism
long-term estrogen-only therapy increased risk of stroke, fractures, and gallbladder disease
Reference – Cochrane Database Syst Rev 2008 Oct 8;(4):CD004143, commentary on earlier version can be found in Evidence-Based Medicine 2006 Jan-Feb;11(1):22, Cochrane for Clinicians summary of earlier version can be found in Am Fam Physician 2006 Nov 1;74(9):1501
Women’s Health Initiative (WHI) trial:
long-term use of HRT has more risks than benefits in healthy postmenopausal women
Women’s Health Initiative (WHI) is largest randomized trial of HRT, stopped early due to risks exceeding benefits
16,608 postmenopausal women ages 50–79 were randomized to HRT (equine estrogens 0.625 mg/medroxyprogesterone acetate 2.5 mg [Prempro]) vs. placebo orally daily for mean 5.2 years (range 3.5–8.5 years)
discontinuation of study drug occurred in 42% HRT and 38% placebo patients, while addition of HRT through personal clinician was started in 6.2% HRT and 10.7% placebo patients; intention-to-treat analysis was performed, so results likely underestimate “per protocol” results
no differences in overall mortality or endometrial cancer
results reported as annualized percentages (event rates per year of therapy)
NNH or NNT reported as number treated with HRT for 1 year
adverse outcomes more common with HRT than placebo
0.37% vs. 0.3% coronary heart disease events (NNH 1,428), differences related to nonfatal myocardial infarctions
0.29% vs. 0.21% stroke (NNH 1,250)
0.38% vs. 0.3% invasive breast cancer (NNH 1,250)
0.34% vs. 0.16% venous thromboembolic event (NNH 555)
0.16% vs. 0.08% pulmonary embolism (NNH 1,250)
1.7% vs. 1.51% absolute excess in risk (NNH 526) based on global index of death, coronary heart disease event, stroke, pulmonary embolism, breast cancer, endometrial cancer, colorectal cancer, or hip fracture
adverse outcomes less common with HRT than placebo
0.1% vs. 0.16% colorectal cancer (NNT 1,667)
0.1% vs. 0.15% hip fracture (NNT 2,000)
0.09% vs. 0.15% vertebral fracture (NNT 1,429)
1.47% vs. 1.91% any osteoporotic fracture (NNT 228)
Reference – JAMA 2002 Jul 17;288(3):321
editorial can be found in JAMA 2002 Jul 17;288(3):366, commentary can be found in BMJ 2008 May 10;336(7652):1033 (commentary can be found in BMJ 2008 May 24;336(7654):1148)
for considerable commentary Click for Details
editorial commentary can be found in BMJ 2002 Jul 20;325(7356):113 full-text (correction can be found in BMJ 2002 Aug 24;325(7361):435), BMJ 2002 Nov 2;325(7371):1036 full-text, BMJ 2002 Nov 23;325(7374):1243 full-text
HRT did not have clinically meaningful effect on health-related quality of life in WHI trial (N Engl J Med 2003 May 8;348(19):1839), editorial can be found in N Engl J Med 2003 May 8;348(19):1835, commentary can be found in ACP J Club 2003 Nov-Dec;139(3):60, Am Fam Physician 2004 Jan 15;69(2):423, N Engl J Med 2004 Feb 5;350(6):622
HRT may increase risk of ovarian cancer
WHI trial had mean follow-up 5.6 years
annualized risk of invasive ovarian cancer was 0.04% with HRT vs. 0.03% with placebo (not statistically significant)
HRT did not increase risk of endometrial cancer (0.06% vs. 0.07% annualized risk) but, due to vaginal bleeding, more women required endometrial biopsies (33% vs. 6%, NNH 3.7)
Reference – JAMA 2003 Oct 1;290(13):1739, commentary can be found in JAMA 2004 Jan 7;291(1):42
3 years after halt of WHI trial, HRT associated with higher rates of cancer but not cardiovascular disease
based on postintervention phase of randomized trial
15,730 women from WHI trial were followed for mean 2.4 years after halt of trial
8,052 women from HRT group
7,678 women from placebo group
comparing HRT vs. placebo during postintervention phase
all-cause mortality 2.9% vs. 2.6% (not significant)
malignancies (including invasive breast cancer, endometrial or colorectal cancer) in 3.5% vs. 2.8% (p < 0.05, NNH 142)
invasive breast cancer in 0.09% vs. 0.08% (not significant)
total cardiovascular disease events in 4.3% vs. 4.2% (not significant)
any fracture in 4.2% vs. 4.5% (not significant)
Reference – JAMA 2008 Mar 5;299(9):1036, commentary can be found in JAMA 2008 Jun 18;299(23):2744, ACP J Club 2008 Aug 19;149(2):11
HRT not associated with increased risk for lung cancer but may increase risk of mortality from lung cancer
based on postintervention phase of WHI randomized trial
16,608 women from WHI trial were followed for mean 2.4 years after halt of trial
comparing HRT vs. placebo during postintervention phase
lung cancer in 0.16% vs. 0.13% per year (not significant)
non-small-cell lung cancer in 0.14% vs. 0.11% per year (not significant)
small-cell lung cancer in 0.02% vs. 0.02% per year (not significant)
death from lung cancer in 0.11% vs. 0.06% per year (p = 0.01)
death from non-small-cell lung cancer in 0.09% vs. 0.05% per year (p = 0.004,
death from small-cell lung cancer in 0.02% vs. 0.01% per year (not significant)
poorly differentiated non-small-cell cancer in 0.04% vs. 0.02% (p = 0.03)
non-small-cell lung cancer distant metastases in 0.06% vs. 0.03% (p = 0.04)
no significant differences in local or regional non-small-cell lung cancer or in any other cancer grade of non-small-cell lung cancer
Reference – Lancet 2009 Oct 10;374(9697):1243, editorial can be found in Lancet 2009 Oct 10;374(9697):1217, commentary can be found in Lancet 2010 Jan 9;375(9709):117
vasomotor symptoms may worsen after stopping estrogen therapy
based on postintervention phase of WHI randomized trial
3,496 women from WHI trial continued assigned intervention to trial closure and completed symptom surveys 7 years after study initiation and after stopping intervention (306 days after trial closure)
after stopping intervention vasomotor symptoms reported by 9.8% estrogen group vs. 3.2% placebo group (p < 0.05)
among women with no moderate or severe symptoms at baseline, hot flashes reported after stopping intervention in 7.2% estrogen group vs. 1.5% placebo group
Reference – Menopause 2010 Sep-Oct;17(5):946
hormonal replacement therapy (HRT) associated with increased risk of breast cancer at 11 years
based on additional follow up of WHI trial
12,788 women with no prior hysterectomy followed for mean 11 years
comparing HRT vs. placebo
invasive breast cancer incidence 0.42% per year vs. 0.34% per year (hazard ratio [HR] 1.25, 95% CI 1.07–1.46)
breast cancer mortality 0.03% per year vs. 0.01% per year (HR 1.9, 95% CI 1–4.04)
all-cause mortality after breast cancer diagnosis 0.05% per year vs. 0.03% per year (HR 1.57, 95% CI 1.01–2.48)
HRT associated with increased risk of node-positive breast cancer (HR 1.78, 95% CI 1.23–2.58)
Reference – JAMA 2010 Oct 20;304(15):1684, editorial can be found in JAMA 2010 Oct 20;304(15):1719
estrogen alone (compared to placebo) does not have benefits which outweigh risks overall (level 1 [likely reliable] evidence)
10,739 postmenopausal women aged 50–79 years with prior hysterectomy randomized to conjugated equine estrogen 0.625 mg (Premarin) vs. placebo orally daily for mean of almost 7 years; study stopped early by NIH since there was no benefit in primary outcome of reducing cardiovascular disease
comparing estrogen vs. placebo groups
3.33% vs. 3.67% had coronary heart disease death or myocardial infarction (not significant)
2.98% vs. 2.17% had stroke (NNH 123)
1.90% vs. 1.44% had venous thromboembolism (NNH 217 but not quite statistically significant)
1.77% vs. 2.28% had invasive breast cancer (NNT 196 but not quite statistically significant)
1.15% vs. 1.07% had colorectal cancer (not significant)
0.72% vs. 1.18% had hip fracture (NNT 217)
0.73% vs. 1.18% had hip fracture (NNT 222)
7% vs.7.52% had any fracture (NNT 192)
5.48% vs. 5.32% overall mortality (not significant)
13% vs. 13% had any of the above outcomes (except only counting hip fractures instead of vertebral or any fractures)
Reference – JAMA 2004 Apr 14;291(14):1701, editorial can be found in JAMA 2004 Apr 14;291(14):1769, NIH News Release 2004 Mar 2, commentary can be found in JAMA 2004 Aug 11;292(6):683, summary can be found in Am Fam Physician 2005 Jan 15;71(2):371
no significant differences in most outcomes after 10 years (level 2 [mid-level] evidence)
based on 10.7 year follow-up of 7,645 women from above study
postintervention annualized risk for outcome comparing estrogen vs. placebo
coronary heart disease 0.64% vs. 0.67% (not significant)
stroke 0.36% vs. 0.41% (not significant)
deep vein thrombosis 0.17% vs. 0.27% (hazard ratio 0.63, 95% CI 0.41–0.98)
invasive breast cancer 0.26% vs. 0.34% (not significant)
hip fracture 0.36% vs. 0.28% (not significant)
no significant differences in mortality, coronary heart disease mortality, myocardial infarction, pulmonary embolism or colorectal cancer
Reference – JAMA 2011 Apr 6;305(13):1305, editorial can be found in JAMA 2011 Apr 6;305(13):1354 (correction can be found in JAMA 2011 Jun 15;305(22):2418)
estrogen alone did not have clinically meaningful effect on health-related quality of life in WHI trial (Arch Intern Med 2005 Sep 26;165(17):1976), commentary can be found in BMJ 2005 Oct 22;331(7522), commentary can be found in Am Fam Physician 2006 Apr 15;73(8):1453, commentary can be found in Evidence-Based Medicine 2006 May-Jun;11(3):76
WISDOM trial:
HRT increases risk of cardiovascular disease and venous thromboembolism in older postmenopausal women (level 1 [likely reliable] evidence)
based on randomized trial
5,692 postmenopausal women aged 50–69 years (mean age 63 years) who had ≥ 80% compliance during 12 week run-in period were randomized to hormone replacement therapy vs. placebo
women with uterus were randomized to combination HRT (Prempro) vs. placebo daily
women without uterus and unwilling to take placebo were randomized to Prempro vs. estrogen alone (Premarin) 0.625 mg orally daily
women without uterus and willing to take placebo were randomized to Prempro vs. Premarin vs. placebo daily
medroxyprogesterone acetate 5 mg (Premique) orally daily given to women with uterus and within 3 years of last period, women aged 50–53 years, and older women with unacceptable breakthrough bleeding
median follow-up 11.9 months due to early closure of trial
major cardiovascular disease defined as unstable angina requiring hospitalization, myocardial infarction (fatal or non-fatal), or sudden coronary death
comparing combination therapy vs. placebo in analysis of 4,385 women
0.32% vs. 0 had major cardiovascular disease (NNH 312)
rate of major cardiovascular disease 26.9 vs. 0 per 10,000 person-years (p = 0.016, NNH 371 person-years)
1% vs. 0.14% had venous thromboembolism (NNH 116)
rate of venous thromboembolism 85.1 vs. 11.5 per 10,000 person-years (p < 0.001, NNH 136 person-years)
1.8% vs. 2.6% had osteoporotic fractures (NNT 125)
rate of osteoporotic fractures 155.3 vs. 226.2 per 10,000 person-years (p = 0.07)
no significant differences in rates of cerebrovascular disease, cancer, or death
comparing combination HRT vs. estrogen alone in 1,641 women
no significant differences in major cardiovascular disease, venous thromboembolism, cancer, osteoporotic fracture or death
some combination HRT women counted in this analysis and in analysis compared to placebo
of 11 women who had major cardiovascular events, 9 were > 64 years old and had other cardiovascular risk factors
Reference – WISDOM trial (BMJ 2007 Aug 4;335(7613):239 full-text), editorial can be found in BMJ 2007 Aug 4;335(7613):219, commentary can be found in Evid Based Med 2008 Apr;13(2):52
Heart and Estrogen/Progestin Replacement Study (HERS) trial:
HERS trial is randomized trial of HRT vs. placebo for 4 years in 2,763 postmenopausal women with coronary artery disease
no significant effect on cardiovascular events
hormone use increased risk of venous thromboembolic events (NNH 60 for 4 years) and gallbladder disease (NNH 62.5 for 4 years)
HRT only improves quality of life in women with postmenopausal symptoms, otherwise worsens quality of life
among women with flushing at baseline, hormones associated with improved mental health and fewer depressive symptoms compared to placebo
among women without flushing at baseline, hormones associated with declines in physical function and energy/fatigue compared to placebo
age and clinical illness had greater affect on quality of life measures than hormone use
Reference – JAMA 2002 Feb 6;287(5):591, editorial can be found in JAMA 2002 Feb 6;287(5):641, commentary can be found in JAMA 2002 May 1;287(17):2210, Evid Based Nurs 2002 Jul;5(3):83, ACP J Club 2002 Nov-Dec;137(3):105, Evid Based Ment Health 2002 Nov;5(4):112
Additional quality of life data:
ultra-low-dose transdermal estradiol NOT associated with differences in cognition or quality of life
based on 2-year randomized placebo-controlled trial in 417 postmenopausal women (only 16% of whom had hot flashes at baseline)
Reference – Arch Neurol 2006 Jul;63(7):945 in J Watch Online 2006 Aug 17
interactive multimedia decision aid reduced decisional conflict and increased patient involvement in decision-making in randomized trial of 205 women considering HRT (BMJ 2001 Sep 1;323(7311):490 full-text), editorial can be found in BMJ 2001 Sep 1;323(7311):466, commentary can be found in BMJ 2002 Feb 2;324(7332):296
decision analysis finds that HRT worsens quality of life in women without menopausal symptoms and may benefit women with menopausal symptoms (depending on assumed valued of symptoms), with benefit reduced as breast cancer risk increases (BMJ 2004 Feb 14;328(7436):371 full-text), editorial can be found in BMJ 2004 Feb 14;328(7436):357