Nelson HD. Assessing Benefits and Harms of Hormone Replacement TherapyClinical Applications. JAMA. 2002;288(7):882–884. doi:10.1001/jama.288.7.882
Author Affiliations: Division of Medical Informatics and Outcomes Research and Evidence-based Practice Center, and Department of Medicine, Oregon Health and Science University, and Medical Service, Veterans Affairs Medical Center, Portland, Ore.
Scientific Review and Clinical Applications Section
Editor: Wendy Levinson, MD, Contributing Editor.
An estimated one third of postmenopausal women in the United States
use hormone replacement therapy (HRT) to treat symptoms of menopause and prevent
chronic conditions. In the context of this widespread use, evidence has been
growing about the potential harms of HRT, particularly regarding long-term
use. Physicians and patients are often confused about how to use results of
studies in individual cases. This article applies the current state of evidence
for the benefits and harms of HRT to management decisions in 4 clinical situations.
Patient preferences, as well as evidence, are important for these decisions.
Benefits and harms need to be readdressed periodically to apply newly published
evidence and to reassess emerging risks, comorbidities, and needs of individuals.
A 51-year-old woman seeks your advice on initiating hormone replacement
therapy (HRT). She has been having irregular menstrual periods throughout
the past 6 months and has had increasingly frequent episodes of hot flashes
that sometimes disturb her sleep. She has been using a soy product that initially
controlled her symptoms; however, its effect has diminished. She has no other
major health problems, is a nonsmoker, and has no strong family history of
cancer or cardiovascular disease, although some relatives have osteoporosis.
She needs advice for 2 issues: appropriate use of HRT for the short-term
relief of menopausal symptoms and the role of HRT in preventing chronic conditions.
It is important to separate these issues because short-term (<5 years)
and long-term (≥5 years) benefits and harms differ.1
Hormone replacement therapy effectively relieves menopausal symptoms
and can be used only as long as symptom control is necessary. Estrogen is
combined with a progestin for a woman with a uterus to prevent endometrial
hyperplasia and endometrial cancer, although unexpected vaginal bleeding requires
evaluation. The daily combined regimen is common in the United States, although
other regimens are also used. Several formulations, doses, and regimens are
effective and should be individualized.
Potential short-term benefits include improvement of hot flashes, sleep
disturbances, urogenital atrophy, and possibly mood2
and aspects of cognition,3 although these effects
may be a consequence of improved sleep. Hormone replacement therapy can improve
or maintain bone density,4 although short-term
use does not prevent fractures in the future.5
Short-term harms include a 1.8-fold increased risk for cholecystitis,6 3.5-fold increased risk of a thromboembolic event
in the first year,7 and possibly increased
risk of stroke and myocardial infarction (MI). Data from the Women's Health
Initiative (WHI),8 a randomized controlled
trial of HRT and primary prevention in postmenopausal women, indicate increased
rates of stroke, coronary heart disease (CHD), and thrombosis among women
randomized to HRT compared with those taking placebo. It is unknown how much
HRT increases the risk for cardiovascular outcomes for women who smoke, are
obese, or have hypertension or for other cardiovascular risk factors. Women
in the Heart and Estrogen/progestin Replacement Study (HERS), a secondary
prevention trial of HRT and cardiac events in women with preexisting CHD,
experienced worsening urinary incontinence throughout a 4-year period of HRT
use.9 Observational studies suggest a potential
increase in breast cancer risk with short-term use, although these are not
definitive.1 The WHI reported significantly
increased breast cancer risk after 5 years of use.8
Women with undesirable menopausal symptoms might consider the risk-benefit
ratio acceptable for short-term use.
The second issue for this patient concerns long-term benefits and harms.
Some women find that even though hot flashes are no longer present, HRT provides
other positive effects for them, such as improved mood, relief of symptoms
of urogenital atrophy, and sense of well-being, although these outcomes vary
among individuals. Long-term HRT users continue to have higher bone density
than past or never users,5 and observational
studies of long-term use indicate that optimal benefit is obtained when use
is begun early in menopause and continued indefinitely.10
The WHI is the first randomized controlled trial to demonstrate protection
for hip, vertebral, and other fractures with HRT use. The WHI also indicated
reduced risks for colon cancer with HRT use.8
Prevention of dementia with long-term use is suggested by some studies; however,
most of these have important methodologic limitations.3
These benefits must be weighed against increasing evidence of important
harms. Contrary to previous beliefs that HRT could prevent cardiovascular
disease, results of the WHI indicate increased risk for MI and stroke after
5 years of use.8 Observational studies of HRT
and breast cancer suggest a 1.2- to 1.4-fold increased risk in breast cancer
incidence in long-term users but no increased risk of dying of breast cancer.1 The WHI confirmed these findings, reporting a relative
risk for breast cancer of 1.26 (95% confidence interval, 1.00-1.59) after
5 years of use.8 Risk of thromboembolism is
higher in the first year of use and diminishes thereafter to approximately
a 2-fold increased risk.7,8,11
Long-term harms also include a 2.5-fold increased risk for cholecystitis.6 Use of unopposed estrogen has been associated with
ovarian cancer in recent studies.12,13
This patient's risk-benefit ratio may no longer justify HRT use.
A 72-year-old woman has been taking estrogen daily for more than 20
years and is seeking advice about whether to continue.
It was initially begun to control hot flashes after her hysterectomy
and oophorectomy, but she has continued to use it because she heard it was
good for her heart and bones. She has no known heart disease but has elevated
cholesterol levels and diabetes mellitus controlled by oral agents. She has
no other major medical problems but complains of chronic back pain and smokes
1 pack of cigarettes daily.
Until the recent publication of the WHI indicating increased risk for
CHD,8 there were no primary prevention trials
of HRT and cardiac disease in the general population, and results of observational
studies were limited by important biases.14
There is now no justification for using HRT for prevention of cardiac disease,
and the American Heart Association recommends basing HRT decisions on noncoronary
benefits and harms ().15 Optimizing this
patient's diabetes management, treating her hypercholesterolemia, and helping
her to stop smoking would better address her cardiac risk factors.
A woman this age with back pain could have a vertebral fracture. Although
HRT may have helped maintain her bone density throughout the years, she may
now have osteoporosis. Results of a spine radiograph and bone densitometry
would be diagnostic. If a nontraumatic vertebral fracture is identified, or
if her T score is low (eg, ≤2.5), she has osteoporosis, and another treatment,
such as a bisphosphonate, may be indicated. Hormone replacement therapy is
approved for prevention but not treatment of osteoporosis. If she has normal
or slightly reduced bone density and no fractures, continuing HRT would likely
continue to provide osteoporosis prevention but also increase her risks for
cardiovascular disease and breast cancer. Although the findings of the WHI
were for women receiving estrogen combined with progestin therapy and this
woman is receiving estrogen alone, a careful approach would assume comparable
risks and discontinue estrogen until new data prove otherwise. The evidence
on other benefits, such as prevention of dementia3
or colon cancer,16 is suggestive but not strong
enough to continue use for these indications alone. Additional considerations
about discontinuing HRT in a long-term user are discussed in a recent article.17
A 54-year-old woman who has been receiving HRT since natural menopause
2 years ago has a 62-year-old sister who was just diagnosed with breast cancer.
Her sister has been checking into their family history and learned that
an aunt and cousin died of breast cancer. The patient's annual mammogram results
have all been normal.
She has a strong family history of breast cancer—an important
breast cancer risk factor. The contribution of HRT use as an additional risk
factor in women with strong family histories is unclear. The Iowa Women's
Health Study found no increase in risk for women using HRT whether they had
a family history of breast cancer or not,18
although several other studies indicate that risks for average-risk long-term
users could be elevated.1,8 This
patient's risk for invasive breast cancer is estimated to be 2.2% throughout
the next 5 years compared with 1.4% for a woman who is the same age and has
average risk factors estimated with the Gail model
(BOX 2).19 If she decides to discontinue
estrogen, she should taper off use throughout several weeks to avoid provoking
A 66-year-old woman who has been receiving HRT since menopause just
had her first MI. Her evaluation indicated 2-vessel disease and she was treated
medically. All of her medications are reviewed after her discharge from the
hospital, and she is still taking HRT.
Results from HERS indicated that women with known CHD who were randomized
to conjugated equine estrogen and medroxyprogesterone acetate had a 52% increased
risk of MI during the first year of use and increased CHD deaths during the
first 3 years of use.20 After 6.8 years of
follow-up in HERS, there were no differences in rates of primary or secondary
CHD.21 How these results apply to the patient
who has been receiving HRT for several years before her event is not entirely
clear. However, HRT may play a role in promoting recurrent thrombotic events.
Recommendations of the American Heart Association support discontinuing HRT
after an acute event ().15
Women Without Cardiovascular DiseaseBase the decision to use hormone replacement therapy (HRT) on noncoronary
benefits and risks.
Women With Cardiovascular DiseaseHormone replacement therapy should not be initiated for secondary prevention of cardiovascular disease.
Women With Cardiovascular Disease and Taking HRTBase the decision to stop or continue HRT on noncoronary benefits and
Stop HRT after acute events; reinstitution should be based on noncoronary
benefits and risks.
*Adapted from Mosca et al15 according
to the Scientific Statement of the American Heart Association.
What is the age of your patient?
What was the patient's age at her first menstruation?
What was the patient's age when she first gave live birth to a child?
How many of the patient's first-degree relatives—mother or sisters—have had breast cancer?
Has the patient ever had a breast biopsy?
How many positive or negative breast biopsies has the patient had?
Has the patient had at least one biopsy with atypical hyperplasia?
Please indicate the race and ethnicity of the patient, if known.
*The assessment tool was reproduced from the National Cancer Institute
and the National Surgical Adjuvant Breast and Bowel Project Biostatistics