Editorials

Hormone replacement therapy

What shall we tell our patients now?

From BMJ USA 2002;September:484

The controversy about hormone replacement therapy (HRT) was not started but only renewed when the results of the Women's Health Initiative (WHI) were announced in July.¹ In the WHI, 16 608 postmenopausal women ages 50 to 79 were randomized to receive estrogen plus progestin or placebo. Although the trial was planned to last 8.5 years, the estrogen-progestin arm was stopped prematurely after 5.2 years because of increased risks.

The WHI reported a small but significantly higher risk of myocardial infarction (hazard ratio=1.32), stroke (HR=1.41), deep vein thrombosis (HR=2.07), pulmonary embolism (HR=2.13), and invasive breast cancer (HR=1.26) in women receiving continuous conjugated estrogen (0.625 mg) combined with medroxyprogesterone acetate. These relative risks seem small when translated into absolute risks: for every 10 000 women-years of HRT there was an excess of 7 cardiac events, 8 breast cancers, 8 strokes, and 8 pulmonary emboli. On the positive side, there were 6 fewer colorectal cancers and 5 fewer hip fractures.

These benefits and risks, although confirmed and measured with greater precision in this large trial, have been studied and debated for years. Observational studies have reported a link between HRT and breast cancer. ^{2 3 7} The studies suggested that this risk increased with duration of use² and that the estrogen-progestin combination increased risk beyond that associated with estrogen alone.³ The Heart and Estrogen/Progestin Replacement Study (HERS) study,⁴ a randomized, blinded, placebo-controlled trial of women with known coronary heart disease, reported an increase in cardiac events in the first one to two years of estrogen-progestin therapy but a decreased incidence in years four and five. This decreased risk was found not to persist in HERS II,⁵ in which HRT was taken for an average of 6.8 years. HERS also confirmed that thromboembolic risk nearly tripled (RR=2.7) with combined estrogen-progestin.⁶ The Nurses Health Study demonstrated an inverse relationship between HRT and the risk of hip fractures,⁷ and data for 59 002 postmenopausal women in this study revealed a lower risk of colon cancer.⁸ In August the Agency for Healthcare Research and Quality released five systematic reviews (www.ahrq.gov/clinic/3rduspstf/hrt) and two journal articles ^{9 10} that summarize the evidence regarding HRT to prevent cardiovascular disease and other chronic health problems. These reviews support the WHI findings and will serve as background for new HRT recommendations to be released this fall by the US Preventive Services Task Force.

The intense media coverage surrounding the WHI has prompted many women to call their physicians to determine what to do. What do these findings mean for postmenopausal women, some of whom have been taking HRT for years? Although the WHI provides a few answers, it leaves more questions unresolved:

1. Are progestins responsible for the increased risks? Should estrogen be used alone? Should unopposed estrogen be used only for short periods (because of risks for endometrial hyperplasia and cancer)?

2. Is any dose or type of progesterone "safe"? For example, is micronized progesterone better than medroxyprogesterone acetate?

3. Is HRT safer in other preparations (eg, non-conjugated or plant-derived estrogens) or doses?

4. Is short-term use of HRT for relief of menopausal symptoms safe? For how many months or years can it be used without increasing the risk of breast cancer? Even one year may be unsafe if the evidence from HERS, in which women with CHD experienced an increased risk of myocardial infarction in the first year of HRT, is generalizable to women without CHD.

5. Can women safely take estrogen following hysterectomy? The WHI may provide an answer in a few years because this study arm is still active, but a hysterectomy does not circumvent the risks of thrombosis and of breast and ovarian cancer that accompany estrogen use.

6. What is the best way to discontinue HRT? Options include a dose taper, a day taper, or a combination.

What should the clinician do in the face of these uncertainties? First, each woman should be counseled individually. Women's needs, desires, expectations, and risk factors are different. Although it would be much easier to counsel uniformly all women to take or not take HRT, patient heterogeneity makes such a stance inappropriate. For example, the risk of osteoporosis is much lower for a 60 year old African-American runner who weighs 150 pounds and does not smoke than is the risk for a sedentary 60 year old Asian-American woman who weighs 100 pounds.

Second, women should not take HRT if their aim is to reduce cardiovascular risk. HRT should not be used for primary or secondary prevention of cardiovascular disease until we better understand the mechanism by which it promotes thrombosis and myocardial infarction.

Third, women should not use HRT indefinitely to alleviate menopausal symptoms. Vaginal dryness, hot flashes, or mental and mood changes are not indications for long-term HRT.

Fourth, HRT does decrease the risk of osteoporosis, and taken for 7-10 years will decrease the risk of hip fractures. HRT is thus a treatment option for patients with or at high-risk for osteoporosis, but bisphosphonates and selective estrogen receptor modulators (eg, raloxifene) may be better alternatives.

Rebekah Wang-Cheng, clinical professor of medicine, associate editor of BMJ USA. 

Medical College of Wisconsin (rwang{at}mail.mcw.edu)

Jo Ann Rosenfeld, assistant professor of medicine, associate editor of BMJ USA. 

Johns Hopkins University School of Medicine (jrosenfe{at}jhmi.edu)