Postmenopausal Hormone Therapy: A Response to the Critics

Special Report By Leon Speroff, MD

Editor’s Note: In this OB/GYN Clinical Alert Special Report, Dr. Speroff takes a timely look at current controversy involving postmenopausal hormone therapy. Please also notice the inclusion of Dr. Speroff’s list of references at the conclusion of the report.

It seems like everyone is jumping on the bandwagon: use postmenopausal hormone therapy for less than 5 years to treat menopausal symptoms. This position is based upon 3 arguments:

  1. The data are insufficient to believe that estrogen will protect against fractures;
  2. Recent studies indicate that estrogen will not reduce the risk of coronary heart disease;
  3. Postmenopausal hormone therapy used for longer than 5 years will increase the risk of breast cancer.

Let’s address each of these issues, and let’s bend over backward to be objective, unlike some recent publications by critics of postmenopausal hormone therapy.

Estrogen and Fractures

The argument against estrogen is based upon the contention that conclusions should be based upon randomized clinical trials, and only bisphosphonates have been demonstrated to reduce the risk of fractures.1-3 The facts are as follows: Estrogen and bisphosphonates are equally potent in their beneficial effect as measured by bone density.1,4 Case-control and cohort studies indicate that estrogen will reduce the rate of fractures with the same efficacy as bisphosphonates.5-9 And most importantly, there are 2 randomized trials (always ignored by the critics) documenting a beneficial effect of hormone therapy on fracture incidence. In a Finnish study, a 71% reduction in nonvertebral fractures was documented within 5 years of treatment, a result comparable to that predicted by the case-control and cohort data.10 In a Danish ongoing randomized trial, there was a 39% reduction in the overall fracture rate after 5 years.11 In both of these randomized trials, the women did not already have osteoporosis at the beginning of treatment. If you want to play the critic’s game and be selective in your use of the literature, it would be appropriate to argue that bisphosphonates have been proven to be effective in reducing fractures only in women with documented osteoporosis, whereas hormone therapy has been proven to be effective in the primary prevention of osteoporosis and fractures. Likewise, raloxifene has been demonstrated to reduce fractures only in women with already demonstrated osteoporosis, and the reduction was observed only with spinal fractures, and not with hip or other-site fractures.12

Where does that leave us with bone? Hormone therapy prevents bone loss and fractures at all sites, even though long-term treatment is necessary to achieve this effect.8,13 Bisphosphonates prevent bone loss, probably are effective in the primary prevention of osteoporosis and fractures, and for sure, reduce the fracture risk in women with osteoporosis. Raloxifene prevents bone loss (although it is less potent than estrogen and bisphosphonates14) and reduces spinal fracture risk in women with osteoporosis. Add to these conclusions the fact that hormone therapy is the least expensive of the treatment options.

Estrogen and the Cardiovascular System

The evidence of the secondary prevention trials is impressively uniform that hormone therapy will not reduce the risk of a subsequent cardiovascular event in women who already have significant, clinically apparent atherosclerosis.15-17 It is not apparent whether the "small increase" in cardiovascular events reported in the on-going Women’s Health Initiative was influenced by the women in the study who had atherosclerosis upon entry (it is not a pure primary prevention trial). The American Heart Association recommended that hormone therapy not be initiated in women with atherosclerosis with the expectation that treatment will reduce the risk of a subsequent cardiovascular event.18 I have no problem with that statement. We know statins are the drug of choice for these women, and furthermore, this group of women represents a very small part of our practice. I do have a problem with the second statement from the American Heart Association that said: no recommendation can be made regarding primary prevention until we have the results of randomized clinical trials. You can’t practice medicine based only on randomized clinical trial data! If we based our recommendations only on clinical trial data, we would never tell our patients who smoke to stop smoking. There is an enormous collection of biological and observational data indicating that hormone therapy given to apparently healthy women will reduce the risk of coronary heart disease.19 A primary prevention trial recently reported progression of atherosclerosis in the placebo group and slight improvement in the estrogen-treated group.20 It seems to me that there is an emerging theme: it takes a healthy endothelium to respond to estrogen. Animal studies, including the primate, indicate that as the endothelium progressively is involved with atherosclerosis, it loses its ability to respond to estrogen.21,22 Thus, there are good reasons to believe that estrogen will provide primary prevention of coronary heart disease. The major criticism, the "healthy user effect," will not be answered until we have results from the ongoing large trials. However, the difference in health comparing users and non-users has not been great enough to explain the observed beneficial effect. We also await clinical trial results to answer this important question: are the results in the secondary prevention trials due to a lack of an estrogen effect or is the beneficial effect of new statin treatment obscuring the positive action of estrogen? In the primary prevention trial noted above, the beneficial effect of estrogen could be documented only in women not using statins.20

In a very recent publication by one of the critics, it was suggested that because the risk of venous thromboembolism increases with age, the venous thrombosis risk with hormone therapy can be expected to be higher in elderly women.23 However, the risk of venous thrombosis is almost exclusively concentrated in the first two years of use (data from the HERS trial with the critic as senior author24), and, therefore, long-term users have little, if any, risk of this side effect as they get older.

Hormone Therapy and Breast Cancer

It is almost fashionable to conclude that hormone therapy for longer than 5 years is associated with a slight increase in the risk of breast cancer, referring either to the collaborative re-analysis of the world’s literature published in 1997 or the Nurses’ Health Study.25, 26 Of course, there are no randomized clinical trial data here. When we have only observational data, clinicians want to see uniformity, agreement, and consistency among the studies (like that seen with smoking and lung cancer or oral contraceptives and ovarian cancer). With postmenopausal hormone therapy and breast cancer, we have the opposite situation, a lack of uniformity, agreement, and consistency in over 60 studies. Sometimes clinicians believe that all the recent studies have consistently indicated an increased risk; however, that is not the case. The positive studies (increased risk) have received the publicity; the negative studies (no increased risk) have been ignored (and the negative studies impressively outnumber the positive studies).

In the practice of medicine, we make medical judgments every day. I define a "medical judgment" as making a decision when you don’t have all the facts you need to make that decision. That is the art and skill of medicine, of being a clinician. This is one of those situations. Until we have randomized trial data (no earlier than 2006), the clinician must present the confusing picture to the patient. In my view, the evidence is consistent either with a slightly increased risk of breast cancer with long-term treatment or hormonal stimulation of pre-existing tumors (an effect that is consistent with the uniformly observed reduced risk of dying of breast cancer in hormone users who develop the disease).

I believe it is appropriate to make the following three statements to patients:

  1. The evidence does not indicate a major increase in risk of breast cancer associated with the long-term use of postmenopausal hormone therapy;
  2. A small increase in risk of breast cancer or an increase with long duration of use is unlikely;
  3. The use of hormone therapy does not increase the risk beyond that already associated with a positive family history of breast cancer (a positive family history is not a contraindication).


There continues to be good reason to believe that long-term use of postmenopausal hormone therapy is associated with preventive health benefits for the bones and the cardiovascular system. Inaccurate and non-objective reporting in both the lay press and professional journals is unfairly enhancing the appropriate fear of breast cancer in our patients. It is time that some of us speak up to balance the scales. Postmenopausal hormone therapy is the most cost-effective treatment option available and the only one with a broad spectrum of beneficial effects for older women.

Finally, let me share an approach that I have found to be effective with patients. I like to tell patients that there is an enormous amount of research ongoing with postmenopausal women and hormone therapy. My obligation is to review each year the new information. Based on that review, the patient makes a new short-term decision. Thus, repetitive short-term decisions allow the patient to avoid being uncomfortable by the thought of a long-term commitment. I truly believe this approach improves long-term continuation rates.


1. Harris ST, et al. JAMA. 1999;282:1344-1352.

2. Black DM, et al. J Clin Endocrinol Metab. 2000;85: 4118-4124.

3. McClung MR, et al. N Engl J Med. 2001;344:333-340.

4. McClung MR, et al. Ann Intern Med. 1998;128: 253-261.

5. Weiss NC, et al. N Engl J Med. 1980;303:1195.

6. Ettinger B, et al. Ann Intern Med. 1985;102:319.

7. Cauley JA, et al. Ann Intern Med. 1995;122:9-16.

8. Michaësson K, et al. BMJ. 1998;316:1858-1863.

9. Grodstein F, et al. Epidemiology. 1999;10:476-480.

10. Komulainen MH, et al. Maturitas. 1998;31:45-54.

11. Mosekilde L, et al. Maturitas. 2000;36:181-193.

12. Eastell R, et al. The effects of raloxifene on incident vertebral fractures in postmenopausal women with osteoporosis: 4-year results from the MORE trial. In press.

13. Schneider DL, et al. JAMA. 1997;277:543-547.

14. Bjarnason NH, et al. Osteoporosis Int. 1998; 8(Suppl 3):11.

15. Hulley S, et al. JAMA. 1998;280:605-618.

16. Herrington DM, et al. N Engl J Med. 2000;343: 522-529.

17. Viscoli CM, et al. N Engl J Med. 2001;345:1243-1249.

18. Mosca L, et al. Circulation. 2001;104:499-503.

19. Grodstein F, Stampfer M. Prog Cardiovasc Dis. 1995; 38:199-210.

20. Hodis HN, et al. Ann Intern Med. 2001;135:939-953.

21. Hanke H, et al. Atherosclerosis. 1999;147:123-132.

22. Holm P, et al. Circulation. 1999;100:1727-1733.

23. Grady D. JAMA. 2002;287:2130-2137.

24. Grady D, et al. Ann Intern Med. 2000;132:689-696.

25. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet. 1997;350:1047-1059.

26. Colditz GA, Rosner B. Am J Epidemiol. 2000;152: 950-964.