The Women's Health Initiative (WHI) is a NIH sponsored research project of 40 predominately university research centers formed to study several preselected disease risks of postmenopausal women. From 1993 to 1998 WHI enrolled 161, 809 well women between ages 50 and 79 years in this observational study. A subset of 8506 women volunteered to receive combination continuous hormonal replacement (HRT) of 2.5 mg of the synthetic progestin, medroxyprogesterone acetate (MPA) (Provera), along with 0.625 mg of conjugated equine estrogen (Premarin). A matched control group of 8102 women received a placebo.

At the time the study was designed continuous HRT had moved to the front of alternatives for HRT. With protracted use most women become amenorrheic avoiding the most frequent side effect and largest compliance problems of unwanted uterine bleeding. Continuous therapy with estrogen alone (no progestin opposition) unequivocally had been shown to increase the risk of endometrial cancer. The dose of estrogen prescribed was the one most commonly used. There was concern that lower doses may not provide sufficient protection against osteoporosis. There was reasonable evidence that estrogen may be cardioprotective and an oral route offered the most beneficial effects on lipid metabolism.

Several of the most pressing and previously unresolved issues about HRT was its effect on the cardiovascular system, breast and other female cancer risks, and osteoporosis. Despite a very large number of studies on each of these topics, insufficient data was generated either because of experimental design, length of study, or number of women studied, to draw conclusions about the risk/benefit of HRT for chronic disease. Metaanalysis had shown a possible slight increased risk of breast cancer and equivalent findings on cardiovascular disease. Although shown to delay bone resorption, HRT had not been proven to reduce fracture risk. The age after which HRT should not be started and at what age HRT should be stopped was not known. Overall, the benefits of HRT for most women were considered greater than the risks. It is in this environment that the WHI study was designed in 1991-1992.

Although the WHI study was designed for a longer observation time, the study arm receiving continuous HRT regimen was discontinued when preset safety limits were exceeded. This may or may not be at the level of a statistically proven adverse effect. A second trial of HRT using estrogen alone in women with previous hysterectomy is to continue until 2005. The results showed a statistically significant increase of breast cancer, cardiovascular disease including stroke and thromboembolism, with a decreased risk of osteoporosis and colon cancer. “All cause mortality”w as not affected during the trial. The authors of the writing group conclude, "the risk benefit profile is not consistent with the requirement for a viable intervention for primary prevention of chronic diseases…." The results so widely circulated in the press have caused extreme controversy and unrest with patient and health care providers alike. There are a number of issues about the trial and its conclusion that warrant scrutiny.

1. Although the numbers of enrolled is impressive and the design appears sound, the WHI study can not supercede and countermand the voluminous works from numerous clinical researchers on the subject. The study makes a significant contribution to the literature on the subject and its results should be taken seriously. It is presumptuous for the WHI authors to state that theirs was the "first definitive study" or use a term like "global" when it was a directed study on a relatively small number of health and lifestyle concerns. Decisions to treat or not to treat patients with HRT cannot be made on this study alone.
   
2. There are a number of statistical concerns with the study. Although the WHI study is of prospective randomized design and the largest of studies yet performed, the results may not be statistically substantiated and the questions posed are not satisfactorily answered. Did it have sufficient enrollment? Are the individuals seen truly representative of the "normal" population, or different from the possibly select population seeking HRT counseling from physicians? Was the study truly blinded? Was the dropout (attrition) rate too high? Can the statistical analysis stand up to the scrutiny of reanalysis by an independent observer?
   
3. If the statistics are correct, it is very likely that most non-statisticians will not be able to critically review the findings of the WHI. This is true of both health care providers and consumers. For example, there is a significant difference in perception between reporting that use of HRT results in a 41% increase in the risk of stroke and reporting the risk of stroke is increased by 8 per 10,000 using HRT for one year. Breast cancer was increased by 26% or 8 more women per 10,000 will be diagnosed. Previous studies have indicated that more cancer is found in women on HRT but these women had improved 5-year survival rates. The short duration of the study makes it unlikely that the cancers were formed after therapy was started. It is likely that estrogen may be cancer promoting, but this is different than cancer forming.
   
4. Two thirds of the participants between ages 60 and 79 years and 21% were over age 71. Many of the users of HRT are younger and seeking relief from menopausal symptoms, not disease prevention. Women with severe menopausal symptoms were excluded from the study. Risks may be not only related to the individual health and genetic history, but also to age. It should be stressed that even if there is a statistically significant increased risk of serious medical diseases or even death, a small incremental risk may be worth taking for many women. The WHI illustrates that risks may be age dependent and the study even implies that therapy should be limited to younger women. However, a much better age stratification of study results is needed for the proper evaluation and application of the results.
   
5. HRT is not synonymous with estrogen replacement. Thus far, there is no way to distinguish the effects of estrogen from those of progestins in the WHI results. Since the estrogen alone arm of the WHI trial is continuing, it can be argued against a major effect of estrogen, If it were estrogen alone, breast cancer risk should also be increased. Perhaps too much emphasis has been placed on the actions of estrogen and not enough on that of progestins. Progestins have portent effects on vasculature and breast either acting alone or as an additive effect. Progestins may be as likely as estrogen to cause the adverse effects.
   
6. The WHI study used what is increasingly thought to be a relatively high dose of both estrogen and progestin. The dose of estrogen prescribed was the one most commonly used. There was concern that lower doses may not provide sufficient protection against osteoporosis or be insufficient to control uterine bleeding. Therapeutic benefit is possible with as little as 25% of the amount of estrogen used in the WHI study. Presently, many physicians are prescribing and many women using lower doses than those used in the WHI trial. There may be a dose response curve for both the negative and positive finding in the WHI study.
   
7. With protracted use of continuous HRT, most women become amenorrheic avoiding the most frequent side effect and largest compliance problem of unwanted uterine bleeding. Only a single continuous regimen was used in the WHI study. Cyclic or sequential regimens were not evaluated.
   
8. Synthetic progestins are not the same as progesterone. Micronized progesterone is available and may have different effects on vasculature and breast.
   
9. While bioassays have shown equivalence, equine estrogens (Premarin) used in the WHI study, while "natural", is not the same as estradiol or estrone alone.
   
10. An oral route offers the most beneficial effects on lipid metabolism and this was thought to be the basis of estrogen’s now debated cardioprotective effects. Alternative delivery systems e.g. transdermal, vaginal, injectable depot or buccal, may have different effects on vasculature.
    
11. As a consequence of the WHI report, there have been numerous negative implications about the alliance or allegiance of the medical profession and pharmaceutical firms. By a system of collusion designed for big business, this carelessly places patients at risk. It has been inferred that clinicians and scientists cannot be separated from the interests of drug companies. The immediate result has been decreased confidence in physicians trying earnestly to give the best care to their patients. Even if another study completely refuting the WHI study were reported, there would still be the residual belief for a significant number of women that HRT is harmful. Patients are being driven toward alternative therapies with not only unproven benefits, but also unknown risks.
    
12. The study has governmental sponsorship and the force of a governmental decree. The manner in which the study findings were reported and released to physicians and the media is problematic. While the statistically significant benefits are reported, the WHI study authors seem to give more credibility to the risks than benefits of HRT without explaining how their value judgement were derived. The difficulties of how to transmit the WHI data fairly and expeditiously are considerable, but there would seem to be a better way. Many patients were calling their physicians asking questions about the importance of the study without the general medical community aware of the report.
    

Women can now expect to live at least one third of their lives after menopause. The menopausal transition with its associated estrogen deficiency can be a troublesome time. No therapy other than estrogen has been as effective for the treatment of the symptom of menopause. It appears that the long-term consequences of estrogen use may not be as beneficial as previously thought and that here may be inherent risks of its use in some women. HRT may not be a panacea, but neither is it the poison that is suggested in the current backlash created by the WHI study. As important as the WHI data might be, the effect of the study has been much greater than the information it provides. There are major questions left unanswered and it is possible that the results of the WHI study do not meet necessary statistical scrutiny. Regardless, the WHI should not be considered as absolute nor should it supercede all accumulated clinical and research experience. It is difficult to predict whether the overall effect of the WHI study will be judged in the future as positive or negative. A balanced approach is certainly mandated. No one doubts that each patient should be treated individually. Hormonal manipulation strategies are still evolving. The objectives of both health care provider and consumer are the same - a longer and better life.

Sam Thatcher
Johnson City TN
September 1, 2002

BACK HOME