Riverview Physicians for Women

Call us directly: 804.524.0890

439 Jennick Drive, Colonial Heights, Virginia 23834 View Location


Evaluation of

The Women's Health Initiative


       The question is being asked frequently since the release of the results of the much publicized Women’s Health Initiative: “Where do we stand now?”  The answer quite simply is the same place we stood before the results were released.  Before we look more closely at this simplistic answer we need to look at what we knew before the study and how we got here. What women don’t know, and have not been told, about this study could cause significant harm to roughly 6 million women if they don’t empower themselves with the truth.


What We Know:


q      Together, Oral Hormone Contraception (OC’s) and Hormone Replacement Therapy (HRT) constitute perhaps the greatest advance in Women’s Health Care.


q      Premarin is perhaps the most studied drug in use today.


q      Hormone Replacement Therapy (estrogen alone or combined) has a long-standing history of approximately 40 – 50 years of proven safety and efficacy; is well tolerated with few side-effects; and is relatively inexpensive.


q      Some twenty years ago it was acknowledged that Premarin 0.625 mg was the lowest effective dose for the prevention of Osteoporosis.  In an effort to minimize the risk of estrogen therapy it was recommended to start at the lowest effective dose and increase it accordingly until response was achieved.  (As a result of the efforts to minimize exposure to estrogen it is now believed 0.3 mg of Premarin is protective.)


q      FemHRT (1mg Norethindrone Acetate (NA) / 5 mcg. Ethinyl Estradiol (EE)) has an estrogen dose equivalent to Premarin 0.625 mg, Prempro, Activella, and 17B-Estradiol (1mg).


q      LoEstrin 1/20 OCP is related to FemHRT (1mg NA/20mcg EE)


q      OC’s, with 35 and 50 mcg Ethinyl Estradiol, therefore by comparison to LoEstrin and FemHRT, have 7 – 10 times more estrogen than traditional doses of HRT.


q      Oral Contraceptives have been demonstrated to reduce the risk of Benign Breast Disease and do not increase the risk of Breast Cancer.


q      No well-designed prospective study has conclusively demonstrated an increased risk of Breast Cancer associated with HRT. The vast volume of estrogen use and the failure to conclusively demonstrate increased Breast Cancer risk has lead some to argue that if a risk exists it is of marginal magnitude.


q      It is implausible then to propose that HRT, at one-seventh to one-tenth the dose of OC’s, would increase the risk of Breast Cancer when OC’s do not.  Though implausible, that doesn’t mean the relationship isn’t in fact correct, but if true then an inverse relationship exists in which case the highest risk exists with no estrogen at all!


q      Menopause is the manifestation of the hypo-estrogenic consequences of ovarian failure; whether spontaneous or surgical.


q      Contrary to attempts by some to glorify menopause as a Natural Stage of a Woman Life, which should be celebrated, menopause is in fact a Hormone Deficiency Condition and should be recognized and managed as such, just as one would Diabetes or Hypothyroidism. Until relatively recent times, the typical life expectancy was 35-40 years of age.  Most women died before reaching menopause.  Therefore, menopause is a condition of relatively recent advent; made manifest by increased longevity as a consequence of our affluent society with its improvements in living conditions and health care.


q      Though symptoms may diminish or resolve, one does not “get over” menopause.


q      As soon as hormone therapy is stopped, bone loss resumes and quite often symptoms recur.


q      Recently concerns have arisen regarding the affect of Medroxyprogesterone Acetate (MPA) on the beneficial effects of HRT.


q      Evista (a selective estrogen) is being promoted as a safer replacement for ERT even though it is known to increase the risk of blood clots and it is too soon to know its cardiovascular consequences.


q      The Heart and Estrogen Replacement Study (HERS) demonstrated that compared to no hormone treatment, women who suffered a heart attack (myocardial infarction, MI) and were then placed on ERT / HRT experienced an increased risk of repeat MI during the first year, experienced the same rate of MI as non-users during the second year and then showed a reduction in MI rates after the third year.  This would indicate the safe continuation of ERT for women who have already received ERT for more than 3 years, so long as other risk factors have not arisen.


q      Long before the HERS study was published it was felt to be prudent to withhold ERT for 6 months from patients with thromboembolic events.


q      ESTROGEN IN THE PREVENTION OF ATHEROSCLEROSIS TRIAL (EPAT):  The primary finding from EPAT is that women who were randomized to unopposed estrogen had a significant reduction in subclinical atherosclerosis compared to women who received placebo.


q      Because the majority of women entering menopause are asymptomatic but have early vascular disease, EPAT suggests that early unop­posed estrogen may have a profound impact on reduc­ing the progression of atherosclerosis and its clinical sequelae; heart attacks and strokes.


q      The Premarin only arm continues to run because the same adverse trends were not noted.

How Did We Get Here?

        Initially HRT was used to treat only symptomatic women.  As experience accumulated it was recognized that treated patients demonstrated a decreased incidence of osteoporosis, fractures and associated complications.  Emphasis was placed on determining risk factors for osteoporosis and identifying patients at risk.  It was found that thin boned, fair skinned women of northern European and Asian extraction were at highest risk.  The indications for HRT management were expanded to include the prevention of osteoporotic disease.  Women of African extraction, with thick sturdy bone mass, were not felt to be at risk and therefore would not benefit from HRT. As further experience accumulated it was recognized that women treated to prevent osteoporosis exhibited fewer cardiovascular events; hypertension, heart attacks and strokes were reduced.  This was recognized as a significant potential benefit for those of African heritage and once more treatment indications were expanded.  Subsequently improved mental function was documented and the possible benefit for Alzheimer’s disease was proposed.


          Against the framework of these successes was the criticism that large, prospective, randomized, double- (or even triple-) blinded studies were lacking.  The possibility that improved outcomes were the result of selection bias (only healthy women were being placed on HRT) or surveillance bias (increased contact with health care providers) was raised.


          To address to validity of the observed benefits and to better assess the risks the need for a large, prospective, randomized, blinded study was recognized.  Enter the Women’s Health Initiative.  This highly touted study was promised to be the study to answer the questions definitively, once and for all.  Expectations were high and the results anxiously awaited.  Then the disappointment came as the study was ended prematurely with the report of concerning trends towards increased risk of Cardiovascular Disease, Heart Attacks, Strokes and Blood Clots.  Hopes were dashed.  New management strategies were proposed. 

But wait!  What did the study really show?  This very important study, which was to set matters straight had major flaws.  The expert commentators would certainly expose and discuss these flaws! … or would it? 


Study Design:

            The stated purpose of the research group was to investigate the primary prevention of cardiovascular disease in postmenopausal women using Estrogen (ERT) or Combined Estrogen-Progestin Hormone Replacement Therapy (HRT).  The study began in 1993.  Women 50 to 80 years of age were recruited into the study.  Menopausal status was assumed if the woman had undergone surgical removal of both ovaries, had not had periods for 6 months or had a prior history of HRT use.  After a 3 month wash out period women with prior HRT use were randomized into treatment or control groups.

The initial protocol allowed for randomization of women with an intact uterus into either estrogen only or combined estrogen-progestin treatment arms.  This act was a deviation from the standard of care as it was already recognized that protection with combined progestin therapy was necessary to protect women, who had not undergone hysterectomy, from the increased risk of endometrial cancer associated with unopposed estrogen.  This was not corrected until 1995 when these women were converted to combined HRT regimen.

Assessment of baseline cardiovascular status was made by patient questionnaire…Do you have cardiovascular disease?  Participants in the treatment group then received Premarin 0.625 or Prempro. 


The Catch!

            Remember that criticism of past results for protective benefits related to the lack of a large, prospective, randomized, double blinded, placebo controlled study.  The biggest obstacle to providing this “gold standard” study design is the very nature of the problem being studied….It’s symptomatic!  If one randomizes symptomatic participants into a placebo control group, one would logically expect a significant portion of participants to remain symptomatic.  Conversely, the treatment group will experience relief of symptoms.  This has two immediate consequences.  First is un-blinding the study, the patient is aware of receiving medication or placebo.  More importantly is migration of the control group into the treatment group as participants seek relief of their symptoms. 


Discussion of Study Design:

            In order to prevent or reduce the incidence of cardiovascular disease (or any other process for that matter) you must intervene before the process is established.  It is obvious, and should not need to be pointed out explicitly, that one cannot prevent that which is already established.  The point being, if one intends to prevent cardiovascular disease one must act at an appropriately early point in the process, as early as possible. 

            As most women typically become menopausal between 45 and 55 years of age that would appear to be the most prudent point to intervene for prevention purposes.  Given that this spans a 10 year period for the youngest menopausal patients one would likely want to limit inclusion to not more that 3 years from onset of first menopausal symptoms (or surgery) to minimize the chance of including participants with preexisting disease, though this would introduce the possibility of recall bias on the part of the participant.  The only other option to reduce the possibility of participants with preexisting significant disease would be to perform Cardiac Catheterization studies on all participants.  Obviously this would introduce an unacceptable degree of risk and expense.

            The research group instead elected to recruit women aged 50 – 79 years of age and excluded anyone with hypo-estrogenic symptoms.  As noted above, participants were screened for cardiovascular status by questionnaire.  It is a well-documented phenomenon that persons undergoing cardiac catheterization who demonstrate significant degrees of coronary artery stenosis may be asymptomatic.

            The presence of symptoms would reflect relatively fresh menopausal status.  Not only did the group recruit participants from a pool that carried a very high probability of preexisting cardiovascular disease, they excluded the very participants they should have studied and who had the best chance for successful intervention.

Furthermore, the investigators contaminated their control group by randomization for inclusion of past HRT users into the non-treatment group; after a 3-month “washout” period. This would imply that the principal researchers believe the effects of HRT are transient and gone by three months.  This would also ignore the potential benefit already obtained from past use.  One quarter of the study participants had a history of prior hormone therapy.

The final study group had an average age of 63 years at the time of enrollment in the study.  This mean one half of the women studied were 63 to 80 years old.  Remember that the study was trying to prevent cardiovascular disease with the onset of menopause.  Less than 10% of participants were in their 50’s.  Thus by the very design of the study it was impossible to show a protective benefit from estrogen; the study was doomed to fail before the first patient was enrolled.

            One would not expect to observe a preventive effect on cardiovascular disease in a group of participants who have a high prevalence of preexisting disease, especially when compared to a group containing members who have achieved protective benefit from past use.  Yet this esteemed group of researchers claims the failure to detect protective benefit is proof of the lack of benefit.  An increased occurrence of adverse events is not unexpected in a group of participants with established vascular disease exposed to estrogen as was demonstrated by the results of the Heart and Estrogen Replacement Study (HERS).  Granted the HERS results were not available at the time of the study design but they were widely known by the time of release and interpretation of these WHI findings, which would serve to guide the interpretation by these researchers.

            Participants in the treatment group received only one dosage level of only one type of estrogen and only one type of progestin.  This presupposes that all estrogens, all progestins, and all delivery methods (patch, pill, injection or cream) are equally effective.  It also presupposes that the minimum effective dose for prevention of osteoporosis is adequate for the prevention of cardiovascular disease. 

Where is the evaluation of 17B-estradiol?  Where are the results for FemHRT and Activella.  How well do Vivelle, Climara, Esclim, Combipatch or the other Transdermal patches perform?  How do birth control pills rate? What benefit does Premarin or Estradiol vaginal cream provide?  How about combination therapy with other progestins such as Norethindrone Acetate or Micronized Progesterone?  Should we monitor response to treatment by checking Hormone levels?


Discussion of Results:

            The study was ended because of concerns that an unfavorable trend was developing.  To stop the study for safety concerns and report that adverse events were noted is appropriate and commendable.  To state the trend needed further investigation would be appropriate, however, this group made the de facto statement that the claimed benefits have been disproved and a new era of menopausal management is dawning.

The investigators state that from their data they are unable to tell if the adverse effects are related to the estrogen or the progestin, yet the Premarin only arm continues to run because the same adverse trends were not noted.  All things being equal otherwise, with the only difference being presence or absence of Provera (Medroxy-Progesterone Acetate, MPA), it would seem intuitively obvious that, if indeed there is a problem, it would lie with the MPA.  Most certainly this possibility deserves a statement that it warrants further investigation.  Of most reassuring importance is the fact that these “unfavorable tends” are not apparent in the Premarin (estrogen) only arm of the study (and that it is being allowed to continue), a fact that is being down played in the discussions in the press.  Thus, a simplistic answer to women who have undergone hysterectomy is that the results of this study does not apply to them.

           Indeed, in a later interview, one of the principal investigators stated, “what women need is a doctor who is a patient advocate not a hormone advocate”.  It will be those doctors who, because they truly are patient advocates, will continue to be hormone advocates; who will continue to educate their patients about the benefits of ERT/HRT, while acknowledging the risks and certainly discussing the severe limitations of The WHI Study.

While unfavorable trends were reported it is noted they did not reach statistical significance; though some commentators are citing the results as statistically significant.  Even if they had reached significance, the statistically significant results of an invalid study are nonetheless invalid.  The only results that reached statistical significance were reduction of osteoporosis / fractures and reduction of colorectal cancer.  Again this study confirms what was already known, the reduction of osteoporosis / fracture and colorectal cancer.

            To put the reported risk of increased Breast Cancer into perspective it must be remembered that a woman’s baseline risk of Breast Cancer, no other risk factors considered, is 12%.  The observed excess incidence was 8 cases per 10,000 women.  This equates to 0.08%. 


[Since the initial release of the study it has since been reported that the excess occurrence of breast cancer noted in years 4 and 5 of the study abruptly returned to baseline risk in year 6.  That is to say that after the 5th year there was NO difference in cancer rates between the treatment and non-treatment groups of the study.


The possibility that Estrogen CAUSES breast cancer would be a “dose and duration” phenomenon.  That is, the higher the dose and the longer the duration of exposure, the higher the incidence of breast cancer should be.  If this were the case the occurrence of excess cases of cancer would continue to increase every year, for ever. 


It has long been argued that Estrogen does NOT CAUSE breast cancer but rather facilitates the growth of undiagnosed, pre-existing cancers that were too small to detect at the time the patient enrolled in the study.  Under the influence of the estrogen growth of the tumors was promoted and allowed detection earlier than the corresponding tumors in the non-treatment group giving the appearance of increased incidence, when in fact it was earlier detection.  Supporting this rationale is the finding that women diagnosed with breast cancer while taking estrogen have earlier disease with more favorable tissue histology and better survival rates.


This PROMOTION of preexisting tumors is the only explanation for the abrupt return to baseline risk at year 6 of the WHI study.  This finding would mean that not only does estrogen not increase the risk of getting cancer but also improves the outcome of women who have taken estrogen.]


But to know the Risk / Benefit ratio, we need to know reliably the protective benefits of ERT/HRT; the most effective agent(s) and dose(s). This study clearly does not provide that information. 

What information does this study provide? 

  1. It confirms statistically significant protection against osteoporosis / fractures
  2. It revealed an unexpected protection against colorectal cancer. 
  3. It suggests continued concern about the impact of MPA on the effects of HRT. 
  4. It cannot speak to cardiovascular risk because of the design flaws.

As more women live a significantly longer period of their lives in postmenopausal status it is imperative we obtain the right information to make proper management recommendations.  Women not only deserve a properly performed study, they should demand it.  I demand it for the sake of my family members and my patients’ welfare.  We should all demand it, not because we are hormone advocates but because we are patient advocates.

So what have we learned from this study?  How should we change our recommendations to our patients and family?  Where do we stand now?  Quite simply, we are just where we were in June of 2002 before the WHI results were announced.  As discussed above, and which seems so simply and painfully obvious, the WHI is an Invalid study design and therefore the major conclusion that there is no cardio-protective effect is invalid.  We still desperately need the study that was promised and we all so eagerly await.  This area is too important to allow sloppy work to pass as important, monumental information and put the welfare of millions of women at risk.

          This study does not usher in a new era, it is not a “Landmark” study, it does not warrant a major revision in management or recommendations, it certainly does not cause a paradigm shift!  Health care providers should not be advocating the wholesale abandonment of a proven safe and effective treatment. The WHI findings do reinforce the concept learned from HERS, that patients must be carefully evaluated for existing cardiovascular disease prior to initiating HRT.  Perhaps some women really can be “too old” to benefit from starting HRT, if say by age 80 the probability of CVD is high enough to be a significant risk.  But that does is not to say that a woman would have to stop long term HRT just because she turned 80 years old.  Or maybe we need only find safer progestins.

           Why would we consider the recommendation of a multi-drug approach which is untried, and more importantly, unproven; which is comprised of expensive medications, which have significant side effects

           Were this study sponsored by anyone other than the National Institutes of Health it would have never achieved this level of notoriety and much less likely even to have been published. To allow this study to cast aside the accumulated weight of evidence and experience with HRT is unwarranted and inappropriate. 

It is unexpectedly fortuitous that this study was ended early.  Not because of the adverse trends reported but because it provides an important opportunity to start over with an appropriate study design, well thought out and executed to explore an issue of paramount importance.  It is time to publicly retract the results and ensuing commentary.  The concern for adverse cardiac events raised by this study may have some benefit in the long run if for no other reason than providing the motivation for those subjects randomized into the non-treatment control group to stay there and not seek alternative treatment of their symptoms.  The new study should consider a four-armed approach: Placebo control, Estrogen only, combined Estrogen-Progestin, and what could be termed “Novel” management to consist of the alternative management approaches now being advocated.