Time to reflect on the Women’s Health Initiative (WHI) Study

David H. Barlow

Editor-in-Chief

With 25 years experience of the field of reproductive medicine I can think of no time in my career that hormone replacement therapy (HRT) has been free of controversy. At times the HRT debate has been overshadowed by other issues in the reproductive medicine field, usually in association with developments in assisted reproduction, but throughout my career I have noted unusual polarization of opinion concerning HRT. It is a field of therapeutics on which the views can transcend the usual basis on which we judge medicines. Most medicines are judged on the basis of the best objective data and their clinical role determined by that evidence without emotion coming into play. With HRT, however, there appears to be, for many, an almost partisan support or opposition to HRT.

 In the late 1970s working in the menopause clinic at Glasgow Western Infirmary in the adjacent bay to a cardiovascular clinic, the difference of attitude to HRT between the two groups could best be described as a gulf. The only cardiovascular issues for HRT at that time were concern on the part of the physicians that HRT caused hypertension and thrombosis. Coronary heart disease (CHD) was not part of the agenda. On our part we believed that HRT was not associated with hypertension or thrombosis, arguing that it was metabolically different from the combined oral contraceptive pill from which these risks were extrapolated. Subsequent data have confirmed a lack of a significant adverse blood pressure effect but it was only 15 years later that our confidence over thrombosis risk was reversed (Daly et al., 1996Go).

The interest in a possible link between CHD and postmenopausal estrogen use emerged as a result of the publication of a sequence of observational studies through the 1980s and 1990s suggesting reduced CHD risk in these women. For many years the only observational study suggesting a lack of benefit was the Framingham Study (Wilson et al., 1985Go). Against this it was possible to cite a couple of dozen studies suggesting reduced CHD events in HRT users, or ex-users (Stampfer and Colditz, 1991Go). Through the 1980s the focus on the mechanisms that might mediate a benefit concentrated primarily on lipid metabolism and as a result we saw HRT companies argue with each other over which product offered the best lipid profiles. The study of mechanisms became more sophisticated in the 1990s as a range of methodologies made possible functional studies of blood vessels and their cellular components. These studies reinforced the concept that estrogen exerts a beneficial effect on blood vessel physiology with respect to CHD risk. Additionally the 1990s also brought a major series of studies of CHD in the monkey model (Mikkola and Clarkson, 2002Go) and human angiographic data that substantiated the same view (Sullivan et al., 1990Go).

The 1990s coincided with the emergence of the importance of ‘evidence based medicine’ which emphasized the importance of best-objective evidence, especially randomized controlled clinical trials, and the importance of studies based on clinical event end points rather than surrogate measures. The new orthodoxy was straightforward to comply with where the topic was a common clinical event that could be affected over a relatively short time span. A good example of this with HRT, is the use of HRT to relieve vasomotor symptoms where studies of fewer than one hundred women of six months duration, and with placebo control, are sufficient to demonstrate efficacy. The HRT studies on osteoporosis had been based on bone density changes as a surrogate for fracture but the newer drugs, notably the bisphosphonates and raloxifene, were put through very large placebo controlled trials of several years duration in women at increased risk of fracture in order to provide the fracture end point data that the regulatory agencies had started to demand. No HRT manufacturer appeared prepared to undertake such trials, licences for use in osteoporosis being already in place. As a result in recent years the centre stage in osteoporosis management has been occupied by the bisphosphonates influenced by their now superior evidence for a reduction in fractures based on trials rather than the observational studies on which the HRT fracture evidence was based.

By the 1990s it was common to argue that the most important health effect of HRT in population terms was the reduction in CHD because this is the predominant cause of death and morbidity in older women overshadowing the positive effect on osteoporosis risk and the gradually accepted adverse effect on breast cancer risk and venous thromboembolism. For HRT the question of whether the case required randomized trial support as ‘proof’ of the effect was much argued over. Any trial aimed at testing the ability of HRT to prevent CHD would be very expensive because it would inevitably be of long duration and very large because of the relatively low clinical event rates in a primary prevention setting. For strong advocates of HRT the huge investment was sometimes criticised as unnecessary, partly because of practicality and partly because the case for HRT benefit in CHD was so strong. For those with an anti-HRT perspective the lack of clinic trial evidence was a key weakness undermining the whole case for HRT in prevention of CHD, especially once there was clinical end point trial evidence that statins were effective in reducing CHD risk in a secondary prevention setting. With no HRT manufacturer apparently interested in embarking on the huge expense of a major clinical end point trial it is possible that the situation could have continued indefinitely but the climate in the United States was right for public funding to be used to support major trials to address the key questions.

The most likely route to an answer in the relatively short term was a placebo controlled CHD secondary prevention trial studying the effect of HRT in women at high risk of CHD clinical events because they had recently experienced myocardial infarction. This is a different scenario from the primary prevention of CHD in women with healthy coronary arteries but it promised to yield a result in a trial of shorter duration and involving smaller numbers of women than a primary prevention trial. It was, however, still a major undertaking. As well as the funding of the HERS trial (the CHD secondary prevention trial) funding was allocated for a huge long-term placebo controlled HRT trial in healthy women (The Women’s Health Initiative Study). This would be powered to act as a CHD primary prevention trial but also would address the spectrum of positive and negative effects of HRT when used in the long-term. In both these US trials the HRT formulation selected used the predominant estrogen used in US HRT, conjugated equine estrogens (CEE) and the progestogen, medroxyprogesterone acetate (MPA), which had emerged as a potentially cardiovascular system (CVS) friendly progestogen because of its lipid profile. The regime in both studies was to be continuous estrogen and continuous progestogen because this offers the potential for women on treatment being bleed free. This would maximize the chance that the treatment allocation could remain blinded for as many women as possible and it was thought that continuous combined HRT was likely to emerge as the most popular regime for long-term HRT in postmenopausal women, as has indeed been the case.

The HERS trial reported in 1988 (Hulley et al, 1998Go) and failed to demonstrate secondary prevention of CHD over four years. Indeed the clinical outcome was adverse in the first year in the HRT arm. This unexpected result lead to extensive debate and many questions were raised which have partly been addressed by the subsequent publication, a few months ago, of the continuation of HERS as HERS II (Grady et al, 2002Go). The failure to demonstrate CHD benefit in the extension was a further disappointment to those who suspected that the trend, after the first year of HERS, in favour of the HRT group might continue beyond year 4. Within weeks of the HERS II publication came the premature cessation of the estrogen/progestogen arm of the WHI Study at five years (Writing Group for the Women’s Health Initiative Investigators, 2002Go). The termination had been triggered by thresholds of adverse outcome having been crossed (breast cancer combined with the global outcome statistic). The estrogen versus placebo aspect of the trial continues since an adverse threshold has not been reached.

With this publication we have the first very large long-term placebo controlled randomized trial of HRT. It provides trial outcome data that could previously only be estimated from observational studies. These data include demonstration of significant benefit as well as significant hazard. On the positive side the fracture data (clinical vertebral, hip and other fractures) confirm significant reductions in fractures of which the hip fracture reduction is the most important new information. Similarly the previous epidemiological estimates of a reduction in colon cancer (Grodstein et al., 1999Go) is confirmed. On the negative side the risk of breast cancer is very much in keeping with the estimates provided by the meta-analysis of observational studies (Collaborative Groups on Hormonal Factors in Breast Cancer, 1997Go). The endometrial cancer data did not reveal a significant difference in risk between HRT and placebo but with a trend to lower risk on the continuous combined HRT, again in keeping with the epidemiological clues (Weiderpass et al., 1999Go).

The central focus of debate is necessarily on the cardiovascular events. The expected venous thromboembolism risk is confirmed and the previously ambiguous observational data on stroke are reflected in an adverse risk of stroke in the HRT group. The expected reduction in CHD events in the HRT group in this primary prevention CHD study was not confirmed. There were significantly more CHD events in the HRT group suggesting an increase in risk of CHD in HRT users. The publication was followed by major professional, media and public interest. Those who have had an inherent anti-HRT stance can feel confirmed in their view and many have expressed this. Those who have an inherent confidence in the positive effect of HRT are asking questions about the trial’s findings requesting details which should emerge but which could not be included in the first publication. Over the next year or two we can hope that outstanding issues may be clarified.

We have subsequently had the publication of a meta-analysis (Beral et al, 2002Go) which places the WHI findings in the context of the other RCT data. The conclusion is confirmation that HRT does not reduce the risk of coronary heart disease and that the effect is neutral at relative risk 1.11 (95% CI = 0.96–1.30). In the same meta-analysis an increased risk of stroke was reported as relative risk 1.27 (95% CI = 1.06–1.51)

As an Editor of the Cochrane Menstrual Disorders and Subfertility Group I would wish to confirm my belief that the well conducted and adequately powered randomized controlled trial is the best approach to getting close to discovering whether treatments are effective or not. In the Cochrane approach there is great stress placed on defining the patient group being treated, on defining the treatment intervention and on defining the research questions posed by a trial.

The WHI study is extremely important to the reproductive medicine community and there is huge interest in understanding its findings. In forming a view on what we can learn from this major study I take the view that it is inherently valid for the intervention used in the population studied but we then need to ask what is it telling us. How far can the findings be extrapolated, beyond the intervention used, to related interventions and how should the data be interpreted beyond the overall patient group studied to other groups, or subgroups, of women. We have been given an overall balance of hazards and benefits for the overall 50 to 79 year age group. We know that between 50 and 79 years there are major changes in the incidence of different conditions. I would like to understand this balance as it applies to a woman 50 to 60 years old and whether this appears to be different from that of a woman in the 70 to 79 year age-group. I would like to know the age-specific balance of risks and benefits and to be able to put these in the context of age-specific absolute risk. I can then provide my patients with more meaningful advice.

There are many questions to be addressed concerning the WHI publication and it is to be expected that these will be covered in meetings in the near future and in further publications. The questions that are being raised cover many facets of the study and its analysis and include:

(i) The study plan: Since the breast cancer result was predicted by meta-analysis and was a major contributor to the stopping decision, was this appropriate?

(ii) The patient group: To what extent was this a truly healthy population. The older women, in particular, could have had subclinical disease. Was this really a primary prevention trial? Did the younger age groups have the same levels of hazard as the older women? Did the necessary unblinding of many cases because of bleeding, affect the conclusions which can be drawn?

(iii) The intervention: To what extent can the findings be considered to be a class effect for HRT? Since there are data showing differences in CVS biology between different forms of HRT (CEE versus estradiol; MPA versus some other progestogens) can the CVS effect be considered a class effect? Can any conclusion be drawn concerning the effect of HRT doses that are different from the trial dose?

(iv) The analysis: Which methods of analysis are the most appropriate? Is the correlation testing adjustment method the most appropriate? Hormone replacement therapy remains controversial but we have never before had so much data to discuss. National advisory bodies are beginning to issue revised guidance on the use of HRT. It is only with time to reflect will it be clear whether the established place of HRT in women’s health care will change. How are the new data to be incorporated into treatment strategies and are these strategies to be different following WHI and the other recent data?

Human Reproduction would like to stimulate debate on this very important study. With the current diversity of views on HRT and we would like to encourage contributions from across the full spectrum of opinion.


    References
 Top
 References
 
Collaborative Groups on Hormonal Factors in Breast Cancer (1997) Breast cancer and hormone replacement therapy: Collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet, 350, 1047–1059.[CrossRef][ISI][Medline]

Beral, V., Banks, E. and Reeves, G. (2002) Evidence from randomized controlled trials on the long-term effects of hormone replacement therapy. Lancet, 360, 942–944.[CrossRef][ISI][Medline]

Daly, E., Vessey, M.P., Hawkins, M.M., Carson, J.L., Gough, P., and Marsh, S. (1996). Risk of venous thromboembolism in users of hormone replacement therapy. Lancet, 348, 977–980.[CrossRef][ISI][Medline]

Grady, D., Herrington, D., Bittner, V., Blumenthal, R., Davidson, M., Hlatky, M., Hsia, J., Hulley, S., Herd, A., Khan,S. et al. (2002) Cardiovascular disease outcomes during 6.8 years of hormone therapy. Heart and estrogen/progestin replacement study follow up (HERS II). JAMA, 288, 49–57[Abstract/Free Full Text]

Grodstein, F., Newcomb, P.A. and Stampfer, M.J. (1999). Postmenopausal hormone therapy and the risk of colorectal cancer: a review and meta-analysis. Am. J. Med., 106, 574–582.[CrossRef][ISI][Medline]

Hulley, S., Grady, D., Bush, T., Furberg ,C., Herrington, D., Riggs, B., and Vittinghoff, E. (1998) Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/progestin Replacement Study (HERS) Research Group. JAMA, 280, 605–613.[Abstract/Free Full Text]

Mikkola, T.S. and Clarkson, T.B. (2002) Estrogen replacement therapy, atherosclerosis, and vascular function. Cardiovasc. Res., 53, 605–619.[CrossRef][ISI][Medline]

Stampfer, M.J. and Colditz, G.A. (1991) Estrogen replacement therapy and coronary heart disease: a quantitative assessment of the epidemiologic evidence. Prev. Med., 20, 47–63.[ISI][Medline]

Sullivan, J.M., Zwaag, R.V., Hughes, J.P., Maddock, V., Kroetz, F.W., Ramanathan, K.B. and Mirvis, D.M. (1990) Estrogen replacement and coronary artery disease. Effect on survival in postmenopausal women. Arch. Intern. Med., 150, 2557–2562.[Abstract]

Weiderpass, E., Adami, H.O., Baron, J.A. Magnusson, C., Bergstrom, R., Lindgren, A., Correia, N. and Persson, I. (1999) Risk of endometrial cancer following estrogen replacement with and without progestins. J. Natl Cancer Inst., 91, 1131–1137.[Abstract/Free Full Text]

Wilson, P.W., Garrison, R.J., and Castelli, W.P. (1985) Postmenopausal estrogen use, cigarette smoking, and cardiovascular morbidity in women over 50. The Framingham Study. N. Engl. J. Med., 313, 1038–1043.[Abstract]

Writing Group for the Women’s Health Initiative Investigators (2002) Risks and benefits of estrogen plus progestin in healthy postmenopausal women. Principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288, 321–333.[Abstract/Free Full Text]





This Article
Extract
FREE Full Text (PDF )
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (6)
Request Permissions
Google Scholar
Articles by Barlow, D. H.
PubMed
PubMed Citation
Articles by Barlow, D. H.