Can a Study of a Million Women be Wrong?

Hormone Replacement Therapy and Breast Cancer Link Questioned

/ Author:  / Reviewed by: Joseph V. Madia, MD

It used to be standard practice: a woman stopped having her menstrual cycles and she went on hormone replacement therapy (HRT) to relieve symptoms of menopause.

That practice was abruptly halted in 2002 when part of a large study - the Women's Health Initiative - was stopped because of a reported link between HRT and increased incidence of breast cancer.

Another even larger study in England - the Million Women Study - reached the same conclusions and also announced that HRT increased a postmenopausal woman's risk of developing breast cancer.

A recent review is now questioning the design, methods, and results of the Million Women Study.

A Harvard professor agrees that the link between HRT and breast cancer may be more media hype than fact.

The studies

Million Women Study

The Million Women Study (MWS) is a national study of a number of aspects of women’s health, involving more than one million United Kingdom (UK) women aged 50 and over. The effects of hormone replacement therapy use is the main focus of the study.

The MWS produced four reports in 2003, 2004, 2006 and 2011. It's the largest study of its kind and one of three major research efforts that led to warnings about the safety of long-term HRT use.

This study evaluated newly diagnosed cases of breast cancer in women invited to receive breast screening between 1996 and 2001.

Women's Health Initiative

The Women's Health Initiative (WHI) was conducted in the United States and included various clinical trials and an observational study, which involved a total of 161,808 generally healthy postmenopausal women.

The clinical trials tested the effects of postmenopausal hormone therapy and other health issues such as diet and vitamin intake. The trial looking at HRT had two studies that looked at:

  • estrogen-plus-progestin therapies (a combination known to prevent endometrial cancer) in women with a uterus
  • estrogen-alone study of women without a uterus

Both studies found that more women taking HRT developed breast cancer than women taking a placebo or no HRT at all. The number of women who developed breast cancer in the WHI was higher in women taking estrogen plus progestin, and this was the study arm that was stopped in 2002.

Disputing these findings

A group of researchers led by Professor Samuel Shapiro of  Department of Public Health and Family Medicine at the University of Cape Town Medical School in Cape Town, South Africa, recently reviewed the Million Women Study.

Researchers found the design of the study to be flawed which in turn skewed the results. The authors' main concerns included the following:

  • Since the study participants were taking part in a breast cancer screening program and were invited to participate in the MWS study, these women were probably more likely to be HRT users and already aware of breast changes.
  • This "detection bias" led to higher numbers of breast cancers being found.
  • Breast cancer can grow slowly and could have been in a woman's body for years before it was detected, regardless of her HRT use.
  • It's generally believed that it takes about a decade for a tumor to reach the size of 1 cm.
  • In the MWS, the time from a participant's enrollment until breast cancer detection was about 1.5 years.
  • Authors say that the likelihood of a fatal breast cancer being 22 percent higher in women using HRT during this timeframe was "biologically implausible."
  • Even 1-2 extra cases per 1000 women would have invalidated the overall results.
  • The authors also say there was crucial data missing from the MWS.
  • In the third report (2006),  follow-up data on HRT use was not available for 57%-62% of the study.

Quantity doesn't guarantee quality

“The name ‘Million Women Study’ implies an authority beyond criticism or refutation," the authors wrote.

“Yet the validity of any study is dependent on the quality of its design, execution, analysis and interpretation. Size alone does not guarantee that  the findings are reliable,” they said.

Symptomatic of a larger problem

dailyRx asked our Contributing Expert, Daniel B. Kopans, M.D., professor of radiology at  Harvard Medical School and senior radiologist - the Breast Imaging Division - Massachusetts General Hospital, to comment on this study.

He agrees with the authors and says that this review points out a larger problem.

"I am an expert in breast cancer detection and diagnosis and not in hormone replacement therapy. However, the paper by Shapiro et. al....raises important questions about the biases of investigators, analysis of data, peer review, press releases and the immediate impact of scientifically questionable conclusions," Dr. Kopans writes in an email to dailyRx.

"As with this paper, and the importance of media attention, it has become common for journals to send out press releases prior to the publication of controversial topics. This means that the media will "break" their stories on the same day that the paper is published.

"The information, and its interpretation, that has been reviewed by a limited number of individuals, will be widely disseminated before it has been exposed to the medical and scientific communities and more rigorous review. Not infrequently, misinformation gets disseminated. Like ripples on the water from a dropped stone, the misinformation spreads and cannot be recalled.

"As the authors point out, the results of the Million Women Study and the Women's health Initiative study were rapidly disseminated. As a result, the use of hormone therapy dropped precipitously. It now appears that the conclusions were not justified by the methodologies of the studies and the analyses by the authors.

"There are important issues that bear on the conclusions that were not provided to the public at the time of publication. This also occurred with the Women's Health Initiative. The authors of that study were clearly intent on concluding that hormone therapy increases the risk of breast cancer. They played down the fact that estrogen use, in their study, actually, reduced the risk of breast (although a borderline result) while the small increase in risk from combined HT was emphasized.

"The WHI authors failed to provide a context for that risk and it was made to sound huge. In fact, a 26% increase in breast cancer risk for women with a baseline risk of 2/1000 per year amounts to 2.52 cancers/1000 women per year. The authors also failed to provide a frame of reference such as the fact that this risk was less than the risk from the routine consumption of alcoholic beverages.

"There are times when the conclusions and dissemination of information is an emergency, but most studies, such as these, do not represent such an emergency. It would be preferable (although unlikely) for these papers to be published and digested by the medical and scientific communities before dissemination by the media so that the problems can be identified (such as with the Million Women Study) before the public is misled," Dr. Kopans concludes.


This review was published January 16, 2012 in The Journal of Family Planning and Reproductive Health Care, a specialty journal of the BMJ (British Medical Journal) Group.

Authors disclose that they are currently consulting or have previously consulted with the product manufacturers mentioned in their study.


Review Date: 
January 16, 2012