Editor’s Note: A paper published in the July edition of the American Journal of Public Health claims that the landmark 15-year Women’s Health Initiative (WHI) study harmed women by advising those who’d undergone hysterectomies to avoid estrogen therapy. The paper, by Philip Sarrel and his colleagues, claims that in women ages 50–59 (i.e., those closest to menopause), between 19,000 and 91,000 more deaths have occurred over a 10-year period than would have occurred had more women in this age group been taking estrogen.
The finding has caused an uproar, resulting in widespread media coverage and causing many women and clinicians to question the WHI’s guidance. We asked Sylvia Wassertheil-Smoller, Ph.D., a WHI principal investigator at Einstein (one of the 39 national study sites), to provide her perspective on the latest study.
A main goal of the WHI was to assess whether estrogen therapy was safe for postmenopausal women. The WHI had two separate clinical trials, with a total of about 27,000 women participating: one for women with uteruses and one for women without uteruses.
In trial #1, women with uteruses were randomized to receive estrogen plus progestin or a placebo. These women had to receive progestin along with the estrogen to protect their uteruses from developing cancer. In trial #2, women without uteruses were randomized to take either estrogen alone or a placebo.
In 2002 the bombshell dropped: it turned out that in trial #1, the estrogen-plus-progestin group had a greater risk for serious health problems than the placebo group, and this trial was stopped early as a result. Subsequently, prescriptions for hormones dropped markedly. Everyone was upset (especially the pharmaceutical makers of the hormone pills, as well as some doctors who fervently believed that all postmenopausal women should take hormones). Soon a few doctors came up with the “timing hypothesis,” saying that the WHI results were due to the older age of the women in the trials—and arguing that if women were given estrogen starting right at menopause, it would be protective rather than harmful.
The Sarrel article set out to bolster that hypothesis by looking at one subgroup of WHI women: younger women ages 50–59 who were enrolled in trial #2, in which women without uteruses received either estrogen alone or a placebo.
In that subgroup of younger women, those taking the placebo experienced higher mortality compared with women taking estrogen. Sarrel and his colleagues applied that excess mortality among those taking the placebo to estimate the number of women in the U.S. who’d had hysterectomies and might have been taking estrogen but weren’t; they then came up with a hypothetical number of excess deaths due to the drop in estrogen use.
There are several important reasons that this approach and the resulting findings are not valid.
- Estimates are not hard data and can mislead. The researchers have no new data but instead have carried out mathematical extrapolation based on estimates. To demonstrate their assertion, they would need to check whether mortality rates rose as prescriptions for hormones fell—the sort of analysis that showed that breast cancer incidence in the U.S. actually dropped after the results of the WHI estrogen-plus-progestin trial became public.
- The WHI found that stroke risk and dementia risk with estrogen alone are higher than with a placebo, and similar to the risks for estrogen plus progestin. These are devastating diseases. We don’t know what the mortality or stroke and dementia excess would be for women who received estrogen alone between ages 50 and 59 and continued to take it as they got older, when their risk of these diseases would continue to increase. The risk/benefit picture might change.
- Women older than 59 have greater absolute risks for stroke and dementia than younger women do. But Sarrel’s message could lead women to continue hormone therapy well past the age of 59, greatly increasing their risk for these serious health problems. Sarrel and colleagues would therefore be doing the same thing that they fault the WHI for doing—getting the wrong message out.
To understand the results of various studies, people must understand the nuances regarding what a study can and cannot answer. It seems to me that the authors of this paper do not illuminate nuances; they do not point out that the two trials were different—and therefore fail to note that not all postmenopausal women should routinely receive hormones. Nor do they point out that, in both trials, hormone therapy is consistently shown to increase risk for stroke and dementia. We do not know—and neither do Sarrel and his colleagues—what would happen to women who started hormone therapy at age 50 and continued it into older ages, as was the usual case when the WHI was begun.
The provocative findings of this paper—which are based on estimates rather than new data—are an overreach with potentially negative consequences. They cloud the public’s understanding of risks and benefits.