Breast Cancer Risk in the WHI Estrogen-Only Arm

Abstract & Commentary

By Leon Speroff, MD, Editor, Professor of Obstetrics and Gynecology, Oregon Health and Science University, Portland, is Editor for OB/GYN Clinical Alert.

Synopsis: The estrogen-only arm of the WHI reports a reduction in invasive breast cancer.

Source: Stefanick ML, et al. Effects of conjugated equine estrogens on breast cancer and mammography screening in postmenopausal women with hysterectomy. JAMA. 2006;295:1647-1657.

The updated breast cancer results in the canceled estrogen-only arm of the Women’s Health Initiative (WHI) are based on 237 cases of invasive breast cancer and 55 cases of cancer in situ, diagnosed by the February 29, 2004, date of study cancellation.1 Overall, a reduction in invasive breast cancer in the treated group did not reach statistical significance (HR = 0.80; CI = 0.62-1.04). However, when the group adherent to treatment (accounting for the 54% drop-out rate) was analyzed, there was a statistically significant 33% reduction in invasive breast cancer (HR = 0.67; CI = 0.47-0.97). No effect was observed on in situ disease. The reduction in invasive breast cancer was confined to ductal carcinomas (no significant effect was seen with lobular cancer) and in women who had no prior exposure to hormones. A protective effect of estrogen treatment was observed in women without a history of benign breast disease, and in women without first-degree relatives with breast cancer. No interaction was seen with body mass index; oophorectomy status; age at menarche, first birth, or menopause; or with prior oral contraceptive use. The invasive breast cancers in the treated group were slightly larger (but the standard deviation was very wide), and there was no significant difference in the number of positive nodes. The treated group of women had more mammograms requiring follow-up, leading to more biopsies or aspirations.


This updated report from the WHI differs from the initial report from the cancelled estrogen-only arm in that a reduction in invasive breast cancer achieved statistical significance in those women who adhered to their estrogen treatment. The most important and difficult-to-answer questions are the following:

  1. Do the differences between the estrogen-progestin and estrogen-only arms of the WHI reflect an adverse impact of progestins on breast cancer risk?
  2. Do the epidemiologic studies reflect an impact of hormones on pre-existing tumors?

The cancelled estrogen-progestin arm of the WHI reported an increase in invasive breast cancer that reached statistical significance in the 5th year of exposure.1 Even the WHI investigators have cautioned clinicians to avoid comparing the 2 trial arms because the participants in the 2 arms were considerably different.2 In regards to breast cancer risk, the women in the estrogen-only arm had a higher rate of previous exposure to hormones and for longer durations. In the current WHI report, the investigators performed 2 calculations that led them to conclude that differences in the patient characteristics in the 2 arms could not explain the different conclusions. First, the 5-year Gail breast cancer risk estimates of the participants in the 2 arms were similar. Second, The rate of invasive breast cancer per year was similar in the placebo arms of the 2 trials. Nevertheless, the differences in previous hormone exposure continue to be reason for caution because the estrogen-only arm found a reduction only in women who had not previously used hormones.

To this date, not a single study has reported that hormone use increases the risk of ductal in situ breast cancer. If hormone exposure was causing the growth of new tumors, one would expect a difference in in situ disease as well as invasive disease. Is it possible that the apparent differences in estrogen-progestin and estrogen-only exposure are due to a greater differentiation effect of estrogen-progestin? In a Kaiser cohort study from Southern California, a reduction in breast cancer mortality was statistically significant in users of estrogen-progestin, but not in users of estrogen-only.3

The breast cancer results in the WHI do not allow us to answer the above questions with any confidence. Exposure to estrogen-progestin either has a greater risk of breast cancer, or pre-existing tumors respond differently to various hormone regimens, accounting for differences in epidemiologic reports.

Many of the WHI conclusions are the result of statistical manipulations. For example, the investigators concluded that there was no interaction with oophorectomy status, despite the fact that the treated group had fewer bilateral oophorectomies. How confident can clinicians be with conclusions derived from Cox proportional hazard models? In another analysis, there was no trend with time for the risk of invasive breast cancer. What does this mean? Is an effect on pre-existing tumors seen only early during exposure? Keep in mind that all studies that have reported an increase in breast cancer with hormone therapy have found the increase very rapidly, within a few years. This contrasts with the well-recognized relatively slow growth of tumors until they reach a clinically detectable stage.

Rather than bringing clarity to this troublesome problem, the recent studies only raise more questions. Until we have more definitive data, I am comfortable telling patients that there is either a small increase in breast cancer risk with hormone therapy or the studies reflect an effect on pre-existing tumors. I also believe it is appropriate to share with patients the strong possibility that hormone users have lower grade, lower stage disease with better outcomes.


  1. Chlebowski RT, et al. Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women. The Women’s Health Initiative Randomized Trial. JAMA. 2003;289:3243-3253.
  2. Stefanick ML, et al. The Women’s Health Initiative postmenopausal hormone trials: overview and baseline characteristics of participants. Ann Epidemiol. 2003;13(9 Suppl):S78-S86.
  3. Chen W, et al. Mortality following development of breast cancer while using oestrogen or oestrogen plus progestin: a computer record-linkage study. Br J Cancer. 2005;93:392-398.