Influence of Estrogen Plus Progestin on Breast Cancer and Mammography in Healthy Postmenopausal Women

Abstract & Commentary

Synopsis: In an intent-to-treat analysis, the use of conjugated equine estrogen combined with continuous medroxyprogesterone acetate by menopausal women increased the hazard ratio of invasive breast cancer to 1.24.

Source: Chlebowski RT, et al.JAMA. 2003;289:3243-3253.

The aim of the present analysis is to "better understand the relationship between breast cancer and exposure to estrogen plus progestin." The data are from the Women’s Health Initiative. A total of 8506 women were randomized to the conjugated equine estrogen plus continuous medroxyprogesterone acetate arm, and 8102 women started in the placebo arm. The data from the women who are receiving conjugated equine estrogen alone are not available for this analysis. Randomization appears to have been achieved for important variables such as smoking, alcohol use, prior hormone use, age of menarche, parity, risk of breast cancer, and so forth. There were 349 incident invasive breast cancers and 84 in situ breast cancers; 42% of the participants using CEE+MPA and 38% of those using placebo stopped their medication. Also, 6.2% of those in the CEE+MPA arm and 10.7% in the placebo arm started using another hormone preparation while in the study (drop-ins). Participants were followed regardless of adherence.

The hazard ratio for invasive breast cancer was 1.24 with an unweighted confidence interval of 1.01-1.54 (P = .003) and an adjusted confidence interval of 0.97-1.59 (nonsignificant). When nonadherent women were omitted from the analysis, the hazard ratio became 1.49 (CI, 1.13-1.96). The hazard ratio for in situ breast cancer was 1.18 (CI, 0.77-1.82), which is statistically nonsignificant. The breast cancers in the women using CEE+MPA were slightly larger (1.7 vs 1.5 cm), but the biological significance of this is unknown because only 4 women who had breast cancer died in each arm. More disturbing is the finding that women in the CEE+MPA arm were more likely than those in the placebo arm to have advanced stage (regional/metastatic 25.4% vs 16.0%, P = .04) and to be node positive (25.9% vs 15.8%, P = .03). Also, more women in the CEE+MPA arm had abnormal mammograms (9.4% vs 5.4%). Histologic type and estrogen and progesterone receptor status were similar in both groups. More reliable molecular markers of tumor prognosis such as cyclins were not determined. Prior hormone users were at somewhat lower risk of breast cancer when compared to those without prior use (refer to Table 2 in the source article). Chlebowski and colleagues favor the interpretation that CEE+MPA stimulates breast cancer growth and delays breast cancer diagnosis by mammography. However, they note in the discussion that the data and safety monitoring board indicated on May 31, 2002, that there was not an increase in breast cancer in the conjugated equine estrogens-only arm. 

Comment by Sarah L. Berga, MD

The results of this study need to be considered in conjunction with the manuscript that follows next in this issue of JAMA (see the following abstract & commentary). The present manuscript is carefully and clearly written. The discussion is interesting and balanced. Chlebowski et al note that the data "cannot inform questions regarding risk associated with other oral or topical menopausal hormone therapies" and that "the rates of discontinuation of study medications in both study groups are limitations." Figure 1 in the article suggests a divergence of risk for invasive breast cancer that appears by year 4 of use, but the hazard ratio for year 6 is lower in the CEE+MPA arm than that observed in years 4 and 5, so it is not clear if this early risk is sustained. If not, then it might not be "real." Also, while the data suggest that the risk for a breast cancer with a poor prognosis is increased by CEE+MPA use, the lack of data regarding mortality and the absence of more reliable molecular markers of prognosis preclude a firm interpretation. As I recently noted in an earlier review,1 clinical staging does not correlate well with outcome, while the tumor marker cyclin E does.2 In short, this detailed analysis of the WHI data suggests that use of CEE+MPA modestly increases the risk of invasive breast cancer. We are told that as of last year, there was not an increased risk of invasive breast cancer in the CEE-only arm. If true, this would strongly implicate the progestin component as the source of the increased risk of invasive breast cancer. Even if the risk of death is not increased in those with invasive breast cancer, women with more advanced stages will be subjected to more aggressive evaluation and intervention for the treatment of breast cancer. Given that we cannot guarantee that other hormonal approaches do not carry the same risk, the increased probability of more procedures alone might serve as a deterrent to use other hormonal preparations. On the other hand, if other types of menopausal hormone therapy confer some of the putative benefits of menopausal hormone use, such as neuroprotection from dementia or preservation of habitus and sexual responsivity but with a null risk of breast cancer, it would be tragic that fear alone could motivate so many women to be denied these and other benefits. Clearly, this is not a time for dogmatism.


1. Berga S. OB/GYN Clinical Alert. 2003;19(9):66-67.

2. Keyomarsi K. N Engl J Med. 2002;347:1566-1567.