15 April, 2013:
WHI investigators keep hammering the issue of hormone therapy (HT) and related increased risk of breast cancer. Their 'money' is actually on what they call the estrogen + progestin (E+P) users, since they cannot claim a real risk in the estrogen-alone users. So now they bring new data from the WHI observational study [1]. Here is the Abstract:
In the WHI randomized trial, estrogen plus progestin increased both breast cancer incidence and mortality. In contrast, most observational studies associate estrogen plus progestin with favorable prognosis for breast cancers. To address differences, a cohort of WHI observational study participants with characteristics similar to the WHI clinical trial was studied. We identified 41,449 postmenopausal women with no prior hysterectomy and mammogram-negative within 2 years who were either not hormone users (n = 25,328) or estrogen and progestin users (n = 16,121). After a mean of 11.3 (SD = 3.1) years, with 2236 breast cancers, incidence was higher in estrogen plus progestin users than in non-users (0.60% vs. 0.42%, annualized rate, respectively; hazard ratio (HR) 1.55, 95% confidence interval (CI) 1.41–1.70, p < 0.001). Women initiating hormone therapy closer to menopause had higher breast cancer risk with linear diminishing influence as time from menopause increased (p < 0.001). Survival after breast cancer, measured from diagnosis, was similar in combined hormone therapy users and non-users (HR 1.03, 95% CI 0.79–1.35). On a population basis, there were somewhat more deaths from breast cancer, measured from cohort entry (HR 1.32, 95% CI 0.90–1.93, p = 0.15), and more all-cause deaths after breast cancer (HR 1.65, 95% CI 1.29–2.12, p < 0.001) in estrogen plus progestin users than in non-users.
Thus the investigators concluded that, consistent with WHI randomized trial findings, estrogen plus progestin use is associated with increased breast cancer incidence. Because prognosis after diagnosis on combined hormone therapy is similar to that of non-users, increased breast cancer mortality can be expected.
Comment
These data look convincing at first glance; however, there are so many methodological and statistical issues embedded in the small details that raise serious questions on possible potential biases and the reliability of the results and conclusions. A group of IMS members are now writing a response to that article, to be sent as a Letter to the Editor, but here are some relevant points to be considered.
First, it is time now to stop looking at the WHI study as investigating the effects of E+P. The WHI used a specific hormonal product, namely the combination of conjugated equine estrogen + medroxyprogesterone acetate (CEE+MPA), at a specific dose, in a continuous regimen, and administered orally. A class effect should not be assumed based on well-established data demonstrating differences in the metabolic effects of other types of estrogens and progestogens, and with other dosages and other modes of administration. Furthermore, in the E3N study, different breast cancer risks were found for different hormonal preparations [2]. The authors of this paper, while acknowledging the contrary findings of reduced risk with CEE-alone, fail to provide a rationale for their claim that E+P is associated with risk when E-alone is not. The above principles were addressed in the 2011 IMS recommendations on the use of HT [3]. The currently formulated, updated version of the IMS recommendations, soon to be finalized and published, will discuss this topic in more detail.
Second, as a general rule of thumb, in the case of a HR of 1.5 or so (the range found in the current WHI-HT-breast cancer analyses, even when statistically significant), it is virtually impossible to discriminate between bias and causation, and one should be very careful in making decisive clinical declarations, especially in the set-up of a methodologically complicated study such as the WHI. Just for examples, the following points should be considered:
(1) Since 2002, when the preliminary E+P results from the WHI clinical trial were released and dispatched globally to the public, later studies, especially those involving the evaluation of breast cancer risk, might be biased as participating women were already well informed about the alleged risks, which may have an impact on their decision to continue taking hormones and on the level of clinical aggressiveness in evaluating suspicious lesions in women with known exposure to HRT.
(2) Inclusion criteria for the WHI observational study comprised a negative mammography during the 2-year period prior to enrollment; data for overall survival were collected at baseline and then only once in 3 years, calling into question the interval and time trends described. Such circumstances may certainly be a cause for bias.
(3) The use of the gap-time calculations (time elapsed from menopause until the start of hormone use) is problematic because WHI participants were recruited in the 1990s, and many observational study women joined it because they had been on HT for years and did not want to discontinue treatment in order to be randomized in the parallel placebo-controlled clinical trial. When those women started HT in the 1970s, 1980s and 1990s, it was common for HT to be given near the time of menopause and continued for the long term. In this case, a short time gap, older age and longer duration of use are all related. So older women are more likely to have had shorter gap times, and therefore short gap times are a surrogate for age, which is well-known to be the strongest risk factor for breast cancer [4]. For prior users, these analyses do not distinguish between short and long durations of use, even though short gap times provide greater opportunity for long-term use. As for current users, only 17% of the WHI clinical trial women randomized to CEE+MPA had a gap time of < 5 years, whereas the corresponding figure in the WHI observational study was 75%. In the latter situation, which is the basis of the current paper, the gap time construct is hopelessly intertwined with age: the effect of short gap time would be more accurately described as the association of older age and long duration of use.
Last, but not least, is the way in which data were reported by the WHI investigators. Phrasing sometimes makes the music. They write: 'there were somewhat more deaths from breast cancer, measured from cohort entry (HR 1.32, 95% CI 0.90–1.93, p = 0.15)', while in fact they should have said that death from breast cancer was similar in the hormone users and non-users, since CIs crossed 1, and the p value was non-significant.
Nonetheless, the WHI was a very important study, a milestone in menopause medicine, and we should not attempt to wipe its results from the table. On the other hand, we have to recognize its pitfalls and potential biases while evaluating the individual benefit–risk balance of HT.
Amos Pines, MD
Tel-Aviv University School of Medicine, Tel-Aviv, Israel
and
Robert D. Langer, MD, MPH
Principal Scientist and Medical Director, Jackson Hole Center for Preventive Medicine
References
1. Chlebowski RT, Manson JE, Anderson GL, et al. Estrogen plus progestin and breast cancer incidence and mortality in the Women's Health Initiative observational study. J Natl Cancer Inst 2013 Mar 29. Epub ahead of print
http://www.ncbi.nlm.nih.gov/pubmed/23543779
2. Fournier A, Berrino F, Clavel-Chapelon F. Unequal risks for breast cancer associated with different hormone replacement therapies: results from the E3N cohort study. Breast Cancer Res Treat 2008;107:103-11.
http://www.ncbi.nlm.nih.gov/pubmed/17333341
3. Sturdee DW, Pines A; International Menopause Society Writing Group. Updated IMS recommendations on postmenopausal hormone therapy and preventive strategies for midlife health. Climacteric 2011;14:302-20.
http://www.ncbi.nlm.nih.gov/pubmed/21563996
4. Langer RD. Re: Estrogen plus progestin therapy and breast cancer in recently postmenopausal women. Am J Epidemiol 2009;169:784-5
http://www.ncbi.nlm.nih.gov/pubmed/19208722
