Search:
Menopause Live - IMS Updates
InFocus

Date of release: 21 January, 2009

A rapid decline in the incidence of breast cancer following cessation of menopausal hormone therapy: sense or nonsense?


At the Breast Cancer Symposium held in San Antonio, Texas, in December, 2008, Chlebowski and colleagues [1] referred to correlated declines in the use of menopausal hormone therapy (HT) and the incidence of breast cancer, as reported in an analysis of the SEER database by Ravdin and colleagues [2], and they presented data from the Women’s Health Initiative (WHI) randomized, controlled trial (RCT) and observational study on changes in incidence following cessation of estrogen plus progestin (E + P) (RCT) or HT (observational study). In both WHI studies, the incidence of breast cancer declined following cessation of use, and the findings were interpreted as suggesting that ‘… cessation … is associated with a rapid reduction … [in breast cancer incidence] … which is not explained by mammography utilization change (sic) and [the findings] support the hypothesis that the recent reduction … seen in certain age groups is predominantly related to a decrease in menopausal therapy use (sic).’

Comment

To date, only the Symposium abstract [1] has been published, and the only additional information available is a copy of the slides presented at the symposium. Ordinarily, it would therefore be inappropriate to comment until the findings have been published in full. However, since the San Antonio presentation has received wide publicity, here we make provisional comments, subject to revision.  
In both the correlational data of Ravdin and colleagues [2] and in the WHI data [1], whatever brought about the declines in the incidence of breast cancer, it cannot have been cessation of the use of HT. Turning first to the paper by Ravdin and colleagues [2], contrary to what they claimed, their data did not show a correlated decline in the use of HT and breast cancer incidence. The first WHI report was published on July 17, 2002 [3]. Yet, in the report of Ravdin and colleagues, in the relevant age group (≥ 50 years), the decline in breast cancer incidence commenced 3 years earlier, in 1999 (their Figure 1). In addition, Li and Daling [4], who also analyzed the SEER database, and Glass and colleagues [5], who analyzed the Kaiser Permanente cohort, have reported a decline in incidence that commenced in 1997/98. Glass and colleagues also observed a corresponding decline of 4% in screening compliance between 2000 and 2004 – a change that could have accounted for part or the entire drop in incidence observed by Ravdin and colleagues [2]. 
In the report of Ravdin and colleagues [2], among women aged ≥ 50 years, the initial decline in incidence was followed by a plateau during the next 3 years, and only then by the decline to which they referred. Even if it is assumed that there were two declines, only the second of which was related to the use of HT, that decline commenced in mid-2002, at the same instant as the WHI findings were reported. As explained below, on pathological and clinical grounds, that sequence of events is impossible. And still further, correlated declines in HT use and breast cancer incidence have not been confirmed in other studies [6-11]. In those countries in which screening remained stable (e.g. Scotland [9]), a recent decline in incidence was not recorded. In the US, heterogeneous changes in screening practices, and in risk factors, could well have accounted for the observed decline in breast cancer incidence [10,11]. In short, the underlying assumptions that led to the WHI analysis were not valid.
Turning next to the San Antonio presentation [1], no epidemiologically identified association can be interpreted as causal if it is incompatible with established pathological and clinical evidence. Pathologically, kinetic studies have established that, on average, more than a decade (and more likely two decades) must elapse from the time of induction to the time of malignant transformation from a pre-malignant state to clinically detectable breast cancer [12]. Hence, what is at issue is promotion, not initiation, and, once promotion has been accelerated, it cannot thereafter be decelerated. In addition, once transformed, breast cancer cannot then instantaneously ‘untransform’. For the hypothesis of Chlebowski and colleagues to be tenable, it has to be assumed that any promotional effect of HT on breast tumor growth stops instantaneously as soon as HT is stopped and that early breast cancer can be reversed. And once malignant, breast cancer cells are not so hormone-dependent that they can immediately revert to a benign or pre-malignant state. Thus cessation of the use of HT cannot bring about an instantaneous and rapid decline in the incidence of breast cancer.
Clinically, identical breast cancers can be diagnosed in different women at widely separated times, sometimes years apart, depending on factors such as the frequency and thoroughness of breast examinations and mammographies, the interpretation of which, in turn, can be influenced by awareness of exposure status, and on other factors such as clinical acumen, education, access to medical care, culture, and so on. Hence, even under the assumption of a reduced risk following cessation of HT use, the incidence of breast cancer cannot decline immediately: there has to be a ‘latent interval’ before a perceptible decline becomes evident.  
Based on these considerations, whatever the explanation, cessation of the use of E + P or HT cannot explain the declines in breast cancer incidence reported by Chlebowski and colleagues [1]. At most, the increased risks observed in the WHI data among current users, if causal, can only have been due to acceleration of breast cancer promotion, not induction, and, even then, based on kinetic considerations, it stretches plausibility to suggest that E + P can accelerate the promotion to clinically evident breast cancer within 2 years, as in the WHI RCT [3]. In addition, that study was beset by multiple methodological defects [13,14], and it did not establish that even the current use of E + P or HT increases the risk of breast cancer (although other epidemiological studies (but not all) are more suggestive [15]). And when it comes to the further claim, now made, that cessation of the use of E + P or HT brings about an instantaneous reduction in the incidence of breast cancer, that sequence of events is impossible.
As has been pointed out in earlier critiques of the analysis of E + P use in the RCT, there are also more plausible alternative explanations of the WHI findings [11,12]. The differences in incidence rates of breast cancer among E + P and placebo recipients were in all instances sufficiently small to be readily accounted for by bias and confounding, and the same limitations now apply to the WHI analysis of past use.  
More particularly with regard to the possibility of detection bias among women who previously used and did not use E + P or HT, there were no differences in the use of mammography. The WHI investigators therefore claimed that detection bias had been excluded as a possible explanation for the declining incidence of breast cancer following cessation of therapy. However, even in the best of hands, the sensitivity and specificity of mammography are poor, and more common diagnosis of breast cancer among currently and previously exposed women who underwent mammography could readily have explained the small differences that were observed. There was also quantitative evidence to support that likelihood: the number of biopsies was greater in the exposed and previously exposed than in the non-exposed women [1], indicating that the physicians who examined the mammograms were aware that most of the former had received active treatment. Thus, against the background of poor specificity, they could selectively have diagnosed early breast cancers among the exposed women that would otherwise have gone undetected. In addition, following cessation of exposure, progressively diminishing concern about the possibility of breast cancer could also have explained the seemingly progressive decline in incidence.
Based on the evidence reported in the WHI Symposium abstract [1] and presentation, the claim that the risk of breast cancer progressively declines as soon as HT ceases cannot be defended.

Comentario

Samuel Shapiro
Department of Public Health and Family Medicine, University of Cape Town, South Africa

Anne Gompel
Unité de Gynécologie Endocrinienne, Hôtel-Dieu de Paris, Université Paris Descartes, Paris, France

    References

  1. Chlebowski RT, Kuller L, Prentice R, et al. Breast cancer after stopping estrogen plus progestin in postmenopausal women in the Womens Health Initiative. Presented at San Antonio Breast Cancer Symposium, December 2008, Abstract 64.


  2. Ravdin PM, Cronin KA, Howlader N, et al. The decrease in breast-cancer incidence in 2003 in the United States. N Engl J Med 2007;356:1670-4. Published April 19, 2007.
    http://www.ncbi.nlm.nih.gov/pubmed/17442911

  3. Rossouw JE, Anderson GL, Prentice RL, Writing Group for the Womens Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Womens Health Initiative randomized controlled trial. JAMA 2002;288:321-33. Published July 17, 2002.
    http://www.ncbi.nlm.nih.gov/pubmed/12117397

  4. Li CI, Daling JR. Changes in breast cancer incidence rates in the United States by histologic subtype and race/ethnicity, 1995 to 2004. Cancer Epidemiol Biomarkers Prev 2007;16:2773-80. Published December, 2007.
    http://www.ncbi.nlm.nih.gov/pubmed/18086787

  5. Glass AG, Lacey JV Jr, Carreon JD, Hoover RN. Breast cancer incidence, 1980-2006: combined roles of menopausal hormone therapy, screening mammography, and estrogen receptor status. J Natl Cancer Inst 2007;99:1152-61. Published August 1, 2007.
    http://www.ncbi.nlm.nih.gov/pubmed/17652280

  6. Kliewer EV, Demers AA, Nugent ZJ. A decline in breast cancer incidence [Letter to the Editor]. N Engl J Med 2007;357:509-10. Published August 2, 2007.
    http://www.ncbi.nlm.nih.gov/pubmed/17674464

  7. Zahl P-H, Maehlen J. A decline in breast cancer incidence [Letter to the Editor]. N Engl J Med 2007;357:510-11. Published August 2, 2007.
    http://www.ncbi.nlm.nih.gov/pubmed/17674462

  8. Mueck AL, Wallwiener D. Brestkrebsrate und HRT-verordnungen: differierende date aus USA und Europa. Frauenarzt 2007;48:812-15.


  9. Vaidya JS. Re: Declines in invasive breast cancer and use of postmenopausal hormone therapy in a screening mammography population. JNCI 2008;100:598-9. Published April 16, 2008.
    http://www.ncbi.nlm.nih.gov/pubmed/18398100

  10. Pines A, Sturdee DW, Birkhaeuser MH, et al.; International Menopause Society. HRT in the early menopause: scientific evidence and common perceptions. Climacteric 2008;11:267-72. Published August, 2008.
    http://www.ncbi.nlm.nih.gov/pubmed/18645691

  11. Gompel A, Rozenberg S, Barlow DH; the EMAS Board members. The EMAS 2008 update on clinical recommendations on postmenopausal hormone replacement therapy. Maturitas 2008;61:227-32. Published November 20, 2008.
    http://www.ncbi.nlm.nih.gov/pubmed/19028033

  12. Dietel M, Lewis MA, Shapiro S. Hormone replacement therapy: pathobiological aspects of hormone-sensitive cancers in women relevant to epidemiological studies on HRT: a mini review. Hum Reprod 2005;20:2052-60. Published August, 2005.
    http://www.ncbi.nlm.nih.gov/pubmed/15932918

  13. Shapiro S. Risks of estrogen plus progestin therapy. A sensitivity analysis of findings in the Womens Health Initiative randomized controlled trial. Climacteric 2003;6:302-10. Published December, 2003.
    http://www.ncbi.nlm.nih.gov/pubmed/15006251

  14. Shapiro S. Adverse neoplastic and cardiovascular outcomes of HRT: the validity of the evidence. Endocrine 2004;24:203-10. Published August, 2004.
    http://www.ncbi.nlm.nih.gov/pubmed/15542886

  15. International Agency for Research on Cancer (IARC). Combined Oestrogen-Progestin Contraceptives and Combined Oestrogen-Progestin Menopausal Therapy. Volume 91. France: Lyon, 2008.