One hundred years ago, George Beatson described partial clinical responses in three premenopausal women with advanced breast cancer following surgical oophorectomy.’ His conclusion that ‘we must look … to the ovaries as the seat of the exciting cause of (breast) cancer’ laid the foundations for the current understanding of the oestrogen sensitivity of this disease and the concept underlying endocrine breast cancer therapy that tumours are deprived of oestrogen derived from the plasma or synthesized locally in breast tissue itself. Epidemiological, in vitro and in vivo experimental evidence has since confirmed the role of endogenous oestrogens in the promotion of breast cancer cell growthand recent prospective observational studies have shown significant associations between elevations in circulating serum oestradiol and its metabolites and the risk of developing postmenopausal breast cancer or breast cancer recurrence.Ovarian ablation and tamoxifen (a mixed oestrogen agonist and antagonist) have been shown to improve significantly the disease-free and overall survival of women with oestrogen receptor positive breast cancer; the more widespread use of adjuvant therapy is considered to account for the significant (10%) reduction in breast cancer mortality in England and Wales observed between 1985 and 1995. Attention has, furthermore, been focused on the link between oestrogens and breast cancer, given that tamoxifen and the selective oestrogen receptor modulator (SERM) raloxifene have both been associated with a decrease in the incidence of Oestrogen receptor-positive breast cancer.The benefits of hormone replacement therapy in relieving oestrogen deficiency symptoms and in the longer term conferring protection against osteoporosis and probably the primary prevention of arterial disease are recognized. Despite these obvious benefits the overwhelming evidence for the oestrogen dependency of breast cancer implies that Hormone replacement therapy exposure will increase the risk of developing breast cancer and disease recurrence. However, review of available clinical data suggests that this may not be the cause and demonstrates that breast cancer is a disease that defies theoretical predictions.

Indirect evidence that hormone replacement therapy may not stimulate the growth of breast cancer cells

Although endocrine therapy has an established role in the treatment of breast cancer, our understanding of the basic mechanisms responsible for the efficacy of such treatment is incomplete given the observation that mean oestradiol levels in breast tumours from premenopausal and postmenopausal women do not differ significantly despite a 10-fold difference in respective circulating oestradiol levels.Tamoxifen is an effective treatment in premenopausal women with Oestrogen receptor-positive disease even though it induces a hyperoestrogenic state, in which serum oestradiol levels can exceed those observed during the peak of the follicular phase of the menstrual cycle by 2 — 3-fold.

The assumption that oestrogens alone are responsible for the development and progression of breast cancer can be questioned further in view of the paradoxical behaviour of breast cancer when exposed to high serum levels of exogenous or endogenous oestrogen, where prognosis does not appear to be adversely affected. For example, pharmacological doses of oestrogens were used in the palliation of postmenopausal women with advanced breast cancer before the advent of tamoxifen, and in randomized adjuvant breast cancer trials diethylstilbestrol has been shown to have an equivalent disease-free and overall survival benefit compared with tamoxifen in postmenopausal women, irrespective of tumour Oestrogen receptor status. There is also some evidence that the prognosis for breast cancer patients who become pregnant after a diagnosis of breast cancer may be improved, although the possibility that those becoming pregnant were a self-selected, good prognosis group at low risk of recurrence cannot be excluded. In a series of 227 patients aged under 35 years at diagnosis who received chemotherapy, lower rates of recurrence and death were reported in 25 women who subsequently became pregnant.Von Scholtz et al compared 50 women becoming pregnant after diagnosis with a control group of 2000 women. The hazard ratio, after adjustment for nodal status and age, was 0.48 (95% confidence interval 0.02-1.29) for the group who became pregnant. In isolation, none of these arguments can be used as evidence that Hormone replacement therapy will not influence breast cancer risk. They do, however, reinforce the view that the aetiology of this disease is complex and poorly understood.

Does oestrogen replacement therapy increase the risk of developing breast cancer?

Recent prospective observational studies have shown a significant association between the risk of developing postmenopausal breast cancer and increases in circulating serum oestrogens. This increase in risk, however, occurs with very small increments in serum oestrogens, none of which exceed the upper limit of the normal postmenopausal range, i.e. <100 pmol/1. Collectively, these studies suggest that exposure to oral, transdermal, or low dose implant (i.e. 25 ug oestradiol) replacement therapy that results in a mean range of serum oestradiol of 200 — 360 pmol/l will increase the risk of developing postmenopausal breast cancer. However, if the risk of developing breast cancer were related simply to levels of circulating oestrogens, it is surprising that the numerous clinical studies undertaken to evaluate the effect of Hormone replacement therapy on the risk of this disease have produced contradictory findings. Since all these studies have been case — control or cohort in design and therefore lacked appropriate randomized controlled groups for comparison, they are open to the influence of bias in patient selection, recall, surveillance, and interview techniques. Increased mammographic surveillance, for example, could explain the increase in the reported incidence of ductal carcinoma in situ in Hormone replacement therapy compared with non-Hormone replacement therapy users.21 To allow for the effect of systematic errors, confounders, and biases, it has been recommended that individual epidemiological studies should not be considered persuasive of an association unless the lower limit of the calculated 95% confidence interval falls at least above a threefold risk. No single study investigating the association between Hormone replacement therapy and breast cancer risk has reported an increase in risk of that magnitude.

Only one placebo-controlled, randomized trial of Hormone replacement therapy has been completed in which breast cancer was a primary end-point. Here, a sequential preparation was prescribed (i.e. conjugated oestrogens 2.5 mg/day and medroxyprogesterone acetate 10 mg/day) and was not shown to have any significant effect on the risk of developing breast cancer. Despite the long follow-up of 22 years, however, patient numbers were far too small for these results to be considered conclusive.

In an attempt to clarify this controversy, several meta-analyses have been undertaken. Overall, these suggest that current long-term use of Hormone replacement therapy (i.e. > 10 years) is associated with a slight increase in the risk of developing breast cancer (range of relative risk 1.2 — 1.3). The most recently published re-analysis of 51 individual studies, worldwide,calculated that for every year that Hormone replacement therapy is used it increases the relative risk of developing breast cancer by 1.023 (95% confidence interval 1.011-1.036) and that with more than 5 years’ use (although the median duration use of Hormone replacement therapy in this subgroup of women was in fact 11 years) the relative lifetime risk of developing breast cancer is 1.35 (95% confidence interval 1.21-1.49). Based on these calculations, it has been estimated that if Hormone replacement therapy is commenced at the age of 50 years, 5 years’ continuous use will be associated with two extra breast cancers, 10 years’ use with six extra breast cancers, and 15 years use with 12 extra cancers per 1000 women. A similar pattern of risk with the use of Hormone replacement therapy was also calculated for women with a family history of breast carcinoma, but the 99% confidence intervals were very wide and encompassed unity. Failure of investigators in individual trials to accurately document family history renders it impossible to determine whether women investigated were truly at an increased risk of breast cancer or not. Tumours developing in women at high risk of breast cancer owing to inherited mutations in the BRCA1 and BRCA2 genes, which account for 75% of familial disease, may have a hormone resistant phenotype in that they are usually high grade, Oestrogen receptor and progesterone receptor negative and therefore resistant to tamoxifen therapy.Reproductive factors, however, do appear to be important in the development of these tumours. Rebbeck et al reported a significant decrease in the incidence of breast cancer in women with BRCA1 mutations who underwent bilateral oophorectomy for the prophylaxis of ovarian cancer. However, the use of’add-back’ Hormone replacement therapy by some women to prevent the unwanted effects of premature oestrogen deficiency did not appear to negate the benefit of oophorectomy for breast cancer risk. The collaborative re-analysis did not examine the role of Hormone replacement therapy in women at high risk of developing breast cancer due to benign breast disease, which encompasses a range of conditions of the breast, of which only atypical ductal or lobular hyperplasia are associated with a significant 3 — 4-fold increase in the risk of developing breast cancer. Most studies have not categorized benign breast disease and therefore it is very difficult to interpret the available data. In their meta-analysis, Dupont and Page concluded that there was no definitive evidence to support the contraindication of Hormone replacement therapy in women with a history of benign breast disease. This is further supported by their retrospective cohort study, in which accurate categorization of benign breast disease was made.

Table Meta-analyses of Hormone replacement therapy and breast cancer risk.

No of studies Any Hormone replacement therapy use Duration of use
relative risk (95% Cl) relative risk (95% Cl)
Not stated 1.01 (0.95-1.08)
28 1.07 (1.00-1.05)
16 1.0 1.30 (1.20-1.60) > 15 years
10 1.0 1.23 (1.04-1.51) > 10 years
37 1.06 (1.00-1.12) 1.20 (no Cl) >S years
Combined Hormone replacement therapy (n = 3 studies) 0.99 (0.72-1.36) ever use
31 1.40 (1.20-1.63) current use 1.23 (1.08-1.40) > 10 years
Combined Hormone replacement therapy (n = 4 studies) 1.13 (0.78-1.64) ever use
51 1.35 (1.21-1.49) > 5 years
Combined Hormone replacement therapy(not stated) 1.53 (SE 0.33) > 5 years

Combined Hormone replacement therapy and breast cancer risk

The role of progestins in the aetiology of breast cancer remains to be clarified. It has been hypothesized that progestins will confer protection against the development of breast cancer or breast cancer recurrence. This is based on the observations that the relative risk of developing breast cancer is increased fivefold in women with luteal phase progesterone deficiency and that the survival of premenopausal women with early stage breast cancer may be increased if they are operated on during the luteal phase of the menstrual cycle. Alternatively, it has been postulated that progestins may increase the risk of developing breast cancer, since most proliferative breast activity occurs during the luteal phase of the menstrual cycle, when both oestrogen and progesterone levels are elevated. Complicating matters further is the fact that it is not known whether the class of progestin prescribed or the pattern of progestin prescription (i.e. continuous or combined) is relevant in this context.

Theoretical arguments favour the avoidance of the 19-nortestosterone derivatives, since, compared with the C21 progesterone derivatives, the former exhibit relatively greater androgenic and oestrogenic properties. However, the 19-nortestosterone derivative lynestrenol significantly reduces the Oestrogen receptor content of cellular aspirates from women with benign breast disease, suggesting an inhibition of oestrogen stimulation of breast epithelial cells in vivo. Furthermore, breast cancer risk has not been reported to be increased in women treated solely with this class of progestin for benign breast disease (relative risk 0.48, 95% confidence interval 0.25-0.90).

The hypothesis that continuous combined, rather than sequential Hormone replacement therapy will confer protection against the development of breast cancer is based on in vitro data in which the continuous application of progestins with oestrogens induces a sustained inhibitory effect on oestrogen-driven cell replication.This inhibitory response appears to be mediated via a variety of cellular pathways, including an increase in the enzymatic conversion of oestradiol to oestrone sulphate, promotion of apoptosis, inhibition of the proto-oncogenes c-myc and c-fos, and a decrease in the breast cancer growth factor cathepsin D. As progestins down-regulate cellular progesterone receptor (progesterone receptor) content, continuous application of unopposed progestin will attenuate this inhibitory effect. The collaborative re-analysis reported that current long-term use of combined Hormone replacement therapy might confer a greater lifetime risk of developing breast cancer than would oestrogen alone.However, the number of breast cancer cases upon which this risk estimate was based was too small for this finding to be considered definitive. Subsequent to this, five observational studies have been published which are all consistent in demonstrating an association between long-term exposure to combined Hormone replacement therapy and an increase in the risk of developing breast cancer. Unfortunately, since the individual data are expressed in different ways, it is impossible to determine whether this risk is equivalent to that for unopposed oestrogen, or of a greater magnitude. Risk has been reported to be increased with the use of cyclic rather than continuous administration of C21 progesterone derivatives but the converse has been reported with the 19 non-testosterone-derived progestins. The small numbers in these subgroup analyses prevent any firm recommendations from being made.

Since most women prescribed Hormone replacement therapy have not had a hysterectomy, this is obviously an area that warrants further evaluation.

Hormone replacement therapy, breast density, and mammography

Subset analysis of the placebo-controlled, randomized Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial has confirmed findings of numerous observational studies reporting an increase in mammographic breast density with exposure to postmenopausal Hormone replacement therapy. The results from the PEPI Trial furthermore suggest that mammographic density is increased to a greater degree in women prescribed combined Hormone replacement therapy preparations. After 12 months, density increased by 3.5% (95% confidence interval 1.0-12.0%) in women treated with conjugated equine oestrogen (conjugated oestrogens) alone compared with increases of 19.4% (95% confidence interval 9.9-28.9%) and 23.5% (95% confidence interval 11.9-35.1%) in women prescribed conjugated oestrogens plus continuous medroxyprogesterone or conjugated oestrogens plus cyclic medroxyprogesterone, respectively. It is worth noting, however, that the 95% confidence intervals were wide, particularly for the use of combined therapy.

Although increased breast density is associated with an increase in the risk of developing breast cancer (>75% breast density, relative risk for developing breast cancer 4.35, 95% confidence interval 3.1 — 6.1) and the use of postmenopausal unopposed or combined Hormone replacement therapy is associated with an increase in breast epithelial density,whether either can be used as an accurate surrogate measure for the risk of developing breast cancer or breast cancer recurrence in women taking Hormone replacement therapy is unresolved. Hormone replacement therapy does appear to reduce the sensitivity and specificity of mammography but it has been reported that cessation of Hormone replacement therapy for as little as 2 weeks before mammography could be sufficient to overcome this screening problem.

The Million Women Study, a survey of Hormone replacement therapy use in women attending for mammographic breast cancer screening in the United Kingdom, will not provide definitive data about breast cancer risk in the absence of a randomized control group for comparison (Institute of Cancer Research, Epidemiology Unit/NHS Breast Screening Programme). There is no evidence that healthy women on Hormone replacement therapy require more frequent mammograms than are received through the National NHS Breast Cancer Screening Programme.

The influence of Hormone replacement therapy on breast tumour biology

The recent collaborative re-analysis demonstrated that the small increase in the risk of developing breast cancer associated with Hormone replacement therapy exposure disappears completely within 5 years of it being stopped. In conjunction with case reports of breast tumour regression following withdrawal of Hormone replacement therapy, this suggests that Hormone replacement therapy promotes the growth of pre-existing breast cancers rather than initiating carcinogenic change in the breast. There is no evidence at present, however, to suggest that Hormone replacement therapy stimulates exclusively the growth of Oestrogen receptor-positive disease.

Evidence of the effect of Hormone replacement therapy on breast tumour biology and extrapolation to clinical events is unclear. Most studies are flawed in that Hormone replacement therapy was not taken up to the day of breast biopsy. It has been shown quite clearly in studies of preoperative hormonal manipulation of primary breast cancers that significant changes in breast tumour receptor status and proliferation rates (i.e. Ki67) can occur after periods as short as 1 week. Hormone replacement therapy (both unopposed and combined preparations) was administered up to the day of biopsy in the study of Hargreaves et and shown to induce breast tissue progesterone receptor content. However, the lack of any expected associated increase of breast cell proliferation is probably explained by the methodology used, which precludes exclusion of a significant effect on the latter.

In contrast, percutaneous administration of progestin has been shown to inhibit oestrogen-induced breast epithelial cell proliferation. Although assessment of cellular proliferation was more robust in this latter study, the application of percutaneous oestrogen and progestin was via the breast, which results in a significant increment in breast oestradiol levels. As such a difference in breast oestradiol levels has not been seen in comparisons between premenopausal and postmenopausal women, this suggests that administration through the breast is the significant issue but it cannot be related directly to conventional Hormone replacement therapy.

The identification of two distinct subtypes of the Oestrogen receptor, ERα and ERβ, in the rat, mouse, and human has resulted in a revaluation of the molecular basis for oestrogen activity. At present, the significance of their different tissue distributions and ligand selectivities is unknown and requires further extensive study. How this will influence our understanding of the effect of Hormone replacement therapy on breast cell proliferation remains to be answered.

The effect of Hormone replacement therapy on breast cancer mortality

With the exception of findings from the Nurses Health Study, use of Hormone replacement therapy before a diagnosis of breast cancer does not appear to have an adverse effect on cause-specific or overall mortality. Breast tumours diagnosed in women with a history of current or past use of Hormone replacement therapy tend to be less clinically advanced in that they are smaller in size and better differentiated but, although some investigators have found a predominance of lymph node negative disease in women exposed to Hormone replacement therapy, others have not.

Table Mortality in breast cancer patients with a history of previous Hormone replacement therapy exposure at time of breast cancer diagnosis.

Cancer mortalityrelative risk (95% Cl) Breast cancer mortalityrelative risk (95% Cl)
0.53, p < 0.007 (Cl not stated)
0.22 (no Cl) 0.73 (0.44-1.22)
0.68 (0.52-0.87)
0.70 (0.55-0.85) 0.76 (0.45-1.06)
0.80 (NS) 0.81, p > 0.05 (Cl not stated)
No reduction (p > 0.01)
0.71 (0.62-0.81) current use 0.76 (0.56-1.02) current use
1.04 (0.90-1.13) past use 0.83 (0.63-1.09) past use
0.75 (0.48-1.17) use < 5years
0.79 (0.83-1.67) use > 5years
0.2 (0.1-0.3) use < 5 years
0.7 (0.5-0.9) use 5-9 years
0.7 (0.5-0.9) use > 10years
0.84 (0.75-0.94)
0.59 (0.40-0.87) natural menopause
0.76 (0.54-1.09) surgical menopause
0.92 (0.55-1.54) family history of breast
Hormone replacement therapy 92% v no Hormone replacement therapy 86% (p = 0.07)
0.78 (0.63-0.96) use < 5 years
0.77 (0.56-1.08) use > 5years
0.5 (0.3-0.8) current use at diagnosis
2.2 (0.9-5.2) > 144 months after diagnosis

Gapstur et al observed that Hormone replacement therapy use is associated with an increase in the incidence of invasive breast cancer with a favourable histology (i.e. tubular and papillary cancers) but not with the more common invasive ductal or lobular carcinomas, which have a worse overall outcome. The beneficial effect of Hormone replacement therapy on subsequent breast cancer mortality seems to occur irrespective of tumour hormone receptor or axillary lymph node status. Although prior use of Hormone replacement therapy has been reported to be associated with a small increase in the incidence of ipsilateral breast cancer recurrence, the incidence of distant metastatic disease is significantly reduced and mortality is unaffected.

This apparent favourable effect of Hormone replacement therapy could be influenced by surveillance bias, in that women requesting Hormone replacement therapy tend to have more of an interest in general disease prevention activities such as breast cancer surveillance, or that women with breast cancer are not generally prescribed Hormone replacement therapy and will be among the non-users in observational studies. However, these mortality studies suggest that, even if Hormone replacement therapy increases the incidence of breast cancer or promotes disease recurrence, it may not have a detrimental effect on breast cancer versus all-cause mortality, which is the important end-point.

The use of hormone replacement therapy in women previously treated for breast cancer

Tamoxifen and Hormone replacement therapy — antagonistic or synergistic?

Hormone replacement therapy and the breast: Conclusion

Endogenous oestrogens are implicated in the aetiology of breast cancer, but the role of exogenous hormones remains unclear. Theoretical predictions that Hormone replacement therapy will increase breast cancer risk and disease progression have not been substantiated by observational data but it is probable that Hormone replacement therapy promotes the growth of pre-existing breast cancers rather than increasing the incidence of the disease per se. This growth promotion, however, does not appear to be associated with an increase in breast cancer mortality and may reflect a selective biological effect on the growth of less aggressive tumours. Two large randomized trials of Hormone replacement therapy in healthy women are now underway in the United States (the Women’s Health Initiative, which was set up by the National Institutes of Health) and the United Kingdom (the MRC WISDOM study; the Women’s International Study of Long-Duration Oestrogen use after the Menopause). In conjunction with planned randomized trials in breast cancer survivors, it is hoped that, ultimately, more reliable data on the long-term benefits and risks of Hormone replacement therapy will be provided. Until such data are available, clinicians must rely on the findings of observational studies, but in this it is essential that the limitations incurred by designs of such studies are appreciated.


Selections from the book: “Hormone Replacement Therapy and the Menopause”, 2002

Share →