Caroline is 49, and since her periods stopped a year ago, she has suffered from debilitating menopausal symptoms, with hot flashes, vaginal dryness and mood swings. However, as both her parents died from cancer, she has so far refused HRT as she has heard that it can cause cancer.
The debate about the influence of hormone therapy after menopause on breast cancer rates has resulted in a large volume of both clinical and basic research. The implication derived from the mass media is that hormone therapy can cause or increase the risk of breast cancer. Invariably this is the first question a menopausal woman wishes to discuss before she is to initiate hormone therapy, and not uncommonly, it is the point of note that she wishes to address if she is to continue with long-term usage. The etiology of breast cancer is multifactorial and is likely to involve more than just one factor when it does occur. Breast cancer will develop in about 12% of women who live to 90 years of age. The large majority of breast cancers are sporadic in nature and over 80% of breast cancers in postmenopausal women occur in women who have never taken hormone therapy. Factors that have been shown to be associated with an increased risk for breast cancer include obesity, alcohol intake, no full-term pregnancy, first full-term pregnancy > 35 years of age, never having breastfed an infant, early menarche < 12 years, late menopause > 50 years, family history of breast cancer particularly in first-degree relatives, genetic mutations (BRCA1 or BRCA2) and increased breast density. The medical literature is replete with a variety of observational studies that have shown an increase, a decrease or a null-effect on the risk for breast cancer in menopausal women receiving estrogen (E)-only therapy or estrogen and progestogen (E and P) combined therapy. The evidence from the observational studies is not consistent for an association of hormone therapy and breast cancer. The most compelling aspect from these studies is that hormone therapy acts as a promoter of breast cancer tissue and not an initiator of the disease, resulting in an increase in risk of breast cancer only being seen after 5 years of use; and that after stopping the hormone therapy, the risk rapidly returns to that observed in never-users of hormone therapy. The data suggest stimulation rather than induction of neoplasia. The Women’s Health Initiative (WHI) Study initially did not find a statistically significant increase in the occurrence of invasive breast cancer in E and P combined therapy users (hazard ratio (HR) 1.26, 95% confidence interval (CI) 1.00–1.5). There was a null-effect for breast cancer in women who had never used hormone therapy (HR 1.06, 95% CI 0.81–1.38), which was in agreement with the data previously published from the Collaborative Group report in women who had used hormone therapy for less than 5 years (HR 1.05, 95% CI 0.99–1.12). The overall risk of breast cancer in the WHI trial was elevated only in women who had previously used hormone therapy. Further analysis of the WHI using centrally adjudicated data and adjusted 95% CI showed that breast cancer was not significantly increased (HR 1.24, 95% CI 1.02–1.50). These data translate to four extra women developing breast cancer per 1,000 women taking combined hormone therapy for 5 years, or less than 0.1% per annum, compared with 1,000 women not taking hormone therapy. Support for the fact that it is the progestogen which is the more significant factor than estrogen in the risk for breast cancer comes from the E-only arm of the WHI, which showed an absolute reduction of eight per 10,000 person-years in the annual rate of newly diagnosed invasive breast cancers, which was not statistically significant (HR 0.77, 95% CI 0.59–1.01). A significant risk reduction was seen among women who were fully compliant with their treatment throughout the study (HR 0.67, 95% CI 0.47–0.97), which persisted after discontinuing the hormone therapy for up to 10.7 years follow-up. The level of breast cancer protection associated with E-only therapy did not vary by duration of use and among women with breast cancer diagnoses – both overall mortality and breast cancer mortality were significantly lower in the estrogen users. Fewer women died from any cause after breast cancer diagnosis than among the controls. A small increase in risk for breast cancer in lean younger women may be associated with unopposed E-only therapy. The Million Women Study reported that duration of hormone therapy is related to the incidence or occurrence of breast cancer. Both E-only and E and P therapy users had an increased risk for breast cancer but the risk decreased rapidly after stopping the therapy (at 1 year relative risk (RR) 1.03, 95% CI 0.92–1.12). Micronized progesterone and dydrogesterone in E and P combined therapy may be associated with a better risk profile for breast cancer than other synthetic progestogens. More data are still needed to determine whether mode of administration of hormone therapy may impact variably on breast cancer risk .
Tibolone is an analog of the progestogen, norethynedrel. After ingestion it is converted to three metabolites that provide estrogenic, progestogenic and androgenic effects. This selective tissue estrogenic activity regulator (STEAR) is effective in treating menopausal symptoms and seems to have the least likelihood of causing breast tenderness or increasing breast density. It has only been the Million Women Study that has suggested an increased risk for breast cancer (RR 1.45, 95% CI 1.25–1.67) in tibolone users, which was similar to that seen in women using E-only therapy and significantly less than that seen in women using E and P therapy within the study. This finding proved to be a surprise, and was in direct contrast to the findings of the LIFT study, a randomized placebo-controlled trial using tibolone in a fracture prevention trial where a significant reduction in invasive breast cancer was shown (HR 0.32, 95% CI 0.13–0.80) .
There are no data to support that testosterone hormone therapy impacts on the risk for breast cancer or for any of the gynecologic cancers.
In general, women having a first-degree relative with a history of breast cancer are at an increased risk of disease themselves, especially if the affected first-degree relative developed breast cancer before 40 years of age. A family history of a first-degree relative with breast cancer diagnosed after 50 years of age doubles the risk of breast cancer, but if the family member was diagnosed with breast cancer before 40 years of age, it increases the risk for developing breast cancer 6-fold. A second-degree relative with breast cancer increases risk slightly (RR 1.2–1.5). Even though there are only limited data, the data that are available suggest that family history does not impact negatively on risk for breast cancer if hormone therapy is taken. Family history and hormone therapy appear to have independent and no interacting effects on the risk of invasive breast cancer. Data from long-term cohort studies have not shown a further added risk with hormone therapy in those women who had a family history of breast cancer. Provided there are appropriate indications and there are no other contraindications, hormone therapy should not be withheld simply because of the family history of breast cancer.
Baseline mammographic density correlates with breast cancer risk, although this is independent of breast cancer association with hormone therapy. About 8% of the general population will have extremely dense breast tissue and their risk for breast cancer is increased (RR 1.4) compared with women with average breast density. About 39% of women will have heterogeneously dense breasts and these women have a relative risk of 1.2 for breast cancer. Screening sensitivity falls from 88% in breasts composed almost entirely of fat tissue to 62% in women with extremely dense breast tissue, whilst it is 69% in women with heterogeneously dense breasts. Digital mammography has shown to be more effective than film mammography in women with extremely dense breasts.
Unopposed E-only therapy in a woman with an intact uterus causes endometrial hyperplasia, the precursor of most endometrial cancers, and the longer the duration of use, the greater the related increase in the risk of endometrial cancer. About 10% of women taking unopposed E-only therapy will develop complex hyperplasia after 1 year, as will 50% after 2 years and 62% after 3 years. Of importance, is the fact that this risk for endometrial cancer remains increased for many years after stopping the unopposed E-only therapy. Even after 15 years or more without estrogen therapy, there is still a significantly increased risk for endometrial cancer (RR 5.8, 95% CI 2.0–17.0), a concept directly opposite to that found with breast cancer, where the impact of E and P therapy on risk for cancer diminishes rapidly on cessation of therapy and returns to baseline by 5 years after stopping the therapy. Any estrogen therapy given to a woman with an intact uterus must therefore be accompanied by giving progestogens for 10–14 days per cycle as sequential regimens which allow monthly withdrawal bleeding, or as continuously combined regimens which avoid withdrawal bleeding, to counteract the proliferative effect of unopposed estrogen on the endometrium. Many studies have shown the beneficial effect of continuously combined E and P therapy in protecting the endometrium, particularly if used for longer than 5 years. The risk of endometrial cancer decreases significantly with increasing duration of continuous combined E and P therapy, with a 20% decrease during the first 5 years of use (odds ratio (OR) 0.8, 95% CI 0.5–1.3) and an 80% decrease if used for more than 5 years (OR 0.2, 95% CI 0.1–0.8). In contrast this is not seen with sequential E and P therapy in that the risk for endometrial cancer is not increased with short-term use, but with more than 5 years of use there appears to be an increased risk (< 5 years of use OR 1.5, 95% CI 1.0–2.2; > 5 years of use OR 2.9, 95% CI 1.8–4.6). It may therefore be prudent to change all women on sequential therapy to continuous combined E and P therapy after 5 years of use. The long-cycle E and P sequential regimen where the progestogen is given only 3-monthly to reduce the frequency of bleeding and so improve compliance is less likely to prevent endometrial hyperplasia and is therefore not suitable for long-term use. Both the WHI (HR for endometrial cancer 0.8, 95% CI 0.48–1.36) and the Million Women Study (MWS) showed that continuous combined E and P therapy protected the endometrium. The MWS showed a significant 29% decrease in the risk for endometrial cancer in women using combined hormone therapy, although this was not seen for sequential therapy. Endometrial cancer has rarely been reported in women using continuous combined E and P therapy. In the majority of such cases there commonly was a history of unopposed estrogen therapy, sequential therapy use with less than 10 days of progestogen or other risk factors such as a family history of endometrial cancer. The newer low-dose regimens containing continuous combined 0.3 mg conjugated equine estrogen and 1.5 mg of medroxyprogesterone acetate or 0.5 mg estradiol and 0.1 mg norethisterone acetate produce very acceptable bleeding profiles with no apparent evidence of endometrial stimulation as assessed by ultrasound, and at the same time providing symptom relief. The primary role for progestogen use is to prevent endometrial proliferation brought about by the estrogen, and it therefore stands to reason that if it is administered directly to the endometrial cavity it will give high local concentration of progestogen in the endometrium and lower circulating levels than if administered systemically. Intrauterine delivery of progestogen is a logical route of administration. The intrauterine system, Mirena®, delivers 20 µg of levonorgestrel per day, provides effective endometrial suppression and is a very plausible mode of delivering progestogen for continuous combined therapy in postmenopausal women.
Tamoxifen reduces the risk of developing breast cancer and of recurrences in breast cancer survivors, but it also affects the risk for developing endometrial cancer. Tamoxifen has an agonistic effect on the endometrium, increasing the risk of endometrial hyperplasia, endometrial polyps and endometrial cancer. The risk for endometrial cancer is increased significantly (RR 2.70, 95% CI 0.94–3.75), with women > 50 years of age having a greater risk compared with women < 50 years of age. Increased duration of use impacts adversely on risk. Other malignancies shown to be increased include gastrointestinal (RR 1.31, 95% CI 1.01–1.69), fallopian tube cancer, carcinosarcoma and sarcoma of the uterus.
In future, the anticipated wider use of tissue selective estrogen complex compounds (TSEC) in the treatment of menopausal symptoms hopes to avoid the use of progestogens and their role in the development of breast cancer. Bazedoxifene, a SERMS, with concomitant conjugated estrogen as the TSEC, has not been shown to cause endometrial hyperplasia or breast stimulation (no breast tenderness or increased breast density), although at the same time, it is very successful in treating vasomotor symptoms, preventing osteoporosis, reversing vulvo-vaginal atrophy, and improving sleep and quality of life .
Ovarian epithelial cells express estrogen and progesterone receptors, and the receptor expression differs between normal and malignant cells. This gives reason to question whether sex hormones might have a role in ovarian carcinogenesis. The impact of hormone therapy on ovarian cancer risk has been studied in a number of different studies, including meta-analyses, pooled analyses and the WHI randomized clinical trial. The results have been less consistent than those on breast and endometrial cancer, with as many studies showing an increase in risk as there are studies that show a decrease in risk or a null effect on risk. Of the three pooled analyses published, two did not show any increase in risk (RR 1.1, 95% CI 0.9–1.4), although one study did for E-only therapy users (RR 1.7, 95% CI 1.3–2.2), whilst in the only one that analyzed the impact of E and P on risk for ovarian cancer, there was a null effect irrespective of duration of therapy use. Most of the case-control studies have detected a modest non-significant increase in ovarian cancer risk (RR 1.3 for E-only therapy and RR 2.1 for E and P therapy respectively). The magnitude of risk increased with duration of use and was detected in users of E-only therapy as well as in the women using E and P therapy. Of the five cohort studies published, two did not detect an association between hormone therapy and ovarian cancer (RR 0.9) whereas two of the others did (RR 1.3 and 1.7 respectively). In one study mortality from ovarian cancer was increased. In these studies, both the risk of ovarian cancer and the risk of dying from the ovarian cancer was related to the duration of use and increased with longer use of the hormone therapy. Of the different histologic subtypes to be affected, endometrioid adenocarcinoma shows the strongest association with hormone therapy, with mucinous adenocarcinoma being the least likely to be affected. In the WHI randomized trial, the risk for ovarian cancer in women using E and P was mildly increased after 5.6 years of follow-up (HR 1.6) compared with women using placebo. Recently, ovarian cancer risk in postmenopausal women using estradiol-progestin therapy was analyzed in a large Finnish study of over 224,000 women followed up for over 12 years. Ovarian cancer risk was not elevated in women who had used E and P therapy for less than 5 years (standardized incidence ratio (SIR) 1.21, 95% CI 1.06–1.37), and for > 5 years of use the risk for ovarian cancer was similar (SIR 1.26, 95% CI 0.94–1.64). The risk did not differ between sequential or continuous E and P regimens, or between oral or transdermal E and P formulations. The risk elevations for E and P use for > 5 years was only seen for serous adenocarcinoma (SIR 1.56, 95% CI 1.33–1.80). In summary, long-term E-only therapy may be associated with a small attributable risk of ovarian cancer of 0.7 per 1,000 women users for 5 years, whilst either a significantly smaller increase, or no increase at all, is seen with combined E and P hormonal therapy. The risk is so small that it is unlikely to influence prescribing habit.
The data pertaining to the impact of hormone therapy on risk for borderline tumors is sparse and only consists of case-controlled studies. The majority of the studies have not detected any association, although in one study the only significant association found was between ever-use of unopposed estrogen and serous borderline tumors (OR 2.1), whereas in women with a BMI > 26 the risk was close to unity .