Introduction:
Two new studies (1-2) regarding breast cancer risk from hormone replacement therapy gained major media attention. These articles are generating a large number of questions from patients whether they should continue to use hormone replacement therapy and particularly estrogen-progestin combinations. The message from the media is that there is a 20 % increase in risk of breast cancer in menopausal women taking estrogen alone and a 40% increase in risk with estrogen plus a progestin. The risk increases by 1% per year of use for estrogen alone and 8% per year for the combination. At face value these data are alarming to patients. However, the practical significance of the findings is widely misunderstood by the media, patients, and their physicians. The increase in risk reported is the relative risk and not the attributable risk to the patient. For a full understanding of the key issues involved in these results, one needs to distinguish between the terms relative risk, absolute risk, and attributable risk.
Definition of Terms:
Relative risk represents the ratio of risk of breast cancer in women taking or not taking HRT. The term does not take into account the actual frequency of breast cancer in the group being considered.
Absolute risk is determined by multiplying the usual rate of breast cancer in the group being considered by the relative risk. For example, the average 50 year old women has an average risk of developing breast cancer of 2.52 per 100 women over a ten year period. A 10% increase in relative risk from estrogens alone would increase her absolute chances of getting a breast cancer to 2.77 per 100 women.
Attributable risk refers to the number of women who would develop a breast cancer that would not otherwise have occurred without use of estrogen replacement. Using the example above, the difference between breast cancer risk of 2.52 per 100 and 2.77 per 100 represents the increased risk attributable to estrogen or 0.252 per 100 women. Stated in another way, 1 in 397 women would develop a breast cancer that would not have ordinarily occurred.
What are the attributable risks of breast cancer from HRT?
This depends upon how common breast cancer is at each age group and how much the relative risk increases with each year of use of HRT. The recent studies report that relative risk goes up linearly with duration of HRT usage. For example, with estrogens alone the relative risk goes up by 1% per year and with estrogen /progestin combinations the relative risk increases by 8% per year. The implications of this are substantial when comparing short term with long term use of HRT (Figures 1 and 2).
Let’s consider the implication with regard to attributable risk in a 50-year old women. If that patient took estrogen for only two years, the relative risk would increase by a total of 2% (1 percent per year over two years). On average, 2 in 400 50-year old women will have a new breast cancer diagnosed over a two year period. If the relative risk increases by 2%, she will have a 2.04 in 400 chance of getting a breast cancer. The attributable risk due to HRT is then 0.04 per 400 women which represents a 1 in 9925 chance that a women will get breast cancer because of taking estrogen. If we make the same calculations for an estrogen/progestin combination, the actual risk is that 1 in 1241 women will get a breast cancer that she would not otherwise have had. This small increase in risk occurs even though the relative risk is increased by 8% per year over two years or 16% in total. These examples show that short term use of HRT increases the actual risk of breast cancer only slightly.
Let’s now consider taking HRT for ten years. If this is estrogen alone, the risk is increased by 1% per year or 10% altogether over ten years. The normal rate of breast cancer is 2 per 400 women over two years or 10 per 400 at ten years in a 50 year old woman. A 10% increase would make this 11 per 400. The attributable risk related to estrogen is a 1 in 397 extra chance of getting breast cancer. If she took an estrogen /progestin combination, the relative risk increases by 8% per year or 80% overall. She now has a 1 in 50 extra chance of getting breast cancer. This risk is now becoming more meaningful.
Trade off between risks and benefits:
The decision to take HRT is a trade off between risks and benefits. One can calculate the benefits attributable to HRT in the same way that one calculates attributable risks. For the 50 year old woman, this predominantly involves reduction of risk of first cardiovascular event (i.e. primary prevention of heart disease). Based upon the best available data, the attributable benefit from estrogen alone is 0.37 per 100 women taking estrogen for a ten year period. This means that I in 270 women of age 50 will have a cardiovascular event prevented by taking estrogen for this period of time. Prevention of osteoporosis and fracture is another benefit. However, the precise number of women in whom fractures are prevented is difficult to calculate because most fractures occur when women are much older.
A series of calculations can determine breast cancer risk/cardiovascular benefit statistics for women under a variety of clinical circumstance. Table I lists these various calculations which can be used in guiding patients through a decision process regarding HRT.
Examples to use for explaining risk to patients:
Informed patients have seen the news reports about relative risk of breast cancer from use of HRT. It is difficult to explain to them the complex nature of the terms used in these reports and particularly the concept of " relative risk." We have found that the best way for patients to understand relative risk is by thinking about specific examples. A document written specifically for patients and available on our Web page spells out this approach. We do not use the word "attributable risk" but rather "actual risk". Our patients are not epidemiologists and are not familiar with the scientific terminology. In the following section we will review use of these examples which were developed to help women understand the issues.
Let’s consider 1000 women who start HRT at age 50 and another 1000 who decide not to take this medication. After 10 years, we can determine how many women in each group have developed breast cancer. We find that 33 of the 1000 women taking HRT and 30 of the 1000 women not on HRT have developed breast cancer. The relative risk of breast cancer in the women taking HRT is increased 10%. This is the case because the number 33 is 10% higher than the number 30.
We explain to the patient that use of the word "Relative risk" can sometimes cause a lack of understanding of actual risks. Most people think that a 10% increase in relative risk means that 10% of women actually will get breast cancer. However, the word "relative risk" does not consider the actual number of women at all but only the relative differences among groups of women. It also doesn’t emphasize that some women will get breast cancer even if they don’t take HRT.
It is helpful to illustrate how relative risk can be identical but the actual number of women affected substantially different. As an example, let’s say that we identify 1000 sixty year old women who exercise daily and another 1000 who do not. We want to find out the risk of not exercising. We then determine how many of these women have had a heart attack ten years later. We find that 150 in the exercise group and 165 in the non-exercise group have had a heart attack. This is also a 10% increase in relative risk of heart attack (165 divided by 150). The important thing to notice is that the increased relative risk (10%) is the same for both examples given above —the HRT breast cancer example and the exercise and heart disease example. However, if we think about the actual number of women, there are 3 extra cases of breast cancer from HRT and 15 extra cases of heart attack in the non-exercisers. The actual risk is then 5 times higher in the heart disease example even though the relative risk is the same. These two examples show how the term relative risk does not reflect how many people are actually affected. An extreme illustration of how relative risk and actual risk may differ can be seen from another example. In the 1950’s, women were frequently given a hormone called DES during pregnancy to prevent miscarriage. In the women taking DES during pregnancy, their daughters were later found to have a 32 fold (3200%) increase in relative risk of cancer of the vagina (9). This is an extremely rare form of cancer which normally only affects 1 in 32,000 women. The actual risk that vaginal cancer would develop in daughters of DES users was only 1 in 1000. Here is an example where the relative risk is extremely high (relative risk 3200%) but the actual risk (1 in 1000) is quite low.
Practical Use of attributable risks and benefits in decision making:
Short term use of ERT/HRT for menopausal symptoms
Use of ERT or HRT for less than two years causes only a negligible increase in risk of breast cancer in a 50 year old women. For estrogen alone, 1 in 9925 would develop a breast cancer attributable to estrogen. For the combination of estrogen plus a progestin ( HRT), this risk would be 1 in 1241 women. Consequently, a woman could be encouraged to take ERT or HRT short term for menopausal symptoms without a great deal of concern regarding risk of breast cancer.
Long term use of ERT/HRT to prevent heart disease or osteoporosis
The risk of developing breast cancer as a result of taking estrogens long term increases. For the average 50 year old woman, the risk attributable to estrogen alone is 1 in 397 women. For the 60 year old, the respective risk is 1 in 286. This risk is generally outweighed by the cardiovascular benefits with 1 in 270 and 1 in 152 cardiovascular events prevented respectively for the 50 and 60 year old. Prevention of osteoporosis and fractures are additional benefits.
The JAMA study (1) suggests that the risk of breast cancer increases with use of estrogen plus a progestin. For the 50 year old starting HRT, the breast cancer risk attributable to hormonal therapy would be 1 in 50. For the 60 year old, the risk increases to 1 in 36. In the average woman, this risk would not be outweighed by the benefits of preventing cardiovascular events (i.e. heart attacks prevented is1 of 270 50-year old women and 1 in 152 60-year old women). Nonetheless the risk is relatively small in absolute terms. A 50-year old women taking an estrogen/progestin combination as HRT for 10 years has a 96% chance of not getting a breast cancer versus a 98% chance for those not taking HRT.
Other factors which modify risk:
The two most recent studies of HRT and breast cancer risk suggest that only thin patients (i.e. BMI of <24.4 in one study and <25 in the other) experience an increased risk of breast cancer from either estrogen alone or the combination of estrogen and a progestin (1, 6). When limiting analysis to thin women, only those taking HRT long term(i.e. greater than 5 years) had a statistically significant increase in breast cancer risk. Obese women did not have an increased risk of breast cancer attributable to estrogens. Family history of breast cancer, early age of menarche, late age of child bearing, high fat diet, obesity, increased breast density on mammograms, and certain benign breast lesions increase the underlying risk of developing a breast cancer.
Take home message from all available studies:
Definitive data to prove a causal relationship between HRT and breast cancer risk must await randomized, controlled trials such as the Women’s Heath Initiative which will be completed in approximately the year 2005. Accumulating data from several studies suggest that the risk of breast cancer increases with increasing duration of use and is higher with an estrogen–progestin combination than with estrogen alone. This area remains controversial since not all studies support these conclusions (3). While awaiting definitive proof, it is prudent to accept the data from these new studies reporting an increased risk of breast cancer attributable to estrogens and a greater risk from estrogen plus a progestin (Table I). It is recognized the two new studies (1-2) are not perfect. However, three prior studies also suggested that addition of a progestin to estrogen increases breast cancer risk (4-6). A large meta-analysis suggested that HRT increases the relative risk of breast cancer linearly with duration of use. At present, the weight of available evidence supports a link between HRT use and breast cancer.
Alternatives to use of ERT or HRT in women at high risk of developing breast cancer:
A number of alternatives do exist which can be used in place of systemic estrogen to ameliorate problems related to estrogen deficiency (9). Vaginal estrogen can be used to treat the symptoms of urogenital atrophy without increasing systemic estrogen levels to a measurable degree. The SSRI class of drugs can alleviate symptoms of depression. These agents as well as vitamin E and clonidine may alleviate hot flashes to some degree. For maintenance of bone density and prevention of osteoporosis and fractures, the bisphosphonates, and raloxifene, a SERM ( Selective Estrogen Receptor Modulator) and calcitonin can be beneficial. For prevention of heart disease, the HMG-CoA-reductase inhibitors (statins) are proven to be effective. These can be chosen in place of estrogens alone or estrogen –progestin combinations in patients at high risk of breast cancer or fearful of taking HRT. No alternatives to estrogen exist for prevention of Alzheimer’s disease, colon cancer, or macular degeneration---diseases for which preliminary but not definitive evidence exists for the beneficial effects of HRT.
Practical approach to decision making:
Individual risk assessment: A new approach to decision making regarding HRT starts by assessing the individual patient’s risk of developing breast cancer and placing her into the high, intermediate or low risk category. This risk assessment utilizes the Gail model which is based upon responses to 7 questions:* age, race, age of menarche, age first live birth, number of first degree relatives with breast cancer, number of previous breast biopsies, and diagnosis on breast biopsy. These calculations can be made by calling the Cancer Center hot line ( 434-924-29610, by obtaining a computer program from the NIH, or a hand held calculator from the Astra-Zeneca corporation. Definitions of high, intermediate, and low risk are included in Table II. Choices of therapy are then specifically tailored to the individual patient based upon risk versus benefit.
High risk patients: These women are best managed by referral to a high risk breast clinic and treatment with a SERM such as tamoxifen for breast cancer prevention. Alternatives to estrogen can be used for treatment of hot flashes, urogenital atrophy, and mood changes. Tamoxifen acts as an estrogen on bone to decrease risk of osteoporosis and the liver to lower LDL cholesterol levels. Bisphosphonates and statins may be used in addition in patients at high risk of osteoporosis or heart disease.
Intermediate risk patients: Alternatives to estrogen as outlined above are considered in these patients. Short term use of estrogen or an estrogen-progestin combination for two years or less is associated with minimal increased risk of breast cancer and is a reasonable choice. Long term HRT may be reasonable but its risks and benefits should be weighed against the alternatives for HRT or watchful waiting.
Low risk patients: HRT would be usually be recommended for short term treatment of symptoms such as hot flashes, urogenital atrophy, and mood changes or depression. The benefits of estrogen alone for long term prevention of heart disease or osteoporosis usually outweigh the risks. With a combination of estrogen with a progestin for ten years, the benefit /risk ratio narrows and women may choose alternatives to HRT.
Questions
-
What is your race?
-
What is your age?
-
When was your first menstrual period?
-
What was your age at the time of your first live birth of a child?
-
How many first degree (i.e. mother, daughter, sister) relatives do you have with breast cancer?
-
How many prior breast biopsies have you had?
-
Did any biopsy show atypical ductal hyperplasia?
|
|
References:
- Schrairer C, Lubin J, Troisi R, Sturgeon S, Brinton L, Hoover R. Menopausal Estrogen and Estrogen-Progestin Replacement therapy and Breast Cancer Risk JAMA 283:485-491, 2000 and editorial page 534, January 26, 2000
- Ross RK, Paganini-Hill A, Wan PC, Pike MC. Effect of Hormone Replacement Therapy on Breast Cancer Risk: Estrogen versus Estrogen plus Progestin JNCI 92:328-332,2000)
- Magnusson C, Baron JA, Correia N, Bergstrom R, Adami HO, Persson I. Breast cancer risk following long-term oestrogen – and oestrogen-progestin replacement therapy Internation Journal of Cancer 81:339-44, 1999
- Colditz CA, Rosner B for the Nurses’ Health Study Research Group. Use of estrogen plus progestin is associated with greater increase in breast cancer risk than estrogen alone Am J Epidemiol 147:S45, 1998 ( abstract).
- Bergkvist L, Adami H-O, Persson I, Hoover R, Schairer C. The risk of breast cancer after estrogen and estrogen-progestin replacement N Eng J Med 321:293-297, 1989
- Collaborative group on hormone factors in breast cancer. Breast cancer and hormone replacement therapy; collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer Lancet 350:1047-59, 1997
- Santen RJ, Petroni GR. Relative versus attributable risk of breast cancer from estrogen replacement therapy. Journal of Clinical Endocrinol and Metab 84:1875-1881, 1999
- Edelman DA. Diethylstilbestrol exposure and the risk of clear cell cervical and vaginal adenocarcinoma. Int J.Fertil 34:251-255, 1989
- Pinkerton, J.V., Santen, R.J. Alternatives to the use of estrogen in post-menopausal women. Endocrine Reviews, 20:308-320, 1999
|
