A major concern for those who prescribe postmenopausal estrogen therapy and for women who take this treatment is the potential increased risk of breast cancer. The biologic basis for estrogen's promotive role in breast cancer is well documented: estrogen promotes breast cell mitosis,1 women exposed to more years of endogenous estrogen due to early menarche or late menopause have higher risks of breast cancer,2,3 and higher estradiol serum levels have been linked to a higher risk of breast cancer.4 Use of postmenopausal estrogen therapy has been intensely scrutinized for an associated risk of breast cancer, but recent meta-analyses have found no overall increase in risk with short term use.5-7 Most epidemiologic studies of postmenopausal estrogen use have had limited statistical power to determine the risk of long-term use. The average duration of postmenopausal estrogen use in case-control and cohort studies is less than 10 years.8 Three meta-analyses of studies that have data on long-term estrogen use7,9,10 have concluded that breast cancer risk increased with duration of therapy, and after 10 years of use, the estimated increase was 25-30%.
Because long-term treatment with estrogen is required to provide skeletal and cardiovascular benefits, women are presently being advised to continue postmenopausal estrogen treatment for at least 15 years, and possibly for life, i.e., 30 or more years. Because the effects of such long-term use on breast cancer risk are unknown, we have undertaken a study of breast cancer incidence in a cohort of women previously selected for long-term estrogen use, an average of 17 years. We compared the incidence and clinical presentation of breast cancer in these estrogen-using women to a cohort of similar women who did not use long-term estrogen.
In 1980, from a review of computerized pharmacy records at Kaiser Permanente Medical Center, San Francisco, for the years 1968-1971, two cohorts were identifed; one, composed of women who had used postmenopausal estrogen for at least 5 years, and the other, age-matched women who had not used estrogen for as long as 1 year.11,12 We found 1,110 women, born between 1900 and 1915, who had filled at least two prescriptions for an oral estrogen preparation. Included in the estrogen cohort were those subjects who satisfied the following criteria: the date of menopause was documented by either bilateral oophorectomy or spontaneous cessation of menses; and estrogen replacement therapy, equivalent to at least 0.3 mg conjugated estrogens, was begun within 3 years of menopause and taken for at least 5 years. Because our original purpose was to study osteoporotic fracture, we excluded those subjects who used thyroid preparations in doses exceeding 2 grains daily; we also excluded those who used anticonvulsants or glucocorticoids, or had chronic alcoholism, chronic renal or hepatic disease, hyperparathyroidism or hypoparathyroidism, insulin-dependent diabetes mellitus, hyperthyroidism, or other conditions known to adversely affect skeletal integrity (immobilization, malnutrition, or severe debilitating chronic disease of any sort). Black women were excluded because they were not considered prone to osteoporotic fractures. Two hundred forty-five estrogen users satisfied the criteria for inclusion. Nonusers were women matched for age and length of membership in the health plan who were found from the same computerized pharmacy records to have filled a prescription for a medication other than oral estrogen. They also satisfied all inclusion and exclusion criteria, except that none used estrogen for as long as 1 year. We excluded from further analyses women who, before the index pharmacy visit, had been diagnosed with any cancer except squamous skin neoplasm. The remaining 232 estrogen-using women and 222 nonusers are the subject of this analysis.
Information concerning medical history was obtained from both medical charts and computer data bases. Follow-up began with a visit to a Kaiser Foundation Health Plan pharmacy in 1968-1971 and continued until the end of 1995 or until subjects were no longer health plan members. Exposure to estrogen was determined by medical record review. Date of breast cancer diagnosis, mortality date, and cause of death were obtained from the California Tumor Registry, California death certificates, or computer data bases and were verified in all cases from medical records. Only invasive breast cancer was analyzed; one case of carcinoma in situ was excluded (in an estrogen user). Potential breast cancer risk factors considered were gynecologic and reproductive history, education, alcohol use, marital status, body mass index, smoking status, and family history of breast cancer. We categorized these variables, present at the start of follow-up, as follows: age at menarche <12 years (yes/no); age at menopause >53 years (yes/no); nulliparity (yes/no); schooling >12 years (yes/no); alcohol use none, 1-2 per day, or ≥3 drinks per day; obesity, body mass index >27 (yes/no); current smoker (yes/no); and routine breast cancer surveillance, both clinical and radiologic examinations.
Breast examination data were used to determine two time-dependent variables—whether a routine mammogram (or routine physical examination of the breast) was done in the prior 3 years or not. Using a Cox proportional hazards model, we determined the hazard ratios and 95% confidence intervals for estrogen use alone and in combination with other variables related to breast cancer risk. Using the product-limit method, we estimated the crude distribution of time of breast cancer diagnosis by estrogen use status.
A total of 454 women were studied. Our previous publication provides detailed information on demographic variables.12 The 232 women using estrogen were followed an average of 31.1 years after menopause, and they had taken estrogen for 17.2 of those years. The 222 nonusers were followed the same length of time after menopause as estrogen users. At the end of 1995, 72% of estrogen users and 57% of controls were still active health plan members. By end of follow-up, 9.6% of estrogen users were still taking estrogen. During the years each individual used estrogen, the mean dose exceeded 1.25 mg in 3%, was at least 0.625 mg in 81%, and was less than 0.45 mg in 6%. Progestin use was limited to 5.6% of estrogen users.
Table 1 shows prevalence of breast cancer risk factors among estrogen users and nonusers. Compared to nonusers, women using estrogen were on average 1.5 years older at menopause, were more likely to be nulliparous, were more likely to consume alcoholic beverages, but were similar in other breast cancer risk variables. In any year of follow-up between 1965 and 1995, the likelihood of a woman in this cohort having a routine clinical breast examination was about 50%. However, the mean frequency of routine breast examination was about 50% greater in estrogen users compared to nonusers (0.92 ± 0.47 vs. 0.61 ± 0.39 examinations per year, respectively; p < 0.001). During the same years of follow-up, the frequency of routine radiographic breast examination varied considerably in response to promotion of mammography screening. In each of the years 1965 through 1975, when annual multiphasic examinations (including mammography) were promoted for health plan members, on average about 20% of the cohort underwent mammography. In each of the years 1976 through 1982, when multiphasic examinations were not promoted, this percentage fell to 5%. After 1981, when mammography per se was promoted, the percentage of the cohort undergoing mammography rose progressively and peaked at about 30% in 1987—there-after remaining stable. During the entire follow-up period, the mean frequency of mammography was about 50% higher among estrogen users than nonusers (0.22 ± 0.16 vs. 0.15 ± 0.13 mammograms per year, respectively; p < 0.001).
Figure 1 shows the probability of breast cancer diagnosis during follow-up. The difference in the probabilities of cancer developing among estrogen users versus nonusers became apparent after 5-10 years and increased with time. The average age at breast cancer diagnosis was 69 years for both estrogen users and nonusers. The relative risk (RR) and 95% confidence intervals (95% CI) for estrogen use were 2.8 (95% CI 1.3-5.9). Addition of breast cancer covariates to a multivariate proportional hazards model reduced the RR for estrogen use by 29% and the resulting 95% confidence interval included 1.0 (RR 2.0, 95% CI 0.9-4.5). In this multivariate model, the only variable that was statistically significant was mammography within the previous 3 years, RR 5.4 (95% CI 2.3-12.8).
Survival rates for breast cancer that developed among estrogen users and nonusers were similar; 17 of 26 users (65%) and 5 of 9 nonusers (55%) were still alive 5 years after diagnosis. Additionally, the proportions of affected women dying of breast cancer were similar in estrogen users and nonusers, 7 of 26 (27%) and 3 of 9 (33%), respectively.
Evidence from this study suggests a strong association between long-term estrogen therapy and breast cancer in postmenopausal women. Our finding of a twofold risk after multiple adjustments is consistent with others who have examined long-term use; the case-control study of Hoover and coworkers13 showed a relative risk of 2.0 when estrogen was taken for >15 years; the case-control study of Ewertz and coworkers14 showed a relative risk of 2.3 for >12 years of estrogen use; and the cohort study of Bergkvist and coworkers15 found a relative risk of 1.7 after ≥9 years of use. Using data from four large follow-up studies, Steinberg and coworkers10 calculated an RR of 1.45-1.49 for 10 years of estrogen use. If the slope of time versus risk of cancer calculated in this meta-analysis is extrapolated to 17 years of estrogen exposure, a relative risk of 1.77-1.83 would result. Others dispute a link between estrogen use and breast cancer risk, stating either that there is an inconclusive relationship between estrogen and breast cancer,16 or that the risk is not as high as reported above.17-20
Our study is limited in that clinical information was obtained largely from medical records. Many of these data were self-reported on multiphasic examinations, increasing the potential for error, especially in reporting age of menarche and other recalled variables. We lacked complete information on age at first birth and on family history of breast cancer, but both are unlikely to greatly affect risk of breast cancer among elderly women.21 The estrogen these women took is somewhat different from the estrogen most women in the U.S. take today. The dose was larger, and was not usually combined with progesterone. Larger doses of estrogen have generally been associated with an increased risk for breast cancer. Progestin's effect is still disputed: it has been hypothesized that progestins could protect the breast from adverse effects of estrogen,22 but the largest study to date showed that progestins did not alter the increased risk observed among estrogen users.23
A limitation of many previous reports has been the reliance on case-control designs of short-term estrogen use. Our cohort study includes women who have received postmenopausal estrogen therapy for at least 5 years and in most cases considerably longer. The importance of long-term follow-up is demonstrated by our finding the divergence of breast cancer incidence only after 5-10 years of follow-up. This follow-up of elderly women is fairly inclusive, because it tracks the women to near the end of their life expectancy, when breast cancer risk is highest. A strength of our study is that we were able to gather information on many covariates and adjust for them.
A higher prevalence of health-seeking behavior24 and breast examinations may cause detection bias for breast cancer, increasing the apparent incidence and reducing case mortality. Our finding that estrogen users are more likely to have mammographic surveillance confirms results of the Nurses’ Health Study.25 We found that mammographic surveillance was strongly and independently associated with breast cancer diagnosis, whereas they found a smaller difference in surveillance and less impact on risk. Despite the increased surveillance favored upon estrogen users in our study, they did not appear to show major differences in mortality from the disease. The small number of breast cancers in our study preclude an analysis of the association between estrogen use and fatal breast cancer.
In conclusion, our study suggests that long-term postmenopausal estrogen therapy increases the risk of breast cancer, especially after 5-10 years of use. This must be considered in light of the many benefits conferred upon postmenopausal women by estrogen, including our recent report that estrogen reduces all-cause mortality.12 Even a small time-related increase in risk is made relevant by the fact that one in nine women will ultimately develop breast cancer in her lifetime26 and that 32% of women aged 50-65 use estrogen.27 Furthermore, until quite late in life; women have about the same death rate from breast cancer as from heart disease.28 We suggest that a woman's breast cancer risks be seriously considered when she is prescribed long-term estrogen.
This research was supported by the National Cancer Institute, National Institutes of Health, grant R35CA49761.
1. Pike MC, Spicer DV, Dahmoush L, Press M. Estrogens
, progesterone, normal breast cell proliferation, and breast cancer
risk. Epidemiol Rev
2. Hsieh C-C, Trichopoulos D, Katsouyanni K, et al. Age at menarche, age at menopause, height and obesity as risk factors for breast cancer
: associations and interactions in an international case-control study. Int J Cancer
3. Ewertz M, Duffy SW. Risk of breast cancer
in relation to reproductive factors in Denmark. Br J Cancer
4. Toniolo PG, Levitz M, Zeleniuch-Jacquotte A, et al. A prospective study of endogenous estrogens
and breast cancer
in postmenopausal women. J Natl Cancer Inst
5. Steinberg KK, Thacker SB, Smith J, et al. A meta-analysis of the effect of estrogen replacement therapy on the risk of breast cancer
6. Dupont WD, Page DL. Menopausal estrogen replacement therapy and breast cancer
. Arch Intern Med
7. Sillero-Arenas M, Delgado-Rodriguez M, Rodigues-Canteras R, Bueno-Cavanillas A, Galvez-Vargas R. Menopausal hormone replacement therapy and breast cancer
: a meta-analysis. Obstet Gynecol
8. Grady D, Rubin SR, Petitti DB, et al. Hormone therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med
9. Grady D, Ernster VL. Invited commentary: does postmenopausal hormone therapy cause breast cancer
? Am J Epidemiol
10. Steinberg KK, Smith SJ, Thacker SB, Strough DF. Breast cancer
risk and duration of estrogen use: the role of study design in meta-analysis. Epidemiology
11. Ettinger B, Genant HK, Cann CE. Long-term estrogen replacement therapy prevents bone loss and fractures. Ann Intern Med
12. Ettinger B, Friedman G, Bush T, Quesenberry C. Reduced mortality associated with long-term postmenopausal estrogen therapy. Obstet Gynecol
13. Hoover R, Gray LA Sr, Cole P, et al. Menopausal estrogens
and breast cancer
. N Engl J Med
14. Ewertz M. Influence of non-contraceptive exogenous and endogenous sex hormones on breast cancer
risk in Denmark. Int J Cancer
15. Bergkvist L, Adami H-L, Persson I, Hoover R, Schairer C. The risk of breast cancer
after estrogen and estrogen-progestin replacement. N Engl J Med
16. Stanford JL, Weiss NS, Voigt LF, et al. Combined estrogen and progestin hormone replacement therapy in relation to risk of breast cancer
in middle-aged women. JAMA
17. Wingo PA, Layde PM, Lee NC, et al. The risk of breast cancer
in postmenopausal women who have used estrogen replacement therapy. JAMA
18. Palmer JR, Rosenberg L, Clarke EA, et al. Breast cancer
risk after estrogen replacement therapy: results from the Toronto breast cancer
study. Am J Epidemiol
19. Kaufman DW, Palmer JR, de Mouzon J, et al. Estrogen replacement therapy and the risk of breast cancer
: results from the case-control surveillance study. Am J Epidemiol
20. Brinton LA, Hoover RN, Fraumeni JF Jr. Menopausal oestrogens and breast cancer
risk: an expanded case-control study. Br J Cancer
21. Bain C, Speizer FE, Rosner B, Belanger C, Hennekens CH. Family history of breast cancer
as a risk factor for the disease. Am J Epidemiol
22. Wren BG, Eden JA. Do progestins reduce the risk of breast cancer
? A review of the evidence. Menopause
23. Colditz GA, Hankinson SE, Hunter DJ, et al. The use of estrogens
and progestins and the risk of breast cancer
in postmenopausal women. N Engl J Med
24. Barrett-Connor E. Postmenopausal estrogen and prevention bias. Ann Intern Med
25. Colditz GA, Stampfer MJ, Willet WC, et al. Prospective study of estrogen replacement therapy and risk of breast cancer
in postmenopausal women. JAMA
26. American Cancer Society. Cancer facts and figures—1991. Atlanta, GA: American Cancer Society; 1992.
27. Utian WH, Schiff I. NAMS-Gallup survey on women's knowledge, information sources, and attitudes to menopause and hormone replacement therapy. Menopause
28. Grodstein F, Stampfer MJ, Colditz GA, et al. Postmenopausal hormone therapy and mortality. N Engl J Med