American Journal of Epidemiology Copyright © 1991 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved
Vol. 134, No 12 Printed in U S.A.
ORIGINAL CONTRIBUTIONS Estrogen Replacement Therapy and the Risk of Breast Cancer: Results from the Case-Control Surveillance Study
David W. Kaufman,1 Julie R. Palmer,' Jacques de Mouzon,2 Lynn Rosenberg,1 Paul D. Stolley,3 M. Ellen Warshauer,4 Ann G. Zauber,5 and Samuel Shapiro1
To examine the relation of noncontraceptive estrogen use to the risk of breast cancer among postmenopausal women, the authors conducted a case-control study: 1,686 cases were compared with 2,077 hospital control subjects, of whom 1,120 had nongynecologic cancers and 957 had nonmalignant (also non-gynecologic) conditions. Data were obtained from 1980 to 1986, by interview of subjects in hospitals in the United States and Canada. The relative risk estimate for any use of replacement estrogens unopposed by progestogens was 1.2 (95% confidence interval (Cl) 1.0-1.4), after adjustment for age and type of menopause; when all known risk factors for breast cancer were taken into account in a multivariate analysis, the estimate was similar. For use of at least 15 years duration, the estimate was 0.9 (95% Cl 0.5-1.9). Most of the unopposed use was of conjugated estrogens: overall, the relative risk (95% Cl) was 1.3 (1.0-1.6); for durations of 15 or more years, it was 0.9 (0.4-1.9); for use of 5 years followed by a latent interval of 15 or more years, it was 1.3 (0.7-2.4); and for current use it was 1.1 (0.7-1.6). There was no evidence of increased breast cancer risk when the conjugated estrogen users were divided according to dose. There was little use of estrogens opposed by progestogens; the relative risk estimate was 1.7 (95% Cl 0.93.3). The results of this large study provide no evidence that the use of unopposed conjugated estrogens increases the risk of breast cancer, even after long durations of use or long latent intervals, but the possibility of a modest increase (less than a doubling) could not be excluded. There were insufficient data to evaluate the effects of nonconjugated estrogens and of combined estrogen and progestogen therapy. Am J Epidemiol 1991;134:1375-85. breast neoplasms; estrogen replacement therapy
Editor's note: For a discussion of this paper and the paper immediately following by Palmer et ai, see the Invited Commentary on p. 1396 and the reply by the authors on p. 1401.
Since 1976, the relation between estrogen replacement therapy and the risk of breast cancer has been the subject of numerous epidemiologic studies (1-29). Most have shown no overall increase in risk, although
Received for publication January 16, 1991, and in final form May 20, 1991. ' Slone Epidemiology Unit, School of Public Health, Boston University, School of Medicine, Brookline, MA 2 INSERM U292, Hopital de Bicetre, Le Kremlin Bicetre Cedex, France. 3 Department of Medicine, Clinical Epidemiology Unit, The University of Pennsylvania School of Medicine, Philadelphia, PA 4 Cancer Care and Research Program, New York Hos-
pital, Department of Public Health, Cornell Medical Center, New York, NY. 5 Department of Epidemiology and Biostatistics, Memonal Sloan-Kettering Cancer Center, New York, NY. Reprint requests to Dr. David W. Kaufman, Slone Epiderniology Unit, 1371 Beacon Street, Brookline, MA 02146. Supported by contract nos. 223-76-3016, 233-80-3001, and 226-82-0007 and cooperative agreement nos. UO1 FD01222 and FD-U-000082 from the Food and Drug Administration Additional support was provided by contract
1375
1376
Kaufman et al.
in three studies (15, 19, 26), there was a modest overall association (up to an 80 percent increase). There have also been inconsistent elevations in risk in various subgroups of women (1, 2, 5-7, 9, 11, 15, 17-19, 24, 26). Some studies have suggested that the risk is increased after long durations of use (9, 14, 21, 24). Two other studies have shown no elevation for 10 or more years of use (13, 18), and most studies have had insufficient data to evaluate long-term use or have yielded equivocal results. Two reports (21, 24) further suggested that the risk may be increased among women who had received estrogens opposed with progestogens, although two others (28, 29) suggested that it may be reduced. The most recently published study raised the possibility of a modest increase in risk (36 percent) among current estrogen users (27). In 1984, we published the results of a large case-control study in which there was no increase in the risk of breast cancer overall among users of conjugated or other noncontraceptive estrogens; there was also no increase in risk after long durations of conjugated estrogen use, or after use in the distant and grant nos NO1-CP-71029, NO1-CB-74099, and R37CA-45762 from the National Cancer Institute. The Slone Epidemiology Unit has also received support from Hoffman-LaRoche, Inc., Nutley, NJ; CIBA-GEIGY Corporation, Summit, NJ; Hoechst AG, Frankfurt, West Germany; McNeil Pharmaceutical, Spring House, PA; Merrell Dow Pharmaceuticals, Inc., Cincinnati, OH; Ortho Pharmaceuticals Corporation, Raritan, NJ; and Sterling Drug, Inc., New York, NY. The authors are indebted to the many physicians who allowed them to interview their patients, the nurseinterviewers who collected the data, Marguerite Angeloni for coordinating the study, and Leonard Gaetano for assistance with data management The following hospitals participated in this study Boston: Beth Israel Hospital, Mt Auburn Hospital, Newton-Wellesley Hospital, Sancta Maria Hospital, Tufts New England Medical Center, New York: Brookhaven Memorial Hospital, Lenox Hill Hospital, Memorial Sloan-Kettenng Cancer Center, New York Hospital, Philadelphia' American Oncologic Hospital, Crozier Chester Medical Center, Hahnemann University Hospital, Hospital of the Medical College of Pennsylvania, Hospital of the University of Pennsylvania, Lankenau Hospital, Montgomery Hospital, Pennsylvania Hospital, Presbyterian Hospital, Thomas Jefferson University Hospital; Baltimore: The Johns Hopkins Hospital; Tucson- University of Arizona Health Science Center; Kansas City. Kansas University Medical Center; San Francisco: University of California, San Francisco Moffitt Hospital; London, Ontario: St. Joseph's Hospital.
past, or among subgroups of users (13). We report here the results of a new study of postmenopausal women, conducted with the same methods as the previous investigation. MATERIALS AND METHODS
The present data were collected in our Case-Control Surveillance Study, in progress since 1976 (30). Specially trained nurses stationed in hospitals in major metropolitan areas, primarily in the eastern United States, interview patients with various cancers and a wide variety of other conditions. Information is obtained on demographic factors (e.g., age, years of education), habits (e.g., cigarette smoking, alcohol consumption), relevant medical history (e.g., family history of cancer, history of major illnesses), and medication use. The medication history is obtained by asking about indications for use. For each drug, details of timing, duration, and frequency of use are obtained for each episode. In addition, for conjugated estrogens, the dose is recorded; photographs of conjugated estrogen pills are used to assist recall. Details of the patients' diagnoses are obtained from the hospital record after discharge. Our previous study of estrogens and breast cancer (13) was also derived from CaseControl Surveillance; the present analysis is confined to subjects enrolled subsequent to that investigation (some had already been interviewed, but we had not yet obtained their discharge information). All cases and controls were interviewed from January 1980 through September 1988 in hospitals in Boston, New York City, Philadelphia, Baltimore, Tucson, San Francisco, and London, Ontario. Of the patients approached for interview, 4 percent refused to participate. Cases
Postmenopausal women aged 40-69 years with breast cancer diagnosed not more than 6 months before the present hospital admission and with no concurrent or previous
Estrogens and Breast Cancer
cancer were eligible as cases. There were 1,686 cases; the median age was 59 years, and 86 percent were white. Controls
Postmenopausal women aged 40-69 years with malignant and nonmalignant conditions judged to be unrelated to estrogen use were eligible as controls. There were 1,120 cancer controls, including patients with colorectal cancer (n = 550), malignant melanoma (n = 200), and various other cancers {n = 370) (e.g., Hodgkin's disease, stomach, kidney, and leukemia) diagnosed within 6 months of admission, and no other concurrent or previous cancer. Women with cancers of the female genital tract were not eligible. There were 957 noncancer controls, including patients admitted for trauma (n = 422), acute infections (n = 124), disk disorders (n = 108), and various other conditions (H = 303) (e.g., bowel obstruction and pyelonephritis), and no history of cancer. Women with nonmalignant gynecologic conditions were not eligible. Thus, the total control series comprised 2,077 women; the median age was 59 years, and 83 percent were white. Classification of noncontraceptive estrogen use
Estrogen use initiated shortly before admission could have been after the onset of breast cancer or the control conditions, and hence not have been etiologically relevant. Therefore, the main focus of the analysis was on use that began at least 18 months before admission. Cases and controls in the latter category were subdivided according to whether they took unopposed estrogens (no concurrent use of progestogens) as replacement therapy, unopposed estrogens for other indications (regulation of menstrual periods, pregnancy maintenance, suppression of lactation, and breast conditions), or opposed estrogens (concurrent progestogen use of at least one month's duration). Vaginal estrogen preparations were not considered. Those who took unopposed estrogens for replace-
1377
ment, by far the largest group, were further subdivided into the following categories: those who took conjugated estrogens only, those who took nonconjugated estrogens only, those who took unspecified estrogens only, and those who took more than one type of estrogen. Table 1 shows estrogen use among the controls according to diagnostic category, with rates standardized to the overall age distribution of the control series. The rates were reasonably similar among the various categories, and the rates among cancer controls and noncancer controls as combined groups were closely similar. Therefore, the cancer and noncancer controls were combined into a single group for all analyses. Data analysis
Relative risks and 95 percent confidence intervals were estimated for various categories of estrogen use, with never use of noncontraceptive estrogens as the reference category. To control confounding by age and type of menopause (natural, due to hysterectomy with retention of at least one ovary, due to bilateral oophorectomy), relative risks were estimated from stratified data by the Mantel-Haenszel procedure (31). Confidence limits for these estimates were calculated by the method of Miettinen (32). To simultaneously take into account confounding by numerous factors, multiple logistic regression was used (33). In addition to categories of estrogen use, the factors included in the multivariate models were age (decade), age at menarche (15 years), age at first pregnancy (30 years), parity (0, 1-4, >5), age at menopause (50 years), type of menopause (natural, due to hysterectomy with retention of at least one ovary, due to bilateral oophorectomy), family history of breast cancer (yes, no), history of cystic breast disease (yes, no), race (white, other), religion (Jewish, other), years of education (12), body mass index (wt(kg)/ht(m)2) (29), oral contraceptive use (never, 5 years), alcohol consumption (1 drink/month - 1 drink/day, >1 day), year
1378
Kaufman et al.
TABLE 1. Estrogen use* among 2,077 hospital controls according to diagnosis in a study conducted in seven US and Canadian cities, 1980-1988 Cancer controls
Estrogen use
Never used estrogens Unopposed, for replacement Unopposed, for other indications Opposed by progestogens Any use that began s18 months before interview
Noncancer controls
Colon/ rectum (n = 550)
Malignant melanoma (n = 200)
Other cancer (n = 370)
All cancer controls (n = 1,120)
Trauma (n = 422)
Other (n = 535)
All noncancer controls (n = 957)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
424
(78)
153
(77)
292
(79)
869
(78)
345
(82)
406
(77)
751
(79)
9
(16)
35
(17)
61
(16)
191
(17)
61
(15)
99
(18)
160
(17)
19
(3)
6
(3)
11
(3)
36
(3)
10
(2)
13
(3)
23
(2)
2
(0)
3
(1)
2
(1)
7
(1)
3
(1)
6
(1)
9
(1)
10
(2)
3
(1)
4
(1)
17
(2)
3
(1)
11
(1)
14
(1)
* Rates of use are standardized to the overall age distribution of the control series
of interview (1980-1983, 1983-1984, 19851988), and geographic region (New York, Boston, Philadelphia, other). RESULTS
Eighteen percent of the 1,686 cases and 17 percent of the 2,077 controls used unopposed estrogens for replacement (table 2); the overall stratified relative risk estimate was 1.2 (95 percent confidence interval (CI) 1.0-1.4), and for use of at least 15 years, it was 0.9 (0.5-1.9; based on 16 cases and 25 controls). The median durations of use in cases and controls were 22 and 24 months, respectively. Three percent of both cases and controls had used unopposed estrogens for indications other than replacement, giving a stratified relative risk estimate of 1.0 (0.71.4). One percent of each group used opposed estrogens, with an overall stratified estimate of 1.7 (0.9-3.3), and an estimate of 1.5 (0.7-3.3) for use of at least one year (16 cases, 12 controls). In all instances, the stratified and multivariate estimates were closely similar. Among the women who took estrogens for replacement, the relative risk estimates for users of conjugated estrogens, other named estrogens, unspecified estrogens, and more than one type were all close to 1.0
(table 2). There were insufficient data to estimate relative risks for specific nonconjugated estrogens. Because conjugated estrogens accounted for over 90 percent of named compounds, users of unspecified estrogens were considered to be conjugated estrogen users in all further comparisons presented. When the analysis was confined to identified conjugated estrogens, the results were similar. The use of conjugated estrogens for replacement is shown in table 3 according to duration of use. All relative risk estimates were close to 1.0. For use that lasted 15 or more years, the stratified estimate was 0.9 (0.4-1.9). Again, the stratified and multivariate estimates were closely similar. The use of conjugated estrogens for replacement is shown in table 4 according to time elapsed since last use. There was no pattern of increased risk, and most relative risk estimates were close to 1.0. In particular, the stratified estimate for use ending within 12 months of admission (current use) was 1.1 (0.7-1.6); the corresponding estimate for use of at least 5 years' duration (based on 37 exposed cases and 44 exposed controls) was 1.3 (0.8-2.1). The only statistically significant result was a multivariate estimate of 0.5 (0.2-0.9) for use that ended 36-59 months before admission; the stratified es-
Estrogens and Breast Cancer
1379
TABLE 2. Estrogen use among 1,686 breast cancer cases and 2,077 hospital controls in a study conducted in seven US and Canadian cities, 1980-1988
Estrogen use
Never used estrogens Unopposed, for replacement Conjugated estrogens only Other estrogens onlyj Unknown type of estrogen only More than one type of estrogen Unopposed, for other indications§ Opposed by progestogens Ever use that first began 3 Oral contraceptive use Yes No Age (years) at menopause 50 Type of menopause Natural Due to hysterectomy Due to bilateral oophorectomy History of cystic breast disease Yes No Family history of breast cancer Yes No Body mass index (weight/height2) 29
Cases (no.)
Controls (no.)
Use>5 years/never use
Use>5 years/never use
Stratified relative risk estimate! (95% confidence interval)
2/110 31/557 48/613
13/180 37/586 49/809
0.4$ (0.1-2.0) 1.1$ (0.6-1.8) 1.3$ (0.8-1.9)
70/1,152 11/141
85/1,381 17/239
1.1 (0.8-1.6) 1.2 (0.5-2.8)
28/568 30/504
55/800 29/542
1.5 (0.8-3.1) 0.9 (0.5-1.5)
24/250 32/579 25/457
21/281 43/612 37/711
1.3 (0.7-2.6) 0.9 (0.5-1.4) 1.3 (0.7-2.2)
9/161 72/1,132
11/208 91/1,412
1.1 (0.4-3.1) 1.1 (0.8-1.6)
14/191 36/469 28/588
37/329 41/610 24/599
0.8 (0.4-1.5) 1.1 (0.7-1.8) 1.2 (0.7-2.2)
28/957 12/221
26/1,123 27/317
1.3§ (0.8-2.2) 0.6§ (0.3-1.2)
41/115
49/179
1.3§ (0.8-2.2)
20/243 61/1,050
10/140 92/1,480
1 2 (0.5-2.8) 1.1 (0.7-1.5)
18/195 63/1,098
9/113 93/1,507
1.7 (0.6-4.5) 1.0 (0.7-1.4)
19/154 52/796 9/324
20/314 64/933 18/325
2.2 (1.1-4.6) 1.1 (0.7-1.6) 0.6 (0.2-1.3)
* First used at least 18 months before admission, not opposed by progestogens; includes unknown estrogens t Relative to never use in each category; adjusted for age and type of menopause by the Mantel-Haenszel procedure (31). t Adjusted for type of menopause. § Adjusted for age.
Estrogens and Breast Cancer
remains possible that the results were confounded by factors not taken into account. The general absence of association in the present study is similar to the findings of our previous investigation (13). The present results are also consistent with those from a case-control study carried out in Toronto, Canada (presented in the accompanying report) (37), with the exception that a significant relative risk estimate of 2.5 was found in that study for 15 or more years duration of use among women aged 60-69 years. The corresponding estimate from the present study was 1.2. The Toronto study was methodologically similar in that the interview was virtually identical to that used in the CaseControl Surveillance Study. It was based on a different population, however, and used neighbor controls. Further data will be needed before it can be judged whether the positive association in older women in that study is real or is due to chance. There have been many previous studies (1-29), of both the case-control and followup type, on the relation of exogenous estrogen use to the risk of breast cancer, and it is beyond the scope of this paper to comment on them in detail. Most have shown no overall effect of estrogen use. In three studies (15, 19, 26), there were modest but statistically significant increases in risk that ranged from 60 percent to 80 percent. There have also been increases in risk reported in various subgroups of women classified according to various risk factors for breast cancer (1-7, 9, 11, 15, 17-19, 24, 26). The elevated relative risk estimates have generally been based on small numbers, and there has not been any consistent pattern of association across studies. Of greater concern, data from four studies (9, 14, 21, 24) have suggested that the risk may be increased for long durations of estrogen use. Three of the studies were of the case-control type. One of these (9) was based on relatively small numbers of users. Another (14) was large, but the cases and controls were derived from a breast cancer screening program, and it is possible that women with symptoms of breast cancer who were also long-standing estrogen users may
1383
have been the most likely to participate in the program. The third (24) was a large population-based study from Denmark, in which a significant trend of increasing risk with increasing duration of noncontraceptive estrogen use was observed, with a relative risk estimate of 2.3 for use for at least 12 years. Virtually none of the use was of conjugated estrogens, which accounted for most of the estrogen use in the present study; thus, the results may not be directly comparable. Finally, in a follow-up study conducted in Sweden (21), there was a 70 percent increase in risk among women who had used estrogens, primarily estradiol compounds, for at least 109 months. Estradiols are rarely used in the United States, and again the results may not be comparable with those of the present study, although it is by no means clear that estradiols would have a different effect on breast tissue than conjugated estrogens (38). Both of the Scandinavian studies had elevated relative risk estimates for estrogen therapy opposed by progestogens, but these were not statistically significant, and contrast with the results of a follow-up study (28) and a small clinical trial (29), which were consistent with a reduced risk for opposed therapy. Among other large case-control studies, there was not a statistically significant association with long-term use in the Toronto study (36) or in the Cancer and Steroid Hormone (CASH) Study (20), but the relative risk point estimates were somewhat elevated. The risk did not appear to be increased in the subgroups that have been previously reported. In neither of these studies was there much information concerning nonconjugated estrogens, or opposed therapy. Finally, two recent follow-up studies (26, 27) have shown modest elevations in risk for current users, and the investigators have proposed that their findings are compatible with a promotional effect. In the Adventist Health Study (26), after 7 years of followup, the relative risk estimate for breast cancer in relation to current use of estrogens at the time of enrollment was 1.69; there was no duration effect. In the Nurses' Health
1384
Kaufman et al.
Study (27), the relative risk estimate for the current use of conjugated estrogens (defined as exposure within the preceding 2 years) was 1.36; there was no trend in the risk according to duration of use or dose. In both studies, the selective diagnosis of breast cancer among current or recent estrogen users could readily have contributed to the associations. Estrogen users are advised to examine their breasts regularly, and to attend regular examinations (including screening mammographies). In the Nurses' Health Study (27), the proportions of tumors over 2 cm in size were the same in exposed and nonexposed cases, but this is not necessarily evidence against selection bias: it is only when the cancer becomes locally invasive or metastatic that the diagnosis becomes inevitable, and hence independent of selective detection by screening. In addition, the observations that the risk was elevated for durations of exposure of less than one year, and the lack of duration or dose effects, detract from the biologic plausibility of the findings. By contrast, there was no association with current use in the CASH Study (20), where the data were sufficient to exclude a risk increase of 30 percent for current users. Nor was there evidence of an increase in risk for current conjugated estrogen users in the present study, although the upper confidence limit was insufficient to exclude the recent findings. Based on the present results and those of other large epidemiologic studies, we conclude that unopposed conjugated estrogens taken for hormone replacement do not appreciably increase the risk of breast cancer. The possibility remains open that the risk is modestly increased after very long durations of use, and that opposed therapy, and other noncontraceptive estrogens, increase the risk. Further studies should focus on these latter questions.
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
15. 16. 17. 18.
19.
20.
REFERENCES 1. Casagrande J, Gerkins V, Henderson BE, et al. Brief communication: Exogenous estrogens and
21.
breast cancer in women with natural menopause. JNCI 1976;56:839-41. Hoover R, Gray LA, Cole P, et al. Menopausal estrogens and breast cancer. N Engl J Med 1976,295:401-5. Sartwell PE, Arthes FG, Tonascia JA. Exogenous hormones, reproductive history, and breast cancer. JNCI 1977;59:1589-92. Brinton LA, Williams RR, Hoover RN, et al. Breast cancer risk factors among screening program participants. JNCI 1979;62:37-44. Ross RK, Paganini-Hill A, Gerkins VR, et al. A case-control study of menopausal estrogen therapy and breast cancer. JAMA 1980;243:1635-9. Jick H, Walker AM, Watkins RN, et al. Replacement estrogens and breast cancer. Am J Epidemiol 1980;l 12:586-94. Brinton LA, Hoover RN, Szklo M, et al. Menopausal estrogen use and risk of breast cancer. Cancer 1981 ;47:2517-22. Kelsey JL, Fisher DB, Holford TR, et al. Exogenous estrogens and other factors in the epidemiology of breast cancer. JNCI 1981;67:327-33. Hoover R, Glass A, Finkle WD, et al. Conjugated estrogens and breast cancer risk in women. JNCI 1981;67:815-20. Lawson DH, Jick H, Hunter JR, et al. Exogenous estrogens and breast cancer. Am J Epidemiol 1981;114:710-13. Hulka BS, Chambless LE, Deubner DC, et al. Breast cancer and estrogen replacement therapy. Am J Obstet Gynecol 1982;143:638-44. Hiatt RA, Bawol R, Friedman GD, et al. Exogenous estrogen and breast cancer after bilateral oophorectomy. Cancer 1984,54:139-44. Kaufman DW, Miller DR, Rosenberg L, et al. Noncontraceptive estrogen use and the risk of breast cancer. JAMA 1984;252:63-7. Brinton LA, Hoover R, Fraumeni JF Jr. Menopausal oestrogens and breast cancer risk: an expanded case-control study. Br J Cancer 1986;54:825-32. La Vecchia C, Decarli A, Parazzini F, et al. Noncontraceptive oestrogens and the risk of breast cancer in women. Int J Cancer 1986,38:853-8. McDonald JA, Weiss NS, Daling JR, et al. Menopausal estrogen use and the risk of breast cancer. Breast Cancer Res Treat 1986;7:193-9. Nomura AMY, Kolonel LN, Hirohata T, et al. The association of replacement estrogens with breast cancer. Int J Cancer 1986;37:49-53. Buring JE, Hennekens CH, Lipnick RJ, et al. A prospective cohort study of postmenopausal hormone use and risk of breast cancer in US women. Am J Epidemiol 1987; 125:939-47. Hunt K, Vessey M, McPherson K, et al. Long-term surveillance of mortality and cancer incidence in women receiving hormone replacement therapy. Br J Obstet Gynecol 1987;94:620-35. Wingo PA, Layde PM. Lee NC, et al. The risk of breast cancer in postmenopausal women who have used estrogen replacement therapy. JAMA 1987;257:209-15. Bergkvist L, Adami H-0, Persson I, et al. The risk of breast cancer after estrogen and estrogenprogestin replacement. N Engl J Med 1989,321:293-7.
Estrogens and Breast Cancer
22. Thomas DB, Persing JP, Hutchinson WB. Exogenous estrogens and other risk factors for breast cancer in women with benign breast diseases. JNCI 1982;69:1017-23. 23. Rohan TE, McMichael AJ. Non-contraceptive exogenous oestrogen therapy and breast cancer. Med J Aust 1988;148:217-21. 24. Ewertz M. Influence of non-contraceptive exogenous and endogenous sex hormones on breast cancer risk in Denmark. Int J Cancer 1988;42:832-8. 25. Dupont WD, Page DL, Rogers LW, et al. Influence of exogenous estrogens, proliferative breast disease, and other variables on breast cancer risk. Cancer 1989;63:948-57. 26. Mills PK, Beeson WL, Phillips RL, et al. Prospective study of exogenous hormone use and breast cancer in Seventh-day Adventists. Cancer 1989,64:591-7. 27. Colditz GA, Stampfer MJ, Willett WC, et al. Prospective study of estrogen replacement therapy and risk of breast cancer in postmenopausal women. JAMA 1990,264:2648-53. 28. Gambrell RD Jr, Maier RC, Sanders BI. Decreased incidence of breast cancer in postmenopausal estrogen-progestogen users. Obstet Gynecol 1983,62:435-43. 29. Nachtigall LE, Nachtigall RH, Nachtigall RD. et al. Estrogen replacement therapy. II. A prospective etudy in the relationship to carcinoma and cardiovascular and metabolic problems. Obstet Gynecol 1979;54:74-9.
1385
30. Slone D, Shapiro S, Miettinen OS. Case-control surveillance of serious illnesses attributable to ambulatory drug use. In: Colombo F, Shapiro S, Slone D, et al, eds. Epidemiological evaluation of drugs. Amsterdam: Elsevier North-Holland Inc, 1977:5982. 31. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies. JNCI 1959;22:719-48. 32. Miettinen OS. Estimability and estimation in casereferent studies. Am J Epidemiol 1976; 103:22635. 33. Schlesselman JJ. Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982. 34. Key TJA, Pike MC. The role of estrogens and progestogens in the epidemiology and prevention of breast cancer. Eur J Cancer Clin Oncol 1988;24:29-43. 35. Kelsey JL. A review of the epidemiology of human breast cancer. Epidemiol Rev 1979:1:74-109. 36. Helmrich SP, Shapiro S, Rosenberg L, et al. Risk factors for breast cancer. Am J Epidemiol 1983;117:35-45. 37. 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 1991; 134:1386-95. 38. Stevenson JC, Whitehead MI. Breast cancer and estrogen replacement. N Engl J Med 1990;322:201-2.
Vol. 134, No 12 Printed in U S.A.
ORIGINAL CONTRIBUTIONS Estrogen Replacement Therapy and the Risk of Breast Cancer: Results from the Case-Control Surveillance Study
David W. Kaufman,1 Julie R. Palmer,' Jacques de Mouzon,2 Lynn Rosenberg,1 Paul D. Stolley,3 M. Ellen Warshauer,4 Ann G. Zauber,5 and Samuel Shapiro1
To examine the relation of noncontraceptive estrogen use to the risk of breast cancer among postmenopausal women, the authors conducted a case-control study: 1,686 cases were compared with 2,077 hospital control subjects, of whom 1,120 had nongynecologic cancers and 957 had nonmalignant (also non-gynecologic) conditions. Data were obtained from 1980 to 1986, by interview of subjects in hospitals in the United States and Canada. The relative risk estimate for any use of replacement estrogens unopposed by progestogens was 1.2 (95% confidence interval (Cl) 1.0-1.4), after adjustment for age and type of menopause; when all known risk factors for breast cancer were taken into account in a multivariate analysis, the estimate was similar. For use of at least 15 years duration, the estimate was 0.9 (95% Cl 0.5-1.9). Most of the unopposed use was of conjugated estrogens: overall, the relative risk (95% Cl) was 1.3 (1.0-1.6); for durations of 15 or more years, it was 0.9 (0.4-1.9); for use of 5 years followed by a latent interval of 15 or more years, it was 1.3 (0.7-2.4); and for current use it was 1.1 (0.7-1.6). There was no evidence of increased breast cancer risk when the conjugated estrogen users were divided according to dose. There was little use of estrogens opposed by progestogens; the relative risk estimate was 1.7 (95% Cl 0.93.3). The results of this large study provide no evidence that the use of unopposed conjugated estrogens increases the risk of breast cancer, even after long durations of use or long latent intervals, but the possibility of a modest increase (less than a doubling) could not be excluded. There were insufficient data to evaluate the effects of nonconjugated estrogens and of combined estrogen and progestogen therapy. Am J Epidemiol 1991;134:1375-85. breast neoplasms; estrogen replacement therapy
Editor's note: For a discussion of this paper and the paper immediately following by Palmer et ai, see the Invited Commentary on p. 1396 and the reply by the authors on p. 1401.
Since 1976, the relation between estrogen replacement therapy and the risk of breast cancer has been the subject of numerous epidemiologic studies (1-29). Most have shown no overall increase in risk, although
Received for publication January 16, 1991, and in final form May 20, 1991. ' Slone Epidemiology Unit, School of Public Health, Boston University, School of Medicine, Brookline, MA 2 INSERM U292, Hopital de Bicetre, Le Kremlin Bicetre Cedex, France. 3 Department of Medicine, Clinical Epidemiology Unit, The University of Pennsylvania School of Medicine, Philadelphia, PA 4 Cancer Care and Research Program, New York Hos-
pital, Department of Public Health, Cornell Medical Center, New York, NY. 5 Department of Epidemiology and Biostatistics, Memonal Sloan-Kettering Cancer Center, New York, NY. Reprint requests to Dr. David W. Kaufman, Slone Epiderniology Unit, 1371 Beacon Street, Brookline, MA 02146. Supported by contract nos. 223-76-3016, 233-80-3001, and 226-82-0007 and cooperative agreement nos. UO1 FD01222 and FD-U-000082 from the Food and Drug Administration Additional support was provided by contract
1375
1376
Kaufman et al.
in three studies (15, 19, 26), there was a modest overall association (up to an 80 percent increase). There have also been inconsistent elevations in risk in various subgroups of women (1, 2, 5-7, 9, 11, 15, 17-19, 24, 26). Some studies have suggested that the risk is increased after long durations of use (9, 14, 21, 24). Two other studies have shown no elevation for 10 or more years of use (13, 18), and most studies have had insufficient data to evaluate long-term use or have yielded equivocal results. Two reports (21, 24) further suggested that the risk may be increased among women who had received estrogens opposed with progestogens, although two others (28, 29) suggested that it may be reduced. The most recently published study raised the possibility of a modest increase in risk (36 percent) among current estrogen users (27). In 1984, we published the results of a large case-control study in which there was no increase in the risk of breast cancer overall among users of conjugated or other noncontraceptive estrogens; there was also no increase in risk after long durations of conjugated estrogen use, or after use in the distant and grant nos NO1-CP-71029, NO1-CB-74099, and R37CA-45762 from the National Cancer Institute. The Slone Epidemiology Unit has also received support from Hoffman-LaRoche, Inc., Nutley, NJ; CIBA-GEIGY Corporation, Summit, NJ; Hoechst AG, Frankfurt, West Germany; McNeil Pharmaceutical, Spring House, PA; Merrell Dow Pharmaceuticals, Inc., Cincinnati, OH; Ortho Pharmaceuticals Corporation, Raritan, NJ; and Sterling Drug, Inc., New York, NY. The authors are indebted to the many physicians who allowed them to interview their patients, the nurseinterviewers who collected the data, Marguerite Angeloni for coordinating the study, and Leonard Gaetano for assistance with data management The following hospitals participated in this study Boston: Beth Israel Hospital, Mt Auburn Hospital, Newton-Wellesley Hospital, Sancta Maria Hospital, Tufts New England Medical Center, New York: Brookhaven Memorial Hospital, Lenox Hill Hospital, Memorial Sloan-Kettenng Cancer Center, New York Hospital, Philadelphia' American Oncologic Hospital, Crozier Chester Medical Center, Hahnemann University Hospital, Hospital of the Medical College of Pennsylvania, Hospital of the University of Pennsylvania, Lankenau Hospital, Montgomery Hospital, Pennsylvania Hospital, Presbyterian Hospital, Thomas Jefferson University Hospital; Baltimore: The Johns Hopkins Hospital; Tucson- University of Arizona Health Science Center; Kansas City. Kansas University Medical Center; San Francisco: University of California, San Francisco Moffitt Hospital; London, Ontario: St. Joseph's Hospital.
past, or among subgroups of users (13). We report here the results of a new study of postmenopausal women, conducted with the same methods as the previous investigation. MATERIALS AND METHODS
The present data were collected in our Case-Control Surveillance Study, in progress since 1976 (30). Specially trained nurses stationed in hospitals in major metropolitan areas, primarily in the eastern United States, interview patients with various cancers and a wide variety of other conditions. Information is obtained on demographic factors (e.g., age, years of education), habits (e.g., cigarette smoking, alcohol consumption), relevant medical history (e.g., family history of cancer, history of major illnesses), and medication use. The medication history is obtained by asking about indications for use. For each drug, details of timing, duration, and frequency of use are obtained for each episode. In addition, for conjugated estrogens, the dose is recorded; photographs of conjugated estrogen pills are used to assist recall. Details of the patients' diagnoses are obtained from the hospital record after discharge. Our previous study of estrogens and breast cancer (13) was also derived from CaseControl Surveillance; the present analysis is confined to subjects enrolled subsequent to that investigation (some had already been interviewed, but we had not yet obtained their discharge information). All cases and controls were interviewed from January 1980 through September 1988 in hospitals in Boston, New York City, Philadelphia, Baltimore, Tucson, San Francisco, and London, Ontario. Of the patients approached for interview, 4 percent refused to participate. Cases
Postmenopausal women aged 40-69 years with breast cancer diagnosed not more than 6 months before the present hospital admission and with no concurrent or previous
Estrogens and Breast Cancer
cancer were eligible as cases. There were 1,686 cases; the median age was 59 years, and 86 percent were white. Controls
Postmenopausal women aged 40-69 years with malignant and nonmalignant conditions judged to be unrelated to estrogen use were eligible as controls. There were 1,120 cancer controls, including patients with colorectal cancer (n = 550), malignant melanoma (n = 200), and various other cancers {n = 370) (e.g., Hodgkin's disease, stomach, kidney, and leukemia) diagnosed within 6 months of admission, and no other concurrent or previous cancer. Women with cancers of the female genital tract were not eligible. There were 957 noncancer controls, including patients admitted for trauma (n = 422), acute infections (n = 124), disk disorders (n = 108), and various other conditions (H = 303) (e.g., bowel obstruction and pyelonephritis), and no history of cancer. Women with nonmalignant gynecologic conditions were not eligible. Thus, the total control series comprised 2,077 women; the median age was 59 years, and 83 percent were white. Classification of noncontraceptive estrogen use
Estrogen use initiated shortly before admission could have been after the onset of breast cancer or the control conditions, and hence not have been etiologically relevant. Therefore, the main focus of the analysis was on use that began at least 18 months before admission. Cases and controls in the latter category were subdivided according to whether they took unopposed estrogens (no concurrent use of progestogens) as replacement therapy, unopposed estrogens for other indications (regulation of menstrual periods, pregnancy maintenance, suppression of lactation, and breast conditions), or opposed estrogens (concurrent progestogen use of at least one month's duration). Vaginal estrogen preparations were not considered. Those who took unopposed estrogens for replace-
1377
ment, by far the largest group, were further subdivided into the following categories: those who took conjugated estrogens only, those who took nonconjugated estrogens only, those who took unspecified estrogens only, and those who took more than one type of estrogen. Table 1 shows estrogen use among the controls according to diagnostic category, with rates standardized to the overall age distribution of the control series. The rates were reasonably similar among the various categories, and the rates among cancer controls and noncancer controls as combined groups were closely similar. Therefore, the cancer and noncancer controls were combined into a single group for all analyses. Data analysis
Relative risks and 95 percent confidence intervals were estimated for various categories of estrogen use, with never use of noncontraceptive estrogens as the reference category. To control confounding by age and type of menopause (natural, due to hysterectomy with retention of at least one ovary, due to bilateral oophorectomy), relative risks were estimated from stratified data by the Mantel-Haenszel procedure (31). Confidence limits for these estimates were calculated by the method of Miettinen (32). To simultaneously take into account confounding by numerous factors, multiple logistic regression was used (33). In addition to categories of estrogen use, the factors included in the multivariate models were age (decade), age at menarche (15 years), age at first pregnancy (30 years), parity (0, 1-4, >5), age at menopause (50 years), type of menopause (natural, due to hysterectomy with retention of at least one ovary, due to bilateral oophorectomy), family history of breast cancer (yes, no), history of cystic breast disease (yes, no), race (white, other), religion (Jewish, other), years of education (12), body mass index (wt(kg)/ht(m)2) (29), oral contraceptive use (never, 5 years), alcohol consumption (1 drink/month - 1 drink/day, >1 day), year
1378
Kaufman et al.
TABLE 1. Estrogen use* among 2,077 hospital controls according to diagnosis in a study conducted in seven US and Canadian cities, 1980-1988 Cancer controls
Estrogen use
Never used estrogens Unopposed, for replacement Unopposed, for other indications Opposed by progestogens Any use that began s18 months before interview
Noncancer controls
Colon/ rectum (n = 550)
Malignant melanoma (n = 200)
Other cancer (n = 370)
All cancer controls (n = 1,120)
Trauma (n = 422)
Other (n = 535)
All noncancer controls (n = 957)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
No.
(%)
424
(78)
153
(77)
292
(79)
869
(78)
345
(82)
406
(77)
751
(79)
9
(16)
35
(17)
61
(16)
191
(17)
61
(15)
99
(18)
160
(17)
19
(3)
6
(3)
11
(3)
36
(3)
10
(2)
13
(3)
23
(2)
2
(0)
3
(1)
2
(1)
7
(1)
3
(1)
6
(1)
9
(1)
10
(2)
3
(1)
4
(1)
17
(2)
3
(1)
11
(1)
14
(1)
* Rates of use are standardized to the overall age distribution of the control series
of interview (1980-1983, 1983-1984, 19851988), and geographic region (New York, Boston, Philadelphia, other). RESULTS
Eighteen percent of the 1,686 cases and 17 percent of the 2,077 controls used unopposed estrogens for replacement (table 2); the overall stratified relative risk estimate was 1.2 (95 percent confidence interval (CI) 1.0-1.4), and for use of at least 15 years, it was 0.9 (0.5-1.9; based on 16 cases and 25 controls). The median durations of use in cases and controls were 22 and 24 months, respectively. Three percent of both cases and controls had used unopposed estrogens for indications other than replacement, giving a stratified relative risk estimate of 1.0 (0.71.4). One percent of each group used opposed estrogens, with an overall stratified estimate of 1.7 (0.9-3.3), and an estimate of 1.5 (0.7-3.3) for use of at least one year (16 cases, 12 controls). In all instances, the stratified and multivariate estimates were closely similar. Among the women who took estrogens for replacement, the relative risk estimates for users of conjugated estrogens, other named estrogens, unspecified estrogens, and more than one type were all close to 1.0
(table 2). There were insufficient data to estimate relative risks for specific nonconjugated estrogens. Because conjugated estrogens accounted for over 90 percent of named compounds, users of unspecified estrogens were considered to be conjugated estrogen users in all further comparisons presented. When the analysis was confined to identified conjugated estrogens, the results were similar. The use of conjugated estrogens for replacement is shown in table 3 according to duration of use. All relative risk estimates were close to 1.0. For use that lasted 15 or more years, the stratified estimate was 0.9 (0.4-1.9). Again, the stratified and multivariate estimates were closely similar. The use of conjugated estrogens for replacement is shown in table 4 according to time elapsed since last use. There was no pattern of increased risk, and most relative risk estimates were close to 1.0. In particular, the stratified estimate for use ending within 12 months of admission (current use) was 1.1 (0.7-1.6); the corresponding estimate for use of at least 5 years' duration (based on 37 exposed cases and 44 exposed controls) was 1.3 (0.8-2.1). The only statistically significant result was a multivariate estimate of 0.5 (0.2-0.9) for use that ended 36-59 months before admission; the stratified es-
Estrogens and Breast Cancer
1379
TABLE 2. Estrogen use among 1,686 breast cancer cases and 2,077 hospital controls in a study conducted in seven US and Canadian cities, 1980-1988
Estrogen use
Never used estrogens Unopposed, for replacement Conjugated estrogens only Other estrogens onlyj Unknown type of estrogen only More than one type of estrogen Unopposed, for other indications§ Opposed by progestogens Ever use that first began 3 Oral contraceptive use Yes No Age (years) at menopause 50 Type of menopause Natural Due to hysterectomy Due to bilateral oophorectomy History of cystic breast disease Yes No Family history of breast cancer Yes No Body mass index (weight/height2) 29
Cases (no.)
Controls (no.)
Use>5 years/never use
Use>5 years/never use
Stratified relative risk estimate! (95% confidence interval)
2/110 31/557 48/613
13/180 37/586 49/809
0.4$ (0.1-2.0) 1.1$ (0.6-1.8) 1.3$ (0.8-1.9)
70/1,152 11/141
85/1,381 17/239
1.1 (0.8-1.6) 1.2 (0.5-2.8)
28/568 30/504
55/800 29/542
1.5 (0.8-3.1) 0.9 (0.5-1.5)
24/250 32/579 25/457
21/281 43/612 37/711
1.3 (0.7-2.6) 0.9 (0.5-1.4) 1.3 (0.7-2.2)
9/161 72/1,132
11/208 91/1,412
1.1 (0.4-3.1) 1.1 (0.8-1.6)
14/191 36/469 28/588
37/329 41/610 24/599
0.8 (0.4-1.5) 1.1 (0.7-1.8) 1.2 (0.7-2.2)
28/957 12/221
26/1,123 27/317
1.3§ (0.8-2.2) 0.6§ (0.3-1.2)
41/115
49/179
1.3§ (0.8-2.2)
20/243 61/1,050
10/140 92/1,480
1 2 (0.5-2.8) 1.1 (0.7-1.5)
18/195 63/1,098
9/113 93/1,507
1.7 (0.6-4.5) 1.0 (0.7-1.4)
19/154 52/796 9/324
20/314 64/933 18/325
2.2 (1.1-4.6) 1.1 (0.7-1.6) 0.6 (0.2-1.3)
* First used at least 18 months before admission, not opposed by progestogens; includes unknown estrogens t Relative to never use in each category; adjusted for age and type of menopause by the Mantel-Haenszel procedure (31). t Adjusted for type of menopause. § Adjusted for age.
Estrogens and Breast Cancer
remains possible that the results were confounded by factors not taken into account. The general absence of association in the present study is similar to the findings of our previous investigation (13). The present results are also consistent with those from a case-control study carried out in Toronto, Canada (presented in the accompanying report) (37), with the exception that a significant relative risk estimate of 2.5 was found in that study for 15 or more years duration of use among women aged 60-69 years. The corresponding estimate from the present study was 1.2. The Toronto study was methodologically similar in that the interview was virtually identical to that used in the CaseControl Surveillance Study. It was based on a different population, however, and used neighbor controls. Further data will be needed before it can be judged whether the positive association in older women in that study is real or is due to chance. There have been many previous studies (1-29), of both the case-control and followup type, on the relation of exogenous estrogen use to the risk of breast cancer, and it is beyond the scope of this paper to comment on them in detail. Most have shown no overall effect of estrogen use. In three studies (15, 19, 26), there were modest but statistically significant increases in risk that ranged from 60 percent to 80 percent. There have also been increases in risk reported in various subgroups of women classified according to various risk factors for breast cancer (1-7, 9, 11, 15, 17-19, 24, 26). The elevated relative risk estimates have generally been based on small numbers, and there has not been any consistent pattern of association across studies. Of greater concern, data from four studies (9, 14, 21, 24) have suggested that the risk may be increased for long durations of estrogen use. Three of the studies were of the case-control type. One of these (9) was based on relatively small numbers of users. Another (14) was large, but the cases and controls were derived from a breast cancer screening program, and it is possible that women with symptoms of breast cancer who were also long-standing estrogen users may
1383
have been the most likely to participate in the program. The third (24) was a large population-based study from Denmark, in which a significant trend of increasing risk with increasing duration of noncontraceptive estrogen use was observed, with a relative risk estimate of 2.3 for use for at least 12 years. Virtually none of the use was of conjugated estrogens, which accounted for most of the estrogen use in the present study; thus, the results may not be directly comparable. Finally, in a follow-up study conducted in Sweden (21), there was a 70 percent increase in risk among women who had used estrogens, primarily estradiol compounds, for at least 109 months. Estradiols are rarely used in the United States, and again the results may not be comparable with those of the present study, although it is by no means clear that estradiols would have a different effect on breast tissue than conjugated estrogens (38). Both of the Scandinavian studies had elevated relative risk estimates for estrogen therapy opposed by progestogens, but these were not statistically significant, and contrast with the results of a follow-up study (28) and a small clinical trial (29), which were consistent with a reduced risk for opposed therapy. Among other large case-control studies, there was not a statistically significant association with long-term use in the Toronto study (36) or in the Cancer and Steroid Hormone (CASH) Study (20), but the relative risk point estimates were somewhat elevated. The risk did not appear to be increased in the subgroups that have been previously reported. In neither of these studies was there much information concerning nonconjugated estrogens, or opposed therapy. Finally, two recent follow-up studies (26, 27) have shown modest elevations in risk for current users, and the investigators have proposed that their findings are compatible with a promotional effect. In the Adventist Health Study (26), after 7 years of followup, the relative risk estimate for breast cancer in relation to current use of estrogens at the time of enrollment was 1.69; there was no duration effect. In the Nurses' Health
1384
Kaufman et al.
Study (27), the relative risk estimate for the current use of conjugated estrogens (defined as exposure within the preceding 2 years) was 1.36; there was no trend in the risk according to duration of use or dose. In both studies, the selective diagnosis of breast cancer among current or recent estrogen users could readily have contributed to the associations. Estrogen users are advised to examine their breasts regularly, and to attend regular examinations (including screening mammographies). In the Nurses' Health Study (27), the proportions of tumors over 2 cm in size were the same in exposed and nonexposed cases, but this is not necessarily evidence against selection bias: it is only when the cancer becomes locally invasive or metastatic that the diagnosis becomes inevitable, and hence independent of selective detection by screening. In addition, the observations that the risk was elevated for durations of exposure of less than one year, and the lack of duration or dose effects, detract from the biologic plausibility of the findings. By contrast, there was no association with current use in the CASH Study (20), where the data were sufficient to exclude a risk increase of 30 percent for current users. Nor was there evidence of an increase in risk for current conjugated estrogen users in the present study, although the upper confidence limit was insufficient to exclude the recent findings. Based on the present results and those of other large epidemiologic studies, we conclude that unopposed conjugated estrogens taken for hormone replacement do not appreciably increase the risk of breast cancer. The possibility remains open that the risk is modestly increased after very long durations of use, and that opposed therapy, and other noncontraceptive estrogens, increase the risk. Further studies should focus on these latter questions.
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
15. 16. 17. 18.
19.
20.
REFERENCES 1. Casagrande J, Gerkins V, Henderson BE, et al. Brief communication: Exogenous estrogens and
21.
breast cancer in women with natural menopause. JNCI 1976;56:839-41. Hoover R, Gray LA, Cole P, et al. Menopausal estrogens and breast cancer. N Engl J Med 1976,295:401-5. Sartwell PE, Arthes FG, Tonascia JA. Exogenous hormones, reproductive history, and breast cancer. JNCI 1977;59:1589-92. Brinton LA, Williams RR, Hoover RN, et al. Breast cancer risk factors among screening program participants. JNCI 1979;62:37-44. Ross RK, Paganini-Hill A, Gerkins VR, et al. A case-control study of menopausal estrogen therapy and breast cancer. JAMA 1980;243:1635-9. Jick H, Walker AM, Watkins RN, et al. Replacement estrogens and breast cancer. Am J Epidemiol 1980;l 12:586-94. Brinton LA, Hoover RN, Szklo M, et al. Menopausal estrogen use and risk of breast cancer. Cancer 1981 ;47:2517-22. Kelsey JL, Fisher DB, Holford TR, et al. Exogenous estrogens and other factors in the epidemiology of breast cancer. JNCI 1981;67:327-33. Hoover R, Glass A, Finkle WD, et al. Conjugated estrogens and breast cancer risk in women. JNCI 1981;67:815-20. Lawson DH, Jick H, Hunter JR, et al. Exogenous estrogens and breast cancer. Am J Epidemiol 1981;114:710-13. Hulka BS, Chambless LE, Deubner DC, et al. Breast cancer and estrogen replacement therapy. Am J Obstet Gynecol 1982;143:638-44. Hiatt RA, Bawol R, Friedman GD, et al. Exogenous estrogen and breast cancer after bilateral oophorectomy. Cancer 1984,54:139-44. Kaufman DW, Miller DR, Rosenberg L, et al. Noncontraceptive estrogen use and the risk of breast cancer. JAMA 1984;252:63-7. Brinton LA, Hoover R, Fraumeni JF Jr. Menopausal oestrogens and breast cancer risk: an expanded case-control study. Br J Cancer 1986;54:825-32. La Vecchia C, Decarli A, Parazzini F, et al. Noncontraceptive oestrogens and the risk of breast cancer in women. Int J Cancer 1986,38:853-8. McDonald JA, Weiss NS, Daling JR, et al. Menopausal estrogen use and the risk of breast cancer. Breast Cancer Res Treat 1986;7:193-9. Nomura AMY, Kolonel LN, Hirohata T, et al. The association of replacement estrogens with breast cancer. Int J Cancer 1986;37:49-53. Buring JE, Hennekens CH, Lipnick RJ, et al. A prospective cohort study of postmenopausal hormone use and risk of breast cancer in US women. Am J Epidemiol 1987; 125:939-47. Hunt K, Vessey M, McPherson K, et al. Long-term surveillance of mortality and cancer incidence in women receiving hormone replacement therapy. Br J Obstet Gynecol 1987;94:620-35. Wingo PA, Layde PM. Lee NC, et al. The risk of breast cancer in postmenopausal women who have used estrogen replacement therapy. JAMA 1987;257:209-15. Bergkvist L, Adami H-0, Persson I, et al. The risk of breast cancer after estrogen and estrogenprogestin replacement. N Engl J Med 1989,321:293-7.
Estrogens and Breast Cancer
22. Thomas DB, Persing JP, Hutchinson WB. Exogenous estrogens and other risk factors for breast cancer in women with benign breast diseases. JNCI 1982;69:1017-23. 23. Rohan TE, McMichael AJ. Non-contraceptive exogenous oestrogen therapy and breast cancer. Med J Aust 1988;148:217-21. 24. Ewertz M. Influence of non-contraceptive exogenous and endogenous sex hormones on breast cancer risk in Denmark. Int J Cancer 1988;42:832-8. 25. Dupont WD, Page DL, Rogers LW, et al. Influence of exogenous estrogens, proliferative breast disease, and other variables on breast cancer risk. Cancer 1989;63:948-57. 26. Mills PK, Beeson WL, Phillips RL, et al. Prospective study of exogenous hormone use and breast cancer in Seventh-day Adventists. Cancer 1989,64:591-7. 27. Colditz GA, Stampfer MJ, Willett WC, et al. Prospective study of estrogen replacement therapy and risk of breast cancer in postmenopausal women. JAMA 1990,264:2648-53. 28. Gambrell RD Jr, Maier RC, Sanders BI. Decreased incidence of breast cancer in postmenopausal estrogen-progestogen users. Obstet Gynecol 1983,62:435-43. 29. Nachtigall LE, Nachtigall RH, Nachtigall RD. et al. Estrogen replacement therapy. II. A prospective etudy in the relationship to carcinoma and cardiovascular and metabolic problems. Obstet Gynecol 1979;54:74-9.
1385
30. Slone D, Shapiro S, Miettinen OS. Case-control surveillance of serious illnesses attributable to ambulatory drug use. In: Colombo F, Shapiro S, Slone D, et al, eds. Epidemiological evaluation of drugs. Amsterdam: Elsevier North-Holland Inc, 1977:5982. 31. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies. JNCI 1959;22:719-48. 32. Miettinen OS. Estimability and estimation in casereferent studies. Am J Epidemiol 1976; 103:22635. 33. Schlesselman JJ. Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982. 34. Key TJA, Pike MC. The role of estrogens and progestogens in the epidemiology and prevention of breast cancer. Eur J Cancer Clin Oncol 1988;24:29-43. 35. Kelsey JL. A review of the epidemiology of human breast cancer. Epidemiol Rev 1979:1:74-109. 36. Helmrich SP, Shapiro S, Rosenberg L, et al. Risk factors for breast cancer. Am J Epidemiol 1983;117:35-45. 37. 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 1991; 134:1386-95. 38. Stevenson JC, Whitehead MI. Breast cancer and estrogen replacement. N Engl J Med 1990;322:201-2.
