A meta-analysis published in 2000, summarizing 15 case-control studies, concluded there was no overall association between estrogen replacement therapy (ERT) and risk of epithelial ovarian cancer.1 However, there was statistical heterogeneity among the studies examined, with an odds ratio of 1.3 (95% confidence interval [CI]=1.0–1.6) for a subset of 4 U.S. studies with community controls. Subsequently, 2 cohort studies2,3 and 24,5 of 34–6 case-control studies have reported a positive association between ERT and ovarian cancer occurrence or mortality. The association was strongest for unopposed ERT of long duration, with an approximately doubling of ovarian cancer risk for long-term ERT use compared with nonuse in the 2 cohort studies and 1 case-control study.5
To add to the few prospective studies on this topic, we analyzed the association of ERT with epithelial ovarian cancer incidence in the Iowa Women’s Health Study. An early publication on ovarian cancer in the IWHS, with only 97 events, reported no association with ERT.7 However, there was limited power to address long-term ERT use.7 Our hypothesis was that ERT of long duration would be associated with increased ovarian cancer incidence.
The Iowa Women’s Health Study Cohort
In 1986, we mailed a questionnaire to nearly 100,000 randomly selected women age 55 to 69 years who had a valid Iowa driver’s license in 1985. A total of 41,836 returned the questionnaire. Compared with nonrespondents, respondents were on average 3 months older, had 0.4 kg/m2 lower body mass index, and had lower incidence rates of several smoking-related diseases.8 Follow-up questionnaires were mailed in 1987, 1989, 1992, and 1997 to confirm participants’ residence, vital status, and other characteristics.
Women were queried about ERT by asking whether they had ever used “pills, other than birth control pills, that contain estrogen or other female hormones” and how long they took them (with response categories of <1 month, 2 to 6 months, 7 to 12 months, 1 to 2 years, 3 to 5 years, and >5 years). Type of ERT was not assessed. Each follow-up questionnaire also asked about current ERT (as a yes/no variable) but not duration.
Women were asked at baseline the age at which they first menstruated, whether they currently had menstrual periods (within the past year), and, if not, the age at which their periods stopped completely. In addition, they were asked whether they had had their uterus or 1 or both ovaries surgically removed. Hysterectomy and oophorectomy status was updated in the 1992 follow-up questionnaire. Participants were asked at baseline whether they had ever been pregnant, and for each pregnancy (up to 10) their age, pregnancy duration, and outcome (live birth, stillbirth, miscarriage, ectopic, induced abortion). Participants also were asked whether they had ever taken birth control pills. However, given the age of this cohort, few women had used oral contraceptives and use was not related to ovarian cancer risk.7 History of tubal ligation was not assessed. Participants were asked to report current weight and height, from which body mass index (in kilograms per square meter [kg/m2]) was computed. A paper measuring tape was sent for a friend, spouse, or relative to measure the circumference of the waist (1 inch above the umbilicus) and the hips (maximum); waist/hip ratio was calculated. The self-measured or self-reported anthropometric measures obtained by this protocol were valid and reliable.9
The baseline questionnaire assessed smoking status and amount smoked. Leisure physical activity level was assessed in 2 ways. Participants were asked a general question about regular physical activity that has been used for over 3 decades by the Gallup Poll: “Aside from any work you do at home or at a job, do you do anything regularly—that is, on a daily basis—that helps keep you physically fit?”10 In addition, participants completed 2 questions asking how often they participated in moderate physical activity (for example, bowling, golf, light sports or physical exercise, gardening, taking long walks) and vigorous physical activity (for example, jogging, racket sports, swimming, aerobics, strenuous sports). We combined these to form a 3-level physical activity index (low, moderate, high) based on frequency and intensity of activity. This activity index was associated inversely with mortality from coronary heart disease in this cohort,11 indicating that it has reasonable predictive validity.
To ascertain prevalent cancer, subjects were asked whether they had ever been diagnosed by a physician as having any form of cancer other than skin cancer and to specify the site. Participants were asked whether their mother, maternal or paternal grandmothers, aunts, sisters, or daughters had cancer diagnosed, including ovarian cancer.
Ovarian cancer cases were ascertained through the State Health Registry of Iowa, part of the National Cancer Institute’s Surveillance, Epidemiology, and End Results Program (SEER), via an annual computer match of name, maiden name, and date of birth. Primary site, morphology, extent of disease, and date of diagnosis were obtained for each incident cancer case from 1986 through 2000. Only cases diagnosed within the state of Iowa are identified and included among cases. This analysis included ovarian cancers classified as “common epithelial tumors” by the World Health Organization Histologic Classification of Ovarian Tumors.12
We considered women to be at risk from January 1986 through December 31, 2000, or until they were diagnosed with ovarian cancer, died, moved out of Iowa, or were otherwise lost to follow up. Deaths in Iowa were found by the State Health Registry. Deaths outside of Iowa were found by the 1987, 1989, 1992, and 1997 mailed follow ups, and, for nonrespondents, via the National Death Index. For deaths outside of Iowa, the censoring date was the midpoint between the date of last contact in Iowa and the date of death. Women known to have moved from Iowa were censored at the date of the move, if known, or at the midpoint between the date of last contact in Iowa and first known date out of Iowa, or the end of the follow-up period. The outmigration rate among this cohort was approximately 1% per year.
We excluded from the analysis women who reported at baseline a history of cancer other than skin cancer (N = 3830) or bilateral oophorectomy (N = 8064) (1455 women reported both baseline cancer and bilateral oophorectomy). We also excluded women who developed nonepithelial ovarian neoplasms (N = 16). After these exclusions, 31,381 women remained in the at-risk cohort. We ran analyses using the updated oophorectomy information from the 1992 follow up, censoring the 1219 women who reported an interim bilateral oophorectomy. Data from the State Health Registry led to the identification of 223 incident cases of epithelial ovarian cancer over the 15 years of follow up.
We categorized the baseline characteristics of women into natural categories or quartiles. We computed age-adjusted and multivariate-adjusted relative risks (RR) and their CIs through proportional hazards regression models. The multivariate model was developed in the following fashion. We started with variables that had previously been found to be possible risk factors in this cohort,7 including physical activity and smoking, even if they were not traditional risk factors for ovarian cancer. We examined the association of these covariates with ERT and, because most showed at least a weak association, most were included in multivariate models. In a supplemental analysis, exclusion of physical activity and smoking from the model, and addition of oral contraceptives, did not materially affect the results.
We tested the assumptions of proportional hazards regression and found them not to be violated. In supplemental analyses, current ERT use (yes, no) during follow up was modeled as a time-dependent variable.
At baseline, 8% of these 55- to 69-year-old women reported currently taking ERT, and 24% reported former use. Compared with women not using ERT (Table 1), current users at baseline tended to be less overweight, have a lower waist/hip ratio, be more physically active, and be more likely to have had a hysterectomy and unilateral oophorectomy. During the entire 15 years of observation, 20% of women used ERT at some time, often briefly.
During follow up, we identified 223 incident epithelial ovarian cancers. Compared with never users, the age-adjusted RR of ovarian cancer was 1.4 (CI = 0.9–2.2) for current ERT users at baseline and 2.0 (CI = 1.2–3.4) for current ERT users of more than 5 years (Table 2). After excluding women who reported a unilateral oophorectomy before baseline (34,924 person-years and 8 cases excluded), these RRs were 1.6 (CI = 1.0–2.5) for current use and 2.4 (CI = 1.4–4.1) for current use greater than 5 years. In contrast, women who at baseline either formerly used ERT or were currently users for less than 5 years did not have increased incidence of ovarian cancer compared with never users. When treated as a time-dependent covariate, use versus nonuse of current ERT at any time during follow up carried a RR of 1.5 (CI = 1.0–2.1), similar to the RR for baseline use.
Adjustment for other baseline characteristics somewhat strengthened the association between current ERT use at baseline and ovarian cancer. Compared with never-use of ERT, multivariate-adjusted RRs were 1.7 for current use and 2.5 for current use greater than 5 years (Table 2). The multivariate-adjusted RR of ovarian cancer for time-dependent current ERT also was 1.7. All of these multivariate-adjusted RRs were nearly identical after excluding women who reported a partial oophorectomy before baseline (data not shown).
Finally, results were similar in an analysis limited to the 151 incident cases who died with ovarian cancer listed as the underlying cause. For this death-based analysis, the multivariately adjusted RRs were 2.0 (CI = 1.1–3.5) for current (versus never) use of ERT at baseline and 2.8 (CI = 1.4–5.8) for current use greater than 5 years (versus never).
In this cohort of older women followed for 15 years, participants who currently used ERT at baseline in 1986 (or at any time during follow up) had a 72% higher multivariate-adjusted RR of epithelial ovarian cancer compared with never users. Incidence was 2½ times higher among women who at baseline were users of ERT for more than 5 years compared with never users. We did not have data on whether women were using progestins with ERT, but other data from the mid-1980s suggested that no more than 20% of current ERT users took combination hormonal replacement therapy.13,14 Thus, our baseline findings pertain mostly to unopposed ERT, although more combination therapy could have been initiated during follow up.
Our findings seem consistent with the 2 other recent cohort studies on ERT and ovarian cancer. The large American Cancer Society cohort study,2 initiated in 1982, investigated mortality from ovarian cancer; that study did not have information on whether the ERT included progestin. Compared with never users, the RR of ovarian cancer death was 1.5 (CI = 1.2–2.0) for current ERT use and 2.2 (CI = 1.5–3.2) for current ERT use more than 10 years. Long duration of use also elevated risk in former ERT users.2 In the Breast Cancer Detection Demonstration Project,3 time-dependent analyses yielded a RR of incident ovarian cancer of 1.6 (CI = 1.1–3.0) for unopposed ERT, and RRs were 1.8 for 10 to 19 years and 3.2 for 20 or more years of ERT use. In contrast, there was no clear association of ovarian cancer incidence with estrogen plus progestin use.3 Taken together, these 3 prospective studies and some case-control studies1,4,5 suggest that long-term unopposed ERT could at least double the risk of ovarian cancer.
The mechanisms by which ERT might cause ovarian cancer are unknown. Pituitary gonadotropins (including follicle-stimulating hormone and luteinizing hormone) increase during menopause as a result of lack of suppression from estradiol produced by the ovary; these could act as a promoters on affected ovarian tissue.15 If so, it might be expected that ERT would be protective, because estrogen suppresses gonadotropin secretion.16,17 However, a prospective study found lower levels of gonadotropins in women who subsequently developed ovarian cancer.18 Another theory suggests that ovarian cancer is the result of estrogen stimulation through autocrine or paracrine mechanisms.19,20 The surface epithelium of both normal ovary and malignant ovarian tumors has been shown to express estrogen receptors.21,22 Thus, for postmenopausal women, ERT use (particularly unopposed) could stimulate the ovary and lead to malignancy. A third theory is that long-term use of ERT is not causal, but rather a marker of some hormonal condition related to low levels of estrogen that predisposes to ovarian cancer.
Two studies have noted an unexplained increase in ovarian cancer in U.S. women in recent decades.23,24 We speculate that increased ERT use could be a contributing factor if long-term ERT use is truly associated with an increased ovarian cancer risk.
Our study had some limitations. First, the participants were postmenopausal and almost all were white, limiting generalizability to other population subgroups. Second, we lacked data on type of ERT. Third, our data on duration of ERT use at baseline did not provide for categories of use beyond 5 years, and duration of ERT use during follow up was not explicitly ascertained. Fourth, there were only 223 cases of ovarian cancer, yielding potentially unstable estimates for smaller ERT exposure groups, as well as limiting power to detect weak associations for some exposure groups or to explore potential effect modifiers. Fifth, it is theoretically possible that women taking ERT of long duration had greater medical screening for ovarian cancer than did nonusers; this would tend to elevate the RR. However, most ovarian cancers are not detected by routine screening, but rather are found when they become symptomatic. Moreover, results were similar for ovarian cancer incidence and death, arguing against this bias. Finally, this was an observational study and could suffer from unknown confounding biases. However, no clinical trial of ERT has lasted long enough or had enough ovarian cancer events to address this hypothesis.
In summary, long duration of ERT use postmenopausally could increase the risk of epithelial ovarian cancer.
1.Coughlin SS, Giustozzi A, Smith SJ, et al. A meta-analysis of estrogen replacement therapy and risk of epithelial ovarian cancer. J Clin Epidemiol
2.Rodriguez C, Patel AV, Calle EE, et al. Estrogen replacement therapy and ovarian cancer mortality in a large prospective study of US women. JAMA
3.Lacey JV Jr, Mink PJ, Lubin JH, et al. Menopausal hormone replacement therapy and risk of ovarian cancer. JAMA
4.Chiaffarino F, Pelucchi C, Parazzini F, et al. Reproductive and hormonal factors and ovarian cancer. Ann Oncol
5.Riman T, Dickman PW, Nilsson S, et al. Hormone replacement therapy and the risk of invasive epithelial ovarian cancer in Swedish women. J Natl Cancer Inst
6.Sit AS, Modugno F, Weissfeld JL, et al. Hormone replacement therapy formulations and risk of epithelial ovarian carcinoma. Gynecol Oncol
7.Mink PJ, Folsom AR, Sellers TA, et al. Physical activity, waist-to-hip ratio, and other risk factors for ovarian cancer: a follow-up study of older women. Epidemiology
8.Bisgard KM, Folsom AR, Hong C-P, et al. Mortality and cancer rates in nonrespondents to a prospective study of older women: 5-year follow-up. Am J Epidemiol
9.Kushi LH, Kaye SA, Folsom AR, et al. Accuracy and reliability of self-measurement of body girths. Am J Epidemiol
10.The Gallup Poll: Public Opinion 1987
. Wilmington, DE: Scholarly Resources Inc; 1988;316:
11.Folsom AR, Kaye SA, Sellers TA, et al. Body fat distribution and 5-year risk of death in older women. JAMA
12.Serov SF, Scully RE, Sobin JH. International Histological Classification of Tumors. No. 9. Histological Typing of Ovarian Tumors
. Geneva: World Health Organization; 1973.
13.Hemminki E, Kennedy DL, Baum C, et al. Prescribing of noncontraceptive estrogens and progestins in the United States, 1974–86. Am J Public Health
14.Nabulsi AA, Folsom AR, White A, et al. Association of hormone replacement therapy with various cardiovascular risk factors in postmenopausal women. N Engl J Med
15.Stadel B. The etiology and prevention of ovarian cancer. Am J Obstet Gynecol
16.Larsson-Cohn U, Johannsson ED, Kagedal B, et al. Serum FSH, LH, and oestrone levels in post-menopausal patients on oestrogen therapy. Br J Obstet Gynaecol
17.Aedo AR, Le Donne M, Landgren BM, et al. Effect of orally administered oestrogens on gonadotrophin levels in post-menopausal women. Maturitas
18.Helzlsouer KJ, Alberg AJ, Gordon GB, et al. Serum gonadotropins and steroid hormones and the development of ovarian cancer. JAMA
19.Fathalla MF. Incessant ovulation—a factor in ovarian neoplasia? Lancet
20.Cramer DW, Welch WR. Determinants of ovarian cancer risk, II: inferences regarding pathogenesis. J Natl Cancer Inst
21.Lau K, Mok S, Ho S. Expression of human estrogen receptor-alpha and -beta, progesterone receptor and androgen receptor mRNA in normal and malignancy ovarian epithelial cells. Proc Natl Acad Sci U S A
22.Lindgren P, Backstrom T, Mahlck CG, et al. Steroid receptors and hormones in relation to cell proliferation and apoptosis in poorly differentiated epithelial ovarian tumors. Int J Oncol
23.Gnagy S, Ming EE, Devesa SS, et al. Declining ovarian cancer rates in U.S. women in relation to parity and oral contraceptive use. Epidemiology
24.Tarone RE, Chu KC. Age–period–cohort analyses of breast-, ovarian-, endometrial-, and cervical-cancer mortality rates for Caucasian women in the USA. J Epidemiol Biostat