OBJECTIVE: To estimate the association of the estrogen and progestin potency of combined oral contraceptive pills (OCPs) with epithelial ovarian carcinoma risk.
METHODS: This population-based case–control study included 745 women with incident, histologically confirmed epithelial ovarian carcinoma and 943 controls, matched on age and ethnicity. Data were collected using a standard questionnaire, picture albums, and calendars. The association of OCP potency with epithelial ovarian carcinoma risk was modeled using unconditional logistic regression.
RESULTS: When compared with women who never used hormonal contraception, users of OCPs with low estrogen (equal to or less than 0.035 mg ethinyl estradiol) and low progestin (less than 0.3 mg norgestrel) were at significantly reduced risk of ovarian carcinoma (odds ratio 0.19; 95% confidence interval 0.05–0.75). The risk among these women was lower than among users of estrogen or progestin of high potency, but the difference was not statistically significant. However, in a subset of 205 women who reported exclusive use of norethindrone, users of 0.5 mg or less, had a significantly reduced risk of ovarian cancer compared with women using 10 mg of this progestin. Increased norethindrone dose resulted in a significant increase in ovarian carcinoma risk, indicating a dose–response association.
CONCLUSION: Combined OCPs were effective at decreasing the risk of epithelial ovarian carcinoma, with the strongest risk reduction associated with low-potency formulations.
LEVEL OF EVIDENCE: II
Combined oral contraceptive pills, especially those with low-potency estrogen and progestin components, are effective at reducing the risk of epithelial ovarian carcinoma.
From the 1Cancer Epidemiology Program, Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Hawaii; and 2Department of Obstetrics and Gynecology, John A. Burns School of Medicine, University of Hawaii, Honolulu, Hawaii.
Supported in part by Public Health Service grant R01-CA-58598 and by contracts N01-CN-67001 and NO1-CN-25403 from the Department of Health and Human Services, National Institutes of Health.
Corresponding author: Dr. Galina Lurie, Cancer Research Center of Hawaii, University of Hawaii, 1236 Lauhala Street, Honolulu, HI; e-mail: glurie@CRCH.hawaii.edu.
Numerous epidemiologic studies have shown that oral contraceptive pill (OCP) use reduces the risk of epithelial ovarian carcinoma.1 During the past 30 years, substantial changes have been made in the estrogen and progestin content of OCPs, as well as in the schedule of their use, aimed at decreasing the undesirable side effects.2 Low-dose OCPs have been proven to be effective in the suppression of ovulation, although some investigators have speculated that their role in reducing ovarian cancer risk is less effective than that of older formulations that contained higher doses of hormones.3–5
The few earlier studies6–10 showed that both high- and low-dose formulations were protective. Formulations with low estrogen and low progestin had similar estimated risk reduction when compared with high estrogen and high progestin formulations in a study by Ness et al.11 In contrast, Schildkraut et al4 found that high-potency OCP formulations provided significantly greater reduction in risk than formulation with estrogen and progestin of low potency. Most recently, Pike et al5 reported a greater (nonsignificant) protective effect of OCPs with high dose of progestin, especially in combination with low estrogen dose. Inconsistencies in the results of studies of OCPs and ovarian cancer risk might reflect assessment of different formulations available at the time of exposure and the effect of time since first use. Categorization of OCPs by estrogen and progestin potency has also differed among studies. Numerous considerations complicate interpretation of available data on progestin potency, including the interaction between estrogen and progestin and the complex characteristics of progestins that might possess various degrees of progestational, estrogenic, and androgenic activity.12 Although assessment of the potency of progestin is difficult13,14 and has been a subject of controversy,15–18 we believe that certain generalizations can be made about the potency of progestins based on previous publications.14,18–21 In our study, we defined the potency of each progestin based on the subnuclear glycogen vacuolization assay, which was found to be superior in defining the progestational activity of hormones to the delay of menses assay17,20 and more consistent with clinical observations.22
We investigated the effect of OCP formulations containing various doses of estrogen and progestin on epithelial ovarian carcinoma risk among women who provided complete information on OCP use and who were exclusive users of the formulations that belonged to the same potency group. We also had a unique opportunity to examine the association of different doses of one progestin, norethindrone, on the risk of ovarian cancer in a relatively large subset of women who used only OCPs containing norethindrone with no intraindividual variation in its dose.
MATERIALS AND METHODS
This population-based case–control study was conducted in Hawaii and Los Angeles, California, and included women aged 18 years or older, who were diagnosed with histologically confirmed primary epithelial ovarian carcinoma between 1993 and 2005. Eligible cases were identified through the rapid-reporting systems of the Hawaii Tumor Registry and the Cancer Surveillance Program at the University of Southern California. Information on tumor stage, grade, and histology was obtained from pathology and surgical reports. Eligibility criteria for controls included age 18 years or older, residency in Hawaii or Los Angeles County for a minimum of 1 year, no prior history of ovarian cancer, and having at least one intact ovary. Control participants were randomly selected from participants in an annual survey of representative households in Hawaii23 and by random-digit dialing in Los Angeles.24 Control pools were supplemented with women aged 65 years or older through random sampling from lists obtained from the Health Care Finance Administration. We used a frequency-matching approach to ensure comparability of cases and controls by age and ethnicity. The pool of controls was stratified by ethnicity and 5-year age groups, and approximately one control was randomly selected to match to each case to achieve similar distributions of age and ethnicity in study and comparison groups.25 A total of 1,778 women, 772 (43.4%) cases and 1,006 (56.6%) controls, were recruited. The response rate was 63.7% for cases and 70.7% for controls. The study was approved by the Institutional Review Boards of the University of Hawaii and the University of Southern California. All study participants signed detailed consent forms.
Interviews were conducted by a multilingual interviewing staff and took approximately 2.5 hours to complete. Interviewers were uniformly trained and supervised to standardize interviewing and coding techniques. Quality control and performance of the interviewers was monitored by a repeat interview by project coordinators of a random sample of 15% of participants on a random 5% of the questions. A structured pretested questionnaire was used for data collection including sociodemographic and health-related information, menstrual, reproductive, and gynecologic histories, and exogenous hormone use.26 A description of specific OCP, content and dosage, reason for use, time interval of use, and adverse effects was recorded for each episode of OCP use. Monthly calendars were used to facilitate recollection of detailed information on reproductive history and episodes of hormone use. Interviewers used photo albums containing more than 100 products to aid in the identification of specific OCPs.
The “potency” of the estrogen component of each formulation was defined by the dose of ethinyl estradiol (E2) or mestranol in mg of ethinyl E2 equivalent. The potency of mestranol was considered to have 50% of the potency of ethinyl E2.27 The OCP formulations with estrogen content 0.035 mg or more ethinyl E2 were categorized into “high estrogen” potency group; OCPs with estrogen dose less than 0.035 mg were considered as “low estrogen” potency formulations. Progestational potency of OCPs was estimated based on the results of the work of Grant and Philips as described in Dickey and Stone20 and Dickey.28 The progestational activity of 1 mg of each progestin was expressed in milligrams of norgestrel equivalent—0.1 mg for norethynodrel, 0.38 mg for norethindrone, 0.44 mg for norethindrone acetate, 0.53 mg for ethynodiol diacetate, 0.40 mg for chlormadinone acetate, 2.0 mg for levonorgestrel, and 0.12 mg for medroxyprogesterone acetate20—and was multiplied by a daily dose of the progestin in each formulation. Progestins with a dose of less then 0.3 mg norgestrel were classified as low, as in a study by Pike et al.5
Overall, 340 cases (36%) and 610 controls (64%) reported using some type of hormonal contraceptive. We excluded the participants who used parenteral contraceptives (13 cases and 36 controls); 9 cases and 14 controls who reported ever using sequential contraceptives; and 5 cases and 13 controls who used progestin-only pills. We did not assess these types of contraceptives separately because they were not used exclusively. A total of 317 cases (36.5%) and 551 controls (63.5%) reported exclusive use of combination OCPs. For 347 women (154 cases; 194 controls), it was not possible to verify a specific brand name or dose of OCP formulation used in each episode. These women were classified as “unknown OCP” users. Women who reported exclusive use of the same potency formulation during all episodes (up to six episodes reported) were classified into four groups by potency: “high estrogen and high progestin,” “high estrogen and low progestin,” “low estrogen and high progestin,” “low estrogen and low progestin.” Women who used formulations with both high and low potency of either one of the hormone components were combined into the separate group of “various potency” OCP users. We also created categories of “high progestin and low progestin,” and “high estrogen and low estrogen,” collapsing the data over the estrogen and progestin potency categories, respectively.
Statistical analysis was performed using the Statistical Analysis System software (SAS 8.02, SAS Institute Inc., Cary, NC). Unconditional multiple logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (95% CIs) for the association of OCP categories by potency of estrogen and progestin with the ovarian carcinoma risk while adjusting for age (continuous variable), ethnicity (white, Asian, other), study site, education (12 or fewer years, 13–14 years, 15 years or more), family history of ovarian cancer among first-degree female relatives (yes or no), gravidity (0, 1, 2–3, 4 or more), age at last pregnancy in gravid women (younger than 25 years, 25–29 years, 30–34 years, 35 years or older), history of a tubal ligation procedure (yes or no), menopausal status (premenopausal compared with postmenopausal) and type of menopause (natural compared with induced), age at menopause (52 years or younger, older than 52 years), menopausal hormonal therapy (estrogen alone, progesterone alone, combination of estrogen and progesterone), duration of OCP use (less than 5 years, more than 5 years), and time since first OCP use (less than 5 years, more than 5 years). The log likelihood test was used in selecting variables for inclusion into the final model. Age at first or last pregnancy (full term or incomplete) was not associated with epithelial ovarian carcinoma risk, but we included age at last pregnancy as a covariate into all final models to adjust for the age at last exposure to high levels of naturally occurring sex steroids. Including number of all pregnancies and age at last pregnancy, regardless of its duration, instead of parity and age at last full-term pregnancy resulted in a better model fit based on the log likelihood test. Such variables as perineal area talc powder use, age at menarche, body mass index (calculated as weight in kilograms divided by the square of height in meters), history of clinically diagnosed fertility problems in a participant, fertility drugs use, use of other steroid hormones, ovarian surgery, history of diabetes mellitus, history of endometriosis, alcohol consumption, and smoking were not associated with epithelial ovarian cancer risk in our analyses and did not change the association between OCP use and ovarian carcinoma when included in the models. To evaluate the association of norethindrone dose with ovarian cancer risk, the trend variable was assigned four levels, as follows: 1=0.4–0.5 mg; 2=1 mg; 3=2 mg, and 4=10 mg. We chose to use uniformly distributed scores rather than the dosage level, because the inherent linearity assumption fit better with these scores.29 The attributable risk, the fraction of ovarian cancer that would have been avoided if all women had used OCPs, was computed by the method of Bruzzi, which assumes a representative sample of cases30; a 95% CI was computed based on normal theory and the standard error based on influential function methods.31 All P values were derived from two-tailed statistical tests. Statistical significance was considered at a P value less than .05.
The distribution of cases and controls by participant characteristics included in the statistical models of ovarian cancer risk are presented in Table 1. The mean age (55.6 years; range 18 to 93 years) and ethnic distribution of the study participants was similar in cases and controls. Cases had fewer years of education, were more likely to have a family history of ovarian cancer, had fewer pregnancies (including full-term and incomplete), and were less likely to have a tubal ligation, to be premenopausal, and to use combined estrogen and progesterone menopausal hormones. Among women for whom complete information related to OCP use was available, 40% used combinations with a low estrogen potency and, of these, only 10% of women used formulations with a progestin component potency less than 0.3 mg norgestrel. Although time since first use and time since last use for these women were shorter, the duration of use appeared to be longer. No differences were observed in age at first or last use among women using OCPs of different potency. Overall use of OCPs was associated with 50% reduction in epithelial ovarian risk (OR 0.51, 95% CI 0.26–0.98) and provided an almost 40% risk reduction when duration of use was considered (OR 0.62, 95% CI 0.48–0.81). A significantly reduced risk of epithelial ovarian carcinoma was observed in all categories of OCP by potency, with ORs ranging from 0.62 for high estrogen and high progestin to 0.19 for low estrogen and low progestin users (Table 2) when compared with participants who never used hormonal contraception. The attributable risk for not using any combined OCP was estimated at 42% (95% CI 13–70%), and the attributable risk for not using low estrogen and low progestin formulations was 73% (95% CI 0.18–1.00%). That is, up to 42% of ovarian cancers might have been avoided if all women used some form of combined OCP and an estimated 73 % of ovarian cancers might have been avoided if all women used OCP formulation of low estrogen and low progestin. Although the odds of ovarian cancer were lower in users of low potency OCPs than in users of high potency OCPs, the difference in risk reduction among them was not statistically significant (Table 2). In analyses that included only invasive epithelial ovarian carcinoma (604 women), use of OCP was associated with a significant 46% lower risk (95% CI 0.41–0.70) of malignancy. All users of OCPs had significantly lower odds of developing invasive ovarian carcinoma when compared with never users (Table 3). Women who used low estrogen and low progestin formulations had the lowest OR for invasive ovarian carcinoma risk, although differences among groups were not statistically significant (Table 3). Subgroup analyses were performed among women aged younger than 55 years because they were the only group who were exposed exclusively to low potency formulations. Results were similar to those found for the entire group of participants (Tables 2 and 3). We further restricted the analyses to women who were exclusive users of combined monophasic OCP formulations with a single progestin, norethindrone, with no interindividual variation in dose (76 cases; 129 controls). The risk of developing ovarian carcinoma was significantly lower among users of low-dose (0.5 mg or less) norethindrone than among women who used high doses (10 mg) of this progestin. A positive monotonic increase in risk with increasing norethindrone dose was observed (P for trend was .02) (Table 4). A slightly greater reduction in the risk of ovarian carcinoma was found among women who used a lower dose of ethinyl E2 (adjusted for norethindrone dose) (Table 4).
The principal finding of this investigation was that use of combined OCPs with low estrogen (0.035 mg or less ethinyl E2) and progestin (less than 0.30 mg norgestrel) potency provided significant reduction in epithelial ovarian carcinoma risk. This risk reduction was consistently lower than that of higher-potency formulations in all of our analyses, including subgroups of younger women and women diagnosed with invasive carcinoma. In a subset of exclusive users of norethindrone, we evaluated changes in ovarian carcinoma risk with dose rather than using potency equivalents and showed a significant increase in the ORs for ovarian carcinoma risk, indicating a dose–response association; users of a dose 0.5 mg or less had significantly lower risk than users of 10 mg of this progestin.
The reduction of ovarian carcinoma risk associated with OCP use in our investigation was comparable to estimates found in other cohort and case–control investigations.1 Our results were inconsistent with those of Schildkraut et al,4 who reported that high-dose preparations provided greater reduction in ovarian cancer risk than low-dose OCPs. We also found no evidence that the use of high-dose progestin combined with low-dose estrogen resulted in the greatest reduction of the ovarian carcinoma risk as reported by Pike et al.32 Our results were in agreement with those of Ness et al11 and other investigators who showed that low-dose OCPs were no less effective than high-dose formulations.6–8,10
Several hypotheses have been suggested to explain the protective effect of OCP use against ovarian cancer risk.33–36 The similar associations of high and low-dose OCP formulations with the risk of ovarian carcinoma are consistent with the incessant ovulation and gonadotropin hypotheses given that combined OCPs, even with a low dose, provide effective ovulation suppression. Using ultrasonography to measure follicular size and endometrial thickness, researchers have shown that low-dose estrogen (0.020 mg ethinyl E2) and progestin (0.1 mg levonorgestrel) formulations effectively suppress ovulation.37,38 The slightly greater reduction in ovarian carcinoma risk among women using low-dose OCPs might have resulted from improved compliance with their use. The adverse effects and other complaints associated with OCP use have been mitigated by changes in the estrogen and progestin content of OCP formulations.39,40 According to the hormonal hypothesis,36 excessive androgen stimulation of the ovarian surface epithelium could lead to increased ovarian cancer risk, whereas progesterone is considered to be protective. Recently Rodriguez et al41 showed that treatment with OCPs containing levonorgestrel (or levonorgestrel in combination with ethinyl E2) dramatically induced apoptosis in the ovarian epithelium. Progesterone-induced apoptosis by an extrinsic pathway involving caspase-8 was recently demonstrated in normal ovarian epithelial cells and in ovarian cancer cell cultures by other investigators.42 The greater reduction in ovarian carcinoma risk associated with high-potency progestin OCPs in contrast to other formulations found by Schildkraut et al6 is in agreement with this hypothesis. However, studies of the dose-dependent effects of synthetic progestins on the biochemistry of the endometrium showed that low doses of progestins were sufficient to counteract the proliferative effects of estrogens.43,44 In a study by Siddle et al,43 increasing daily dosage of norethindrone to 10 mg daily was counterproductive, because DNA synthesis in glands and stroma was less effectively suppressed than with lower dosages. The same results were obtained by Whitehead et al44: the 10 mg dose of norethindrone was less effective than the 1, 2.5, or 5 mg daily doses at opposing estrogen stimulation. These findings might be attributed to the depletion of progesterone receptor sites, with increased levels of progestins resulting in a loss of biologic response. In addition, progesterone can both stimulate and inhibit cell proliferation and is involved in the regulation of a considerable number of genes about which little is known.45 Furthermore, progesterone receptors are induced by estrogen in most target tissues, so it is not possible at this time to distinguish progesterone-specific effects on the ovary from those of estrogen.
An important strength of this study was its population-based approach to participant selection and accurate case ascertainment through cancer registries that are part of the Surveillance, Epidemiology, and End-Results Program of the National Cancer Institute,46 with case ascertainment close to 99 %. Information regarding OCP use and other risk factors were collected by trained personnel using detailed interviewer-administered questionnaires, comprehensive photo albums of OCPs, and calendars to assist in recollection. Complete information on OCP use was obtained for a substantial proportion of women, including hormone dose and time-related factors related to OCP use. Furthermore, enough women exclusively used formulations with progestin doses lower than 0.3 mg norgestrel to make it possible to examine the association of these formulations with ovarian carcinoma risk. This study had a power of 80% at a critical level of 5% (two-tailed) to detect as statistically significant ORs between 0.40 and 0.73 when the categories were compared against never users in Tables 2 and 3. The range was 0.36–0.57 when the analysis was limited to women aged younger than 55 years. When formulation potency levels were compared among OCP users, the minimum detectable ORs ranged between 0.37 and 0.63 for all women and 0.22 and 0.58 for women aged younger than 55 years. We also had an opportunity to investigate the effect of progestin dose in a relatively large subgroup of women who used the same progestin norethindrone, with no intraindividual variation in dose. The results of this subgroup analysis did not depend on defining progestational potency and therefore was free of misclassification bias. Although the numbers were smaller, the comparison is very important as it compares dosage within the same type of progestin and confirms the earlier findings across all types of progestin. The comparisons using never users as the reference group resulted in minimal detectable ORs ranging from 0.38 to 0.68. Although comparisons of norethindrone formulations were limited to detection of very protective effects (0.1–0.3), significant differences between groups were still observed.
One of the limitations of our study is a possibility of a nonresponse bias. Response rates among cases were comparable to other population-based studies of ovarian cancer and were partially explained by short survival among cases with advanced disease. We analyzed the association of OCP use with risk by stage at diagnosis of ovarian cancer and found that the patterns of risk were similar, suggesting that the bias due to rapidly fatal cancer was not substantial in our study. A further limitation of this analysis was our reliance on patient recall. We did not review pharmacy records in this study, based on cost and resource considerations. Because we were interested in lifetime OCP use, this would have required review of records from multiple providers. However, in our previous studies of agreement between interview information and physician records, we found that women could recall estrogen use with a high degree of accuracy.47,48 Most women used several types of OCPs in their lifetimes from various potency groups, and it was therefore difficult to categorize them for the purpose of this analysis. Furthermore, 347 women who did not know specific OCP types were classified as “unknown OCP” users. We performed statistical analyses both including and excluding unknown contraceptive pill users and found similar estimates of risk. A further important limitation was that estrogenic and progestational components of the OCP have unique pharmacologic features and are not completely comparable.13,14,21,49
Oral contraceptive pills have been used by women for more than 40 years and, in addition to the contraceptive benefits, provide substantial protection against ovarian cancer, partially explaining the declining ovarian cancer rates in U.S. women.50 Compliance to OCP prescription is affected by undesirable adverse effects that have been mostly attributed to high doses of estrogenic and progestational compounds.39,40 The fact that low-dose OCPs provide significant reduction in ovarian cancer risk might be important for clinicians and women in selecting an OCP with the most favorable contraceptive and noncontraceptive benefits. Future studies that include larger numbers are needed to further evaluate the association of low-dose OCP formulations with ovarian cancer risk.
1. La Vecchia C. Oral contraceptives and ovarian cancer: an update, 1998-2004. Eur J Cancer Prev 2006;15:117–24.
2. Gerstman BB, Gross TP, Kennedy DL, Bennett RC, Tomita DK, Stadel BV. Trends in the content and use of oral contraceptives in the United States, 1964–88. Am J Public Health 1991;81:90–6.
3. Whittemore AS. Personal characteristics relating to risk of invasive epithelial ovarian cancer in older women in the United States. Cancer 1993;71 suppl:558–65.
4. Schildkraut JM, Calingaert B, Marchbanks PA, Moorman PG, Rodriguez GC. Impact of progestin and estrogen potency in oral contraceptives on ovarian cancer risk. J Natl Cancer Inst 2002;94:32–8.
5. Pike MC, Pearce CL, Peters R, Cozen W, Wan P, Wu AH. Hormonal factors and the risk of invasive ovarian cancer: a population-based case-control study. Fertil Steril 2004;82:186–95.
6. The reduction in risk of ovarian cancer associated with oral-contraceptive use. The Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. N Engl J Med 1987;316:650–5.
7. Rosenblatt KA, Thomas DB, Noonan EA. High-dose and low-dose combined oral contraceptives: protection against epithelial ovarian cancer and the length of the protective effect. The WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Eur J Cancer 1992;28A:1872–6.
8. Rosenberg L, Palmer JR, Zauber AG, Warshauer ME, Lewis JL Jr, Strom, BL, et al. A case-control study of oral contraceptive use and invasive epithelial ovarian cancer. Am J Epidemiol 1994;139:654–61.
9. Royar J, Becher H, Chang-Claude J. Low-dose oral contraceptives: protective effect on ovarian cancer risk. Int J Cancer 2001;95:370–4.
10. Sanderson M, Williams MA, Weiss NS, Hendrix NW, Chauhan SP. Oral contraceptives and epithelial ovarian cancer. Does dose matter? J Reprod Med 2000;45:720–6.
11. Ness RB, Grisso JA, Klapper J, Schlesselman JJ, Silberzweig S, Vergona R, et al. Risk of ovarian cancer in relation to estrogen and progestin dose and use characteristics of oral contraceptives. SHARE Study Group. Steroid Hormones and Reproductions. Am J Epidemiol 2000;152:233–41.
12. American Medical Association Drug Evaluations. 5th ed. Philadelphia (PA): W.B. Saunders; 1983.
13. Stanczyk FZ. Pharmacokinetics and potency of progestins used for hormone replacement therapy and contraception. Rev Endocr Metab Disord 2002;3:211–24.
14. Stanczyk FZ. All progestins are not created equal. Steroids 2003;68:879–90.
15. Edgren RA, Sturtevant FM. Potencies of oral contraceptives. Am J Obstet Gynecol 1976;125:1029–38.
16. Edgren RA. Progestational potency of oral contraceptives: a polemic. Int J Fertil 1978;23:162–9.
17. Dickey RP. Reply to paper by Dr. Edgren on “progestational potency of oral contraceptives: a polemic.” Int J Fertil 1978;23:170–4.
18. Edgren RA. Oral contraceptive “potencies” and deep venous thromboembolism. Int J Epidemiol 1992;21:420–1.
19. Dorflinger LJ. Relative potency of progestins used in oral contraceptives. Contraception 1985;31:557–70.
20. Dickey RP, Stone SC. Progestational potency of oral contraceptives. Obstet Gynecol 1976;47:106–12.
21. Benagiano G, Primiero FM, Farris M. Clinical profile of contraceptive progestins. Eur J Contracept Reprod Health Care 2004;9:182–93.
22. Dickey RP, Dorr CH. Oral contraceptives: selection of the proper pill. Obstet Gynecol 1969;33:273–87.
23. Oyama N, Johnson DB, Hawaii Health Surveillance Program Survey Methods and Procedures. R & S Report No 54. Honolulu (HI): State Department of Health Research and Statistics Office; 1986.
24. Waksberg J. Sampling methods for random digit dialing. J Am Stat Assoc 1978;73:40–6.
25. Schlesselman JJ. Case–control studies: design, conduct, analysis. New York (NY): Oxford University Press; 1982.
26. Tung KH, Goodman MT, Wu AH, McDuffie K, Wilkens LR, Kolonel LN, et al. Reproductive factors and epithelial ovarian cancer risk by histologic type: a multiethnic case-control study. Am J Epidemiol 2003;158:629–38.
27. Bolt HM, Bolt WH. Pharmacokinetics of mestranol in man in relation to its oestrogenic activity. Eur J Clin Pharmacol 1974;7:295–305.
28. Dickey RP. Managing contraceptive pill patients. 8th ed. Durant (OK): Essential Medical Information Systems; 1994.
29. Rothman KJ, Greenland S. Modern epidemiology. Philadelphia (PA): Lippincott-Raven; 1998.
30. Bruzzi P, Green SB, Byar DP, Brinton LA, Schairer C. Estimating the population attributable risk for multiple risk factors using case-control data. Am J Epidemiol 1985;122:904–14.
31. Graubard BI, Fears TR. Standard errors for attributable risk for simple and complex sample designs. Biometrics 2005;61:847–55.
32. Pike MC, Spicer DV. Hormonal contraception and chemoprevention of female cancers. Endocr Relat Cancer 2000;7:73–83.
33. Fathalla MF. Factors in the causation and incidence of ovarian cancer. Obstet Gynecol Surv 1972;27:751–68.
34. Stadel BV. Letter: The etiology and prevention of ovarian cancer. Am J Obstet Gynecol 1975;123:772–4.
35. Cramer DW, Welch WR. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. J Natl Cancer Inst 1983;71:717–21.
36. Risch HA. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst 1998;90:1774–86.
37. Spona J, Feichtinger W, Kindermann C, Wunsch C, Brill K. Inhibition of ovulation by an oral contraceptive containing 100 micrograms levonorgestrel in combination with 20 micrograms ethinyl E2. Contraception 1996;54:299–304.
38. Crosignani PG, Testa G, Vegetti W, Parazzini F. Ovarian activity during regular oral contraceptive use. Contraception 1996;54:271–3.
39. Sitruk-Ware R. New progestagens for contraceptive use. Hum Reprod Update 2006;12:169–78.
40. Sulak PJ. Contraceptive redesign: new progestins/new regimens. J Fam Pract 2004;suppl:S3–S8.
41. Rodriguez GC, Walmer DK, Cline M, Krigman H, Lessey BA, Whitaker RS, et al. Effect of progestin on the ovarian epithelium of macaques: cancer prevention through apoptosis? J Soc Gynecol Investig 1998;5:271–6.
42. Syed V, Ho SM. Progesterone-induced apoptosis in immortalized normal and malignant human ovarian surface epithelial cells involves enhanced expression of FasL. Oncogene 2003;22:6883–90.
43. Siddle NC, Townsend PT, Young O, Minardi J, King RJ, Whitehead MI. Dose-dependent effects of synthetic progestins on the biochemistry of the estrogenized post-menopausal endometrium. Acta Obstet Gynecol Scand Suppl 1981;106:17–22.
44. Whitehead MI, Townsend PT, Pryse-Davies J, Ryder T, Lane G, Siddle N, et al. Actions of progestins on the morphology and biochemistry of the endometrium of postmenopausal women receiving low-dose estrogen therapy. Am J Obstet Gynecol 1982;142:791–5.
45. Graham JD, Clarke CL. Physiological action of progesterone in target tissues. Endocr Rev 1997;18:502–19.
46. Surveillance, Epidemiology, and End Results Program. Summary Staging Guide for the Cancer Surveillance, Epidemiology, and End Results Reporting (SEER) Program. Bethesda (MD): U.S. Department of Health and Human Services, Public Health Services, National Institute of Health; 1986.
47. Goodman MT, Nomura AM, Wilkens LR, Kolonel LN. Agreement between interview information and physician records on history of menopausal estrogen use. Am J Epidemiol 1990;131:815–25.
48. Coughlin SS, Pickle LW, Goodman MT, Wilkens LR. The logistic modeling of interobserver agreement. J Clin Epidemiol 1992;45:1237–41.
49. Hammond GL, Rabe T, Wagner JD. Preclinical profiles of progestins used in formulations of oral contraceptives and hormone replacement therapy. Am J Obstet Gynecol 2001;185 suppl:S24–S31.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
50. Gnagy S, Ming EE, Devesa SS, Hartge P, Whittemore AS. Declining ovarian cancer rates in U.S. women in relation to parity and oral contraceptive use. Epidemiology 2000;11:102–5.