Epidemiologic Perspective of Ovarian Cancer
M. Steven Piver
Table Of Contents
M. Steven Piver, MD
FACTORS RELATED TO THE DEVELOPMENT OF OVARIAN CANCER|
FACTORS PROTECTIVE AGAINST OVARIAN CANCER
In the United States, ovarian cancer is the sixth most common cancer in women and the second most common malignancy of the female genital tract. The incidence of ovarian cancer increases with age, as follows:
Age<30: 5 in 100,000
Worldwide, the incidence of ovarian cancer varies among geographic regions. Ovarian cancer tends to occur among the most affluent women in the most highly industrialized countries. The incidence of ovarian cancer is significant in Scandinavia, Israel (American-born or European-born residents), and North America. The frequency of ovarian cancer is low among East Indians and Africans and African Americans and Latinos in the United States. Although the incidence of ovarian cancer is lower in Japan and Hong Kong compared with the United States, in the 1990s it increased 15% to 25% every 5 years in both of these countries.2
The cause of ovarian cancer is unknown. Epidemiologic studies suggest that environmental or dietary factors or both may play a role. Advances in molecular biology suggest that genetics play a larger role than previously suspected. Despite major advances made in treatment, ovarian cancer continues to have the highest fatality-to-case ratio of all gynecologic malignancies.3 A better understanding of potential risk factors for the development of ovarian cancer may help prevent or at least decrease the incidence of this chronic, fatal disease.
|FACTORS RELATED TO THE DEVELOPMENT OF OVARIAN CANCER|
Hygienic Perineal Talc
Because it is related to asbestos, talc has been implicated as a cause of ovarian cancer.4 Besides the chemical similarities between talc and asbestos, cosmetic talcum powder manufactured before 1976 contained significant amounts of the three types of asbestos: tremolite, chrysolite, and crocidolite.5,6 Studies by Heller and associates5 and Henderson and colleagues7 showed that talc particles may be found in normal and malignant tissues. Conceptually, talcum powder applied to the perineum can reach the ovaries via retrograde flow. The association between talc and ovarian cancer remains equivocal. Whittemore and associates8 could not show an association between genital talcum powder exposure and ovarian cancer; however, Cramer and coworkers9 reported a relative risk (RR) of 1.9 (p < .03) for ovarian cancer among talc users. The RR was increased further in women who used talcum powder on perineal napkins and dusting powder on the perineum (Table 1). Wong and coworkers10 reported, however, in a case-control study that included 704 controls and 430 cases of ovarian cancer no increased risk for ovarian cancer with the use of perineal talc. There was no increased risk with 20 or more years of use of perineal talc. In a meta-analysis of six studies, Harlow and colleagues11 calculated a modest risk for ovarian cancer development among talc users: RR 1.3 (95% confidence interval [CI], 1.1 to 1.6) (Table 2).11 Although these case-control studies showed a small increased risk for development of ovarian cancer with perineal talc use, in 2000 the scientific advisory panel to the National Toxicology Program, a branch of the National Institutes of Health, voted 7 to 3 against recommending that perineal talc be listed as a human carcinogen.12 The panel next meets in 2002.
CI, confidence interval.
(Harlow BL, Cramer DW, Bell DA, et al: Perineal exposure to talc and ovarian cancer risk. Obstet Gynecol 80:19, 1992)
The major dietary distinction in industrialized nations is the high intake of meat and animal fat, which has been reported to be associated with an increased incidence of ovarian cancer in several studies.13 Shu and associates14 reported a significant (p <.01)dose-response relationship between the intake of fat from animal sources and ovarian cancer risk in a case-control study of 172 epithelial ovarian cancer cases and 172 controls. In a prospective study of 20,305 women, Helzlsouer and colleagues15 showed that women with cholesterol levels greater than 200 mg/dL are at increased risk for ovarian cancer.
Lactose, found naturally only in milk, is the only source of dietary galactose. Ovarian tissue has a relative abundance of galactose-metabolizing enzymes, and experimental studies on rats indicated that galactose-rich diets can have an adverse effect on the ovaries.16,17,18,19 Young women with galactosemia, an inborn error of galactose metabolism, experience hypergonadotropic hypogonadism and early ovarian failure.20 Cramer and colleagues21 proposed that increased dietary galactose consumption and low serum levels of galactose-1-phosphate uridyltransferase, an important enzyme in the metabolism of galactose, may be linked to an increased risk of ovarian cancer. In a case-control study, the investigators found that women who consumed significantly more yogurt and cottage cheese, the only foods consumed regularly that distinguished the cases from the controls, had three times the incidence of ovarian cancer than the controls. Women with ovarian cancer were reported to have significantly lower galactose-1-phosphate uridyltransferase levels than controls (Table 3). These authors proposed that in women with an inherited abnormality of galactose-1-phosphate uridyltransferase, the ovaries would be exposed to galactose for a prolonged period, increasing the ovarian cancer risk in these women compared with women who can metabolize galactose efficiently.
Cl, confidence interval; RR, relative risk.
(Cramer DW, Harlow BL, Willett WC, et al: Galactose consumption and metabolism in relation to the risk of ovarian cancer. Lancet 2:66, 1989)
Several other studies found no association between galactose consumption and the incidence of ovarian cancer. In a case-control study of 450 patients, Risch and associates22 showed that galactose consumption did not increase the risk of epithelial ovarian cancer (Table 4). In a 2000 report of the Nurses’ Health Study of 80,326 women who completed a dietary questionnaire, Fairfield and coworkers23 reported, however, a 44% greater risk for the development of invasive epithelial ovarian cancer for women with the highest consumption of lactose compared with women with the lowest consumption (RR, 1.44; 95% CI, 1.01 to 2.07; p for trend 0.07). The largest contributors to dietary lactose were skim and low-fat milk.3 Another, perhaps better relationship between milk consumption and ovarian cancer was proposed by Mettlin and Piver,24 who evaluated this association in terms of animal fat content in milk. The RR for ovarian cancer was 3.1 for women drinking more than one glass of whole milk each day compared with women who never drank whole milk (Table 5). Women who drank only whole milk were at increased risk for ovarian cancer compared with women who drank skim or 2% fat milk. These findings suggest that animal fat, described as the fat content in milk, may be the source of increased risk.
OC, oral contraceptives.
(Adapted from Risch HA, Jain M. Marrett LD, et al: Dietary lactose intake, lactose intolerance, and the risk of epithelial ovarian cancer in Southern Ontario [Canada]. Cancer Causes Control 5:540, 1994)
Infertility and Fertility Drugs
Infertility and the use of fertility drugs have been found to increase the risk of ovarian cancer.25,26,27,28 In a case-control study of 3837 women evaluated for infertility, Rossing and colleagues25 observed an increased ovarian cancer risk among infertile women (standardized incidence ratio [SIR], 2.5; 95% CI, 1.3 to 4.5); this study suggested that women with infertility resulting from abnormalities in ovulation were at greater risk for ovarian cancer (SIR, 3.7; 95% CI, 1.4 to 8.1) than women with infertility resulting from tubal abnormalities (SIR, 3.0; 95% CI, 0.4 to 10.8). Similarly, Whittemore and coworkers26 reported that it is not the estimated number of years of ovulation, the number of pregnancies, or the lack of use of oral contraceptives that increases a woman’s risk for ovarian cancer, but rather the inability to conceive among ovulating women who have sexual intercourse for 10 or more years, unprotected by any type of contraception (Table 6).
The use of ovulation-inducing agents is increasing in the United States; in the 1980s and 1990s, the number of clomiphene citrate prescriptions dispensed annually nearly doubled.29 In the late 1990s, researchers raised concerns regarding the relationship between ovulation-inducing agents and the development of ovarian cancer. Rossing reported25 an increased risk of ovarian cancer among women who had ever used clomiphene citrate (SIR, 3.1; 95% CI, 1.4 to 5.9) and a further increase in risk of ovarian cancer among women who used clomiphene for 12 or more cycles (SIR, 11.1; 95% CI, 1.5 to 82.3). In a nationwide case-control study from Israel, Shushan and associates28 reported an adjusted odds ratio (OR) of 3.19 (95% CI, 0.86 to 11.82) among women who used human menopausal gonadotropins alone and an adjusted OR of 1.42 (95% CI, 0.65 to 3.12) among women who used human menopausal gonadotropins alone or combined with clomiphene citrate. In an analysis of fertility drug use from eight case-control studies conducted between 1989 and 1999 in the United States, Denmark, Canada, and Australia, Ness and coworkers30 reported, however, no increased risk for the development of invasive epithelial ovarian cancer among infertile women taking clomiphene citrate or human menopausal gonodotropin.
Estrogen Replacement Therapy
Estrogen replacement therapy (ERT) has been used to relieve menopausal symptoms such as vaginal dryness and hot flashes; studies have shown that ERT has significant protective effects benefiting the cardiovascular system, bones, and lipoprotein profile.31,32,33 The association between ERT and ovarian cancer is inconclusive, however.34,35,36,37,38,39 Results from the Collaborative Ovarian Cancer Group showed no risk of ovarian cancer among ERT recipients (OR, 1.1; 95% CI, 0.59 to 2.0), and several studies agreed with these findings.34,39 Cramer and colleagues37 reported an association, although not statistically significant, between ERT and the development of ovarian cancer (RR, 1.56; 95% CI, 0.85 to 2.87). The most controversial report on ERT and ovarian cancer is that of the Department of Epidemiology and Surveillance Research of the American Cancer Society.40 These investigators did not report on the association between the use of ERT and the risk of developing ovarian cancer but rather that among postmenopausal women, the use of ERT in the 1970s and early 1980s for 10 or more years doubled the risk of dying from ovarian cancer (Table 7). The authors assumed that women took estrogen alone (without progesterone) and admitted that they did not know which estrogen product the women took.
It is commonly accepted that cancer results from a series of genetic alterations.41,42 In 1972, Knudson and Strong43 proposed the two-hit mutation model for the development of cancer (Fig. 1). According to this concept, in inherited cancers, the first mutation (single-gene alteration) is prezygotic, making the fetus predisposed to the development of cancer from conception onward. Cancer develops in the somatic cells (in this case, the ovary), however, only when a postzygotic second mutation occurs after birth. In contrast, nonhereditary cancers occur when both mutations occur postzygotically in the somatic cell line (the ovary). Knudson and Strong’s hypothesis helps explain why hereditary cancers frequently occur 10 to 20 years earlier than nonhereditary cancers (i.e., because the first mutation has occurred prezygotically).
Genetic changes in two categories of normal cellular genes—proto-oncogenes and tumor-suppressor genes—have been proposed to cause cancer.41,42,44 Genetic changes (e.g., point mutations, translocations, amplifications) in proto-oncogenes facilitate the transformation of a normal cell to a malignant cell by producing an altered or overexpressed gene product. Products of proto-oncogenes include cellular growth factors, growth factor receptors, enzymes, and DNA-binding proteins.45,46 Tumor-suppressor genes are proposed to prevent cancer.42,44,45,46,47,48 Inactivation or loss of these genes contributes to the development of cancer because of the lack of a functional gene product. Dysfunction of tumor-suppressor genes could result in abnormal cell division, abnormal gene expression, or the increased ability of cells in tissue to detach.
Activation of several proto-oncogenes (particularly K-ras, h-ras, c-myc, and HER-2/neu) has been assayed by various techniques in ovarian tumors and in established ovarian tumor cell lines.49,50,51 These techniques have measured the ability of isolated tumor DNA to transform the murine fibroblast cell line (NIH/3T3) and the amplification of proto-oncogene messenger RNA and protein by tumors; data from these investigations showed that activation of proto-oncogenes occurs in ovarian tumors. The significance of proto-oncogene activation and the role, if any, it plays in the development of ovarian cancer remain to be determined.
Until more recently, there was little evidence in the literature that the origin of some ovarian cancers was either genetic or familial. In 1979, we reported a family in which epithelial ovarian cancer developed in five members, spanning three generations: a grandmother, three daughters, and a granddaughter.52 Shortly thereafter, five members of a second family, spanning two generations (three sisters, a first cousin, and her daughter), came under our care for epithelial ovarian cancer.53 Our review of the literature revealed that before 1970, familial ovarian cancer had been reported in only five families. In contrast, during the 1980s, 26 more families were documented as having familial ovarian cancer. In response to this seemingly real increase in the incidence of familial ovarian cancer, and because of the need for appropriate genetic counseling, a Familial Ovarian Cancer Registry was established at Roswell Park Cancer Institute in 1981 to document the number of cases occurring in the United States and to study the mode of inheritance. It was postulated that if an autosomal dominant mode of transmission was established, genetic counseling for prophylactic oophorectomy at an appropriate age might lead to a decrease in the death rate from ovarian cancer in such families.
The Registry’s first report in 1984 documented 201 cases of ovarian cancer in 94 families.54 After the initial collection of 76 cases in 1981, the average number of cases enrolled between 1982 and 1996 was 189 per year (range, 32 to 635). Because of an increased awareness of a possible genetic link generated by the media, the Registry now receives daily information on new families with familial ovarian cancer. In 1990, the Registry was renamed the Gilda Radner Familial Ovarian Cancer Registry (GRFOCR) in honor of the comedic actress, Gilda Radner, who died of ovarian cancer. Through December 31, 2001, the Registry has enrolled 1652 families with two or more close relatives with ovarian cancer (Fig. 2). The Registry has made it clear that familial ovarian cancer is not a rare occurrence, as originally conceived in the four decades from 1930 to 1970.
Three conditions are associated with familial ovarian cancer: (1) site-specific ovarian cancer, in which the familial risk is restricted to ovarian cancer; (2) breast/ovarian cancer syndrome, with clustering of ovarian and breast cancer cases in extended pedigrees; and (3) cancer syndrome family, also known as Lynch syndrome II, which consists of hereditary nonpolyposis colorectal cancer (with most of the cancers occurring in the proximal colon) in association with ovarian or endometrial carcinoma or both.55 In all of these syndromes, inheritance seems to follow an autosomal dominant pattern with variable penetrance. In addition, several genetic factors are associated with increased gonadal dysgenesis (dysgerminoma),56 Peutz-Jeghers syndrome (mucocutaneous pigmentation or gastrointestinal polyposis, granulosa cell tumor, and cystadenoma),57 and basal cell nevus syndrome (fibroma).58
Hall and associates59 in 1990 mapped by genetic linkage a locus for familial ovarian cancer to the long arm of chromosome 17 in the interval 17q12-21. Before the cloning of BRCA1 by Miki and coworkers60 in 1994, the GRFOCR participated in the Breast Cancer Linkage Consortium Study of 145 breast/ovarian cancer families. 61 These families consisted of three or more cases of early onset (<60 years old) breast or ovarian cancer. It was estimated that 92% of the families in which there was no male breast cancer and with two or more cases of ovarian cancer were linked to BRCA1. The GRFOCR participated in another international study of nine families with site-specific ovarian cancer defined as three or more cases of epithelial ovarian cancer and no case of breast cancer diagnosed at less than 50 years of age. This study documented that 89% of such site-specific ovarian cancer families are linked to BRCA1.62,63
In 1995, Wooster et al64 cloned BRCA2, which previously had been mapped by genetic linkage to chromosome 13q12. It is now estimated that BRCA1 and BRCA2 are responsible for 5% to 10% of all epithelial ovarian cancers with 52% BRCA1, 32% BRCA2, and 16% other yet undefined genes.65 In the United States, lifetime risk of developing epithelial ovarian cancer is 1 in 55 or 1.8% compared with 28% to 44% for BRCA1 carriers and 27% for BRCA2 carriers.
Because familial ovarian cancer (site-specific) is thought to be inherited in an autosomal dominant pattern with variable penetrance, sisters and daughters in families with a history of familial ovarian cancer (two or more first-degree relatives, mother/sister/daughter, who share one half of one’s genes) may have a 50% chance of disease development. This figure compares with a 1.8% (1 in 55) in women without a family history. It has been recommended that such women undergo a prophylactic oophorectomy by age 35 if they have completed childbearing. A word of caution: There have been reports of women with a family history of ovarian cancer who have been diagnosed with intra-abdominal carcinomatosis histologically similar in appearance to ovarian adenocarcinoma (also known as extraovarian papillary carcinoma of the peritoneum) after prophylactic oophorectomy.66 In some cases, the prophylactically removed ovaries were cancerous on retrospective review67; however, in other cases of extraovarian papillary carcinoma, the prophylactically removed ovaries were cancer-free. It is postulated that whatever causes ovarian cancer also affects the peritoneum to form an extraovarian papillary carcinoma. From 1981 to July 1982, the GRFOCR had enrolled 931 families totaling 1222 cases of familial ovarian cancer. Of these, 324 had undergone prophylactic oophorectomy.68 Of these 324 women, 6 or 1.8% developed primary peritoneal carcinoma indistinguishable from epithelial ovarian cancer 1, 2, 5, 13, 15, and 27 years after prophylactic oophorectomy. Although these data suggest that development of primary peritoneal carcinoma after prophylactic oophorectomy for a family history of familial ovarian cancer is not a common event, they do point out that all women who undergo such preventive surgery need to be under lifetime surveillance for the possible development of primary peritoneal carcinoma.
|FACTORS PROTECTIVE AGAINST OVARIAN CANCER|
Epidemiologic studies have shown repeatedly that the use of oral contraceptives reduces a woman’s risk for epithelial ovarian cancer. Whittemore and associates34 showed that women who used oral contraceptives had an RR of 0.66 (95% CI, 0.55 to 0.78), a 40% reduction compared with population controls. The risk of ovarian cancer decreases as the duration of using oral contraceptives increases, and this protective effect occurs even with a short duration of use. Using meta-analysis and regression equations, Gross and Schlesselman69 proposed that women with a positive family history of ovarian cancer can reduce their risk to a level below that of a woman with no family history of ovarian cancer and who has never used oral contraceptives. Similarly a nulliparous woman who uses oral contraceptives for 5 years can reduce her risk to the level of a parous woman who has never used oral contraceptives (Table 8).
OC, oral contraceptive.
(Adapted from Gross TP, Schlesselman JJ: The estimated effect of oral contraceptive use on the cumulative risk of epithelial ovarian cancer. Obstet Gynecol 83:419, 1994)
Pregnancy consistently has been found to have a protective effect against ovarian cancer.34,70,71 Data from the Collaborative Ovarian Cancer Group showed a 40% risk reduction after the first term pregnancy and a 14% risk reduction with each successive pregnancy (Table 9).34 In a large prospective study of 121,700 women in the Nurses’ Health Study, Hankinson and coworkers72 documented that parous women have a reduced ovarian cancer risk relative to nulliparous women (RR, 0.84; 95% CI, 0.77 to 0.91 per pregnancy). Whittemore and associates34 also showed that failed pregnancies (defined as miscarriages, abortions, ectopic pregnancies, and stillbirths) also decreased the risk of ovarian cancer but had a lesser effect than that of full-term live births.
CI, confidence interval.
*Pregnancies of at least 20 weeks’ gestation.
†Adjusted for age, study, and oral contraceptive use.
(Adapted from Whittemore AS, Harris R, Itnyre J, and the Collaborative Ovarian Cancer Group: Characteristics relating to ovarian cancer risk: Collaborative analysis of 12 US case-control studies: Part II. Invasive epithelial ovarian cancer in white women. Am J Epidemiol 136:1184, 1992)
Two theories have been offered to explain the decreased incidence of ovarian cancer associated with oral contraceptive use and multiparity or, conversely, why nulliparous women and, more significantly, women who have tried to conceive but have failed are at higher risk for ovarian cancer. In 1948, Garder73 proposed the concept termed by Cramer and Welch74 the excess gonadotropin secretion theory. This theory held that ovarian cancer results from continually high levels of gonadotropin acting on the ovaries as a result of ovarian failure or the inability of the ovarian-pituitary feedback mechanism to inhibit the excretion of these high levels of gonadotropin on the ovary. The incessant ovulation theory of Fathala75 proposed that continuous ovulation causes repeated trauma to the ovary and leads to the development of ovarian cancer, noting that the risk increases in proportion to the number of times a woman ovulated in her lifetime. Conversely, most factors that are associated with a lower risk of ovarian cancer (e.g., oral contraceptive use, pregnancy) suppress ovulation. Berchuck and colleagues76 showed that patients with ovarian cancer who had an overexpression of p53 had a significantly greater number of lifetime ovulations (OR, 7.7; 95% CI, 1.4 to 41.2); these results strengthen the hypothesis that proliferation of ovarian epithelium in response to ovulation may increase the risk of spontaneous mutations in p53.
Tubal Ligation and Hysterectomy
Interruption of the female genital tract, either by bilateral tubal ligation or hysterectomy, has been shown to diminish the risk of ovarian cancer.77,78 In a prospective study of 121,700 registered nurses in the Nurses’ Health Study with 12-year follow-up, a strong inverse relationship between bilateral tubal ligation and ovarian cancer was reported (RR, 0.33; 95% CI, 0.16 to 0.64).77 Similarly, in a case-control study of 300 cases and 606 controls, Cornelison and coworkers79 showed a 50% risk reduction in women who had tubal ligations. Hysterectomies seem to reduce the risk of ovarian cancer by 50%.77,78,79,80 It has been hypothesized that tubal ligations and hysterectomies decrease the risk of ovarian cancer by preventing the ascending migration of potential carcinogens.
The author expresses his appreciation to Dr. Cheung Wong, for his contribution to previous editions of this chapter.
8. Whittemore AS, Wu ML, Paffenbarger RS, et al: Personal and environmental characteristics related to epithelial ovarian cancer: Part II.Exposure to talcum powder, tobacco, alcohol, and coffee Am J Epidemiol 128:1228, 1988
33. Farish E, Fletcher CD, Hart DM, et al: The effects of conjugated equine oestrogens with and without a cyclical progestin on lipoproteins and high density lipoprotein subfractions in postmenopausal women. Acta Endocrinol 113:123, 1986
34. Whittemore AS, Harris, , Itnyre J, and the Collaborative Ovarian Cancer Group: Characteristics relating to ovarian cancer risk: Collaborative analysis of 12 US case-control studies: Part II. Invasive epithelial ovarian cancers in white women Am J Epidemiol 136:1184, 1992
65. Ford D, Easton DF, Stratton M, et al: Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families.The Breast Cancer Linkage Consortium Am J Hum Genet 62:676, 1998
68. Piver MS, Jishi MF, Tsukada Y, et al: Primary peritoneal carcinoma after oophorectomy in women with a family history of ovarian cancer: A report of the Gilda Radner Familial Ovarian Cancer Registry. Cancer 71:2751, 1993