Steroidal Contraception and Cancer Risk

The obstetrician-gynecologist who prescribes steroidal contraception for patients must be aware of potential cancer risks associated with these medications. These risks include a slightly increased risk of certain cancers (notably breast, cervix, and liver), as well as a reduced risk of ovarian and endometrial cancer. Schlesselman has calculated that the effect of 8 years of oral contraceptive use among 100,000 women in the United States would result in 151 additional cases of breast cancer, 125 additional cases of cervical cancer, and 41 additional cases of liver cancer, counterbalanced by 197 fewer cases of endometrial cancer and 193 fewer cases of ovarian cancer.¹ He concluded that the net effect of oral contraceptive use upon the risk of cancer among U. S. women was negligible.

Since the introduction of oral contraceptives in the 1960s, dosages and formulations have changed dramatically. Most oral contraceptives prescribed nowadays have a much lower dose of estrogen than the original pill. Various estrogens and progestins are now available. Much of the epidemiologic literature evaluating the impact of oral contraceptives on a variety of health outcomes is based on the experience of women whose contraceptives contained relatively high doses of estrogen. We do not know, therefore, whether the same risks and benefits associated with the older pills are to be found with the newer formulations. Goldzieher has argued that women should take high-dose oral contraceptives for at least two years to ensure adequate protection against ovarian and endometrial cancer.²

Injectable steroidal contraception and subdermal implants have also emerged as a popular option in the last decade or so. Both depot medroxyprogesterone acetate (DMPA) and levonorgestrel (Norplant) are widely used in the United States and elsewhere. Although less is known about the long-term health effects of these products, the obstetrician-gynecologist should be able to counsel patients appropriately on the use of these formulations. Kaunitz, in a review of the effects of DMPA on cancer risk, concluded that the only impact of DMPA was to decrease the risk of endometrial cancer.³ Data on the progesterone-releasing intrauterine device (IUD), Progestasert, should also be considered.

Worldwide, breast cancer is the leading cause of cancer death among women. We know that various reproductive factors, including age at menarche, age at first childbirth, and age at menopause affect at woman's risk of breast cancer. What is the impact of steroid contraception about breast cancer?

In 1991, Thomas published a metaanalysis using the previously reported epidemiologic studies on the subject. It was concluded that overall oral contraceptive users were not at higher risk for developing breast cancer as compared with the controls.⁴ Even users with a family history of breast cancer had no increased risk. However, a slight but significantly higher risk of breast cancer diagnosis was demonstrated in users younger than 45 years (relative risk [RR], 1.16). Long-term use of oral contraceptives in this age group further increased the risk for this cancer (RR, 1.42). Similar results were also reported by Schlesselman who found the relative risk of breast-cancer diagnosis in oral-contraceptive users younger than 45 years increased from 1.18 to 1.25 and 1.29 after 4.8 and 12 years of use respectively, compared with age-matched nonusers.¹

In women older than 45 years, oral contraceptive use has not been found to increase the risk of breast cancer diagnosis. Data from the Cancer and Steroid Hormones Study (CASH), analyzed by Wingo and coworkers, showed a slight decrease in the risk with oral contraceptive use (RR, 0.9) in women between 45 and 65 years.⁵ In the metaanalysis of Schlesselman, it was also found that women between the ages of 45 and 60 had only a nonsignificant trend to increased risk with increasing duration of oral contraceptive use.¹

The Collaborative Group on Hormonal Factors in Breast Cancer performed a metaanalysis of 54 studies in 25 countries including a total of 53,000 women with breast cancer and more than 100,000 controls.^6,7 Overall, women currently taking oral contraceptives seemed to be at slightly increased risk of breast cancer (RR, 1.24). This risk progressively declined, however, after the last use so that 10 years or more after discontinuation, women faced no increased risk. Of note, oral contraceptive users who developed breast cancer were more likely to have disease confined to the breast than nonusers who developed breast cancer. No increase in the risk of breast cancer associated with increasing dose, duration of use, type of estrogen, or progestin was demonstrated. Similar findings were also reported in a recent case-control study by Ursin and associates, involving 744 women diagnosed with breast cancer or cancer-in-situ, at age 40 or younger. They found that long use (12 years), use at an early age, and use of high-estrogen pill, were associated with slight, nonsignificant increase in risk.⁸

There are fewer data on the potential association between breast cancer and DMPA. Skegg and coworkers have reported a pooled analysis of two large case-control studies in women with breast cancer.⁹ Overall, they found no increased risk of breast cancer associated with DMPA, even for women who used DMPA for more than 5 years. They did, however, note a twofold increase in the risk for recent users, mainly younger than 35 years of age. Shapiro and partners found no increased risk of breast cancer associated with DMPA in a case-control study conducted in South Africa among 419 cases and 1625 controls.¹⁰

The epidemiologic data given suggest an increased risk of breast cancer diagnosis in current and recent oral contraceptive users. Women older than 45 and those who stopped using the pill before 10 or more years are not at higher risk. Because oral contraceptive users belong usually to age groups among whom breast cancer incidence is smaller, the excess number of breast cancer cases attributable to oral contraceptive use would be small and do not increase the overall lifetime risk of breast cancer. Moreover, these tumors are usually early and localized.^6,7 These data support the hypothesis that the additional breast cancers diagnosed in oral contraceptive users can be the result of better surveillance, or of the stimulation of growth of small pre-existing lesions caused by steroids, or both.

Although these considerations are quite reassuring, special attention has been addressed to the possibility that breast cancer occurring before the age of 35 may be partly due to an inherited mutation (BRCA 1/2). Although family history of breast cancer has not been found to modify risk in oral contraceptive users, it was recently suggested that mutation carriers may be at heightened risk for breast cancer if they are also oral contraceptive users.⁸

Cervical cancer is the sixth commonest malignant tumor in women in the United States and the leading cause of cancer deaths among women in many countries of the third world. What is the impact of oral contraceptives on cervical intraepithelial neoplasia and cancer? Both case-control and cohort studies have demonstrated an increased risk for cervical neoplasia associated with oral contraceptive use.¹¹ Even after controlling for confounding factors, such as sexual activity, infection with human papillomavirus (HPV), and screening bias, most studies have shown an increased risk, rising to about twofold excess for users of 5 or more years. More recent case-control studies have demonstrated similar findings. Daling and associates, in a study in Washington State, found that oral contraceptive use was important only if first use occurred at an early age (younger than 17 years).¹² Preinvasive cervical neoplasia has also been related to oral contraceptive use. Ylitalo and colleagues noted a fourfold increase in risk for cervical carcinoma in situ associated with current oral contraceptive use.¹³ Among their study subjects, the risk of cervical neoplasia increased with duration of use. Gram and colleagues reported that current users of oral contraceptives had an RR of 1.5 for cervical intraepithelial neoplasia.¹⁴ Kjellberg and partners, however, found that in multivariate analysis HPV infection and cigarette smoking, but not oral contraceptive use, remained significant predictors of cervical dysplasia.¹⁵ Oral contraceptives have also been linked to an increased risk for cervical adenocarcinomas. In the study of Thomas and coworkers, the risk was greatest among long-term users and users of high progestin potency products.¹⁶ Passage of time appeared to diminish the risk. Lacey and colleagues have reported an association between oral contraceptive use and cervical adenocarcinoma-in-situ but not invasive adenocarcinoma.¹⁷ In the Oxford Family Planning Association cohort study of 17,000 women, current or recent long-term users of contraceptives were at increased risk for all types of dysplasia, carcinoma-in-situ, and invasive cancer.¹⁸ The odds ratio (OR) was 3.34 for 49 to 72 months of use, 1.69 for 73 to 96 months of use, and 2.04 for 97 or more months.

Two case-control studies have examined an association between DMPA use and cervical neoplasia. One of these studies, conducted in Costa Rica, found no increased risk of cervical neoplasia or cancer associated with DMPA.¹⁹ In the second study, conducted in Thailand, Mexico, and Kenya, there was slightly increased risk of cervical carcinoma-in-situ associated with DMPA, but no increased risk of cervical carcinoma.²⁰ A third cohort study, conducted in New Zealand, found similar rates of cervical neoplasia among DMPA users compared with women using oral contraceptives and intrauterine devices for contraception.²¹

Contraceptive implants are designed to release progestins at a steady rate. Preliminary studies have shown no change in cervical cytology among implant users over 2 years.²² After 2 years, in a cohort of 60 patients, there was no evidence of an increased risk for cervical neoplasia or cancer. We do not yet have any data regarding the long-term effects of these implants on the risk of cervical cancer.

The mechanisms by which oral contraceptives might increase a woman's risk for cervical neoplasia have not been firmly identified. Proposed mechanisms include the promotion of growth of neoplastic lesions, changes in cervical mucus which might facilitate HPV infection, a broadening of the transformation zone that puts more cells undergoing metaplasia at risk for HPV infection, decreased immune response to HPV infection, and increased HPV replication.^23,24,25 Further in vitro, animal, and human studies are required to elucidate the biologic link between oral contraceptives and cervical neoplasia.

There have been numerous case reports linking oral contraceptive use to liver cancer. Case-control studies suggest that the risk is minimal for duration of use less than 5 years. For women who use oral contraceptives for longer than 5 years, the risk of liver cancer may be increased.²⁶ A more recent case-control study in Europe found no increased risk associated with either short- or long-term use of oral contraceptives.²⁷ It is possible that the more modern oral contraceptives, with a lower estrogen dose, pose less risk than the older formulations. Prolonged use of oral contraceptives also puts women at increased risk for focal nodular hyperplasia of the liver.²⁸ A WHO study in countries with high prevalence of hepatitis infection did not reveal any increase in the risk of liver cancer associated with short-term oral contraceptive use.²⁹ However, the possibility of interactions between chronic hepatitis B and C and long-term oral contraceptive use cannot be excluded.³⁰

A recent large case-control study has confirmed a hypothesis suggested by earlier reports that oral contraceptive use might increase the risk of gestational trophoblastic disease.³¹ In this study involving 235 cases and 413 control subjects, overall RR associated with oral contraceptive use was 1.9. This risk increased with duration of use, from 1.9 for those who used oral contraceptives for 3 to 6 years to 2.5 for those with 10 or more years of use. The mechanism by which oral contraceptives might increase the risk of gestational trophoblastic disease is unknown. It has been hypothesized that long-term use of oral contraceptives might damage the ova or interfere with meiosis and yield ova with absent or nonfunctioning nuclei.

Many case reports have been published about women who developed prolactin-secreting pituitary adenomas after exposure to estrogen, both endogenous, from pregnancy, and exogenous. However, most case-control studies have not found an increased RR associated with oral contraceptives.³² Even the large cohort studies of oral contraceptive users conducted in the United States and Great Britain did not find any increased incidence of pituitary adenomas.

La Vecchia and associates have reported on a metaanalysis of 13 epidemiologic studies of thyroid cancer, based on 2132 cases, and 3301 controls.³³ They found that current oral contraceptive users have a significantly increased risk of thyroid cancer (OR, 1.5). There was no relation with duration of use, age at first use, or use before first birth. OR declined with increasing time since stopping (OR, 1.1 for more than 10 years since stopping). It has been hypothesized that oral contraceptives may plan a role in promotion, rather than initiation, of thyroid cancer.

Both case-control and cohort studies have consistently demonstrated that oral contraceptives are associated with a decreased risk of ovarian cancer.³⁴ Protection from ever-use is around 40%, rising to over 50% for use longer than 5 years.^35,36 A protective effect in seen both among white and black women.^37,38 In the CASH study, this protective effect was seen in women who had used oral contraceptives for as little as 3 to 6 months.³⁹ It was independent of the type of oral contraceptive used as well as the histologic type of ovarian cancer. In addition, the protection appears to last for 15 years after the use ended. In similar findings, Rosenberg and associates in a study of 441 cases and 2065 controls found a risk of 0.6 for women who used oral contraceptives as long as 15 to 19 years after stopping oral contraceptive use.⁴⁰ In the Royal College of General Practitioners Oral Contraception study, users of oral contraceptives had a 40% (RR, 0.6) decreased incidence of ovarian cancer compared with never-users.⁴¹ The Oxford Family Planning Association contraceptive study found a RR of 0.4 associated with oral contraceptive use.⁴² Women with a longer duration of use had a greater protection against ovarian cancer (RR of 1.0 for use up to 48 months, compared with RR of 0.3 for users of 97 months or more). The Nurses' Health Study, however, found parity to be the only reproductive factor that had a substantial independent association with ovarian cancer.⁴³ The impact of long-term oral contraceptive use only retained borderline significance after adjustment for other risk factors. Rosenblatt and colleagues have reported finding a slightly lower OR for ovarian cancer among women who had taken high-dose oral contraceptives, compared with those who had taken lower dose oral contraceptives (OR 0.68 compared with OR 0.81).⁴⁴ This suggests that the degree of protection may be slightly lower for newer low-dose formulations.

One case-control study has examined the association between DMPA and risk of ovarian cancer.⁴⁵ They found no significant difference in the risk of ovarian cancer between DMPA users and nonusers.

It has been hypothesized that the protection may be secondary to a decreased number of lifetime ovulations in women on oral contraceptives or progestin-induced apoptosis of abnormal ovarian epithelial cells, or both. Harlow and coworkers have reported that women who consume more than 11 g/day of lactose are the most likely to benefit from oral contraceptive use after age 30 years.⁴⁶ This same group had previously reported an increased risk of ovarian cancer associated with increased lactose intake.

Narod and associates have reported that oral contraceptives appear to protect women at inherited risk for ovarian cancer.⁴⁷ In a case-control studying 207 women with hereditary ovarian cancer and 161 of their sisters as controls, they found an OR of 0.5 for ever-users of oral contraceptives. This protection was noted among both BRCA1 and BRCA2 mutation carriers. Similar findings were reported by Godard and associates.⁴⁸

As with ovarian cancer, both case-control and cohort studies have consistently demonstrated an association between oral contraceptive use and decreased risk of endometrial cancer.⁴⁹ In the CASH study, a 40% lower risk of endometrial cancer was observed in women who used oral contraceptives for at least 12 months.⁵⁰ The protection lasted for 15 years or more after cessation of use. Metaanalysis of eight case-control studies and two cohort studies found that 1 year of oral contraceptive use reduced risk to 77%, 2 years to 62%, 4 years to 49%, and 12 years to 30%.^51,52 The Oxford Family Planning Association has subsequently reported an RR of 0.1 for ever-users of oral contraceptives compared to never-users.⁴² The degree of protection appears to be related to the dose of progestin, with formulations containing higher doses of progestin conferring increased protection. Voigt and colleagues, conversely, found no impact of progestin potency on RR of endometrial cancer.⁵³ Weiderpass and associates found that progestin alone pills reduced risk to a greater degree than pills with both estrogen and progestin.⁵⁴ Stanford and coworkers have reported that the protection from endometrial cancer persisted for as long as 20 years after discontinuing oral contraceptives.⁵⁵

One case-control study has examined the risk of endometrial cancer among DMPA users. Overall, DMPA users were 80% less likely to be diagnosed with endometrial cancer than nonusers.⁵⁶ The protection persisted at least 8 years after the last use of DMPA.

Contraceptive implants are designed to release progestins at a steady rate. Preliminary studies have shown significant reduction in endometrial thickness associated with the use of contraceptive implants.²² After 2 years, in a cohort of 60 patients, there was no evidence of an increased risk for endometrial hyperplasia or endometrial cancer. We do not yet have any data regarding the long-term effects of these implants on the risk of endometrial cancer.

The progesterone-releasing intrauterine device (Progestasert) releases progesterone at a constant rate for more than 1 year. It has been used to treat atypical endometrial hyperplasia and well-differentiated endometrial carcinoma. We do not have any data on the long-term effects of Progestasert use on the risk of subsequent endometrial cancer.

These findings are consistent with the hypothesis that endometrial hyperplasia and endometrioid endometrial carcinoma arise in a hormonal milieu characterized by an excess of estrogen relative to progestin. Topical or systemic progestins can reverse this hormonal imbalance. We have no data as yet whether women at inherited risk for endometrial cancer (hereditary nonpolyposis colorectal cancer [HNPCC], mismatch repair) have the same level of benefit from progesterone-containing contraception as women not at inherited risk.

Both case-control and cohort studies have suggested an association between oral contraceptive use and decreased risk of colon cancer.

Fernandez and partners have reported a case-control study conducted in Italy with 803 women with colon cancer, and 429 women with rectal cancer, and 2793 controls.⁵⁷ Ever use of oral contraceptives was associated with an OR of 0.63 for colon cancer and 0.66 for rectal cancer. Duration of use of oral contraceptives was inversely related to risk of colon cancer, but not of rectal cancer. Martinez and colleagues have reported a 40% lower risk of colorectal cancer in the cohort Nurses' Health Study over 1,012,280 person-years of follow-up.⁵⁸ Troisi and associates, however, did not find any association between past oral contraceptive use and risk of colorectal cancer among 57,529 women who were part of a breast cancer screening program between 1973 and 1980 (388,555 person-years).⁵⁹ The biologic mechanism by which oral contraceptives might decrease the risk of colorectal cancer is not known. Freedman and associates have reported that colon cancers that developed in parous women or women who took oral contraceptives were less likely to have p53-positive tumors than those cancers that developed in nulliparous women or women who had never used oral contraceptives.⁶⁰ These findings suggest that biologic heterogeneity may explain the inconsistent epidemiologic data.

Steroidal contraception, in the form of oral contraceptives, injectable contraceptives, and progestin-releasing intrauterine devices, can affect a woman's risk of cancer. The hormone-releasing intrauterine devices, which do not cause any change in systemic hormonal levels, probably only affect a woman's risk of endometrial cancer. Injectable contraceptives, which release progestins, appear to increase the risk of breast cancer among current users and decrease the risk of endometrial cancer among current and ever-users. Oral contraceptives, however, are associated with a mixed picture of cancer risk and cancer protection. Because the baseline incidence of cancer among child-bearing women is low, the attributable risk of cancer due to oral contraceptives is small. Nonetheless, it is reasonable to conclude that oral contraceptives increase a woman's risk of breast cancer, preinvasive cervical neoplasia, cervical cancer, liver cancer, and gestational trophoblastic disease as well as decrease her risk of endometrial, ovarian, and possibly colorectal cancer. Combined with protection from unwanted pregnancy, as well as the other health benefits associated with steroidal contraception, the benefits of steroidal contraception outweigh the risks for the general population. Obstetrician-gynecologists should be able to counsel their patients appropriately on these issues, as well as be prepared to evaluate promptly any signs or symptoms that might suggest the development of neoplasia.