Egypt Ministry of Health

Published with support from the United States Agency for International Development (USAID), Global, G/PHN/POP/CMT, under the terms of Grant No. HRN-A-00-97-00009-00. The opinions expressed herein are those of the authors and do not necessarily reflect the views of the US Agency for International Development or the Johns Hopkins University.

Currently more than 100 million women rely on the pill. It is the top modern family planning method among married women in half of countries surveyed. The pill is most popular in Western Europe, where half of married women use it. It is least used in China, India, and Japan.

A great many women use the pill at some point in their lives. Outside India and China, half of married women who have ever used family planning have relied on the pill at some time. In the US 80% of all women born since 1945 have used the pill. A method so widely used deserves continuing attention from health care programs, providers, and researchers.

Oral contraceptives (OCs) offer a variety of other health benefits. For example, by reducing menstrual bleeding, OCs help prevent iron deficiency anemia, which is common and often serious in developing countries. OC use halves the risk of cancers of the uterine lining and of the ovary. Some protection persists for many years after OC use stops. Because estrogen-progestin OCs stop ovulation (release of an egg), they prevent pregnancy outside the womb, which can be life-threatening. Some evidence suggests that OCs reduce risk of colorectal cancer, too.

Compared with earlier, higher-dose pills, current low-dose formulations have considerably lowered the risk of heart attack, stroke, and blood clots in the deep leg veins attributed to OC use. Research has better defined which women would face appreciable risk of heart attack or stroke if they used OCs—women over age 35 who smoke or who have high blood pressure. For all other women, using OCs is clearly safer than childbearing in both developing and developed countries.

Cervical cancer is even harder to study than breast cancer. It may never be clear whether methodological problems in research or an actual cause-and-effect relationship explain recent observations of a slight increase in risk for long-term OC users. Condoms and careful choice of a sex partner offer the sexually active woman the best protection from human papillomavirus, the primary cause of cervical cancer.

Background

Over the 40 years since oral contraceptives (OCs) were first marketed, they have symbolized modern contraception and have remained the most widely used hormonal method worldwide. OCs provide millions of women with effective, convenient, and safe protection from pregnancy.

OCs also have been the most studied of any family planning method. The study of OCs continues, with new epidemiologic research reported from around the world.

First introduced in 1960, the pill was the leading contraceptive in the US by 1965 (150). By 1970 an estimated 8 to 10 million US women were using the pill, as were an equal number in other developed countries (355, 434).

In developing countries OCs began to appear in the mid-1960s, but high prices put them beyond the reach of all but a few women (378, 434). In 1967 international donor organizations, led by the US Agency for International Development (USAID) and the Swedish International Development Authority (SIDA), began to make OCs available to developing country governments and international family planning organizations (361, 379). Family planning programs in the developing world then began to supply more women with OCs. By the early 1970s an estimated 20 to 30 million married women in developing countries used OCs (434).

OCs remain popular today. With more than 100 million users, OCs trail only voluntary sterilization and IUDs in worldwide use among married women. Among sexually active unmarried women OCs are the most widely used modern method of family planning.

When the pill was introduced, it satisfied women's need for convenient, safe, and reliable contraception. There were some problems, however. Some pill users experienced such side effects as headaches, nausea, cramps, irregular menstrual bleeding, breast tenderness, or weight gain. These side effects usually are temporary and are not signs of more serious problems. They can be troubling, however, and have led many women to stop using the pill. Also, research in the 1960s and 1970s suggested that estrogen, in the doses used in early OCs, increased the risk of blood clots, stroke, and heart attack (396, 399, 496). Press reports about these findings created repeated "pill scares" and gave OCs an unwarranted aura of danger (150).

Meanwhile, studies found striking evidence of important noncontraceptive benefits of OC use. Most notably, epidemiological studies in the 1980s demonstrated that OCs provide strong protection against endometrial and epithelial ovarian cancer (see Protection from Some Cancers).

Since their introduction, OCs have offered safe contraception for the great majority of women. Still, to reduce common side effects and to minimize the risk of any serious complications, pharmaceutical companies and health care providers have used three approaches:

• To lower the doses of both estrogen and progestin without compromising effectiveness;
• To develop different new progestins;
• To screen women more specifically and to inform them about side effects that they may experience with OCs.

Lauren Goodsmith

A Moroccan health worker hands packets of pills to a client. In Morrocco pill use has risen substantially, from 14% of married women in 1980 to 32% in 1995. With more than 100 million users worldwide, oral contraceptives remain one of the most popular family planning methods.

The reduction of estrogen doses followed early research that related the likelihood of thromboembolic disorders to the size of the estrogen dose (214). US clinical trials found that estrogen doses as low as 20 µg, combined with a progestin, usually limit pregnancy rates to less than 1 per 100 women per year (27, 28, 39, 141, 209, 248, 400, 404, 499, 543). Also, side effects such as nausea, vomiting, cramps, breast discomfort, and headache occurred less often with less estrogen. Initial menstrual bleeding irregularities are more frequent, however (16, 119, 262, 369).

The progestin doses in OCs vary widely because progestins differ greatly in potency by weight (121). Currently, doses of progestins in the norethindrone family—norethindrone, norethindrone acetate, ethynodiol diacetate, and lynestrenol—range from 0.4 to 2 mg. Pills containing the more potent progestins levonorgestrel, desogestrel, and gestodene use doses of 0.05 to 0.15 mg. The different progestins have somewhat different physiological effects and interact differently with estrogens, possibly modifying the effects of both hormones (48, 467).

Research suggests that lower doses do lower risks for some conditions. For example, as lower-dose pills have come into wider use, findings from epidemi-ologic studies suggest that risks of OC-related venous thrombosis, heart attack, and stroke have declined. Significantly increased risk of heart attack and stroke is limited to women over age 35 who smoke or women who have high blood pressure (see Circulatory System Diseases).

RHRU/Baragwanath Hospital

Progestin-only contraceptive pills are a good option for breastfeeding mothers who want oral contraception. This South African brochure explains how progestin-only pills work, their side effects, and advantages and disadvantages.

Multiphasic Combined OCs

The doses in multiphasic OCs change during each pill cycle to keep hormone doses low (78). Like other low-dose OCs, multiphasics appear to provide highly effective contraception when taken correctly. Some clinical trials have observed that multiphasics produce minimal breakthrough bleeding, spotting, and amenorrhea (14, 125, 154, 349). There is little evidence that risks of serious health problems are less with multiphasics than with other low-dose OCs (437).

Also, as many as half of all new users stop using the pill within a year, and many use the pill intermittently for a few months at a time. While some may not need continuous contraception, this discontinuation rate suggests that many women are having difficulties taking the pill regularly or are dissatisfied. Just as researchers and providers have concentrated on making pill use safer, providers need to focus their attention on making pill use easier. The next issue of Population Reports will discuss this need and how it can be met.

Oral Contraceptive Use Worldwide

Oral contraceptives deserve close and continuing attention. Even though newer contraceptives have become available, in most countries, OCs remain among the most popular methods, and in many countries OCs are the most widely used method of all.

In surveyed countries nearly one married woman in every four who has ever used contraception has relied on the pill at some point in her life. Currently, more than 100 million women use OCs. Data both on ever use and on current use of contraception demonstrate the continuing popularity of OCs.

In some countries pill use has been very common. In Brazil nearly 80% of married women have used the pill at some point, as have two-thirds of married women in Costa Rica, Morocco, and Zimbabwe. Between 50% and 60% of all married women have used the pill in diverse countries including Bangladesh, Botswana, Cape Verde, Colombia, the Dominican Republic, Jamaica, Nicaragua, South Africa, Thailand, and Trinidad and Tobago. Among developing regions, the pill has been most widely used in Latin America, where 55% of all married women have used the pill at some time. More than one-third of married women in the Near East and North Africa have used the pill, while not quite 15% have used it in sub-Saharan Africa (see Table 1).

Similarly, many sexually active unmarried women have used the pill. In 12 of 28 countries with surveys, more of these women have used the pill than any other modern family planning method. Overall, in these countries 52% of wo-men who have ever used family planning have relied on the pill at some point—39% of all sexually active unmarried women. In Bolivia, Colombia, the Dominican Republic, Guatemala, Nicaragua, and Zimbabwe, between 50% and 60% of sexually active unmarried women have used the pill.

Experience with the pill is probably even more common in developed countries than in developing countries, although data on ever use are not available for many developed countries. In Canada 86% of women surveyed in 1995 had used the pill (38). In the US 80% of all women born since 1945 have used the pill, according to a 1990 estimate (106). Perhaps the highest level of experience with the pill is among German women: For example, 94% of eastern German women ages 30 to 44 have taken the pill (261).

OCs are the top modern method among married women in 78 of 150 countries with available data and, if China and India are omitted from the world estimate, the most widely used contraceptive method overall. Outside China and India about 12% of married women use the pill. By comparison, 9.5% rely on female sterilization, and almost as many use traditional or folk methods. About 9% use IUDs. Table 2 presents OC use data with and without China and India.)

Considering developing countries alone, 6% of married women use the pill—far fewer than use female sterilization, at 21%, and the IUD, at 13%. When China and India are removed from the use estimate, however, OCs become the most popular method in developing countries, used by 10% of married women compared with 9% relying on female sterilization and 8% on traditional methods. IUD use falls to 7%.

In developed countries OCs are the most widely used method. Some 16% of married women use the pill. Condoms and the IUD tie for second place at 14%, while slightly less than 14% of married women use traditional or folk methods.

Worldwide, among sexually active unmarried women, OCs are even more widely used than among married women. In countries with available data 26% of sexually active unmarried women use the pill. Data on current contraceptive use among unmarried women are available in Africa, Eastern Europe, Latin America, the Caribbean, and most developed nations. OCs are the most popular method among unmarried women in Latin America, North America, and Northern and Western Europe.

Patterns in use of family planning methods vary considerably within and among countries and regions. Differences in availability, access, cost, promotion, program policy, as well as people's preferences, help to explain these differences. Indeed, exceptionally high rates of use for any one method can suggest that access to other methods may be limited.

Near East and North Africa. In this region nearly 10 million women use OCs—13% of the region's 74 million married women. Three of every 10 family planning users are pill users. In Algeria, Iran, Kuwait, Morocco, Oman, Qatar, and the United Arab Emirates, OCs are the most widely used method.

In Algeria 44% of married women relied on the pill in 1995—the highest level of pill use in the developing world, accounting for 84% of all contraceptive use. Iran and Kuwait also report high levels of pill use, at 23% and 24% of married women. In contrast, levels of OC use are low—4%—in Turkey, where overall contraceptive use is 64%, and in Yemen, where overall use is about 21%.

In Morocco pill use has risen substantially, from 14% of married women in 1980 to 32% in 1995. In Egypt, however, OC use fell from 17% of married women in 1980 to 10% in 1995. In the 1990s many Egyptian women shifted to IUDs (188), which have been more promoted. (See Figure 1.)

Latin America and the Caribbean. In Latin America and the Caribbean, OCs are the second most widely used method among married women, following female sterilization. About 14% of married women use the pill—nearly one in every seven married women, or one in every five family planning users.

Some of the world's highest levels of current OC use, as well as ever use, are found in Latin American countries. For example, in Brazil 6 million women use the pill—in numbers fourth in the world after China, Germany, and Indonesia. Some 21% of married women used OCs in 1996, over one-fourth of all Brazilian women using family planning. Brazilian women overwhelmingly use either the pill or female sterilization. Access to the IUD is limited, and a 1994 rating of access to family planning methods found that the condom was readily accessible to less than 80% of couples in Brazil (73).

In some Latin American countries contraceptive use shifted from the 1980s through the mid-1990s. Smaller percentages used OCs as the use of female sterilization and IUDs grew. Overall, OC use rates dropped from one in every six married women in 1987 to one in every seven in 2000. In Colombia, for example, prevalence of OC use fell from 17% of married women in 1978 to 13% in 1995 (see Figure 1). In Mexico OC use declined from 14% in 1982 to 8% in 1995. At the same time, overall contraceptive use increased from 46% to 72% in Colombia and from 48% to 67% in Mexico. Researchers studying Mexican data conclude that the shift from OCs to other methods is occurring because more family planning users are older women choosing IUDs and sterilization once they have had all the children they want (519).

In Latin America the pill is the most popular way for sexually active unmarried women to avoid pregnancy. One-fourth of sexually active unmarried women use the pill. Use is highest in Brazil, at 36%

Sub-Saharan Africa. Pill use accounts for about one-quarter of all contraceptive use among both married and unmarried women in sub-Saharan Africa. Overall, about 15% of married women use family planning, and slightly less than 4% use the pill. Among sexually active unmarried women, about 43% use some contraceptive method, and 10% use the pill.

In some African countries levels of OC use are among the world's highest. For example, in Zimbabwe 33% of married women and 32% of sexually active unmarried women were using OCs in 1994. In Zimbabwe access to the pill is generally good, while access to the IUD and to female and male sterilization is considerably more difficult (73). Also, in Reunion 40% of married women use OCs, in Mauritius 21%, and in Botswana and Cape Verde about 18%. Among sexually active unmarried women, about 17% in both Mali and Niger are using the pill, and about 20% in Botswana and South Africa. Nevertheless, high levels of pill use are more the exception than the rule in sub-Saharan Africa. In five countries of the region, 1% or less of married women use the pill. In another eight countries use is between 1% and 2%.

JHU/CCP

In Bangladesh a village health worker discusses family planning with a client. In this country oral contraceptives have become the most widely used contraceptive, used by about 21% of married women.

Prevalence of OC use is highest in southeast Asia, led by Thailand, where an estimated 27% of married women of reproductive age used OCs in 1993. Despite the low percentage who use the pill, China has more pill users than any other country—about 7.6 million. Indonesia, the world's fourth largest country, has 6.1 million pill users. Few data are available about contraceptive use among unmarried women in Asia.

In a few Asian countries OC use among married women has increased considerably in recent years. In Bangladesh, for example, OCs have become the most widely used contraceptive method, taken by nearly 21% of married women of reproductive age in 1996–97 compared with 3% in 1983 (see Figure 1). Pill use also has grown recently in Sri Lanka and Vietnam.

Eastern Europe and Central Asia. In Eastern Europe and Central Asia 65% of married women use family planning, but only about 6% use the pill. IUDs, traditional methods, and condoms are more widely used. In 8 of 14 countries surveyed in the 1990s, levels of OC use were 5% or lower. Hungary is an exception; some 33% of married women use the pill. OC use is lowest in the Central Asian Republics. Among sexually active unmarried women in the region, just 6.5% use the pill, while larger percentages rely on condoms or traditional methods.

Other regions. In most developed countries the majority of married women rely on either the pill or female sterilization. In developed regions outside Eastern Europe and Central Asia, the pill ranks first, used by 21% of married women. In Western Europe half of all married women are pill users. In North America female sterilization is the choice of 24% of married women. The pill ranks second at 16%.

Japan reports the world's lowest rate of OC use. In 1994, when the most recent survey was taken, less than 0.5% of married women relied on the pill. Low-dose OCs were approved for contraceptive use in Japan only in September 1999. Before this date medium- and high-dose pills were available but only to treat menstrual disorders (276).

Some of the highest levels of pill use in the world are among sexually active unmarried women in developed countries. Outside Eastern Europe and Central Asia, 36% of sexually active unmarried women in developed countries use the pill. Some 45% in Europe use the pill; in North America, 36%.

For country-by-country statistics on OC use, see the internet website of the Johns Hopkins Center for Communication Programs at http://www.populationreports.org/a9/a9suptab.stm.

Benefits of Oral Contraceptives

Oral contraceptives have substantial benefits for women's health. The most important benefit, of course, is highly effective protection against pregnancy. OCs also help prevent ectopic pregnancy (pregnancy outside the uterus) and, by reducing menstrual blood loss, OCs lower the risk of iron deficiency anemia. In addition, they help to protect women from epithelial ovarian cancer and endometrial cancer and also may reduce the risks of bone density loss, ovarian cysts, benign breast disease, and colorectal cancer.

• Effectively prevent unwanted pregnancy, and
• Prevent ectopic pregnancy

A World Health Organization (WHO) study that compared six combined OCs containing 20 to 50 µg of estrogen found no significant differences in effectiveness (533). Generally, among perfect users (women who miss no pills and follow instructions exactly), only 1 in every 1,000 women becomes pregnant in the first year of use (189). Among typical users, however, about 60 to 80 women in every 1,000 will become pregnant during the first year (189, 312).

Progestin-only pills are slightly less effective than combined pills, except for breastfeeding women, for whom they are at least as effective (see sidebar, Progestin-Only OCs for Breastfeeding Women). Still, among perfect users of progestin-only pills who are not breastfeeding, only 5 pregnancies per 1,000 women occur in the first year of use. Typical-use rates are not available (190, 297). While progestin-only pills do not prevent ovulation in about half of menstrual cycles, other effects of progestins provide contraceptive protection. In particular, progestins thicken cervical mucus, making it more difficult for sperm to pass through the cervical canal (146, 297, 498).

Some have speculated that multiphasics might be more likely to allow pregnancy if not used correctly, including taking them out of order (99, 134, 167, 239, 482, 512). Most clinical trials find no difference in effectiveness between multiphasics and constant-dose combined pills, however (124, 142, 371).

Some women forget pills or stop them for a time. This largely accounts for the gap in OC effectiveness between perfect users and typical users. Irregular pill-taking may explain why users of oral contraceptives sometimes experience higher pregnancy rates than users of injectables, IUDs, or implants. According to Demographic and Health Survey data in 15 developing countries in the 1980s, the pregnancy rate among OC users was about 6 per 100 per year—twice the pregnancy rate among IUD users, at 3 per 100 (312).

Long-term effectiveness of the pill requires sustained correct and consistent use. A recent review of 53 reports on contraceptive effectiveness concluded that on average about 7% of OC users are likely to become pregnant in the first three years of use, but the percentage varies depending on whether women take the pill correctly. Among the generally consistent and conscientious users, 3.8% would become pregnant within three years. In contrast, among those who use the pill inconsistently and incorrectly, 7.8% would become pregnant within three years (365).

Preventing ectopic pregnancy. Protection against ectopic pregnancy is a benefit of all contraceptive methods, to varying degrees. Because they consistently stop ovulation, all combined OCs very effectively prevent ectopic pregnancies (117, 336). Ectopic pregnancy, which occurs when a fertilized ovum develops outside the uterine cavity, can be life-threatening (181). Ectopic pregnancy is fairly common. One US study found that ectopic pregnancy was the reason for 1 in every 13 emergency room visits during the first trimester of pregnancy (446). In the US ectopic pregnancy is the leading cause of pregnancy-related death in the first trimester. In 1992 ectopic pregnancies accounted for 2% of reported pregnancies and 9% of all pregnancy-related deaths in the US (478).

• Less iron deficiency anemia, due to lighter menstrual bleeding,
• More regular menstrual cycles,
• Less dysmenorrhea, and
• Less severe premenstrual symptoms.

Because of higher blood iron levels, OC users are less likely than nonusers to develop iron deficiency anemia (147, 328, 466). Also, by preventing unwanted pregnancies, OCs—like other contraceptives—prevent anemia associated with pregnancy (484). In developing countries anemia is a serious health problem among women, many of whom suffer from inadequate diets, parasitic infections, and the strain of repeated pregnancies. As many as half of women of reproductive age in developing countries may have subnormal levels of hemoglobin, the iron-containing pigment of red blood cells (532, 536).

Some 60% to 80% of women who use OCs bleed less heavily during menstruation than before starting OCs, and on average OC users lose 50% to 60% as much blood per cycle as other women (147, 260, 302, 328, 382, 433, 438). A 1992 Swedish study found that a low-dose OC, containing 30 mg ethinyl estradiol and 0.15 mg desogestrel, reduced menstrual blood loss to 56% of previous levels (260).

More regular menstrual cycles. Oral contraceptives generally improve menstrual patterns (179). For example, a UK study of 2,115 women ages 18 to 49 found that most OC users had shorter, lighter periods that occurred at more regular intervals (49). Only 7% of OC users reported irregular periods, compared with 10% of IUD users, 11% of women relying on female sterilization, and 12% of women using other methods or none. The Oxford University/Family Planning Association (Oxford/FPA) cohort study found that, compared with nonusers, OC users or recent OC users (within the previous 12 months) were two-thirds as likely to be referred to a hospital for treatment for irregular periods (495).

IEC, Center, State Information Service, Egypt Ministry of Information

This Egyptian poster reminds pill users, "Don't forget to take a pill every day. If you forget, take it with the pill of the following day."

OCs are a standard treatment for dysmenorrhea (50, 189). A 1990 Swedish study found that users of both monophasic and triphasic low-dose OCs had less severe dysmenorrhea than nonusers (309). Combined OCs appear to be more effective than progestin-only pills at reducing dysmenorrhea (68).

Less severe premenstrual symptoms. Several studies have found that premenstrual symptoms are less severe among OC users than among other women (20, 100, 166, 253, 311, 341, 427). Differences in defining and measuring symptoms, however, make it difficult to compare the effects of different formulations (11, 294, 308, 309). Premenstrual syndrome, a condition caused by natural hormonal changes, begins at the middle of the menstrual cycle and tends to intensify in the last seven days before menstruation. Multiple physical and/or emotional symptoms characterize premenstrual syndrome, such as headache, fatigue, acne, backache, breast soreness, changes in sexual desire, nervousness, difficulty concentrating, irritability, anxiety, and depression. Symptoms subside when a woman begins to menstruate. Most women experience noticeable premenstrual symptoms at some time; 10% or less report severe discomfort (50, 225).

• Endometrial cancer (cancer of the lining of the uterus) and
• Epithelial ovarian cancer.

Combined OCs probably help protect against these cancers by reducing the rate of cell division in the endometrial lining and the ovaries. In the case of the uterine endometrium, the progestin component in the pill is thought to counteract the effects of estrogen, which would otherwise encourage cell division. OCs may protect against ovarian cancer by reducing gonadotropin production by the pituitary gland, thus reducing the effects of gonadotropin stimulation of the surface cells of the ovaries (62, 359).

Endometrial cancer. Even as little as one year's use of combined OCs cuts the risk of endometrial cancer substantially, and protection lasts long after women stop using OCs. A combined analysis of eight case-control studies and two cohort studies found that longer use significantly increased protection (409, 535). One year of OC use reduced risk to 77% of that among nonusers, 2 years to 62%, 4 years to 49%, 8 years to 36%, and 12 years to 30%. Earlier studies reported protection persisting from 3 to 10 years (195, 210, 234, 516.)

It is uncertain whether the degree of protection against endometrial cancer varies with estrogen and/or progestin dose. The 1985 US Centers for Disease Control's Cancer and Steroid Hormone (CASH) study found no relationship between progestin dose and the degree of protection (58). Although the number of women using any one formulation in the CASH study was too small to allow an analysis by formulation, both high- and low-dose pills had a protective effect. In contrast, a 1991 WHO study suggested that protection was greater for users of formulations containing high doses of progestins (393, 460). Although there are no studies on whether progestin-only OCs protect against endometrial cancer, studies of the effects of progestins on the endometrium suggest that it is progestin rather than estrogen that confers the protective effect (243, 297, 360). Moreover, a 1991 WHO study of the progestin-only injectable contraceptive depot medroxyprogesterone acetate (DMPA) found that it protected against endometrial cancer as well as combined OCs (461). Therefore progestin-only OCs may have at least some protective effect (297).

Epithelial ovarian cancer. Combined OCs help protect against epithelial ovarian cancer (12, 59, 62, 92, 200, 299, 325, 391, 474, 515, 524). This finding from the large 1985 CASH study and many earlier, smaller studies has been confirmed over the past decade (180, 345, 386, 390).

In the CASH study women using OCs for 10 years or more reduced their risk of ovarian cancer to 20% of that among nonusers. The CASH study also found that protection against epithelial ovarian cancer persists long after women stop using OCs. Even women who had stopped using OCs 15 or more years earlier faced just half the risk that never-users faced. Each of the 11 pill formulations studied offered similar protection, whether the formulation was high- or low-dose (59).

The protective effect of OCs against epithelial ovarian cancer may grow in importance in the coming years. All studies to date have focused on women younger than 55, since most OC users and former users are in this age group. Ovarian cancer is more common in women over age 60, however. Since the protective effect of OC use apparently persists for many years, widespread OC use may eventually result in a decline in the incidence of this frequently fatal disease. Epithelial ovarian cancer is by far the most common type of ovarian cancer (59).

• Loss of bone density,
• Ovarian cysts,
• Benign breast disease, and
• Colorectal cancer

It has not been demonstrated that the effect of OCs on bone density makes a practical difference. Neither of the two major British cohort studies, the Royal College of General Practitioners study or the Oxford/FPA study, found that pill use helped to protect premenopausal women from bone fractures (87, 492).

Ovarian cysts. Several early studies indicated that high-dose OCs—those containing 50 mg or more of estrogen—protect women from functional ovarian cysts (334, 397, 493). The Oxford/FPA cohort study found that the risk of follicular ovarian cysts in current OC users was about half that in users of nonhormonal methods. The protection from corpus luteum cysts was even greater. Users of combined OCs faced about one-fifth the risk that other women faced (493). Low-dose combined and multiphasic OCs, even though they prevent ovulation effectively, may permit some follicular development and thus offer less protection against cysts (155, 436) or perhaps none at all (74, 207, 258).

Benign breast disease. Studies of women using older, higher-dose formulations found that OC use protected against fibroadenoma and fibrocystic breast disease. OC users had from one-quarter to one-half the risk of nonusers (45, 335, 398).

Protection against benign breast disease may depend on the progestin content of the pill, with more progestin offering more protection. The Oxford/FPA cohort study compared women using pills with the same amount of estrogen but with different amounts of the same progestin. Women using pills containing 2.5 or 3.0 mg of the progestin norethindrone acetate had half the incidence of fibrocystic breast disease as women who used pills with 1.0 mg norethindrone acetate. Also, protection increased with length of pill use (45). Since most OCs now in use contain lower amounts of progestin than in this study, they may offer less protection against benign breast disease (29, 265).

Colorectal cancer. Some studies have found that OC use reduces the risk of colorectal cancer (26, 135, 148, 293, 364). The largest case-control study to date found that women who had ever used OCs reduced their risk of colorectal cancer to 60% of that of nonusers and that OC use for over two years reduced risk to 50% (135). Other studies, however, have found no protective effect (252, 468, 514). Colorectal cancer is the fifth most common cancer among women worldwide (348, 542).

Health Risks of Oral Contraceptives

Modern oral contraceptives are safe for the great majority of women. The health risks of using OCs are much less than the risks of pregnancy and childbearing for almost all women, especially in countries with high maternal mortality rates. Even where maternal mortality is low, pill use is safer than childbearing except for older women who smoke or have high blood pressure (130, 332, 337). Today, with the lower doses in modern pills, the risks of a number of medical conditions appear to be lower than in the past. Also, recent large studies have made it possible to assess the health risks of long-term OC use more accurately and to better identify the groups most likely to experience them.

A major finding of the last decade is the increased risk of heart attack and stroke for older OC users with hypertension. For OC users who do not smoke and do not have high blood pressure, however, the low doses in today's pills appear to minimize these risks.

The major established health risks of OCs are certain circulatory system diseases, particularly heart attack, stroke, and venous thromboembolism. Other health risks include gallbladder disease in women already susceptible to it and rare noncancerous liver tumors. In addition, users and providers of OCs should be aware of possible interactions between OCs and other drugs that might make OCs less effective or modify the effects of the other drugs.

Jamaican National Family Planning Board

Modern oral contraceptives are safe for the great majority of women. The health risks of using OCs are much less than the risks of pregnancy and childbearing for almost all women, especially in countries with high maternal mortality rates. This Jamaican poster points to high effectiveness and few side effects with low-dose oral contraceptives.

Circulatory System Diseases

Heart attack. Ischemic heart disease results from an impediment to circulation that deprives the heart of adequate blood supply. Myocardial infarction—heart attack—is the resulting death of heart muscle cells. Ischemic heart disease can develop gradually from atherosclerosis, in which deposits on the walls of coronary arteries restrict blood flow to the heart muscle, or it can result from a thrombus, or clot, that suddenly blocks a coronary artery. Myocardial infarction is rare in young women who do not smoke or have other clinical risk factors (122, 463, 540).

Using OCs may somewhat increase heart attack risk, but risk is largely limited to older women who smoke or have high blood pressure. Early case-control studies found that the risk of myocardial infarction (heart attack) in current OC users was two to four times greater than in nonusers (286, 287, 288, 290, 431). Recent studies suggest risk less than twice that of nonusers. For example, a recent US case-control study found that OC use increased risk by about 1.7 times. In this study an estimated 5% of myocardial infarction cases were attributable to current OC use, the equivalent of less than three additional heart attacks per 1 million US women in one year of OC use (425). Two other recent case-control studies, one in the US (389) and one in the UK (465), found no significantly increased risk among either current or past OC users. The findings of lower risk may be attributable to use of lower-dose OCs and perhaps also to better screening of potential pill users for risk factors such as smoking (463).

Combining OC use with other risk factors for heart disease, particularly smoking, hypertension, and long-term or uncontrolled diabetes, raises the risk of myocardial infarction substantially (288, 353, 463). Relative risk* increases considerably in pill users who smoke (95, 163, 353, 388, 415, 463). Between 1989 and 1993 the World Health Organization (WHO) conducted a large case-control study of cardiovascular disease and hormonal contraceptives in 21 hospitals in Europe and in developing countries of Africa, Asia, and Latin America. OC users with histories of hypertension or who were heavy smokers were at a much greater risk of heart attack relative to OC users without either of these risk factors (see Table 3).

*Relative risk is a measure of how much a particular factor influences the risk of a specified outcome. For example, a relative risk of 2 associated with a factor means that people with the factor face twice the risk of having a specified outcome that people without the factor have. A relative risk of 0.5 means that people with the factor face half the risk of the specified outcome that people without the factor have (a protective effect).

Among nonsmoking OC users who had their blood pressure checked before starting OCs and had no other risk factors for heart attack, there was no appreciable increase in risk in either European or developing countries (539, 540). The researchers conclude that:

Women who do not smoke, who have their blood pressure checked, and who do not have hypertension or diabetes are at no increased risk of myocardial infarction if they use combined oral contraceptives, regardless of age. There is no increase in the risk of myocardial infarction with increasing duration of use of combined oral contraceptives. There is no increase in the relative risk of myocardial infarction in past users of oral contraceptives. These conclusions appear to apply equally to women in developed and developing countries. (540)

Recent studies have suggested no increased risk of myocardial infarction among users of third-generation OCs with no other risk factors (222, 268, 270). The rarity of myocardial infarction in women of reproductive age, the large number of different OC formulations, and the differing geographic distributions of users of specific formulations will make it difficult to compare various formulations in detail, however (353).

Recent studies—and most older ones—have found that heart attack risk does not increase with duration of OC use. Also, risk does not persist after a woman stops using OCs (3, 10, 95, 104, 288, 289, 375, 415, 425, 431, 442, 443, 465, 539).

Planned Parenthood Federation of America

Smoking increases the risk of heart attack and stroke especially for older women who use oral contraceptives. This US poster warns of the risk and urges pill users not to smoke.

Studies with low-dose OCs in the 1980s and 1990s suggest less overall risk of stroke than did earlier studies (464). The first studies of the health risks of OC use, conducted in the 1960s and 1970s, suggested about a fivefold greater risk of any type of stroke among OC users than among nonusers (81, 213, 224, 263, 264, 355, 375, 441, 490). Now, based on the results of a multi-center WHO study (367) and other recent studies (192, 354, 412), the estimated risk of ischemic stroke among OC users is about 2.5 times greater than the risk among nonusers (131). The more recent studies provide more information on ischemic stroke than on hemorrhagic stroke.

Lower doses appear to have reduced the risk. For example, in the Oxford/FPA cohort study, the relative risk of stroke for OC users appeared to drop as the study progressed. In 1984 the study reported relative risks of 1.5 to 2.0 for subarachnoid hemorrhage and 2.3 to 3.2 for other types of stroke among OC users compared with nonusers (497), down from 5.0 for all types in 1976 (490).

The multicenter WHO study—the largest case-control study of stroke and OCs by far—found an overall relative risk of ischemic stroke of about 3 among OC users. As with heart attacks, other risk factors make a big difference to the risk of OC use (see Table 4). Current OC users who did not smoke, had their blood pressure checked, and did not have high blood pressure were at 1.5 times greater risk than nonusers. In contrast, OC users who smoked faced a higher risk—about two times greater than among nonsmoking OC users in the developing countries studied and about 3.5 times greater in Europe. Current OC users with a history of hypertension faced the greatest risk—about five and four times greater than for other OC users. As in the WHO study of heart attack, the risk of ischemic stroke was lower among women who reported having their blood pressure checked before starting OCs than among those who did not (367).

The WHO study produced conflicting findings on the relationship between dosage and ischemic stroke risk. In Europe lower doses meant lower risks, while in the developing countries the pattern was reversed. The researchers suggest that the opposing patterns reflect different levels of other risk factors for ischemic stroke. In both Europe and the developing countries, however, risk did not rise significantly with continuing OC use, and elevated risk did not appear to persist after women stopped using OCs (367).

Migraine headaches have been linked to a twofold or greater increased risk of ischemic stroke. Several studies have found that OC users with a history of migraine are two to four times more likely to have an ischemic stroke than women with a history of migraine who do not use OCs (69, 272, 273, 274, 411, 475). For example, a case-control study conducted in European hospitals from the WHO study of cardiovascular disease and hormonal contraceptives found that, compared with women not using OCs and having no history of migraine, ischemic stroke was 6.6 times more likely among OC users with a history of migraine, and 2.3 times more likely among nonusers with a history of migraine (69). Studies suggest risk is greater among women who have severe migraine headaches with "aura"—focal neurologic symptoms such as blurred vision, temporary loss of vision, seeing flashing lights or zigzag lines, or trouble speaking or moving (61, 69, 411, 475). The recent studies (61, 69, 273, 475) led a March 2000 meeting of experts convened by WHO to recommend that a woman who has migraine headaches with focal neurologic symptoms should not start combined OCs. The group recommended that a woman age 35 or older should choose another method if possible if she has migraine headaches even without focal neurologic symptoms. Mild or severe headaches that are not migrainous do not rule out OC use (557).

For hemorrhagic stroke, the WHO study found a slightly increased risk among OC users in general (see Table 5). The difference was statistically significant in developing countries but not in the European countries. In both developing and European countries, current OC users age 35 or over had a significantly increased risk of hemorrhagic stroke, with relative risks of 2.5 and 2.2, respectively, compared with nonusers of OCs. The relative risk among current OC users who smoked was three to four times that of nonusers who did not smoke. Also, current users with a history of hypertension faced a substantially higher relative risk than nonusers with no such history. As with ischemic stroke, the duration of OC use did not affect the risk of hemorrhagic stroke, and the elevated risk did not persist after OC use ended. Risk did not vary with estrogen dose or with progestin dose or type (537, 540).

Because, in combination, hypertension and OC use increased the risk of stroke and myocardial infarction far more than would either risk factor alone, WHO medical eligibility criteria for OC use were recently revised to recommend that women who know that they have high blood pressure—systolic pressure of 140 mm Hg or higher and/or diastolic pressure of 90 or higher—or, where blood pressure cannot be evaluated, who have a history of hypertension should choose another contraceptive method (557). (Blood pressure must be properly taken, and one reading is not enough to diagnose high blood pressure.)

In areas where medical services are limited, blood pressure checks for OC users may be impractical. In these service areas, maternal mortality and morbidity tend to be greater risks than any risks associated with OC use (540).

In any case, the benefit of screening potential combined OC users for high blood pressure and withholding OCs from women with high blood pressure would not be substantial if hypertension itself cannot be treated. A recent analysis conducted for WHO pointed out that this screening would prevent only about 10% of stroke and heart attack cases attributable to OC use. In particular, screening women under age 35 for high blood pressure would not prevent an appreciable number of cardiovascular disease cases or deaths attributable to OC use. At the same time, "false positive" diagnoses of hypertension would needlessly prevent some women from using OCs (556).

From a public health perspective, the impact of cardiovascular disease attributable to OC use is slight, particularly since most OC users are young and do not have other risk factors for cardiovascular disease. The analysis for WHO points out that the additional number of cardiovascular disease cases and deaths among OC users depends greatly on age. For example, among one million OC users under age 35, 20% of whom smoke, fewer than 20 deaths annually would be due to OC use. By comparison, among one million OC users over age 35, 20% of whom smoke, 24 to 96 deaths annually would be attributable to OC use, depending on the region (556).

Thromboembolism. Thromboembolism is an obstruction of a blood vessel by a blood clot. The most common thrombo-embolic disorder in OC usersÅknown as venous thrombo- embolism (VTE), or deep vein thrombosisÅinvolves clots that form in veins deep in the leg. These clots sometimes circulate to the lungs, where they become potentially fatal pulmonary embolisms. A multinational study by WHO, conducted from 1989 to 1993 (366), and three smaller studies in the mid-1990s (34, 221, 440) found that modern low-dose OCs pose less risk of thromboembolism than indicated by earlier studies that involved mostly first-generation pills.

The new case-control studies reported a risk of VTE about three times greater among users of second-generation OCs than among women not using OCs and about six times greater among users of third-generation OCs containing desogestrel and gestodene (131). There is debate about whether the difference in risk between second- and third-generation pills is real or due to bias in the studies (88, 131, 132, 194, 223, 267, 269, 275, 368, 440, 447, 503, 540, 554).

When the findings were released in 1995, they caused a "pill scare" in the news media of the UK and other European countries where third-generation OCs are most widely used. Responding to the publicity, many women switched to other OCs or stopped taking pills altogether. In the months following, the number of unintended pregnancies and abortions increased substantially (15, 75, 76, 118, 218, 257, 376, 528).

The estimated risk of VTE is low with all modern low-dose OCs —including those containing desogestrel and gestodene—and all low-dose OCs pose less risk of VTE than previous higher-dose formulations. For users of high-dose OCs, early studies suggested that about 80 cases of VTE per 100,000 women per year could be attributed to OC use (197, 221, 358, 441). By comparison, recent studies estimate that the annual number of VTE cases attributable to second-generation OCs ranges from about 6.5 cases per 100,000 women per year at ages 20 to 24 to 12 cases per 100,000 per year at ages 40 to 44. The number attributable to third-generation OCs ranges from about 16 cases per 100,000 at ages 20 to 24 to 30 cases per 100,000 at ages 40 to 44 (131). VTE risk associated with pregnancy is about 60 cases per 100,000 pregnancies (86).

The risk of death from VTE is slight. Worldwide, among women not using OCs, an estimated 0.6 to 1.2 deaths per million women per year can be attributed to VTE. Based on recent case-control studies, an estimated 1.3 to 2.4 additional deaths per million women per year occur among OC users (131).

Increased blood pressure and hypertension. Many studies have found small but statistically significant increases in blood pressure in women taking combined OCs with 50 mg of estrogen or more (42, 137, 138, 300, 435, 510, 511, 525). Increases averaged about 6 mm Hg for systolic blood pressure and 2 mm Hg for diastolic blood pressure (417). Studies have reported generally comparable increases in blood pressure among users of low-dose combined OCs as well (162, 241, 320, 327, 525, 534).

In some women these increases are enough to lead to a diagnosis of hypertension (blood pressure of 140/90 or higher). Usually, however, blood pressure remains within normal ranges and declines once the woman stops using OCs (55, 71, 137, 241, 510). For example, a recent US cohort study of about 68,000 nurses ages 25 to 42 found that, after adjusting for other possible risk factors, OC users were almost twice as likely to develop hypertension as women who had never used OCs. Risk increased with duration of use but decreased rapidly after women stopped using OCs (71).

Carbohydrate metabolism and diabetes. Combined oral contraceptives may affect carbohydrate metabolism in two ways. The estrogen component is thought to increase glucose levels and suppress the insulin response to them. The higher the dose, the more effect (153, 156, 544). The progestin component has been hypothesized to stimulate overproduction of insulin, a suspected risk factor for atherosclerosis (96, 544, 545).

In low-dose OC users with initially normal blood sugar levels, these responses seldom exceed the normal range. These women face no apparent risk of developing diabetes (143, 152, 158, 171, 185, 385).

Can women with diabetes use combined OCs? Diabetics, whose insulin response to increases in glucose is already suppressed, may still be able to use low-dose OCs, depending on the severity of their diabetes. If they are insulin-dependent, their insulin requirement may increase, although with low-dose pills this does not appear to happen often (48, 305). Diabetics with known vascular disease or women who have had diabetes for over 20 years (and therefore may have suffered vascular damage) generally should choose another family planning method (557). Women with a history of diabetes during pregnancy or a family history of diabetes can safely use combined OCs without special medical supervision (48, 177, 189, 190, 228, 280, 538).

Gallbladder disease. OCs probably do not cause gallbladder disease, but instead they may accelerate the development of gallstones in already susceptible women. Gallstones are caused by abnormally high saturation of bile with cholesterol. Cholesterol saturation is higher in OC users than nonusers, possibly due to estrogen (432, 526).

After finding a higher risk during the early years of OC use, the major cohort studies did not detect any elevated risk of gallbladder disease among long-term OC users. The lack of long-term risk suggests an acceleration effect in women with already high cholesterol saturation (318, 375, 494, 526). An analysis of the results of several studies from the 1970s and early 1980s concluded that OC use is associated with only a slight increase in the risk of gallbladder disease (458).

Population and Community Development Asociation

In Thailand a community-based distributor who travels by boat gives a client a packet of pills. Oral contraceptives are well suited to community-based distribution and social marketing because of their safety and because using them involves no medical procedure.

Because of concerns that OCs may worsen existing gallbladder disease, WHO recommends that women with current symptoms of the disease should choose another method if possible (538).

Noncancerous liver tumors. Noncancerous liver tumors (hepatocellular adenomas), which are rare, are somewhat more frequent in OC users than in nonusers. They can be fatal if untreated (483). Their incidence increases with higher estrogen dose and longer OC use. Studies in the 1970s of women using higher-dose pills estimated that three cases attributable to OCs would occur per 100,000 users per year. With today's low-dose OCs, this rate may be lower (277), but new studies have not been conducted. (See Liver Cancer for discussion of liver cancer and OCs.)

Contraceptive hormones can interact with certain other drugs, reducing the effectiveness of OCs or modifying the effects of the other drugs. Pregnancies and breakthrough bleeding due to interference with contraceptive hormones have been reported in OC users taking:

• The anticonvulsant drugs carbamazepine, ethosuximide, methylphenobarbital, paramethadione, phenobarbital, phenytoin, primadone, and topirimate;
• The antitubercular antibiotic rifampicin; an
• The antitubercular antibiotic rifampicin; an

If a user of low-dose OCs is taking any of these drugs, she can increase her contraceptive protection by using an additional method of contraception while continuing with her daily pill, or she can change to an OC with 50 µg estrogen. If she uses the drug for less than a month, she should continue using her back-up contraceptive method or different pill regimen for at least a week after stopping the drug. If her cycle of 21 contraceptive pills ends during this week, for best protection she can start the next cycle of pills immediately. If she is using 28-day pill cycles, she can skip the seven placebos or iron tablets and start the next cycle of pills immediately (178, 189).

Women who must take any of these drugs for a long time, such as women being treated for tuberculosis, may want to consider another contraceptive method. Alternatively, they may increase their contraceptive protection by taking OCs with 50 µg estrogen for extended periods without interruption (179).

Oral contraceptives can speed up the metabolism of certain other drugs. Increased clinical effects have been observed in OC users taking anti-anxiety benzodiazepine drugs, corticosteriods used against inflammations, and theophylline (a drug used for asthma and some circulatory conditions). Thus OC users may require lower doses of these drugs than other women (18, 189, 418). Low-estrogen OCs generally speed metabolism less than high-dose pills.

Uresolved Health Issues

OCs have proved safe for most women. Still, several important health issues remain unresolved, even though the perspective of these issues has become clearer. Of particular concern are associations between OC use and neoplasia of the cervix and breast. Women who use OCs may have slightly elevated risks of being diagnosed with cervical neoplasia and early occurring breast cancer—risks which disappear within 10 years after discontinuing use. In both cases screening bias cannot be ruled out as an explanation for the apparent risk. The OC users studied tend to have more regular gynecological care than other women, and thus early cancer may be more likely to be detected in these women. OC use has been linked to an increased risk of certain reproductive tract infections, including HIV. While an association is biologically plausible, possible methodological problems make interpretation difficult. Finally, results of recent studies conflict as to whether OCs are linked to hepatocellular carcinoma, a rare form of liver cancer.

Family Planning Association of Kenya

Where treatment is available, cervical cancer screening can benefit all women. This Kenyan poster invites women for cervical screening.

Relative risks of pre-invasive lesions were for the most part less than 2.0—for example, 1.3 for ever-use (546), 1.4 for past users (169), and 1.8 for low-grade lesions (323). A Swedish study found current use of OCs to be associated with a fourfold increase in risk of pre-invasive lesions overall; risk increased with duration of use (547). One recent study, however, found no association between OC use and pre-invasive lesions (80). For women who had used OCs for five years or more, several recent studies report about a doubling of risk compared with women not using the pill (46, 47, 245, 546). Studies that have looked at various grades of pre-invasive lesions, however, have reported inconsistent findings (46, 251, 323).

Epidemiologic findings on invasive cancer. As with pre- invasive lesions, most studies in the past decade have found that long-term OC use is associated with a slight increase in risk of invasive cervical cancer (304). In many of these studies risk increased with duration of use (43). An analysis of 14 studies found relative risks of invasive cervical cancer to be 1.37, 1.60, and 1.77 for 4, 8, and 12 years of OC use (410).

Some evidence suggests that OCs accelerate the progression of precancerous lesions to invasive cancers. Any increased risk may be concentrated in current and recent users (26, 344). One study found that no increased risk persisted beyond 10 years after OC use ended (344). Another study found that OCs increased the risk of invasive cancer only when first used at a young age, especially at age 17 or younger, a crucial time in the development of a woman's reproductive tract (101).

Interpreting the findings. A number of biological mechanisms have been proposed to explain an association between OCs and cervical neoplasia. Currently, no firm evidence favors any one of these mechanisms. It has been suggested that OCs might: (1) promote the growth of existing lesions, (2) change cervical mucus to increase tissue susceptibility to HPV (491), (3) alter immune response to increase susceptibility to HPV, (4) produce a folate deficiency in the cervix that could stimulate development of abnormal lesions (144, 408), and/or (5) enhance genetic replication of HPV (43, 161, 183, 216, 350, 426).

Observed associations between cervical neoplasia and OCs may reflect the difficulties of studying the issue rather than causal relationships. First, OC use may be part of larger behavior patterns that also increase the risk of cervical cancer (430). Second, cervical neoplasia may be more easily detected in OC users than in other women (detection bias). These difficulties are hard to overcome completely. Furthermore, the biological factors that influence the development of cervical cancer are complex.

Behavioral factors. Studies of cervical cancer and OCs may need to take account of both sexual behavior and smoking. Sexual behavior, particularly age of first intercourse, lifetime number of sexual partners, and use of barrier contraception, are known to affect the risk of developing cervical cancer. Younger age at first intercourse and more partners raise risks. Condom use lowers risks. If women choose OCs because they start sexual activity early or have many sex partners, and they do not use condoms, studies would find a noncausal association between OC use and cervical cancer (64).

There is fairly strong evidence associating cervical cancer and cigarette smoking (523, 527). Several studies suggest about a twofold increase in risk for smokers compared with nonsmokers (46, 102, 168, 547). A Danish study suggested that OC users who smoke are at particularly high risk of cervical cancer. Among women using OCs for six years or more, smokers had a relative risk of 6.0 compared with 2.2 among nonsmokers (245).

J. Doug Story for JHU/CCP

This rimary health cetner in Nepal displays a wall painting near the entrance that promotes oral contraceptives. In many clinics worldwide, health care providers help clients choose the contraceptive method that best suits their individual needs.

Biological factors. Cervical cancer develops slowly. Invasive cancer apparently occurs at the end of a slow progression of pre-invasive lesions. But most mild, and many moderate, pre-invasive lesions regress spontaneously (93, 206). Very few progress to invasive cancer (321, 383). Risk factors for progression at each stage—and for progression to invasive cancer—may vary (44). Hypothetically, OC use could have an independent effect or act as a cofactor at any stage. Thus a causal link between OC use and pre-invasive lesions—if established—would not necessarily imply a link to invasive cancer (184, 346). Research needs to explore why some pre-invasive lesions progress while most regress, and what role, if any, OCs play in progression.

Because HPV is the primary cause of cervical cancer (see sidebar, Preventing Cervical Cancer), researchers have looked for a connection between OCs and the risk of acquiring HPV infection. Findings are mixed. Some studies have reported that OC users are significantly more likely to acquire or have an HPV infection (159, 271, 283, 323, 395, 459, 487). Others have not (24, 51, 198, 233, 279, 367, 426, 449, 489). Studies have not consistently controlled for sexual behavior, however.

HPV targets cervical cells that are actively dividing (244). OCs increase cervical ectopy—the extension of sensitive columnar epithelial cells from the cervical canal to the vaginal surface of the cervix. Thus it is possible that OCs could enhance susceptibility of the cervix to HPV infection (216).

Findings from the pooled analysis include:

• Overall, women currently taking OCs or who have quit use within the past 10 years were slightly more likely than nonusers to be diagnosed with breast cancer.
• Risk was greatest for current users and decreased with time between last use and diagnosis. Relative risk was 1.24 for current users, 1.16 for women who had stopped use within one to four years before diagnosis, and 1.07 for women who had stopped use five to nine years before diagnosis.
• There was no additional risk for OC users who discontinued use 10 to 20 years before diagnosis. In some subgroups former OC users faced less risk than nonusers.
• The excess risk of breast cancer diagnoses in OC users was solely for cancers that were localized. OC users actually had a reduced risk of cancers that had spread beyond the breast.
• Women who used OCs before age 20 faced somewhat higher relative risk, when compared with nonusers of the same age, than women who used OCs later in life.
• Whether a woman first used OCs before or after she first gave birth did not appear to make much difference.
• For women with a family history of breast cancer, OC use did not seem to increase risk particularly.
• Duration of OC use did not affect risk.
• Data were limited, but risk did not appear to be related to type of estrogen or progestin, and the only dose-related difference was a reduction in breast cancer among women who had used the highest dose pills more than 10 years before (82, 83).

The relation observed between breast cancer risk and hormone exposure is unusual, and it is not possible to infer from these data whether it is due to an earlier diagnosis of breast cancer in ever-users, the biological effects of hormonal contraceptives, or a combination of reasons. (82)

The finding that the modest additional risk is greatest during OC use and eventually disappears after a woman stops OCs has important public health implications (521). Because most women use OCs when they are young, and breast cancer is extremely rare at young ages, the number of breast cancer cases attributable to OC use would be quite small. By the Collaborative Group's estimate, among 10,000 European or North American women using OCs from ages 16 to 19, an additional 0.5 cases of breast cancer would be diagnosed in the 10 years after these women quit OC use; among those using OCs from ages 20 to 24, 1.5 additional cases; and among those using OCs from ages 25 to 29, 4.7 additional cases. Because of this age gradient, earlier OC use in a population does not lead to more cancers diagnosed overall (82). Generally, the numbers of additional cases would be smaller in developing countries, where breast cancer is less common (83). By 20 years after stopping OC use, there was no significant difference between women who used OCs at these ages and nonusers in the cumulative number of breast cancer cases diagnosed.

Population Services International

A parade float promotes social-marketed Nova brand oral contraceptives and injectables in Pakistan.

Chlamydial infection. Most recent studies find an association between use of OCs and infection with Chlamydia trachomatis, the most common STI (2, 4, 17, 35, 37, 97, 244, 284, 285, 331, 553). Older studies found a two- to threefold increase in risk of chlamydia among OC users (506). Some recent studies have found less increase, about 70% more than nonusers, and others have found no increase in risk (347, 374, 394, 509).

Greater risk of chlamydial infection among pill users could be due largely to cervical ectopy. Cervical ectopy is the extension of sensitive columnar epithelial cells from the cervical canal to the vaginal surface of the cervix. It is known to occur in OC users (219). Cervical ectopy may make columnar cells easier targets for C. trachomatis (17, 30, 94, 392). Several studies confirm a link between ectopy and chlamydial infection (284, 347, 374). These studies, however, could not determine which came first—ectopy or infection—and so it is not clear whether ectopy leads to infection or infection leads to ectopy. Similarly, in studies of chlamydial infection and OCs, it is possible that ectopy simply makes it easier to detect the infection (17, 505).

Chlamydial pelvic inflammatory disease (PID). For more than 10 years there has been considerable speculation about whether or not OCs actually offer some protection against PID caused by the ascent of chlamydial infection from the cervix into the fallopian tubes. Although OC users seem more susceptible to chlamydial infection than other women do, they are less likely to experience symptomatic chlamydial PID. For example, two recent studies found that pill users faced 20% to 30% as much risk of chlamydial PID as women using nonhormonal methods (242, 439).

How would OCs help protect against chlamydial PID? Possible explanations include reduced penetrability of cervical mucus, reduced uterine contractions during menses, and alteration of immunological response (36). Whether any of these mechanisms apply, however, is not certain.

Furthermore, most studies of PID have involved only women hospitalized for PID, which accounts for less than one-quarter of cases (439). Women hospitalized for PID are not representative of all women who have PID. Chlamydial PID, moreover, is less likely to lead to hospitalization than other forms of PID because chlamydial PID tends to be milder and less often noticed (340).

Other reproductive tract infections. Possible associations between OCs and other RTIs have also been studied, although not as extensively as OCs and chlamydial infection. The use of OCs has been reported to increase the risk of gonorrheal infection by 70% (284, 392) and the risk of candidiasis, a common yeast infection that is not usually sexually transmitted, by 50% to 80% (67, 392). Findings on OCs and bacterial vaginosis are conflicting. One study found a significant increase in risk (67); some have found a significant decrease (310, 424); and another found no relationship between OC use and risk of bacterial vaginosis (176). There is some evidence that OCs help protect against Trichomonas vaginalis, a common cause of vaginitis (22), although not all studies have found a protective effect (331).

Human immunodeficiency virus (HIV). Studies fail to show clear and consistent associations between OC use and HIV infection. Studies of the risk factors for HIV infection vary widely in quality and methodology and are difficult to compare directly. For many, OC use was only one of a number of variables that were examined, and the potential relationship between OC use and HIV was not the focus of the study. Individual studies often show elevated, although not statistically significant, risk of HIV infection among OC users. A recent analysis that pooled the results of 28 studies published or presented between 1985 and 1999 found a significant association between use of OCs and the incidence or prevalence of HIV infection. Based on the eight studies considered methodologically most sound, OC use was associated with a 60% increase in risk (504). Studies of OCs and HIV may not be suitable for a pooled analysis, however. One recent review of the literature on the association between OCs and HIV infection concludes that studies to date suffer from methodological limitations that make them inappropriate for combined statistical analysis (445).

Programme de Marketing Social au Cameroun

This poster from Cameroon states: "Happy Family: A child needs someone who gives him a lot of love but also a lot of time. Thanks to Novelle Duo, a reliable and effective contraceptive, you can space births according to your wishes and your family or professional situation." For many women, both high effectiveness and convenience are important reasons for choosing oral contraceptives.

Several recent studies have found a significant association between OCs and HIV among various subgroups after adjusting for a variety of confounding factors. A prospective study of 435 HIV-negative Kenyan sex workers found that over a one-year period OC users were 2.6 times more likely to become HIV positive than women not using OCs (291). Another found a link only among poor women, after taking account of condom use, number of partners, and husband having multiple sexual contacts (202). In contrast, a study of women attending a Nairobi antenatal clinic—a group considered at low risk of HIV infection—found that OC users were 3.5 times more likely to be infected with HIV than women using other methods of contraception or no contraception at all. The association persisted after adjustment for variables such as frequency of intercourse, number of partners, and history of STI symptoms. Few of these women used condoms (428).

A cross-sectional study in Nairobi suggests that OC use increases the risk of HIV infection only among women with genital ulcers. OC use alone did not increase HIV risk. Women who had used OCs longer than 12 months and had genital ulcer disease, however, were 25 times more likely to be infected with HIV than women who did not use OCs and did not have genital ulcers. This finding was based on 16 women who were long-term OC users and had genital ulcer disease, 80% of whom were infected with HIV (362).

Some evidence has led to speculation that HIV-infected OC users could infect their partners more readily than other HIV-infected women. In Kenyan women HIV DNA was found more often in the cervical and vaginal secretions of HIV-infected OC users than in other HIV-infected women. The higher the OC dose, the more likely that HIV DNA was present (315). Another study found that HIV-infected OC users shed significantly more HIV DNA in cervical cells than did other HIV-infected women (77). Not all studies have found a link between OCs and HIV DNA levels, however (249).

The presence of another STI increases the risk of HIV infection by two- to sixfold (56, 91, 108, 114, 205, 232, 238, 256, 303, 339, 363). Bacterial vaginosis, which is often but not always sexually transmitted (201), also has been linked to increased HIV risk (414, 451, 507). If other STIs make women more susceptible to HIV infection, and OCs make women more susceptible to other STIs, then OCs might indirectly increase HIV risk (66, 217).

Studies of OCs and HIV are particularly difficult to carry out and to interpret. As in all studies of family planning methods, ethics prevent randomly assigning women to use various contraceptive methods. Differences in sexual behavior associated with choice of methods then make it difficult to compare users of different methods. Furthermore, all study participants at risk for STIs must be encouraged to use condoms and instructed in their use, making it difficult to study the effects of OCs among women who do not use condoms. Also, uncertainty about when a woman became infected with HIV can make it hard to know if a woman was using OCs at the time of infection. Finally, differences in classifications of OC use patterns make comparisons among studies difficult.

A number of case-control studies in developed countries have detected increased risks of a rare liver cancer, hepatocellular carcinoma, in OC users (139, 196, 254, 324, 342, 455, 548). These studies reported risks among OC users about 2 to 20 times greater than risks among nonusers. The largest of these studies found that women using OCs for eight years or more were four times more likely to develop this liver cancer than nonusers (324). In contrast, a recent study in six European countries (191) and a study of South African women (240) found no increased risk of hepatocellular carcinoma among short-term or long-term users.

City of Toronto, Department of Public Health

For protection against sexually transmitted diseases (STDs), including HIV, the virus that causes AIDS, latex condoms are the best method and can be used along with OCs. This Canadian poster urges pill users to use condoms to prevent AIDS and STDs.

Despite some lingering uncertainties, the benefits of oral contraceptives far outweigh the risks for the vast majority of women. Continuing research has made it possible to identify more clearly the few women who face substantial risks and should choose another method of contraception. Forty years after their introduction, OCs remain popular for their convenience, effectiveness, and safety.

Side-Bars

Progestin-Only OCs for Breastfeeding Women For breastfeeding women who have resumed menstruation, progestin-only pills, or "minipills," are a good option if they want to use a hormonal method.* In contrast to combined pills, there is no risk that progestin-only pills will reduce milk production. Why Progestin-Only Pills? Postpartum women often want to delay another pregnancy, and, indeed, birth intervals of at least two years are healthiest for both siblings (31). Intrauterine and barrier methods offer good postpartum contraception with no effect on lactation. Many women, however, prefer to use OCs. Because combined pills may inhibit milk production, some providers are reluctant to give them to breastfeeding women. If providers will not give OCs to breastfeeding women, however, some women may stop breastfeeding in order to obtain them (21). Progestin-only pills are a good alternative. They have no adverse effects on lactation. Most research has found either that they have positive effects—increasing milk quantity or improving its nutritional quality—or that they have no effect (72, 145, 211, 296, 530, 531). Women who choose progestin-only pills can use them and continue to breastfeed until lactation naturally stops. The main comparative disadvantage of progestin-only pills—higher pregnancy rates than combined pills—is offset by the protection against pregnancy that breastfeeding itself provides; during breastfeeding ovulation is uncommon before menstruation resumes and may be irregular even after menstruation has resumed (70). Also, the bleeding irregularities associated with progestin-only pills may not bother postpartum women because they may be amenorrheic or expect irregular bleeding postpartum (297). Progestin-only OCs may not be the best method, however, for women with a history of gestational diabetes (temporary diabetes that develops only during pregnancy). A recent study of women with a history of gestational diabetes found that those who used progestin-only OCs during breastfeeding were almost three times more likely to develop chronic non-insulin-dependent diabetes than women who used nonhormonal methods. Use of combined OCs did not increase the risk of diabetes for women with a history of gestational diabetes (246). Although combined OCs do affect breast milk, these effects do not seem to harm infants. With combined OCs milk volume usually decreases slightly, even with low estrogen doses (21, 116, 211, 297, 452). Breast milk composition may change, too, although findings vary. Most studies report decreases in mineral content (211, 296). A number of studies have found, however, that reduced milk volume in OC users did not affect their infants' weight gain (57, 208, 452, 529). Studies in Chile reported slower infant weight gain but no other adverse effects on infant health (98, 116, 351). The longest follow-up study found no effects on the health or the physical, intellectual, or psychological development through age eight of Swedish children whose mothers used combined OCs while nursing (329). Progestin-only pills do not adversely affect a mother's milk supply, and women using progestin-only pills breastfeed as long as women using no contraception or a method other than OCs (111, 297, 520, 551). In one study 83% of progestin-only pill users breastfed for four months or longer compared with 40% of combined OC users (90). When to Begin? When can breastfeeding women begin to use progestin-only pills? As a general rule, as soon as six weeks after childbirth, according to the World Health Organization medical eligibility criteria for contraceptive methods (538). If a woman is partially breastfeeding and her child receives much other food or drink, six weeks after childbirth is the best time to start progestin-only pills. If she waits longer, fertility may return (190, 255). In contrast, if a woman plans to breastfeed exclusively or fully for a lengthy period, some providers may advise her to wait and offer her progestin-only pills later. Of course, a program can provide any woman with pills immediately postpartum with instructions about when to start them, if contacting her later might be difficult. In all cases it is important that the woman has access to the pills before she needs them. Most family planning programs prefer not to offer any hormonal contraception in the early postpartum months. This is because trace amounts of contraceptive hormones—usually less than one-tenth of 1% of maternal doses—can reach infants in breast milk. No health risks have been linked to such exposure, however (500, 530, 531). In any case, as noted, fully or nearly fully breastfeeding women who are amenorrheic do not need OCs in the early postpartum period. Fully breastfeeding is more than 98% effective in protecting against pregnancy as long as a mother is: (1) in the first six months postpartum and (2) still amenorrheic (237). This rate—two pregnancies per 100 women in the first six months after childbirth—is about the same as typical OC effectiveness (see Chapter 2.1). Program practices about when to offer progestin-only pills to fully or nearly fully breastfeeding women can be based largely on the breastfeeding patterns of the client population. To protect herself from pregnancy, the client should begin progestin-only OCs when menstruation returns or at six months postpartum, whichever comes first (84, 237, 485). *Postpartum women have little need of contraception for up to six months after giving birth if they have not resumed menstruating and they are fully or nearly fully breastfeeding—that is, breastfeeding often, day and night, with breastfeeds accounting for at least 85% of the baby's feedings (255, 552). Recent studies have reported a high degree of pregnancy protection for at least six months postpartum and somewhat less protection up to 12 months, if menstruation has not resumed (555).

Facts About Pill Use: Did You Know...? The greatest number of married users of the pill is in China (7.6 million), followed by Germany (6.8 million), Indonesia (6.1 million), Brazil (6.0 million), Bangladesh (5.7 million), and the United States (5.6 million). Nearly one-half of married women in Western Europe use the pill. This amounts to three of every five contraceptive users. In the US, an estimated 80% of all women born since 1945 have used the pill at some point in their lives (106). OCs are the most popular method among sexually active unmarried women in sub-Saharan Africa and Latin America. Some 95% of French women have ever used the pill, contrasted with 4% of Japanese women (356). Japan approved the pill for contraceptive use only recently—in September 1999. In Canada 7 of every 10 pill users over the age of 35 have been using the pill for more than 10 years (38).

Preventing Cervical Cancer Cervical infection with some types of human papillomavirus (HPV) appears to cause most, if not all, cases of cervical cancer (126, 372). A recent analysis of 1,000 cervical cancer specimens collected worldwide found evidence of HPV infection in 99.7% of the samples (502). Many women develop HPV infections, but few go on to develop cervical neoplasia. HPV infection usually is transient and clears without treatment (199). Apparently, cancer arises from infections that persist—perhaps those lasting six months or more (203, 383). Avoiding HPV Primary prevention of cervical cancer is the ideal, and that means minimizing exposure to HPV. A woman can reduce her exposure to HPV and other sexually transmitted disease organisms by using a barrier method of contraception—preferably condoms, but perhaps also diaphragms and spermicides—whether or not she also uses another family planning method such as OCs. Abstinence and delaying first sexual intercourse also reduce the risk (173). The behavior of women's sexual partners is important. Men who were young when they first had sexual intercourse, who have multiple sexual partners, or who visit prostitutes regularly increase their partners' risk of cervical cancer significantly (107, 317, 462). It may be particularly difficult for a sexually active woman to avoid HPV. Identifying an uninfected sexual partner—and knowing one's own status—is not possible without testing. Moreover, the types of HPV that cause cervical cancer do not cause warts (235) or any other obvious symptom. At the same time, the virus is very common. Condoms are helpful, but HPV can spread through contact between areas of the body near the anus or genitals that a condom does not cover (423). HPV vaccines are being developed, but the availability of a safe and effective vaccine is probably over a decade away (215, 372). While HPV infection may initiate most or all cervical cancers, cigarette smoking poses an increased risk (523, 527), and avoiding smoking will limit risk. A diet rich in vitamin C may also help (173). Screening Since most women cannot eliminate all chances of exposure to HPV, where feasible, women should be screened for cervical lesions. The Papanicolaou (Pap) smear is the current standard screening method. Pap smears can identify cervical neoplasia at early stages, when treatment is almost always effective. Countries that have instituted national screening programs have seen deaths from cervical cancer decline to one-third or less of previous levels (384). Unfor-tunately, comprehensive Pap screening is practically non-existent in developing countries, where cervical cancer is the most common type of cancer among women. A more feasible screening technique appears to be on the horizon. Visual inspection of the cervix after an acetic acid (vinegar) wash—also known as cervicoscopy, or VIA—offers a low-cost, low-tech alternative to the Pap smear. Lesions appear white after application of vinegar and can be seen with a flashlight (220). In Zimbabwe nurse-midwives using this method accurately detected more than 75% of pre-invasive lesions compared with 44% with Pap smears (479). Similarly, in India paramedical personnel could accurately detect pre-invasive and invasive lesions using VIA (402). In India, VIA was as specific—able to detect accurately women who do not have pre-invasive or invasive lesions—as a Pap smear (402), while in Zimbabwe VIA was less specific than a Pap smear (479). Early detection allows for early treatment with low-cost, easy methods such as cryotherapy—freezing the cervix with a liquid coolant to destroy abnormal tissue—that nurse-midwives and many other health care providers can administer (220).

Figures