NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on the Relationship Between Oral Contraceptives and Breast Cancer. Oral Contraceptives & Breast Cancer. Washington (DC): National Academies Press (US); 1991.

Cover of Oral Contraceptives & Breast Cancer

Oral Contraceptives & Breast Cancer.

Show details

AOral Contraceptives and Breast Cancer: A Review of the Epidemiological Evidence with an Emphasis on Younger Women

Kathleen E. Malone

The possibility of increased breast cancer risk related to oral contraceptive use is a major concern to American women and to the scientific community. Breast cancer incidence in Western countries is relatively high and apparently is increasing. That breast cancer appears to be influenced by other hormonally mediated factors leads to the hypothesis that the high rate of exposure to oral contraceptives among American women may also be associated with this increase.

Examination of cancer incidence data from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute suggests that there has been an overall increase in the incidence of breast cancer, with increases of the largest magnitude occurring among women over age 50 (according to SEER data, approximately 1.2 percent per year since 1974). Age-adjusted incidence rates for breast cancer in women under the age of 50 have also increased since 1973, but the increases have been of a much smaller magnitude—approximately 0.2 percent per year. The use of mammographic screening, which facilitates the detection of cases that might otherwise have gone unnoticed, or at the least detects cases at an earlier point in time, may explain some of this increase, especially in women over age 50. However, because screening recommendations apply mainly to middle-aged and older women, screening may not account for much of the increased incidence in young women.

Most of the early research on this topic (i.e., studies conducted prior to 1980) found no association between oral contraceptive use and breast cancer, reassuring many in the research and clinical communities that there was little or no increased risk (Henderson et al., 1974; Sartwell et al., 1977; Ravnihar et al., 1979; Vessey et al., 1979). Investigations reported in the early 1980s offered little to cause observers to change their minds. The majority of studies, including the largest case-control study in the United States, found no consistent suggestion of elevated risk (Kelsey et al., 1981; Vessey et al., 1981, 1982, 1983; Brinton et al., 1982; CASH, 1983; Janerich et al., 1983; Hennekens et al., 1984; Rosenberg et al., 1984; Stadel et al., 1985; Talamini et al., 1985), although several studies reported elevated risk estimates for particular aspects of oral contraceptive use (Paffenbarger et al., 1980; Pike et al., 1981, 1983; Royal College of General Practitioners, 1981; Trapido, 1981; McPherson et al., 1983; Olsson et al., 1985). There was no consistency among these elevated risk estimates.

The picture seems to have changed since 1986. There are a number of studies during this period that do not support an increased risk related to oral contraceptive use, including more analyses of the Cancer and Steroid Hormone (CASH) Study data and the updated analysis of data from the United States nurses' cohort study (CASH, 1986; Ellery et al., 1986; La Vecchia et al., 1986; Lipnick et al., 1986; Romieu et al., 1989; Stanford et al., 1989). But an increasing number of studies have appeared that suggest an elevated risk in relation to some aspects of oral contractive use (Meirik et al., 1986, 1989; Paul et al., 1986; McPherson et al., 1987; Ravnihar et al., 1988; Kay and Hannaford, 1988; Miller et al., 1989; Olsson et al., 1989; U.K. National Case-Control Study Group, 1989). These studies, especially those with positive findings, have received much publicity and generated renewed concern over the safety of oral contraceptives.

Changing Profile Of Oral Contraceptive Use

Oral contraceptives were introduced in the early 1960s but did not find widespread acceptance until the late 1960s and early 1970s. There are four major types of oral contraceptives. Combination pills, which contain fixed amounts of estrogen and progestin and act by suppressing ovulation, were the first oral contraceptive approved in the United States (in the early 1960s) and have always been the most popular type of pill used. Sequential pills, in which estrogen alone is given for the first two weeks followed by an estrogen-progestin combination during the last six days, were also introduced early but were removed from the U.S. market around 1977 (Piper and Kennedy, 1987). The minipills, or progestin-only pills, introduced in 1972 contain no estrogen and a lower amount of progestin. They do not affect ovulation but rather inhibit ovum transport and implantation by thickening the cervical mucus. They have never had a significant share of the market. Phasic oral contraceptives, combination pills that contain estrogen along with a progestin dose that varies in amount throughout the month, were introduced in 1983 and have since become increasingly popular.

Following the advent of oral contraceptives, two major changes have occurred that must be considered in evaluating the research on oral contraceptives and breast cancer: (1) changes in the formulation of oral contraceptives and (2) changes in the patterns of their use and in the women who use them. Regarding the first change, over the past three decades, the formulations of oral contraceptives have undergone substantial modifications. Both the types and doses of steroids have changed; the doses of both estrogen and progestins have been greatly decreased; several new progestins have entered the market whereas others have been withdrawn; and sequential pills are no longer used. High-estrogen-potency oral contraceptives constituted 94 percent of the United States retail market in 1964, but by 1976 low-estrogen oral contraceptives (less than 50 µg of estrogen), which were introduced in 1967, dominated the market (Piper and Kennedy, 1987). Many of the formulations evaluated in past breast cancer studies are no longer being sold.

The profiles of oral contraceptive users have also changed over time. Oral contraceptives were most commonly used first in the 1960s by married women to space pregnancies and only later by single women as a method to delay a first pregnancy (McPherson and Drife, 1986). Routine prescription to younger single women was not common until the early 1970s in Great Britain and the late 1970s in the United States (McPherson and Drife, 1986). These geographic differences in oral contraceptive prescription patterns have been posited as a possible explanation for some of the differences in study results that have been seen between the two countries. In the United States, only women born since the mid-1940s have had the opportunity to be exposed to long-term use of the pill early in life.

At present the use of oral contraceptives is an extremely prevalent exposure among young women. With the recent epidemic of teenage pregnancies in the United States, routine prescription to teenage girls as young as age 13 is not uncommon. Unpublished data from a case-control study conducted by the author and colleagues of breast cancer diagnosed between 1983 and 1988 in young women born after 1944 in Seattle, Washington, showed that 92 percent of both the cases and controls reported ''ever'' having used oral contraceptives; among women under age 20 the proportion of "ever" use increased to 100 percent.

Biological Plausibility

An important criterion for evaluating causality is the biological plausibility of the relationship. Estrogen causes proliferation of breast tissue and would be expected to increase breast cancer risk by stimulating growth of stem cells and intermediate cells (Thomas, 1984). Progestin causes alveolar cell growth in the estrogen-primed breast, but it also causes differentiation. It is unclear, therefore, whether the predicted net effect would be to increase or decrease breast cancer risk.

The influence of estrogen and progestin on breast epithelium proliferation and differentiation appears to differ with age. Cancer risk is thought to be a function of the number of cells at risk, which varies with age. It is possible to posit that any carcinogenic risk of oral contraceptives may be strongly mediated by age of exposure or by the timing of exposure in relation to other events that are thought to affect epithelial proliferation or differentiation (e.g., menarche, first full-term pregnancy).

The etiology of breast cancer has strong hormonal themes: oophorectomy decreases risk, early menarche and late menopause increase risk, late age at first full-term pregnancy increases risk. These effects seem to last for decades. Thus, if use early in life affects risk, it may be many years before deleterious outcomes are seen. It is possible that the studies conducted thus far have not allowed an adequate interval between exposure to oral contraceptives and the onset of breast cancer, so that even if an association were present, it might not yet be detectable. Thus, vigilance must be maintained and investigation of this issue should continue in the future—even though we might conclude, based on current data, that findings are too inconsistent to be alarming at present.

Epidemiological Studies Of Breast Cancer and Oral Contraceptives

As mentioned earlier, the studies conducted prior to 1980 carry little suggestion of an increased risk for breast cancer in relation to oral contraceptive use. These studies focused on cases diagnosed before 1975; therefore, they included a large proportion of women who, because of their birth years, had little opportunity to have ever used oral contraceptives or to have used oral contraceptives for a long time, and virtually no opportunity for exposure at an early age. For these reasons, as well as the briefness of the time between exposure and diagnosis, studies conducted before 1980 cannot contribute any insight into the current controversy and are largely ignored in this review. Unfortunately, these same difficulties plague some of the studies published in the early 1980s as well. Therefore, although studies from the early 1980s are included in this review, the emphasis is on more recent research.

"Ever" Use of Oral Contraceptives

Table A-1 presents a summary of the findings related to "ever" use of oral contraceptives and risk of breast cancer. Twenty-five of the 30 studies for which a relative risk for "ever" use was reported had relative risks (RR) close to 1.0. Two of the five studies that report elevated risk estimates for "ever" use of the pill are ones for which questions with regard either to basic study design, low response rates, or low exposure prevalence have been raised (Olsson et al., 1989; Miller et al., 1989). Overall, there is no evidence of increased risk of breast cancer in women who meet the criterion of ''ever" using oral contraceptives. The finding of no association between "ever'' use of oral contraceptives and breast cancer risk has been quite consistent throughout the past 20 years of research.

TABLE A-1. Summary of Risk Estimates for "Ever" Use of Oral Contraceptives.

TABLE A-1

Summary of Risk Estimates for "Ever" Use of Oral Contraceptives.

"Ever" use of oral contraceptives is probably too crude a measure to provide much insight into any relationship with breast cancer risk because such use is so common an exposure that it typically encompasses more than 90 percent of all young women. Further, interpretation is difficult because there are subgroups of women who try oral contraceptives but stop—because of side effects—soon after they start. Women who never use oral contraceptives may be a unique subgroup. For instance, they may have an increased family history of breast cancer, health problems that contraindicate the use of oral contraceptives, or an increased awareness or suspicion of undiagnosed infertility; these factors, in turn, may relate to both their decision not to use oral contraceptives and their risk for breast cancer.

Duration of Oral Contraceptive Use

There was little suggestion of increased risk related to long-term use of oral contraceptives in any of the studies published prior to 1986. Of the case-control studies published since that time, seven reported an excess risk for use of long duration. Table A-2 presents a summary of reported risk estimates for long-term use.

TABLE A-2. Summary of Risk Estimates for Lifetime Duration of Oral Contraceptive Use.

TABLE A-2

Summary of Risk Estimates for Lifetime Duration of Oral Contraceptive Use.

The largest case-control study of breast cancer conducted to date, the CASH study, showed no evidence of an association of breast cancer risk and long-term use of oral contraceptives among women aged 20 to 54 (CASH, 1986). Yet despite its size (the study comprised more that 4,700 cases and a similar number of controls in eight geographic regions of the United States) the majority of the women were over age 45. Coupled with the early diagnosis years (1980–1982), it is possible that this study was conducted too early to detect any association between oral contraceptives and breast cancer. The major reproductive years for most of these women occurred before the height of popularity of the pill. A number of the other studies among women up to the age of 65 have also found no pattern related to long durations of use, suggesting no increase in risk related to long-term oral contraceptive use in the aggregate or perhaps no increase in risk with long-term use that does not begin at an early age (Ellery et al., 1986; La Vecchia et al., 1986; Paul et al., 1986; Stanford et al., 1989). A Yugoslavian study of women under age 55, however, found an RR of 2.4 for oral contraceptive use exceeding seven years, as well as a significant dose-response pattern (Ravnihar et al., 1988). Recently, the World Health Organization (WHO) study, a multinational case-control study conducted in three developed and seven developing countries, reported a suggestive dose-response pattern of increasing risk with years of oral contraceptive use (WHO, 1990); this association, however, could well be attributable to a recency effect since risk was highest in current users and steadily declined with time since last exposure.

In recent years, a large number of studies have focused on breast cancer risk in women under age 45. Meirik and colleagues (1986) published the first report suggesting a dose-response relationship with duration of oral contraceptive use among young women: for use of 8 to 11 years, an RR of 1.4 was found; for 12 or more years of use, a 2.2-fold excess risk of breast cancer was found. Paul and colleagues (1986) reported an RR of 4.6 for use of 10 or more years among women aged 25–34. McPherson and coworkers (1987) found an increased RR of 1.8 for use exceeding 11 years among a group of British women through age 45. The fact that U.K. prescription patterns were about five years ahead of those in the United States has been offered as a possible explanation for the earlier emergence of positive studies from the United Kingdom.

A hospital-based case-control study conducted in the northeastern United States (Miller et al., 1989) among women under age 45 ob served a twofold excess risk for oral contraceptive use of five to nine years' duration and a fourfold excess risk for use of 10 or more years. Questions have been raised about Miller's findings because of concern about the appropriateness of the hospitalized control group and because there appears to be a lower proportion of exposed controls than would be expected from national data. A well-conducted study in the United Kingdom (U.K. National Case-Control Study Group, 1989) among women under age 36 recently reported a significant dose-response pattern for duration of use. Oral contraceptive use of 49 to 96 months was associated with a 1.4-fold excess risk, and use exceeding 96 months was associated with a 1.7-fold excess breast cancer risk. This case-control study was one of the few that was able to validate the self-reported data on oral contraceptive use so as to rule out the often-raised criticism of recall bias.

Conflicting observations have been recorded among the three large prospective cohort studies. The largest, the Nurses Health study in the United States (Romieu et al., 1989), found no increase in risk for any duration of use (or for any other aspect of oral contraceptive use except current use). Current use of oral contraceptives was associated with an overall adjusted RR of 1.5. This excess risk was confined to women between ages 40 and 50. Tumors in current users were reported to be larger and to have more lymph node involvement at the time of diagnosis than were tumors in women not currently using oral contraceptives. The Oxford cohort in the United Kingdom has seen no evidence of increased risk related to oral contraceptive use (Vessey et al., 1989). However, the Royal College of General Practitioners cohort in the United Kingdom reported excess breast cancer risk for longer durations of oral contraceptive use, although there was no consistent dose-response pattern (Kay and Hannaford, 1988). In this cohort, there were inconsistent, mildly elevated risks for duration of use among women of all ages. In two subgroups, women ages 30 to 34 and women who were parity 1, much higher risks were seen—as high as a 10-fold excess risk for use of 10 or more years. Both of the cohorts formed in the United Kingdom began recruitment in 1968 and excluded women not involved in a married or living as married relationship (Kay and Hannaford, 1988; Vessey et al., 1989). The Nurses' Health study recruited women aged 30–55 in 1976 (Romieu et al., 1989). All three of these cohorts may have been initiated too early to include many women born recently enough to have had the opportunity to use oral contraceptives at a young age or for a long duration.

Oral Contraceptive Use Before First Full-term Pregnancy or Before Age 25

It is possible that oral contraceptives may be especially influential at or in relation to particular reproductive milestones. Because of the increasing frequency of use of the pill at young ages, and because of the possibility of increased susceptibility of breast tissue to hormonal exposures during young ages when breast development is still ongoing and when endogenous hormone concentrations are still increasing, there has been mounting interest in the evaluation of breast cancer risk in relation to use at young ages. Findings regarding any relationship between use of oral contraceptives before the first full-term pregnancy (FFTP) or before age 25 have been inconsistent. A factor that further complicates the picture is that some investigators report on use before FFTP only among parous women, some consider parous and nulliparous women combined, and others investigate only nulliparous women. Table A-3 summarizes the findings regarding oral contraceptive use before FFTP.

TABLE A-3. Summary of Risk Estimates for Oral Contraceptive Use Before First Full-term Pregnancy.

TABLE A-3

Summary of Risk Estimates for Oral Contraceptive Use Before First Full-term Pregnancy.

The vast majority of investigations of the relationship between oral contraceptive use prior to FFTP and breast cancer risk have focused on women under age 45. Pike's 1981 study was the first to report an adverse effect from oral contraceptive use prior to FFTP, observing an RR of 2.3 for five to eight years of oral contraceptive use preceding FFTP and an RR of 3.5 for more than eight years use before FFTP. One limitation of this study, however, was a low response rate for the cases. A threefold excess risk was reported by McPherson and colleagues (1983) for use exceeding four years' duration before FFTP and was accompanied by the suggestion of a dose-response relationship. The McPherson team's 1987 report observed a twofold excess risk of breast cancer for one to four years of use before FFTP and a 2.6-fold excess risk for use exceeding four years' duration prior to FFTP among nulliparous and parous women combined. Meirik and coworkers' (1986, 1989) reports on Swedish and Norwegian women revealed an increased risk for eight or more years of use before FFTP in the aggregate as well as in both nulliparous and parous women when examined separately (all women: RR = 2.0, confidence interval (CI) = 1.8-4.2; nulliparous women: RR = 4.3, CI = 1.4-13.1; parous women: RR = 1.7, CI = 0.7-4.2). There is some suggestion in Meirik's data that use before FFTP may be related to overall long duration of oral contraceptive exposure. This may well be true for some of the other studies but was not addressed in the tables or text of most other papers.

Two recent studies also observed increased risks for oral contraceptive use prior to FFTP, although both studies have been criticized because of possible design limitations. The hospital-based study by Miller and colleagues (1989), for which possible limitations were mentioned previously, found a 1.6-fold excess risk of breast cancer for oral contraceptive use preceding FFTP that exceeded four years. In the Swedish study by Olsson and coworkers (1989), an RR of 1.8 (CI = 1.0-3.2) was observed for three or fewer years of oral contraceptive use prior to FFTP, an RR of 1.7 (CI = 1.1-3.8) for four to seven years of use before FFTP, and an RR of 2.1 (CI = 0.8-4.7) for use of eight or more years before FFTP. Different interviewers were used for cases versus controls, however, and the response rates were rather low in this study, raising concerns about the findings.

A number of analyses found no suggestion of increased breast cancer risk for oral contraceptive use before FFTP (Paul et al., 1986; Jick et al., 1989; Romieu et al., 1989; WHO, 1990). Overall consideration of oral contraceptive use before FFTP in the complete Cancer and Steroid Hormone study data revealed no suggestion of excess breast cancer risk in the study's first report on the entire data set (Stadel et al., 1985). However, in a recent analysis of a "high-risk" subgroup of the CASH study subjects—nulliparous women with an early age of menarche diagnosed with breast cancer before age 45—an excess risk of breast cancer was seen in relation to increasing duration of oral contraceptive use (Stadel et al., 1988). The risk for oral contraceptive use of 8 to 11 years was 2.7, and the risk for use of 12 or more years was 11.8. A recent letter by Peto (1989) presented a crude reanalysis of the published CASH data that challenged an earlier conclusion of the study of no excess risk for use before FFTP.

Eight studies, which are summarized in Table A-4, have reported on use before age 25. Three studies (Stadel et al., 1985; Miller et al., 1986; McPherson et al., 1987) have shown no indication of a relationship with breast cancer risk, whereas four (Pike et al., 1983; Meirik et al., 1986; Olsson et al., 1989; WHO, 1990) suggest a positive relationship and one (Paul et al., 1986) suggests a protective effect. Pike and coworkers' 1983 study, in an expansion of the 1981 investigation, observed a significant dose-response pattern of increased risk for increased duration of use before age 25. For such use exceeding five years' duration, there was a 4.9-fold excess risk. In the data from Meirik and colleagues (1986), among Swedish and Norwegian women, there was no elevation in risk for use of less than eight years' duration prior to age 25, but there was an increased risk of 2.7 (CI = 0.7-11.0) for use of eight years or more before age 25. The Olsson team's 1989 study, also among Swedish women, reported a suggestive dose-response pattern for increasing duration of oral contraceptive use prior to age 25, with a 1.6-fold excess risk for use of less than three years before age 25, a twofold excess risk for use of three to five years, and a 5.3-fold excess risk for use exceeding five years.

TABLE A-4. Summary of Risk Estimates for Oral Contraceptive Use Before Age 25.

TABLE A-4

Summary of Risk Estimates for Oral Contraceptive Use Before Age 25.

A major challenge in interpreting many of the studies of use at a young age lies in separating the effects related to use early in life from effects associated with longer durations of exposure. More attention needs to be given to this issue in future analyses, particularly in populations in which the majority of the women were born recently.

Duration Since First Use of Oral Contraceptives (Latency)

It has been suggested that long-term latent effects that have been missed might be the alternative explanation for many of the studies with negative findings. More than 10 studies have presented such analyses with no demonstration of a consistent latency pattern (Brinton et al., 1982; Harris et al., 1982; Vessey et al., 1983; Rosenberg et al., 1984; Ellery et al., 1986; Meirik et al., 1986; La Vecchia et al., 1986; Paul et al., 1986; Lee et al., 1987; Ravnihar et al., 1988; Schlesselman et al., 1988). It is possible, however, that these studies were conducted too early to see such an effect.

Use of Oral Contraceptives in High-Risk Subgroups

Although many investigators adjust for high-risk factors (e.g., family history of breast cancer, history of benign breast disease) in their analyses, only a subset have examined oral contraceptive use within each of these strata. Furthermore, of the few studies that have examined oral contraceptive use in each strata, the majority have limited their definition of use to ''ever/never'' and their definition of the high-risk subgroups to fairly crude delineations (i.e., "ever/never" had sister with breast cancer). These approaches are unfortunate because they may well miss important modifying relationships that cannot be detected at such a crude level.

With regard to family history of breast cancer, the bulk of the studies have detected no substantial differences in the risk related to oral contraceptive use for women with and without this factor (Miller et al., 1989; Murray et al., 1989; Romieu et al., 1989). Brinton and colleagues (1982) found no differences in women with and without a mother with breast cancer but did see differences in women with and without a sister with breast cancer.

Elevated risks have been observed for oral contraceptive use among women with a history of benign breast disease (Fasal and Paffenbarger, 1975; Lees et al., 1978; Brinton et al., 1982; Janerich et al., 1983); but more work is needed to evaluate specific histologic types of benign breast disease in terms of both breast cancer risk and relationship to use of the pill. Some past analyses failed to distinguish between oral contraceptive use before and after the diagnosis of benign breast disease (Stadel and Schlesselman, 1986). In addition, not much has been done to examine histologic subgroups of breast cancer for the possibility of differential relationships with oral contraceptive exposure.

Steroidal Potency of Various Formulations

The hormonal contents of oral contraceptives have been examined by a number of classification schemes related to potency, brand, and type of estrogen (Brinton et al., 1982; Harris et al., 1982; Pike et al., 1983; Vessey et al., 1983; Stadel et al., 1985; Ellery et al., 1986; CASH, 1986; Miller et al., 1986; McPherson et al., 1987; Schlesselman et al., 1987; Ravnihar et al., 1988). Not one of these approaches, however, has consistently exhibited a relationship with breast cancer risk.

Issues to Consider in Reviewing the Epidemiological Evidence

In attempting to evaluate the accumulated research, several issues should be considered. First, the design and conduct of each study should be examined to detect possible limitations that could have affected the results. Specific factors such as the fundamental design (case-control or follow-up study) and the proportion of eligible subjects who participated in the study, or in a follow-up design, are salient to interpretation of the findings. Several of the previously mentioned studies suffered from low response rates or large losses to follow-up (Janerich et al., 1983; Kay and Hannaford, 1988). If these losses are great, the validity of the case-control study is compromised because of the possibility of differential exposure distributions in the responders versus the nonresponders. In a follow-up study, similar questions arise concerning the possible differential distribution of disease occurrence. Table A-5 itemizes the following characteristics for most of the studies addressed here: diagnosis dates and the ages of women at the time of diagnosis, proportion of cases and controls who participated, sample size, proportion of cases and controls who reported "ever" using oral contraceptives, and whether controls were secured from hospitals.

TABLE A-5. Summary of Characteristics of Case-Control and Cohort Studies of the Relationship of Oral Contraceptive Use and Breast Cancer Risk.

TABLE A-5

Summary of Characteristics of Case-Control and Cohort Studies of the Relationship of Oral Contraceptive Use and Breast Cancer Risk.

The sample size of a study must be large enough to allow the detection of an effect or to rule out with a certain amount of confidence the presence of an effect. A number of past studies (i.e., Kelsey et al., 1978; Harris et al., 1982; Olsson et al., 1985; Ellery et al., 1986; Lee et al., 1987; Jick et al., 1989) may not have had adequate power to evaluate the relationship of oral contraceptives and breast cancer. In addition, sample sizes need to be even larger to examine the interrelationships of other risk factors with oral contraceptive use (Greenland, 1983; Smith and Day, 1984).

In a case-control study, the method of ascertaining cases and controls should be carefully considered, as well as the degree to which the controls represent the population from which the cases were drawn. Concern has been raised about the validity of conducting studies in hospital settings, especially referral or tertiary care hospitals, because it is virtually impossible to identify the population from which the cases arose. The usual approach for selecting controls is to recruit them from among the other patients in the same hospital who are seeking care for other diseases. The difficulty lies in deciding what conditions these patients can and cannot have in order to be an appropriate comparison group. Any condition known to be associated with the exposure or with the disease should be excluded. As the number of exclusions increases, questions arise as to the select nature of the control group; unfortunately, it is impossible to be sure of the direction of the bias that might result from such selectivity. One could hypothesize that the controls with these other medical conditions that have led to hospitalization are less likely to be oral contraceptive users; in that case, however, the controls would produce an underestimate of the population's use of the pill, and the observed risks would be spuriously high. Quite a few of the reviewed studies were conducted in hospital settings (Paffenbarger et al., 1980; Kelsey et al., 1981; Vessey et al., 1982, 1983; McPherson et al., 1983, 1987; Rosenberg et al., 1984; Talamini et al., 1985; Ellery et al., 1986; La Vecchia et al., 1986; Miller et al., 1986, 1989; Ravnihar et al., 1988).

Most recently, the study by Miller and colleagues (1989) was criticized because of its low proportion of pill-exposed controls; the proportion exposed was lower than that observed in national survey data for women in the age group studied. This result raises questions about bias in the selection of the hospital controls (Spirtas, 1989).

A second issue involves any bias that might arise in the ascertainment of exposure. All of the case-control studies reviewed here relied on interview data, although one—the U.K. National Case-Control Study Group (1989)—was able to review medical records to evaluate recall.

Another study (Olsson et al., 1989) violated a basic design principle by having a different interviewer for cases and controls.

Third, it is necessary to understand and incorporate into analyses (when appropriate) the other important breast cancer risk factors, such as age of first full-term pregnancy, number of live births, age of menarche, family history of breast cancer, and history of benign breast disease. Some of these risk factors, as well as some other factors, may also affect the decision to use oral contraceptives, the age of first and last use, and the lifetime duration of use. These and other factors may modify or be modified by the relationship of oral contraceptives to breast cancer risk.

Fourth, whether the women in a study had the opportunity to be exposed to oral contraceptives deserves attention. The most obvious example of this issue is when a study includes a large number of women who were born early enough that oral contraceptives were not even on the market during their prime reproductive years. All of the studies before 1980 as well as a number of the studies in the early 1980s were constrained in this way. Extreme examples include La Vecchia and colleagues (1986), with only 14 percent of the study subjects having ever used oral contraceptives; Stanford and coworkers (1989), with 24 percent having used them; and Lee and colleagues (1987), with 41 percent having used them.

When birth year and opportunity for exposure to oral contraceptives are examined in the context of the possible latency period for breast cancer, it becomes apparent that many of the already completed studies may have been unable to evaluate the relationship of oral contraceptives and breast cancer, especially in regard to use at younger ages as well as premenopausal disease onset. For example, the studies of radiation effects on breast cancer suggest a latency interval of at least 15 years, which must be factored into efforts to detect an expected association. Presuming oral contraceptives have a promotional effect, the time interval until a detectable lesion is present is unknown. Studies that include women diagnosed before a certain point in time may not have allowed an adequate interval between exposure to oral contraceptives and onset of breast cancer. In this instance, even if an association were present, it might not be detectable in these women.

More subtle examples related to exposure opportunity that need consideration in future analyses include sterilization, hysterectomy, and infertility, all of which necessarily affect the need and timing for any method of contraception.

Discussion

The one conclusive statement that can be made concerning the sum of the epidemiological evidence of a relationship between oral contraceptives and breast cancer is that there has been a remarkable lack of consistency in the findings. However, an increasing number of the recent studies suggest that there are subgroups of women who may be at increased risk of breast cancer owing to their pattern of oral contraceptive use. Of these groups, the one of most concern may be those women with long-term use beginning at a young age. The findings for use before the first full-term pregnancy or before age 25, or just for overall long duration of use are certainly suggestive and do not permit the conclusion that there is no relationship between use of the pill and breast cancer risk. Use of oral contraceptives at a young age has increased over the past 20 years, and the possible risk of breast cancer associated with early use is an important public health issue. Why would we only now be seeing an increased risk for long-term use at a young age? There are several possible explanations:

1.

Women who use oral contraceptives are required to visit their health care provider on a regular basis to secure a renewal of their prescription. Thus, it is possible that oral contraceptive users are seen by the health care system more frequently than other women, and may receive more frequent medical surveillance, including breast exams and even mammograms. This hypothesis has led to the suggestion that women with breast cancer who use oral contraceptives are detected and diagnosed earlier than nonusers. Little is known about the natural progression of breast lesions; thus, the detection of breast cancer through mammography may indeed pick up some cases of breast cancer that might have otherwise gone undetected for many years. (This may be especially true of in situ breast cancers.) Several recent studies have incorporated into their analyses a consideration of such factors as the stage of disease, the frequency of breast self-exams and physician exams, and the frequency of mammograms as a way to account for the impact of increased surveillance (McPherson et al., 1987; Kay and Hannaford, 1988; U.K. National Case-Control Study Group, 1989). Because oral contraceptives have always required prescriptions, it seems unlikely that detection bias could be contributing enough new noise to the analyses of breast cancer and oral contraceptives to account completely for the emergence of these recent positive findings.

2.

In recent years, the controversy about the relationship between oral contraceptive use and breast cancer risk has received much attention in the lay press. Unfortunately, there are indications that positive studies receive greater publicity than negative studies. Recent studies therefore have been conducted in an environment in which many participants are familiar with the hypothesis of interest. Many epidemiological studies rely heavily on a woman's memory to ascertain details of pill use; consequently, the issue of recall bias has been raised as a possible explanation for some of the positive studies. The specific scenario of concern is one in which women with breast cancer, because of their diagnosis, ruminate about possible causes and dwell heavily on an exposure that they believe may have caused their disease (and thus remembering every detail about their use of oral contraceptives). In contrast, typically healthy women without breast cancer, who have no similar drive to ruminate about their history, may fail to remember the details of their exposure as accurately and thus may underreport their use of oral contraceptives. Cohort studies avoid this problem by securing exposure information prior to disease onset.

Women have worried for many years about the relationship of oral contraceptives to adverse health outcomes, and it seems difficult to believe that possible recent increases in levels of worry and awareness might affect recall and thus findings in recent studies. Also, although it is conceivable that this type of selective memory might operate with a fairly unmemorable exposure (e.g., aspirin use 10 years earlier), oral contraceptive use is so strongly linked with other reproductive experiences that it is unlikely that only in recent studies would controls have been underreporting their oral contraceptive use because of an inability to recall exposure.

3.

Some other methodological problem(s) may have been introduced in these recent studies that have yet to be identified, although there is no evidence supporting this hypothesis.

4.

These recent findings could simply be due to chance. The fact that there is some consistency to the positive findings, namely involving prolonged use at an early age, decreases our ability to attribute these findings to chance.

5.

Little is known about the relationship between various formulations of the pill and breast cancer risk, largely because no reliable classification system of the effects on the breast of oral contraceptives and their components has been developed. Many of the women in the recent studies may well have been exposed to both high-and low-dose pills. Although removal of the high-dose pills successfully decreased deleterious cardiovascular and other side effects experienced by pill users, it is not actually known if the newer formulations are necessarily associated with a lower, higher, or unchanged risk of breast cancer. It is possible that there are some aspects of oral contraceptive formulations whose effects on breast cancer risk have only recently become apparent.

6.

Only studies of women born since the mid-1940s could have included large numbers of women exposed to oral contraceptives at a young age. Although no one is sure of the exact latency period for breast cancer, studies of atomic bomb survivors and the risk patterns observed in breast cancer for other factors such as menarche and age at first birth suggest a long latency period of at least 15 years. If hormones have a promotional effect, it is not yet known which time-related factors may be most relevant. Studies conducted in the latter half of the 1980s may be the first conducted in women born recently enough to have used oral contraceptives at a young age and for a long enough duration and whose use preceded the diagnosis of disease by a sufficient amount of time to be consistent with a latent period for breast carcinogenesis. Oral contraceptives may have a differential effect on young breast tissue, and this effect may just have become detectable in the more recent studies. Alternatively, the effect on young breast tissue may not translate into detectable breast cancer until 30 or more years later, in which case current studies will not be able to detect this association.

If risk increases in relation to use at an early age, recent findings supporting this premise may represent only the tip of the iceberg because the cohort of women who could have used the pill at an early age has yet to enter the highest breast cancer incidence age cohorts.

The Future

Many questions remain concerning the relationship between oral contraceptives and breast cancer risk. Gaining insight will entail a collaborative interdisciplinary effort by epidemiologists, clinicians, and basic scientists. Specific future endeavors could include some of the activities noted below.

Appropriate methods should be developed to classify the potency of oral contraceptives. More information is needed on the basic biology of estrogen and progestin as separate components as well as combined products. A better understanding is required of the biological effects of hormones on breast tissue. (Most past classification systems of hormonal effects have been based on effects on the endometrium.)

In the past decade, as the state of the art in epidemiological design has advanced, the criteria for what constitutes a methodologically suitable study have increased. A number of the studies reviewed here were plagued with limitations that, at the very least, render these studies suboptimal. Future case-control studies should be carefully designed so as to avoid some of these pitfalls. More work is needed to address the possibility of recall bias, although it is hard to believe that reporting of long-term use is compromised enough to explain the recent emergence of positive findings. Validation of a reported history of oral contraceptive use through medical record review would be ideal, but is quite difficult in the United States or any country where people are mobile and do not have a centralized source of health care. Case-control or small validation studies conducted in health maintenance organizations, or studies in countries with national health care systems, both of which have lifelong medical records for study subjects, might be the best way to evaluate the question of misclassification of oral contraceptive use because of recall bias. Future case-control studies should be population based, achieve high levels of response, and be designed to minimize the likelihood of introducing bias.

The causes of the increase in breast cancer incidence should be investigated further, and the extent to which increased breast screening contributes to the increased incidence should be evaluated. In turn, for case-control studies, the degree to which increased screening of oral contraceptive users affects case identification needs to be evaluated.

In existing and future data, especially those studies with large enough numbers of subjects, more effort is needed to tease out the effects of long duration of exposure versus exposure at specific ages—as well as other patterns of use that may be associated with adverse or protective effects. In particular, the effect of use at an early age should be further evaluated to distinguish the timing of exposure from the longer duration of exposure that is associated with beginning use at an earlier age. Besides examining oral contraceptive use at young ages, there are unanswered questions concerning use during the perimenopausal years. The recent FDA decision to lift the recommended upper age limit for use of oral contraceptives will probably lead to increased use in the over-40 age group. The relationship between oral contraceptives and breast cancer risk also needs to be examined in other subgroups, such as women with a strong family history of breast cancer and women with previous benign breast disease, using data sets with sufficient numbers of subjects.

In conclusion, the current picture of the relationship between oral contraceptives and breast cancer is rather confusing. It is not clear whether we should be reassured or alarmed, or whether it is just too early to know. Despite the more than 40 epidemiological studies conducted thus far, the end result seems to be more questions than answers.

References

  • Brinton, L. A., R. Hoover, M. Szklo, and F. J. Fraumeni, Jr. 1982. Oral contraceptives and breast cancer. International Journal of Epidemiology 11:316–322. [PubMed: 7152784]
  • (CASH) Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development: Long-term oral contraceptive use and the risk of breast cancer. 1983. Journal of the American Medical Association 249:1591–1595. [PubMed: 6338262]
  • Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development: Oral-contraceptive use and the risk of breast cancer. 1986. New England Journal of Medicine 315(7):405–411. [PubMed: 3736618]
  • Ellery, C., R. MacLennan, G. Berry, and R. P. Sherman. 1986. A case-control study of breast cancer in relation to the use of steroid contraceptive agents. Medical Journal of Australia 144:173–176. [PubMed: 3945217]
  • Fasal, E., and R. S. Paffenbarger. 1975. Oral contraceptives as related to cancer and benign lesions of the breast. Journal of the National Cancer Institute 55:767–773. [PubMed: 1185801]
  • Greenland, S. 1983. Tests for interaction in epidemiologic studies: A review and a study of power. Statistics in Medicine 2:243–251. [PubMed: 6359318]
  • Harris, N. V., N. S. Weiss, A. M. Francis, and L. Polisar. 1982. Breast cancer in relation to patterns of oral contraceptive use. American Journal of Epidemiology 116:643–651. [PubMed: 7137151]
  • Henderson, B. E., D. Powell, I. Rosario, et al. 1974. An epidemiologic study of breast cancer. Journal of the National Cancer Institute 53:609–614. [PubMed: 4369771]
  • Hennekens, C. H., F. E. Speizer, R. J. Lipnick, B. Rosner, C. Bain, C. Belanger, M. Stampfer, W. Willett, and R. Peto. 1984. A case-control study of oral contraceptive use and breast cancer. Journal of the National Cancer Institute 72:39–42. [PubMed: 6363789]
  • Janerich, D. T., A. P. Polednak, D. M. Glebatis, and C. E. Lawrence. 1983. Breast cancer and oral contraceptive use: A case-control study. Journal of Chronic Diseases 36(9): 639–646. [PubMed: 6619259]
  • Jick, S. S., A. M. Walker, A. Stergachis, and H. Jick. 1989. Oral contraceptives and breast cancer. British Journal of Cancer 59:618–621. [PMC free article: PMC2247141] [PubMed: 2713248]
  • Kay, C. R., and P. C. Hannaford. 1988. Breast cancer and the pill—A further report from the Royal College of General Practitioners' oral contraception study. British Journal of Cancer 58:675–680. [PMC free article: PMC2246833] [PubMed: 3219280]
  • Kelsey, J. L., L. R. Holford, C. White, E. S. Mayer, S. E. Kilty, and R. M. Acheson. 1978. Oral contraceptives and breast disease. American Journal of Epidemiology 107:236–244. [PubMed: 629261]
  • Kelsey, J. L., D. B. Fischer, T. R. Holford, V. A. Livolsi, E. D. Moston, I. S. Goldenberg, and C. White. 1981. Exogenous estrogens and other factors in the epidemiology of breast cancer. Journal of the National Cancer Institute 67:327–333. [PubMed: 6943372]
  • La Vecchia, C., A. Decarli, M. Fasoli, S. Franceschi, A. Gentile, E. Negri, F. Parazzini, and G. Tognoni. 1986. Oral contraceptives and cancers of the breast and the female reproductive tract. Interim results from a case-control study. British Journal of Cancer 54(2):311–317. [PMC free article: PMC2001528] [PubMed: 3741766]
  • Lee, N. C., L. Rosero-Bixby, M. W. Oberle, C. Grimaldo, A. S. Whatley, and E. Z. Rovira. 1987. A case-control study of breast cancer and hormonal contraception in Costa Rica. Journal of the National Cancer Institute 79(6):1247–1254. [PubMed: 2961909]
  • Lees, A. W., P. E. Burns, and M. Grace. 1978. Oral contraceptives and breast disease in premenopausal northern Albertan women. International Journal of Cancer 22:700–707. [PubMed: 721325]
  • Lipnick, R. J., J. E. Buring, C. H. Hennekens, B. Rosner, W. Willet, C. Bain, M. P. Stampfer, G. Colditz, R. Peto, and F. E. Speizer. 1986. Oral contraceptives and breast cancer: A prospective cohort study. Journal of the American Medical Association 255:58–61. [PubMed: 3940306]
  • McPherson, K., and J. Drife. 1986. The pill and breast cancer: Why the uncertainty. British Medical Journal 293(6549):709. [PMC free article: PMC1341438] [PubMed: 3094619]
  • McPherson, K., A. Neil, M. P. Vessey, and R. Doll. 1983. Oral contraceptives and breast cancer (letter). Lancet 12/17:1414–1415. [PubMed: 6140506]
  • McPherson, K., M. P. Vessey, A. Neil, R. Doll, L. Jones, and M. Roberts. 1987. Early oral contraceptive use and breast cancer: Results of another case-control study. British Journal of Cancer 56:653–660. [PMC free article: PMC2001913] [PubMed: 3426932]
  • Meirik, O., E. Lund, H. O. Adami, R. Bergstrom, T. Christofferson, and P. Bergsjo. 1986. Oral contraceptive use and breast cancer in young women. A joint national case-control study in Sweden and Norway. Lancet 2:650–654. [PubMed: 2876135]
  • Meirik, O., T. M. Farley, E. Lund, H. O. Adami, T. Christofferson, and P. Bergsjo. 1989. Breast cancer and oral contraceptives: Patterns of risk among parous and nulliparous women—further analysis of the Swedish-Norwegian material. Contraception 39(5):471–475. [PubMed: 2721197]
  • Miller, D. R., L. Rosenberg, D. W. Kaufman, D. Schottenfeld, P. D. Stolley, and S. Shapiro. 1986. Breast cancer risk in relation to early oral contraceptive use. Obstetrics and Gynecology 68:863–868. [PubMed: 3785799]
  • Miller, D. R., L. Rosenberg, D. W. Kaufman, P. D. Stolley, M. E. Warshaver, and S. Shapiro. 1989. Breast cancer before age 45 and oral contraceptive use: New findings. American Journal of Epidemiology 129:269–280. [PubMed: 2912040]
  • Murray, P. P., B. V. Stadel, and J. J. Schlesselman. 1989. Oral contraceptive use in women with a family history of breast cancer. Obstetrics and Gynecology 73:977–983. [PubMed: 2726118]
  • Olsson, H., M. Landin-Olsson, T. Moller, J. Ranstam, and P. Holm. 1985. Oral contraceptive use and breast cancer in young women in Sweden. Lancet 30:748–749. [PubMed: 2858009]
  • Olsson, H., T. R. Moller, and J. Ranstam. 1989. Early oral contraceptive use and breast cancer among premenopausal women: Final report from a study in southern Sweden. Journal of the National Cancer Institute 81(13):1000–1004. [PubMed: 2733043]
  • Paffenbarger, R. S., J. B. Kampert, and H. G. Chang. 1980. Characteristics that predict risk of breast cancer before and after the menopause. American Journal of Epidemiology 112(2):258–268. [PubMed: 7416152]
  • Paul, C., D. C. G. Skegg, G. H. Spears, and J. M. Kaldor. 1986. Oral contraceptives and breast cancer: A national study. British Medical Journal 293:723–726. [PMC free article: PMC1341446] [PubMed: 3094626]
  • Peto, J. 1989. Oral contraceptives and breast cancer: Is the CASH study really negative? Lancet 1:552. [PubMed: 2564078]
  • Pike, M. C., B. E. Henderson, J. T. Casagrande, I. Rosario, and G. E. Gray. 1981. Oral contraceptive use and early abortion as risk factors for breast cancer in young women. British Journal of Cancer 43:72–76. [PMC free article: PMC2010485] [PubMed: 7459241]
  • Pike, M. C., B. E. Henderson, M. D. Krailo, and A. Duke. 1983. Breast cancer in young women and use of oral contraceptives: Possible modifying effect of formulation and age at use. Lancet 2:926–930. [PubMed: 6138501]
  • Piper, J. M., and D. L. Kennedy. 1987. Oral contraceptives in the US: Trends in content and potency. International Journal of Epidemiology 16(2):215–221. [PubMed: 3301706]
  • Ravnihar, B., D. G. Seigel, and J. Lindtner. 1979. An epidemiologic study of breast cancer and benign breast neoplasias in relation to the oral contraceptive and estrogen use. European Journal of Cancer 15:395–405. [PubMed: 436904]
  • Ravnihar, B., M. P. Zakelj, K. Kosmelj, and J. Stare. 1988. A case-control study of breast cancer in relation to oral contraceptive use in Slovenia. Neoplasma 35:109–121. [PubMed: 3352833]
  • Romieu, I., W. C. Willett, G. A. Colditz, M. J. Stampfer, B. Rosner, C. H. Hennekens, and F. E. Speizer. 1989. Prospective study of oral contraceptive use and risk of breast cancer in women. Journal of the National Cancer Institute 81(17):1313–1321. [PubMed: 2769784]
  • Rosenberg, L., D. R. Miller, D. W. Kaufman, S. P. Helmrich, P. D. Stolley, D. Schottenfeld, and S. Shapiro. 1984. Breast cancer and oral contraceptive use. American Journal of Epidemiology 119:167–176. [PubMed: 6695896]
  • Royal College of General Practitioners. 1981. Breast cancer and oral contraceptives: Findings in Royal College of General Practitioners' Study. British Medical Journal 282:2089–2093. [PMC free article: PMC1506508] [PubMed: 6788214]
  • Sartwell, P. E., F. G. Arthes, and J. A. Tonascia. 1977. Exogenous hormones, reproductive history and breast cancer. Journal of the National Cancer Institute 59:1589–1595. [PubMed: 926183]
  • Schlesselman, J. J., B. V. Stadel, P. Murray, S. Lai. 1987. Breast cancer in relation to type of estrogen contained in oral contraceptives. Contraception 36(6):595–612. [PubMed: 3446437]
  • Schlesselman, J. J., B. V. Stadel, P. Murray, and S. Lai. 1988. Breast cancer in relation to early use of oral contraceptives: No evidence of a latency effect. Journal of the American Medical Association 259:1828–1833. [PubMed: 3343791]
  • Smith, P. G., and N. E. Day. 1984. The design of case-control studies: The influence of confounding and interaction effects. International Journal of Epidemiology 13(3):356–365. [PubMed: 6386716]
  • Spirtas, R. 1989. Breast cancer before age 45 and oral contraceptive use: New findings (letter). American Journal of Epidemiology 130:1255–1256.
  • Stadel, B. V., and J. J. Schlesselman. 1986. Oral contraceptive use and the risk of breast cancer in women with a prior history of benign breast disease. American Journal of Epidemiology 123(3):373–382. [PubMed: 3946384]
  • Stadel, B. V., G. L. Rubin, L. A. Webster, J. J. Schlesselman, and P. A. Wingo. 1985. Oral contraceptives and breast cancer in young women. Lancet 2:970–973. [PubMed: 2865503]
  • Stadel, B. V., S. Lai, J. J. Schlesselman, and P. Murray. 1988. Oral contraceptives and premenopausal breast cancer in nulliparous women. Contraception 38(3):287–299. [PubMed: 3168449]
  • Stanford, J. L., L. A. Brinton, and R. N. Hoover. 1989. Oral contraceptives and breast cancer: Results from an expanded case-control study. British Journal of Cancer 60:375–381. [PMC free article: PMC2247191] [PubMed: 2789945]
  • Talamini, R., C. La Vecchia, S. Franceschi, F. Colombo, et al. 1985. Reproductive, hormonal factors and breast cancer in a northern Italian population. International Journal of Epidemiology 14:70–74. [PubMed: 3988443]
  • Thomas, D. B. 1984. Do hormones cause breast cancer? Cancer 53:595–604. [PubMed: 6362820]
  • Trapido, E. J. 1981. A prospective study of oral contraceptives and breast cancer. Journal of the National Cancer Institute 67:1011–1015. [PubMed: 6946243]
  • U.K. National Case-Control Study Group. 1989. Oral contraceptive use and breast cancer risk in young women. Lancet 8645:973–982. [PubMed: 2565515]
  • Vessey, M. P., R. Doll, K. Jones, K. McPherson, and D. Yeates. 1979. An epidemiologic study of oral contraceptives and breast cancer. British Medical Journal 1:1755–1758.
  • Vessey, M. P., K. McPherson, and R. Doll. 1981. Breast cancer and oral contraceptives: Findings in the Oxford-Family Planning Association contraceptive study. British Medical Journal 282:2093–2094. [PMC free article: PMC1506553] [PubMed: 6788215]
  • Vessey, M. P., K. McPherson, D. Yeates, and R. Doll. 1982. Oral contraceptive use and abortion before first term pregnancy in relation to cancer risk. British Journal of Cancer 45:327–331. [PMC free article: PMC2010930] [PubMed: 7041938]
  • Vessey, M. P., J. Baron, R. Doll, K. McPherson, and D. Yeates. 1983. Oral contraceptives and breast cancer: Final report of an epidemiological study. British Journal of Cancer 47:455–462. [PMC free article: PMC2011321] [PubMed: 6849791]
  • Vessey, M. P., K. McPherson, L. Villard-Mackintosh, and D. Yeates. 1989. Oral contraceptives and breast cancer: Latest findings in a large cohort study. British Journal of Cancer 59:613–617. [PMC free article: PMC2247164] [PubMed: 2713247]
  • Yuan, J. M., M. C. Yu, R. K. Ross, Y. T. Gao, B. E. Henderson. 1988. Risk factors for breast cancer in Chinese women in Shanghai. Cancer Research 48:1949–1953. [PubMed: 3349468]
  • (WHO) World Health Organization Collaborative Study of Neoplasia and Steroid Contraceptives. 1990. Breast cancer and combined oral contraceptives: Results from a multinational study. British Journal of Cancer 61:110–119. [PMC free article: PMC1971339] [PubMed: 2404507]

Footnotes

Kathleen E. Malone is a Research Associate at the Fred Hutchinson Cancer Research Center, and a graduate student at the University of Washington, Seattle, Washington.

Copyright © 1991 by the National Academy of Sciences.
Bookshelf ID: NBK234348

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (8.4M)

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...