Present standards for inclusion of women of reproductive age in clinical trials from the Federal Drug Administration and the National Institutes of Health (NIH)1,2 are intended to strike a balance between increasing participation of women of reproductive age who are hormonally intact in clinical trials while limiting the potential for fetal or embryonal risk. They are a response to medical, legal, and ethical concerns about the exclusion of reproductive-age women from research and a subsequent absence of information pertaining to the population of women who would eventually be using the drugs.3 Research grants from the federal government now have substantive oversight to avoid unnecessarily excluding women of reproductive age from trials.4
Liability for negative outcomes of drug use involving fertile women was raised with the tragedies of diethylstilbestrol and thalidomide, which led to regulatory changes requiring proof of efficacy before approval of a drug and to fetal protection policies. Pharmaceutical companies and researchers continue to maintain an interest in avoiding costly liability from teratogenic exposure, although their actual liability is unquestionably limited.5 Conversely, potential liability could also accrue from exclusion when evidence emerges that a marketed drug has more risks, unique risks, or is less effective in women.6
In a unanimous decision in 1991, the United States Supreme Court consideration of International Union versus Johnson Controls Inc.7–9 held that discrimination on the basis of fertility (for fetal protection) was impermissible in a workplace where potential fetal toxins were present. The verdict stated that informed parents should determine how best to protect the potential fetus rather than the company preempting such choice by preventing women from holding certain jobs. The burden for fetal protection was thus placed on exposed individuals. This legal approach has parallels to inclusion of fertile women in clinical trials.
The Federal Drug Administration and NIH guidelines require clinical trials to include a representative sampling of the population, including gender, that will likely use the drug. However, although now included in trial design, women of reproductive age are often excluded if fertile. The requirements for fertility control mechanisms, regardless of the women's sexual activity, impose greater medical burdens and risks for women as a result of contraceptives or other fertility control mechanisms that they might not need or want. Men of reproductive age have no similar restrictions placed on them. Thus, trials fail to address the articulated goal of identifying the impact on the full range of women represented by the population, including hormonally intact women. We review the contraceptive and reproductive requirements in a 36-month sample of Institutional Review Board-reviewed clinical trials. This database was used to assess the gender differences in contraceptive requirements. We discuss implications for women's health research.
Materials and Methods
This is a descriptive study of all clinical protocols presented to the full Institutional Review Board of the Pennsylvania State University College of Medicine, between January 1994 and January 1997. This study was approved by the Institutional Review Board, and two of the authors were members, concurrently (ARL) and before (JMC) study completion. Data collection was completed by one author (JL) with a review of a 20% random sample of the first 100 abstractions for completeness and accuracy. No reabstraction was required, but additional coding was introduced after data collection in order to characterize ambiguous cases into predefined categories. Four hundred ten protocols were coded, of which 96.8% were approved by the Institutional Review Board.
Fifty-three (24%) trials were sponsored by the NIH; 43 (48%) had other government sponsors; and (20.2%) were commercially sponsored. Of the 410 protocols, 69.5% involved diseases or drugs, 8.8% involved a device, and 18.5% involved physiologic processes of potential application to women of childbearing age. No protocol was limited to pregnant women. None of the protocols involved a disease process limited to men. Most protocols involved chronic disease or continuing disease processes (n = 272, 66.3%). Eighty (19.5%) addressed acute processes, and only six (1.5%) addressed rapidly fatal conditions. The average length of study intervention varied widely (Table 1).
Pregnant women were excluded in 95.6% of studies, including studies of physiologic processes or involving Federal Drug Administration-approved drugs such as corticosteroids that are currently used by pregnant women. Exclusions for age less than 18 years were present in 51.2% for men and 66.1% for women. Eight percent of protocols excluded both genders equally over the age of 65.
There were 171 (41.7%) protocols that required contraception or sterility for women, without a stated basis for the requirement. Reproductive exclusions were based on a specific drug in 146 protocols (35.6%), although less than 5% of all drugs being studied were from a class of drugs with demonstrated human teratogenicity (Class X). Only 8.5% of all protocols had no restrictions. Of the studies requiring contraceptive use, 28 protocols included recommendations for male contraception and none required contraception for male subjects. All protocols using drugs with established human teratogenicity required sterility. No other relationship was found between class of drug studied, known teratogenicity, and the type or level of contraception required. The menu of contraceptives and the language used depended more on the individual sponsor than on the drug studied.
A wide variety of contraceptives were recommended (Table 2). One hundred thirty-eight protocols required from women signature certification of contraceptive use. A medical or surgical record of sterility was required in 10.2%. Only 0.7% of protocols requiring contraception did not require signature certification or documentation. Two hundred thirty protocols required signature certification that female participants understood the contraceptive requirements. One hundred one protocols required signature certification regarding contraception at two places on the consent form, and two protocols required three signatures. Two hundred thirty-four protocols required signature certification that participants were not pregnant at enrollment, thereby averaging two to four signatures regarding contraception or current nonpregnant status by women. There were no signature certification requirements for male subjects even for protocols where contraception was recommended. No protocol included provisions for decreased pregnancy risk based on sexual orientation. Only 15 protocols included abstinence as an acceptable form of contraception. None recognized celibate women as satisfying contraceptive requirements. All protocols required pregnancy tests of women before the start of the study.
The specter of potential liability for fetal harm has spurred continued requirements for contraception or sterility for entry of women into clinical trials. This requirement creates disproportionate burdens and access by gender for clinical trials. Nonsterile, hormonally intact men can participate whereas nonsterile, hormonally intact women cannot. Once approved, however, drugs are prescribed and sold virtually without regard for the gender or reproductive capacity of the patient.10
The Belmont Report notes “In balancing these different elements, the risks and benefits affecting the immediate research subject will normally carry special weight. On the other hand, interests other than those of the subject may on some occasions be sufficient by themselves to justify the risks involved in the research, so long as the subjects' rights have been protected. Beneficence thus requires that we protect against risk of harm to subjects and also that we be concerned about the loss of substantial benefits that might be gained from research.”11
There are benefits and harms for women as a class as well as individual participants in clinical trials. Women pay for clinical research through their use of drugs and their payment of federal taxes, so they have a valid claim to an accounting of the benefits and harms of requirements for contraception or sterility. Given that women use more drugs and have more reported side effects than men,12,13 they have a justifiable claim to accurate information about drug metabolism and side effects in intact hormonal systems. The paucity of information regarding pharmakokinetics of normally cycling women remains a major area for unrecognized drug side effects, drug-to-drug interactions, and efficacy interactions, including those with contraceptives.3
The former complete exclusion of fertile women led to more deaths of women with human immunodeficiency virus than men and eventual revision of exclusionist policies.14 The replacement of such exclusions with inclusion of only women who use contraception, particularly hormonal contraceptives, noted in this review is problematic. Whether current Federal Drug Administration policies and guidelines that allow such reproductive restrictions are sufficient to protect the interests of women as a class is a subject of continuing controversy.
The application of sponsor-specific contraception menus to clinical trials for entry of fertile women presents problems. First, there is disproportionate exposure to contraceptives by women who would not otherwise use contraception.15 The menu approach also has the negative side effect of requiring or recommending directly contraindicated contraceptives. For example, oral contraceptives (OC) were part of a menu of contraceptive requirements for a new drug for patients with hepatic failure in an intensive care setting—clearly a contraindication. Second, use of hormonal contraception as part of the menu prevents further explication of variations in pharmakokinetics attributable to normal menstrual cycling. Finally, contraception was unnecessary in one third of the protocols studied (24-hour to 2-day drug use) because timing to menstrual cycles would prevent potential exposure during pregnancy. The use of a menu approach to contraception by trial sponsors, where one of a menu of options is required, potentially increases the risk women take for access to drug trials, while limiting important information for women as a class. In addition, the sense of legal and economic security of researchers and institutions achieved by menu approaches could discourage creative development of other cycle timing approaches to include hormonally intact, fertile women.
We question a priori assumption that there is potential for a fetus to exist that might bear some risk. The most common duration of intervention in protocols we reviewed was less than one menstrual cycle. Unprotected cycle fecundity is at best 20%, so fetal exposure without contraception is unlikely. Furthermore, the women might be celibate at the time of the study, the drug might interfere with reproduction, partners might be sterile or the same sex, the setting might not be conducive for reproduction (such as an intensive care unit), and virtually all studies required negative pregnancy tests for entry. None of these were recognized as valid exceptions to contraceptive requirements in any protocol in this review. The actual risk of fetal exposure was small even without contraception.
Drugs studied by protocols in this review included 5% that were in a class of known human teratogenic drugs, primarily chemotherapeutic and dermatologic agents. All those studies had contraceptive or sterility requirements, including protocols where the treatment or disease itself resulted in sterility. For women losing their fertility because of a malignancy, this is unnecessarily painful.
The elucidation of potential for fetal harm is made more difficult because of the paucity of data regarding fetal risk. Teratogenicity requires exposure at a specific fetal developmental stage, not universally throughout pregnancy, which varies by agent. Most drugs in this review were in classes of drugs never studied in human pregnancy. Since most drugs already approved have not been tested in women of reproductive age, much less pregnant women, the ability to estimate teratogenicity is limited. The actual risk of fetal exposure to medications that might cause fetal harm or pregnancy loss, with cycle-sensitive protocol design and without these menus of contraceptive requirements, is negligible.
The near complete exclusion of pregnant women found in this review illustrates the impact that legal and economic concerns of sponsors about fetal outcome have beyond issues of teratogenicity. An explanation for exclusion stated that vigorous exercise and amount of blood drawn was considered unnecessary stress on pregnancy for a 2-week study on insomnia involving treadmill exercise, venipuncture, and an electrocardiogram. Another study on insomnia with 8 consecutive days of electrocardiographic studies, electroencephalographic studies, sleep deprivation, and venipuncture excluded pregnant women because sleep deprivation and amount of blood drawn was considered unnecessary stress on pregnancy. In this review, concerns about liability for any fetal harm extended to prohibiting studies of physiologic processes with pregnant women doing things they might normally do.
“To respect autonomy is to give weight to autonomous persons' considered opinions and choices while refraining from obstructing their actions unless they are clearly detrimental to others.”11 The present near-universal requirement of contraception or sterility for entry of women into clinical trials calls into question this axiomatic basis of research and health care. There is an underlying assumption that reproductive choice must be controlled for women but not for men. In this review, even when studies were restricted to men because of known teratogenicity of the drug studied, the reproductive control required was for female partners of subjects included the following agreement: “to avoid potential exposure of this drug to developing fetus, I understand that I must take precautions such as the use of condoms or use of birth control pills, diaphragm, tubal ligation, or hysterectomy by my partner.” No mention of abstention, vasectomy, or sterility as a requirement to prevent fetal exposure by male subjects was included.
The potential for fetal harm or altered fertility from drug exposure is not limited to one gender. The disproportionate attention women receive ignores the potential harm to spermatogenesis, fertility, and the impact on offspring for male participants, as was raised with lead in the Johnson Controls case. This disproportionate attention to female subjects has the potential for inadequate exploration of harms to male participants and their offspring. Extending menu contraception requirements to male subjects as a solution, however, would deny both genders adequate information and the acknowledgment of capacity and responsibility to make choices regarding potential exposure to offspring.
The language used in the protocols reviewed, as well as the number of times women must certify their understanding of the fertility requirements, implies that potential fertility was a reason women should not participate. “If there is a chance I could become pregnant during this study, I cannot take part or I must use a highly effective means of birth control” is a common consent for women yet “I understand I must practice an effective means of birth control” was the strongest language used for men. The requirement of multiple signatures certifying comprehension of these requirements implies disbelief in women's ability to make decisions regarding their own health behavior. Furthermore, the exclusion of abstinence as a form of contraception is particularly alarming. Abstinence has been specifically mentioned as an acceptable form of contraception in federal guidelines and is a major staple of education to avoid unwanted pregnancies nationally.
The present approach makes a travesty of the principles of informed consent. The disproportionate treatment of this issue implies that women but not men are incapable of making reasoned choices that adequately protect their own interests, or, more accurately, it assuages the insecurities of the institutions and researchers involved. Excluding abstinence as a contraceptive choice implies that women are too untrustworthy or are constantly at risk of unwanted sexual assault for this to be a reasoned choice. The signatures and restrictions create an environment of distrust of women's decision-making capacity and imply that fertility makes them less desirable research subjects (whereas the fertility of males does not). As an ethical basis for the researcher-subject relationship, implied distrust is undesirable. Yet the incentive for researchers to support sponsor language and requirements rather than to argue with sponsors for more inclusive and reasoned approaches, is powerful. As one researcher noted “they will just take it to an institutional review board that agrees, and we lose.”
Support for an alternative—a single, comprehensive consent for men and women regarding fertility control—comes from Johnson Controls. Differential treatment of employees based on reproductive capacity was not acceptable because adults have the right to hear information about risks and to make their own choices about reproduction. Although fertile women are no longer denied access to research protocols, the sponsors in this review differentiated treatment of research subjects based on gender and reproductive capacity. A better approach would incorporate full disclosure of known or suspected reproductive and fetal risks with a single consent form.16
The effort to control one's choice regarding exposure of fetuses will ultimately fail, as it lies with the actions of the subject—male or female. That decision, and the decision of a woman or family to consider abortion, is based on values and circumstances that individual investigators cannot fully understand, and so they must depend on the individual to deliberate and choose. Because abortion is a deep personal issue for most individuals, and because of society's failure to identify a philosophic, religious, or medical common ground, standard policy has been to avoid discussions of individual rights to make such a choice as an integral part of the inclusion of reproductive-age persons in research. If pregnancy occurs, to a female partner of a research subject or to a female research subject, it is ultimately the pregnant woman's choice about the disposition of her pregnancy and her ability to withhold, confer, or having once conferred, withdraw the status of being a subject from the fetus.17 There is no escape from this final arbitration by the woman herself, despite the burdensome efforts to double, triple, and quadruple assent to fertility control in the research protocol consent forms for women participants.
“An injustice occurs when some benefit to which a person is entitled is denied without good reason or when some burden is imposed unduly. Another way of conceiving the principle of justice is that equals ought to be treated equally.”11 The just application of the burdens and benefits of research requires that those who bear the burden should also have access to the benefits. The disproportionate inclusion of men could potentially be viewed as a greater burden than benefit to men, because men do not use approved drugs to the same degree as women. Men have benefitted from a more complete picture of male physiology with fewer drug reactions and side effects. This gender-biased understanding of physiology has defined standard treatment protocols for cardiovascular disease, transplantation, and other medical entities that have come under question as translatable to female physiology.
Women stand to benefit from information about the bioavailability and metabolism of drugs in varying hormonal environments as well as a more complete understanding of the physiology of disease in women. They have a right to benefit from the ability of research to identify strategies likely to be successful in women for the control and cure of disease. This information should be obtained in the preapproval setting under the same oversight that covers the majority of research on male subjects rather than in Phase IV studies or by retrospective review after considerable exposure of women (including fertile women) in postapproval settings. Equitable research restrictions must be based on actual risk to individual women not the potential risk to a potential fetus.
The Canadian Tri-Council Policy Statement of 1998 summarized the obligations of researchers, institutions, and institutional review boards as having “important roles to play in ensuring a fairer distribution of the benefits and the burdens of research … distributive justice imposes on (them) a duty not to act in a discriminatory fashion.”18 Menu contraceptive requirements skew the population studied and discriminate against male or female study participants. The time has come for a balanced approach to research for both genders with full-disclosure informed consent instead of restrictions and careful elucidation of methods and timing that avoid potential fetal risk and encourage inclusion of hormonally intact women.
1. National Institutes of Health. NIH guidelines on the inclusion of women and minorities as subjects in clinical research. 59 Federal Register 14508. Washington, DC: March 28, 1994.
2. National Institutes of Health Revitalization Act of 1993. United States Public Law. 103rd Congress, first session, p. 107, statute 122. Washington, DC: Government Printing Office, 1993.
3. Merkatz RB, Temple R, Sobel S, Feiden K, Kesler DA. Women in clinical trials of new drugs—A change in Food and Drug Administration policy. N Engl J Med 1993;329:292–6.
4. Office of Research on Women's Health. Recruitment and retention of women in clinical studies. NIH publication no. 95-3756. Washington, DC: National Institutes of Health, 1995.
5. Clayton EW. Liability exposure when offspring are injured because of their parent's participation in clinical trials. In: Mastroianni AC, Faden R, Federman D, eds. Women and health research: Ethical and legal issues of including women in clinical studies. Vol. 2. Washington, DC: National Academy Press, 1994:103–10.
6. Clayton EW. Legal considerations. In: Mastroianni AC, Faden R, Federman D, eds. Women and health research: Ethical and legal issues of including women in clinical studies. Vol. 2. Washington, DC: National Academy Press, 1994:128–70.
7. International Union versus Johnson Controls. 1991:n159 111 Supreme Court 1196.
8. Annas GJ. Fetal protection and employment discrimination—The Johnson Controls case. N Engl J Med 1991;325:740–3.
9. Coyle M. Fetal protection ruling buoys rights groups. National Law J 1991;5:22.
10. Herz SE. Don't test, do sell: Legal implications of inclusion and exclusion of women in clinical disease trials. Epilepsia 1997;38:42–9.
11. The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. The Belmont report: Ethical principles and guidelines for the protection of human subjects of research. Department of Health Education and Welfare Publication no. (OS) 78-0013, (OS) 78-0014. Washington, DC, 1979.
12. Swann JR, Reinstein PF. Annual adverse drug experience report 1993. Washington, DC: Food and Drug Administration, 1993.
13. Schmucker DL, O'Mahony MS, Vesell ES. Women in clinical drug trials: An update. Clin Pharmacokinet 1994;27:411–7.
14. Edgar H, Rothman DJ. New rules for new drugs: The challenge of AIDS to the regulatory process. Milbank Q 1990;68:111–41.
15. Physician's desk reference. 52nd ed. Montvale, NJ: Medical Economics Company, 1998.
16. Nolan MT, Pressman EK, Starklauf BL, Gregory RE. Consent documents, reproductive issues, and the inclusion of women in clinical trials. Acad Med 1999;74:275–81.
17. McCullough LB, Cervenak FA. Ethics in obstetrics and gynecology. Oxford, England: Oxford University Press, 1994:176–8.
18. Tri-Council Policy Statement: Ethical conduct for research involving humans: August 1998, Ottawa, Canada: Medical Research Council of Canada, Natural Science and Engineering Research Council of Canada, and Social Sciences and Humanities Research Council of Canada, 1998.