IRB: Ethics & Human Research
Ethical, Evidence-Based Guidelines for Contraceptive Use in Research
Research ethics review committees commonly approve protocols that require women of childbearing potential to use one or two means of birth control while enrolled in clinical drug trials, and sometimes the means of birth control are specified (e.g., oral contraceptives or an intrauterine device). The purpose of a contraception requirement is to protect a developing fetus from exposure to drugs that might harm it by preventing pregnancy for the duration of the trial. Some commentators contend that it is ethically unacceptable to require women of childbearing potential to use contraception as a condition of trial participation because such a requirement violates women’s autonomy and is inconsistent with women’s well-being, as some forms of contraception carry health risks for women. These risks are all the more significant when contraceptive use is unnecessary, as when there is no risk of pregnancy. For others, mandated contraception for women of childbearing potential is ethically defensible on the grounds that while women might benefit from participating in clinical drug trials, they should not expose a potential fetus to drugs that could be harmful.
In an effort to balance the duty to respect women’s autonomy and well-being and the (contested) duty to protect potential fetuses from possible research harms, many research ethics review committees have developed policies regarding contraceptive use for women who participate in clinical trials. One published policy is that of the IRB at the University of Nebraska Medical Center (UNMC).1This policy uses the U.S. Food and Drug Administration’s (FDA) categories for prescription drug labeling for drugs used during pregnancy “as thresholds for levels of required contraception.”2The underlying reasoning is that the current FDA labeling categories (A, B, C, D, X)3—in place to protectactual fetusesfrom potential harm in the context of therapeutic treatment—can reasonably be used to protectpotential fetusesfrom potential harm in the context of clinical research.
In this paper we describe the FDA’s labeling categories in the order of category A, category B, and categories D and X together. We then turn to category C, as discussion of this category requires the prior discussion of categories D and X. For each category, we present the UNMC IRB’s policy regarding contraceptive use for trial participants that is based on those categories, followed by our policy recommendations, which use the UNMC IRB policy as both a starting point and a template. Our recommendations reflect the evidentiary standards held in common between the FDA’s current labeling rule and the agency’s proposed changes to that rule, which we describe in the final section of the article.4We conclude by summarizing the differences between our proposed policy and the UNMC IRB policy, and by clearly outlining the ethical concerns motivating these differences.Table 1provides a summary comparison of the current FDA categories, the UNMC IRB policy, and our recommendations for an ideal institutional policy regarding the use of contraception by participants in clinical drug trials.
Category A Drugs
Drugs in this category are those in “which adequate and well-controlled studies in pregnant women have failed to demonstrate a risk to the fetus in the first trimester of pregnancy (and there is no evidence of a risk in later trimesters).”5Examples of category A drugs include folic acid and thyroglobulin, a treatment for hypothyroidism.
The UNMC IRB’s policy on contraceptive use by research participants strongly discourages the mandated use of contraception for studies involving drugs in category A, but nevertheless allows investigators to petition the IRB to have contraception as an inclusion criterion. If the proffered justification is accepted by the IRB, then the policy states that the research participant and his/her partner “should use one reliable form of contraception while on study and for a stipulated number of months afterwards.”6The UNMC IRB considers the following forms of contraception to be reliable: condoms with spermicide, diaphragm or cervical cap with spermicide, IUD, and hormonally-based contraception.
We believe this policy choice is inconsistent with the duty to respect the autonomy and well-being of women, and that it cannot be justified by referring to the contested duty to protect potential fetuses from potential harm. For drugs that have been demonstrated to pose no risk to the fetus in adequate, well-controlled studies in pregnant women, it is unclear why any researcher should have the option of mandating contraception (which, as noted above, may entail risks to the pregnant woman). The UNMC IRB’s policy on contraceptive use in research is generally described as a solution to “the dilemma of including women of childbearing potential in research and protecting potential fetuses from harm.”7For drugs with no potential to harm fetuses, however, this dilemma does not exist. As such, allowing researchers to petition the IRB to mandate contraceptive use because the research involves category A drugs seems misguided and in conflict with the ethical rationale underlying the policy, as there is no dilemma. Category A drugs have a research-supported track record of not demonstrating a risk to the fetus. Thus, it follows that “use of birth control” as an inclusion criterion cannot be justified by the fact that the research involves a drug in category A, and the use of such an inclusion criterion should be prohibited.8
In response one might argue that mandating the use of contraceptives in research involving drugs that have not been shown to harm fetuses should nonetheless be an option, “just in case.” On this view, the FDA categories might be imperfect, or there might be some good though idiosyncratic reason to believe the drug could pose an increased risk to the fetus. However, such an argument jettisons the strategy of using the information provided by the FDA as a guideline for policies regarding contraceptive use in research—the underlying rationale for the UNMC policy. Evidence indicating that the possibility of fetal harm is remote for drugs in category A makes this the most robust of all the categories. If this information is not to be trusted and accepted as authoritative, then there is little rational basis for using the FDA policy as a framework.
Ideal Policy for Category A.When previous adequate and well-controlled studies in pregnant women fail to demonstrate that a drug poses a risk to the fetus, IRBs should not require women of childbearing age who participate in research to use contraception.
Category B Drugs
Drugs are placed in this category because there are no adequate and well-controlled studies in pregnant women and animal studies show no risk of fetal harm, or there are animal studies that show an adverse effect, but adequate, well-controlled studies in pregnant women have failed to demonstrate a risk to the fetus.9Examples of category B drugs include acetaminophen and insulin.
The UNMC IRB policy does not require women trial participants to use contraception for trials of drugs in category B. However, the policy permits investigators to mandate contraceptive use, and they are not required to justify this requirement to the IRB.10In our view, allowing researchers to unilaterally mandate the use of contraception in clinical trials of drugs in category B is problematic. As a whole, the institutional policy purports to balance the need to protect potential fetuses from potential harms against the need to include women of childbearing potential in clinical trials while respecting their autonomy. By allowing researchers to require the use of contraception when there is no positive evidence of risk to human fetuses, the balance is lost in favor of unwarranted fetal protection. By giving researchers the green light to require contraceptive use, the policy inappropriately gives priority to protecting potential fetuses from harm rather than to respecting the autonomy and well-being of women trial participants.
Ideal Policy for Category B.A more balanced approach to research involving women who are sexually active and who could get pregnant would require researchers to justify to the IRB the need for abstinence or contraceptive use for trials of drugs for which a) animal studies show no risk of fetal harm and no adequate, well-controlled studies in pregnant women have been conducted, or b) animal studies show some risk of harm, but adequate, well-controlled human trials show no risk to fetuses. If the IRB accepts the justification (and the standard for so doing should be high given the limited evidence of fetal risk from animal studies alone), then the inclusion criteria for the drug trial could legitimately include sexual abstinence or the use of reliable contraception (e.g., condoms with spermicide, diaphragm or cervical cap with spermicide, IUD, or hormonally-based contraception). Under this provision, as required, women research participants should be advised to refrain from attempts to get pregnant and invited to choose between sexual abstinence or the use of reliable contraception. Options for dealing with an unexpected, unwanted pregnancy due to failed contraception should also be discussed as part of the consent process.
Categories D and X
In category D, there is positive evidence of potential fetal harm in adequate and well-controlled or observational studies, but the potential benefit to the woman of taking the drug may outweigh the risk to the fetus. Examples of drugs in this category include irinotecan, a chemotherapy drug. There is also positive evidence of fetal risk for category X drugs from studies in animals or humans, and this risk is deemed to outweigh any potential benefits to women. Drugs in category X are contraindicated for women who are or may become pregnant. Such drugs include isotretinoin (Accutane) to treat cystic acne and the sedative thalidomide, once used to treat morning sickness.
For women research participants who are sexually active and could get pregnant, the UNMC IRB policy requires the use of two reliable forms of contraception for clinical trials of drugs in category D. For clinical trials of drugs in category X, the UNMC IRB policy requires that participants follow the manufacturer’s recommendations for contraception. For Accutane, which is in category X, the manufacturer recommends, among other measures, “that two reliable forms of contraception be used simultaneously unless abstinence is the chosen method.”11Other drug manufacturers may not include abstinence among their recommended contraceptive options for drugs in category X.
Ideal Policy for Categories D and X.Our recommendations for contraceptive requirements for research involving category D and X drugs differ very little from the UNMC IRB requirements. For trials of drugs with positive evidence of fetal risk deriving from well-controlled or observational studies involving pregnant women, but where the benefit for pregnant women may outweigh the fetal risk (category D), we recommend that research participants who are sexually active and could get pregnant follow the drug manufacturer’s recommendations. When such is not available, we recommend abstinence or the use of two forms of contraception.12We recommend the same approach for trials of drugs with positive evidence of fetal risk in well-controlled studies involving pregnant women or animals, where the potential benefits to pregnant women do not outweigh the fetal risks (category X).
Some may be concerned that researchers and research sponsors would be reluctant to undertake clinical trials of drugs with positive evidence of fetal risk if trial participants capable of getting pregnant through sexual activity were allowed to opt for abstinence. However, we believe that such reluctance on the part of researchers and sponsors would be unreasonable. Our recommendations with respect to such drugs are effectively no different from those of the UNMC IRB, which for all research (including research involving category D and X drugs) allows abstinence when participating in sexual activity could lead to pregnancy. The only difference is that we recommend that abstinence be explicitly mentioned as an option in research consent forms. The UNMC IRB consent forms for prospective research participants state that contraceptive requirements only come into effect “If you [the participant] are sexually active or become sexually active and can get pregnant.” This caveat means that contraceptive use in research is not required for those who are not sexually active or for those who are sexually active but cannot get pregnant. In this way, the UNMC IRB research consent form allows for abstinence as an alternative to contraceptive use (irrespective of the drug category) without making a clear statement to this effect.
As stated above, the manufacturer of Accutane explicitly recommends the simultaneous use of two reliable forms of contraception or abstinence during the treatment and for a month afterward.13If the manufacturer of a category X drug can include a clear statement recognizing abstinence as a legitimate precaution against pregnancy, then there is no reason why this legitimate option should not be stated in consent forms for clinical trials of drugs with positive evidence of fetal risk.
During the consent process, prospective research participants should also be provided with information about options for dealing with an unexpected, unwanted pregnancy.
In category C there are no adequate, well-controlled studies in pregnant women, and risk to the fetus has not been ruled out. There may or may not be animal studies that show a risk of fetal harm. Because of these evidence gaps, there is greater uncertainty about fetal effects involving drugs in category C, as compared to drugs in categories A, B, D and X. Examples of category C drugs include the decongestant pseudoephedrine and some antidepressants. In a research context, first-in-human trials of new drugs would be in category C.
According to the UNMC IRB’s policy, investigators can require sexually active women who could get pregnant to use one or two forms of contraception in studies of category C drugs. In making this decision the investigator must take into consideration the population participating in the trial, the nature of the study and its duration, and the perceived risk to potential fetuses. The IRB can disagree with the investigator’s decision to mandate one or two forms of contraception and require the investigator to justify his/her decision.
Again, this policy choice is depicted as striking a balance between respecting the autonomy of research participants and protecting potential fetuses from potential harm. Realistically, however, requiring the use of two forms of contraception tips the balance clearly in favor of fetal protection. The autonomous choices of women research participants are not respected when the research ethics review committee approves a protocol that requires two forms of contraception, given that there is no positive evidence of risk to a potential fetus from human studies (though there may be animal studies showing adverse effects).
A requirement for the use of two forms of contraception suggests that the contraceptive choices of women research participants will likely be unreliable, and so redundancies must be built in. In such a scenario, a category C drug is treated as analogous to a category D or X drug, without the evidence to support such a risk-averse position. This position appears excessive given that “balance” is the stated goal of the UNMC IRB policy on contraceptive use in research. Category C stands between two poles. On the one hand, there are drugs in categories A and B for which there is generally good evidence that contraception is not required. On the other hand, for drugs in categories D and X, there is good evidence supporting a more cautious approach that includes mandating sexual abstinence or two forms of contraception for women who are sexually active and may get pregnant. There is less evidence for category C drugs regarding fetal risk. A balanced position for category C would require abstinence or one reliable form of contraception.
Ideal Policy for Category C.We recommend a policy of sexual abstinence or the use of one reliable contraceptive method as inclusion criteria for trials involving category C drugs. Similarly, any future IRB policy aligned with a new FDA policy should allow abstinence or the mandated use of a single contraceptive for trials of drugs that have not established a risk to the fetus in adequate, well-controlled studies in pregnant women. As with drugs in other categories, all research participants who are sexually active and could become pregnant should be provided with information about options for dealing with an unexpected, unwanted pregnancy.
Though we agree with the UNMC IRB’s insight that the current FDA use-in-pregnancy categories for prescription drugs are a helpful guide for developing a policy regarding contraceptive use during clinical drug trials, we have argued that the IRB’s policy gives priority to protecting potential fetuses at the expense of respecting the autonomy and well-being of women trial participants. We also believe that a sound policy regarding contraceptive use while enrolled in a drug trial should apply to both male and female trial participants. Too often, contraceptive policies focus on women only, with insufficient attention given to potential male contributions to fetal harms. Consider, for example, the possibility that some chemicals binding to the sperm might adversely affect the developing fetus. Having said this, our recommendations focus narrowly on the mandated use of contraception by women of childbearing potential because women rather than men have historically been excluded from enrolling in clinical research. Mandatory contraception in research is part of a wider phenomenon of treating women research participants differently from male research participants.
A second important consideration is that a sound policy regarding the use of contraception while enrolled in a drug trial would not include women who are not at risk of getting pregnant in the category of “women of childbearing potential.” This category is typically understood in generic terms to refer to all females who are postpuberty and premenopause. There are many women in this range, however, for whom a mandatory abstinence or contraception policy is clearly unnecessary. Consider, for example, women who are not sexually active (e.g., nuns) or who are not sexually active in a heterosexual relationship (e.g., lesbians). A policy of mandated contraception for these groups sends a paternalistic message of mistrust that is out of keeping with the general norm of recognizing research participants as persons who are capable of making autonomous decisions.
But what about women who are sexually active in a heterosexual relationship where pregnancy might occur? Our starting assumption is that a sound policy will respect women’s autonomy and well-being. As such, women research participants should be permitted to choose sexual abstinence or a method of birth control that “best coincides with . . . lifestyle, values, and priorities.”14We recognize, however, that available data on a drug’s effects on a gestating fetus can shift the balance between the autonomy and well-being of women trial participants and the interests of potential fetuses.
The FDA’s proposed rule concerning labeling of prescription drugs for use during pregnancy would eliminate the current five letter categories that characterize the fetal harm/benefit profile of drugs, as these categories have long been criticized as being misleading and overly simplistic. For example, some suggest that health care providers might falsely believe that the lettering system somehow corresponds to degrees of risk, rather than differences in the weighting of potential harms and benefits. The letter categories have also been judged to inadequately reflect distinctions between human and animal data, and differences in severity, frequency, and type of potential toxic effects on the fetus.
In place of the letter categories, the FDA proposed a more detailed statement of the risks to the fetus, a summary of clinical considerations, and a comprehensive synopsis of relevant research findings from both animal and human studies. The proposed format is to include three sections.15Section one, thefetal risk summary, will give a risk conclusion describing the likelihood that the drug increases the risk of structural anomalies, fetal and infant mortality, impaired physiologic function, and growth alteration. The risk summary will also state whether the risk conclusion is based on human or animal data. If the conclusion is based on human data, the summary will also outline the most important data concerning the drug’s effects on the fetus, including, for example, information about the incidence, reversibility, or risks associated with certain gestational periods. The second section,clinical considerations, will address what is known about fetal risks due to exposure to the drug before the woman realizes she is pregnant, information about prescribing the drug to pregnant women (dosing adjustments, adverse reactions, recommended interventions, risks from the disease the drug is meant to treat), and information about effects of the drug on the woman and fetus during labor and delivery. The third section ondatawill summarize the human and animal studies on the use of the drug in pregnancy.
It is unclear if the FDA’s proposed rule will take effect as a final rule. Thus, the current letter categories regarding labeling for prescription drug use during pregnancy are binding as federal regulatory requirements. By applying these categories to the research context, the UNMC IRB aims to have the level of required contraceptive protection for women trial participants correspond in some measure to the evidence of risk to potential fetuses, and to the harm-benefit profile of the study drug(s). The UNMC IRB policy on contraceptive use in research is innovative insofar as it focuses on the characteristics of the drugs used in a clinical trial instead of focusing exclusively on the characteristics of the prospective research participants (e.g., women of childbearing potential). Moreover, the policy incorporates a number of important insights: 1) that contraceptive requirements should apply equally to men and women; 2) that contraceptive requirements should not apply to women who are not sexually active or to women who cannot get pregnant; 3) that the levels of protection provided topotentialfetuses should not exceed the levels of protection provided toactualfetuses; and 4) that in the context of research, as in the context of therapy, it is important to weigh potential harms against potential benefits, rather than considering potential harms to the fetus in isolation from potential benefits to the woman.
We agree with the UNMC IRB on these points. Nevertheless, in our view, there are a number of difficulties with some of the policy details. When it is necessary to require sexual abstinence or contraception in clinical research, the requirements should not be disproportionate to the risk of harm to potential fetuses. Excessive contraceptive requirements send the message that it is crucially important that pregnant women not be research participants in order to protect fetuses. This message is harmful and should not be perpetuated. The exclusion of pregnant women from research leads to a dangerous lack of knowledge about the safety and efficacy for pregnant women of drugs used during pregnancy, as well as the risks of harm that trial drugs pose to a developing fetus. Lyerly, Little, and Faden contend that there are at least four dimensions of harm caused by the exclusion of pregnant women from research: 1) evidence about the effectiveness of medical treatments for pregnant women will not be obtained; 2) information about the safety and effectiveness of drugs used on the fetus will not be obtained; 3) pregnant women will be undertreated if they avoid drugs they need for their own health because they are unsure about how their fetus will be affected; and 4) pregnant women will be denied the occasional benefits that derive from trial participation.16
In response to these concerns, our proposed policy on contraceptive requirements in research differs from the UNMC IRB policy in a number of areas. First, in research trials involving drugs for which there is no evidence of harm to the fetus (category A), we eliminate the option of mandating abstinence or contraception solely because the research involves the use of such drugs. Second, in trials of drugs for which a) animal studies show no risk of fetal harm and there are no adequate, well-controlled human studies, or b) animal studies show some risk of harm, but adequate, well-controlled human trials show no risk to fetuses (category B), we allow the researcher to require abstinence or one form of contraception for women who are sexually active and can get pregnant, but only with prior approval from the research ethics review committee. This contrasts markedly with the UNMC IRB policy, which allows investigators to make this decision unilaterally for drugs in category B. Third, in trials of drugs for which there are no adequate, well-controlled studies involving pregnant women and for which animal studies either have not been conducted or show an adverse effect (category C), we recommend that women research participants who are sexually active and can get pregnant be required to abstain from sexual activity or to use one reliable form of contraception. We thereby eliminate the policy option of requiring two forms of contraception for research involving drugs of this type.
With these recommendations, we attempt to find an appropriate balance between the interests of potential fetuses and the autonomy and well-being of women. In very general terms, we propose scaling back some of the more onerous requirements of the UNMC IRB policy for mandating contraception. At the same time, we want to preserve that IRB’s insight that the FDA drug use-in-pregnancy information is a useful guide for contraceptive requirements for women who participate in drug trials. Such requirements in research should accurately reflect the risks presented in the drug use-in-pregnancy guidelines and should only apply to those who are sexually active in heterosexual relationships and could potentially become pregnant.
For personal, financial, or other reasons, contraceptive requirements can be onerous for potential research participants. We should be mindful of this and thus careful about developing policies or procedures that might deter women who are postpuberty and premenopause (irrespective of whether they can bear children) from participating in research.
Further, mindful of the fact that the FDA has published a proposed rule to eliminate the current letter-based labeling categories, our policy recommendations aim to reflect the evidentiary standards held in common between the FDA’s current and future labeling requirements regarding the use of prescription drugs during pregnancy. In this way, our recommendations can still be useful if the proposed regulation is adopted.
We would like to thank the members of the Novel Tech Ethics (https://www.noveltechethics.ca) research team for helpful comments on an earlier draft and Robyn MacQuarrie for help researching and developing the ideas in this article.
Chris Kaposy, PhD,is Assistant Professor of Health Care Ethics, Memorial University of Newfoundland, St. John’s, NL, Canada; andFrançoise Baylis, PhD,is Professor and Canada Research Chair in Bioethics and Philosophy, Dalhousie University, Halifax, NS, Canada.
1. Schonfeld TL and Gordon BG. Contraception in research: A policy suggestion.IRB: Ethics & Human Research2005;27(2):15-20.
2. See ref. 1, Schonfeld and Gordon 2005, p. 16.
3. U.S. Food and Drug Administration. Proposed Rule. Content and format of labeling for human prescription drug and biological products; Requirements for pregnancy and lactation labeling.Federal Register2008;73(104):30831-30868.
4. See ref. 3, U.S. Food and Drug Administration 2008.
5. See ref. 3, U. S. Food and Drug Administration 2008, p. 30832.
6. See ref. 1, Schonfeld and Gordon 2005, p. 17.
7. See ref. 1, Schonfeld and Gordon 2005, p. 15.
8. In cases in which an intervention (e.g., radiation, herb, or other drug) is coupled with a drug that is safe for the fetus, a separate judgment about contraceptive requirements for research participants should be made on the basis of the risk posed by the additional intervention. In such cases, mandated contraception might be justifiable because of the risk associated with the additional intervention, but not because of any risk associated with the drug that is safe for the fetus.
9. See ref. 3, U. S. Food and Drug Administration 2008, p. 30832.
10. See ref. 1, Schonfeld and Gordon 2005, p. 17.
11. Hoffman-LaRoche Limited. Product monograph: Accutane Roche. 2006,https://www.skintherapyletter.com/download/accutane_pm.pdf, p. 5. The product monograph for Accutane also advises that patients “must have two negative pregnancy tests before starting ‘Accutane’ Therapy” and that “all female patients of childbearing potential treated with ‘Accutane’ have regular monthly pregnancy tests during treatment and one month after the discontinuation of treatment” (Hoffman-LaRoche Limited 2006, p. 5).
12. According to an authoritative text on reproductive health, “Using two methods at once dramatically lowers the risk of unintended pregnancy, provided they are used consistently.” Hatcher RA, Trussell J, Nelson A, et al.Contraceptive Technology. 19th rev. ed. New York: Ardent Media, 2007, p. 28. The evidence for this position appears to derive from probabilistic reasoning, rather than from clinical trials comparing the efficacy of two methods of contraception to one method. For example, oral contraceptives have an 8% rate of unintended pregnancy with typical use, and male condoms have a 15% rate of unintended pregnancy with typical use. Hatcher et al. 2007, p. 24. According to the principles of probability, when these methods are combined, there is only a 1.2% chance of unintended pregnancy with typical use. For an example of this kind of reasoning, see Kestelman P and Trussell J. Efficacy of the simultaneous use of condoms and spermicides.Family Planning Perspectives1991;26:226-227, 232.
13. See ref. 11, Hoffman-LaRoche Limited 2006, p. 5.
14. Anderson JR, Schonfeld TL, Kelso TK, Prentice ED. Women in early phase trials: An IRB’s deliberations.IRB: Ethics & Human Research2003;25(4):7-11, p. 8.
15. See ref. 3, U.S. Food and Drug Administration 2008.
16. Lyerly AD, Little MO, Faden R. The second wave. Toward responsible inclusion of pregnant women in research.The International Journal of Feminist Approaches to Bioethics2008;1(2):5-22.
Chris Kaposy and Françoise Baylis, “Ethical, Evidence-Based Guidelines for Contraceptive Use in Research,”IRB: Ethics & Human Research, vol. 32, no. 5: 1-9.