Impact of Current Federal Regulations on the Inclusion of Female Subjects in Clinical Studies
Vanessa Merton
THE IMPACT OF PRESENT LAW
Although clearly regulations that were adopted only 20 or so years ago cannot account for the longer history of exclusion of women from research, some researchers now say that they would be willing to include women in their protocols, were it not for the federal regulations that seem to them (or to their institutional review boards, or their institutions' lawyers) to prohibit or limit women's participation.1 These regulations do indeed pose some obstacles and should be amended, but careful examination suggests that even in their present form, the federal regulations are by no means an insurmountable hurdle to the inclusion of women, including pregnant women, in biomedical protocols. Indeed, some federal regulations support and facilitate the participation of women in clinical research.
Regulations that Appear to Limit Women's Participation in Clinical Research
FDA Definition of "Childbearing Potential" and Required Prior Reproductive Studies
Research protocols commonly contain the exclusionary criterion "pregnant and lactating women and women of childbearing potential," language apparently derived from the Food and Drug Administration's Guidelines for researchers.2 While the Guidelines are not legally binding, research conducted in accordance with the Guidelines qualifies for FDA consideration in a New Drug Application,3 and most investigators take the Guidelines seriously. Certainly no prudent
attorney would recommend their cavalier disregard, since they might well be deemed a "standard of care" for the research community.4
The Guidelines state that women of childbearing potential should be barred from large-scale (Phase III) clinical trials until all three segments of the FDA animal reproduction studies have been completed, and that women may be included in Phase II (controlled trials in several hundred subjects) studies only if "segment II and the female part of segment I of the FDA Animal Reproduction Guidelines have been completed."5 Segment I covers gonadal function, effects of estrous cycles/mating behavior, and early gestation; segment II, teratogenesis; and segment III, the drug's effect on late fetal development, labor and delivery, lactation, and newborn health. (Remarkably, the Guidelines are silent on the question of how the results of the animal reproduction studies ought to affect inclusion or exclusion of women.)
There is no mandate to perform any of the animal reproduction studies ever, and certainly not prior to the conduct of Phase II or III trials.6 The regulation that describes what applications for new drug approval must contain says only that the application should include nonclinical "studies, as appropriate, of the effects of the drug on reproduction and on the developing fetus."7 Whether under this standard the FDA could approve a New Drug Application without animal testing for reproductive effect has never been determined by a court, so far as I can tell (probably because it would be extremely imprudent for a drug manufacturer not to do some such studies at some point prior to marketing),8 but the FDA does not appear ever to have required such studies to precede Phase III trials.9 There is no practical way for me to research this, but I would bet that "as appropriate" has never been interpreted to require animal studies that would elicit adverse reproductive effects mediated through the male animal.10
What this boils down to is that pharmaceutical companies can choose to market drugs with no information about their reproductive impact, so long as the label makes this clear,11 and that the animal studies which the FDA defines as a necessary precursor to large-scale clinical trials with female subjects may never be conducted at all or may be done only in parallel with, not in advance of, clinical testing.
To compound this, the FDA, and thus clinical investigators, define the Guidelines' key phrase "of childbearing potential" in a way that, as I have written in another context, envisions all women as "constantly poised for reproductive activity":12
A woman of childbearing potential is defined as a premenopausal female capable of becoming pregnant. This includes women on oral, injectable, or mechanical contraception; women who are single; women whose husbands have been vasectomized or whose husbands have received or are utilizing mechanical contraceptive devices.13
The breadth of this definition of childbearing potential makes it tantamount to "all fertile women." The FDA limits the universe of women subjects right up through Phase III trials to "women who have been surgically sterilized, women who are postmenopausal, and women who are infertile . . . provided they [infertile women] are willing to use an effective form of contraception during the study, or have been evaluated by a fertility expert and have been found to be infertile, and have been so for greater than five years," to quote one research manual's interpretation.14
The FDA Guidelines do, however, expressly recognize an exception to the animal reproductive study requirement that gives researchers substantial latitude: testing of a drug anticipated to be a lifesaving or life-prolonging measure.15 "Life-prolonging" seems a quite elastic phrase that could cover many situations of clinical research. Under the Guidelines, so long as the lack of reproduction studies is pointed out during the informed consent process, and she is tested for pregnancy and advised of contraceptive measures, the woman of childbearing potential may participate in clinical trials of a potentially life-prolonging intervention, even during Phase I. The nursing mother is specifically mentioned as a potential subject, with analysis of the excretion of the drug or its metabolites in the milk to be determined "when feasible."16 And the one reference to pregnant women in the Guidelines merely notes that fetal follow-up should be carried out if a subject becomes pregnant while on the protocol;17 it does not say anything about immediately terminating the woman from the protocol, a provision frequently found in research design.
The head of the FDA's Office on Drug Evaluation, Dr. Robert Temple, maintains that the Guidelines should not be interpreted by researchers to require the exclusion of women from protocols.18 This observation is a welcome one, but apparently it has not been widely disseminated and is not widely shared by researchers.19 The problem here may be largely one of misperception, which a vigorous effort by the agency could remedy.
DHHS Limitations on Research with Pregnant Women
The FDA Guidelines, then, can be parsed to have relatively little impact on women's participation in research, unless the researcher wants them to. The relevant Department of Health and Human Services (DHHS) regulations, on the other hand, appear to be a far greater constraint. However, they deal exclusively with the pregnant, rather than merely pregnable, woman.
Subpart B of Part 46 of Title 45 of the Code of Federal Regulations is entitled "Additional Protections Pertaining to Research, Development, and Related Activities Involving Fetuses, Pregnant Women, and Human In Vitro Fertilization." Promulgated in the mid-1970s, Subpart B prohibits research "involving" pregnant women unless (1) "appropriate'' studies on animals and
nonpregnant individuals have been completed;20 (2) the purpose of the research is "to meet the health needs" of the woman; and (3) the fetus will be placed at minimal risk or at risk to the minimum extent necessary to meet the woman's health needs.21
So how significant a barrier for the pregnant woman are these regulations? To begin, let us consider the first requirement: completion of "appropriate" studies in animals and nonpregnant people (but presumably, under the Guidelines, not studies in women of childbearing potential). "Appropriate" is what lawyers call a weasel word, a word that gives the decision maker in a situation great discretion. Could this be a reference to teratogenicity studies? General clinical pharmacology studies for toxicity? Or perhaps the sort of in vitro work or uncontrolled case reports that one would ordinarily expect to foreshadow clinical research? It is hard to imagine the government's succeeding in imposing sanctions on a researcher for violating this amorphous provision unless absolutely no work had been done on the intervention on trial before the pregnant woman was permitted to participate.
More troubling is the requirement that the research be intended to "meet the health needs of the mother (sic)."22 I will not try to improve on Robert Levine's trenchant exposition of the conceptual murkiness of the terms "therapeutic research" and "nontherapeutic research" and their especially problematic usage in this context.23 Suffice it to say, this is another phrase that gives the researcher considerable scope. In a longer paper, I have identified the many ways in which participation in a research protocol in and of itself, regardless of the efficacy of the intervention on trial, may serve a subject's health needs.24 I will add only that from the standpoint of psychic health, it takes a severe toll for a pregnant woman (or for any woman) to be told that after having been fully informed of the risks to her potential offspring, she cannot be trusted to decide whether to participate in a protocol, while no restrictions of any kind are placed on the decisions of her male counterpart, whose offspring may be at equal risk.
The final requirement is that the risk to the fetus must be minimal: again, an ill-defined and comparative term.25 "Minimal risk" is defined elsewhere in the regulations to mean that the danger anticipated from the research is not greater, considering both probability and magnitude, than the danger of ordinary daily life or routine physical or psychological tests.26 For a fetus, isn't an amniocentesis, with its half-percent chance of miscarriage, now a routine test? What about chorionic villi sampling? Ultrasound? Would the risk of ordinary daily life include the risk of a parent who smokes or who works where smoking is permitted? Drinks socially? Eats food containing additives? Disregards the prenatal care provider's advice?
But perhaps the more important question is, who should make this assessment? Is there reason to believe that the researcher, or perhaps the IRB, or perhaps a federal bureaucrat, is the best choice to judge the net of harm and
benefit, risk and advantage, that would result from a pregnant woman's participation in a protocol?27 Such a conclusion seems to presume maternal—fetal conflict, to ignore the inextricable link between the pregnant woman and the conceptus at any stage of its development, and to deny the woman's inherent responsibility for that part of her body which may be born a child.28 As the DHHS regulations require, the woman must and should be fully informed about the risks, known and unknown, to her fetus. But nowhere do the regulations say, and no fair and rational reading can impute, that the judgment as to the requisite level of risk, and its proper weight in light of the woman's health needs, should be made by anyone other than her.29 The absence of either administrative or judicial review of this question, as well as of the other questions raised by the DHHS regulations, underscores my conclusion that federal regulations are not trammeling researchers who in their eagerness to recruit and accept women subjects have been testing the boundaries of the law. The participation of women, and indeed of pregnant women, in clinical research is permitted, if complicated, by the federal regulations; the most problematic provisions probably would not survive scrutiny if challenged. But there have been no challenges, which seems to suggest that they are the product, not the cause, of the research tradition that excludes women for quite other reasons.
Federal Requirements that Support the Inclusion of Women in Clinical Research
FDA Premarketing Testing
Two different sources of law require adequate testing of pharmaceutical products. First, federal statutes and FDA regulations require that a New Drug Application demonstrate "adequate, and well-controlled investigations, including clinical investigations . . ."30 and "data demonstrating substantial evidence of effectiveness for the claimed indications."31 Second, it has long been recognized that the obligation to adequately test drugs before beginning to profit from their marketing is grounded in basic tort law principles. Beginning with the debacle of MER/29, the Richardson-Merrell anticholesterol product that blinded many people because the company failed to pursue ocular abnormalities in test animals,32 courts have penalized companies that do not conduct reasonable testing to determine the potential adverse reactions of their products, even if the testing involved was not required by the FDA.33
In terms of the regulations, however, the question is, what do the phrases "adequate and well-controlled investigations" and "data demonstrating substantial effectiveness" mean? For the purposes of this discussion, one significant clue is the FDA's regulation on the labeling of drugs: "Evidence is . . . required to support the
dosage and administration section of the labeling, and modifications for specific subgroups (for example, pediatrics, geriatrics, patients with renal failure)."34 "If evidence is available to support the safety and effectiveness of the drug only in selected subgroups of the larger population . . . , the labeling shall describe the available evidence and state the limitations of the usefulness of the drug."35
Certainly it is reasonable to assert that with respect to the many interventions that have been tested only in male subjects, women are a "subgroup" for whom the data available is of limited utility. It is now uncontroversial that drug metabolism, dose-response reaction, and many other significant values, signs, and markers of clinical effect are different in women than in men.36 A recent report of the Council on Ethical and Judicial Affairs of the American Medical Association documents numerous examples of disparities in providing women major diagnostic and therapeutic interventions, ranging from kidney dialysis and transplantation to diagnosis of lung cancer.37 The Council concluded that these disparities could not be accounted for by biological differences or other benign or neutral variables. It found that "medical treatments for women are based on a male model, regardless of the fact that women may react differently to treatments than men or that some diseases manifest themselves differently in women than in men. The results of medical research on men are generalized to women without sufficient evidence of applicability to women."38 The Council went on to recommend: "Research on health problems that affect both sexes should include male and female subjects. Sound medical and scientific reasons should be required for excluding women from medical tests and studies, such as that the proposed research does not or would not affect the health of women."39
Arguably, then, the FDA regulations should be read to require adequate testing, meaning sample sizes and/or stratified analyses sufficient to detect high-background adverse reactions, in a female population, unless the drug label reveals the paucity of data in women and cautions against undue extrapolation from testing in men.40
DHHS Regulations on Research with Children
One of the issues frequently raised in the course of discussing the enrollment of female subjects in clinical trials is the research sponsor's concern about liability to their offspring.41 In that context, it is sometimes said that this concern cannot be resolved by any sort of "waiver" or "assumption of risk" by parent-subjects on behalf of their offspring, because while parents can consent to their children's participation in all manner of fairly dangerous and nonbeneficial activities outside the medical arena, generally they can consent to medical intervention for their children only if the intervention is intended and expected to be beneficial to the individual child.42 The traditional view,
therefore, has been that parents lack capacity to consent to their child's participation in nontherapeutic research, and therefore, ipso facto, to assumption of the risks of the child's participation in such research, although the one attempt to obtain a judicial declaration that parents have no right either to permit or to compel their children's participation in nontherapeutic research did not succeed.43
The analogy between parental assumption of risk for born, living children who are themselves the subjects of research, and release of a sponsor from responsibility for harm to a subject's unborn or unconceived children, is tenuous at best, but federal regulations become relevant here because of their explicit recognition that parents can consent to nontherapeutic research (and therefore, logically, to assumption of the risks of that research). In 1983 the DHHS promulgated regulations that clearly contemplate and authorize parental consent to a broad range of nontherapeutic research with child subjects.44 It is hard to imagine that the lengthy provisions of these regulations which describe in detail the quality of consent necessary for such research45 will be treated as nugatory in any future liability litigation, rather than as the source of public policy they obviously are.
RECOMMENDED CHANGES IN THE FEDERAL LAW
Full inclusion of women in biomedical research would be somewhat easier to implement if there were some revision of current federal regulations. Because these limitations are not statutory in origin, however, such revision would not require congressional action, but could be undertaken by the relevant administrative agencies at any time.46
Amendment of FDA Regulations
First, the FDA Guidelines should be redrafted to raise identical concerns about the participation of both male and female subjects of reproductive potential, and to allow women to decide for themselves, as men do, about the relative risks and benefits—for them—of participation in Phase I or early Phase II trials.47 FDA regulations require researchers to inform subjects "when appropriate" that the research may involve "risks to the subject (or to the embryo or fetus, if the subject is or may become pregnant) which are currently unforeseeable."48 This section should be amended to require researchers also to inform male subjects who are or may become involved in reproductive activity of the state of knowledge about male-transmitted birth defects and/or effects on male germ cells.49 In many cases the current knowledge will be nil; nothing will be known because the intervention has not been tested on this parameter.
Subjects should be told this, and told also that while instances of adverse reproductive effect for or through the male parent have occurred, too little is yet known to permit quantification of the risk. If women are required to use contraception, then so should men.50
The most efficient method of changing researcher behavior would be to amend the substantive provisions of FDA regulations to require complete testing of new drugs in relevant populations, specifically women of childbearing age, pregnant women, and nursing women.51 At a minimum, the FDA ought to ensure that drug labels state that evidence of both safety and efficacy is lacking for these populations whenever that is the case, and that nothing is known about reproductive hazards for men, which will almost invariably be the case.
Amendment of DHHS Regulations
The only major change needed in the Federal Policy for the Protection of Human Subjects is in the section on equitable selection of subjects, Section III. "Pregnant women" should either be removed from the category of "vulnerable populations" or replaced by "men and women actively engaged in reproduction." To amplify the definition of "equitable," helpful language may be borrowed from a fine consensus document by a working group on principles and policies for clinical research on HIV infection, which concluded that:
No group should be categorically excluded, on the basis of age, gender, mental status, place of residence or incarceration, or other social or economic characteristic from access to clinical trials or other mechanisms of access to experimental therapies. Special efforts should be made to reach out to previously excluded populations. . . . It is inequitable and discriminatory to exclude women, including women of reproductive age, from clinical trials.52
Another component of DHHS regulations that needs revision is Subpart B. Here, the best option would be to delete Section 46.207, the provision dealing with "Activities directed toward pregnant women as subjects." Its language is ambiguous and confusing; the subsection requiring paternal consent is surely unconstitutional even under present standards; and given that about all it permits is activity intended to "meet the health needs of the mother," depending on one's view of the purposes of clinical research, either it is tautological or it describes a null set. The bulk of the other provisions of Subpart B, which govern fetal research, suffer from various infirmities and illogicalities that ought to be corrected, but do not in themselves pose any particular barrier to women's participation in protocols.
National Institutes of Health Policy
The National Institutes of Health can itself, through implementation of its own policy,53 dramatically influence the pattern of excluding women from biomedical research. But to do so, more specific standards than are contained in the current policy are necessary. The NIH/ADAMHA Policy Concerning Inclusion of Women in Study Populations (''NIH Policy") appears to establish a firm presumption that women shall be included in protocols. However, that presumption may be rebutted by showing either (1) that it would be "inappropriate" to include women in a given study; or (2) a "compelling justification" for the exclusion of women. In conjunction, these criteria for exemption from the policy are quite capable of swallowing it whole—the proverbial Mack truck would have no trouble navigating these holes.
For example, the NIH Policy states that "appropriateness" of inclusion of women is in part a function of the "known incidence/prevalence" of a condition among women. Yet one of the major consequences of the exclusion of women from research has been enormous gaps in existing knowledge about the epidemiology of many conditions in women. As the members of the Women in Research Task Force, a bioethics group in which I participated, wrote in a letter to Dr. Kirschstein, then Acting Associate Director of the new NIH Office of Research on Women's Health, "Thus, women could end up being excluded from a study based on data from studies which excluded women in the first place."54 Moreover, women must be leery of terms like "appropriate" when those who will be deciding what they mean remain overwhelmingly products of the mind set and world view that has so often found subordination, denigration, and paternalistic protection of women ''appropriate."
Likewise, with respect to the "compelling justification" for exclusion of women, to exempt a protocol from the NIH Policy, researchers may need only recite the usual claims about "dirty data" or fear of liability. In its Memorandum OER 90-5, NIH refers to two situations that qualify as "compelling": (1) the condition to be studied occurs only in men; or (2) inclusion of women would "jeopardize the health and safety" of a class of subjects. Without clarification, it is not hard to imagine that fetal protection may be the hidden meaning of the latter. I find disquieting, rather than comforting, Dr. Kirschstein's response to the Women in Research Task Force letter, in which she states:
A list of situations that comprehensively accounts for all such justifications for "compelling" exclusion is very difficult to create. . . . One potential basis . . . is the case in which the financial and human costs of conducting research trials are significantly increased or unduly burdensome in comparison to the benefit gained by including a representative number of both genders in the study population for a disease or condition in which the incidence is lower in one gender than in the other. . . . Another potential basis for rebuttal is the situation in which violation
of the established legal rights of a child or potential rights of a fetus are a foreseeable possibility as a result of the mother's participation in a clinical trial. . . . [C]ases in which studies on animals measuring teratogenic or other significant adverse effects are incomplete or inconclusive warrant exclusion justification review and may serve as sufficient grounds for rebutting the inclusion presumption.55
Several questions are raised by this response. First, when Dr. Kirschstein speaks of "financial and human costs of conducting research trials," I wonder whose costs, or costs to whom? For women as a class, the cost of their involuntary nonparticipation in research has long outweighed the benefit. Second, what "potential rights of a fetus" (or, for that matter, established rights of a child) are violated when a woman chooses to take an unapproved drug? Surely those same rights, whatever they may be, are violated when the same woman now takes an approved drug that has never been tested in women, or in pregnant women, but is available on the market, or when the woman fails to follow a doctor's orders, or to exercise regularly, or to do a thousand other things that are ''good" for her—as, once upon a time, thalidomide, diethylstilbestrol, diuretics, X-rays, routine cesarean section, and minimal weight gain were deemed "good" for pregnant women by their physicians. With this language, I am afraid Dr. Kirschstein and the Office of Research on Women's Health, the putative bastion of women's rights and liberties in this process, inadvertently reinforce the coercive, intrusive model of ''maternal-fetal (or child) conflict."56
As for "incomplete or inconclusive" animal studies of teratogenicity or "other significant adverse effects [presumably, upon reproductive outcomes]," so far as I am aware, there is nothing but incomplete, inconclusive data in this area; for a start, the absence of such studies with respect to male-mediated effects on offspring render them all, by definition, partial and inconclusive. If Dr. Kirschstein's statement is taken literally, hardly any research should be permitted.
There is an alternative approach: requiring adequate animal studies of adverse reproductive effects in both male and female animals. Such studies would have to test for all potential male routes of prenatal and preconceptual impact. If NIH is serious about avoiding risk to the offspring of research subjects, the only effective method is either to permit participation only of nonfertile men and women—difficult for large-scale studies, and unlikely to be clinically representative—or to impose much more rigorous controls on reproductive behavior by human study participants, both male and female.57
CONCLUSION
There is much more, beyond fixing specific regulations, that all the federal agencies—NIH, DHHS, and FDA—could do to redress the current situation. Gender-specific and gender-comparative research that will provide better understanding of the etiology and risk factors for disease in women ought to receive priority in funding and other resources. Federal agencies should require that all research be evaluated to determine whether its design will elucidate gender-specific differences and permit systematic analyses of gender-specific variables.58 This requirement cannot be confined to areas of acknowledged gender difference, for example, drugs metabolized through pathways influenced by sex steroid hormones, because the failure to recognize differences may reflect a lack of prior research rather than the right questions having been asked and answered in the negative.59 At a minimum, pharmacokinetic screens of all new drugs should be conducted in both women and men, and animal studies should include female as well as male animals. Whenever animal reproductive studies are conducted, they should seek to determine the incidence of adverse reproductive outcomes through the male parent.
In the eloquent words of the executive director of the Center for Women Policy Studies, "research must study women in their own right, and on their own biological terms, not as if they were defective men."60 Especially when research is conducted with government funds, subject to government control, in government facilities by government-supported or government-trained researchers, there can be no justification for validating the safety and efficacy of an intervention only in men. Nothing less than the health of our daughters is at stake.