The treatment of medical conditions complicating pregnancy is challenged by a lack of information about the safety and efficacy of medications used, or potentially used, by pregnant women. Generally, what constitutes the most effective therapy is identified through systematic research; however, pregnant women are commonly considered ineligible to participate. An unintended consequence is that women and health care providers are often faced with a lack of information to inform treatment decisions for obstetric and nonobstetric complications of pregnancy.
The Institute of Medicine,1 the World Health Organization,2 the American College of Obstetricians and Gynecologists,3 and other leaders agree that pregnant women should be considered eligible for participation in clinical trials. Although there are many situations in which excluding pregnant from research is justified, they are often reflexively excluded even when their inclusion can be ethically and scientifically justified.4 To address this, the U.S. Food and Drug Association (FDA) is preparing a draft Guidance for Industry entitled “Pregnant Women in Clinical Trials: Scientific and Ethical Considerations.”5
Although the consequences of broad-based exclusion of pregnant women have been noted, the extent of the underrepresentation of women in research trials is poorly characterized. The intent of this study was to measure the current state of inclusion and exclusion in industry-sponsored clinical trials and to establish a baseline against which future measurements of the representation of pregnant women in research can be compared. Understanding the patterns and extent of their exclusion will be critical to setting priorities, anticipating barriers, developing policies, and responding appropriately to FDA draft guidance and other policy interventions.
MATERIALS AND METHODS
Our data were compiled from the list of all studies enrolling women of childbearing potential posted on www.ClinicalTrials.gov between 1 October 2011 and 31 January 2012. Because the data on ClinicalTrials.gov are in the public domain, the Institutional Review Board of the University of North Carolina at Chapel Hill classified this study as exempt from approval.
We limited the review to open United States–based phase IV interventional studies that included adult female participants. Because the anticipated FDA guidance is intended for industry, we limited our review to studies that were sponsored by pharmaceutical, biotech, or medical device companies, estimated in one analysis to be 63% of all studies.6 To be included, the study must have been evaluating treatment of conditions that may be experienced by but are not limited to pregnant women and must not have involved the use of a medication that is in the FDA pregnancy categories D or X, for which there is positive evidence of human fetal risk or teratogenicity.
We chose to investigate enrollment in phase IV studies because studies of pregnant women would best be conducted after initial testing had established low risk of toxicity and teratogenicity in animal tests and efficacy in the general population. Having established these data points, phase IV trials are potentially the most appropriate and least controversial for inclusion of pregnant women. Phase IV trials are conducted using drugs that have already received regulatory approval and are currently on the market. They evaluate drug interactions, long-term effectiveness, use in populations other than those for which the drug was initially approved, and other characteristics.7 In addition, FDA pregnancy risk categories for the study drugs in phase IV trials are available to help assess the appropriateness of exclusion of pregnant women. Given these characteristics, phase IV studies represent those most likely to enroll pregnant women and thus to establish the most permissive baseline from which to characterize inclusion.
The ClinicalTrials.gov database was queried using the search function of the web site for the terms: “open studies,” “interventional studies,” “location in the United States,” “studies with female participants,” “adult age group (18–65),” “phase 4,” and “funded by industry.” Variables such as sponsor, drug or device or procedure, condition undergoing study, and exclusion criteria from each study retrieved by the query were manually transposed to an Excel worksheet. The FDA pregnancy categories for all drug products were ascertained and transcribed from www.PDR.net, the electronic version of the Physicians' Desk Reference from PDR Network, a leading publisher of drug labeling information.8
Those studies evaluating the treatment of medical conditions that could not be experienced by pregnant women were excluded, as were studies found to be using medications in FDA pregnancy category D (positive evidence of human fetal risk but benefits of use may outweigh the risk) or category X (positive evidence of animal or human fetal risk and the risk of use clearly outweighs the potential benefit). If there was no mention of pregnancy in the inclusion or exclusion criteria, we contacted a study representative by e-mail or telephone to discern whether pregnant women could be enrolled and, if not, the reason for their exclusion.
Application of the search criteria retrieved a total of 559 industry-sponsored studies from the ClinicalTrials.gov web site over the course of this 4-month period. Of these, one trial (a trial for hemophilia limited to males) was excluded, resulting in a total of 558 clinical trials available for study.
Of these 558, five studies limited enrollment to pregnant women and 98 were found to appropriately exclude pregnant women based on the following criteria: 74 trials utilized at least one drug in FDA pregnancy category D or X; the age range criteria in 17 studies precluded childbearing potential; and the topic of seven studies precluded pregnancy at enrollment, including four on menopause, two on contraception, and one on lactation. Of the remaining 455 clinical trials that could potentially enroll women of childbearing potential, 301 specifically excluded pregnant women, one specifically did not exclude pregnant women, and 153 made no mention of pregnancy in the inclusion or exclusion criteria posted on ClinicalTrials.gov.
One might expect that, if pregnancy was not mentioned in the exclusion criteria, pregnant women could be included. To test that assumption, we attempted to contact study personnel from all 153 of these studies by telephone, by e-mail, or both. Eighty-eight did not respond to the request for information. Of the 65 studies for which clarification about whether pregnant women could be included was obtained, 47 (72%) excluded pregnant women and 18 (28%) did not exclude them from enrollment in the study. Justifications for the pregnancy exclusion elicited from study personnel included that the condition undergoing study had a low prevalence in pregnant women, that there was not yet enough safety data, that pregnant women are considered by FDA to be a “vulnerable population,” and that pregnancy was “normally always an exclusion criteria [sic].” One researcher commented, “no glaucoma treatments have been approved for use in pregnant women so [excluding them] was a no-brainer.”
Overall, of 558 qualifying industry-sponsored studies, only five were designed specifically for pregnant women (a rate of approximately 1%). Of the 367 phase IV clinical trials for which we had specific inclusion and exclusion criteria about pregnancy, 348 (95%) excluded pregnant women and 19 (5%) did not exclude pregnant women from enrollment (5% is a lower limit based on the 88 studies with unknown exclusion criteria) (Fig. 1).
A range of medical conditions that could occur in pregnant women was noted in the phase IV studies evaluated. These included, for example, epilepsy, depression, arthritis, peripheral artery disease, human immunodeficiency virus, and schizophrenia. All of the conditions being evaluated by the clinical trials in our study could occur in pregnant women, although treatment of some, like knee replacement, might be safely postponed until the completion of the pregnancy; however, treatment of others would be important to the ongoing health of a pregnant woman and her fetus.
The topics of the five studies that were limited to women who were pregnant or intending to become pregnant were all pregnancy-related conditions. We found no studies designed specifically to evaluate the treatment of nonobstetric illness during pregnancy.
Acknowledging that the small number of studies that allowed enrollment of pregnant women (n=19) precludes robust statistical analysis, we compared the characteristics of studies that excluded pregnant women with those that permitted their enrollment (Table 1). In approximately 80% of the studies in both groups, the purpose of the trial was to evaluate treatment as opposed to preventive, diagnostic, or supportive care. Regardless of whether they included or excluded pregnant women, the proportion of “open-label” studies was the same as the proportion of blinded studies. Almost half of the studies that permitted the enrollment of pregnant women were nonrandomized (47%) compared with 34% in the trials that excluded pregnant women.
The types of interventions being studied were also compared. There were five major categories: drugs; devices; procedures; biological products; and diet. Excluding diet because of the small number (n=5), the percentage of studies in each category that excluded pregnant women was highest (98%) for drug studies and lowest (81%) for biologics (Table 2). Regarding this point, we constructed a 2×2 table of drugs (case group) compared with biologics (control group) and pregnancy excluded (not exposed) compared with pregnancy included (exposed). In the excluded row, there were 13 biologic studies over a total of 251 pregnancy-excluded studies compared with, in the included row, three biologic studies over a total of nine studies. Using the two-sided Fisher exact test for 13 of 251 (5.2%) compared with three of nine (33.3%), the difference in the proportions was significant (P=.01).
Although it is widely recognized that pregnant women and their interests are underrepresented in health research, the extent and patterns of their exclusion have not been characterized. In this study, we found that only 1% of industry-sponsored studies were designed specifically for pregnant women and 95% of studies of conditions that can affect pregnant women excluded pregnant women from participation. De facto exclusion (pregnant women may not be recruited or enrolled even if their participation is permissible) and an expected higher rate of exclusion in earlier phase studies would predict even greater prevalence of exclusion than our data suggest.
A common explanation for the exclusion of pregnant women from research is the desire to “do no harm,” yet clinical care during pregnancy often requires the use of medications untested in pregnancy. Participation in an ethically designed phase IV clinical study would be very much like treatment in clinical practice with the additional potential benefits of expanded informed consent, enhanced monitoring, and the patient's knowledge that she has contributed to the evidence base and has benefitted other pregnant women.
Another explanation for broad-based exclusion of pregnant women from research may be the overinterpretation of federal guidelines9 that preclude participation unless several requirements are met. Some researchers may presume that pregnant women would not be willing to participate in research even if they were eligible. A few small studies on this subject suggest that pregnant women, like men and nonpregnant women, volunteer for enrollment for many reasons, including to gain treatment for or protection against disease, easier access to the therapeutic intervention, enhanced monitoring of their condition, and altruism.10 One study found that 74% of pregnant women interviewed said that they would participate “if there is a chance that participation in a clinical trial would help their pregnancy and improve their baby's health.”11 Including pregnant patients in clinical trials may not lead to robust evidence about the safety or effectiveness of the intervention in pregnancy because the subgroup of pregnant patients may be too small to provide statistical significance. However, adding these limited data to knowledge acquired from inadvertent pregnancy exposures during earlier phase trials, postmarketing pregnancy registries, and clinical experience will potentially lead us to evidence-based treatment recommendations sooner than current practice. Further, adequate inclusion to garner robust evidence for pregnancy should be a serious consideration in future trial design.
How to safely include pregnant women in clinical research is beyond the scope of our analysis. The draft FDA guidance document is expected to address this and the data analysis challenges wrought by the physiologic and pharmacokinetic changes that occur during pregnancy. Our finding that 53% of the studies that included pregnant women were randomized suggests that researchers conducting both randomized and nonrandomized phase IV trials may be open to the inclusion of pregnant women as study participants. Our finding that a significantly higher proportion of studies of biological agents than of pharmaceuticals included pregnant women (19% compared with 2%) suggests that recent public health messaging about maternal vulnerability to bioterrorism and flu may have increased society's understanding of the need for studies including pregnant women.
This study has several limitations. The clinical trials we evaluated were posted on ClinicalTrials.gov during a specific period in time that may not reflect the same prevalence of exclusion of pregnant women from studies conducted during other time periods. Our study size precluded robust statistical comparison; however, the descriptive data suggest striking patterns of underrepresentation. There may have been scientifically justified reasons for excluding pregnant women from particular studies that this study did not identify. We did not address the exclusion of pregnant women from clinical trials not funded by industry, a considerable proportion (37%) according to one analysis.6 This question is worthy of further research including the reasons for any differences in proportions, if identified.
The rationale for the exclusion of pregnant women was rarely addressed. The one study that explicitly included pregnant women stated that “[drug] carries a category B pregnancy risk factor. Since this is a minimal pregnancy risk category, no special precautions will be taken to determine that the patient is not pregnant.” Because the Institute of Medicine, the American College of Obstetricians and Gynecologists, and other experts assert that pregnant women should be presumed eligible for enrollment unless there is a scientific rationale for their exclusion, researchers should be explicit about justifications for nonenrollment.
In the United States, more than 400,000 women experience serious complications of pregnancy each year,12 and two women die every day from pregnancy-related causes.13 More than 50% of pregnant women use at least one prescription medication,14–17 but the safety of those drugs is largely unknown because of a lack of human data.17,18 We conservatively estimated that approximately 1% of industry-sponsored research is targeted to pregnancy-related conditions, approximately 95% of industry-sponsored trials preclude pregnant women from enrolling, and studies designed to assess the treatment of nonobstetric illness during pregnancy are rare or nonexistent.
National and international experts, including the FDA, agree that there should be greater inclusion of pregnant women in clinical studies to obtain evidence about treatment options during pregnancy. The release of the draft FDA guidance will generate dialogue about appropriate rationales for and barriers to their expanded inclusion. With this effort, we can begin to implement rational and ethically appropriate inclusionary practices that will provide the evidence base pregnant women deserve. Knowing the current extent of their exclusion in industry-sponsored trials will provide a baseline against which to measure future progress.
1. Mastroianni AC, Faden R, Federman D, editors. Women and health research: Ethical and legal issues of including women in clinical studies. Institute of Medicine. Washington (DC): National Academy Press; 1994.
2. Council for International Organizations of Medical Sciences. International ethical guidelines for biomedical research involving human subjects. Available at: http://www.cioms.ch/images/stories/CIOMS/guidelines/guidelines_nov_2002_blurb.htm
. Retrieved April 17, 2011.
3. Research involving women. ACOG Committee Opinion No. 377. American College of Obstetricians and Gynecologists. Obstet Gynecol 2007;110:731–6.
4. Lagakos SW, Gable A, editors. Design considerations: Methodological challenges in biomedical HIV Prevention Trials. Institute of Medicine. Washington, DC: National Academies Press; 2008:104–8.
5. FDA guidance document to-do list. “The pink sheet”. 2011. Available at: http://www.elsevierbi.com/publications/the-pink-sheet?issue=Jan-03-2011
. Retrieved March 22, 2011.
6. Bourgeois FT, Murthy S, Mandl KD. Outcome reporting among drug trials registered in ClinicalTrials.gov. Ann Intern Med 2010;153:158–66.
8. Physicians' Desk Reference online version, PDR.net. Available at: www.PDR.net
. Retrieved October 2011 through January 2012.
9. U.S. Code of Federal Regulations, title 45, part 46, subpart B 2009. U.S. Government Printing Office via GPO access [Cite: 45CFR46]. Available at: http://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.html#subpartb
. Retrieved May 2, 2011.
10. Lyerly AD, Namey EE, Gray B, Swamy G, Faden RR. Women's views about participating in research while pregnant. IRB 2012;34:1–8.
11. Rodger MA, Makropoulos D, Walker M, Keely E, Karovitch A, Wells PS. Participation of pregnant women in clinical trials: will they participate and why? Am J Perinatol 2003;20:69–6.
12. Lyerly AD, Little M, Faden RR. Pregnancy and clinical research. Hastings Cent Rep 2008;38:53.
13. Heron M, Hoyert DL, Murphy SL, Xu J, Kochanek KD, Tejada-Vera B. Deaths: final data for 2006. National Vital Statistics Report 2009. Available at: http://www.cdc.gov/nchs/data/nvsr/nvsr57/nvsr57_14.pdf
. Retrieved April 22, 2012.
14. Andrade SE, Gurwitz JH, Davis RL. Prescription drug use in pregnancy. Am J Obstet Gynecol 2004;191:398–7.
15. Mitchell AA, Gilboa SM, Werler MM, Kelley KE, Louik C, Hernandez-Diaz S. Medication use during pregnancy with particular focus on prescription drugs: 1976-2008. Am J Obstet Gynecol 2011;205:51.el–8.
16. Glover DD, Amonkar M, Rybeck BF, Tracy TS. Prescription, over-the-counter, and herbal medicine use in a rural obstetric population. Am J Obstet Gynecol 2003;188:1039–45.
17. Lo WY, Friedman JM. Evolving knowledge of the teratogenicity of medication in human pregnancy. Obstet Gynecol 2002;100:465–73.
18. Adam MP, Polifka JE, Friedman JM. “Teratogenicity of recently introduced medications in human pregnancy”. Am J Med Genet C Semin Med Genet 2011;176:175–82.