Relatively few drugs, especially those recently approved by the US Food and Drug Administration, have published human pregnancy experience. Although all drugs contain animal reproduction data, these are usually not predictive of human risk. Clinical trials in pregnant women are rarely conducted because of ethical and legal concerns, and it may be many years before sufficient observational data are collected to guide clinical treatment decisions. Because many of these drugs will be used in pregnancy, human data are needed shortly after the drugs come to the market. Lack of human data leads to uncertainty over whether a drug can be safely prescribed for a pregnant patient. Unless there are compelling scientific and ethical reasons to exclude them, pregnant women should be included in phase IV clinical trials (postmarketing studies to obtain additional information, including the risks, benefits, and optimal use of a drug). This paper examines how physicians currently counsel pregnant women when there are no human data and proposes an alternative method in which knowledge regarding risks associated with the use of drugs during pregnancy can be enhanced in a clinical trial setting.
The vast majority of newly marketed drugs have no reported human pregnancy experience. This is because pregnant women are routinely excluded from clinical trials because of the fear of potential fetal harm. In a 3 year period (2011–2013), the Food and Drug Administration (FDA) approved 95 new drugs. Only 3 drugs (clobazam, dimethyl fumarate, and factor XIII concentrate human) had human pregnancy experience at the time of the FDA approval, and that was from their use in other countries before US approval.
Some of these drugs appear to offer better treatment over older therapies, so they could very well have advantages for a pregnant woman. However, if the past is any indication, obtaining human pregnancy data for these drugs will be a daunting task. A study published in 2002 found that of the new drugs approved in the United States between 1980 and 2000, more than 90% had insufficient data to determine the human risk of developmental toxicity (abnormal weight, structural anomalies, functional and/or neurobehavioral deficits, or death).
In a follow-up study conducted 9 years later, 98% of 172 drugs approved in the United States between 2000 and 2010 lacked data or had insufficient data to determine the risk of developmental toxicity in humans. An even more striking finding from this study was that it can take an average of 27 years before there are sufficient data published on a drug to make even a crude estimate of its human teratogenic risk. A study published in 2014 also confirmed the lack of human data. Of the 213 drugs approved by the FDA between 2003 and 2012, 93% had only animal data, and just 5% had human pregnancy data. Moreover, despite FDA encouragement to establish pregnancy registries, 85% of the newly approved drugs did not have a pregnancy registry. Of those that had a registry, 69% were by therapeutic category, not agent specific.
Preclinical teratology studies in animals are required for regulatory approval and are important research tools for determining whether a drug might cause developmental toxicity. These studies often provide a means of identifying teratogenic potential before humans have been harmed. Unfortunately, animal studies do not always reliably predict developmental toxicity in humans. For example, the severe limb defects typically found in the newborns of women who had taken thalidomide during pregnancy could not be produced consistently in animal studies and not at all in rodent studies, even at extremely high doses.
This inability of animal studies to consistently predict the developmental toxicity of a drug in humans leads to uncertainty over whether a drug can safely be used in a pregnant patient. Because of this uncertainty, pregnant women may be treated with ineffective or subtherapeutic medications, or not treated at all, putting them and their fetus at increased risk. The risk to the fetus of undertreating or not treating the mother may be greater than the potential risk of the drug.
Instead of waiting more than 2 decades for human pregnancy data, many health care professionals and researchers are advocating for the conducting of clinical trials specifically designed to address health concerns relevant to pregnant women as an alternative way of gathering data. However, a recent review found that a mere 1% of industry-sponsored studies were designed specifically for pregnant women. When they reviewed studies of conditions that can affect pregnant women, they found that 95% excluded pregnant women from participation. Unless they intend to be used specifically to treat medical disorders unique to pregnancy, the FDA does not require manufacturers to prove safety and efficacy of their drugs in pregnant women. Consequently, when most medications are used in pregnancy, they are prescribed off label for an unapproved indication.
Because the FDA does not regulate the off-label prescribing of drugs, there is little incentive for manufacturers to seek FDA approval for the use of their product during pregnancy. It is easier for manufacturers to simply recommend that their drug not be taken during pregnancy unless clearly needed. Without data from clinical trials to guide them, physicians and their patients are burdened with the task of evaluating the risk (which is usually not known) and the benefit of using a drug during pregnancy.
Identifying potential human teratogens in the absence of human data
How does a clinician respond to a pregnant woman who has been exposed to a new drug when she asks whether taking the drug will harm her embryo or fetus? Without human pregnancy data to inform them, clinicians must rely on data from animal reproduction studies and human pharmacokinetic studies using nonpregnant individuals to estimate the developmental toxicity risk in their patients. These methods and their limitations for assessing risk in humans are discussed in the following text.
Several factors, such as type of drug (chemical and pharmacological properties), dose, route of administration, duration, and gestational time of exposure are important determinants of teratogenicity. Dose and time of exposure are the cardinal principles of teratology. Drugs that are applied topically and have little systemic absorption are associated with a low risk of developmental toxicity. The period of organogenesis is considered the period of greatest sensitivity to most teratogenic agents. Exposures later in pregnancy are more likely to cause growth restriction and functional disorders other than structural defects.
Regarding animal studies, with the exception of drugs inhibiting the renin-angiotensin system and some cardioselective β-blockers, all drugs known to cause human developmental toxicity also cause developmental toxicity in animals. However, the doses used in these animal studies are critical because all drugs will cause this toxicity if the dose is sufficiently high. So can animal data be used to estimate human risk? Well, the answer is partially.
Unfortunately, because of interspecies differences, the congenital anomalies found in preclinical studies can differ from those observed following human exposure. In addition, the doses used and the manner in which the animals were exposed may differ greatly from the typical human experience. The fact that no defects are produced in an experimental animal model does not mean that defects will not occur in humans. For these reasons, animal studies do not always reliably predict developmental toxicity in humans. Therefore, when using the results of animal studies to evaluate teratogenic risk in humans, it is important to consider whether the experimental conditions in which the adverse effects were observed and the types of defects produced are at all relevant to humans.
In a 2014 reference book, after excluding nondrugs and herbs, 1154 drugs were reviewed. Among these, there were 93 drugs known to cause human developmental toxicity (8%). In 75 of these cases (81%), the published animal data suggested potential risk in humans because developmental toxicity was observed at doses that were 10 times or less the human dose based on milligrams per square meter (body surface area [BSA]) or area under the concentration-time curve (AUC) without causing maternal toxicity. In the remaining 18 cases, the animal data suggested low risk (angiotensin-converting enzyme inhibitors and cardioselective β-blockers).
The dose limit of 10 times or less the human dose was agreed upon by a group of clinical teratologists convened at a 2003 workshop organized by the Teratology Society Public Affairs Committee and cosponsored by the Centers for Disease Control and Prevention, the FDA’s Office of Women’s Health, the March of Dimes, Merck Research Laboratories, the National Institute of Environmental Health Sciences, the Neurobehavioral Teratology Society, the Organization of Teratology Information Services, and Pfizer. The primary objective of this workshop was to develop recommendations on how to assess the teratogenic risk in humans using data from animals.
There were also 236 drugs that caused toxicity in animals at the 10 times or less threshold, but there was either no human pregnancy experience in 135 or the limited published experience in 101 had not shown harm in human embryos or fetuses. Thus, the animal data in 311 (236 plus 75) (27%) of the drugs raised the possibility of human developmental toxicity that was eventually confirmed in 75 of the cases (24%). This is consistent with the belief that preclinical developmental toxicity studies in animals have a limited ability to predict human risk.
The nature of maternal disease and the pharmacokinetic characteristics of drugs can be considered to estimate the teratogenic risk. For example, will not treating or undertreating a mother’s disease pose a greater risk to the developing fetus? For less severe maternal disorders such as mild hypertension or psoriasis, discontinuation of treatment during pregnancy has not been associated with adverse effects on the mother or fetus, whereas failure to treat severe or life-threatening conditions, such as diabetes or some mental illnesses, can have serious repercussions for both mother and fetus.
Usually if there are human data for other drugs with the same mechanism of action, this information can be used to estimate the embryo-fetal effects of a new drug that has no human data. For example, all inhibitors (angiotensin-converting enzyme inhibitors, angiotensin II receptor antagonists, and the renin inhibitor) of vasoconstrictor angiotensin II should be expected to have a similar risk of fetal toxicity because this action causes decreased fetal renal blood flow, resulting in impaired kidney function.
The same reasoning can be applied to the cardioselective β-blockers and, to a lesser extent, to the nonselective β-blockers. But there are exceptions; for example, thalidomide is known to cause severe birth defects, whereas glutethimide has not. Moreover, some drugs appear to be more teratogenic than others in the same class because they are more widely used and have been studied in greater depth.
Current system (postapproval clinical trials)
Women who are enrolled in a clinical trial are usually cautioned against pregnancy exposure, and if they do conceive, treatment will be discontinued and they will be dropped from the study. If the new drug is the best medication for her condition, then she will be either untreated or prescribed a second-line agent during her pregnancy. In the worst-case scenario, women who become pregnant during a clinical trial may worry that the fetus has been harmed by the exposure and consider termination of an otherwise wanted pregnancy.
Current system (postapproval clinical trials)
Women who are enrolled in a clinical trial are usually cautioned against pregnancy exposure, and if they do conceive, treatment will be discontinued and they will be dropped from the study. If the new drug is the best medication for her condition, then she will be either untreated or prescribed a second-line agent during her pregnancy. In the worst-case scenario, women who become pregnant during a clinical trial may worry that the fetus has been harmed by the exposure and consider termination of an otherwise wanted pregnancy.
Proposed system: phase IV clinical trial
Collection of data related to the use of drugs during pregnancy can be improved by including pregnant women in phase IV clinical trials. Is it ethical to include women of reproductive potential or who are pregnant in clinical trials? We believe so. In fact, an argument can be made that it is unethical to exclude pregnant women and their fetuses from clinical trials because it could put them at greater risk of being exposed to interventions that are not evidence based and might limit the number of therapeutic options that have been specifically tested for safety and efficacy in pregnant women.
Prescribing decisions for pregnant women should be based on evidence that has been rigorously evaluated in controlled research in the pregnant population, not on evidence that has been derived from males or nonpregnant females.
The current system in which a clinician prescribes a drug that has not been studied in pregnancy to a nonpregnant woman may also be unethical because she may or may not be adequately counseled regarding the potential risk to her embryo and/or fetus should she become pregnant. This opinion is based on our collective experience that pregnant women exposed to medications rarely receive in-depth counseling concerning the potential for embryo-fetal harm before they start taking the drug.
It can also be argued that it is unethical to encourage participation in a pregnancy registry, the goal of which is to provide health care providers with useful information for treating or counseling patients who are pregnant or anticipating pregnancy and then neglect to share this information through publication in a peer-reviewed journal.
The authors propose that interested health care professionals collaborate with professional organizations, such as the Society for Maternal Fetal Medicine, the American College of Obstetricians and Gynecologists, the Organization of Teratology Information Specialists, and the Teratology Society, to encourage Congress to address the need for research in pregnancy and authorize the FDA to offer pharmaceutical companies both regulatory and financial incentives to study their products for use during pregnancy.
The Best Pharmaceuticals for Children Act and the Pediatric Research Equity Act that were passed by Congress in 2002 and 2003, respectively, can serve as a drug development model for the study of drugs in pregnancy, with the appropriate differences. Together these 2 laws authorize the FDA to provide regulatory and financial incentives (such as grants and extension of marketing exclusivity) for pharmaceutical companies to conduct research in children.
Since the enactment of these 2 laws, the number of pediatric clinical trials has substantially increased. In addition to providing new knowledge regarding pediatric indications and dosing, these laws have also enabled manufacturers to overcome the economic, legal, and ethical obstacles that often deter them from conducting pediatric research.
A mandate from Congress, similar to that discussed in previous text, would give the FDA authority to offer incentives to drug manufacturers to conduct studies in pregnant women. Only when there are compelling scientific and/or ethical reasons to exclude pregnant women would this requirement be waived.
To aid in the selection of which drugs to study, a committee composed of experts selected by the FDA would be established. The function of the committee would be to recommend which of the newly approved or older drugs should undergo a phase IV clinical trial in pregnant women or women of reproductive potential. For short-term therapy (less than 1 month), only pregnant women would be enrolled, whereas for long-term therapy (longer than 1 month), both pregnant women and women of reproductive potential would be enrolled. The study design would involve subjects, as defined in the following text, who are randomized to receive either the newly approved drug or the current standard of care. These trials should include dose finding/pharmacokinetic studies in pregnant patients. Most importantly, the studies would be conducted and published, in a timely manner, in peer-reviewed journals.
The recruitment of these subjects should be based on the following criteria:
- 1.
The drug is likely to be used in women of reproductive potential and/or in pregnancy because it appears to be a first-line therapy for its indication.
- 2.
The drug does not cause developmental toxicity or maternal toxicity in animals at doses or exposures that are 10 times or less the human dose exposure based on BSA or AUC.
- 3.
There are no other approved drugs with the same mechanism of action that are known to cause human developmental toxicity.
- 4.
Effective contraception is not required when taking the drug.
Each woman will be counseled, using evidence-based data (eg, risk of their condition to themselves and their embryo-fetus, drugs that are currently available to treat their condition, animal data, and the estimated benefits and risks of the drug under study before enrolling in the clinical trial), by the following:
- 1.
A physician experienced in the treatment of the disease regarding the expected benefits and risks of adverse outcomes from the new drug compared with other drugs or no treatment.
- 2.
A genetic counselor or other specialist with expertise in clinical teratology regarding the embryo, fetal, and/or neonatal risk of the drug compared with the risks of other drugs, including the background risk of a spontaneous abortion, stillbirth, or a birth defect as well as the risk of not being treated.
- 3.
A maternal-fetal medicine specialist or designated obstetrical care provider regarding what will be done to maximize the likelihood of a healthy baby if she is or becomes pregnant during the clinical trial.
- 4.
A dysmorphologist trained to look for congenital anomalies in newborns will evaluate the newborn.
- 5.
The maternal-fetal medicine specialist or designated obstetrical care provider will discuss the benefits of breast-feeding with the patient antenatally. A pediatrician will follow up the baby during the neonatal period and reinforce the discussions of breast-feeding that occurred during pregnancy.
All of the pregnancies will be managed throughout by the maternal-fetal medicine specialist or designated obstetrical care provider, including those of women who become pregnant during the trial and choose to discontinue treatment. The infants will undergo a standardized examination by a dysmorphologist during the neonatal period and again between 1 and 2 years of age. All specialists involved in counseling and follow-up will be neutral as to whether one option is better than the other (equipoise).
Follow-up of pregnancies, data analysis, interpretation, and reporting will be conducted by or guided by an independent center with experience in birth defects research.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
