In 2021, amid the COVID‑19 pandemic, Kristin Wall became pregnant with her second child. Her physician told her that little was known about the COVID-19 vaccine’s safety and effectiveness in pregnant people. Observational data — collected from those vaccinated before they knew that they were pregnant — suggested that the vaccine was safe, so she could have it. Still, she’d have to weigh up the risks and benefits herself.
Women’s health research lacks funding — these charts show how
Wall is an epidemiologist at Emory University in Atlanta, Georgia, and is therefore unusually qualified to assess health risks. But her experience is a common one — and far from unique to COVID-19. When it comes to pregnancy, people must often make important medical decisions with limited or imperfect information, says Wall. Pregnancy changes the body, including how it metabolizes and distributes medications, so a drug with a predictable effect in a non-pregnant person might work very differently in someone who is pregnant.
There is now solid evidence from the huge numbers of people vaccinated, and subsequent clinical trials, that the COVID-19 vaccines that use messenger RNA (mRNA) — as well as influenza vaccines and others that do not use live viruses — are generally safe and effective during pregnancy. Safety is also well established for certain other medications: some anti-allergy drugs, some antidepressants and some antibiotics are considered safe for use in pregnancy. Paracetamol (acetaminophen) — the main ingredient in Tylenol — is also considered safe, despite claims to the contrary by US President Donald Trump. Other drugs, however, are known to carry risks. An example is the anti-epilepsy drug sodium valproate, which can cause severe developmental disorders when taken during pregnancy.
Most people take medications during pregnancy — more than 90% in the United States according to some tallies1. Despite this, the safety and effectiveness of many drugs — including some anti-nausea treatments, some anti-inflammatories and the vast majority of medicines approved over the past decade — remain uncertain because they have never been tested in pregnant people. Fears over adverse consequences and associated policy restrictions mean that pregnant people have been largely excluded from clinical trials. That leaves public-health officials and health-care providers basing recommendations on indirect evidence, including animal studies, trials in non-pregnant people and observational studies of pregnant people taking a drug after it is already on the market.
“The gap is huge,” says Wall. She and others have designed trials that have successfully tested treatments for Ebola, HIV and some pregnancy-specific conditions in pregnant people — collecting reliable data while protecting the health of study participants and their babies. Based on that experience, they have developed tools and guidance to help other researchers to include pregnant people in studies, too. But despite these advances, logistical, regulatory and ethical barriers remain.
What happens if pregnant women stop taking Tylenol?
There’s no evidence yet that Trump-era policies, such as the ‘Make America Healthy Again’ movement, have reduced the inclusion of pregnant people in trials, Wall says. But she adds that the broader increase in anti-pharmaceutical rhetoric might reinforce long-standing hesitancy to study medications in pregnancy — slowing efforts to close evidence gaps.
For Denise Astill, who took sodium valproate during pregnancy before clear guidance advised against its use, real progress “is about equity”. Her twin daughters were diagnosed at the age of four with a wide range of health conditions, including cognitive difficulties, hearing problems, kidney impairment, autism and attention deficit hyperactivity disorder.
Her experience led her to found the charity Foetal Anti-Convulsant Syndrome New Zealand in Auckland, which offers guidance on taking anti-seizure medications in pregnancy. Pregnant people should have the same access to evidence-based care as everyone else, she says. “Can we truly make an informed choice when our health system and research are based on men?”
Thalidomide’s enduring impact
Much of the caution around research in pregnancy stems from a mid-twentieth-century medical disaster. In the late 1950s, a drug called thalidomide was marketed in 46 countries as a morning-sickness treatment, although it had not been tested in pregnant people. Physicians soon saw a surge in the number of babies with missing or shortened limbs born to mothers who had taken the drug during pregnancy.
In the early 1960s, after studies confirmed that it was linked to congenital disorders in more than 10,000 children, thalidomide was withdrawn. The disaster reshaped drug regulation, leading to stricter testing standards that ended up excluding pregnant people.
In 1964, the Declaration of Helsinki — a document on ethical principles for medical research involving humans — classified pregnant people as a ‘vulnerable population’, grouping them with people who are unable to give informed consent, such as children or those with severe cognitive impairments.
In 1977, the US drug regulator barred not only pregnant people but also ‘women of childbearing potential’ from participating in clinical trials. Although this policy began to shift in the late 1980s and early 1990s — when the US National Institutes of Health required women to be included in research — pregnant people continued to be sidelined because they were considered to be a vulnerable group.
Over the past decade, some of these restrictions have eased. The World Health Organization (WHO) now says that pregnant people should be considered ‘complex’ rather than ‘vulnerable’ and be included in trials when possible. But progress remains slow: an analysis of US drug trials between 2008 and 2023 found that fewer than 1% specified that they included pregnant participants2 (see ‘Enrolment gap’), and fewer than 0.4% of trials submitted in the European Union include pregnant people.
Source: Ref. 2
Once a therapy is approved, it can take a long time for evidence to accumulate. In the case of COVID-19 vaccines, so many people were having the shot that evidence built up quickly. By the end of the pandemic, most countries recommended vaccination for pregnant people. But a review of medicines approved by the US Food and Drug Administration (FDA) found that when a drug’s safety in pregnancy is unknown at the time of approval, it takes, on average, 27 years before there’s enough evidence to confidently rate its safety3.
This evidence gap leaves physicians and pregnant people with a difficult choice: treat a condition with drugs for which the risks are uncertain, or leave it untreated, says Alyssa Bilinski, a statistician at Brown University in Providence, Rhode Island. “Sometimes ‘ask your doctor’ is really good advice, because your doctor has a lot of information,” she says. “But ‘ask your doctor’ cannot substitute for underinvestment in research over decades.”
Bilinski’s work suggests that excluding pregnant people from randomized controlled trials can do much more harm than good. In a study last year, she and her colleagues looked at two kinds of risk: testing a drug in pregnant people if the drug turns out to cause harm, and not testing it in pregnant people at all. Their model estimated that if thalidomide had been tested in a trial with 200 pregnant participants, about 33 children would have been harmed1. But such a trial could have prevented about 8,000 congenital disorders, according to the model, or 99.6% of all thalidomide-related cases between 1956 and 1962.
The fraught quest to account for sex in biology research
The same model also suggested that lives would have been saved if pregnant people hadn’t been excluded from early COVID-19 vaccine trials. When the US drug regulator first authorized the vaccines in December 2020, no pregnancy-specific data were available. Although access wasn’t restricted, the US Centers for Disease Control and Prevention did not strongly recommend the vaccine in pregnancy until August 2021 — by which point uptake by pregnant people already lagged behind uptake in the general population.
If pregnancy-specific data from clinical trials had been available, the resulting vaccine uptake and reduction in infections would have prevented about 20% of COVID-related maternal deaths and stillbirths in the United States between March and November 2021, totalling more than 200 deaths, the study concluded.
Many people assume that risk comes only from taking medications, but doing nothing can be even more dangerous, leaving both the pregnant individual and their baby exposed to uncontrolled disease or serious infections, Bilinski says. “What happens absent a trial is quite possibly the worst of both worlds.”
Breaking down barriers
Some of the barriers to including pregnant people in clinical trials are real and others are perceived, says Anna David, an obstetrician and maternal–fetal medicine specialist at University College London. A common myth is that pregnant people won’t volunteer, David says. Her experience shows the opposite: when they are given as much information as possible about a trial, they are able to give informed consent, and many are willing to participate.
Another perceived barrier, says David, is the idea that fetal health can’t be assessed during a trial, which would be essential to identify adverse effects. “That’s completely wrong, because as a fetal medicine specialist, I do that all the time,” she says.
To make it easier to track and compare risks in trials including pregnant people, David and her colleagues developed standardized definitions and grading scales for problems that can affect the person and the fetus, including issues detectable with fetal imaging, such as growth problems and fluid build-up4. Published in 2021, these measures have been incorporated into drug trials in pregnant people, cited in dozens of papers and are being validated by regulatory agencies, David says.
Despite these resources, she adds, pharmaceutical companies are often hesitant to test drugs in pregnant people. That’s partly owing to fear of reputational damage from adverse outcomes. “Imagine headlines saying, ‘new drug has bad outcome’ — you wouldn’t want that,” she says.
Another barrier is the difficulty that researchers face in securing insurance, David says. When deciding whether to cover a trial, insurers often rely on data from past clinical trials to assess risk, including complications such as congenital disorders. But because so few trials are conducted in pregnant populations, there is little information to guide these assessments. That uncertainty leads insurers to decline coverage or to charge high premiums.
Evidence shows that COVID-19 mRNA vaccines are generally safe during pregnancy, but for a lot of medications, such data are scant.Credit: Lauren DeCicca/Getty
Even when insurance is in place, clinical trials in pregnant people are typically more expensive than those in other people, requiring specialized expertise and infrastructure. One estimate by the non-profit organization the Concept Foundation in Geneva, Switzerland (see go.nature.com/4tdnvtm) suggests that developing a medicine specifically for use during pregnancy could cost an extra US$5.7 million compared with drugs for use in other therapeutic areas, with most of the added cost due to safety and efficacy studies.
