Despite making up almost half of the world’s population, women have historically been underrepresented in clinical studies. It was only in 1993 that the NIH Revitalization Act established that women and minorities had to be included in all NIH-funded clinical research. That same year, the FDA issued guidance encouraging the inclusion of women in clinical trials, including early-phase studies. This legacy of exclusion has left critical gaps in our understanding of how diseases and treatments affect women, with potential consequences for drug safety and efficacy.
For decades, medical research conducted primarily on men has been used to draw insights into diseases, drug efficacy, and medical devices — often applying these findings to women without sufficient evidence. However, a woman’s anatomy, genetics, hormonal balances, and organ systems differ significantly from those of men. Given these differences, it is neither accurate nor appropriate to assume that treatments studied predominantly in men will have the same effects on women.
The inclusion of women in clinical trials has significantly increased since the 1990s, but representation is still uneven. In 2022, a large-scale analysis of 1,433 clinical trials involving more than 300,000 participants found that females accounted for just 41.2 percent of participants on average — despite representing 51 percent of the US population. The analysis examined trials across cardiovascular disease, psychiatric disease, and cancer, comparing female participation with the proportion of women affected by each disease. Across all three therapeutic areas, women were underrepresented relative to their expected share of the disease population.
But inclusion is only part of the problem. “While the number of women enrolled in studies has grown, the ability to make meaningful conclusions regarding biological sex differences based upon those enrollments remains limited due to underpowered study designs and an absence of consideration of hormonal status or life stages,” Brian Honeyman, a board-certified physician at Marietta Springs, told DDN.
In other words, enrolling women does not automatically make a trial capable of answering questions about sex differences. So what does meaningful inclusion actually look like — and what might drug development still be missing when it comes to understanding how medicines affect women?
Inclusion isn’t enough
For decades, researchers often treated the male body as the default when studying disease and developing treatments. However, there is substantial evidence that sex can influence disease presentation, pharmacokinetics, treatment response, and adverse effects. That means enrolling women without subsequently examining their outcomes separately can still leave important biological differences hidden.
“Simply including women in a study does not guarantee that the results will provide useful information regarding sex differences. A study must be designed to look at sex differences and the results must be analyzed and reported by sex to create scientifically valid results,” Ayesha Bryant, Clinical Advisor of Alpas Wellness, told DDN.
Despite this, sex-specific analysis is still uncommon. In a 2022 analysis of Australian clinical literature, only seven of 71 mixed-sex studies analyzed the data by sex. Another study from 2024 examined sex outcome comparisons across oncology clinical trials and found that only 472 of 89,221 trials (0.5 percent) reported any sex comparison data. Yet among the trials that did report these comparisons, the researchers found differences in both treatment outcomes and toxicity. Women generally had better post-treatment outcomes in several cancer types, while men experienced fewer side effects. The researchers also identified sex-specific differences associated with particular treatments, including EGFR inhibitors in non-small cell lung cancer and rituximab in non-Hodgkin lymphoma.
Simply including women in a study does not guarantee that the results will provide useful information regarding sex differences. A study must be designed to look at sex differences and the results must be analyzed and reported by sex to create scientifically valid results.
—Ayesha Bryant, Alpas Wellness
This is not isolated to oncology. Sex differences have been found to influence transplant outcomes, antidepressant response, trauma response, Parkinson’s disease, metabolic disorders, and so much more. If clinical outcomes are averaged across an entire study population, a treatment effect that occurs primarily in one sex can become extremely difficult to see.
As evidence accumulates across therapeutic areas, researchers are increasingly recognizing that sex differences are not isolated exceptions but fundamental components of physiology. Rather than introducing unnecessary complexity, incorporating sex into study design offers an opportunity to better understand disease mechanisms, identify more precise therapeutic targets, and improve the likelihood that promising drug candidates will translate successfully into the clinic.
When averaging hides a difference
Not interrogating sex differences can have insidious consequences. One example of this is digoxin therapy, which once was one of the most commonly prescribed medications for heart failure. However, a post hoc analysis published in 2002 found that women taking digoxin had a significantly higher mortality rate than women taking placebo, an effect that was not observed in men. Subsequent analyses suggested that the higher mortality among women may have been linked to higher digoxin concentrations in the blood, even one month after dosing.
The findings highlight the complexity behind how sex can influence a drug's effect . The post hoc analysis suggested that differences in heart failure pathophysiology, hormonal influences, and even hormone replacement therapy could all contribute to how patients respond to the same treatment.
These differences are not unique to digoxin. Drug absorption, distribution, metabolism, and excretion often vary between men and women, affecting how quickly a medicine is cleared from the body and how much of it patients are exposed to.
In 2024, a pharmacokinetic analysis of 99 approved anticancer drugs found significant sex-related differences in clearance for 15 medicines, with some differences exceeding 20 percent. This included chemotherapy agents, targeted therapies, and atezolizumab, an immune checkpoint inhibitor. In general, women were found to clear these drugs more slowly, increasing their drug exposure and potentially contributing to differences in toxicity or treatment response.
Differences in how men and women process medicines are often attributed to something as simple as body weight. However, the researchers found sex differences even after accounting for weight, suggesting that other biological factors may affect how drugs behave in the body. These could include differences in body composition, blood flow to organs, and drug metabolism.
“Researchers need to evaluate whether differences in factors such as metabolism, hormones, immune response, pharmacokinetics and disease progression may affect how a treatment works or how side effects are experienced. Trials should be designed with these questions in mind and, where appropriate, powered to analyze outcomes by sex rather than relying solely on overall results,” Katrina Rice, Chief Delivery Officer at eClinical Solutions, told DDN.
Representation of women across life stages
One of the biggest differences between men and women is that female bodies constantly change as they age. Women are not a biologically static population, and factors such as hormonal changes, pregnancy, and menopause can alter how the body responds to medicines. Yet these changes are rarely accounted for in drug development.
“Sex differences interact with age and reproductive status,” Charul Avachat, a University of Minnesota-trained clinical pharmacologist, told DDN. “A drug's behavior in a 30-year-old woman, a pregnant woman, a postmenopausal woman, and an age-matched man can all differ in clinically relevant ways that a simple male to female comparison would miss entirely.”
That makes life-stage biology an important extension of sex-based research. Avachat recently published a study investigating how age and sex affect the use and pharmacokinetics of an antiseizure medication, lamotrigine. The researchers analyzed electronic health records from nearly 315,000 adults, including almost 24,000 people who had been prescribed lamotrigine. They found that postmenopausal women cleared the drug 22 percent more slowly than younger women and nine percent more slowly than older men.
The findings add to evidence that menopause can alter how the body processes medicines. Several other studies have reported changes in drug clearance or blood concentrations around menopause, including the anesthetic alfentanil, the antipsychotic quetiapine, and the immunosuppressant tacrolimus.
Despite this, relatively few studies have specifically investigated how menopause affects drug pharmacokinetics, leaving clinicians with limited evidence to guide prescribing in this population. Considering that women are more likely to develop chronic conditions and take multiple medications as they age, understanding how menopause affects drug exposure is increasingly important. Without this evidence, clinicians have to rely on dosing recommendations based on younger women or men, potentially increasing the risk of under- or overexposure to medicines.
And this is just one life stage of many. Pregnancy presents an even more striking example.
The pregnancy paradox
Pregnant and breastfeeding women are among the populations with some of the greatest need for drug safety data — yet they have historically been among those most often excluded from clinical research.
More than 90 percent of pregnant women take at least one medication during pregnancy, and 30 percent take five or more medications during pregnancy. This could be for pain management or pregnancy-related symptoms, or for chronic conditions such as epilepsy, depression, and schizophrenia, that still need treatment during and after pregnancy.
However, pregnancy causes substantial physiological changes that can alter drug pharmacokinetics and potentially affect the dose needed to achieve an appropriate exposure. Additionally, these drugs can cross the placenta and reach the developing fetus or pass into breast milk and expose breastfeeding infants. Understanding how pregnancy and lactation affect drug exposure is therefore important not only for ensuring that medicines work effectively for the mother, but also for assessing potential risks to the fetus or infant.
The result is a paradox. Pregnant and breastfeeding women potentially have the most to lose from poorly understood drug exposure, yet they are also among those for whom the least direct clinical trial evidence is available. A 2025 review found that only 0.8 percent of 44,160 clinical trials included pregnant patients, and that pregnant patients were explicitly excluded from 75 percent of trials and not mentioned in another 24 percent.
“The 1993 mandate got women in the door, but pregnant and lactating women, arguably the group with the highest-stakes, and the most urgent need for real data are still functionally excluded from the evidence base almost the whole way through drug development, both before and after approval,” said Avachat. Her own research shows that the gap doesn't close once a drug reaches the market. In her 2023 study in Clinical Pharmacology & Therapeutics, Avachat and colleagues looked at 488 new molecular entities approved by the FDA between 2000 and 2022. Only 12 percent came with post-marketing commitments or requirements to study use in pregnancy or lactation. Most of those relied on observational pregnancy registries that took a median of 10 years to complete. Only one called for pharmacokinetic studies in pregnant women, and nearly all lactation studies were requested only after 2017.
The exclusion of pregnant patients from clinical studies was historically intended to protect fetuses from potentially harmful drug exposures. However, excluding them from clinical research does not eliminate the potential risk. Instead, it can shift that uncertainty into post-marketing use, when medications are prescribed to pregnant patients without the benefit of evidence from controlled clinical trials.
This was seen during the height of the COVID-19 pandemic, when pregnant women were excluded from many vaccine and treatment trials despite being at higher risk of complications from the virus. Following approval, pregnant patients and their providers were left to make treatment and vaccination decisions in the absence of evidence from well-designed trials.
One of the scariest aspects for a pregnant patient is not knowing how much of the drug reaches the fetus, and what happens after it gets there. Avachat's recent research has begun to address this question. In the MONEAD study of 199 women with epilepsy and 207 infants, she and her colleagues measured antiseizure medication concentrations in maternal and umbilical cord blood. The researchers found that several commonly used medications reached the fetal circulation, providing evidence of placental transfer. “Gabapentin was one example where we saw a higher concentration in the umbilical cord than in the mother’s blood, suggesting that the fetus could be exposed to even more of the drug than the mother,” said Avachat.
But knowing that a drug reaches the fetus is only the beginning. The effects of prenatal exposure may not be immediately apparent, making long-term follow-up essential. The MONEAD study also followed children exposed to antiseizure medications during pregnancy and assessed their development up to age six. While there was no difference in verbal abilities at age six between children of women with epilepsy and children of healthy women, the researchers did find associations between third-trimester antiseizure medication concentrations and some neurodevelopmental measures. For example, the researchers observed a potential association between higher concentrations of levetiracetam and lower verbal scores.
Avachat also pointed to the precursor study to MONEAD, known as the NEAD study, which definitively linked prenatal exposure to valproic acid with significantly poorer neurodevelopmental outcomes. "I remember meeting a mother of two at an epilepsy conference who had taken valproic acid during her first pregnancy. Her child was later diagnosed with autism. It's difficult to see this data emerging years after a drug has already been approved." Avachat said.
These studies show the dangers of leaving pregnant women out of clinical research. Without prospective data, clinicians may not know how pregnancy changes drug exposure, how much medication reaches the fetus, or whether effects will emerge months or years after birth. But how can researchers include these vulnerable populations while minimizing the potential for harm?
Collecting the data without exposing patients unnecessarily
The ethical difficulties around pregnancy research are real. Researchers do not want to expose pregnant women or developing fetuses to an experimental drug without sufficient evidence that the potential benefits justify the risks. But exclusion does not have to mean waiting until a medicine has been on the market for decades before gathering meaningful data.
Instead, researchers can use a combination of approaches to build evidence while minimizing unnecessary exposure to medicines. These can include observational studies, pregnancy and lactation registries, pharmacokinetic sampling during routine clinical care, placental models, and computational modelling.
Physiologically based pharmacokinetic, or PBPK, models combine information about a drug's properties with physiological data to simulate how drug concentrations change throughout the body. Researchers can adapt these models to account for physiological changes during pregnancy or lactation and use them to estimate drug exposure in populations that can be difficult to study directly.
“PBPK models incorporate aspects of biology and physiology, such as a faster heart rate, greater blood volume, faster kidney clearance and altered liver metabolism. PBPK models can thus predict how physiological changes affect drug exposure in both the mother and the baby,” said Avachat.
A 2025 study from the ConcePTION project, for example, used PBPK modelling to predict how 11 medicines move from maternal blood into human milk. The researchers developed and tested models against existing clinical data before using them to predict milk concentrations and estimate the amount of each medicine an infant could receive through breastfeeding. The models reproduced observed plasma and milk concentration profiles for the medicines studied.
These models are not limited to predicting how much drug reaches the bloodstream. They can also be used to explore how differences in physiology affect distribution, metabolism, and elimination, as well as how a given drug concentration translates into a therapeutic response or adverse effect. In principle, this could help researchers identify where men and women may require different doses or dosing schedules before testing those regimens in large clinical trials.
These approaches cannot answer every question, and they do not remove the need for clinical data. Models need real-world measurements for development and validation, while laboratory systems cannot fully reproduce the complexity of pregnancy or lactation. But they can help researchers narrow the uncertainty, identify medicines that need closer investigation, and determine where direct clinical studies would provide the most useful information.
From compliance to understanding
The regulatory framework around women's participation in research has changed considerably since 1993. More recently, the Food and Drug Omnibus Reform Act of 2022 introduced Diversity Action Plans for clinical trials, requiring sponsors to outline strategies for improving enrollment diversity. The FDA subsequently issued draft guidance on implementing these plans.
But representation targets cannot, on their own, answer the biological questions that remain. “It is not enough to ask whether women were included in a clinical trial. We also need to ask: Which women were included? Were enough included to draw meaningful conclusions? Were outcomes analyzed by sex? And did we use what we learned to make the treatment safer and more effective for the patients who will ultimately receive it?” said Rice. “Representation gets people into the dataset, but thoughtful study design and analysis are what allow us to learn from their participation.”
Representation gets people into the dataset, but thoughtful study design and analysis are what allow us to learn from their participation.
—Katrina Rice, eClinical Solutions
A trial can enroll women in numbers that satisfy a diversity requirement and still fail to establish whether a drug's efficacy, safety, or pharmacokinetics differ by sex. It can also enroll women without collecting enough information about age, hormonal status, pregnancy, contraception, or menopause to understand why responses might differ.
Therefore, the next stage is to design trials in which the resulting data can actually answer questions about women. That means planning sex-specific analyses from the beginning, ensuring studies are adequately powered to detect meaningful differences, considering age and reproductive status alongside sex, and building pathways for gathering pregnancy and lactation data earlier in development.
It could also mean following participants and their children for longer when exposure during pregnancy is involved, rather than relying primarily on short-term observations.
The goal is not to complicate drug development for the sake of it. It is to make sure that when a medicine reaches patients, the evidence behind it reflects the people who will actually take it.
Read about the need for female inclusion in preclinical research here: https://www.drugdiscoverynews.com/why-sex-matters-in-preclinical-research-17414









