The Pregnancy Safety Gap in the U.S.: Why Most Drugs Lack Good Data for Pregnancy — and What That Means

Snapshot: Roughly 90% of drugs approved between 2010–2019 lack adequate human pregnancy data, leaving clinicians and pregnant people to make high-stakes treatment choices with uncertain safety information. That gap is driven by historical exclusion from clinical trials, liability and regulatory caution, and under-resourced pregnancy-specific research — and it has real consequences: women stop needed medicine, fetal risks are under-characterized, and public-health decision-making is limited. The Washington PostPMC

1) How big is the problem? — the core facts

• A major recent analysis and reporting summarized that about 90% of FDA-approved drugs (2010–2019) do not have human pregnancy safety data, meaning dosing, fetal risk, and maternal pharmacokinetics are unstudied or incompletely described. This leaves the majority of therapeutic choices in pregnancy reliant on animal data, registries, or post-marketing surveillance. The Washington PostScienceDirect

• Practically every pregnancy involves medication exposure: estimates suggest most pregnant people use at least one prescription drug during pregnancy. Yet guidance for many commonly used medicines is thin. PMC

(These figures and coverage have been highlighted in recent coverage arguing that proposed research funding cuts would worsen an already fragile evidence base.) The Washington Post

2) Why pregnant people were excluded — short history and structural causes

• Thalidomide era caution: After teratogenic tragedies in the 1960s, regulators and sponsors adopted protective measures that largely excluded pregnant people and those who could become pregnant from early drug development, to avoid fetal harm and liability. That protectionist reflex became entrenched. The Washington Post

• Ethics + liability + complexity: Pregnant people are a “vulnerable” population in earlier research norms; sponsors worry about legal exposure and IRBs have historically erred on exclusion. Pregnancy introduces PK/PD complexities (see Section 4), fetal endpoints, and long follow-up needs that make trials harder and costlier. U.S. Food and Drug Administration+1

• Underfunded pregnancy science: Federal surveillance, registries, and pregnancy-specific pharmacology programs are comparatively small; recent agency cutbacks and budget pressures further threaten capacity to collect post-market pregnancy data. The Washington PostThe Washington Post

3) Clinical consequences — what uncertainty causes in the clinic

• Medication discontinuation risk: Without robust data, some pregnant people stop essential drugs (antidepressants, anti-seizure meds, antihypertensives, diabetes therapies) out of fear — sometimes with harmful maternal or fetal outcomes (relapse of depression, seizures, hypertensive crisis). Journalistic and clinical reports show this is a common dilemma. The Washington PostMGH Women's Mental Health Center

• Dosing errors & undertreatment: Physiologic changes in pregnancy (see Section 4) can cause underdosing or overdosing if dosing isn’t adjusted — for example, faster clearance can make standard doses subtherapeutic for some antibiotics, antiepileptics, or anticoagulants. PMCWiley Online Library

• Teratogenic blindspots: For certain classes (e.g., valproate), strong teratogenic signals were eventually established via registries and observational data — but often after substantial exposure and avoidable harms in some populations. This shows the value of registries but also their reactive nature. PMCAED Pregnancy Registry

4) Why pregnancy changes the drug rules — key physiologic mechanisms

Pregnancy induces predictable physiologic shifts that affect how drugs behave:

• Volume of distribution increases (expanded plasma volume) — lipophilic drugs may distribute differently.

• Increased renal blood flow & GFR — many drugs are cleared faster (e.g., some antibiotics, lithium), potentially requiring higher or more frequent dosing.

• Altered hepatic metabolism — pregnancy changes CYP enzyme activity (some isoenzymes increase, others decrease), altering metabolism of antidepressants, beta-blockers, and other classes.

• Gastrointestinal transit, plasma protein binding, and placental drug transfer are also altered.

These changes mandate pregnancy-specific pharmacokinetic (PK) and pharmacodynamic (PD) study to choose safe, effective doses. Reviews and PBPK modeling papers demonstrate these effects and argue for pregnancy-tailored PK trials and modeling. PMCWiley Online Library+1

5) What the clinical evidence looks like today: registries, RWE, and examples

Because randomized controlled trials (RCTs) in pregnancy are rare, evidence usually comes from:

• Pregnancy exposure registries (e.g., AED pregnancy registries) that track outcomes for mothers and infants exposed to specific drugs. Registries uncovered valproate’s high teratogenic risk and informed regulation/change in guidance. They are invaluable but often limited by voluntary enrollment and confounding. JAMA NetworkAED Pregnancy Registry

• Large observational cohort studies and meta-analyses — for example, pooled analyses of SSRI exposure show modestly increased risks for some cardiac anomalies or preterm birth in some studies, but signals are complex to interpret because untreated maternal illness itself raises risk. Careful meta-analyses try to disentangle these effects. PLOSPMC

• Pharmacokinetic studies & PBPK modeling that estimate dose changes in pregnancy and help design dosing recommendations when RCTs are infeasible. These models are growing in sophistication and regulatory acceptance. FrontiersWiley Online Library

Illustrative examples

• Valproate (sodium valproate): robust registry and cohort data show high risks (neural tube defects, developmental disorders), leading to strict prescribing warnings and pregnancy-prevention programs in some countries. PMCMGH Women's Mental Health Center

• SSRIs (antidepressants): meta-analyses show small but statistically significant associations with some congenital heart defects for specific SSRIs; however, confounding by indication (maternal depression) complicates interpretation and clinical decisions. Experts emphasize balancing relapse risk vs fetal risk. PLOSMGH Women's Mental Health Center

• Antiepileptic drugs: large registries show drug-specific teratogenic risk profiles (e.g., valproate much higher risk than lamotrigine), guiding drug selection in women of childbearing potential. JAMA Network

6) Patient voices & clinician experience — the lived impact

• Patients report fear and confusion. Coverage and patient-advocacy groups document that pregnant people frequently face stark choices: continue a medication with uncertain fetal risks or stop and risk relapse or medical complications. Many describe mixed messaging from clinicians and inconsistent counseling. The Washington PostMGH Women's Mental Health Center

• Clinicians face tradeoffs. Ob/gyns, psychiatrists, and neurologists often rely on imperfect registry data, older cohort studies, and expert consensus. Shared decision-making with clear communication about benefits, unknowns, and monitoring plans is standard practice but time-consuming. MGH Women's Mental Health Center

7) Regulatory responses and recent advances

• FDA guidance evolution: The FDA has issued guidance encouraging inclusion of pregnant and breastfeeding people in trials where appropriate, and new documents (e.g., ICH E21 / FDA E21 discussions) aim to standardize how to include these groups ethically. The FDA’s web pages and recent draft guidance reflect a shift from blanket exclusion to conditional, scientifically justified inclusion. U.S. Food and Drug Administration+1

• Methodological tools: Physiologically based PK (PBPK) modeling, opportunistic PK sampling (during routine care), and registry expansion are being used to fill gaps faster and with less risk than traditional RCTs. Regulators increasingly accept these approaches as supportive evidence. FrontiersWiley Online Library

• Surveillance & registries: Pregnancy registries (AED registries, vaccine pregnancy registries for new vaccines) remain vital. However, they need sustainable funding and better capture to be fully effective. Recent reporting warns that federal program cuts threaten surveillance capacity. EpiCAREThe Washington Post

8) What needs to change — practical solutions (research, regulation, clinical)

1. Design trials that ethically include pregnancy when preclinical data permit: adaptive, phased inclusion (e.g., after safety in non-pregnant adults) with clear consent frameworks. FDA guidance now supports this approach where appropriate. U.S. Food and Drug Administration+1

2. Scale pregnancy PK studies and PBPK modeling earlier in development to set evidence-based doses, not guesswork. Regulators and sponsors should require pregnancy-relevant PK for drugs likely to be used by women of childbearing age. Wiley Online LibraryFrontiers

3. Fund and modernize pregnancy registries (link to EHR, claims data, perinatal surveillance) to capture outcomes prospectively, reduce selection bias, and enable faster signal detection. EpiCAREThe Washington Post

4. Improve clinician guidance & point-of-care tools (decision aids that quantify maternal vs fetal risk and provide monitoring plans) to support shared decision-making. MGH Women's Mental Health Center

5. Protect and prioritize funding for maternal health research and surveillance so the progress of the last decades is not reversed. Recent media coverage warns that funding cuts would materially harm this progress. The Washington PostThe Washington Post

9) Where this affects patients most (high-priority areas)

• Psychiatry: depression and anxiety—stopping meds can cause relapse with serious maternal and fetal consequences; risk–benefit counseling is essential. MGH Women's Mental Health Center

• Neurology (epilepsy): seizure control is vital for maternal and fetal safety, but some antiepileptics (valproate) have high teratogenic risk; registries guide safer drug choice. AED Pregnancy RegistryJAMA Network

• Cardiology & obstetrics: antihypertensives and anticoagulants require pregnancy-specific dosing and safety data. Wiley Online Library

• Infectious disease & vaccines: pregnant people were often excluded from early vaccine trials (e.g., initial COVID-19 vaccine trials), complicating early rollout and trust; pregnancy-specific safety data are now emphasized in pandemic planning. (See CDC, FDA pages and vaccine pregnancy registry practices.) U.S. Food and Drug AdministrationPMC

10) Quick practical guidance for pregnant people & clinicians

For pregnant people

• Don’t stop a prescribed medication without consulting your clinician; abrupt discontinuation of antidepressants or antiepileptics can be dangerous. Discuss risks, alternatives, and monitoring. MGH Women's Mental Health Center

• Ask whether your provider can enroll you in a pregnancy registry if you take a medication with limited data. Registries improve knowledge for future patients. JAMA Network

For clinicians

• Use available registries, up-to-date meta-analyses, and PBPK-informed dosing where RCT data are absent. Practice shared decision-making and document counseling. Wiley Online LibraryEpiCARE

11) Selected references & further reading (key sources)

• Washington Post: “Research cuts will make pregnancy even more dangerous.” (coverage on 90% drugs lacking pregnancy data and funding threats). The Washington Post

• FDA — “Clinical Trials in Pregnant Women” and E21 inclusion guidance (recent regulatory guidance encouraging appropriate inclusion). U.S. Food and Drug Administration+1

• Reviews on physiologic and pharmacokinetic changes in pregnancy (Frontiers, JCPh, etc.). PMC+1Wiley Online Library

• Antiepileptic drug pregnancy registries and valproate teratogenicity evidence. AED Pregnancy RegistryJAMA Network

• Meta-analyses and reviews on SSRIs and pregnancy outcomes. PLOSPMC