The drug is valproic acid, sold under brand names like Depakote and Depakene. It is one of the most widely prescribed anticonvulsant and mood-stabilizing medications in the world, listed on the WHO’s essential medicines roster, and it can cause devastating harm to a developing fetus. Roughly ten percent of babies exposed to valproate during the first trimester are born with a major birth defect — heart malformations, cleft lip, neural tube defects like spina bifida. Children who survive without structural damage still face a sharp increase in neurodevelopmental disorders, including lower IQ scores, autism spectrum disorder, and ADHD. Despite decades of accumulating evidence, FDA warnings, and a formal WHO advisory issued in May 2023, valproate continues to be prescribed to women of childbearing age at rates that have barely budged in fifteen years. This is not an obscure laboratory finding.
In May 2015, a jury awarded $38 million to the family of a child born with spina bifida linked to the mother’s use of Depakote during pregnancy — $15 million in compensatory damages and $23 million in punitive damages against the manufacturer. Hundreds of similar lawsuits have followed, with new cases still being filed as recently as early 2025. Yet the drug remains in widespread clinical use for epilepsy, bipolar disorder, and migraine prevention. For families navigating dementia care, where seizure disorders and mood disturbances can overlap with cognitive decline, this issue has particular relevance. A woman of childbearing age caring for a loved one with dementia may herself be prescribed valproate for a co-occurring condition, often without adequate counseling about the reproductive risks. This article examines why valproate remains so entrenched in medical practice, what the research actually shows about its effects on unborn children, how regulatory agencies have responded, and what practical steps women and their physicians should consider. We will also look at the broader problem of teratogenic drug exposure during pregnancy and what the legal landscape means for affected families.
Table of Contents
- Why Does a Drug Known to Harm Unborn Children Still Get Prescribed So Often?
- What the Research Shows About Birth Defects and Neurodevelopmental Damage
- The Regulatory Response — Strong Warnings, Weak Enforcement
- What Women and Families Should Actually Do — Alternatives and Tradeoffs
- The Legal Reckoning — Lawsuits, Verdicts, and What They Mean
- A Broader Problem — Teratogenic Drug Exposure During Pregnancy in the United States
- Where This Goes From Here
- Conclusion
Why Does a Drug Known to Harm Unborn Children Still Get Prescribed So Often?
The short answer is that valproate works exceptionally well for certain conditions, and medicine has been slow to reconcile therapeutic effectiveness with reproductive harm. For some forms of epilepsy, particularly generalized seizure disorders, valproate remains one of the most reliable options available. Uncontrolled seizures during pregnancy carry their own serious risks, including injury and death for both mother and child. This creates a genuine clinical dilemma that the FDA has acknowledged by maintaining two different pregnancy risk categories for the same drug: Category X for migraine prevention, meaning it is flatly contraindicated, and Category D for epilepsy and bipolar disorder, meaning the risk is established but the drug may still be necessary in some cases. But the dilemma does not fully explain the prescribing patterns. A study published in JAMA Network Open found that the rate of valproate prescriptions in the United States for non-epileptic patients of reproductive age has not declined over the past fifteen years, despite repeated FDA safety communications and black box warnings.
Only 22.3 percent of valproate treatment episodes among women of childbearing age included even a single day of contraception coverage. That statistic alone reveals a systemic failure: doctors are prescribing a known teratogen to women who could become pregnant, and the basic safeguard of ensuring reliable contraception is being met barely a fifth of the time. Compare this to the strict prescribing protocols required for isotretinoin (Accutane), another Category X drug, where patients must complete monthly pregnancy tests and use two forms of contraception. No equivalent system exists for valproate in most clinical settings. The problem is compounded by the fact that roughly half of all pregnancies in the United States are unplanned. A woman taking valproate for bipolar disorder or off-label for mood stabilization may not know she is pregnant until well into the first trimester, which is precisely when the drug does its greatest structural damage to a developing embryo. A November 2025 study reported by ScienceDaily confirmed that valproate remains the most commonly used anti-seizure medication globally, a finding the researchers themselves described as alarming given the weight of evidence against its use in women of childbearing potential.
What the Research Shows About Birth Defects and Neurodevelopmental Damage
The numbers are stark and have been replicated across multiple large-scale studies. Exposure to valproate during the first trimester carries approximately a ten percent risk of major congenital malformations. The new England Journal of Medicine has documented a twenty-fold increase in neural tube defects among valproate-exposed pregnancies compared to pregnancies with no antiepileptic drug exposure. In concrete terms, the risk of a neural tube defect like spina bifida rises from a baseline of about 0.06 percent to between one and two percent — a rate of roughly one in fifty to one in one hundred exposed pregnancies. A 2010 study found that the risk of spina bifida specifically was increased nearly twelve-fold. The neurodevelopmental consequences are equally concerning and, in some ways, more insidious because they may not become apparent until a child reaches school age. The landmark NEAD (Neurodevelopmental Effects of Antiepileptic Drugs) study found that children exposed to valproate in utero scored approximately seven IQ points lower on average than children exposed to other antiepileptic drugs during pregnancy. That gap does not sound enormous in the abstract, but shifted across a population it represents a meaningful increase in the number of children who fall below the threshold for intellectual disability.
Beyond IQ, exposed children face a six-fold higher risk of neurodevelopmental disorders overall compared to non-exposed children. The risk of autism spectrum disorder is estimated at 4.5 percent, and the risk of childhood autism at 2.5 percent — two to four times the general population baseline of roughly one to two percent. After adjusting for confounders, the increased risk holds: a 2.3-fold increase for ASD and a 1.7-fold increase for ADHD. However, it is important to note a limitation in how these findings should be interpreted at an individual level. Population-level statistics describe averages across large groups. Some women who took valproate during pregnancy had children with no detectable problems; others experienced the worst outcomes. Dose appears to matter significantly, with higher doses carrying greater risk, but no safe threshold has been established. For a woman who has already been exposed during an unplanned pregnancy, these statistics describe probability, not certainty — and a clinical genetics consultation can help clarify individual risk in a way that aggregate data cannot.
The Regulatory Response — Strong Warnings, Weak Enforcement
The FDA has not been silent on this issue. Valproate carries a black box warning, the most serious type the agency can mandate, regarding the risks of birth defects and decreased IQ in children exposed during pregnancy. The agency has classified the drug as Category X for migraine prevention, meaning no woman who is or could become pregnant should take it for headaches under any circumstances. For epilepsy and bipolar disorder, the Category D designation acknowledges the risk while leaving the door open for use when alternatives have failed. In 2025, the FDA issued an updated label for Depakote reflecting the latest cognitive impairment data, and Medication Guides handed out at pharmacies are required to describe the pregnancy risks in plain language. Internationally, the response has been more forceful. The WHO issued a formal statement in May 2023 advising against valproate use in women and girls of childbearing potential.
The European Medicines Agency has gone further than the FDA, requiring a visual warning on the outer packaging and mandating that prescribers obtain signed acknowledgment forms from female patients. In France, where a national scandal erupted after thousands of children were found to have been harmed by in utero valproate exposure, new prescriptions to women of childbearing age were restricted to specialists and required annual renewal with documented informed consent. The United States has taken no comparable enforcement steps. There is no prescriber registration program, no mandatory pregnancy testing, and no required consent form. The FDA has relied on label changes and safety communications — tools that have demonstrably failed to change prescribing behavior over a fifteen-year period. A February 2026 fact sheet from the Massachusetts General Hospital Center for Women’s Mental Health reiterated that valproic acid is not recommended for women of childbearing age, but such recommendations remain advisory rather than binding. The gap between what regulators say and what happens in clinical practice is where the harm continues to occur.
What Women and Families Should Actually Do — Alternatives and Tradeoffs
If you or someone you are caring for is a woman of childbearing age currently taking valproate, the most important step is an honest conversation with the prescribing physician about alternatives — ideally before a pregnancy occurs, not after. For epilepsy, lamotrigine (Lamictal) and levetiracetam (Keppra) are generally considered safer options during pregnancy, though neither is risk-free. Lamotrigine in particular has a substantially better safety profile in pregnancy and is often the first-line recommendation for women who need seizure control. The tradeoff is that for certain seizure types, particularly generalized epilepsy with multiple seizure patterns, these alternatives may not be equally effective. Switching medications carries its own risks, including breakthrough seizures during the transition period. For bipolar disorder, the calculus is different.
Lithium has its own pregnancy concerns, including a small but real risk of cardiac malformations, but its overall teratogenic risk is significantly lower than valproate’s. Certain atypical antipsychotics, such as quetiapine, have relatively more reassuring reproductive safety data, though the evidence base is still evolving. The worst scenario is an abrupt, unplanned discontinuation of valproate upon discovering a pregnancy, which can trigger rebound seizures or a psychiatric crisis. This is why preconception planning is so critical: a gradual, monitored transition to a safer medication, done well before conception, gives both mother and child the best chance at a good outcome. For families involved in dementia caregiving, there is an additional layer to consider. Valproate has historically been used off-label for behavioral symptoms in dementia patients, and a caregiver who is herself a patient may be prescribed the same drug by a different physician. Medication reconciliation across multiple prescribers — making sure every doctor knows every drug the patient is taking — is an essential but often overlooked safeguard.
The Legal Reckoning — Lawsuits, Verdicts, and What They Mean
The legal landscape around Depakote has been building for over a decade. Hundreds of lawsuits have been filed against Abbott Laboratories, which originally manufactured the drug, and AbbVie, the company that took over after Abbott’s 2013 split. The central allegation in most cases is that the manufacturer knew about the reproductive risks earlier and more completely than it disclosed to physicians and patients. The 2015 verdict — $38 million to the family of a child born with spina bifida after in utero Depakote exposure — was the first major trial outcome, and it set the tone for subsequent litigation. The punitive damages component, $23 million, signaled that the jury believed the company’s conduct went beyond mere negligence. New lawsuits continue to be filed as of early 2025, now encompassing not just structural birth defects but also neurodevelopmental injuries including autism and ADHD.
These cases are more complex to litigate because the causal chain is longer and the conditions have multiple potential contributing factors. However, the epidemiological evidence linking valproate to these outcomes has grown strong enough to support legal claims. For families affected by fetal valproate exposure, consulting a medical malpractice or pharmaceutical liability attorney may be worthwhile, particularly if the prescribing physician failed to discuss pregnancy risks or offer contraception counseling. Statutes of limitations vary by state, and in cases involving children, the clock often does not start running until the child reaches the age of majority or the injury is discovered. One important limitation: a lawsuit is not a substitute for medical care. Families dealing with the consequences of fetal valproate exposure — whether structural defects, cognitive delays, or behavioral diagnoses — need a multidisciplinary care team regardless of whether they pursue legal action. The legal process is slow, adversarial, and emotionally draining, and it should be pursued only with clear eyes about what it can and cannot achieve.
A Broader Problem — Teratogenic Drug Exposure During Pregnancy in the United States
Valproate is the most dramatic example, but it is far from the only concerning medication prescribed to pregnant women. A University of Florida study reviewing more than three million pregnancies found that one in sixteen women — approximately 6.25 percent — were exposed to a teratogenic drug during pregnancy. Medication use during the first trimester has increased by more than sixty percent over the last three decades in the United States, driven by rising rates of chronic disease, expanded indications for existing drugs, and the simple reality that more women are on long-term medications when they become pregnant.
This trend intersects with the dementia caregiving world in specific ways. Caregivers are disproportionately likely to experience depression, anxiety, and sleep disorders, all of which are commonly treated with medications that carry varying degrees of reproductive risk. The burden falls most heavily on women, who make up roughly two-thirds of unpaid dementia caregivers in the U.S. A caregiver managing her own mental health while supporting a loved one with cognitive decline may not think to raise the question of pregnancy planning with her psychiatrist, and the psychiatrist may not think to ask.
Where This Goes From Here
The trajectory points toward stricter regulation, but the pace is glacial. Advocacy groups in Europe have pushed successfully for prescribing controls that do not yet exist in the United States, and the American model of relying on label warnings has been exposed as inadequate by more than a decade of unchanged prescribing data.
The most promising developments are coming from medical education and clinical decision support systems — tools built into electronic health records that flag valproate prescriptions for women of childbearing age and prompt a documented conversation about risks and alternatives. For families already affected, the growing body of research is a double-edged sword: it validates their experience but also underscores how preventable many of these injuries were. The question going forward is whether the medical system will treat valproate’s reproductive risks with the urgency they deserve, or whether it will take another generation of harmed children and another round of billion-dollar litigation to force the kind of systemic change that the evidence has demanded for years.
Conclusion
Valproic acid is an effective medication that has helped millions of people manage epilepsy and bipolar disorder. It is also one of the most dangerous drugs a pregnant woman can take, responsible for a ten percent rate of major birth defects, a significant drop in childhood IQ scores, and a markedly elevated risk of autism, ADHD, and other neurodevelopmental conditions. The FDA, the WHO, and leading medical institutions have all issued clear warnings. Yet prescribing patterns in the United States have barely changed, contraception counseling is provided in fewer than a quarter of cases, and half of all pregnancies remain unplanned. The system is failing at nearly every checkpoint where it could intervene.
If you are a woman of childbearing age taking valproate, or if you are a caregiver helping manage someone else’s medications, do not wait for the system to catch up. Talk to a physician about alternatives now, before a pregnancy forces the conversation under emergency conditions. If you or your child has already been affected by fetal valproate exposure, seek out a developmental pediatrician or neuropsychologist for evaluation, and consider whether a legal consultation is appropriate. The evidence is no longer ambiguous. The only question is whether it will be acted upon in time.
