The Medicines and Healthcare products Regulatory Agency (MHRA) has reinforced and extended its prescribing restrictions for valproate medicines including sodium valproate (Epilim; Sanofi and generics), valproic acid, and semisodium valproate (Depakote; Viatris) in women and girls of childbearing potential, following a review of compliance data from the Valproate Prevent programme and updated evidence on the neurodevelopmental risks associated with in utero exposure. The current regulatory position, confirmed and updated through MHRA Drug Safety Update communications published in 2023 and 2024, requires that valproate must not be used in women or girls of childbearing potential unless a Pregnancy Prevention Programme (PPP) is in place and the conditions of the programme are actively met.
This article sets out the current prescribing action, the evidence basis for the restrictions, and the practical obligations that fall on prescribers, pharmacists, and clinical teams.
Current Prescribing Action, What Is Required Now
Valproate is contraindicated in women and girls of childbearing potential unless all of the following conditions are met: the patient is enrolled in the MHRA Valproate Pregnancy Prevention Programme; two forms of effective contraception are in use (or the patient is abstinent); the patient has been informed about the risks of valproate in pregnancy using the MHRA-approved patient card and educational materials; and the prescriber has signed an annual risk acknowledgement form confirming that the patient has received appropriate counselling.
The PPP applies from the point of initiation of valproate therapy and must be maintained continuously for any woman or girl of childbearing potential who continues treatment, regardless of how long they have been on valproate or how well their condition is controlled. There is no exception based on treatment duration, diagnostic indication, or clinical history.
The prescribing restriction applies across all licensed indications for valproate: epilepsy, bipolar disorder, and, where it was previously used off-label migraine prophylaxis. Use of valproate for migraine prevention in women and girls of childbearing potential is not appropriate and must be discontinued where it is occurring.
The current MHRA position also requires that any prescriber who cannot confirm that all PPP conditions are being met should not initiate or continue valproate in a woman or girl of childbearing potential, and should arrange urgent specialist review if valproate is the only effective treatment option for a severe or refractory condition.
Signal Classification and Evidence Basis
This is a confirmed adverse drug reaction not a signal under investigation and not spontaneous reporting data alone. The teratogenicity and neurodevelopmental toxicity of valproate in pregnancy have been established through multiple studies, regulatory reviews, and post-marketing surveillance programmes across the UK, EU, and internationally. The evidence base is not contested.
The teratogenic risk associated with valproate exposure in the first trimester includes a significantly increased rate of major congenital malformations, including neural tube defects, cardiac defects, cleft palate, limb abnormalities, and urinary tract anomalies. The absolute risk of major congenital malformations in valproate-exposed pregnancies is estimated at approximately 10%, compared with approximately 2% to 3% in the general epilepsy population not taking valproate and approximately 2% in the general population. The risk is dose-dependent higher maternal valproate doses are associated with higher malformation rates but no dose has been established as safe in pregnancy.
The neurodevelopmental risk is a confirmed adverse drug reaction with a distinct evidence profile from the teratogenic risk. Children exposed to valproate in utero have a significantly increased risk of autism spectrum disorder, intellectual disability, delayed language acquisition, and reduced performance on cognitive assessments compared with children born to mothers with epilepsy not treated with valproate during pregnancy. The NEAD study (Meador et al., New England Journal of Medicine, 2009; n = 311 children; follow-up to age 6 years) reported that children exposed to valproate in utero had a mean IQ approximately 8 to 11 points lower than children exposed to lamotrigine, carbamazepine, or phenytoin, a difference that was statistically significant and clinically meaningful. Subsequent cohort studies and registry data including data from the EUROCAT network and the UK Epilepsy and Pregnancy Register have confirmed and extended these findings.
The risk of autism spectrum disorder in valproate-exposed children is estimated at approximately 3-fold higher than in unexposed populations in several registry analyses, though absolute risk estimates vary across studies due to differences in case ascertainment, diagnostic criteria, and confounding by indication. The MHRA considers this an established risk rather than an association under investigation.
Real fact: UK registry data from the Epilepsy and Pregnancy Register show that valproate is associated with a major congenital malformation rate of approximately 10% in exposed pregnancies, more than three times the background rate in women with epilepsy taking other antiepileptic drugs.
Why Restrictions Have Been Progressively Tightened
The MHRA's Valproate Prevent programme was introduced in 2018 following the EMA PRAC review that concluded existing risk minimisation measures were insufficient to prevent foetal exposure. The programme introduced mandatory educational materials, annual risk acknowledgement forms, patient cards, and the PPP framework. Despite these measures, subsequent NHS audit data collected through 2020 to 2023 revealed that a significant proportion of women of childbearing potential prescribed valproate in primary and secondary care had not completed all PPP steps in some cases had not been counselled on the risks at all, and in some cases were not using effective contraception.
The MHRA's 2023 Drug Safety Update (DSU reference: 2023-04) specifically addressed compliance gaps in primary care, noting that valproate prescriptions originating from GP practices without specialist initiation were associated with the lowest rates of PPP compliance. The update required that GPs prescribing valproate on a repeat basis must confirm PPP compliance at every prescription issue and must not issue prescriptions for women of childbearing potential without documented evidence of current PPP participation. This is an active prescribing obligation at every repeat prescription event, not a one-time counselling requirement at initiation.
Practical Obligations for Prescribers and Pharmacists
For neurologists, psychiatrists, and GPs prescribing valproate in women or girls of childbearing potential, the obligations are specific and non-discretionary. At every prescription point: confirm the patient is enrolled in the PPP; confirm two effective methods of contraception are in use or that the patient is abstinent; check that the signed annual risk acknowledgement form is current (valid for 12 months from signature date); provide or confirm provision of the current version of the MHRA patient card.
For pharmacists dispensing valproate to women or girls of childbearing potential, the obligation includes verifying that the prescription carries confirmation of PPP status where this is required under local dispensing guidance, counselling the patient on the risks at each dispensing event if no recent counselling is documented, and refusing to dispense where there is reasonable clinical concern that PPP conditions are not being met with an immediate referral back to the prescriber.
For specialist centres managing women with epilepsy or bipolar disorder of reproductive age who wish to become pregnant, the clinical pathway should involve preconception counselling with a specialist in neurology or psychiatry, a structured review of whether valproate is the only effective treatment option, and where an alternative antiepileptic or mood stabiliser can be used a managed transition before conception is attempted. Abrupt discontinuation of valproate in women with well-controlled epilepsy carries a significant seizure risk; any medication review must be conducted under specialist supervision with seizure risk assessment integrated into the management plan.
What to Watch For
The MHRA's ongoing monitoring of PPP compliance will generate further enforcement actions if audit data continue to show gaps in primary care prescribing practice. Prescribers who cannot demonstrate PPP compliance documentation for women of childbearing potential on their valproate prescribing lists are exposed to prescribing standard complaints and, in cases where harm results, potential medicolegal consequence.
The EMA's PRAC review of valproate remains an open process, with periodic updates to the risk minimisation measures as new compliance data emerge. Post-Brexit, MHRA's positions track closely with but are independent of EMA PRAC conclusions any further tightening at EMA level will trigger a parallel MHRA review.
For women currently taking valproate who have been inadequately counselled about these risks, the clinical priority is not to create alarm but to ensure that accurate information is provided now, that contraceptive status is reviewed, and that a specialist appointment is arranged to reassess the treatment plan.
The weight of evidence on valproate and pregnancy has been accumulating for over two decades. The MHRA's position is not precautionary it is a proportionate response to a well-characterised and preventable harm. The prescriber's obligation is not merely to be aware of the risk but to actively prevent it at every prescribing interaction.
