USE OF SODIUM VALPROATE IN PREGNANCY

Please note – This monograph has received an interim update. Selected figures have been amended; the published data remains current to December 2025 and the guidance is unchanged.

Sodium valproate is an antiseizure medication used either as sole or adjunctive treatment for all forms of epilepsy. Sodium valproate and the related compounds valproate semisodium and valproic acid are also used in the treatment of acute manic episodes in bipolar disorder, and off-licence for migraine prophylaxis.

Sodium valproate (referred to hereafter as valproate or VPA) is an established human teratogen and should therefore only be used in women of childbearing potential when there is no safer effective alternative to treat epilepsy and following a discussion of the risks and benefits. Valproate should not be used to treat bipolar disorder in pregnancy. All healthcare providers should refer to the MHRA guidance documents before prescribing valproate (as a new or repeat therapy) to girls, pregnant women, or women of childbearing potential. The fetal risks associated with use of valproate in pregnancy are assumed to apply to all valproate-containing compounds. Where use of valproate during pregnancy or in women of childbearing potential is clinically necessary, use of the lowest effective dose is advised.

Valproate monotherapy in the first trimester is generally associated with an absolute risk of infant congenital malformation (CM) of up to 11% (i.e. more than three times increased above background rates). Malformations caused by in utero valproate exposure include neural tube defects (NTD), particularly spina bifida, cardiac defects, orofacial clefts and hypospadias. Malformations of the skull, limb and digits have also been documented. Although there is no ‘safe’ threshold dose, a dose-response relationship has been observed, with a number of studies identifying 1,000mg/day as the dose above which CM risk significantly increases. CM risk of up to 62.1% has been reported with valproate doses of 3,000mg/day.

Only one study has assessed the CM risk associated with dosing regimens that may avoid high peak plasma levels, such as divided dosing or use of modified release preparations. No significant difference in risk was observed compared to once daily dosing. Use of divided doses or slow-release preparations is, however, advised where possible, based on a theoretical benefit.

Overall, there is no strong evidence that valproate use in polytherapy is associated with a risk of congenital anomaly higher than that observed at the same dose used in monotherapy but data are limited and require confirmation.

Valproate impairs the production of the active metabolite of folate. It is therefore recommended that women who take valproate also take high dose folic acid (5mg daily) prior to conception and throughout the first trimester. Women should, however, be counselled that while folic acid reduces the overall risk of NTDs in the general population, there is no strong evidence that it protects against NTDs and other adverse outcomes specifically caused by valproate.

Valproate is also a neurobehavioural teratogen that has adverse effects on features, including offspring IQ and motor skills. Data suggest that the likelihood of autism spectrum disorder may be up to three-fold increased following prenatal valproate exposure, compared to the ~1% incidence in the background population. Studies have noted that between 19% and 62% of valproate-exposed children require additional educational support. Individuals with neurodevelopmental disorders caused by prenatal valproate exposure may not exhibit overt structural anomalies.

The relationship between maternal valproate dose and offspring neurodevelopment is not fully established. A 2014 Cochrane systematic review identified a potential association between poorer cognitive outcome and daily valproate doses in excess of 800 to 1,000mg.

Data relating to miscarriage are limited but high doses of valproate (>750mg/day) have been associated with a possible increased risk. There is weak evidence that folate supplementation may offer a protective effect with doses of valproate <1,000mg/day.

The available data suggest that valproate exposure may be associated with low infant birth weight and additional fetal growth monitoring in the third trimester may therefore be indicated. There is no evidence of an increased risk of preterm delivery. Data relating to stillbirth are limited, and although two large studies have suggested a crude likelihood of ~1% (compared to expected background risk of ~0.5%), lack of statistical analysis means that this outcome requires further assessment before conclusions regarding causality can be drawn.

Use of any centrally acting drug throughout pregnancy or near delivery may lead to poor neonatal adaptation syndrome (PNAS). There are reports of adverse neonatal symptoms following in utero exposure to valproate, although the absolute risk is unquantified. Withdrawal symptoms are likely to be more severe following exposure to more than one CNS acting drug. Delivery should be planned in a unit with facilities for neonatal care.

Regular clinical review is indicated for pregnant women who continue treatment with valproate as plasma concentrations may change as pregnancy progresses and dose adjustment may be required to maintain therapeutic levels.

Periconceptual paternal valproate use is not linked to offspring congenital anomaly. Valproate exposure has been associated with impaired male fertility that may be reversible on cessation of treatment. A single study indicated a potential association with neurodevelopmental impairment in the children of periconceptually exposed men, but this finding was not replicated when part of the original dataset was systematically reanalysed, including comprehensive sub-analyses designed to account for residual confounding. As a precautionary measure, the UK Medicines and Healthcare products Regulatory Agency (MHRA) advises that men should avoid fathering a pregnancy while using valproate and for three months following treatment. UKTIS pragmatically advises that, as the initial signal suggesting altered neurodevelopment in the offspring of exposed fathers has not been reproduced, prescribers should consider the potential risks of valproate discontinuation and the likelihood that an alternative treatment will be effective when carrying out a medication review for a man planning to father a pregnancy. For further information, or patient specific risk assessments, please contact UKTIS.

Women exposed to valproate during early pregnancy should be offered detailed anomaly scans from 16 weeks gestational age, although some anomalies linked to valproate will not be detected by ultrasound. Patients should also be counselled about the increased risk of adverse neurodevelopmental outcomes, which are often not noted at least until after infancy. Other risk factors may also be present in individual cases which may independently increase the risk of adverse pregnancy outcome. Clinicians are reminded of the importance of consideration of such factors when performing case-specific risk assessments.