Article Text

other Versions

Cohort study
Prenatal exposure to sodium valproate is associated with increased risk of childhood autism and autistic spectrum disorder
  1. Amanda Wood
  1. Department of Psychology, University of Birmingham, Birmingham, UK
  1. Correspondence to: Dr Amanda Wood, Department of Psychology, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK;

Statistics from

Commentary on: OpenUrlCrossRefPubMedWeb of Science

Implications for practice and research

  • Anticonvulsant drugs increase the risk of poor neurodevelopmental outcome.

  • Women taking sodium valproate should be counselled about potential risks well in advance of pregnancy.

  • Further research is required to understand whether lower doses reduce the risk of autism to population levels.

  • Prospective longitudinal studies assessing all children are necessary to clarify the full extent of risk associated with prenatal medication exposures.


The clinical management of women with epilepsy during pregnancy is complicated by the potential teratogenicity of medications needed to control seizures. Past research about the relationship between prenatal exposure to antiepileptic drugs and major congenital malformations has improved understanding of the immediate effects for the child, with recent studies highlighting the long-term impact on cognitive development.1 ,2 This study addresses concerns raised previously3 about the risk of autism spectrum disorders (ASD) in children exposed to sodium valproate (VPA) during pregnancy.


This Danish population-based data-linked study reviewed the National Population Register with prescription data, psychiatric register and birth records. Records were collated for 655 615 eligible children, with 508 exposed to VPA and 2136 to other anticonvulsants.

A diagnosis of autism was identified by searching for relevant codes from the standard International Classification of Diseases (ICD-10) codes for ASD or childhood autism (CA) in mothers who filled prescriptions for VPA from 1996 to 2006. The results are reported as adjusted HR and CI.

The primary outcome was the HR of autism associated with prenatal exposure to VPA. Regression analyses controlled for known risk factors for autism and examined the influence of maternal epilepsy on autism risk.


The average age of children at follow-up was 8.84 years (range 4–14 years). An increased risk of ASD (HR=2.9, 95% CI 1.7 to 4.9) and CA (HR=5.2, 95% CI 2.7 to 10.0) was found in children exposed to VPA during pregnancy. The risk of ASD was elevated compared to non-exposure when VPA was taken in the first trimester as well as in children whose mothers filled prescriptions only after the first trimester. An increased risk of ASD/CA in the children of women without epilepsy who took VPA for other conditions as well as in children without major congenital malformations was also reported. The level of risk was similar for high (>750 mg/day) and low-dose VPA exposure.


This study highlights the importance of long-term monitoring in children exposed to medications during pregnancy. The use of population-based data avoids potential selection bias in voluntary registers and the study's scale permits subanalyses to explore clinically relevant questions. This is the strongest evidence to date that VPA taken in pregnancy, and not epilepsy itself, is an important risk for autism in children. The important question of timing of exposure remains equivocal because of insufficient data about VPA exposure limited to the second and/or third trimesters. Furthermore, only ‘prescriptions filled’ are reported which may not be an accurate reflection of the dose taken, which precludes specific interpretations of dose-related risks.

Data-linkage studies may under-report cases that meet the criteria for ASD or CA. Recent data suggest that direct assessment of children at school age detects previously undiagnosed cases of autism.4 In accordance with earlier studies3 only children with an existing diagnosis at the time of study follow-up contributed to risk estimates; further research with prospective recruitment and assessment of autism is required. An accurate estimate of the true rate of neurodevelopmental disorders is important if we are to understand how best to identify women using VPA whose children are at greatest risk of ASD/CA.

Women face difficult decisions regarding treatment during pregnancy. In practical terms, preconceptual counselling is imperative in women with epilepsy, as well as women required to take medication to treat chronic health conditions. The updated UK National Institute for Health and Care Excellence (NICE) guidelines for epilepsy highlight this need, with a view to balancing the potential risk to the unborn child against the mother's personal circumstances (eg, implications of poor seizure control for driving, education and occupation).5 Although the VPA prescription rates reported in this database and others6 have reduced in the last decade, women whose seizures are not well-controlled by newer medications continue to use VPA.


View Abstract


  • Competing interests AW is a member of the Executive and Scientific Advisory Committee of the Australian Pregnancy Register for Women with Epilepsy and Allied Disorders. The Australian Pregnancy Register receives financial support for its operations from the pharmaceutical industry, including Sanofi-Aventis, UCB, Janssen, Novartis and Pfizer Inc, as well as past support from GlaxoSmithKline.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.