Intended for healthcare professionals


Teratogenicity of antiepileptic drugs

BMJ 2006; 333 doi: (Published 21 September 2006) Cite this as: BMJ 2006;333:615
  1. David P Breen (davebreen{at}, foundation year 2 doctor in colorectal surgery,
  2. Richard J Davenport, consultant neurologist
  1. Department of Colorectal Surgery, Western General Hospital, Edinburgh EH4 2XU
  2. Department of Clinical Neurosciences, Western General Hospital, Edinburgh EH4 2XU

    Women should consider stopping, minimising, or switching drugs before pregnancy

    Prescribing for women with epilepsy is complicated by the potential teratogenicity of antiepileptic drugs. Current guidelines recommend that the most effective drug should be chosen before conception and prescribed at its lowest effective dose, ideally as monotherapy.12 But which antiepileptic drug is safest in pregnancy?

    Early research on the safety of antiepileptic drugs in pregnancy was unreliable. Several countries set up pregnancy registries in the late 1990s, and data from these registries are now appearing.

    To date the UK Epilepsy and Pregnancy Registry has recruited more than 3500 women, of whom 72% were given antiepileptic monotherapy. The overall rate of major congenital malformation in women given antiepileptic drugs during pregnancy was 4.2%, compared with 3.5% in women with epilepsy who were not given such drugs.3 By three months of age, infants exposed to sodium valproate monotherapy during gestation had the highest frequency of major congenital malformation (6.2%), confirming similar findings from an Australian register.4 Lamotrigine monotherapy was associated with a 3.2% frequency of malformation, but in a multivariate analysis this frequency was not significantly different from that seen with valproate monotherapy. The risk with lamotrigine at doses above 200 mg a day was similar to that of valproate doses of ≤ 1000 mg/day. Carbamazepine was associated with the lowest frequency of major congenital malformation (2.2% for monotherapy). Polytherapy with antiepileptic drugs was associated with a significantly higher frequency of major malformation than monotherapy (6% v 3.7%).

    The North American Pregnancy Registry found that valproate monotherapy was associated with a 10.7% frequency of major congenital malformation. This represents an increased relative risk of 7.3 compared with a control group from one US teaching hospital's malformation surveillance programme.5 The US registry had previously reported an increased risk of malformation in babies exposed to phenobarbital6—an important finding as this antiepileptic drug is still widely used in many countries.

    These registries provide observational data only, and include many variables (such as type of epilepsy, seizure frequency, and drug compliance) that may influence results. Furthermore, major congenital malformations are defined differently by each registry, so results are not directly comparable.

    The potential for antiepileptic drugs to cause developmental delay in childhood is even more difficult to measure than major congenital malformation. Adab and colleagues found that valproate monotherapy in pregnancy was associated with decreased verbal IQ when compared with carbamazepine or phenytoin monotherapy, and that this was dose related.w1 They also reported that 30% of children exposed to valproate needed special educational support in school, compared with 3-6% of those exposed to monotherapy with other antiepileptic drugs.w2 Similar results were reported in a Finnish study.w3 A further study has, however, shown adverse neurodevelopmental effects in children exposed to a variety of antiepileptic drugs during gestation, not only valproate.w4 The neurodevelopment effects of antiepileptic drugs (NEAD) study is currently investigating behavioural outcomes in children exposed to antiepileptic drugs in pregnancy (

    Overall, evidence shows that valproate is associated with a higher frequency of major congenital malformation than other antiepileptic drugs and seems more likely to cause neurodevelopmental delay. Carbamazepine is associated with the lowest incidence of major congenital malformation, while lamotrigine is probably not as safe as previously thought. No reliable data are available on the safety of most of the newer antiepileptic drugs in pregnancy.

    What does all of this mean for doctors and their patients? Ideally, women with epilepsy should plan pregnancy in advance and discuss this with their doctor and epilepsy nurse. All women should take folic acid 5 mg a day before conception and for at least the first trimester of pregnancy.

    Doctors should consider whether individual patients need antiepileptic drugs at all during pregnancy. Women with focal seizures alone, women who have been seizure-free for at least two years, or women who have infrequent generalised seizures may prefer to stop their antiepileptic drugs, but this must be assessed on an individual basis. If treatment is needed during pregnancy, carbamazepine seems to be the safest option.

    While most authorities recommend either valproate or lamotrigine as first line treatment for idiopathic generalised epilepsy, little evidence exists to support this notion,w5 except that some types of epilepsy—such as juvenile myoclonic epilepsy—may be exacerbated by carbamazepine. Women already established on valproate or high dose lamotrigine face a difficult decision when planning pregnancy. Their options include maintaining current treatment, lowering the dose as far as possible, or considering swapping to carbamazepine. Women taking more than one drug should consider switching to monotherapy, although many do not alter their antiepileptic drug regimen during pregnancy.w6 In practice, many pregnancies are unplanned, so doctors and epilepsy nurses should discuss this issue with all women of childbearing potential when epilepsy is first diagnosed.

    It remains to be seen whether any of the newer antiepileptic drugs will prove to be less teratogenic. In the meantime, we should use the best available evidence to guide our decisions and the advice we give to women with epilepsy, and reassure them that they are likely to have a successful uncomplicated pregnancy.


    • Competing interests RJD has received consultancy fees or speaker's fees (or both) from Pfizer (gabapentin, phenytoin), GSK (lamotrigine), Janssen Cilag (topiramate), and Sanofi (valproate).

    • Embedded Image References w1-w6 are on


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    View Abstract