New Data on Valproic Acid and Malformation Risk

August 15, 2010

August 2010, ObGyn News, Lee S. Cohen, MD

Concerns regarding fetal exposure to valproic acid have been longstanding. Many teratovigilance programs around the world have identified an increased risk for congenital malformations associated with first trimester exposure to valproic acid, particularly neural tube defects. Data from the North American Antiepileptic Drug Pregnancy Registry indicate that the risk of neural tube defects following monotherapy with valproic acid during the first trimester is about 10%. Until recently, fewer data have been available on other malformations associated with fetal exposure to this antiepileptic medication.

A large study published in June attempted to further delineate the risks of major malformations associated with first trimester exposure to valproic acid monotherapy, using a two-step novel approach. First, investigators identified a teratogenic signal, that is, a spectrum of malformations that were more common than expected in offspring exposed to valproic acid during the first trimester based on a series of cohort studies. Then they sought to confirm the signal for a subset of those malformations by conducting a population-based, case-control study using a very large population-based database of congenital anomalies by determining the frequency of valproic acid exposure associated with those malformations (N. Engl. J. Med. 2010;362:2185-93).

In the eight-cohort studies, there were 118 major congenital malformations in 1,565 pregnancies, during which women took valproic acid (a rate of 7.5%). Of those, 14 malformations were found more commonly in valproic acid-exposed offspring: The case-control study used the antiepileptic study database established by European Surveillance of Congenital Anomalies (EUROCAT) of over 3 million live births and about 98,000 major congenital malformations in 14 European countries from 1995 through 2005. Comparing 37,154 cases of any of the 14 malformations with 39,472 controls with other types of major malformations and 11,763 controls with malformations associated with chromosomal abnormalities, first trimester exposure to valproic acid was associated with an increased risk of 6 of the 14 malformations, compared with the two control groups: spina bifida (a 12.7 greater risk), atrial septal defect (2.5), cleft palate (5.2), hypospadias (4.8), polydactyly (2.2), and craniosynostosis (6.8).

Despite the challenges involved when conducting these types of analyses—including inconsistent criteria used to categorize anomalies across studies or registries—the findings reported in this article are highly consistent with findings from previous studies and are, therefore, confirmatory while also refining the risk estimates for major malformations associated with valproic acid other than neural tube defects.

The authors for the EUROCAT Antiepileptic Study Working Group concluded that their results provided further support for the American Academy of Neurology’s recommendation to avoid treatment with valproic acid during pregnancy. Neurologists have certainly been aware of the reproductive risks of valproic acid for many years and typically choose alternative anticonvulsants with comparable efficacy for women of reproductive age.

The situation is somewhat different in psychiatry, however, because valproic acid is a mainstay of therapy for the treatment of bipolar disorder. In previous columns, I have discussed the use of valproic acid, as well as lithium, lamotrigine, and atypical antipsychotics for women of reproductive age with bipolar disorder.

These include a review of the evidence associating in utero exposure throughout pregnancy to valproic acid with neurobehavioral sequelae in children (Ob.Gyn. News, Dec. 15, 2004). I also reviewed the reproductive safety of first trimester exposure to lithium, which is associated with a 0.05% risk (1 in 2,000) of the cardiac malformation Ebstein’s anomaly; and the still sparse reproductive safety data for the newer atypical antipsychotics, such as olanzapine, quetiapine, and risperidone (Ob.Gyn. News, April 15, 2008). First trimester exposure to lamotrigine, another treatment option for bipolar illness, may be associated with a 0.9% risk of oral clefts based on at least one registry but has not been associated with overall increase risk of major congenital malformations. However, lamotrigine monotherapy does not seem to be as effective for the full spectrum of patients clinicians see with bipolar disorder, particularly those with predominantly manic-like symptoms.

In our program, we counsel hundreds of women of reproductive age each year who have bipolar disorder about the reproductive safety of medication options. Given its teratogenicity and behavioral teratogenicity, valproic acid should be considered contraindicated during pregnancy for the treatment of bipolar disorder with only the most rare exceptions, if any.

One option to consider with respect to treatment of bipolar disorder during pregnancy is lithium, where the teratogenic risk is established but is quantified and small (0.05%).

Another option might even include a class of medications where the reproductive safety is not yet clearly defined. Although this goes against the maxim “a known is better than an unknown,” medicines such as atypical antipsychotics, which do not seem to be associated with teratogenic risks based on limited experience, may be a better choice than a medicine—namely, valproic acid—where the quantified risk for a variety of malformations is consistently very, very great based on systematically conducted studies. As an example, we might use lithium or lamotrigine or even lamotrigine plus an atypical antipsychotic because the alternative—valproic acid—is so teratogenic.

Considering the declining use of lithium to treat bipolar disorder over the past few years and its known teratogenicity, and the very indicting data regarding valproic acid, we have increasing interest in identifying reproductive safety data of other medicines used to treat the illness, such as atypical antipsychotics. Studies are underway to address these issues, including a recently established national pregnancy registry for atypical antipsychotics (see For now, we treat patients with bipolar disorder who plan to get pregnant on a case-by-case basis, realizing that it might be reasonable to use medicines with quantifiable risks or a treatment that does not have a safety signal, compared with drugs with clearly significant risk for both organ dysgenesis and behavioral teratogenicity.