SEATTLE—Pregnant women with epilepsy who take antiepileptic drugs (AEDs) have an increased risk of congenital malformations in their offspring, with certain drugs posing a higher risk than others, reported Torbjörn Tomson, MD, PhD, at the 62nd Annual Meeting of the American Epilepsy Society. Valproate was strongly associated with an increased risk of congenital malformations, while other AEDs such as lamotrigine and carbamazepine also showed an increased risk.

A pooled analysis of studies revealed that the prevalence of malformation in offspring of women who were taking AEDs is about 6%, whereas malformation prevalence in the general population is about 0.2%, similar to the rate among women with epilepsy who are not taking AEDs, said Dr. Tomson, Associate Professor and Senior Staff Physician at the Karolinska Institute in Stockholm. In addition, polytherapy with AEDs has a higher malformation rate than monotherapy with AEDs (6.8% vs 4%).

Comparing Malformation Rates Among AEDs
To compare different AEDs with incidence of malformations, Dr. Tomson and colleagues retrospectively analyzed data from epilepsy and pregnancy registries. In the North American Antiepileptic Drug Pregnancy Registry, the highest malformation rate was associated with valproate (10.7%), followed by phenobarbital (6.5%) and lamotrigine (2.3%). The rate for lamotrigine was comparable to that of carbamazepine. In addition, valproate had a malformation rate that was about seven times higher than the external control population. Phenobarbital had a malformation rate five to six times higher than the control population. Lamotrigine and carbamazepine rates were not significantly higher than control population rates.

In the UK Epilepsy and Pregnancy Register, the malformation rate with valproate was 6.3%, whereas lamotrigine and carbamazepine malformation rates were significantly lower. The Lamotrigine Pregnancy Registry, the Finnish Drug Prescription database, the Swedish Medical Birth registry, and the UK, North American, and Australian AED and pregnancy registries all showed a pattern of high malformation rates with valproate.

Is There a Dose-Response Relationship?
In the UK database, a high dosage of lamotrigine (200 mg/day) yielded a malformation rate similar to that observed with the medium dose of valproate (600 to 1,000 mg/day). Malformation rates greatly increased with a lamotrigine dose of about 400 mg/day. At this dose, malformations were greater than 6%. However, a dose-effect relationship was not confirmed within the lamotrigine registry. Malformation rates did not seem to be higher at dosages of 400 mg/day or less, though there are few exposures at this dose, noted Dr. Tomson.

The North American registry also did not reveal a dose-related difference in malformation rates for lamotrigine. Mean doses of lamotrigine were similar for mothers of offspring with malformations, compared with those of children without malformations (345 vs 390 mg/day).

Valproate had a clearer dose-effect relationship, said Dr. Tomson. A low-risk dose of valproate (less than 1,000 or 1,100 mg/day) produced a malformation rate range between 2% and 5.4%.

An Increased Risk of Clefts
According to the North American registry, five clefts among 684 therapy exposures with lamotrigine resulted in a prevalence of 7.3 per 1,000 exposed, about 10 times higher than expected in the control population. Prevalence from the other registries varied considerably—from 1.5 to 11.5 per 1,000 exposed—a reflection of the small number of patients exposed to therapy. Overall prevalence was 3.9 per 1,000 exposed (2.5 per 1,000 exposed, excluding North America).

In contrast, case-control data from the European Surveillance of Congenital Anomalies study found no specific increased risks of facial clefts relative to other malformations with lamotrigine monotherapy. “This does not excuse that there is an increased risk of clefts with lamotrigine,” Dr. Tomson pointed out.

Confounding Factors
“We have to remember that these are population studies,” stated Dr. Tomson. “Women were not randomized to specific treatments.” To control for confounding factors, the researchers compared children of a group of pregnancies in the European and International Registry of Antiepileptic Drugs in Pregnancy who had a healthy outcome with offspring who had four specific types of malformations—oral clefts, heart defects, hyperspadia, and neural tube defects.

Valproate was used in about 20% of the no anomaly group, whereas it was used among 40% of the mothers of children with neural tube defects and hyperspadia. A family history of defects was prevalent in 4% of those without malformations; this was four times higher in the group with heart defects and oral clefts. In the no anomaly group, 30% of fathers had a high level of education. In the group with oral clefts, this rate was significantly smaller. “This suggests that more factors than AEDs play a role as confounders in the risk of birth defects,” Dr. Tomson stated.

The investigators tried to assess independent variables such as drug treatment, education level of the parents, type of epilepsy, family history of defects, maternal age, parity, and seizure control during pregnancy. Unfortunately, this yielded unreliable and unstable estimates, indicating too few pregnancies.

“We need to know more about the less frequently used monotherapies and different types of combinations,” Dr. Tomson commented. More research is also needed to define what causes specific malformations and the influence of confounding factors, he concluded.

—Marguerite Spellman

Do AEDs Increase the Risk for Birth Defects?