Date of Award

January 2015

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Lewis B. Holmes

Second Advisor

Abha Gupta

Subject Area(s)

Medicine, Toxicology, Obstetrics and gynecology


This study aimed to determine if prenatal exposure to carbamazepine (CBZ), lamotrigine (LTG), or valproate (VPA) monotherapies would be associated with adaptive behavior impairments in exposed children, and if these impairments would present in a dose-dependent fashion.

Recruitment letters were mailed to 1,032 women enrolled in the North American Antiepileptic Drug Pregnancy Registry who had taken LTG, VPA, or CBZ as monotherapy throughout pregnancy to suppress seizures. The response rate was 48%. Adaptive behavior of 253 children (104 LTG-exposed, 98 CBZ-exposed, and 51 VPA-exposed), 3- to 6-years-old, was measured using the Vineland-II Adaptive Behavior Scales, administered to each mother by phone. Mean Adaptive Behavior Composite (ABC); domain standard scores for communication, daily living, socialization, and motor skills; and adaptive levels were analyzed across drug groups and correlated with first trimester drug dose. Lower Vineland-II scores indicate greater impairment.

After adjusting for maternal age, education, epilepsy type, prenatal seizures, folate use, cigarette and alcohol exposure, gestational age, and birth weight, the mean ABC score for VPA-exposed children was 95.4 (95% CI [91, 100]), versus 101.1 (95% CI [98, 104]) and 103.5 (95% CI [101, 106]) for CBZ- and LTG-exposed children, respectively (ANOVA; p=0.016). Performance among the three groups differed significantly in all domains except daily living, with VPA-exposed children scoring lowest and LTG-exposed children scoring highest in every category. VPA-exposed children were most likely to perform at an adaptive level that was low or moderately low (standard scores ≤ 85) in each category. Generalized linear models showed higher VPA dose to be associated with significantly lower ABC (p=0.021), socialization (p=0.009), and motor (p=0.041) scores, with a trend toward significance in the communication domain (p=0.055).

Unlike CBZ and LTG, prenatal VPA exposure was associated with adaptive behavior impairments in a pattern consistent with autistic symptoms. The dose-dependent effect suggests that VPA should be avoided during pregnancy and that VPA-exposed children should be routinely referred for early intervention services.