The associations between maternal epilepsy and anticonvulsant drug therapy with the risk of oral clefts in the offspring were investigated using data from a population-based case-control study. Cases included 238 infants with cleft lip +/- cleft palate (CLP) and 107 infants with cleft palate (CP) ascertained through the Metropolitan Atlanta Congenital Defects Program (MACDP) between 1968 and 1980. Controls included 3029 population-based normal infants. Histories of maternal epilepsy and drug therapy during pregnancy were compared between cases and controls using maternal interviews and reviews of hospital medical records. Maternal epilepsy was associated with increased risk of nonsyndromic CLP (OR = 3.78, 95% C.I. 1.65-7.88), and less with CP (OR = 1.75, 95% C.I. 0.20-6.99). Therapy during pregnancy was associated with the greatest excess risk (CLP OR = 7.77, C.I. 2.02-26.0; CP OR = 3.61, C.I. 0.08-26.5). The use of polytherapy was associated with the highest risk (CLP OR = 10.5, C.I. 1.52-59.9). Adjustment for potential confounding variables in the study did not change these findings. In this well-defined population, maternal epilepsy and its treatment account for a small proportion of nonsyndromic oral clefts (attributable fraction CLP = 3.3%, CP = 0.9%).