Objective: To study the relationships between maternal valproate dose in pregnancy and the pattern of various fetal malformations. Methods: Analysis of data in the Australian Register of Antiepileptic Drugs in Pregnancy collected from 1999 to 2012. The specific type of fetal malformation in offspring exposed to valproate in utero was correlated with the dose of valproate taken by the mother in the first trimester. Results: Compared with other malformations, the mean dose of valproate taken during the first trimester was higher in mothers whose offspring had spina bifida (2,000 ± 707 vs 1,257 ± 918 mg/d) and hypospadias (2,417 ± 1,320 vs 1,235 ± 715 mg/d) (both p < 0.05). The overall mean maternal valproate dosage taken by women in the Register decreased over the last 5 years of the study period. This was paralleled by a statistically significant decrease in the rate of occurrence of spina bifida and hypospadias, but not other malformations. Conclusions: Human fetal malformations associated with valproate exposure during pregnancy do not all seem to bear the same quantitative relationship to drug dose, and reduction in valproate dose in earlier pregnancy is likely to offer greater dividends in protecting against spina bifida and hypospadias than against other types of fetal malformations.