Influencing clinical practice regarding the use of antiepileptic medications during pregnancy: modeling the potential impact on the prevalences of spina bifida and cleft palate in the United States.

Selected antiepileptic drugs (AEDs) increase the risk of birth defects. To assess the impact of influencing AED prescribing practices on spina bifida and cleft palate we searched the literature for estimates of the association between valproic acid or carbamazepine use during pregnancy and these defects and summarized the associations using meta-analyses. We estimated distributions of the prevalence of valproic acid and carbamazepine use among women of childbearing age based on analyses of four data sets. We estimated the attributable fractions and the number of children born with each defect that could be prevented annually in the United States if valproic acid and carbamazepine were not used during pregnancy. The summary odds ratio estimate for the association between valproic acid and spina bifida was 11.9 (95% uncertainty interval (UI): 4.0-21.2); for valproic acid and cleft palate 5.8 (95% UI: 3.3-9.5); for carbamazepine and spina bifida 3.6 (95% UI: 1.3-7.8); and for carbamazepine and cleft palate 2.4 (95% UI: 1.1-4.5) in the United States. Approximately 40 infants (95% UI: 10-100) with spina bifida and 35 infants (95% UI: 10-70) with cleft palate could be born without these defects each year if valproic acid were not used during pregnancy; 5 infants (95% UI: 0-15) with spina bifida and 5 infants (95% UI: 0-15) with cleft palate could be born without these defects each year if carbamazepine were not used during pregnancy. This modeling approach could be extended to other medications to estimate the impact of translating pharmacoepidemiologic data to evidence-based prenatal care practice.