Background Since the thalidomide tragedy in the 1960s, many women are concerned with the use of drugs in pregnancy. Antiepileptic drugs (AEDs) are commonly used to treat epilepsy and bipolar disorders in women of child bearing age. However, some AEDs are associated with an increased risk of major congenital malformations. Untreated epilepsy and bipolar disorders in pregnancy may lead to devastating consequences for both mother and child. This leaves women and health care professionals with a dilemma as whether to continue AED medication in pregnancy. Using data from THIN, a UK primary care database, we examined pregnancy as a determinant for ceasing AED treatment.

Methods A cohort study of pregnant women receiving AEDs in UK primary care was conducted. We identified women who were prescribed AEDs in the three months before pregnancy. Time to last consecutive AED prescription in pregnancy was estimated whereby discontinuation of therapy was defined by a gap of more than three months between prescriptions. We identified a control group of twice as many non-pregnant women receiving AEDs. These women were randomly selected, but stratified within five year age bands and matched on indication for AEDs as for the pregnant women Cox’s regression was used to compare the likelihood of discontinuing AEDs between pregnant and non-pregnant women.

Results In total, 934 women were prescribed AEDs three months before pregnancy. Pregnant women with epilepsy were twice as likely to cease AEDs compared to non-pregnant women (Hazard Ratio (HR):2.00, 95% confidence interval (CI:1.62–2.47). Of 745 women with epilepsy, 601 (80.7%) continued treatment into pregnancy and 465 (62.4%) to the end of the second trimester. Of 1,490 non-pregnant women with epilepsy, 1,242 (83.4%) and 1,071 (71.9%) continued for comparable time periods.

Pregnant women with bipolar disorder or depression were three times as likely to cease AEDs compared to non-pregnant women (HR:3.07, 95% CI:2.04–4.62). Of 54 pregnant women with bipolar disorder, 27 (50.0%) continued into pregnancy, and only 8 (14.8%) to the end of the second trimester. In 108 non-pregnant women with bipolar disorder, 82 (75.9%) and 58 (53.7%) continued for comparable periods.

Conclusion Pregnancy is a determinant for the discontinuation of AEDs during pregnancy, especially in women with bipolar disorder or depression despite the potentially severe consequences associated with not treating the underlying illness in pregnancy.