Cancer Risk in Children of Mothers With Epilepsy and High-Dose Folic Acid Use During Pregnancy

Key Points Question Is prenatal exposure to high-dose folic acid (≥1 mg daily) associated with risk of cancer in children born to mothers with epilepsy? Findings In this cohort study of 3 379 171 children, compared with no high-dose folic acid use in pregnancy for mothers with epilepsy, use among those with epilepsy was associated with an increased risk of cancer in their children. Meaning Findings suggest that cancer risk in children should be considered in the risk-benefit analysis of folic acid supplementation for pregnant women with epilepsy.


eTable 1. Ethical Approvals
Ethical approvals were necessary to retrieve and manage data from the national social and health registers. This included approvals from the Data Protection Agency in Denmark; the Norwegian Inspectorate and the Regional Ethics Committee for Medical Research in Western Norway; and the Regional Ethical Board at Karolinska Institutet in Sweden. Below is an overview of approvals.  Test for interaction between antiseizure medication and high dose folic acid on the risk of cancer in the offspring was performed in children to mothers with and without epilepsy. Level of significance was set to p<0.05. Test for interaction was based on the fully adjusted model as presented in Table 2 by adding an interaction term to the analysis.

Test for interaction between antiseizure medication (ASM) and high dose folic acid on cancer risk in children to mothers with and without epilepsy
Test for interaction p-value Children to mothers with epilepsy 0.76 Children to mothers without epilepsy 0.48 Abbreviations: ASM: Antiseizure medication.

eTable 4. Risk of Childhood Cancer During the First 10 Years of Life in Children Born to Mothers With and Without Epilepsy Filling Prescriptions for High-Dose Folic Acid
The risk of childhood cancer during the first ten first years of follow-up after birth was investigated for children born to mothers with and without diagnosis of epilepsy, filling prescription for high dose folic acid before and during pregnancy. The risk of cancer was 3.2 [95% CI 1. 2-8.7] in children during the first ten years of follow-up if they were born to mothers with epilepsy filling prescription of high dose folic acid, compared to mothers with epilepsy that did not have such a prescription.
Association between maternal epilepsy, filled prescription of high dose folic acid, and risk of childhood cancer in the offspring up to age 10 years We calculated the HR for the most common childhood cancer type, leukemia, in children to mothers with or without epilepsy filling for high dose folic acid (1mg or 5mg folic acid). There was not sufficient no. of exposed cancer cases to report HRs for other cancer subtypes. The average delivered dose of folic acid was calculated based on prescription fills for 1 mg and 5 mg of folic acid, filled between 90 days prior to the date of last menstrual period and until birth. This was categorized into >0 mg to <4 mg daily and >4 mg daily. These categories were examined for mothers with and without epilepsy, comparing both groups with mothers that had no fill for high dose folic acid. We were unable to make more categories due to few cancer cases among the exposed children. The estimated mean dose in mothers with and without epilepsy were 4.3mg and 2.9mg, respectively. Therefore, we reran the analysis as presented in eTable 7, but here comparing use of mean daily prescribed dose of folic acid >0 to <3mg and >3mg. The numbers have for this table been rounded to closest ten due to close similarity with eTable 7 with the possibility to calculate difference with less than 5 exposed cancer cases. The risk of childhood cancers with maternal fills for ASM was examined. This was performed regardless of high dose folic acid use or not to be able to obtain a sufficient number of exposed cancer cases to report risk estimates.

Association between average daily dose of folic acid and risk of cancer in the offspring to mothers with and without epilepsy
Hazard rates (HR) of cancer in offspring of women with prescription fills for antiseizure-medication (ASM) regardless of high dose folic acid prescriptions, stratified for maternal epilepsy and types and combinations of antiseizure medications. Birth year, sex of the child and source country were applied as stratum for all models.
Counts have been rounded to closest ten due to close similarity to counts presented in eTable 7.

Hazard rates (HR) of cancer in offspring of women with prescription fills for antiseizure-medication (ASM) regardless of high dose folic acid prescriptions, stratified for maternal epilepsy and types and combinations of antiseizure medications.
Type of ASM Filled prescription of ASM

Hazard rates (HR) of cancer in offspring of women with prescription fills for antiseizure-medication (ASM) regardless of high dose folic acid prescriptions, stratified for maternal epilepsy and types and combinations of antiseizure medications.
Type of ASM Filled prescription of ASM We stratified children born to women with and without epilepsy with no further stratification on whether they were exposed to high dose folic acid or antiseizure medication, to examine any effect of maternal epilepsy. Three adjustment models were performed as described in the statistical analysis. We did not observe any association between maternal epilepsy and risk of childhood cancer in the offspring (aHR=1.0, 95% CI 0.7-1.4). Abbreviations: aHR: adjusted hazard ratio; HR: hazard ratio a) aHR1: Adjustment for maternal age and education b) aHR2: Including adjustment for antiseizure medication exposure. c) aHR3: Including adjustment for maternal BMI, prior births with congenital anomalies, smoking during pregnancy, number of hospitalizations.

Association between maternal epilepsy and risk of childhood cancer in the offspring
Birth year, sex of the child and source country were applied as stratum for all models.

eTable 11. Restrictions
For the main analysis in Table 2, showing risk of cancer in children born to mothers with epilepsy and with and without high dose folic acid exposure, we reran the analysis with separate restrictions on covariates that might have influenced the association. This included maternal cancer before pregnancy, maternal tuberous sclerosis, maternal diabetes mellitus one year before birth. Maternal fill for valproate or carbamazepine included both monotherapy and combination with other ASMs. We also included separate exclusions on children with major congenital anomaly and chromosomal abnormality.