Objectives: To evaluate the risk of childhood brain tumor occurrence in relation to epilepsy and anticonvulsant use.
Study design: As part of a multicenter case-control study of pediatric brain tumors, maternal report on epilepsy occurrence before diagnosis of her child's brain tumor was collected for 540 cases and compared with 801 control children. Mothers also reported on any long-term (> or = 2 weeks) use of medications by her child before the date of tumor diagnosis (or a comparable reference date for controls) and these medications were classified according to whether they contained barbiturates.
Results: As expected, because seizures are often an early brain tumor symptom, a strong association was observed between epilepsy and brain tumor occurrence (odds ratio, OR = 6.2; 95% confidence limit, CL = 2.9, 14). The association remained elevated even after a > or = 10-year interval between diagnoses of epilepsy and brain tumor (OR = 4.7; CL = 0.8, 48). Elevated odds ratios were observed both for epileptic children who were treated with anticonvulsants containing barbiturates (OR = 5.8; CL = 2.2, 18) and for those not treated with barbiturates (OR = 7.9; CL = 1.7, 74), relative to nonepileptic children.
Conclusion: Whereas most of the brain tumor risk associated with epilepsy may be due to occult tumors, the finding of an elevated risk many years after diagnosis of epilepsy is of interest.