Abstracts

Rationale: Behavior problems are more prevalent in children with epilepsy than in healthy children or those with other chronic illnesses. Past studies show that behavior often improves following epilepsy surgery; however, the role of surgical side remains unclear. Further, few studies have used an epilepsy control group to determine if observed changes are attributable to surgery or merely related to passage of time. This study examined differences in behavior outcome among children who underwent right- or left-sided surgeries and a nonsurgical control group of children with epilepsy on medical therapy. Methods: Caregivers for 136 children with intractable epilepsy (mean age=10.9 years) completed the Child Behavior Checklist (CBCL) before and ~6 months after surgery. Caregivers for 76 children with medically managed epilepsy (mean age=10.2 years) also completed this measure at similar time intervals, without intervening surgery. First, base rates of clinically elevated behavior problems were examined at baseline using chi-square analyses. Then, ANOVAs compared mean differences in change scores on the CBCL Syndrome Scales between the three patient groups (right surgery, left surgery, no surgery) followed by post hoc comparisons (Scheffé adjustment). Finally, individual patient scores were classified as borderline/clinical or normal, and loglinear models examined changes in this clinical classification among the three groups over time. Results: There was a high prevalence of baseline behavioral problems among all three groups across all CBCL scales (15% - 46%), except Rule-Breaking Behavior (5-8%). Social Problems and Aggressive Behavior were significantly higher in right surgical patients (42% and 31%) than left surgical (26% and 22%) or nonsurgical (22% and 15%) patients. ANOVAs revealed significant group differences in mean change scores only on the Thought Problems subscale; children who underwent right-sided resections showed greater improvements than the nonsurgical group. Examination of change in clinical category among individual patients using loglinear models revealed no statistically significant differences between the three groups. However, there was a trend on the Attention Problems scale (p=.09); nonsurgical patients had a higher rate of de novo attention problems than surgical patients (20% versus 10-13%). Qualitatively, many children in the borderline/clinical group improved to the normal range of behavior at follow-up, regardless of treatment group (right surgery = 22-75%, left surgery = 19-83%, no surgery = 42-67%). De novo behavior problems were relatively rare (2-20%), occurring most often following left-sided resections (10-15%). Conclusions: Consistent with existing literature, there was a high prevalence of behavior problems among all the pediatric epilepsy patients in this sample. Many children with pre-existing behavior problems improved to the normal range over time regardless of group, suggesting that behavioral improvement is not directly attributable to surgery. Encouragingly, the occurrence of de novo behavior problems was relatively rare.