Reexamining the effects of epilepsy surgery on IQ in children: use of regression-based change scores.

Prior studies have found no adverse effects of pediatric epilepsy surgery on IQ. However, empirical techniques such as regression models, designed to account for confounding factors such as practice effects and test-retest reliability and able to provide a standardized method for evaluating outcome, have not been used in studying change after pediatric epilepsy. The goal of this study was to demonstrate the regression technique while empirically measuring the effect of epilepsy surgery on IQ in a group of pediatric patients. Predictors of retest IQ (e.g., baseline IQ, retest interval, demographics, epilepsy severity) were evaluated in a control group with intractable seizures (N = 23) assessed twice with the WISC-III. The resulting equation was used to evaluate IQ changes in a second group of children who underwent epilepsy surgery (N = 22). In controls, baseline IQ was a strong predictor of retest IQ. Number of AEDs was inversely related to retest IQ. Based on the control regression, four children (18%) in the surgical sample obtained significantly higher than expected postsurgical IQ scores and one child (5%) obtained a lower than expected IQ score. This study demonstrates that regression-based techniques yield informative estimates on outcome and may be an improvement over prior methods of measuring change after pediatric epilepsy surgery.