Abstracts

Rationale: Existing research on memory outcome following pediatric epilepsy surgery has yielded varied results, likely due to methodological differences across studies and failure to account for factors that can impact test-retest change in the absence of any surgical effect (e.g., test reliability, practice effects, measurement error). Reliable change indices (RCIs), which control for these important artifacts, were recently published for children with epilepsy. The current study made use of these RCIs to examine base rates of clinically meaningful change in memory performance following temporal lobe resection (TLR) in children with pharmacoresistant seizures. The study also examined differences in outcome as a function of surgical side and identified factors associated with memory change following surgery.Methods: Fifty-two children completed the Children’s Memory Scale (CMS) before and approximately 7 months after TLR (left= 28; right=24). Children were included if they were: 6 to 16 years old, underwent TLR that included mesial structures, and were right-handed or left hemisphere dominant for language. We calculated change scores for each of the CMS Indices and then categorized them as “decline” or “no decline” based on published RCIs (90% confidence interval) for children with epilepsy. We calculated base rates of memory decline for each of the CMS indices and utilized chi-squares to examine differences in memory outcome as a function of surgical side. We then used T-tests and chi-squares to identify factors associated with verbal memory decline following left TLR.Results: The base rate of decline was 12% for Visual Immediate Memory and 23% for Visual Delayed Memory. There was no significant difference between right and left surgical groups in the proportion of patients who demonstrated visual memory declines (p value range = .284-.310) While the overall base rate of verbal memory decline was similar (ranging from 10 to 21% across indices), declines in verbal memory occurred much more frequently following left TLR (18-36%) than right TLR (0-4%)(p value range = .029-.005). Declines in verbal memory following left TLR were most consistently associated with older age at seizure onset (Cohen’s d range = 1.09 to 1.63) and higher preoperative verbal memory scores (Cohen’s d range = .93 to 1.54).Conclusions: This is the first known study using RCIs developed on a large sample of children with epilepsy to assess clinically meaningful change in memory after TLR in children with pharmacoresistant seizures. While many children demonstrated little change in memory performance following TLR, there was a large subset of patients who showed postoperative declines in memory performance. Visual memory declines were observed following both right and left TLR, while verbal memory declines were clearly more common following left TLR. These results should provide an impetus to develop models to predict cognitive outcome following pediatric epilepsy surgery and be considered when counseling parents and older children about the potential risks of TLR.