Abstracts

Rationale: The proportion of patients enjoying long-term post-operative seizure freedom after TLE surgery remains suboptimal. Better outcomes are reported in children, possibly owing to different etiologies and more extended resections. Individual prediction of seizure relapse is hampered by the weak odds ratios associated with known predictors of TLE surgery. These issues highlight the need for multicenter studies using robust statistical methods. We then conducted a retrospective multicenter study involving eight Italian epilepsy surgery centers in order to compare seizure outcome after TLE surgery and their predictors in children (< 16 years old) versus adults (>16 years old).  Methods: We asked all participating centers to fill in a database including information on general characteristics, EEG, neuroimaging, and surgery. In order to characterize the sample, we analyzed all collected data using descriptive statistics. We focused on hypothesis testing (equality test for means and proportion) in order to assess possible differences between the two groups (children and adults) for all outcomes and variables of interest. To take in account the hierarchical nature of the data (patients belonging to the same center can share common characteristics) and to assess the presence of potential confounders and identify possible interactions, we carried out multilevel multivariate regression analyses, where the level will be defined by the center to which data belong. Results: We included 114 children (51 F) and 397 adults (197 F) which significantly differed for age at seizure onset, frequency of seizures and status, pre- and post-operative EEG, type of intervention, complications, histology, completeness of resection, AED withdrawal. The duration of the FU ranged from 24 to 288 months. 351 patients (68.7%, 262 adults) were completely seizure free (Engel's class Ia ) 2 years after surgery and 319 (62.9%, 235 adults ) at last FU. In the whole sample, predictors of class Ia were abnormal post-operative EEG (P: 0.010), completeness of resection  (p: 0.008), age at surgery <16 years old (p:0.001)at 2 years FU and completeness of resection (p:0.002), abnormal postoperative EEG (p:0.007), type of surgery (lesionectomy, p:0.005 and anterior temporal lobectomy, p: 0.011) and age at surgery <16 years old (p:0.012) at last FU. In children, predictors of Engel’s class Ia were completeness of resection  (p: 0.0000) and preoperative executive functions scores  (p: 0.020) at 2 years FU and frequency of status (p=0.005) and completeness of resection  (p: 0.000) at last FU. In adults, predictors for Engel’s class Ia were abnormal post-operative EEG (worsened p: 0.009 or improved p: 0.004) at 2 years FU and type of surgery (lesionectomy, p:0.0033 and anterior temporal lobectomy, p: 0.013) and abnormal postoperative EEG (worsened  p:0.018 or improved p:0.002) at last FU.  Conclusions: Conclusions: This study contributes to elucidate differences in predictors of surgical outcome in children versus adults, thereby highlighting the relevance of the completeness of resection and the need for early surgery. Postoperative EEG should be taken in account as prognostic indicator of unfavorable outcome in the medication management after epilepsy surgery.  Funding: Epilepsy LICE Foundation