Abstracts

Rationale: Describe pos surgical EEG findings and try to establish its prognostic value with regard to epilepsy surgery outcome. Methods: We included patients suffering from pharmaco resistant epilepsy who had undergone epilepsy surgery between April 2005 and July 2010 and completed 1 year of postoperative follow-up. Epilepsy surgery subtypes were classified into: Temporal Resections (TR): standard anterior temporal lobectomies (ATL) and ATL associated with adjacent neocortical resections; Neocortical Resections (CORT), hemispherectomy and hemispherotomy (HEM). We classified outcome according to Engel Score: Favorable outcome (Engel I), unfavorable outcome (Engel>I). EEG findings were further classified in two groups: Non epileptiform EEG (NE-EEG) and epileptiform EEG (E-EEG). We calculated the sensitivity, specificity, predictive values and accuracy for the presence of IED with respect to seizure outcome (Engel I vs Engel>I). We used Fisher exact tests to compare the patients with favorable and unfavorable seizure outcomes with respect to the presence of E-EEG or NE-EEG in the different resection subtypes; a p value of<0.05 was considered significant. Results: Fourty four patients were included, median time of follow up was 37 months (12-75). Twenty five patients had favorable outcome: 13 had undergone TR, of them 12 (92%) had NE-EEG, and 1 E-EEG; 12 had undergone CORT, 6 (50%) had E-EEG. The difference between EEG findings in both types of resection was statistically significant, (p<0,01). Thirteen patients had unfavorable outcome: in 3 TR was performed, 1 patient had E-EEG and 2 NE-EEG; 10 had undergone CORT, 7 of them had E-EEG (70%), and 3 NE-EEG. In 6 patients HEM was performed, 5 had favorable outcome, 4 of whom had E-EEG. In TR post-surgical EEG sensitivity is 33%, specificity: 92%, negative predictive value: 85%, positive predictive value: 50% and accuracy: 81%. In CORT, post-surgical EEG sensitivity was: 70%, specificity: 50%, negative predictive value: 66%, positive predictive value: 53% and accuracy: 59%. Conclusions: A NE-EEG was statistically significant associated with favorable outcome in patients who had undergone TR. This finding was also supported by post surgical EEG high specificity and negative predictive value in this type of resections. We observed a trend toward more frequent E-EEG in patients with unfavorable outcome in whom CORT was performed. No relationship was found between postsurgical EEG and outcome in patients who had undergone HEM.