Abstracts

Rationale: Few studies have evaluated postoperative changes in various language domains in Temporal Lobe Epilepsy associated with Hippocampal Sclerosis (TLE-HS) [1]. We evaluated one year postoperative language changes in 25 unilateral left (UL) TLE-HS patients after Anterior Temporal Lobectomy (ATL) in semantic (SF) and phonemic fluency (PF), visual confrontation naming for objects (VN-O), verbs (VN-V) and proper nouns (VN-PN), responsive naming (RN) and word comprehension (WC), as well as clinical features associated with postoperative language decline. Methods: 28 UL-TLE-HS patients undergoing pre-surgical evaluation (moment A) underwent language assessment. 25 (89.3%) patients were evaluated one year after ATL (moment B) in those domains, using different VN, RN and WC items to control for learning. Healthy controls were assessed for language performance: 21 at A (CA) and 20 at B (CB). Subjects were right-handed, aged 18-55 years, had 8 or more education years, measured (patients) or estimated (controls) IQ>70, without comorbidities that could interfere with cognition. Patients, CA and CB did not differ in gender, age or education. CA and CB did not differ in any language test. We compared language scores between patients and CA, and between patients and CB with t tests. In domains where patients performed worse than controls, we used Receiver Operating Characteristic (ROC) curve, adjusting for education, to yield cut-off scores with best sensitivity and specificity (Youden's J statistic), as well as the Area Under the Curve (AUC) (Table 1). We then performed paired t tests to determine which language functions showed improvement or decline among patients (Table 2). We evaluated clinical features associated with domains showing a decline in language function: if no different than controls, performance was coded as 0; if different, performance was coded as 0 if above or as -1 if below or equal the cut-off score. Change in function was calculated as the difference between B and A (-1 or 0), followed by chi2 and t tests (Stata v.13). Results: Preoperatively, UL-TLE-HS patients performed worse than controls in SF, PF, VN-V and VN-PN. On one-year postoperative follow-up, patients showed decreased WC and improved VN-PN. We found a significant association between decreased WC and age at epilepsy onset (AEO) (p=.03). AEO for patients that declined decline was 13.4 years vs 8.3 years for those that did not decline. We found no association between decreased WC and demographic, other disease and treatment related factors. Conclusions: Decline in word comprehension has not been previously appreciated after left ATL, and is consistent with the recently recognized role of the dominant anterior temporal lobe in comprehension. Younger onset epilepsy may protect patients from postoperative WC decline probably due to language network reorganization. Reference: [1] M. Schwarz, E. Pauli, Postoperative speech processing in temporal lobe epilepsy: Functional relationship between object naming, semantics and phonology, Epilepsy & Behavior 16(4) (2009) 629–633. Funding: São Paulo Research Foundation (FAPESP - Fundação de Amparo à Pesquisa do Estado de São Paulo), CInAPCe (Cooperação Interinstitucional de Apoio a Pesquisas sobre o Cérebro) project 2005/56464-9.