How predictable is cognitive performance in Brazilian patients with pharmacoresistant mesial temporal lobe epilepsy?

The objective of the study was to investigate the independent association between clinical, demographic, psychiatric, radiologic, electrophysiological, and pharmacologic variables and cognitive performance of Brazilian patients with pharmacoresistant mesial temporal lobe epilepsy (MTLE). Ninety-three patients with pharmacoresistant MTLE related to hippocampal sclerosis (HS) were included in the study. Multiple linear regressions were done to identify predictor variables for 24 cognitive tests. Independent variables analyzed were sex, hand dominance, age, years of education, marital status, work activity, history for an initial precipitant injury (IPI), family history of epilepsy, lesion side, antiseizure medication (ASM) treatment type, ASM serum levels, benzodiazepine (BDZ) treatment, age at epilepsy onset, disease duration, monthly frequency of seizures, and Hospital Anxiety and Depression Scale (HADS) scores. Years of education was an independent and positive predictor in 22 of the 24 cognitive tests evaluated. Male sex was also a positive predictor of one cognitive test. Variables negatively associated with cognitive performance were left side lesion (10 tests), disease duration (5 tests), polytherapy (3 tests), ASM serum levels (3 tests), and BDZ treatment or not working (1 test each). The regression model explained between 6% and 44% of the cognitive test scores variation. In Brazilian patients with pharmacoresistant MTLE-HS, up to 44% of cognitive test scores variation is predictable by clinical, demographic, psychiatric, radiologic, electrophysiology, and pharmacological variables. The identification of predictors of cognitive performance may be helpful for better planning of patient care.

Informations de copyright

Copyright © 2020 Elsevier Inc. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of competing interest The authors have no conflict of interest, source of funding, or financial ties to disclose, and no current or past relationship with companies or manufacturers who could benefit from the results of the present study.