Abstracts

Rationale: The development of high resolution of neuroimaging and improvement in classification and treatment of epilepsies most likely have an impact on the relationship between aetiologies and outcomes as compared to the last decades. In this study, we aimed to re-address some prognostic factors of seizure outcome in epilepsy overall. Methods: In this cohort of 1,177 consecutive patients from a tertiary hospital, we investigated prognostic factors for recurrent seizures including the impact of high-resolution MRI, age, seizure onset and definition of epilepsy etiology. Between April 2,013 to September 2014, we evaluated 1,177 consecutive patients from our outpatient epilepsy clinics (640 women; 893 with recurrent seizures (Ref-Sz), 284 seizure-free (at least 1 year free of seizures, Sz-free). The median age was 43 years (range 13-87), and the age of onset was 10 years (0.1-77). High-resolution epilepsy MRI protocol was performed for 1,131 patients and visually analyzed by imaging epilepsy experts. Statistical analyses included non-parametric Mann-Whitney tests, Chi-square, and logistic regression with seizure outcome including age and etiology as predictor variables. Etiology was divided into MTLE-HS, tumor, cortical malformation, IGE, gliosis, unknown, and others. Results: While age did not differ between groups (p=0.5), Ref-Sz group had earlier seizure onset (median 9 years versus 12 years in Sz-free; p<0.001). MRI analysis revealed an epileptogenic lesion in 806/1,131 (71.3%; 864 Ref-Sz/267 Sz-Free). Ref-Sz occurred in 678/806 (84%) of MRI-positive and in 186/325 (57%) of MRI-negative, with an OR of 4 for the association between Ref-Sz and MRI positive (?2 1df =93; p<0.0001, CI95% 3-5.3). The logistic regression showed an association between seizure outcome, age, and etiology (omnibus chi-square=178, df=15, p<0.0001). In comparison to IGE (30% with Ref-Sz) (Figure 1), the OR for developing Rf-Sz was of 18 for MTLE-HS (87% Ref-Sz), 30 for cortical malformations (92% Ref-Sz) 15 for tumours (85% Ref-Sz), 5 for unknown (64% Ref-Sz), 6 for gliosis (66% Ref-Sz) and 7 for “other” aetiologies (72% Ref-Sz) (all p< 0.001). Patients older than 55 years were more likely to present controlled seizures, except MTLE-HS, which presented with more Ref-Sz regardless of the age. Conclusions: In the light of modern MRI, our results show that age of onset, MRI, and etiology are important prognostic factors. The fact that the presence of a positive MRI is a prognostic factor highlights the need for specific epilepsy MRI protocols when evaluating patients with epilepsy. The analysis based on etiology showed that the presence of brain malformation had the highest OR for having uncontrolled seizures and IGE the lowest. It also highlights the importance of age defining epilepsy prognosis. Funding: São Paulo Research Foundation - FAPESP, Grant # 2013/07559-3