Characteristics and predictive biomarkers of drug resistant epilepsy -- study in Georgia

Aim: The authors conducted a case-control study to estimate predictive factors for timely identification of patients at higher risk for developing drug resistant epilepsy.Methods: The retrospective case-control study was conducted among people diagnosed as having drug resistant epilepsy (cases) and their controls, identified as having drug-responsive seizures. All participants were admitted to the tertiary Epilepsy Center at the Institute of Neurology and Neuropsychology (Tbilisi, Georgia) during 2011. The data on demographic features and disease characteristics were analyzed. Multiple logistic regression analysis was used to identify independent risk factors for development of intractable epilepsy. Results: A total 334 patients were identified; 84 (34%) met the criteria for drug resistant epilepsy. One hundred and sixty-four age- and gender-matched controls with drug-responsive epilepsy were identified. Relative to the control group, the drug resistant seizure group had increased frequency of perinatal pathology (24% vs. 12%), febrile seizures (22% vs. 12%), seizure frequency at disease manifestation (62% vs. 19%), occurrence of convulsive seizures (84% vs. 70%), electroencephalo-graph (EEG) epileptiform discharges (94% vs. 77%), polytherapy (90% vs. 12%), multilobar lesions (30% vs. 16%), hippocampal sclerosis (18% vs. 5%), and malformations of cortical development (8% vs. 2%). Multivariate analysis indicated four factors with independent predictive value for development of intractable epilepsy: frequency of seizure, polymorphism of seizure, polytherapy, and epileptiform EEG abnormalities. The presence of all four factors in combination resulted in a 98% of probability of developing drug resistant epilepsy. Conclusion: Several factors appear to have prognostic value in identifying the risk for drug resistant epilepsy. These factors may prove useful in non-specialized health care settings for timely identification of individuals with elevated risk for drug resistant epilepsy. However, retrospective design and possible recall bias must be considered when interpreting or extrapolating these results