Abstracts

Rationale: Drug-resistant epilepsy (DRE) is a major public health concern with up to a third of individuals with epilepsy considered medically refractory. Understanding the factors affecting the long-term seizure trajectory of individuals with DRE can help clinicians formulate rational treatment plans and provide prognostic information. Previous prognosis studies have focused in newly diagnosed1 or chronic/late-stage intractable epilepsy patients2 and none have utilized the recent ILAE definition of DRE (i.e., failure of 2nd AED due to inefficacy).3 It is unclear what factors affect better or worse seizure trajectories in individuals with DRE. We conducted a large retrospective natural history study of DRE. The objective is to identify which clinical factors predict better seizure trajectories.Methods: From an ongoing, large, retrospective, longitudinal observational study of AED response and tolerability, we identified incident drug-resistant epilepsy patients. Inclusion criteria included failure of the two AEDs due lack of seizure control (i.e., persistent seizures despite 2nd AED at Columbia, and initiation of the 3rd AED at Columbia). Observation time began at the time of start of their 3rd AED. If patients went on to receive epilepsy surgery after meeting the inclusion criteria, their observation data were censored on the date of epilepsy surgery. Outcome definition: Seizure remission is defined as absence of any seizure (including auras), lasting 12 months or longer. Data Analysis: Continuous demographic variables will be summarized with appropriate measures of central tendency and dispersion. Categorical variables will be summarized using frequency distributions. Using graphically-depicted seizure remission/relapse trajectory, patient s outcomes was rated by an expert epileptologist using a five-category ordinal scale. Using ordinal logistic regression, we will determine which clinical factors predict a better outcome (i.e., greater remission time). Results: Among 3419 patients in the ongoing study, 739 met our inclusion criteria (with demographics described in Table 1). The mean observation time in these individuals was 42.7 months (range 12-198 months). Of these 739 patients, 514 (70%) never achieved a 12 months seizure remission during their follow up, while 225 (30%) attained one or more 12 months seizure remission at some point during the observation period. Of these 225 individuals, 48 (21%) had multiple episodes of 12 months remission. The length of remission varied with 88 (39.1%) reaching between 12-24 months, 57 (25.3%) reaching between 24-36 months, and 80 (35.6%) reaching at least a 36 months. Conclusions: Although the majority of DRE patients (70%) never attain a period of sustained seizure remission from the time point that they are considered to have developed DRE, our findings demonstrate the fluctuating health trajectory for a third of the patients with DRE who experience multiple periods of seizure remission interrupted by seizure relapse.