Seizure latency and epilepsy localization as predictors of recurrence following epilepsy surgery

The primary purpose is to determine whether the time between epilepsy surgery and first seizure recurrence can estimate the timing of the next seizure event for temporal and extratemporal epilepsy. A secondary endpoint aimed to compare temporal and extratemporal epilepsy surgery and examine which subgroup has a higher hazard of subsequent seizure recurrence.

Data used were from a retrospective database at Thomas Jefferson University Hospital. Records were stratified into temporal (n = 943) and extratemporal (n = 125) surgeries. Analyses were done using SAS and utilized Cox proportional hazards models while controlling for demographics and clinical factors. The primary predictor of time between surgery and first recurrence was treated as a nominal variable binned into six segments, whereas secondary endpoints used a categorical predictor of epilepsy location while controlling for seizure latency.

Generally, as seizure latency following surgery increased, the time between first seizure and second seizure increased. These results were statistically meaningful in the temporal set (Wald chi-squared = 40.4715, df = 5, p < .0001). Outcomes could also be interpreted based on predictor group; for instance, if Seizure 1 occurred 1-2 months following surgery in the temporal set, the median number of days until the next seizure was 35.5 days (95% confidence interval [CI] = 21-89 days). Secondary analysis showed that temporal lobe epilepsy had a lower hazard of a second seizure than extratemporal lobe epilepsy (89.2% reduction in hazard; 95% CI = .015-.795).

This analysis provides a framework to use initial seizure latency to predict the median number of days until the next seizure event, while stratifying based on epilepsy location and controlling for multiple variables. It also suggests that the hazard of seizure recurrence in temporal lobe epilepsy is lower than in extratemporal lobe epilepsy.