Seizure outcome and its predictors after temporal lobe epilepsy surgery in patients with normal MRI

Abstract

To characterize seizure outcomes following temporal lobe epilepsy (TLE) surgery in patients with normal preoperative brain magnetic resonance imaging (MRI). We reviewed adult patients with pharmacoresistant epilepsy and normal MRI who underwent TLE surgery (1996-2009). Seizure outcomes were analyzed using survival and multivariate regression with Cox proportional hazard modeling. Two analyses were performed using two favorable outcome definitions: complete seizure freedom and Engel classification. Sixty-four patients were analyzed (mean follow-up 4.1 years; range 1-14.5 years). Most had a standard anterior temporal lobectomy (84%) and unremarkable pathology (45%). At 1 year, the chance of complete seizure freedom was 76% [95% confidence interval (CI) 71-81%] comparable to an 81% (95% CI 76-86%) chance of Engel score of 1. With longer follow-up, a progressively broadening significant discrepancy between the two outcome measures was observed. The chance of complete seizure freedom was 66% (95% CI 61-71%) at 2 years, and 47% (95% CI 40-54%) at 7 years and beyond, whereas the respective chances of achieving an Engel 1 classification were 76% (95% CI 70-82%), and 69% (95% CI 63-75%) at similar time points. Seizure outcome as defined by either measure was worse in patients with higher baseline seizure frequency (adjusted risk-ratio 2.7 when >12 seizures/month; p = 0.01) and with preoperative generalized tonic-clonic seizures (adjusted risk ratio 10.8; p = 0.0006). Memory measures declined with dominant hippocampus resections. A normal MRI should not prevent presurgical evaluations in patients with suspected TLE, as favorable long-term postoperative seizure outcomes are possible. Proposed mechanisms of epileptogenicity and seizure recurrence in this group are discussed.