Temporal lobe epilepsy surgery failures: predictors of seizure recurrence, yield of reevaluation, and outcome following reoperation

Abstract

Object The authors provide a systematic analysis of electroclinical characteristics in patients with persistent seizures following temporal lobe epilepsy (TLE) surgery and identify ideal candidates for reoperation. Methods The authors retrospectively reviewed the records of 68 adult patients (mean follow-up 8.7 years) who underwent a video electroencephalography evaluation and high-resolution imaging after failed TLE surgery performed between 1990 and 2004 at The Cleveland Clinic. Multivariate logistic regression analyses were performed to identify predictors of the yield of a repeat evaluation, location of the recurrence focus, and outcome following reoperation. Results Although a focus of recurrence was identified in 44 patients, only 15 underwent reoperation, and only 6 of these became seizure free. Localized foci of recurrence were successfully identified in patients with early (within 1 postoperative year) and frequent (≥ 4 per month) recurrent seizures (yield of 100% if both conditions were fulfilled). Predictors of contiguity of the focus of recurrence to the initial surgical bed were variable depending on the type of the initial surgery: patients with baseline contralateral temporal spiking were 6 times (OR 6.34, p < 0.05) more likely to experience seizure recurrence from the contralateral temporal lobe after a “standard” temporal lobectomy, while the need to use subdural electrodes and the timing of recurrence were more significant following limited temporal resections. The focus of recurrence was distant to the original surgical bed when subdural electrodes were used prior to first surgery (OR 28.0, p = 0.01) or when seizures recurred early (within < 6 postoperative months; OR 12.5, p = 0.04). With reoperation, only patients with mesial and basal extension of the temporal resections became seizure free. Interestingly, seizure freedom was achieved with medical therapy alone in 42% of patients with a nonidentifiable recurrence focus as opposed to 4% of those with an unoperated identifiable focus. Conclusions The timing and frequency of recurrent seizures following unsuccessful TLE surgery provide useful guidelines for the yield of a surgical reevaluation, and potentially for the mechanisms of surgical failure.