Clinical significance of subclinical seizures in epilepsy presurgical evaluation: A systematic review and meta-analysis.
TL;DR
Overall, the presence of SCSs alone was not associated with a poor outcome, but the spatial non-concordance with the seizure-onset zone was highly predictive for poor seizure outcome.
Abstract
Objective
Subclinical seizures (SCSs) remain an understudied aspect of presurgical evaluation in patients with drug-resistant epilepsy (DRE), with uncertain prevalence, distribution among epilepsy types, and predictive value for surgical outcomes. This systematic review and meta-analysis aims to explore the occurrence of SCSs, their concordance with the seizure-onset zone (SOZ), and their association with postoperative seizure outcomes.
Methods
A systematic search of Medline, Embase, and Web of Science was performed to identify studies reporting SCSs during scalp electroencephalography (EEG) monitoring or intracranial EEG (iEEG) monitoring in patients with DRE undergoing epilepsy surgery. Data on SCS prevalence, concordance with SOZ, and postoperative outcomes were extracted. Pooled proportions and odds ratios (ORs) were estimated using random-effects generalized linear mixed models. Subgroup analyses evaluated the proportions of SCSs in temporal vs extratemporal epilepsy.
Results
Thirteen studies comprising 1174 patients met the inclusion criteria. The pooled prevalence of SCSs was 32% (95% confidence interval [CI] 8-73) during scalp EEG monitoring and 50% (95% CI 38-62) during iEEG monitoring. Concordant SCSs with SOZ occurred in 37% of patients undergoing iEEG. Non-concordant SCSs were associated with a nearly fivefold higher odds of poor postoperative seizure control (OR 4.82, 95% CI 1.80-12.96; p = 0.002). SCSs were more frequent in temporal lobe epilepsy than in extratemporal epilepsy (57% vs 44%, p = 0.046). Within temporal lobe epilepsy, the highest prevalence was observed in the mesiotemporal region (69%). Overall, the presence of SCSs alone was not associated with a poor outcome, but the spatial non-concordance with the SOZ was highly predictive for poor seizure outcome.
Significance
SCSs are common in presurgical evaluation, particularly in temporal lobe epilepsy. Although their presence alone is not prognostic, non-concordant SCSs are associated with worse postoperative outcomes. Systematic assessment, especially using iEEG, may improve delineation of the epileptogenic zone and surgical planning in DRE.