Epilepsy surgery outcomes in children with tuberous sclerosis complex: a systematic review and meta-analysis

Patients with tuberous sclerosis complex (TSC), while considered genetically homogeneous, are clinically heterogeneous and present unique challenges for epilepsy surgery evaluation and treatment. Thus, the authors' goal was to identify factors associated with seizure outcomes in children with TSC who had undergone epilepsy surgery.

The PubMed, MEDLINE, Embase, CINAHL, and Web of Science databases were queried for relevant articles. Patients from the TSC Alliance registry and Children's of Alabama were also included. Eligible studies were those reporting individual participant data on the seizure outcomes of pediatric patients with TSC who had undergone epilepsy surgery. Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed, and a meta-analysis was performed using a random-effects logistic regression model. Individual factors and seizure outcome following epilepsy surgery at the latest reported follow-up were tested for association. The primary outcome was a good seizure outcome, defined as Engel class I or II, International League Against Epilepsy class 1, 2, or 3, or ≤ 3 seizures per year following surgery.

From 44 studies eligible for systematic review and 2 additional sources, there were 2058 patients. Of these, 1338 (65%) patients had available surgical outcome data and a median follow-up of 2.5 years (IQR 1.0-5.0 years). Overall, 878 (66%) patients had a good seizure outcome. Such an outcome was associated with ≤ 1 visible tuber on MRI (OR 4.8, p = 0.01, 95% CI 1.44-15.94, I2 = 0%); however, a good seizure outcome was less likely in patients with no or mild cognitive delay (OR 0.52, p = 0.02, 95% CI 0.30-0.09, I2 = 11%) or in those with no or unifocal interictal scalp EEG abnormality (OR 0.36, p = 0.01, 95% CI 0.16-0.80, I2 = 24%). Preoperative IQ (treated as a continuous variable, OR 0.98, p = 0.009, 95% CI 0.96-0.99, I2 = 0%) had little to no effect on seizure outcome. Two (5%) studies met the criteria for a low risk of bias, 28 (64%) met the criteria for a moderate risk of bias, and 14 (32%) met the criteria for a serious risk of bias.

The authors identified several factors associated with seizure outcomes in TSC patients who had undergone epilepsy surgery. However, the study findings should be interpreted with caution, as they represent an aggregation of largely retrospective cohort or case studies with a high potential for bias. Systematic review registration no.: CRD42023393588 (www.crd.york.ac.uk/prospero/).