Corpus callosotomy in Lennox-Gastaut syndrome: Effectiveness, safety, and predictors of outcome: A systematic review and meta-analysis.

BACKGROUND: Lennox-Gastaut syndrome (LGS) is a severe, drug-resistant developmental and epileptic encephalopathy for which corpus callosotomy (CC) can be a particularly effective option for seizure control. However, this benefit must be weighed against the tradeoff of permanent corpus callosum disconnection. Although use of CC declined with the rise of neuromodulation, interest has resurged with minimally invasive techniques, including laser interstitial thermal therapy and endoscope-assisted approaches. However, updated LGS-specific outcomes following CC have not been systematically synthesized across surgical approaches and age groups. METHODS: We conducted a systematic review and meta-analysis (PROSPERO: CRD420261381787) adhering PRISMA 2020 guidelines. PubMed, EMBASE, OVID/MEDLINE, and CENTRAL were searched from inception to March 2026. The primary outcome was ≥ 50% seizure reduction. Secondary outcomes included ≥ 75% reduction, seizure freedom, complication rates, and non-seizure outcomes. Pre-specified subgroup analyses evaluated extent of disconnection and approach (open vs laser vs endoscopic). Meta-regression assessed age at surgery, age at seizure onset, follow-up duration, extent and approach of surgery. We also performed sensitivity analyses excluding high-bias or influential studies, assessed publication bias, and graded certainty of evidence. RESULTS: After screening 1235 articles, 48 (822 patients; 46 retrospective, 2 prospective; no randomized trials) studies from 19 countries were included. The pooled ≥ 50% responder rate was 69.6% (95% CI: 63.1-75.5%; I² = 24.1%; k = 48). The ≥ 75% responder rate was 38.2% (95% CI: 29.8-47.4%; I² = 62.9%; k = 47), and seizure freedom was 20.4% (95% CI: 16.0-25.7%; I² = 0.6%; k = 20). No significant differences were observed by extent or approach. In meta-regression, older age at surgery was the only significant independent predictor of ≥ 50% response in the multivariable model (β = 0.045; p = 0.020). Non-seizure outcomes were limited and qualitatively synthesized. The pooled proportion of complications was 15.8% (95% CI, 10.4% - 23.2%). No publication bias was detected, though all outcomes carried very low certainty ratings. SIGNIFICANCE: CC achieves ≥ 50% seizure reduction in ~70% of LGS patients. Outcomes were not influenced by surgical extent or approach. Evidence is limited to observational studies with very low certainty, highlighting the need for prospective, standardized, multicenter data.