Objective Hemispheric neurosurgery is an established treatment for severe epilepsy caused by extended unilateral brain pathology. However, it is still an unresolved question at which age surgery should best be performed. In light of decreasing plasticity and the cumulative impact of seizures and anticonvulsants on neurodevelopment, early surgery appears preferable.

Methods We retrospectively investigated the medical, cognitive-behavioural and psychosocial long-term outcome (follow-up: 9.4 years (1.1–19.4)) of hemispherectomy as a function of age at surgery (early: <7 years/intermediate: 7–16 years/late: >16 years) based on a structured postal survey in a large patient sample (N=61/81, return rate: 75%).

Results At follow-up, 45 (74%) patients were seizure free. Presurgical levels of intelligence were below average in most patients (79%) and postsurgical cognition largely resembled the presurgical capacities. Best seizure outcome was obtained for early surgery patients (90% seizure free). Patients with late surgery, however, exhibited higher presurgical and postsurgical intelligence and better psychosocial achievements. Binary logistic regression identified better presurgical intelligence and higher age at surgery as positive predictors of postsurgical intelligence. Lower presurgical intelligence and postsurgical seizure freedom predicted intellectual pre-post improvements.

Conclusions Our data confirm the efficacy and cognitive safety of hemispheric surgeries performed at different ages. Eligible older and high functioning patients can be perfect candidates. Presurgical intelligence serves as indicator of the functional integrity of the contralateral hemisphere, which mainly determines postsurgical cognition and psychosocial outcome. Seizure freedom promotes cognitive improvement. As many unknown factors determined age at surgery, our retrospective data neither question early surgeries nor suggest postponing surgery.