PURPOSE: Epilepsy surgery is performed based on the assumption that medical refractory epilepsy will continue. Rarely seizure freedom occurs before surgery is performed, while the patient is being evaluated as an epilepsy surgery candidate. The aim of this study was to describe the number of children withdrawn from an epilepsy surgery programme due to unexpected seizure improvement. METHODS: We retrospectively studied 173 children under 18 years with medical refractory epilepsy referred for epilepsy surgery between 1996 and 2010. Medical records were reviewed in 2012 and 2015. RESULTS: At the first evaluation point in 2012, 13 patients were withdrawn from the epilepsy surgery programme due to unexpected marked improvement. In 2015, 6 of them were still seizure free. They had unexpected seizure freedom due to change in AED treatment (n=3) or after a febrile episode (n=3). The mean number of years they had had seizures was 3.4 years (range 0.6-6.2 years) and the number of seizures at inclusion was 209 per month (range 6-750 per month). The duration of follow-up was 6.6 years after inclusion into the epilepsy surgery programme (range 4.0-13.0 years). The aetiology of the epilepsy for these patients was heterotopia (n=1), focal cortical dysplasia (n=3), infarction (n=1) and unknown, with normal MRI (n=1). They all had an IQ in the normal range. Two of the remaining 7 children were operated later. CONCLUSION: Unexpected seizure control may occur during epilepsy surgery evaluation.