To detect change in mental development or intelligence over two years following hemispherectomy in children with pharmacologically intractable epilepsy. Seventeen infants and preschoolers (median age at epilepsy onset of 0.0 years and at hemispherectomy 1.5 years; epilepsy duration of 0.2-2.6 years) and 12 older children (median age at onset of 1.0 year and at hemispherectomy 8.3 years; epilepsy duration of 1.1-11.7 years) with pharmacologically intractable seizures due to developmental, acquired or progressive pathology. Prospective study with consecutive inclusion of children, fixed assessment intervals (shortly before and 6, 12 and 24 months after hemispherectomy) and assessment using developmental scales and intelligence scales. Dependent variables included mental developmental index (MDI), mental age (MA) and mental developmental delay (MDD) in younger children and intelligence quotient (IQ) in older children. Mental development had arrested or deteriorated prior to hemispherectomy in 14 children (82%) assessed with developmental scales. In 14 children, it was not possible to more precisely determine MDI than "below the lowest MDI that the test manual provided" either before or after hemispherectomy. MA, however, increased in 16 children. Overall, IQ changed negligibly over two years after hemispherectomy, although an individual approach revealed variability. Children with Rasmussen encephalitis did not recover from the significant presurgical deterioration of intelligence. Removal of the affected hemisphere enables epileptic children, even those with severe mental delay, to further develop mentally.