Abstract All patients over the age of 60 who experienced seizures between 1979–83, were registered. The number of deaths was registered until July 2, 1985. Included were 162 patients who received no anti-epileptic drugs prior to the study period; 87 patients had established epilepsy at the time of admission. The number of deaths among previously untreated patients significantly exceeded expectation. Mortality was not significantly correlated to severity of epilepsy. In patients with brain tumor all but one died within the first year. Mortality among patients with postapoplectic seizures was significantly higher than expected being especially during the first year. Numbers of deaths among patients with seizures of unknown cause did not differ from the expected, neither did causes of death. Numbers of deaths in patients with established epilepsy at the time of admission was significantly higher than expected although none had malignant tumours and only 4 had postapoplectic seizures thus illustrating the influence of selecting patients with chronic active epilepsy. Eleven patients died suddenly and unexpectedly of unknown cause, which was more than expected. These patients were found dead under circumstances compatible with death occurring during seizure. Epilepsy was mentioned on the death certificate in only one case, indicating that the frequency of sudden, unexpected death among epileptics could easily be underestimated.