Background Information on the effects of early treatment of seizures on mortality is scarce. The authors assessed the survival of patients with a first generalised tonic–clonic seizure, randomised to immediate treatment (treated) versus treatment only in the event of seizure recurrence (untreated), over a 20-year period.

Methods The authors followed 419 patients. The median follow-up was 19.7 years (range 0.2–21.5) for a total of 7867 person-years.

Results 40 persons (9.6%) died during follow-up, 19 (8.9%) treated and 21 (10.3.%) untreated. The probability of surviving was 100% at 1 year, 97% (95% CI 95% to 99%) at 5 years, 94% (91–97) at 10 years and 91% (87–95) at 20 years in treated patients and 100%, 98% (95–100), 97% (94–99) and 89% (85–94), respectively, in untreated patients (p=0.7). After adjustment for treatment of first seizure and putative risk factors (gender, age, seizure type, previous uncertain seizures, family history of seizures, pre-, peri- and postnatal risk factors, remote aetiological factors for epilepsy, abnormal neurological examination, CT or MRI abnormalities, EEG abnormalities and acute treatment), only the presence of aetiological factors for epilepsy predicted a higher mortality (HR 3.4, 95% CI 2.5 to 4.3%; p<0.01). Patients with remote aetiological factors and who did not achieve 5-year remission had the poorest survival.

Conclusion Starting antiepileptic treatment immediately after the first generalised tonic–clonic seizure or only after seizure recurrence did not affect survival over the following 20 years.