Mortality rates are increased among people with epilepsy, and may be highest in those with uncontrolled seizures. Because epilepsy surgery eliminates seizures in some people, we used an epilepsy surgery population to examine how seizure control influences mortality. We tested the hypothesis that patients with complete seizure relief after surgery would have a lower mortality rate than those who had persistent seizures. Three hundred ninety-three patients who had epilepsy surgery between January 1986 and January 1996 were followed after surgery to assess long-term survival; 347 had focal resection or transection, and 46 had anterior or complete corpus callosotomy. A multivariate survival analysis was performed, contrasting survival in those who had seizure recurrence with survival of those who remained seizure free. Standardized mortality ratios and 95% confidence intervals were calculated. Overall, seizure-free patients had a lower mortality rate than those with persistent seizures. This was true for the subset of patients with localized resection or multiple subpial transection. No patients died among 199 with no seizure recurrence, whereas of 194 patients with seizure recurrence, 11 died. Six of the deaths were sudden and unexplained. Most patients who died had a substantial reduction in postoperative seizure frequency. The standardized mortality ratio for patients with recurrent seizures was 4.69, and the risk of death in these patients was 1.37 in 100 person-years, whereas among patients who became seizure free, there was no difference in mortality rate compared with the age- and sex-matched population of the United States. Elimination of seizures after surgery reduces mortality rates in people with epilepsy to a level indistinguishable from that of the general population, whereas patients with recurrent seizures continue to suffer from high mortality rates. This suggests that uncontrolled seizures are a major risk factor for excess mortality in epilepsy. Achieving complete seizure control with epilepsy surgery in refractory patients reduces the risk of death, so the long-term risk of continuing medical treatment appears to be higher than the risk of epilepsy surgery in suitable candidates.