A cohort analytic study of a broad set of outcomes was performed in a consecutive series of 248 adults and adolescents who underwent evaluation for surgery for intractable epilepsy at the University of California, Los Angeles (UCLA), U.S.A. from 1974 to 1990. In the analysis, 202 were designated as "surgery" patients and 46 were assigned to a "nonsurgery" group, based primarily on whether surgery for a localized surgical region was undertaken. Follow-up at an average of 6 years revealed that 14 (7%) of the surgery and 9 (20%) of the non- surgery group had died (p < 0.01). Results were similar after adjusting for baseline differences between groups. These multivariate models also showed that female gender was significantly associated with lower mortality. A significantly higher proportion of the deceased experienced two or more seizures over the latest year of follow-up compared with survivors (p < 0.01). Specific causes of death remain to be determined.