Summary: Purpose:We studied overall and cause-specific mortality rates in a large cohort of patients with epilepsy compared with mortality rates of the general population in the same geographic area.Methods: The cohort consisted of all patients (N = 9,061) aged >15 years admitted with a diagnosis of epilepsy for inpatient care in Stockholm during the years 1980-1989. All patients were followed in the National Cause-of-Death Register, from which the causes of death were obtained, until December 31, 1992. Thus, 53,520 person-years were observed. Mortality rates were compared with those of the general population of Stockholm.Results: We observed 4,001 deaths in the cohort, compared with an expected number of 1,109 deaths in the general population. This yielded a standardized mortality ratio (SMR) of 3.6[95% confidence interval (CI) 3.5-3.71. Although highest in the younger patients, the SMR was significantly increased in all age groups. The excess mortality rate in the cohort was due to a wide range of causes of death, including malignant neoplasms [SMR 2.6 (2.4-2. Conclusions: Our results demonstrate that this large subgroup of patients with a diagnosis of epilepsy, once hospitalized and discharged, is a population at risk, with an excess mortality rate due to several different causes. Key Words: Epilepsy-Mortality-Standardized mortality ratio-Epidemiology-Cause of death.Patients with epilepsy have consistently been found to have a higher death rate than the general population (1-5). Although overall mortality rate has been analyzed in several studies, few researchers have investigated causespecific mortality rates in patients with epilepsy.In previous studies based on patients in institutions (6,7) and referral centers (3,s) researchers have analyzed a highly selected population with severe epilepsy. Studies (9,lO) based on cause-of-death registers have comprised large cohorts of patients. However, the validity of the diagnosis and the patients selection are uncertain in these studies because epilepsy is seldom entered as a diagnosis on death certificates. On the other hand, population-based studies, which provide valid information representative of the epilepsy population in general, have, on the whole, comprised a comparatively small number of patient-years and deaths (1,2,4). Therefore,