The aim of this study was to evaluate the prevalence and prognostic significance of epilepsy in 1028 patients diagnosed in the computer tomography (CT) era with histological low- or high-grade intracranial gliomas. Survival analysis included Kaplan-Meier plots, log-rank tests, logistic regression and Cox's analysis as implemented in the SPSS statistical package. Epilepsy was a positive univariate (P < 0.0001) and multivariate, (P < 0.03) prognostic factor for survival in the total patient group (n = 1028, relative risk of death 0.83, 95% confidence interval (CI) 0.70-0.98) as well as in the high-grade patient group (n = 649, relative risk of death 0.80, 95% CI 0.66-0.96), but not in the group of low-grade glioma patients (P > 0.2). The prevalence of epilepsy in glioblastoma patients was 251/512 (49%), 95/137 (69%) in anaplastic gliomas, and 322/379 (85%) in patients with low-grade gliomas, with 97 of the 102 T1 low-grade subgroup (95%) having epilepsy, indicating that the presence of epilepsy may select patients for early radiological diagnosis. The frequency of epilepsy at presentation decreased with age in high-grade glioma patients, and increased with age in low-grade glioma patients to a plateau in the fourth decade of life (P < 0.01). The prevalence of epilepsy in patients with histological intracranial gliomas varied with patient age and tumour histology, with low-grade patients having the highest prevalence. Epilepsy was a significant positive prognostic factor except in patients with low-grade gliomas, and may select low-grade patients for early diagnosis.