We conducted a case-control study to evaluate the preclinical association between epilepsy, diabetes, and stroke and primary adult brain tumors. We first identified all 1,501 low-grade glioma, 4,587 high-grade glioma (HGG), and 4,193 meningioma cases reported to the Swedish Cancer Registry from 1987 to 1999. Next, controls (137,485) were randomly selected from the continuously updated Swedish Population Registry and matched to cases diagnosed that year on age and sex. Finally, cases and controls were linked to the Swedish Hospital Discharge Registry (1969-1999). We found that > or =8 years before HGG diagnosis (or control reference year) there was an elevated risk of HGG among people discharged with epilepsy [odds ratio (OR), 3.01; 95% confidence interval (95% CI), 1.73-5.22]. Two to 3 years before HGG diagnosis, this risk increased (OR, 5.33; 95% CI, 3.58-7.93) and was especially strong among people ages <55 years (OR, 13.49; 95% CI, 6.99-25.94). During this 2- to 3-year prediagnostic period, we also found an increased risk of HGG among people discharged with meningitis (OR, 3.02; 95% CI, 1.06-8.59) or viral encephalitis (OR, 12.64; 95% CI, 2.24-71.24). Results are similar for glioblastoma multiforme, low-grade glioma, and meningioma. In contrast, risk of HGG among people discharged with diabetes or stroke does not increase until year of brain tumor diagnosis. The occurrence of excess epilepsy > or =8 years before HGG diagnosis suggests a relatively long preclinical phase, but excess diabetes or stroke appear late in HGG development.