The HHP began in 1965 as a longitudinal study of cardiovascular disease and stroke among 8006 men of Japanese ancestry, 45 to 68 years of age, living on the island of Oahu. Investigation of neurodegenerative disease in surviving participants began in 1991 with the HAAS. Since 1965 the cohort has been followed through 6 standardized examinations with follow up on all hospitalizations and mortality through surveillance of hospital and death records. Medical histories, physical measurements, laboratory values, demographic, socioeconomic, and lifestyle data collected at the first three examinations, 1965 through 1974, were used for this study. Measurement of risk factors before onset of clinical disease provides risk estimates based on exposure status. Between 1965 and 1998, 9 participants, 58 to 80 years old, developed GBM, an incidence of 6.21100,000 person-years. All cases were histologically confirmed. Analysis of a large number of potential risk factors showed GBM cases were associated with increased dietary glucose and carbohydrate levels, low alcohol consumption including wine, and solvent exposure, especially carbon tetrachloride. Incidence rates tended to be higher with increasing adiposity and serum cholesterol levels. No associations were observed with blood pressure, heart rate, pulmonary function, Japanese diet, recurrent herpes labialis, medication use for diabetes or hypertension, smoking, and pesticide or metal exposure. The previous reports of GBM risk with chlorinated aliphatic hydrocarbon exposure are supported. Dietary glucose and carbohydrate risk may be related to acrylamide formation during food preparation. The database of a non-neoplastic epidemiological study may be used to identify potential GBM risk factors.