The chronic effects of cigarette smoking upon regional cerebral blood flow (rCBF) are unclear. The present study investigated the effects of smoking on rCBF assessed by intravenous 133Xe injection in 40 asymptomatic individuals. Analysis was performed using linear regression analysis and multiple regression analysis, creating a model of the rCBF as a function of cerebrovascular risk factor. The risk factors included smoking status, smoking index, age, gender, mean arterial blood pressure, total cholesterol, fasting blood glucose, hematocrit and STT change and left ventricular hypertrophy in electrocardiogram. The male-to-female ratio was higher in the smoking group (18/2) than that in the non-smoking group (2/18). Among the smokers, mean hematocrit was significantly higher than in the non-smokers (p < 0.01). There was no significant difference in rCBF and other variables between the two groups. Linear regression analysis revealed significantly negative correlations between smoking index and CBF in the whole brain (r = -0.33; p < 0.05), right hemisphere (r = -0.34; p < 0.05), right parietal cortex (r = -0.36; p < 0.05), right occipital cortex (r = -0.34; p < 0.05) and left parietal cortex (r = -0.33; p < 0.05). In other variables, age, male sex, hematocrit, left ventricular hypertrophy and STT change showed significantly negative correlations with rCBF. To reduce the effect of confounding variables in our assessment of the dose-related response, a multivariate regression analysis was carried out treating rCBF as a dependent variable and risk factors as independent variables. In the final model, age, hematocrit, male sex and presence of BCG changes remained as independent negative predictors for rCBF.(ABSTRACT TRUNCATED AT 250 WORDS)